• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Food

  • Print
  • Share
  • E-mail

External Peer Review of the FDA/CFSAN Draft Report Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans (October 3, 2008) and Update (November 28, 2008)

July 2009

Introduction

On December 3, 2008, FDA contracted Versar, Inc., to facilitate an external peer review of its Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans (S/RA) and the November 28, 2008 update to the S/RA. The peer reviewers were independently selected by Versar, Inc. The peer reviewers were:

Susan S. Baker, MD, Ph.D.
Digestive Diseases and Nutrition Center
Children's Hospital of Buffalo

Cathy A. Brown
Athens Veterinary Diagnostic Laboratory
College of Veterinary Medicine
University of Georgia

James V. Bruckner, Ph.D.
College of Pharmacy
University of Georgia

Frederick J. Kaskel, MD, Ph.D.
Department of Pediatrics
Montefiore Medical Center of AECOM

Perry Martos, Ph.D., CIH
Department of Chemistry and Biochemistry
University of Guelph

We greatly appreciate the peer reviewers' comments and suggestions, as well as their willingness to provide them on a very tight deadline. The goal of this peer review was to obtain a thorough and meaningful assessment of the S/RA and update and to obtain input on the reasonableness of judgments made from the scientific evidence. The result should be an independent determination by each peer reviewer as to the appropriateness of (a) the assumptions made and hypotheses postulated, (b) the methodology utilized, (c) the quality and relevance of the data and information, (d) the accuracy of the analytic results, and (e) whether the conclusions reached are supported. The reviewers' responses to the specific charge questions are provided below verbatim without attribution to the specific reviewer.

Peer Review Comments in Response to Charge Questions and FDA Response

GENERAL IMPRESSIONS

  • Reviewer #1. The report titled, Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans, offers a succinct, comprehensive and clear definition of the problem, defines the risk for melamine and assesses the intake of melamine per capita from milk-derived food ingredients. The document is clearly written.

    In September 2008, the FDA learned of infant formula contamination with melamine linked to 13,000 hospitalizations and 3 deaths in China. All occurred in children less than 3 years of age, no illness or death occurred in adults. Renal stones and insufficiency were described. The melamine was found only in the powdered formula and concentrations of the melamine varied from 0.1 ppm to 2,500 ppm (the level of detection is 1 ppm in some foods).

    Melamine should be found only in very small amounts in food from environmental contamination as there is no approved melamine use in direct addition to human or animal food in the US, nor is it permitted to be used as a fertilizer in the US.

    Because infants have a unique physiology, rapid growth and development, and totally rely on a single food, for bottle fed infants it is formula, the FDA could not establish a level of melamine and its analogues that does not raise public health concerns.

    For food and food ingredients other than infant formula, the report estimated that levels of melamine and its analogues below 2.5 ppm in foods do not raise public health concerns.

    The estimation of this risk was based on a 100 fold safety factor between the no-observed-adverse-effect level (NOAEL) from animal data (63 mg/kg bw/d) and the tolerable daily intake (TDI) for humans. The FDA then applied an additional 10-fold factor, yielding a combined safety factor of 1000-fold to compensate for uncertainty in the determination of safety/risk. To estimate the amount of melamine consumed/day, the FDA used a 60 kg person and estimated the dietary intake at 3 kg/day. These estimations raise questions as to why 60 kg was used in the calculation since the average weight of the US adult population is 80 kg, 86.6 for men and 74.4 for women. Further, since most consumption of milk is by young children, a weight more appropriate to a toddler or child might be useful. The USDA estimates the intake of food at 2.1 kg/day for adults. For young children the intake on a per kg basis is much higher to compensate for growth. Finally, this estimate of risk does not include vulnerable members of the population such as young children, pregnant or lactating women, the elderly and those with chronic disease.

    Because the incident of melamine contamination was of milk and milk-derived ingredients, the document further estimates US per capita melamine intakes from milk-derived food ingredients as a percent of the TDI/10 if melamine were present at 2.5 mg/kg in the milk. From this estimation based solely on milk, the report concludes that levels of melamine and its analogues below 2.5 ppm in foods other than infant formula do not raise public health concerns. The report does not address the possibility that the food matrix of foods other than milk might be different and the risk might change because of the matrix. Further, the report notes that trichloromelamine, a compound that readily decomposes to melamine, is approved for use as a sanitizing agent on food processing equipment and utensils except for milk containers and equipment. Thus, it is likely that milk and milk derived foods have a lower contamination of melamine than other products.

    The animal study, on which the risk is estimated, was a short term study, 13 weeks. The report is thus strongly focused on what appears to be acute risk. There is no acknowledgement of the risk that might occur over a long period of time, such as a lifetime, if that risk is compounded by other factors or ingredients in foods or in the environment, if long term exposure might lead to a chronic disease or alter the response to other substances in the food or environment, or has an effect on treatments of disease.

  • Reviewer #2. The Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans (October 3, 2008) and Update (November 28, 2008) are concise reviews of the current state of the field regarding the potential for melamine and its analogues contamination in the food supply including infant formula. The initial report cited the epidemiologic evidence from China indicating the wide range of concentrations of melamine in powdered infant formulas and the reports of melamine detection in confections and beverages in other countries. The estimated level of melamine in food resulting from its current uses is less than 15 µg/kg (0.015 ppm). A review of the possible mechanisms for renal injury from melamine and its analogues based on animal studies was provided including the NOAELs for rats and mice. It emphasized that the NOAEL for stone formation of melamine toxicity is 63 mg/kg bw/d in a 13-week rat study, a value that is used with human exposure assessments to provide an estimate of human safety/risk. However, this animal study evaluated the exposure to melamine alone in the feed and not in combination with its analogues. In collaboration with numerous other federal agencies, the FDA developed a TDI for melamine and its analogues during the pet food contamination of 2007 indicating 0.63 mg/kg bw/d as the TDI.

    In the present exposure event, the potential risk for toxicity from consumption of infant formula contaminated with melamine and its analogues is far higher than last year's risk of toxicity to humans from consumption of animals that had been inadvertently fed contaminated feed. The report concluded that this information cannot be applied to the current situation because the contaminated product represents the totality of caloric exposure for most of these infants; the exposure is chronic over months; the persons ingesting the products are infants and toddlers whose renal systems are not yet fully developed; and the exposure is not mitigated by previous passage through the digestive system of an animal. In addition, they summarized the significant gaps in the scientific knowledge regarding melamine and its analogue's toxicity regarding infants including:

    1. The impact of the presence of more than one melamine analogue which has the potential to increase the toxicity of infant formula.
    2. The consequences of continuous use of these infant formulas as the sole source of nutrition.
    3. The possibility that these formulations can be fed as a sole source of nutrition to premature infants with immature kidney function and even greater intake of infant formula per unit body weight for a longer time period.

    With these considerations in mind, the report indicated that the US FDA cannot establish a level of melamine and its analogues in these products that does not raise public health concerns.

    They described the procedure used to calculate a level of melamine and its analogues in food that does not raise public health concerns and used a 100 fold safety factor between the lowest NOAEL from animal data and the TDI for humans yield 0.63 mg melamine and its analogues/kg-bw/d. Furthermore, the potential for increased toxicity from the combined exposure of melamine and cyanuric acid resulted in the FDA applying an additional 10-fold safety factor to yield a combined safety factor of 1000-fold. Again, these approximations are conservative and additional data is needed for their verification.

    The example summarized in Table 1 that U.S. consumers would be exposed to only 1.1% of the melamine TDI/10 if all of the major milk-derived ingredients listed in the table were contaminated at a melamine level of 2.5 mg/kg (2.5 ppm) assuming an average per capita ingredient intake, is an attempt to emphasize that melamine and its analogues below 2.5 ppm in foods other than infant formula do not raise public health concern.

    The safety/risk assessment did not address the potential public concerns from food containing just melamine or one of its analogues. Because FDA has found infant formula where just melamine or just cyanuric acid was present, it is updating the safety/risk assessment. These findings were in U.S.-manufactured infant formula products, and only extremely low levels of melamine or cyanuric acid have been detected in them ranging from 0.137 ppm of melamine in one product to 0.247 ppm of cyanuric acid in another, that are up to 10,000 times less than the levels of melamine reported in Chinese-manufactured infant formula. The Update applied the TDI/10 of 0.063 mg melamine/kg-bw/d to total melamine consumed per day for a 3 kg infant to yield 0.189 mg melamine/infant/day.

    In order to estimate the level of melamine that does not raise public health concerns, FDA used a worst case scenario in which all of an infant's total daily dietary intake (0.15 kg powdered infant formula) is contaminated with melamine. The new total amount of melamine/infant/day approximated 1.26 mg melamine/kg food. In summary, if 100% of the diet were contaminated at a level of 1.26 ppm of melamine, an infant's daily intake would equal 0.063 mg/kg bw/d and to add an additional margin of safety the value is decreased to 1.0 ppm melamine. The safety/risk assessment assumes the analogues to have equal effect. Thus, levels of melamine or one of its analogues alone below 1.0 ppm in infant formula do not raise public health concerns.

    FDA is sponsoring ongoing animal studies to assess the potential toxicity from co-ingestion of melamine and cyanuric acid and will update its interim safety/risk assessment as appropriate. The information provided appears accurate and unbiased and the gaps in the data base are clearly emphasized. The presentation of the data is clear and the conclusions drawn are rational in view of the current state of knowledge in the field.

  • Reviewer #3. A major limitation of this report is that it relies primarily on data concerning the toxicity of melamine ingested by itself. While the report also addresses the apparent cytotoxicity of cyanuric acid, a true assessment of the risks of melamine/cyanuric acid (CYA) requires new/additional toxicology studies. Toxicologic studies of melamine in dogs and rats clearly demonstrate the relative lack of toxic effects of melamine alone. Aware of these studies, veterinary scientists hypothesized the presence of cofactors as the reason for the apparent nephrotoxicity associated with melamine ingestion in the 2007 outbreak of pet food associated nephrotoxicity. After detecting cyanuric acid in addition to melamine within kidneys of affected animals, studies focused on the interaction between these related compounds, as both chemicals are relatively nontoxic by themselves. When melamine and cyanuric acid were combined in acidic fluid (as is present in distal renal tubules) in vitro, crystals formed with a structure identical to those observed in renal tissue. This crucial interaction resulting in toxicity was further demonstrated in vivo in 2007, as cats fed either melamine or cyanuric acid alone suffered no ill effects while those fed a combination of melamine/cyanuric acid developed acute renal failure with the deposition of insoluble melamine/cyanuric acid crystals (melamine cyanurate) within renal tubules. Similar results were obtained with a toxicologic study in rats in 2008. Subsequent studies in fish and pigs indicate that these species also develop similar renal tubular crystals and renal failure after ingesting a combination of melamine and cyanuric acid. As stated in the rat melamine/cyanuric acid combination toxicity study (Dobson, 2008), "The dose-response curve and NOAEL for the mixture is almost certainly different from that of either compound alone." Therefore, while the accuracy of information concerning melamine toxicity is sound, conclusions about the risks of the combination of melamine and cyanuric acid cannot yet be drawn.
  • Reviewer #4. The recent difficulty in protecting human health from potentially contaminated food is complicated by the fact that so many commonly consumed products could have been contaminated by at least 4 (four), it is assumed, equally toxic compounds viz. melamine and analogues. Literature information has revealed the extraordinarily low toxicity of melamine and cyanuric acid, individually, in some animal species, which was probably one of a few reasons why melamine was postulated as a protein adulterant/ Kjeldahl spoof for economic profit, but also given the possibility to freely cycle and get rid of waste from the production of melamine based products.

    In my opinion, the interim S/RA clearly presents and summarizes scientific information that can be reviewed in the literature, with the corollary that human (adult) exposure to those animals fed a diet containing melamine and/ or analogues is not likely to cause a risk to their health. Further, that exposure to a fairly wide range of different food ingredients and foods considering the various scenarios will not result in a public health concern for adults. The risk was the potential formation of melamine:cyanurate agglomerates in the kidneys of those people who consumed potentially adulterated or tainted food. The S/RA presents the information and develops a sound conclusion that the risk of humans (adults) developing agglomerates of melamine and cyanuric acid is not a concern, and I agree based on the information in the document and in the literature.

    Infant exposure to melamine and/ or analogues was also addressed in the various documents. In that case though, I have concerns with the S/RA given the infant's underdeveloped kidneys, and with the argument that only high concentrations of (pure) melamine and (pure) cyanuric acid can form agglomerates in tissues and organs. Considering the latter point, an experiment was carried out which quite surprisingly demonstrated that 25 ppb of each melamine and cyanuric acid can agglomerate in solution (data presented in point 12 of this document); while this does not in any way conclude risk for infants, it does illuminate the possibility of low level agglomeration in vivo, particularly since kidney function is not fully developed in infants and that a model on the clearance of melamine and cyanuric acid has not been developed. An assumption in question is that pure melamine and cyanuric acid were used to adulterate processed protein products. Is it possible that a waste stream cocktail of melamine and analogue products was used to adulterate our food, and not pure melamine or pure cyanuric acid? This complexity creates a dilemma in establishing a safe level for all of the compounds, particularly for infants, given that the waste stream could have widely different compositions over time and include compounds not related to the production of melamine products. It could easily take a decade of concerted effort to uncover the complicated involvement among even the four compounds named herein, let alone their potential side products or co-contaminants from the waste stream, which have gone largely undetected due to the primary focus on kidney failures as a result of melamine:cyanurate agglomerates, formed therein. Infants could be faced with an additional bioburden besides melamine and analogues even with the proposed S/RA. The fact that infants have been deliberately exposed to the potential myriad of compounds resulting from the chemical adulteration of infant formula with melamine and analogues has resulted in our global contempt of those processes. All efforts to protect the most susceptible demographics must be employed, particularly given the underdeveloped kidneys of infants.

    Overall, the scientific rationale and conclusions of the S/RA for adults and infants are generally sound, and as such I support the current S/RA proposal based on our current scientific database on the topic, and particularly so since it is apparent that our best efforts have been implemented to protect the public from future occurrences, viz. demonstrated vigilance in testing. The regulatory bodies have a formidable task before them. I do recommend lowering the proposed S/RA for infants to 0.25 ppm based on various points presented herein and I have comments and suggestions for consideration that could hopefully help in shoring up the document and to consider some future research.

  • Reviewer #5. Please see responses to charge questions and specific comment section.

FDA Response: Overall the reviewers appear to have a generally positive impression of the S/RA. Although some additional unique issues were identified in this section that the peer reviewers believe need to be addressed or clarified, the majority of the responses in this section raised issues that were identified in the 12 charge questions. These included: the basis for the assumptions used in the exposure assessment and their applicability to adults and young children; effects of the food matrix on toxicity, toxic effects of long-term exposure, and underdeveloped infant kidneys and toxicity. In addition to the response below, please see FDA's responses to those issues under each of those questions.

Following the October 3, 2008 S/RA, FDA received information from the Chinese Ministry of Health indicating that virtually pure melamine (99%) was associated with the Chinese infant formula contamination event. Additional information and discussion are provided in response to Charge Questions #2, #7 and #8 below.

Reviewer #1 stated that it would be more appropriate to use a body weight of 80 kg, as they believed this is more reflective of the average weight of the U.S. population.

The use of a default body weight of 60 kg for calculating dietary exposure is a common practice as an overall average value for individual body weight when estimating per capita exposure in the general population. This default value is used by international organizations involved with food safety.1 Another common practice used in per capita dietary exposure estimates for the general population is to use as an overall value for individual food consumption per day is 3 kg.2

If 80 kg were used instead of 60 kg as an estimate of average body weight, the level of melamine that would not cause a public health concern would be 3.4 mg/kg, as compared to 2.5 mg/kg when 60 kg bw is used. Thus, the use of 80 kg bw is less conservative than the approach used to estimate exposure to melamine in the S/RA.

Reviewer #1 also commented that a weight more representative of a toddler should have been used as part of the S/RA. As described in detail in our response to Charge Question #7, we have performed a refined intake estimate for young children (i.e., 2, 3, 4, and 5 year olds and 2-5 year olds as a group) presuming that 100% of milk and dairy products consumed by these population subgroups contained melamine at a level of 2.5 mg/kg. This analysis demonstrates that even with the application of these conservative assumptions, the refined intake estimate for young children was essentially equal to the TDI/10.

FDA agrees with Reviewer #3 with regard to all of the issues and points discussed above as we have pointed out in both the October 3rd and November 28th S/RAs and in our responses to various specific questions in this document. It is clear from the studies referred to by Reviewer #3 that the combined toxicity of melamine and cyanuric acid given together occurs at a significantly lower dose than when each of these substances are given separately. However, for risk assessment purposes we must determine the threshold dose for early evidence of kidney toxicity. We are attempting to find a dose that does not result in even the minimal formation of microcrystals of melamine:cyanurate. Studies are planned to investigate the use of kidney biomarkers of early, reversible toxicity. If these biomarkers can be used successfully, then we can do longer term toxicity studies to assess the potential for chronic exposure to elicit minimal to significant toxicity and monitor the long term functionality of the kidney. As is common in S/RA, not all of the pertinent and necessary scientific data are available. In this circumstance FDA used safety/uncertainty factors as an interim approach to assess toxicity based on the available data. The magnitude of the safety factor selected, 1000-fold, is expected to account for even the added toxic potency resulting from combined exposure to melamine and cyanuric. Furthermore, FDA has used conservative estimates of dietary exposure as an additional hedge on the side of safety while estimating the risk of exposure to melamine. It is anticipated that the conservative use of these latter two risk assessment tools will not underestimate the combined toxicity of melamine and cyanuric acid ingested together to mediate kidney toxicity. However, we recognize that this can only be confirmed by additional studies.

FDA will carefully evaluate the additional references supplied by the peer reviewers and any pertinent issues will be addressed in the revision to the S/RA. In addition, FDA will modify portions of the S/RA to enhance the transparency of presentation of data and their explanation as suggested by the peer reviewers as discussed below under the responses to the individual charge questions

Charge Question 1: Does the S/RA report transparently identify and clearly justify the target populations of the assessment? Does it explain how data used were identified and what criteria were used to determine the suitability of the data? Are these criteria appropriate?

  • Reviewer #1. The S/RA report transparently identifies and clearly justifies infants as one of the target populations. There is another population that is alluded to in the report and that population is not transparently identified nor clearly justified. The second population is the population that does not consume infant formula. Under the section SAFETY/RISK ASSESSMENT FOR FOOD AND FOOD INGREDIENTS OTHER THAN INFANT FORMULA, the calculations are based on adult weight and so one assumes that the second population refers to adults. No mention is made of children whose weights and food intakes/kg bw/day would be very different from those used in this document to assess a level of melamine that does not raise public health concerns. Further, the population that consumes the most milk or milk containing foods is toddlers and young children. Milk consumption decreases as children reach the teenage years, a time when their body weight would approximate the 60 kg used in this report to estimate risk. Other vulnerable groups are not acknowledged such as pregnant or lactating women, the elderly or those with a chronic disease. It would be helpful if the report clearly identified who is included or excluded from the non-formula consuming group and the reasons for their status.

    The draft report identifies the data that were used and references it. There is no discussion of how the data were identified and what criteria were used to determine the suitability of the data. An issue that is important to infants and children, in addition to their small size, is the fact that their organ function may be different than that of adults. The draft report specifically refers to the immaturity of the infant renal system. This fact raises the question of the appropriateness of the animal models on which toxicity was based. Were these adult animals, growing animals, infant animals? Can the organ functions of these test animals be assumed to be comparable to the function of infants or children? If not, then it will be difficult to extrapolate information from adult animals to infants and growing young children.

    A similar argument must be made for other vulnerable groups such as pregnant or lactating women, the elderly and those with chronic diseases.

  • Reviewer #2. The S/RA report is transparent and clearly identifies and justifies the target populations of the assessment. The explanation is appropriate of how data and criteria were used to determine the suitability of the data. The criteria are appropriate.
  • Reviewer #3. The target populations have not been adequately identified in the report. While some specific populations at risk of consumption of contaminated formula or its ingredients (as listed in the table) are identified, other groups are not considered. This risk assessment is primarily responding only to instances of melamine/cyanuric acid adulteration and toxicity that have already occurred, rather than discussing the reason these chemicals are in the food and therefore identifying what foods might be expected to be contaminated. At the bottom of page 1 and the start of page 2 there is a discussion of sources and expected levels of melamine in food as part of normal processing. These levels are extremely low, so a hypothesis as to the source of the melamine/cyanuric acid is required in order to then predict which other foods may be affected and, consequently, determine risks and develop strategies to identify potentially contaminated food prior to entry into the human (or nonhuman animal) food supply. It is clear that melamine has been, apparently for a long period of time, intentionally added to food ingredients originating in China for which the economic value of the product is based on protein content. Melamine is a relatively inexpensive nonprotein nitrogen compound that falsely elevates the apparent protein content. As melamine alone, based on toxicologic studies performed in rats, dogs, and more recently in cats is nontoxic, this apparently common practice of adding melamine to protein-containing food products would likely have gone unnoticed if cyanuric acid (again, a relatively nontoxic compound by itself) did not occur as a co-contaminant. This co-contamination/adulteration occurred in 2004 and 2007 in pet food and in infant formula in 2008, resulting in nephrotoxicity in large numbers of dogs, cats, and infants. It is also reasonable to assume other less widespread or less severe (lower levels of contamination) instances of this melamine/cyanuric toxicosis have occurred in both pet and human populations. Therefore, products anticipated to contain melamine include infant formula, powdered milk, glutens (wheat or corn), rice protein, powdered protein supplements, and food containing these ingredients such as pasta, processed cereal, jarred baby food, and bread. The second risk factor, in addition to which population might be exposed to contaminated product, is level of exposure. As clearly shown with dogs, cats, and infants, populations that consume large quantities of a single processed food are more likely to suffer severe toxic affects. Dogs and cats are typically fed only or primarily a single pet food, so contamination of pet food with cyanuric acid and melamine resulted in a massive outbreak of renal failure that was frequently fatal. Similarly, infants are fed primarily formula (and processed baby food, another potential source of melamine/cyanuric acid), so contamination of infant formula with cyanuric acid and melamine resulted in a massive outbreak of renal disease with a number of reported fatalities. These populations must be assumed to be identifiers ("canaries"). Vegetarians are similarly at risk, though the nature of their diet makes distal tubular fluid and urine alkaline, potentially reducing crystal formation. Acute toxicity resulting in uremia is easily identified. It appears likely that low level toxicity in adult humans has occurred and that this may be a potential cause of chronic kidney disease. The effects of intermittent, low level exposure should be considered and studied in animal models.
  • Reviewer #4. Yes, the S/RA report transparently identifies and justifies the target populations of the assessment: adults, toddlers and infants. Further, it is quite clear how the data were used to reach their conclusions and particularly the criteria selected demonstrating suitability of the data. I believe the criteria are appropriate, except for infants where an additional safety factor is suggested (more below). Also, I do have comments about the analytical aspects of the work, particularly in the ppb to low ppm concentrations of melamine and cyanuric acid, which is related to the formation of melamine:cyanurate agglomerates in the ppb to low ppm concentration ranges (discussed with data in point 12 herein).

    Question: The health of the hogs that contained trace levels of melamine is not mentioned in the report, other than comments regarding their kidneys. The reader must assume they were healthy given they were slaughtered and from comments about kidney health. Would it be a stretch to indicate the animals were slaughtered because they appeared healthy overall? Perhaps a comment on their health at the time of slaughter is warranted to provide some indication of the general impact of exposure to melamine and/ or analogues from feed. The same question is asked about the other animal species that were slaughtered.

  • Reviewer #5. It is clear that the target populations are infants and children. There was no account of how utilized data were identified, nor what criteria were used to determine their suitability.

FDA Response: Although it was clear to the authors of the S/RA that we were considering either infants and their exposure to infant formula or non-infants (all ages) and their exposure to potentially contaminated milk and milk products used in manufacture of foods, we were apparently not adequately transparent in the document's discussions. The populations included in the S/RA and why they were selected will be clearly described in the revision to this S/RA as requested by the peer reviewers. FDA agrees with Reviewer # 3 that there are other populations and conditions that may affect the toxicity of these triazine substances; however, some of the needed data will be forthcoming from future research that FDA and other organizations perform. The scope of FDA's interim risk assessment was primarily to address the risk to the target population: healthy individuals consuming infant formula and milk derived products. It was not within the scope of that risk assessment to include special populations, especially because information on the impacts of melamine or its analogues on those with compromised renal function, or other special health conditions, was lacking. If additional information becomes available, FDA will re-evaluate these populations.

FDA agrees with the comment of Reviewer #3 that exposure is a risk factor to consider. However, we believe that the exposure that was used is sufficiently conservative, even for subpopulations that may have restrictive diets, such as vegetarians. With the exception of infants, we are not aware of any sub-population that would consume large quantities of a single processed food over an extended period of time. There is considerable variability in the human diet, in terms of both food supply and the sources that are used to manufacture food.

Reviewer #1 commented that the exposure estimate may not have adequately accounted for toddlers and young children. As a first approximation, FDA chose 60 kg to act as the surrogate body weight (bw) for all life stages from a 2 year-old toddler to the eldest consumer. FDA believes that this assumption, together with other assumptions used in the S/RA, is appropriately conservative to compensate for the consideration that a child's intake is higher than an adult's on a weight of food consumed per body mass basis and therefore, it is not necessary to use the body weight of younger individuals. As described in detail in our response to Charge Question #7, we have performed a refined intake estimate for young children (i.e., 2, 3, 4, and 5 year olds and 2-5 year olds as a group) presuming that 100% of milk and dairy products consumed by these population subgroups contained melamine at a level of 2.5 mg/kg. This analysis demonstrates that even with the application of these conservative assumptions, the refined intake estimate for young children is essentially equal to the TDI/10.

At this time, the scientific literature does not provide data on the potential susceptibility of special populations, such as pregnant or lactating women, the elderly or those with chronic disease. FDA hopes to perform animal studies that will provide some insight into the range of dose-effects that may exist for these populations.

In their general comments, Reviewer #1 commented that the S/RA does not consider the possibility that different food matrices might involve a greater exposure to melamine and its analogues, and thus more risk than milk. As discussed above, FDA assumed in the S/RA that all of the melamine and its analogues present in a person's diet (assuming that one-half of a person's diet was contaminated with melamine and its analogues) were available for exerting their effect. Therefore, we believe that the S/RA accounts for any matrix variations that might exist among different foods. Furthermore, FDA is unaware of any studies that show that the food matrix has an impact on the bioavailability of melamine or its analogues.

Reviewer #5 indicated that the explanation of how the data used in the S/RA was identified and the criteria that were used to determine the suitability of the data was insufficient. FDA used all available, relevant published studies, of which we were aware at the time the S/RA was prepared. Relevant studies were defined as animal feeding studies that focused on nephrotoxicity as a result of exposure to melamine, cyanuric acid, or the combination of both compounds. When multiple studies were available, the study with the lowest No-Observed-Adverse-Effect-Level (NOAEL) was used.

In terms of the comment that the S/RA appears to be focused strongly on acute risk, a 13-week or subchronic feeding study in rodents is considered to be an intermediate duration toxicological study. There are other longer duration study results available, but CFSAN selected the 13-week study because it provided the lowest No-Observed-Adverse-Effect Level (NOAEL) and so the most conservative estimate of risk. Later information from the Chinese government made it clear that some of the infants may have been exposed for up to 2 years. If data on low level chronic toxicity of melamine were available we would have included this in our analysis. The FDA is not aware of any evidence of chronic toxicity associated with low level dietary exposure to melamine. The study source, as asked for by reviewer #5 is in the published National Toxicology Program (NTP) report can be found on the web at http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr245.pdf.

In response to the questions posed by Reviewer #4 in regards to the health of the hogs mentioned in the report, the hogs in the 2007 risk assessment were clinically normal. In addition, FDA's subsequent study using fish and hogs confirmed that the animals feed melamine at 400 mg/kg for 3 days remained in good health and were clinically normal. (Reimschuessel et al, 2008).

Charge Question 2: Was the risk assessment methodology appropriate regarding the use of the animal-based determination of the no-observed-adverse-effect level (NOAEL), application of safety factors, and determination of infant and adult population exposure to melamine and its analogues and the risk characterization?

  • Reviewer #1. I must defer to the toxicologist who will review the document as this question is out of my area of expertise. However, I do wish to raise a couple of points.

    The rationale for using a 60 kg person to estimate the amount of melamine consumed per day is not clearly stated. The average weight of males and females in the US is 86.6 kg and 74.4 kg respectively.1 This is, on average, approximately a 30% difference in weight from what was used to estimate safety/risk.

    The basis for estimating the weight of a typical diet as 3 kg, 50% liquid and 50% solid, is not clearly identified. The USDA estimates total food intake at 2.1 kg/day2 for adults. This is of course very different for infants, young children, pregnant or lactating women, and other vulnerable groups.

  • Reviewer #2. The risk assessment methodology was appropriate based upon the current state of animal-based data and the determination of NOAEL, application of conservative safety factors, and the determination of both infant and adult population exposure to melamine and its analogues.
  • Reviewer #3. The methodology for determining NOAEL and risk characterization was not appropriate, as the levels identified were based on exposure to melamine alone. Extrapolation from the data obtained at toxicologic studies using melamine alone yields false results, as clearly demonstrated by 2007 S/RA report in which the no-observed-adverse-effect level for melamine and its derivatives was determined to be 63 mg/kg bw/d. While this NOAEL may be accurate for melamine alone, it has been clearly established that the toxic dose of melamine and cyanuric acid together for cats (64 mg/kg bw/d) is equivalent to the NOAEL for melamine alone. This does not represent a species variation, as melamine or cyanuric acid administered separately to cats does not produce intra-renal crystals and renal disease. Dividing a potentially lethal dose of melamine/cyanuric acid by a factor of 1000, as done in this report, does not insure safety. As all of the NOAEL determinations were based on the TDI of melamine alone (with an "uncertainty factor" included to account for the effects of melamine when combined with cyanuric acid), they are not an accurate assessment of the current risk. Furthermore, the recommendations for safe levels should take into account the effects of cumulative, low level, chronic exposure. Histologic renal injury represents a more sensitive endpoint for chronic exposure studies rather than uremic death from a single dose.
  • Reviewer #4. Yes to almost all aspects of the question. Given the NOAEL and the applied safety factors, and our current understanding of the mode of action of regarding the "toxicity" of melamine:cyanurate, the potential adverse risk to the adult population appears highly unlikely with the proposed safe level. With respect to exposure to infants, the statement "Thus, the US FDA cannot establish a level of melamine and its analogues in these products that does not raise public health concerns" was acceptable, but then the suggestion to monitor at the safe level of 1 ppm and with the safety factor of 1000 (considering diet and weight, etc.) for infants this should not raise public health concerns. I believe more could be considered in addressing the safe level for infants, particularly considering (a) the state of current analytical methods and that realizing method detection limits of 50 ppb and less is not a stretch, (b) with the possibility that mothers, who may be exposed to the safe level of melamine, could in fact introduce melamine to their infant's diets through breast milk, (c) that melamine and analogues used for adulteration could be quite impure, and (d) the fact that agglomerates of melamine and cyanuric acid can form at the part per billion (ppb) level (in vitro to date).

    Comments and considerations about the analytical methods:

    The FDA has successfully developed analytical methods, based on liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) for melamine and cyanuric acid, but will all labs testing infant formula for melamine and cyanuric acid be required to follow the FDA method? Specifically, we have seen credible labs develop erroneous extraction methods and mixed standards preparation (additional comments provided in point 12 of this review). My review of the published FDA studies indicates that acid is added to the extracting solutions and in the preparation of standards. Without acid, i.e., water alone, solvent alone, or water and solvent, underestimated levels of melamine and cyanuric acid in sample will occur. We have independently used 10% formic acid for extraction of melamine and cyanuric acid from food since 2007 (data not published, 2007). What I would be concerned with are those labs not required to follow the FDA method employing LC-MS/MS, given the consideration they are considered credible. Also, accreditation of a lab does not in itself identify a credible lab; rather it captures in writing what was and is done during an analytical procedure. Successful participation in round robins with a fairly stringent tolerance around the expected concentrations should be required for a lab to support the testing and identification of food essentially free of melamine and analogues, particularly cyanuric acid, for food destined for infants. The lynch pin of the regulatory program is based on the analytical chemistry and the methods used, and a bias in extraction and analysis will only thwart attempts at protecting the public at large.

    The "safe" level for infants:

    Given the uncertainties indicated above, would it be possible to propose a safe level of 0.25 ppm melamine and analogues in infant formula? While this level is only 4 times lower than the proposed level of 1 ppm, it will require labs to ensure their method detection limits are at least 0.050 ppm. Analytical chemists will agree, based on proficiency tests, that significant bias and precision among highly competent labs is possible. The additional safety factor and protocol could then provide sufficient argument that significant measures of protection have been taken, particularly for the most susceptible demographic. I am not aware of data on melamine concentrations in breast milk as a result of the mother's potential exposure to melamine at the safe level proposed by the FDA. We also need recognize that potential adulteration of formula could involve melamine waste, rather than high-grade melamine, which begs the question about the safe level. By ensuring the safe level for melamine is as low as possible for infants, given our "best available technology," we can help minimize their exposure to a potential myriad of compounds. Finally, perhaps melamine:cyanurate agglomeration is formed at the ppb level in vivo, potentially having no impact on the health of the infants, but without data we cannot be sure at this time.

  • Reviewer #5. The methodology used for calculation of the adult and infant formula TDIs and maximum food contaminant levels are appropriate, as are the starting point and basic assumptions. The critical study, from which the 63 mg melamine/kg value was obtained, should be identified and described in more detail. It is not possible in an information vacuum to judge the quality of neither the study nor FDA's justification for selection of the endpoint and NOAEL.

FDA Response: FDA understands the point raised by Reviewer #1 and agrees that for the purpose of clarity, the rationale for the values assumed for body weight (60 kg) and food consumption (3 kg) should be better described. The rationale for the use of 60-kg as the body weight was discussed in FDA's response to Charge Question 1. Likewise, the rationale for the use of 3 kg of food is discussed in FDA's response to Charge Question 7.

Regarding the comments raised by Reviewer #3, FDA agrees that we do not have the necessary test data from co-exposure to melamine and one of its analogues to estimate the toxicity of combined exposure with confidence; however, FDA is supporting the design and implementation of studies that will provide a better and more representative estimate of expected kidney toxicity when these substances are ingested at the same time.

FDA agrees with Reviewer #4 that it is desirable to include additional factors into the S/RA (see additional discussion under question #9, 11 below). In the final analysis, the decision concerning which analytical method is best to use will be indicated by the results of the toxicity testing, especially the data collected from the toxicity studies in which melamine and one or more of its analogues are given to the test subject at the same time. Whatever the "threshold" for crystal formation is, the analytical assays must be sensitive at the threshold level to provide adequate evidence of safety. The current LC/MS/MS methods on the FDA website can reliably measure 0.25 ppm melamine and cyanuric acid in infant formula. Comments and considerations about analytical methods provided by this reviewer are consistent with our views and are discussed in more detail below in our response to the Question 12 comments from the peer review group

Charge Question 3: We identified three primary sources of uncertainty regarding the data on melamine and its analogues: 1) extrapolation of animal data to humans; 2) variation of response within humans; and 3) lack of adequate data on how toxicity changes (increases) with concurrent exposure to melamine and cyanuric acid. There is also uncertainty relating to the estimation of the extent of exposure to melamine and its analogues that we accounted for using a "worst case" exposure estimate. Are each of these sources of uncertainty in the S/RA identified and characterized? Are there others that should be highlighted? If so, what approach would you advise in characterizing those uncertainties?

  • Reviewer #1. The S/RA report clearly identifies uncertainty regarding the data on melamine and its analogues with respect to extrapolation of animal data to humans and the lack of adequate data on how toxicity changes (increases) with concurrent exposure to melamine and cyanuric acid. The report identifies the uncertainty for infants because of immature renal function and a single food as their total nutrition source. Uncertainty might exist for other life stages, such as the elderly, pregnant or lactating women, or those suffering with a chronic disease.

    Other substances that might have a synergistic relationship with melamine to cause disease are not identified in the S/RA. Is there certainty that other environmental or dietary substances would not interact with melamine to potentiate the harm melamine causes or the possible harm the other substances could cause?

  • Reviewer #2. The three primary sources of uncertainty regarding the data on melamine and its analogues are appropriate, as well as that relating to the estimation of the extent of exposure to melamine and its analogues that are accounted for using a worst case exposure estimate which are conservative. One suggestion for consideration is to develop a scenario that emphasizes the age-specific changes in both renal function and drug metabolism/excretion that may play an important role in susceptibility to renal injury secondary to melamine and its analogues. In an attempt to obtain useful information in this area, animal studies using newborns and maturing offspring should be performed. The application of both the normal renal physiology in these models and the use of biomarkers (NGAL and KIM-1) to identify early renal and tubular injury would be of assistance.
  • Reviewer #3. The first part of this question is simple to answer. Humans are animals; the question is better framed in the context of species differences. Nonhuman species, particularly primates (there was a report of an infant gorilla in China that developed renal failure after consuming contaminated infant formula) and dogs, are used routinely as animal models of human disease; dogs and rats are typically used in toxicologic studies as demonstrated with the melamine toxicity data referenced in this report. The similarities in anatomy and physiology between different mammalian species allow for these comparisons to routinely occur. Lessons learned in human medicine are often used in the treatment of patients treated by veterinarians, and knowledge gained through the study of diseases occurring in various animal species is used by physicians treating their patients. This concept of "one medicine" is scientifically sound and is beneficial for both people and nonhuman species. Melamine/cyanuric acid appears to be excreted unchanged (ie unmetabolized) by the kidney. Filtration of an inert substance by the kidney would be expect to occur similarly across species, and has been shown to occur similarly in dogs, cats, pigs, and fish. All mammals have similar renal function, and differences in handling of melamine/cyanuric acid would not be expected. Interestingly, the renal handling of melamine and cyanuric acid also appear to be similar in fish. This data indicates that the probability that renal handling of melamine/cyanuric acid is unique in humans and, different than in other mammalian species, is remote. Intraspecies variability in response to the clinical effects of melamine/cyanuric acid toxicosis has not been clearly demonstrated. There was no apparent age or sex-related susceptibility to this toxicosis in dogs and cats. Similarly, an increased susceptibility in younger infants compared to older infants consuming contaminated formula has not been reported. However, it is likely that individuals with pre-existing renal disease would be more susceptible to developing renal disease following lower melamine/cyanuric acid exposure levels (duration or amount), as smaller degrees of renal damage in these patients would potentially result in clinical renal disease. The third area of uncertainty, i.e. the effects of the combination of melamine and cyanuric acid rather than their individual effects, is of critical importance and is, in fact, not an uncertainty. It is known that melamine and cyanuric acid together cause toxicity, and this toxicity has not been evaluated in this report. The NOAEL and TDI have not been determined for the combination of melamine and CYA; these values can absolutely not be extrapolated from data based on their individually nontoxic constituents. Doing so is misleading and falsely minimizes the risk posed by simultaneous exposure to these compounds.

    I would address the uncertainties as follows:

    1. The first uncertainty is scientifically unfounded; to suspect that the effects of melamine/CYA reported in a number of mammalian species would not apply to humans is foolhardy. To suggest humans have kidneys that are physiologically different than those in other mammals (the differences in renal physiology between fish and dogs are far greater than those between dogs and people) may convey a false sense of security to the public - i.e. that's an "animal" problem that doesn't occur in people. Unfortunately, the recent occurrence in China clearly demonstrates the danger of this approach.
    2. The second uncertainty is also not expected to be very significant, other than in regards to greater exposure in a subpopulation associated with the ingestion of a limited variety of food such that they are ingesting a greater amount of melamine/cyanuric acid. Because melamine/CYA are not significantly metabolized by the body and toxicity is due to renal excretion/filtering of the compounds, patient age would not be expected to be a significant variable. If available, epidemiologic evaluation of the children involved in the infant formula contamination should answer this question. Younger infants may be more likely to have the tainted formula as their sole dietary constituent versus older infants, but if this data can be interpreted in a manner that normalizes for melamine/CYA intake as a proportion of the total diet, the question of age and risk can be addressed.
    3. The lack of adequate data assessing the toxicity of melamine/CYA together is the most crucial shortcoming of this entire report. The toxicity of these compounds together was demonstrated by numerous studies more than a year ago; the lack of response to this situation is difficult to understand. Toxicity studies using varying levels and duration of melamine/cyanuric acid together are required. Although higher levels will be associated with acute renal failure and possibly death (as shown in cats receiving an acute dose of 64 mg/kg bw/d), evaluation of lower levels is also critical. This will require more sensitive assaying of renal function in these animals (periodic glomerular filtration rate determination, etc) as well as more chronic effects that would need to be assessed both functionally and histopathologically. In particular, chronic subclinical exposure (either continuous or periodic) may result in chronic renal tubular injury, manifested as chronic tubulointerstitial nephritis with scarring (fibrosis).

     
  • Reviewer #4. In my opinion, most of the sources of uncertainty in the S/RA have been identified and well characterized in considering available animal data. Protection factors in consideration of all previous animal studies and current ones carried out by the FDA and other organizations such as Davis, are based on the objective to prevent the formation of melamine:cyanurate agglomerates large enough to cause kidney dysfunction. What should also be considered is the possibility that adulteration with melamine and analogues may not occur with essentially pure compounds, particularly given they are compounds destined for manufacturing and not for human consumption. Further, that a concentrated waste cocktail of melamine and analogues could be used for adulteration, thereby including all concentrated melamine waste ingredients.

    Within the 90th percentile, I think we can all agree that the proposed safe level for adults will not create a public health concern; however, there is still question regarding the possibility of melamine:cyanurate accumulation for individuals with (highly?) compromised kidney function or in particular for infants where their kidneys have not yet fully developed.

  • Reviewer #5. It is not clear in the first paragraph of the Safety/Risk Assessment that the 100-fold safety factor is comprised of a 10-fold interspecies uncertainty factor (UF) and a 10-fold intraspecies UF.

FDA Response: FDA agrees with the comments of Reviewers # 1, and 2 and anticipates that future research will help to reduce or eliminate some of the uncertainties that were identified. FDA also agrees with Reviewer #4 that the toxicity of a "cocktail" of melamine and its analogues resulting from use of the waste products of chemical synthesis by industry is a possible source of triazine exposure. FDA agrees that when it appears feasible that a susceptible population may be affected, the response of their kidneys to melamine and its analogues exposure should be assessed.

FDA agrees with the issues raised by Reviewer #3 that need to be addressed by future research. However, FDA does not agree with the comment that since "melamine/CYA are not significantly metabolized by the body and toxicity is due to renal excretion/filtering of the compounds, patient age would not be expected to be a significant variable." FDA maintains that age is a significant variable. In the case of the 2008 melamine contamination event in China, the primary contaminant in milk and infant formula was melamine alone. The primary clinical finding in the infants has been kidney/bladder stones. In rats exposed to melamine alone, stones also formed and were composed primarily of melamine and uric acid (Heck and Tyl 1985, Ogasawara et al 1995). This also appears to be the case in the stones formed by infants in China (Sun et al. 2008, Shen et al. 2008). This response may depend on the levels of circulating uric acid. Most experimental animals have uric acid oxidase, which humans lack, and thus have different levels (lower) of uric acid. Thus, the extrapolation of animal experimental results to humans is subject to some uncertainty.

Similarly, uric acid levels vary in humans. Infants have higher levels at birth compared with adults and excretion rates for uric acid change during development. In a recent report from China, preterm infants were 4.5 times as likely to have stones as term infants (Guan et al. 2009). Certain disease states also cause altered uric acid metabolism. In addition, diseases and age affect renal competence which can also impact response to melamine exposure. FDA maintains that age and renal function can be important factors in potential risk variations among humans. Thus, the S/RA considers this to be an important uncertainty.

Guan N, Fan Q, Ding J, Zhao Y, Lu J, Ai Y, Xu G, Zhu S, Yao C, Jiang L, Miao J, Zhang H, Zhao D, Liu X, Yao Y. Melamine-Contaminated Powdered Formula and Urolithiasis in Young Children. N Engl J Med. 2009 Feb 4. [Epub ahead of print]

Heck HD,Tyl, RW. The induction of bladder stones by Terephthalic Acid, Dimethyl Terephthalate, and Melamine (2,4,6-Triamino-s-triazine) and its relevance to risk assessment. Regulatory Toxicology and Pharmacology 5: 294-313. 1985

Ogasawara H, Imaida K, Ishiwata H, Toyoda K, Kawanishi T, Uneyama C, Hayashi S, Takahashi M, Hayashi Y. Urinary bladder carcinogenesis induced by melamine in F344 male rats: correlation between carcinogenicity and urolith formation. Carcinogenesis 1995; 16:2773-2777. Shen Y Division of Urology, Department of Pediatric Surgery, Beijing Children's Hospital Affiliated to Capital Medical University, Beijing 100045, China (personal communication)

Sun N, Shen Y, Sun Q, Li XR, Jia LQ, Zhang GJ, Zhang WP, Chen Z, Fan JF, Jiang YP, Feng DC, Zhang RF, Zhu XY, Xiao HZ. Melamine related urinary calculus and acute renal failure in infants. Zhonghua Er Ke Za Zhi. 2008;46(11):810-5.

Charge Question 4: Does our risk assessment use appropriately large safety factors (or other exposure-based factors) in the risk assessment to adequately cover the variance in exposure that is possible across all age groups?

  • Reviewer #1. The appropriateness of the assumptions used to estimate risk is best assessed by the toxicologist who is reviewing this document.
  • Reviewer #2. The safety factors are appropriate and conservative in the risk assessment to cover the variance in exposure that is possible across all age groups. Again, additional information using animal models is clearly needed to distinguish the susceptibility of the different ages to renal injury secondary to melamine and its analogues.
  • Reviewer #3. While the safety factors seem large, they are based on data not applicable to intoxication with melamine/cyanuric acid together. As any amount of melamine/cyanuric acid may result in crystal formation in renal tubules, it is likely that there is no "safe" level for these two compounds together. Therefore, the risk assessment does not adequately cover the risk in any population.
  • Reviewer #4. The safety factors are large enough for the adult population, but I have concerns with potential exposure to melamine and cyanuric acid from breast milk to infants fed both breast milk and infant formula or just breast milk. Since most of the ingested melamine and cyanuric acid are rapidly cleared through body fluids, is it not possible that the aqueous fraction of breast milk could contain melamine from the mother's exposure to safe levels of melamine and what would be the contribution to the infant? Further, the concept that only large concentrations of melamine and cyanuric acid can form agglomerates in vivo requires some additional consideration (see Section 12 of this document).

    In consideration of the above, and given the discussion around the ease with which labs "validate" their methods to 1 ppm, there is indication of the ready possibility to lower the safe level to 0.25 ppm; given all methods demonstrate an excursion around any average value, which is true of the safe level as well, we could ensure the safe level is at least 3x below 1 ppm, and with rounding it would be down to 0.25 ppm. Analytically, most labs, if not all, would ensure the method detection limit would be anywhere from 10% to 20% of this level, which should not be a challenge for labs using methods based on LC-MS/MS. In this way uncertainty with the method detection limit will present the smallest overlap to the safe level. A recent method detection limit study (in my lab) for melamine in infant formula estimated 0.03 ppm or 30 ppb, using an acidified extraction, centrifugation and dilution method direct to LC-MS/MS. I can share this information with the FDA if there is interest. In this study, we ran 8 replicate spikes of infant formula at 0.25 ppm melamine. An appropriate 6-point calibration curve was used for quantification. The standard deviation of the mean was multiplied by 2.996 (for n-1 df). Recoveries even at this low level ranged from 49 to 54% with approximately 5% RSD. A similar study for cyanuric acid yielded a method detection limit of ~0.10 ppm. This method is very fast and quite inexpensive to run. A melamine N15 recovery standard is not used at this time.

  • Reviewer #5. I believe that inclusion of an additional 10-fold UF and the conservative estimates of total dietary intake provide adequate margins of safety for infants and children. If an even more conservative TRI were desired, another 10-fold childrens' UF could be applied.

FDA Response: FDA agrees with Reviewers #2 and #3. It may not be feasible to determine the risk in "all" populations and "all" situations, but FDA understands that the circumstances indicating the greatest chance of toxicity should be determined. FDA will also re-assess the S/RA recommendations in terms of the safety factor, as recommended by Reviewer #5, when the results of the toxicity studies of melamine and/or its analogues combined are determined and available.

FDA is developing protocols to examine the potential of melamine being excreted in milk, and there are other, non-FDA sponsored studies ongoing to assess this possibility. These studies should address the concerns raised by Reviewer #4 regarding the potential exposure to nursing infants via mother's breast milk.

In terms of Reviewer #3's view that the level of detection/quantitation is sufficiently developed to enable the FDA to lower the acceptable level of melamine and/or its analogues in food products from 1 to 0.25 ppm, FDA would be required to support that decision with a sufficiently strong argument based on toxicity data. When toxicity results are available from the ongoing studies looking at the combined toxicity for melamine and cyanuric acid, FDA will carefully re-evaluate its present S/RA.

Charge Question 5: Are there additional scientific/ technical studies available and pertinent to this risk assessment that were not considered?

  • Reviewer #1. The search engines available to me failed to identify other scientific/technical studies that are directly relevant to the risk assessment as identified in this report. However, melamine can be a neoplastic agent and that aspect of risk was not addressed in this report. There are reports of a dietary substance that may have a mitigating effect on the development of melamine induced preneoplastic lesions.
  • Reviewer #2. The literature has been reviewed and the available data summarized.
  • Reviewer #3. a. Reimschuessel et al. Evaluation of the renal effects of experimental feeding of melamine and cyanuric acid to fish and pigs. Am J Vet Res 69:1217-1228, 2008. This study demonstrates the lack of renal pathology induced by these compounds alone versus the extensive renal crystal formation which occurred in both species when these compounds were given simultaneously.

    1. Osborne CA, et al. Melamine and cyanuric acid-induced crystalluria, uroliths, and nephrotoxicity in dogs and cats. Vet Clin Small Animal Practice 39:1-14, 2008. This study describes findings at a large veterinary Urolith center, discusses the findings of unique uroliths in dogs and cats beginning in 2002. These were identified as uric acid monohydrate; recent investigation suggests that these are actually melamine cyanurate. In addition, an article describing similar unique uroliths occurring in people and also diagnosed as uric acid monohydrate was published in 2005 (see reference below).
    2. Schubert G, et al. Uric acid monohydrate - a new urinary calculus phase. Urol Res 2005; 33:231-8. The scientific techniques used to identify uroliths (typically make use of the physical characteristics of the crystals compared to known standards) are different than the scientific techniques used to identify toxic chemicals such as melamine cyanurate.
    3. Dobson RLM et al. Identification and characterization of toxicity of contaminants in pet food leading to an outbreak of renal toxicity in cats and dogs. Toxicol Sci 106(1):251-262, 2008. In this study, rats were given up to 100 mg/kg ammeline or ammelide alone, or a mixture of all 4 compounds (ammeline, ammelide, melamine, and CYA). Neither ammeline or ammelide alone produced any renal effects, but the mixture produced significant renal damage with intratubular crystals of melamine-cyanuric acid. Provides further evidence that these two apparently innocuous compounds result in adverse effects in combination due to the formation of an insoluble precipitate (melamine cyanurate) in renal tubules causing progressive tubular obstruction and degeneration. The dose of 400 mg/kg was based on extrapolation of the dose in the pet food associated cases - the total fraction of triazines in the contaminated gluten was 10-13%, wet cat food assumed to contain up to 10% gluten, following manufacturer's feeding guidelines the dosage of melamine/CYA were calculated to be 360-430 mg/kg/day. This paper also discusses chemical interactions between melamine, CYA, ammeline, and ammelide), and discusses how the melamine/CYA crystals may behave once ingested. The authors also comment that the nature of the toxicity may pose a challenge for traditional risk assessment methods, which are based on assessment of the toxicity of individual compounds. "The dose-response curve and NOAEL for the mixture is almost certainly different from that of either compound alone."

     
  • Reviewer #4. Some additional references are included, but I assume the FDA has already reviewed these. The references highlight the fact that melamine:cyanurate formed at a stoichiometric ratio of 1:1 and that while the agglomerates are birefringent, an x-ray diffraction of them was not possible without some additional chemical modifications. In our lab, analysis of "crystals" from kidneys of animals that died from the pet food issue of 2007 revealed they were birefringent, yet attempts at acquiring x-ray diffraction revealed they were amorphous. Perhaps this isn't significant for the food safety concern, but it does remind us that agglomeration of melamine and cyanuric acid is an equilibrium process with hydrogen bonding and not covalently bonded as with most crystals, and is fully reversible. In that regard, it is good to keep in mind that such an equilibrium can be affected by the local environment such as temperature, pH, ionic strength, etc., and that agglomeration between melamine and cyanuric acid could form in different organs of the body depending on those conditions. This consideration led to our lab's thinking on melamine and cyanuric acid agglomeration in 2007. In addition, the formation of melamine:cyanurate agglomerates in vitro can occur in the ppb concentration range (see point 12).

    Perhaps this consideration can be useful as future research is carried out in this area. Further, the literature reveals a high temperature study was carried out demonstrating no impact on the agglomeration of melamine and cyanuric acid.

  • Reviewer #5. I have integrated a number of additional publications into my comments and listed them separately under Additional References.

FDA Response: Reviewers #1, 3, 4 and 5 provided references or potential topics to be considered. All the provided references will be reviewed and FDA will consider them in the revision melamine and its analogues S/RA if they are pertinent and/or raise issues that have not been discussed in the S/RA.

FDA is aware of the existence of both of the studies demonstrating that melamine can be a neoplastic agent, and that melamine neoplasia can be mitigated by a dietary substance. A discussion of the impact of chronic triazine exposure will be provided in the revision of the S/RA for melamine and its analogues.

FDA agrees with the data provided by Reviewer #4 indicating that because melamine-cyanuric acid complexes are dissociable, various induced changes on the renal environment should be studied. This might be helpful in increasing the rate and completeness of triazine excretion since this could be valuable for the possible prevention or remediation of crystal accumulation and/or kidney/bladder stone formation in animal and human subjects.

Charge Question 6: Is the no observed adverse effect level (NOAEL) an appropriate point of departure for calculating the margin of exposure? Do the data support the value and endpoint chosen?

  • Reviewer #1. The validation of the use of the NOAEL as an appropriate point of departure for calculating the margin of exposure is best assessed by the toxicologist who will review this report. In general, NOAEL is the level of exposure of an organism at which there is no biologically or statistically significant increase in the frequency or severity of any adverse effects in the exposed population when compared to a non exposed population. As noted, my interpretation of this is that any adverse effect is included. The potential for melamine to cause cardiac problems is not addressed in the S/RA. Whether or not it is appropriate to include the risk of the development of neoplastic lesions from melamine exposure is an assessment best made by the toxicologist reviewer.
  • Reviewer #2. Yes, the NOAEL is an appropriate application for calculating the margin of exposure. The data support the value and endpoint chosen.
  • Reviewer #3. The available data are insufficient to calculate NOAEL. As discussed in 3c, toxicologic studies using melamine/CYA in combination are required.
  • Reviewer #4. Yes to both questions, considering our best information at this time. Given the toxic endpoint is considered the point at which melamine and cyanuric acid can sufficiently agglomerate thereby creating obstructions in the kidney, the NOAEL appears appropriate. Therefore, it is currently an appropriate point of departure for calculating the margin of exposure. My comment to consider pertains to the purity of melamine and analogues. For example, in developing the NOAEL, the assumption is that essentially pure melamine was used to adulterate the various foods and that melamine and analogues are equally "toxic;" however, not including the potential toxicity of melamine and analogue interactions, a complete assessment of the potentially used waste stream from melamine production may not have included other toxicants, e.g., heavy metals or inorganics, PCBs, dioxins, and others. There doesn't appear to be any reference(s) available that discusses this issue. While it could create a complexity to the S/RA, addressing it at least indicates awareness of it, which I'm sure the FDA is, and also a potential action plan to study it.
  • Reviewer #5. As noted above, I was unable to locate the critical study and evaluate the data.

FDA Response: FDA found no descriptions of cardiac damage in the numerous studies that have been performed with melamine, as discussed by Reviewer #1. Long term studies have been conducted in rats and mice (NTP 1983) in which a complete necropsy and microscopic examination of lesions found in any of the organs was done. There was no reference to any kind of cardiac toxicity due to melamine in those studies. Likewise, most studies of cyanuric acid toxicity also do not report cardiac lesions. In one study of guinea pigs, however, researchers reported changes after cyanuric acid exposure "dystrophy of the heart muscle" in some of the animals. This is not, however, a finding reported in other species. Mazaev, 1962

Mazaev, V.T. (1962) On the question of the content of cyanuric acid and its monovalent sodium salt in the water of reservoirs. Gigiena i Sanitaria, 27: 13-19.

FDA agrees that the assessment for neoplastic potential should be evaluated and added to the overall assessment.

FDA understands Reviewer #3 and #4's comments that data are insufficient for calculating a NOEL for combined exposure to melamine and one or more of its analogues. FDA is currently conducting preliminary experiments to determine a NOEL in experimental animals. This information will be used when revising the risk assessment

FDA agrees with Reviewer #4 that the risk assessment does not incorporate other non-triazine toxicants in the assessment. That was beyond the scope of the current risk assessment. There are no plans at present to do studies on the combined toxicity of melamine and its analogues with other environmentally persistent chemicals, such as, PCBs, heavy metals, dioxins and others. It would be reasonable for our analytical chemists to perform additional analyses to evaluate the possible presence and concentrations of these agents in infant formula or other food products. There are, however, stringent requirements for purity of food ingredients that are added to food products, e.g., food-grade quality of components as indicated by the Food Chemicals Codex standards. There is no evidence to date that other substances were implicated in the toxicity noted in pets during 2007 or in the children exposed in China in 2008.

Reviewer #5 commented that they were unable to locate the critical study. The study in the published National Toxicology Program (NTP) report can be found on the web at http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr245.pdf.

Charge Question 7: Were scientific assumptions that are not strictly linked to the data, such as the assumption that 50% of the diet was contaminated, explained and appropriate?

  • Reviewer #1. The calculation that if 50% of the diet could be contaminated with 2.5 ppm of melamine and its analogues and the daily intake would be 0.063 mg/kg bw/day is clear. There are many assumptions inherent in this calculation and the reason these assumptions were made is not clear. For example, the average diet was set at 3 kg/day, the contamination was 3.78 mg, a 60 kg individual was used and 50% of the diet was contaminated (why assume only 50% is contaminated?). The thinking is circular.
  • Reviewer #2. Yes, this assumption was explained and is an appropriate conservative assumption.
  • Reviewer #3. Assumption 1: Melamine present at 2.5 mg/kg concentration. This concentration of melamine in formula was cited as being greater than 2500 ppm in some milk samples. The highest reported concentration should be used, so a higher concentration may be more appropriate. Levels of melamine in wheat gluten or rice protein should also be reviewed to ascertain possible levels in these materials.

    Assumption 2: A dietary assumption level of 50% was indicated, but an explanation for how this figure was derived was not given.

    Assumption 3: That a 10-fold safety factor applied to the TDI is sufficient to account for the "increased toxicity results" from combined exposure to melamine and cyanuric acid. This assumption is flawed - toxicity results from the interaction of the chemicals, it is not a reflection of increased toxicity of two toxins but is due to creation of a toxin due to mixing of two nontoxic compounds. The basis of determination of the TDI and NOAEL (which is equal to the lethal dose of the combined chemicals) for melamine alone and attempting to apply this to the toxic combination of chemicals is incorrect.

  • Reviewer #4. Yes, those scientific assumptions were clearly explained and appropriate. If adulteration were to occur, it would require a significant source of melamine and analogues and a rationale of the need to adulterate target products for the benefit of those organizations facilitating the adulteration for the purpose of spoofing the Kjeldahl analysis. Further, the range and extent of potential products that could be adulterated appears comprehensive.

    If adulteration were to occur with the common protein based "products" that were identified by the FDA in the S/RA and where those products could make up to 50% of a person's daily diet, the general public should agree that the FDA had appropriately considered their diet with the corollary that they are being protected. For infants, given their underdeveloped kidney function, the largest practicable safety factor should be considered; particularly at this time since there is insufficient information to assess the potential long-term damage to those infants as they mature.

  • Reviewer #5. The assumptions in the risk assessment were justifiably conservative under these particular circumstances and therefore appropriate.

FDA Response: Reviewers 2, 4 and 5 expressed their views that the assumptions used were appropriate. Reviewers 1 and 3 stated that some of the assumptions used in the calculations in the S/RA were not clearly explained. For the sake of clarity, we will further explain our reasoning for the following: 1) the assumption that an adult's total daily dietary intake of solid and liquid food is 3 kg; 2) the assumption that 50% of the diet contained melamine; 3) the use of an adult body weight of 60 kg; 4) that a dietary intake of 3.78 mg of melamine and its analogues per person per day does not raise public health concerns; 5) the assumption that certain milk and milk-derived food ingredients contained melamine at a level of 2.5 mg/kg; and 6) the assumption that a 10-fold safety factor applied to the TDI is sufficient to account for the "increased toxicity results" from combined exposure to melamine and cyanuric acid.

  1. 1) The assumption that an adult's total daily dietary intake of food is 3 kg.

    The choice of 3 kg of food as an adult's daily intake of food is a conservative assumption as it is reflects the average consumption of adult males aged 25-30 years, who typically represent the subpopulation with the highest total daily dietary food consumption. This value is derived from information published in an article detailing the revision of the food list used in FDA's Total Diet Study (TDS) (Pennington, J.A.T. (1983) Revision of the Total Diet Study food list and diets. J Am. Diet. Assoc., 82, 166-173.) Using data from the 1977-78 USDA Nationwide Food Consumption Survey (NFCS), the Second National Health and Nutrition Examination Survey (NHANES II), and the list of 234 foods contained in the revised TDS, the article estimated total daily food intake for 8 population groups. Of the 8 population groups, males aged 25-30 years showed the largest daily food intake of 3,075 grams of food per day. Rounding this number down to 3 kg yields a conservative estimate of daily food intake for adults. As this number is based on older data, we recently performed a similar calculation for males aged 25-30 years using data from the 2003-2004 NHANES survey. Using all food categories in the 2003-2004 NHANES survey yielded an average daily intake of food for males 25-30 years of age of 2,950 grams. Thus, we believe that it is valid to conservatively estimate a total daily dietary intake of food for adults of 3 kg.

  2. 2) The assumption that 50% of the diet contained melamine.

    At the time of the development of the S/RA, the primary focus of melamine and melamine analogue contamination was in milk and milk-derived ingredients. As such, it did not seem appropriate to assume that 100% of the diet was contaminated with melamine or its analogues, as 100% of the diet would not contain milk or milk-derived products. We presumed that the 50% contamination assumption would cover milk and milk-derived products and foods containing them as ingredients, yet allowed enough conservatism to cover other products as well. To verify that milk and milk products account for less than 50% of an individual's daily dietary intake, we performed an average intake estimate for 25-30 year old males using milk and dairy food consumption data from the 2003-2004 NHANES survey. We estimate that less than 15% of the average 25-30 year old males' diet is comprised of milk and dairy products. Since the focus of the S/RA was milk and dairy products, we also considered the intake of the population subgroup with the highest proportion of their diet from milk and dairy products. Based upon the available data, 2 year olds consume a higher proportion of their daily diet as milk and dairy products compared to other population subgroups. Using 2003-2004 NHANES data we estimate that 33% of the diet of the average 2 year old is from milk and dairy products. In addition, data from our sampling of food products in the domestic food supply generally indicate that melamine is found at extremely low levels in products, including infant formula, that are consistent with background levels from approved uses of melamine. For example, from the use of sanitizers (see FDA Response to Charge Question 8, below). These data indicate that the prevalence of melamine in the U.S. food supply is extremely low. Considering all the data presented above, our assumption that 50% of the diet contains melamine is sufficiently conservative to account for melamine intake from milk and dairy products as well as other foods that could contain melamine.

  3. 3) The use of an adult body weight of 60 kg.

    As a first approximation, FDA chose 60 kg to act as the surrogate body weight (bw) for all life stages from a 2 year-old toddler to the eldest consumer. However, the peer review of the FDA 2008 melamine S/RA expressed concern that this approach may not have been adequate to protect young children. Based on these concerns, we have performed an additional exposure estimate focused on young children. The results of this second analysis confirmed that our first approach is adequate for establishing a level of melamine in foods below which does not raise public health concern.

Due to the potential that young children may consume a higher portion of their total diet from milk and milk products than other segments of the population (e.g., adults), we performed an intake estimate that focused on actual food consumption data rather than the broader assumption of the total amount of food consumed. Specifically, estimates were performed to determine the potential melamine intake from the consumption of milk and dairy products for five population subgroups (2 year olds, 3 year olds, 4 year olds, 5 year olds, and 2-5 year olds). Estimates were performed using 2-day food consumption data from the 2003-2004 National Health and Nutrition Examination Survey (NHANES). Foods included in the intake estimate were: fluid milk (including evaporated and condensed); yogurt; flavored milk and milk-drinks; milk-based meal replacements; creams and cream substitutes; milk-based desserts, sauces and gravies (including ice cream and pudding); and cheeses. Only those individuals that consumed any one or all of the target foods on both survey days were included in the estimate; therefore, this estimate represents an "eaters-only" intake. Mean food intakes from the consumption of milk and dairy products are presented in Table 1, below, on a grams of food consumed per person per day (g/p/d) basis, as well as a grams per kilogram body weight per day (g/kg bw/d) basis.

 

Table 1. Mean food intake for young children from consumption of milk and dairy products based on 2003-2004 NHANES data
Population SubgroupAvg. Body Weight
(kg/person)
Percent of Population consuming one or more of the foods included in the estimateMean Food Intake
(g/p/d)
Mean Food Intake
(g/kg bw/d)
2 year olds13.7399.2%480.235.14
3 year olds16.298.1%431.726.64
4 year olds18.298.7%455.825.01
5 year olds22.399.5%479.521.47
2-5 year olds17.298.7%464.727.08

In the S/RA, the TDI/10 for melamine was determined to be 0.063 mg/kg bw/d. The S/RA estimated that one-half of the diet of a 60 kg person consuming 3 kg of food per day would need to contain 2.5 mg/kg melamine in order to reach the TDI/10. We have calculated the average intake of melamine for each of the population subgroups assuming that all of the milk and dairy products consumed contain melamine at a level of 2.5 mg/kg. The results of these calculations are shown in Table 2. As an example, the calculation for the 2-5 year-old population group is shown below:

Melamine intake assuming 2.5 mg/kg contamination in food =
= (mass of food consumed per day x melamine level) / (body weight x 1000 g/kg) 
= (464.7 g/p/d x 2.5 mg/kg) / (17.2 kg bw x 1000 g/kg) = 0.068 mg/kg bw/d

 

Table 2. Melamine intake for young children based on average food intake from consumption of milk and dairy products
Population SubgroupPortion of fluid milk as a percentage of the total average amount of milk and dairy products consumed4Average melamine intake assuming that milk and dairy products contain 2.5 mg/kg melamineFraction of the TDI/105
2 year olds90%0.089 mg/kg bw/d1.4
3 year olds83%0.067 mg/kg bw/d1.1
4 year olds78%0.063 mg/kg bw/d1.0
5 year olds83%0.054 mg/kg bw/d0.86
2-5 year olds82%0.068 mg/kg bw/d1.1

Thus, the average melamine intake from milk and dairy products containing melamine at a level of 2.5 mg/kg for the population subgroups of 3, 4, and 5 year olds, and the combined 2-5 year olds is essentially equal to the TDI/10 for melamine that was stated in the S/RA. For 2 year olds, the intake is slightly greater than the TDI/10, however, we note that it is highly unlikely that 100% percent of milk and dairy products consumed by and individual over an extended period of time would contain melamine at a level of 2.5 mg/kg in foods sold in the United States.

The recent incident in China was associated primarily with contaminated fluid milk and infant formula manufactured from melamine contaminated milk. Fluid milk is not traded internationally and no infant formula manufactured in China may be legally sold in the United States (U.S.). As part of our response to this incident, FDA confirmed with U.S. infant formula manufacturers that they do not source dairy ingredients from China. In addition, FDA, along with State and local authorities, inspected Asian markets in the U.S. and confirmed that no infant formula produced in China was being sold to consumers. Thus, the source of melamine contamination in foods associated with the recent incident in China that potentially could be sold in the U.S. was from dairy products or dairy derived ingredients and not fluid milk or infant formula. Our subsequent analysis of samples of various foods (including infant formula and fluid milk) offered for sale in the U.S. confirmed this conclusion. Furthermore, as shown in Table 2, fluid milk accounts for the majority of the intake of milk and dairy products consumed by young children. Therefore, the assumption that 100% of the milk and dairy products would contain melamine at a level of 2.5 mg/kg is overly conservative, as fluid milk is not imported for consumption in the U.S., and would therefore not be expected to be adulterated under the current situation.

  1. 4) The assumption that a dietary intake of 3.78 mg of melamine and its analogues per person per day does not raise public health concerns.

    The establishment in the S/RA of 3.78 mg of melamine and its analogues as a level that does not raise public health concerns is a direct result of the use of a tolerable daily intake/10 (TDI/10) of 0.063 mg/kg-bw/d and the assumption of a 60 kg adult body weight. Multiplying these two numbers together yielded a value of 3.78 mg that was determined to be a level that did not raise public health concerns, as stated in the S/RA.

  2. 5) The assumption that certain milk and milk-derived food ingredients contained melamine at a level of 2.5 mg/kg.

    It is true that melamine levels higher than 2.5 mg/kg were found in certain foreign food products. One of the goals of the S/RA was to determine a contamination level in food that would not raise a public health concern. As such, dividing the melamine intake level deemed not to raise a public health concern for adults (3.78 mg) by the total weight of food estimated to contain melamine (1.5 kg) yielded a melamine contamination level of 2.5 mg/kg. At the time of the development of the S/RA, milk and milk derived ingredients were the primary foods of concern in terms of contamination with melamine and its analogues. As a result, it was deemed important to verify through a separate method of intake estimation (in this case U.S. per capita intake of milk-derived food ingredients) that a level of 2.5 mg/kg of melamine and its analogues would not raise public health concerns. Table 1 of the S/RA shows that approximately 1% of the TDI/10 is reached for a 60-kg person consuming milk-derived food ingredients containing melamine and its analogues at a level of 2.5 mg/kg.

  3. 6) The assumption that a 10-fold safety factor applied to the TDI is sufficient to account for the "increased toxicity results" from combined exposure to melamine and cyanuric acid.

    FDA agrees that for situations where co-ingestion of melamine and one of its analogues occurs, the NOAEL and previously-derived TDI may not be appropriate. The results of the toxicity tests in animal models for combined melamine and cyanuric acid exposure will be helpful in supporting any update of the TDI, necessary in future updates to the risk assessment.

Charge Question 8: Are the exposure scenarios addressed (infant formula and consumption of milk-derived non-infant formula food ingredients) representative and comprehensive, considering the public health risk evaluated?

  • Reviewer #1. The infant formula exposure scenario is clear. Formula (for non breast fed infants) is the sole source of nutrition for 4-6 months of life and some organ systems are not fully matured at this age. The S/RA notes that under these conditions, a level of melamine and its analogues that does not raise public health concerns cannot be estimated for infant formulas.

    The scenario for those not consuming infant formula is less clear. Please see response to question number 2 above. In addition, the conclusion that levels of melamine and its analogues below 2.5 ppm in foods other than infant formula do not raise public health concerns is based solely on the estimate of melamine exposure from milk. The report does not address the possibility that the matrix of foods other than milk might be different and the risk might change because of the matrix. Further, the report notes that trichloromelamine, a compound that readily decomposes to melamine, is approved for use as a sanitizing agent on food processing equipment and utensils except for milk containers and equipment. Thus, it is likely that milk and milk derived foods have a lower contamination of melamine than other food products.

  • Reviewer #2. Yes, they are appropriate starting points for exposure scenarios and may be modified as additional information is acquired on the risk and effects of melamine and its analogues exposures.
  • Reviewer #3. The exposure scenarios are not comprehensive because a reason for the contamination was not discussed and therefore, foods at risk were not identified. Two factors determine risk and the potential severity of disease once adulteration of the food occurs. The first risk factor is likelihood of eating a particular adulterated food. Foods that would be predicted to be adulterated are those in which financial worth of the product is based on protein content. Therefore, all processed foods or ingredients obtained from China (and perhaps elsewhere) in which protein content is recorded are potentially affected. These include all products listed in the table in the report and those additional products listed in the answer to question 1. The second risk factor is associated with the level of exposure/amount of melamine/CYA ingested. High risk groups who would be expected to eat these types of food and in significant quantities, include vegetarians (who would ingest more vegetable-based protein than the general population), infants, those who might eat the same type of cereal daily (particularly children), and people taking powdered protein supplements. In addition, people on high protein diets would be more likely to eat products with added protein material. Essentially all mixed-source agricultural feed and all pet foods are at risk for contamination, as protein content is a factor in pricing of final product and ingredients.
  • Reviewer #4. Yes, the exposure scenarios are representative and comprehensive. A question I have is it possible that economic factors could require individuals to use protein based supplements that are less costly than infant formula such as skim milk powder to feed their infants? Have we considered breast milk as a vehicle that could expose infants to melamine? Knowing that melamine is not biotransformed and is readily excreted, if a mother was exposed to melamine at the proposed safe levels, could that melamine migrate to breast milk? Our main concerns to date have been to mitigate "crystal" formation in kidney, given our best information indicates that neither melamine nor cyanuric acid appear to be toxic to certain animal species.
  • Reviewer #5. The infant exposure scenarios were representative and reasonable.

FDA Response: Reviewer #1 indicated that the level of melamine and its analogues in food that does not raise public health concerns (<2.5 mg/kg) was based solely on the estimate of melamine exposure from milk. We do not believe that this statement accurately reflects the S/RA. The melamine level of 2.5 mg/kg was determined using the 2007 estimate of the TDI for melamine and its analogues as a starting point dividing this by an additional safety factor of 10 yielding a TDI/10 and the presumption that one-half of a person's total daily dietary intake contains melamine and its analogues. Reviewer #1 stated that the safety and risk assessment did not address the possibility that the risk may change based on the matrix of foods other than milk. As noted in our response to Reviewer #1 in Charge Question 1, we assumed in the safety and risk assessment that 100% of melamine and its analogues present in the food were available for exerting its effect, and that the safety and risk assessment therefore accounts for any matrix variations that might exist among different foods. Reviewer #1 was also concerned that foods other than milk and milk derived foods could contain a greater level of melamine due to the use of trichloromelamine as a sanitizing agent. Although trichloromelamine can be used as a sanitizer on processing equipment and utensils for food and beverages, and therefore provide another source of melamine in food, it is not approved for use on milk containers and equipment. The conditions of use for trichloromelamine require that there is adequate draining of the sanitizing solution (i.e., trichloromelamine) prior to contact with allowed food products. Thus, only residual levels of trichloromelamine would be in contact with food. We believe that our presumption that 50% of an individual's diet contains melamine is adequately conservative to address incidental exposure, as any potential exposure from the use of trichloromelamine as a sanitizing agent would be significantly lower than that from direct melamine contamination of food.

Reviewer #3 indicated that the exposure scenarios were not comprehensive because a reason for the contamination was not discussed and as a result, the at-risk foods were not identified. We agree that the S/RA did not directly consider all possible exposure scenarios. At the time that the S/RA was developed, the source of melamine introduction into food was not completely understood. FDA assumed that the reason for the contamination incident was attempts by food producers and distributors to fraudulently treat their product so that it would appear (test) as being superior. FDA assumed that everyone that could be exposed would be exposed; both infants and adults. As a result we chose to perform what we considered to be a conservative intake estimate based on the assumption that 50% of an adult's daily food intake contained melamine and its analogues.

In addition, Reviewer #3 was concerned about vegetarians and other high risk groups that would be expected to consume a greater amount of foods containing vegetable-based protein (e.g., wheat gluten). As noted above, we conservatively assumed that everyone would consume food containing melamine and that 50% of an individual's daily diet contained melamine. We believe that this would provide a sufficiently conservative intake estimate to account for vegetarians and other groups that could have an increased intake of foods containing melamine. As indicated above (see Charge Question 7), less than 15% of an average 25-30 year old male's diet is comprised of milk and dairy products. This leaves 35% of an individual's diet to potentially contain melamine and its analogues from the consumption of foods other than milk and milk products. However, data indicate that melamine or cyanuric acid is present only at extremely low levels in U.S. products that do not appear to have been formulated with any adulterated ingredients. While vegetarians may have the potential to consume a greater amount of food that could be contaminated with melamine and its analogues (e.g., foods containing melamine contaminated wheat gluten) compared to the general population, a vegetarian's diet is still sufficiently widely varied to include foods that would not contain melamine and its analogues. Therefore, we believe our assumption that 50% of the diet contains melamine and its analogues is sufficiently conservative to account for a vegetarian diet. If further information becomes available, especially regarding a vegetarian diet, we can refine our intake estimate accordingly.

FDA agrees with Reviewer #4 that the excretion of melamine into milk (either human or animal) is a major knowledge gap. The ability for melamine or it's analogues to be secreted in milk has not been documented. FDA is currently preparing protocols to examine excretion into milk in ruminants.

Charge Question 9: What are the prioritized research needs in relation to melamine and melamine related compounds, including the research that is needed to reduce the uncertainty associated with the S/RA?

  • Reviewer #1. 1. Identify an appropriate test animal. Rats are traditionally used and have offered important information for the development of this report. However, the report clearly states that species specific rates of elimination of the compound can account for some of the differential toxicity. Do rats excrete melamine at the same rate as humans? Would a primate, pig or other animal be more appropriate source of information on excretion rates?

    1. There is some information on the effect of high doses in rats for a relatively short period of time and for dogs for a longer time. What is the long term effect of smaller doses of melamine for long periods of time? Are rats or dogs the best test animal to use?
    2. What is the actual food content of melamine in the US food supply? Are there different levels in different foods? Relative to the melamine content of other foods, where does milk lie?
    3. Besides cyanuric acid and other melamine analogues are there other substances in the environment or food that melamine might interact with to potentiate or ameliorate it s toxicity?
    4. How is the toxicity of melamine different during different times of life (growing infants, children, teenagers, pregnant and lactating women, the elderly, those with a chronic disease, etc.)? Does the estimate of safety/risk change for these more vulnerable populations?

     
  • Reviewer #2. Additional animal data are needed to define appropriate models that take into account age-specific factors relating to renal function and development and the differences in drug metabolism and elimination. The models should also include studies on biomarkers of acute kidney injury such as the role of NGAL and KIM-1 in identifying early injury. If at all possible, obtaining biosamples from humans exposed to these agents during the recent contaminations would permit some analyses for biomarkers that might serve as a template for future investigations.
  • Reviewer #3. a. Acute and chronic toxicologic studies to determine the effects of a combination of melamine and cyanuric acid in order to determine safety guidelines.

    1. If possible, development of protein analytical tests that would measure only protein N, thereby abrogating the impetus for adding melamine and cyanuric acid to food.
    2. If these analytical tests cannot be developed, toxicologic assays to test for the presence of both melamine and cyanuric acid prior to entrance of any ingredients/food into the food supply.

     
  • Reviewer #4. Suggest characterizing the potential sources of melamine that could be used for adulteration. All implications relate to "pure" melamine or analogues. It would be useful to know that the waste from the production of melamine-based products, if indeed it is used to adulterate food, does not contain toxicants. Heavy metals screens, PCBs, Dioxins, Furans, etc. are suggested. This risk assessment relies heavily on the assumption that only melamine or cyanuric acid and/ or analogues would be used as chemical adulterants for the purpose of protein supplementation, not a waste by-product made up of various ingredients. Is it possible that commercial grade melamine would be used to adulterate food instead of its intended application for manufacturing products? What about the quality of water in the manufacturing process? Are Chinese based water production systems free from known toxins and pollutants not permitted in the US present that could end up in the left over residue?

    1. Consider studying breast milk for melamine and analogues, particularly for those potentially exposed to melamine at the proposed level for adults. It may very well be that the FDA has already done this or considered it. A dog study could help us understand the potential for melamine excretion via breast milk.
    2. Studies related to cytochrome P-450 activity with and without melamine, cyanuric acid, in combination with the two, etc. are warranted. Specifically, it would interesting to note whether the activity increased, decreased, or did not change in the presence of melamine:cyanurate agglomerates. Is it possible those agglomerates can act as a catalyst due to large surface area or as an adsorptive site where reactive intermediates can be protected, thereby increasing their apparent half lives? For example, do melamine and/ or cyanuric acid and/ or melamine:cyanurate agglomerates have any effect, synergistic, scavenging or no effects for known toxicants, such as pyrolysis products realized from frying foods? Can a complex form among them thereby increasing the half life of pyrolysis products, such as with TRP-1, etc? Research into the stability of reactive intermediates in the presence of melamine, cyanuric acid and melamine cyanurate agglomerates might be considered. Is it possible that micro-agglomerates of melamine:cyanurate can form in the lung, where the surface of those agglomerates can act as adsorptive sites, for example in considering a Langmuir-Rideal mechanism? In this case, one or more molecules can bind to the agglomerate, accumulate and then react with oxygen in the lung. The products could then be released from the surface into the blood. Can melamine:cyanurate form in the liver and catalyze reactions there as well?
    3. Develop in vitro models where melamine is located in a bulk solution and cyanuric acid solution is located in a permeable tube. What are the kinetics of crystal formation and what factors do they depend on? Perhaps a rapid bioassay could be developed from this work using the urine of patients or test subjects where the formation of crystals on a permeable tube filled with cyanuric acid or melamine would be dipped in urine samples, allowed to sit within the urine for a specified amount of time, removed, soaked in 95% ethanol then analyzed with a polarized light microscope. This could help us quickly ascertain whether melamine and/or cyanuric acid were present in the diet. The method wouldn't work for excreted micro crystals in this simple assay.
    4. The question about renal function and accumulation of potential melamine and cyanuric acid in infants and toddlers may not be too dissimilar than considering renal clearing of drugs, specifically significant would be the renal plasma flow, glomular filtration and active secretion. It could be assumed then that people on medication, such as diuretics could fall outside the norm considered in the S/RA. Perhaps clearing rate studies for infants and for those people on specific medications and with a diet high in processed protein products could help yield insight regarding their risk.
    5. We understand that melamine:cyanurate agglomerates are dissociated at low pH in stomach, absorbed, distributed, not metabolized, then eliminated. What about the pH of urine? What about individuals with a very high pH, low fluid diet? Can they "accumulate" melamine and cyanuric acid even at the safe level?
    6. Can melamine be detected as a metabolite of any pesticide residues? Perhaps cyromazine usage could illuminate this point. If a mother was exposed to cyromazine, could melamine be found in her breast milk?

     
  • Reviewer #5. A top research priority is to conduct dose-response renal toxicity studies in animals given melamine and cyanuric acid orally in combination. The most important need is accurate determination of an acute NOAEL and LOAEL for melamine cyanurate, to use in an updated and more accurate risk assessment. A subacute gavage and a feeding experiment should also be performed to characterize the nephrotoxic potential of melamine cyanurate ingested over a period of days and weeks. Toxic potency should be established/contrasted in two or three species, though reports to date indicate that fish and mammals (including humans) are similarly affected. Age-dependency is another research priority to assess in an appropriate animal model. Epidemiological investigations of afflicted populations in China may yield useful information about age- and dose-dependency. Detailed toxicokinetic studies in animals will be needed to establish relationships between different doses and dosage regimens, kidney concentrations of the toxicants, and the magnitude of crystalluria and nephrotoxicity. Different combinations of melamine and other analogues should be similarly studied. Finally, mechanism(s) of renal injury should be examined, in order to provide information for subsequent studies to provide ways to prevent and to treat the malady.

FDA Response: The reviewers generally recommended 5 areas where research is needed to reduce the uncertainty:

  1. Determine the acute and chronic toxicological effects of the combination of melamine and cyanuric acid, including a no effect level (NOEL) dose.
  2. Determine how age (infants/elderly), metabolic variations (pregnancy/lactation) and disease states (chronic renal disease/medications) affect toxicity of the individual or combined compounds.
  3. Identify appropriate animal models and biomarkers to study the toxicity, especially considering the potential variations listed in #2.
  4. Determine the actual content of melamine and related compounds in US food supply.
  5. Develop better analytical tests for proteins in foods to abrogate the impetus for adding melamine and analogues to foods.

During 2007 and 2008, FDA scientists began a number of studies to provide more information to risk assessors.

Area 1: Acute and chronic toxicological effects, NOEL determination

In 2007, FDA's scientists demonstrated that fish and pigs exposed to a 1:1 combination of melamine and cyanuric acid produce renal crystals morphologically and chemically identical to those in cats that died from contaminated pet food (Reimschuessel et al 2008). Those studies were expanded in 2008 to use fish as a non-mammalian model to determine a NOEL dose. Fish have been shown to be good non-mammalian models for renal injury (Reimschuessel 2002) and by using them as a first tier model we are able to reduce the number of mammals used in early stages of this research.

Using techniques and exposure protocols similar to those developed in the fish study, FDA scientists are conducting a pilot NOEL study in weanling pigs and young rats. A preliminary 7 day NOEL dose has been identified in pigs and will be bracketed in a larger study in spring 2009. The pilot rat study will be conducted at FDA in spring - early summer 2009.

FDA will also begin a comprehensive multi-year NTP study of melamine and it's analogues in rats. This study is designed to refine and confirm the NOEL pilot data and evaluate chronic effects of exposure to the individual and combined compounds.

In addition, since melamine was found in fish feeds in 2007, and there is a general lack of information about the depletion of melamine from edible tissues, FDA began a study in 2008 to determine the depletion of melamine in 2 species of fish. The data from this study will be submitted for publication in spring 2009. These data will help address the potential accumulation of melamine and cyanuric acid in tissues.

Area 2: Age, metabolic variations, diseases (potentially sensitive populations)

The ongoing pilot pig study at FDA is using weanling animals. The pilot study in rats at FDA will also use juvenile animals. More information on age, metabolic variations and disease effects is definitely needed. Studies to address transfer of melamine during lactation and pregnancy are under consideration. FDA is also aware of several lines of investigation being done at University laboratories and welcomes researchers to contact us about their study protocols and any data they wish to share. In addition, epidemiologic information from China may help our understanding of the effects of melamine with respect to populations of differing ages and disease states, for example, Guan et al. 2009 report that preterm infants were more likely to develop stones than full term infants.

Area 3: Special animal models and biomarkers

FDA studies are currently using non-mammalian fish models, weanling pigs and immature rats. The recent melamine episode in China occurred in infants and young children and was primarily due to the dilution of milk and the addition of fairly pure melamine alone. The primary effect was the formation of renal and urinary tract stones composed of a combination of melamine and uric acid. It would be desirable to have information from animal models that have elevated uric acid concentrations. Humans and non-human primates have serum uric acid concentrations ten to twenty times higher than other mammals because they lack the enzyme uric acid oxidase (Fanelli et al. 1974, Watanabe et al. 2002). Normal uric acid concentrations in infants are higher than those of older children or adults (Stapleton 1983). Therefore, it will be important to investigate the comparative effects of melamine alone and in combination with its analogs in animal models with elevated uric acid. Potential models include the rat uric acid nephropathy models (Conger et al 1976, Conger 1990, Stravric et al 1969, Stavric and Nera 1978), or animals that lack uric acid oxidase such as Dalmatian dogs (Roch-Ramel et al 1976, Berger and Yu 1970), or primates (Fanelli and Beyer 1974, Johnson 2002). FDA welcomes additional information from researchers currently using such models.

FDA also agrees that identification of early markers of renal injury or subtle changes due to melamine alone or in combination with its analogues would also be desirable. The NTP study being planned at FDA will evaluate a number of markers of renal injury, such as KIM-1, β2-microglobulin, Cystatin C, Clusterin and Trefoil Factor-3. FDA is also conducting studies to examine the conditions for crystal formation and dissolution in vitro. The effects of pH will also be examined.

Area 4: Determine content of melamine in US foods.

Since 2007, FDA has tested a number of U.S. products that were targeted based on the potential use of melamine-adulterated ingredients from China. In cases where elevated concentrations of melamine were detected, these products have been recalled. Only extremely low levels (< 1 ppm) of melamine or cyanuric acid have been detected in U.S. products, such as infant formula, that do not appear to have used any adulterated ingredients from China. We agree that the background level of melamine in the U.S. food supply should be determined. In order to address this issue, studies are currently underway to assess the levels of melamine in certain food products. However, given the lack of any significant findings of contamination, except in foods associated with the use of adulterated Chinese ingredients, FDA believes it should continue to focus its resources on analyzing food samples that are targeted to the Chinese contamination incidents rather than the entire U.S. supply. As further resources become available, we can expand our focus to include the entire food supply.

Area 5. Develop better analytical tests for proteins in foods

We agree that such tests would be very valuable to help food producers monitor the quality of their products. FDA is participating in the planning of a United States Pharmacopeia (USP) workshop on food protein adulteration that will focus on trying to identify improved tests for proteins in foods and food ingredients.

In addition to the 5 major areas there were also questions and concerns raised by the individual reviewers.

Reviewer 1 voiced concern about environmental issues and FDA agrees that any change in the environment or various host factors may potentially increase or decrease the toxicity of these substances. Future study results may shed some light on these issues.

Reviewer 4 asked if contaminated melamine products could have been used to adulterate foods. FDA definitely has considered this a possibility. This may be where the dual exposure (melamine and cyanuric acid) in pet food products began in the 2007 Chinese wheat gluten adulteration incident. With regard to the water used in China to prepare or dilute the milk or formula, this information is currently unavailable. The effects of pesticides such as cyromazine and their metabolism and subsequent disposition of melamine in human breast milk are currently unknown

Reviewer 4 also suggested studies be conducted on P-450. FDA believes that there is limited potential for these substances to affect liver metabolizing enzymes because they undergo little or no metabolism themselves before being excreted.

Reviewer 4 asked if agglomerulates of melamine-cyanuric acid can increase their half-lives and/or form in the lung. FDA is not aware of any such effect. The pathology of lungs from rats exposed to melamine and cyanuric acid was examined in Dobson et al; no crystals were found by light microscopy.

Berger L, Yü TF. Urinary excretion of uric acid after instantaneous intra-renal-arterial injection in mongrel and Dalmatian dogs. Mt Sinai J Med. 1970 37(4):351-8.

Conger JD, Falk SA, Guggenheim SJ, Burke TJ. A micropuncture study of the early phase of acute urate nephropathy. J Clin Invest 1976; 58:681-9.

Conger JD. Acute uric acid nephropathy. Med Clin North Am. 1990. 74(4):859-871.

Fanelli GM, Beyer KH. 1974. Uric acid in non-human primates with special reference to its renal transport. Ann. Rev. Pharmacol. 14: 355-364

Guan N, Fan Q, Ding J, Zhao Y, Lu J, Ai Y, Xu G, Zhu S, Yao C, Jiang L, Miao J, Zhang H, Zhao D, Liu X, Yao Y. Melamine-Contaminated Powdered Formula and Urolithiasis in Young Children. N Engl J Med. 2009 Feb 4. [Epub ahead of print]

Reimschuessel, R. A fish model of renal regeneration and development. (ILAR), National Research Council, National Academy of Sciences, Washington, D.C. Inst. Lab. Animal Resources, 42: 4: 285-291. 2002

Reimschuessel, R., Gieseker, C., Miller, R.A., Rummel, N., Ward, J., Boehmer, J., Heller, D., Nochetto, C., De Alwis , H., Bataller, N., Andersen, W., Turnipseed. S. B., Karbiwnyk, C. M. Satzger, R D., Crowe, J., Reinhard, M.K., Roberts, J.F., and Witkowski, M. Evaluation of the renal effects of experimental feeding of melamine and cyanuric acid to fish and pigs. Am. J. Vet. Res 69(9):1217-28.. 2008

Roch-Ramel F, Wong NL, Dirks JH. Renal excretion of urate in mongrel and Dalmatian dogs: a micropuncture study. Am J Physiol. 1976 231(2):326-31.

Stapleton FB. Renal uric acid clearance in human neonates. J Pediatr. 1983, 103(2):290-4.

Stavric B, Johnson WJ, Grice HC. Uric acid nephropathy: an experimental model. Proc Soc Exp Biol Med 1969; 130:512-516.

Stavric B, Nera EA. Use of the uricase-inhibited rat as an animal model in toxicology. Clin Toxicol. 1978;13(1):47-74.

Watanabe S, Kang DH, Feng L, Nakagawa T, Kanellis J, Lan H, Mazzali M, Johnson RJ. Uric acid, hominoid evolution, and the pathogenesis of salt-sensitivity. Hypertension. 2002 Sep;40(3):355-60.

Charge Question 10: Are you aware of any data, not discussed in the S/RA that would assist FDA in further characterizing the risk associated with contamination of infant formula by melamine and its analogues?

  • Reviewer #1. I have not identified additional data specific to the charge, as I understand it, for this document. I did find the Berkeley site, http://potency.berkeley.edu/chempages/MELAMINE.html, of interest.
  • Reviewer #2. No. However, a review of both animal models and human data regarding acute kidney injury in newborn and developing animals and infants, especially the premature ages, would be a logical next step in determining what populations are most at risk from these agents.
  • Reviewer #3. Urolith analysis suggests that melamine cyanurate may have been misdiagnosed as uric acid monohydrate (or that melamine cyanurate acts as a nidus for additional urolith formation) in dogs, cats, and people, possibly for a number of years. If uric acid monohydrate uroliths are indeed associated with melamine cyanurate, then it would seem that subclinical exposure might lead to urolith formation in people. This emphasizes the need for chronic low level toxicity studies in addition to acute toxicity experiments.

    Schubert G, et al. Uric acid monohydrate - a new urinary calculus phase. Urol Res 2005; 33:231-8.

  • Reviewer #4. No, I am not aware of any data that could further assist the FDA in further characterizing the risk associated with contamination of infant formula by melamine and analogues.
  • Reviewer #5. See response to charge question 5 and specific comments.

FDA Response: Reviewer #3 focused their comments on the urolith analysis. Urolith analysis of the stones found in children in China exposed primarily to melamine alone indicates that the stones are a combination of melamine and uric acid (Sun et al 2008, Shen et al. 2008). These stones appear to be similar to those which form in rats given high doses of melamine for prolonged periods of time (Heck and Tyl 1985, Ogasawara et al 1995).

The crystals formed in pets during the 2007 recall were primarily caused by ingestion of both melamine and cyanuric acid (Dobson et al 2008, Reimschuessel et al 2008). In general, stones were not associated with that episode, although a previous case of feed contamination may have contributed to stone formation (Brown et al. 2007, Jeong et al 2006, Yhee et al, 2009). The exact composition of the contaminants in that case was not known, but tissues contained melamine and cyanuric acid.

FDA welcomes additional information regarding melamine and stone formation in humans and animals. FDA will also review the references provided by Reviewer #5 under Additional References for relevance and new information, and add those which are relevant.

Brown CA, Jeong KS, Poppenga RH, Puschner B, Miller DM, Ellis AE, Kang KI, Sum S, Cistola AM, Brown SA. Outbreaks of renal failure associated with melamine and cyanuric acid in dogs and cats in 2004 and 2007.J Vet Diagn Invest. 2007 19(5):525-31.

Dobson, R. L. M., Motlagh, S., Quijano, M., Cambron, R. T., Baker, T. R. Pullen, A. M., Regg, B. T., Bigalow-Kern, A. S., Vennard, T, Fix, A., Reimschuessel, R. , Overmann, G., Shan, Y, Daston, G. P. Identification and characterization of toxicity of contaminants in pet food leading to an outbreak of renal toxicity in cats and dogs. Toxicological Sciences 106:251-262. 2008.

Heck, HD'A, Tyl, RW. The induction of bladder stones by Terephthalic Acid, Dimethyl Terephthalate, and Melamine (2,4,6-Triamino-s-triazine) and its relevance to risk assessment. Regulatory Toxicology and Pharmacology 5: 294-313. 1985

Jeong WI, Do SH, Jeong da H, Chung JY, Yang HJ, Yuan DW, Hong IH, Park JK, Goo MJ, Jeong KS. Canine renal failure syndrome in three dogs. J Vet Sci. 2006 Sep;7(3):299-301.

Ogasawara H, Imaida K, Ishiwata H, Toyoda K, Kawanishi T, Uneyama C, Hayashi S, Takahashi M, Hayashi Y. Urinary bladder carcinogenesis induced by melamine in F344 male rats: correlation between carcinogenicity and urolith formation. Carcinogenesis 1995; 16:2773-2777.

Shen Y Division of Urology, Department of Pediatric Surgery, Beijing Children's Hospital Affiliated to Capital Medical University, Beijing 100045, China (personal communication) Sun N, Shen Y, Sun Q, Li XR, Jia LQ, Zhang GJ, Zhang WP, Chen Z, Fan JF, Jiang YP, Feng DC, Zhang RF, Zhu XY, Xiao HZ. Melamine related urinary calculus and acute renal failure in infants. Zhonghua Er Ke Za Zhi. 2008;46(11):810-5.

Yhee JY, Brown C, Yu CH, Kim JH, Poppenga R, Sur JH. Retrospective Study of Melamine/cyanuric acid-induced Renal Failure in Dogs in Korea between 2003 and 2004. Vet Pathol. 2009 Jan 21. [Epub ahead of print]

Charge Question 11: Do the S/RA results objectively support the conclusions with respect to the risks to the general population, or to infants and toddlers whose renal systems may not be fully developed, from consumption of melamine-contaminated foods? Are there other sensitive subpopulations that FDA should address in its S/RA?

  • Reviewer #1. Please see my responses to numbers 1 and 2 above.
  • Reviewer #2. Yes; the S/RA has taken into account this vulnerable population at risk for acute kidney injury secondary to consumption of melamine-contaminated foods. Another sensitive subpopulation that should be evaluated is lactating mothers who may also be placing the nursing offspring at risk of injury from melamine and its analogues. Is there any information on the transfer of these agents via breast milk?
  • Reviewer #3. The risks to infants and toddlers are clearly documented. The fact that these infants were exposed to large quantities of melamine/cyanuric acid (as formula would be their major dietary constituent) is likely the major reason for the observed outcome. Renal "immaturity" may not be as important of a factor, although critical review of the age distribution of these affected children would better answer this question.

    While anyone eating processed foods with ingredients from China (where most adulteration has occurred and the addition of melamine is thought to be a common practice) has some risk (not yet defined), vegetarians, children eating large amounts of cereal, and people on high protein diets (including in particular those ingesting protein supplements) are potentially at higher risk.

  • Reviewer #4. Yes, the S/RA objectively supports the conclusions for adults, but the question about renal function and development in infants must be weighed more heavily in developing the S/RA for infants. An additional safety factor of 4 is suggested, from 1 ppm to 0.25 ppm, thus allowing for the possibility mother's milk may contain melamine and analogues and given the possibility that a waste fraction from melamine production systems could be used to adulterate protein based foods. Further, by ensuring the safe level is 0.25 ppm, labs will then be required to demonstrate a method detection limit of at least 0.050 ppm thereby further assisting the regulatory program in protecting the potentially most sensitive individuals in the population. Realizing a method detection limit of 0.050 ppm (50 ppb) melamine in food products, such as infant formula, is not a challenge for current techniques and analytical instrumentation. While this exercise may appear academic, it will create greater confidence for analytical data generated at 0.25 ppm. Finally, the possibility that melamine:cyanurate agglomerates can be formed in vivo (see below) even at low concentrations should be considered an additional reason to increase the safety factor.
  • Reviewer #5. This S/RA provides a minimum of experimental detail and little scientific support for its choices of critical studies and endpoints, metabolism and toxicokinetics, pathology and functional changes, and intra- or interspecies variability. It is a "bare bones" RA, possibly due to haste with which it was prepared? Might persons (adults) with preexisting renal dysfunction be another sensitive subpopulation?

FDA Response: Regarding the questions raised by Reviewer #5 on the scope of the S/RA, FDA did prepare these S/RAs under the pressure of constrained time; however, most of these studies are the same ones that have been repeatedly selected as the basis for S/RA of melamine and/or cyanuric acid by other international organizations, such as the World Health Organization, the European Food Safety Agency, and the Canadian Health Protection Branch. FDA believes that much of the impression of incompleteness with respect to the overall S/RA is due to the fact that it has only been recently recognized worldwide that humans are potentially at risk from the unregulated, illegal use of these substances as unapproved food additives. Thus, there was minimal relevant and scientifically rigorous data available regarding the exposure to melamine and/or its analogues, either singly or from exposure to two of these compounds in food at the same time. Most experimental studies were conducted with quite elevated concentrations or in only a few animals.

Reviewer #2 questioned the possibility of the transfer of melamine via breast milk from exposure of the mother to melamine-contaminated foods. FDA is developing a protocol to determine whether melamine is secreted into bovine milk and if so, to what extent.

For Reviewer #1's comments, please see FDA's response to Charge Questions 1 and 2 above. For Reviewer # 4's comments, please see FDA's response to Charge Questions #2 and #4 above.

Charge Question 12: Do you have any additional comments that would assist FDA in refining the S/RA?

  • Reviewer #1. No.
  • Reviewer #2. Overall, the reports are appropriate summaries of the current state of our knowledge regarding the risk for exposures to melamine and its analogues in adults and infants. However, the approximations are based on limited animal data that do not reflect the potential renal toxicity of combined exposures and also the vulnerable infant population with limited renal capacity to excrete these products. Research is needed in animal models of the newborn and developing age groups, and also verification of biomarkers of renal injury from these agents would assist in identifying those humans at risk after exposure.
  • Reviewer #3. No.
  • Reviewer #4. Analytical Standards and Validated Methods:
    In our laboratory, experiments were carried out using separate 50 ppb solutions of melamine and cyanuric acid in acetonitrile. The individual solutions were injected into an LC-MS/MS, with clear detection of both compounds (MRM transitions, data not shown). Contents of the two separate vials were then mixed (500µL from each), and the mixture was vortexed, and then injected, resulting in 25 ppb of each melamine and cyanuric. Neither compound was detectable following injection (Figure 1, TOP chromatogram). Following that, one drop of neat formic acid was added to the mix, and the solution was injected (Figure 1, BOTTOM chromatogram). Both melamine and cyanuric acid were fully detected, at approximately half of the original response. This was a little surprising in that we didn't expect to see agglomeration at such a low concentration of each compound in acetonitrile, indicating that hydrogen bonding between those molecules is thermodynamically more favorable than their free solubility in solution at what could be considered infinite dilution. But, this also suggests that perhaps melamine:cyanurate agglomerates can form at significantly lower than expected concentrations in vivo, which may very well be harmless for adults but may be more serious for infants.

    Also, when prescribed (reported approach to preparing melamine / cyanuric acid mixed standards) working standard concentrations of melamine and cyanuric acid were mixed in acetonitrile:water (50:50) (20 ppm), visible particles were detected, that were birefringent. Scanning electron microscope (SEM) images of some of those particles are presented in Figure 2. Besides the obvious fact that melamine:cyanurate agglomerates can form at low concentrations of melamine of cyanuric acid, I also see this information as an opportunity to develop a rapid screening test for melamine and cyanuric acid from samples, where polarized light microscopy can be used to screen out suspect positives. It may not work in all matrices, but it seems to work for infant formula. More on this below.

    Points to consider from the above:

    1. We have consistently used acid to prepare standards of melamine and cyanuric acid to ensure a complex between melamine and cyanuric acid is not formed (we have been doing so since first developing our method in May 2007, data not published). If there is a lab preparing mixed standards of melamine and cyanuric acid, even at 25 ppb, then a check on whether a complex is formed should be carried out by simply adding one drop of formic acid to the vial and establishing whether a larger signal is realized for melamine and cyanuric acid relative to the analysis without acid. Our analytical standards were from a very common supplier located in the US.
    2. That without the addition of acid to the extracting solvent, mixtures of melamine and cyanuric acid won't be readily extractable, yielding artificially low apparent concentrations of melamine and cyanuric acid, and potential analogues. This feeds in to point 11 of the charge questions.
    3. That it is possible to create particles of melamine and cyanuric acid, in vitro, that are not spiked in shape, while in the presence of organic solvents, in the 10's of ppm. This is interesting since we have not seen spiked agglomerates in vivo and perhaps part of the reason for that observation is related to interaction with hydrophobic surfaces. Certainly the overall shape of the agglomerates almost suggests micellar structures were formed.

     
  • Reviewer #5. No.

FDA Response: Regarding Reviewer #2's recommendation for research evaluation of the combined effects of melamine and its analogues and the effects of age and lactation - please see FDA's response to Charge Question 9 on research needs.

Reviewer #4 provides some useful analytical information that has also been noted by other researchers concerning formation of melamine:cyanurate agglomerates at low concentrations in organic solvents such as acetonitrile and in combination with water. The addition of acid ensures that complexes are not formed between melamine and cyanuric acid. We agree that analytical methods need to be evaluated to ensure that they will completely extract melamine and cyanuric acid from the food or product being analyzed, and to ensure that their mixed melamine and cyanuric acid standards do not form an insoluble complex in their solvent.

SPECIFIC OBSERVATIONS

  • Reviewer #1. Please note, my copy was not paginated and so I added page numbers

    1. Page 4, SAFETY/RISK ASSESSMENT FOR INFANT FORMULA, paragraph 3, point number 3 states that "The possibility that these formulations can be fed as the sole source of nutrition to premature infants with immature kidney function and even greater intake of infant formula per unit body weight for a longer time period than term infants." While this statement is accurate, it suggests that the renal function of term infants is mature. Although the glomular filtration rate is adequate for healthy term infants, it does not approximate adult rates until about three years of age. Renal tubular reabsorption and urine acidification is less at birth, and for several months thereafter, than for adults. This immature function of the term infant kidney can contribute to fluid and electrolyte abnormalities in infants who are sick or fed an inappropriate diet (Goldsmith DI, Novello AC. Clinical laboratory evaluation of renal function. IN: Edelman CM (editor) Pediatric Kidney Disease. Little Brown Co. 1992, Boston, p 471)

      The nephrologist who reviews the document would be better positioned to comment on this, but I would like to suggest that the wording be changed to reflect that kidney function of the term infant is also immature.

    2. SAFETY/RISK ASSESSMENT FOR FOOD AND FOOD INGREDIENTS OTHER THAN INFANT FORMULA, page 5, paragraph one.

    The second from last line should read: "...tolerable daily intake (TDI)..."

    Additional Comments on Update Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans:

    1. I re-reviewed the document, Update Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans, and the comments above also apply to this document. For example, my response to charge question 7 is applicable to this document. The S/RA makes several assumptions about weight and food consumption for infants and those estimates are not consistent with actual infant weight and formula consumption. The average birth weight of healthy term, infants in the US is ~3 kg (NCHS 2000 growth curves) and by 6 months of age the average healthy, term infant weighs ~7 kg. One hundred fifty grams (0.150 kg) of powdered infant formula is approximately 34 oz of a 20 Kcal/oz formula. A typical 3 kg infant would be consuming ~16 oz (100 Kcal/kg/day). A typical infant born weighing 3 kg would gain ~2 kg over the first 4 months of life and ~4 kg over the first 6 months of life. Most nutritionists would recommend that solids not be introduced until ~4-6 months.

      The document acknowledges that for infants who are not breastfed, infant formula is the sole nutrient source for these many months and the infant would have no exposure to other foods; the formula would be the sole nutrient source. This is a critical concept.

    2. The assumptions used to calculate risk are just assumptions and not based on data, save a rat study. While I did not have access to the SINGLE study on which all the estimations of risk are made, most likely the rats in that study were adult rats, not infant rats, see response to query number 1.
    3. This update notes there is a "...high degree of uncertainty with regard to the determination of safety/risk. Given these conditions, FDA has applied an additional 10-fold safety factor…" There is no data to support this additional safety factor. A 10-fold safety factor could be a valid assumption, or for infants, the combination of melamine and its analogues might pose a risk that is a thousand fold higher and the 10-fold safety factor is inadequate. There is simply no information to allow a scientific guess.
    4. Importantly, and I did not address this in the original report, is the lack of acknowledgement of the possibility of accumulation of melamine alone or in combination with its analogues (or other environmental factors) in the immature infant system and how that possibility should be considered in the estimation of risk.
    5. This document does not acknowledge that there may be other substances in the infant environment that might potentiate (or mitigate) the risk of melamine and its analogues in the infant diet.
    6. Finally, none of the information that I reviewed addressed what "does not raise public health concerns" actually means. Is the assumption that "does not raise public health concerns" is synonymous with risk?
  • Reviewer #2. Research should be directed at:

    1. Developing animal models of renal toxicity and/or acute kidney injury secondary to melamine and its analogues that include newborn and developing species.
    2. Information needs to be obtained on identifying the risk factors for acute kidney injury following exposure to melamine and its analogues. These include verification of biomarkers of acute kidney injury such as NGAL and KIM-1.
    3. Examining the toxicology of melamine and its analogues in lactating models and humans for potential risks from contaminated breast milk.

     
  • Reviewer #3. Despite a clear understanding of the interaction between melamine and cyanuric acid in causing nephrotoxicosis in multiple animal species, the human medical community and food regulatory industry has continued to concentrate on the risk and effects of melamine alone. Consequently, the risk was assumed to be minimal. It was not until the recent outbreak of renal disease associated with consumption of infant formula that it was considered that serious disease would also result if people ingested the same toxins that were ingested by animals. Even with the milk associated toxicosis, melamine but not cyanuric acid was cited as the causative toxin, and discussions in the media are still directed at melamine alone. Media reports concentrated on relatively benign "stones", which is misleading as melamine-CYA stones form when crystals which pass through the renal tubules (causing tubular obstruction and injury which, if severe, is irreversible) and accumulate with other material in the renal pelvis and urinary bladder. In general, renal physiology is not significantly different between people, dogs, rats, and cats (and these animals are frequently used as animal models of human disease or to assess toxicosis), so to assume that compounds causing significant mortality and renal disease in these species would not result in similar disease in people is scientifically unsound.

    While the current S/RA report does not adequately address the human health risks of melamine/CYA, the decision to have this document reviewed by a panel of unbiased senior scientists is applauded as a step in the right direction. Suggestions put forth by these reviewers should result in needed studies to better define the problem, assess risks, and help identify strategies to prevent melamine/CYA from entering the food chain.

  • Reviewer #4. ND, LOD (MDL) and LOQ:
    For the analytical aspects of the reports, it would be a good idea to clarify the concepts around limit of detection (LOD). I found explanations for both the limit of detection and limit of quantitation indicated in the documents, both suggesting the same definition "the LOD is the level below which a contaminant cannot be measured with current analytical methods." It is my understanding that the "not detected" (ND) level, which is lower than the limit of detection, is the point at which an analyst can truly assert the absence of a target analyte and that at, and below, the ND an analyst cannot, as an example, use signal averaging to increase the signal to noise ratio of a potential target analyte signal to a discernable level. Up from there, the limit of detection is a statistically estimated term related to the "ND" (usually 3 times the average noise of multiple blank injected matrix samples) but where uncertainty around that level is quite high, often resulting in 50% RSD for spikes at the LOD and where recovery values are unpredictable. That isn't to say data at the LOD will be used, rather it proves that when the LOD is realized, a false negative will be realized about 50% of the time. We have found that use of LC-MS/MS can result in the possibility that ions (from an eluted and ionized compound) below the statistically estimated limit of detection can be identified, yet not quantified, given the correct tR and MRM transitions and their ratios. This sometimes occurs given the large uncertainty around the limit of detection. Certainly, analysts would agree that a fixed method detection limit cannot be established and that it can be higher or lower on any given batch of analyses on any given day.

    An example where there appears to be contradiction of this in the documents is presented on page 15 of the S/RA report, paragraph two, line 6, where it is indicated that the concentration "ranged from 9 to 12 ppb". If the LOD as defined is 10 ppb and is argued as "the point below which the compound cannot be identified", then reporting 9 ppb would not have been possible. Also, on page 22, Appendix III, paragraph 1, lines 9 and 10, another term "lower limit of detection" is now mentioned. All references to limit of detection in all of the documents need to be clarified and harmonized.

    Suggestion: Perhaps it would be also useful to use the term Limit of Quantitation (LOQ), which is 3.3 times greater than the LOD. I appreciate the U.S. EPA's approach to estimating concentrations near the ND level and suggest that perhaps the documents could see value in those approaches.
    LOD issues, some specific pages:
    Page 4: The LOD is the level below which a contaminant cannot be measured with current analytical methods. The LOD statements are all over the map.

    Page 15 of the S/RA, 2nd paragraph, "a method has been validated for poultry tissue at the 50 ppb." Expansion of this comment is necessary.

    Odds and ends comments and words in referring to LOD, LOQ, measured, validated, accurately quantified:

    Page 22 of the S/RA, 2nd last line, "at a lower limit of detection of 10 ppb."

    Page 14: "conservative estimate of the LOD," yet statement before indicates "and while the method may be able to detect lower levels, the 50 ppb values is used in the current document (and the S/RA) as "conservative LOD." Should reword to better indicate the fact 50 ppb was the highest level of melamine in those samples.

    Page 4, "the LOD is the level below which a contaminant cannot be measured with current analytical methods." Does the word "measured" indicate a general idea around identification and confirmation with quantitative analyses? Need one strategy developed for limit of detection throughout the documents.

    Page 20 of the report, Table A-4. The LOD and the LOQ have been used interchangeably at this point. The LOD is not the limit of quantitation and neither is the LOQ; need some clarification to help sort out the concepts.

    Page 22, "...none of the fish were found to contain melamine at a lower limit of detection of 10 ppb."?

    Page 15: accurately quantified, should read quantified. The limit of quantitation (LOQ) is 3.3 times greater than LOD which is 3 times greater than ND.

    Page 16, "validation level of the melamine assay of 50 ppb." Does not make sense, need to correct it.

    Page 17, "or private labs." Were those labs required to meet the standards of the FDA methods? Did they use methods based on GC-MS or LC-MS/MS?

    I noticed there was no mention in any of the documents regarding the uncertainty in describing the limit of detection. Uncertainty around limit of detection / method detection limit can overlap uncertainty at the "safe level" if the LOD / MDL is not at least 5 times lower than the safe level.

    S/RA assessment for infant formula, 3rd paragraph, 6th line "not mitigated by previous passage through the digestive tract of an animal." Not a clear statement. Does this give a potentially greater exposure than would be truly realized?

    Use of acid for extraction and standards preparation:

    If numerous labs did not consider the issue, then round robins or proficiency testing would not reveal an issue with trueness amongst those labs. Of course, this isn't a concern for high concentrations, since the difference between 1,000 ppm and 2,000 ppm would not be significant in light of the proposed safe levels, but reporting levels of 1 ppm or less, with a less than adequate extraction and analysis with mixed standards would be significant; the corollary is that the target threshold for safety could be compromised and that only those data labs using acid in their extraction solvent and standards should be considered.

    Significant Figures:

    Throughout the document, in the various tables, there is not a consistent approach to reporting the data in consideration of significant figures. For example, in Table A-1, results are reported up to 4 significant figures for cyanuric acid, yet in Table A-6 results are reported with 1 to 3 significant figures. Suggestion: round up or down as required, while considering confidence levels for those values, which will help with data presentation.

    Positive vs. Negative results:

    In those cases where results are reported as positive or negative, it may be useful to consider a magnitude of "positive" - a semi-quantitative value. Further, is negative below 10 ppb? What is the limit of quantitation or limit of detection in those cases where samples were reported Negative?

    Bioaccumulation

    Comments in the document indicate levels of melamine dropped over time as the animals were withdrawn from sources of melamine contaminated feed. This statement (interestingly) implies some level of bioaccumulation through chemical potential, but does suggest that withdrawal from contaminated food products is necessary. Is it possible that the bioaccumulation, albeit very low in concentration, is realized by melamine:cyanurate agglomerate formation in the tissue?

    Additional references to consider

    Hydrothermal Synthesis of Organic Channel Structures: 1:1 Hydrogen-Bonded Adducts of Melamine with Cyanuric and Trithiocyanuric Acids. Ranganathan, Anupama; Pedireddi, V.R.; Rao, C.N.R. Chemistry Physics of Materials Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, India. Journal of the American Chemical Society (1999), 121(8), 1752-1753. Publisher: American Chemical Society, CODEN: JACSAT ISSN: 0002-7863. Journal written in English. CAN 130:267044 AN 1999:109783 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    Aq. solns. of cyanuric acid and melamine were mixed in a Teflon flask, and the mixt. was kept in a stainless steel bomb. The bomb was sealed and maintained in a furnace at 180°. Rectangular platelike crystals of good quality sepd. from the soln., upon cooling the bomb to room temp.over a period of 4h. X-ray crystallog. data were detd. for the cyanuric acid- melamine (1:1) complex. Packing anal. shows that cyanuric acid and melamine are held together by N-H O and N-H N hydrogen bonds yielding the hexamer (rosette). The hexamers are arranged in two dimensions to form planar sheets, the structures being exactly as predicted by G.M. Whiteside (1991, 1995). A significant feature is that the planar sheets are stacked in three dimensions to give channels with a diam. of 4 .ANG., comparable to the cavities in cryptands. Crystal structure of 1:1 trithiocyanuric acid-melamine complex, obtained by a similar procedure reveals features similar to the cyanuric acid-melamine complex.

    Supramolecular Membranes. Spontaneous Assembly of Aqueous Bilayer Membrane via Formation of Hydrogen Bonded Pairs of Melamine and Cyanuric Acid Derivatives. Kimizuka, Nobuo; Kawasaki, Takayoshi; Hirata, Kiyoko; Kunitake, Toyoki. Department of Chemistry and Biochemistry Graduate School of Engineering, Kyushu University, Fukuoka, Japan. Journal of the American Chemical Society (1998), 120(17), 4094-4104. Publisher: American Chemical Society, CODEN: JACSAT ISSN: 0002-7863. Journal written in English. CAN 129:86491 AN 1998:270289 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    Amphiphilic hydrogen bond networks consisting of alkylated melamines and ammonium headgroup-appended cyanuric acids are stably dispersed in water as supramol. membranes. Electron micrographs of these aq. dispersions indicated the formation of supramol. assemblies of mesoscopic dimension. Their aggregate morphologies, mol. orientation, and thermal characteristics are markedly dependent on the chem. structure of constituent mols. Self-supporting multilayer films were obtainable by casting the aq. dispersions, like the conventional aq. bilayer. X-ray diffraction of the cast films indicated that hydrogen bonded pairs of ammonium-appended cyanuric acid and double-chained melamine adopted the bilayer structure. On the other hand, complementary pairs formed from single-chained melamines adopted partially or completely interdigitated bilayers. The long period of the former cast film (ca. 9 nm) is in good agreement with the thickness of disklike aggregates obsd. in electron microscopy, indicating that bilayer structures mediated by complementary hydrogen bonds are maintained in water. Thermal characteristics of aq. dispersions were investigated by differential scanning calorimetry and spectroscopically by using 1,6-diphenyl-1,3,5-hexatriene as a fluorescence probe. The obsd. spectral characteristics indicate that supramol. membranes display phase transition from a highly ordered state to a liq. cryst. phase. In addn., reversible dissocn. and irreversible segregation of complementary pairs proceeded at higher temps. These supramol. membranes are the first example of water-sol. supermols. directed by complementary hydrogen bonds and constitute a new family of amphiphilically designed supramol. assemblies.

    Estimating the Entropic Cost of Self-Assembly of Multiparticle Hydrogen-Bonded Aggregates Based on the Cyanuric Acid•Melamine Lattice. Mammen, Mathai; Shakhnovich, Eugene I.; Deutch, John M.; Whitesides, George M. Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA, USA. Journal of Organic Chemistry (1998), 63(12), 3821-3830. Publisher: American Chemical Society, CODEN: JOCEAH ISSN: 0022-3263. Journal written in English. CAN 129:86492 AN 1998:325352 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    The entropic component of the free energy of assembly for multiparticle hydrogen-bonded aggregates is analyzed using a model based on balls connected by rigid rods or flexible strings. The entropy of assembly, ΔS, is partitioned into translational, rotational, vibrational, and conformational components. While previously reported theor. treatments of rotational and vibrational entropies for assembly are adequate, treatments of translational entropy in soln. and of conformational entropy-often the two largest components of ΔS-are not. This paper provides improved ests. and illustrates the methods used to obtain them. First, a model is described for translational entropy of mols. in soln. (ΔStrans(sol)); this model provides phys. intuitive corrections for values of ΔStrans(sol) that are based on the Sackur-Tetrode equation. This model is combined with one for rotational entropy to est. the difference in entropy of assembly between a 4-particle aggregate and a 6-particle one. Second, an approx. anal. of a model based on balls connected by rods or strings gives an approx. est. of the max. contribution of conformational entropy to the difference in free energy of assembly of flexible and of rigid mol. assemblies. This anal., although approx., is easily applied by all types of chemists and biochemists; it serves as a guide to the design of stable mol. aggregates, and the qual. arguments apply generally to any form of self-assembly.

    Self-Assembly of Hydrogen-Bonded Polymeric Rods Based on the Cyanuric Acid•Melamine Lattice. Choi, Insung S.; Li, Xinhua; Simanek, Eric E.; Akaba, Ryoichi; Whitesides, George M. Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA, USA. Chemistry of Materials (1999), 11(3), 684-690. Publisher: American Chemical Society, CODEN: CMATEX ISSN: 0897-4756. Journal written in English. CAN 130:237254 AN 1999:98011 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    This paper describes the self-assembly of hydrogen-bonded polymeric rods based on the lattice of cyanuric acid and melamine (CA•M). Data from 1H NMR spectroscopy, IR spectroscopy, gel permeation chromatog. (GPC), and transmission electron microscopy (TEM) are interpreted as indicating that the self-assembly of a bisisocyanuric acid (bisCA) and a bismelamine (bisM) formed polymeric nanorods [(bisCA)n(bisM)n] composed of parallel CA•M rosettes. The TEM results suggest that these rods aggregate as bundles. The length of the bundles ranged from 100 to 1500 nm, and their diam. was in the range from 15 to 500 nm.

    Cyanuric acid and melamine: A platform for the construction of soluble aggregates and crystalline materials. Simanek, Eric E.; Li, Xinhua; Choi, Insung S.; Whitesides, George M. Harvard University, Cambridge, MA, USA. Editor(s): Sauvage, Jean-Pierre; Hosseini, M. Wais. Comprehensive Supramolecular Chemistry (1996), 9 595-621. Publisher: Elsevier, Oxford, UK CODEN: 63GFA6 Conference; General Review written in English. CAN 125:194740 AN 1996:510889 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    A review with .apprx. 50 refs.

    Design, preparation, and characterization of hydrogen bonded supramolecular aggregates based on the cyanuric acid-melamine lattice. Mathias, John P.; Simanek, Eric E.; Seto, Christopher T.; Whitesides, George M. Dep. Chem., Harvard Univ., Cambridge, MA, USA. Macromolecular Symposia (1994), 77(International Symposium on New Macromolecular Architectures and Supramolecular Polymers, 1993), 157-66. CODEN: MSYMEC ISSN: 1022-1360. Journal written in English. CAN 121:107878 AN 1994:507878 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    Symposium proceedings. Mol. self-assembly is described as a strategy for synthesizing large hydrogen-bonded supramol. aggregates that are based on the planar lattice formed of equimolar portions of cyanuric acid and melamine. The aggregates described in this paper contain four to ten individual mols., are stabilized by networks of 18 to 54 hydrogen bonds, and have mol. wts. in the range 2.7 to 6.5 kDa. Each of these aggregates has been characterized using 1H NMR spectroscopy, gel permeation chromatog., and vapor pressure osmometry. Phys.-org. chem. is used to (i) improve the authors' criteria for the design of self-assembling structures and (ii) develop techniques for characterizing non-covalently bound aggregates in org. soln.

    Self-organization via hydrogen-bridge bonds: preparation of a supramolecular aggregate from ten molecules. Mathias, John P.; Simanek, Eric E.; Seto, Christopher T.; Whitesides, George M. Dep. Chem., Harvard Univ., Cambridge, MA, USA. Angewandte Chemie (1993), 105(12), 1848-50 (See also Angew. Chem., Int. Ed. Engl., 1993, 32(12), 1766-9). CODEN: ANCEAD ISSN: 0044-8249. Journal written in German. CAN 120:298591 AN 1994:298591 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    The title aggregate was obtained from one mol. of a sym. 1,3,5-trisubstituted benzene deriv. contg. nine melamine units and nine mols. of isocyanuric acid deriv. I. NMR spectra and gel permeation chromatograms of the aggregate were shown. The aggregate exists in parallel rosettes with 54 stabilizing H bonds.

    1,3,5-trisubstituted benzene derivitive.

    Molecular self-assembly through hydrogen bonding: supramolecular aggregates based on the cyanuric acid-melamine lattice. Seto, Christopher T.; Whitesides, George M. Dep. Chem., Harvard Univ., Cambridge, MA, USA. Journal of the American Chemical Society (1993), 115(3), 905-16. CODEN: JACSAT ISSN: 0002-7863. Journal written in English. CAN 118:254892 AN 1993:254892 CAPLUS (Copyright (C) 2008 ACS on SciFinder (R))

    Abstract

    Reaction of the tris(melamine) derivs. hubM3 (C6H3-1,3,5-[CONHC6H4-3-N(CH2C6H4-4-CMe3)COC6H3-2-NHC3N3(NH2)(NHCH2CH2CMe3)-5-Br]3) and flexM3(C6H3-1,3,5-[CO2(CH2)3OCOC6H4-2-NHC3N3(NH2)(NHCH2CH2CMe3)]3) with R1CA (neohexyl isocyanurate) and R2CA (3,3,3-triphenylpropyl isocyanurate), resp., in CHCl3 yields structurally well-defined supramol. aggregates hubM3(R1CA)3 and flexM3(R2CA)3. These structures were characterized using 1H NMR, 13C NMR, and UV spectroscopy, gel permeation chromatog., and vapor pressure osmometry. FlexM3 is a conformationally flexible analog of hubM3. The greater degree of preorganization that is build into the mol. structure of hubM3 compared to flexM3 makes hubM3(R1CA)3 a more stable aggregate than flexM3(R2CA)3. These self-assembling structures are the first step in a program to design, synthesize, and develop methods to characterize supramol. complexes that are held together by networks of noncovalent interactions.

    Chromatogram: Melamine and cyanuric acid in acetonitrile, 25 ppb each, NO acid.
    Chromatogram: SAME vial as Top, but WITH 1 drop of formic acid added to the vial.

    Figure 1: Top chromatogram. Melamine and cyanuric acid in acetonitrile, 25 ppb each, NO acid. Bottom chromatogram, SAME vial but WITH 1 drop of formic acid added to the vial.

    SEM of Melamine: Cyanurate Agglomerates formed from mixing equal parts of melamine and cyanuric acid to 20 ppm each, in acetonitrile: water (50:50).

    Figure 2: SEM of Melamine:Cyanurate Agglomerates formed from mixing equal parts of melamine and cyanuric acid to 20 ppm each, in acetonitrile:water (50:50). The addition of formic acid, to 0.1 %, reversibly breaks down the agglomerates to melamine and cyanuric acid. The insert is for 6.00K magnification.

  • Reviewer #5. Background

    pgr. 3, lines 4 & 5: Is the estimated level 15 µg/kg expressed per kg of food? This should be stated to avoid confusion with an exposure (i.e., dosage) level of 15 µg/kg body weight. Is there an authoritative source that can be cited for this value? The accuracy of the value is important, as it is the baseline intake, to which amounts in contaminated foods will be added. It would be good to provide a range of baseline values and to state how reliable these data are.

    It would be worthwhile to provide a little background information on potential human exposure to cyanuric acid. Allen et al. (1982) state that cyanuric acid and chlorinated isocyanurates are added to water in swimming pools to stabilize the chlorine content to photochemical reduction, and thereby extend its bactericidal action. Canelli (1974) provides more information on the chemical, bacteriological and toxic properties of these additives.

    Toxicological Study Results

    pgr. 1: The S/RA report needs to be more comprehensive and provide more scientifically-rigorous support/documentation of key points. Studies conducted to date do indicate that melamine is very poorly metabolized by each species tested. This statement should be backed up by specific research findings and referenced. Mast et al. (1983), for example, reported that melamine was not metabolized by male F-344 rats. The plasma and urinary excretion half-lives were found to be 2.7 and 3.0 hours, respectively. In the only apparent human kinetics study, Allen et al. (1982) recovered 98% of ingested melamine as the parent compound, and determined the urinary excretion half-life to be ~ 3 hours. It is important here to point out: (1) the apparent toxicokinetic (TK) relevance of rats to humans; (2) the likelihood that melamine will not accumulate in either species under these experimental conditions.

    It is stated in line 6 & 7 that "Some species excrete melamine more slowly than other species." As I have noted above, this does not appear to be the case for mammals. This is pertinent when extrapolating animal toxicity testing data to humans. If melamine excretion in fish is relatively slow (cite a reference), the chemical might be anticipated to accumulate in the fish's tissues and thus present a potential problem if they are consumed by humans.

    It is stated in the paragraph's last line that "one of the bases for differential toxicity to these substances is species-specific rates of elimination." Again, the data I have located indicate that this is not the case for rodents and humans. There are, however, relatively few TK data for any species. TK and metabolism studies are needed in several species, including humans.

    pgr. 2, lines 3 & 4: There is little information on the toxicity of ammeline, ammelide and certain other melamine analogues. This is not true for cyanuric acid. Hammond et al. (1986) published a review of toxicological investigations of cyanurate and some of its chlorinated derivatives. Several research groups have recently assessed the acute and subacute nephrotoxicity of cyanuric acid alone and in combination with melamine in several species (Puschner et al., 2007; Reimschuessel et al., 2008). Also, Dobson et al. (2008) conducted renal toxicity experiments with a series of dosages of ammeline and ammelide in rats.

    pgr. 3, lines 1 & 2: I would not describe the work of Dobson et al. (2008) on the composition of melamine cyanurate crystals as preliminary. Thompson et al. (2008) have also recently analyzed the crystalline structure.

    pgr. 3, lines 2 - 4: It could be pointed out that melamine alone has been demonstrated to produce bladder stones and ensuing bladder carcinoma in rats and mice. These adverse effects required chronic dietary ingestion of very large amounts of the chemical (Melnick et al., 1984; Ogasawara et al., 1995; Okumura et al., 1992). It is worthy of note that melamine is non-genotoxic, and acts by causing chronic irritation and proliferative lesions.

    pgr. 3, lines 6 & 7: The last sentence pertains to crystalluria in cats, yet reference #2 described fish and pigs. The manuscript of Cianciolo et al. (2008) might be cited instead.

    pgr. 5: It is not sufficient to merely cite the IUCLID Dataset and the OECD SIDS Analysis as sources of the lowest LD50 and NOAELs. The page numbers should be specified, so readers can identify and read about the actual studies from which the values were taken. It should be noted that the selected studies were not published in the peer-reviewed literature, if this is the case.

    pgr. 5, lines 4 - 7: Several NOAELs are listed here, but the toxicity endpoints(s) is(are) not specified for each value. Were different indices monitored to yield the NOAELs or was crystalluria the sole parameter? Again, page numbers in the aforementioned documents should be included in the References. It is essential for readers to be able to examine/assess the critical study and supporting studies upon which safety/risk assessments are based.

    First Bullet: Again, citation of a long document or textbook is not adequate for readers to locate information on specific investigations. Was 1,200 mg melamine/kg given to the dogs? More data should be included about the findings of Dobson et al. (2008), as their's is one of the most definitive studies to date. The 64 mg/kg total dose was given to just one cat. Single cats also received total doses of melamine and cyanuric acid of 121 and 181 mg/kg. A number of dose-dependent adverse morphological changes were present in the kidneys. Nevertheless, the pathology findings of Dobson et al. must be considered as preliminary.

    Second Bullet: The question of mechanism(s) of melamine-cyanurate renal damage should be discussed in greater detail. Obstruction of kidney tubules by crystals may not be the sole mode of action. Dobson et al. (2008) stated that physical blockage was responsible for renal failure. They based their conclusion on the lack of cytotoxicity of melamine, cyanuric acid and other triazines (singly) on cultured dog and cat kidney cells. Other researchers (e.g., Puschner et al., 2007) have speculated that a combination of factors (e.g., inflammation, tubular obstruction by proteinoceous material secondary to casts, cell death and obstruction by the crystals) may be necessary for development of acute nephropathy.

    Third Bullet: It would be preferable to limit the references here to the NTP (1983) carcinogenesis bioassay and a second account of this bioassay by Melnick et al. (1984), as well as the subchronic investigations of Ogasawara et al. (1995) and Okumura et al. (1992). The title of the NTP report (Technical Report on the Carcinogenesis Bioassay in F344/N Rats and B6C3F1 Mice) should be included in its reference.

    Fourth Bullet: State which indices of toxicity were evaluated in the 13-week rat study and by whom the study was conducted. Was a LOAEL identified? What were the criteria for selecting this particular study over another?

    Fifth Bullet: Again, insufficient detail is given about this investigation and where to find it.

    Safety/Risk Assessment for Infant Formula

    pgr. 3, line 4: Several months do not constitute a chronic human exposure. It would be preferable to use the word subchronic.

    Last Sentence: Amend this sentence to reflect the updated infant safety assessment.

    Safety/Risk Assessment for Food and Food Ingredients Other Than Infant Formula

    pgr. 1: It should be pointed out that the customary safety, or uncertainty factor (UF) of 100 is comprised of an interspecies UF of 10 (in case humans are 10-fold more susceptible than the test animal) and an intraspecies UF of 10 (in case there is as much as 10-fold variability in susceptibility within the human population). These full 10-fold default factors must be used, because so few data are available.

    pgr. 3: It is important here to emphasize that Puschner et al. (2007) reported acute renal failure in a cat given as little as 32 mg melamine/kg and 32 mg cyanuric acid/kg. Admittedly, only one cat was given this dosage combination. Undoubtedly, the true NOAEL and LOAEL for renal tubular damage are far lower for melamine + cyanuric acid, hence the necessity of inclusion of the third UF of 10.

    Note that 2.5 mg/kg food = 250 ppm.

    Consideration in the future should be given to derivation of a TDI for food products contaminated with melamine and its analogues, that is a TDI based on results of dose-response experiments in which animals are administered different doses and combinations of melamine and congeners. The current high level of uncertainty should be emphasized.

    It may be worth pointing out that the renal crystalline occlusion syndrome is common to multiple mammalian species (i.e., humans, dogs, cats, pigs, rats and mice), as well as fish. It will be helpful through research to quantify interspecies differences in susceptibility, and to learn whether TK and/or toxicodynamic factors are contributory.

    Update - Interim Safety and Risk Assessment for Melamine and Its Analogues in Food for Humans

    pgr. 2, last line: It should be stated here that infants' and toddlers' renal systems are not fully developed. Several chapters in section XVI of the text entitled Fetal and Neonatal Physiology by Polin et al. (2004) described the maturation of kidney structure and function. Human newborns produce urine that is much less concentrated than in adults. Children almost reach the adult's maximum ability to concentrate urine by ~ 2 years (Polacek et al., 1965). Neonates and young children may therefore be less susceptible (on a mg/kg basis) to urolithiasis, since it has been proposed that melamine and cyanuric acid do not polymerize until their concentration exceeds a critical point as they move down the osmotic gradient in the nephron (Dobson et al., 2008). Nevertheless, caution must be exercised, until research can clarify whether susceptibility is age-dependent.

    ADDITIONAL REFERENCES

    Canelli E. Chemical, bacteriological and toxicological properties of cyanuric acid and chlorinated isocyanurates as applied to swimming pool disinfection. Am J Public Health 1974; 64: 155-162.

    Cinciolo RE, Bischoff A, Ebel JG, Van Winkle TJ, Goldstein RE, Serfilippi LM. Clinicopathologic, histologic, and toxicologic findings in 70 cats inadvertently exposed to pet food contaminated with melamine and cyanuric acid. J Am Vet Med Assoc 2008; 233: 729-737.

    Hammond BG, Barbee SJ, Inoue T, Ishida N, Levinskas GJ, Stevens MW, Wheeler AG, Cascieri T. A review of toxicology studies on cyanurate and its chlorinated derivatives. Environ Health Perspect 1986; 69: 287-292.

    Mast RW, Jeffcoat AR, Sadler BM, Kraska RC, Friedman MA. Metabolism disposition and excretion of [14C] melamine in male Fischer 344 rats. Food Chem Toxicol 1983; 21: 807-810.

    Melnick RL, Boorman GA, Haseman JK, Montali RJ, Huff J. Urolithiasis and bladder carcinogenicity of melamine in rodents. Toxicol Appl Pharmacol 1984; 72: 292-303.

    Ogasawara H, Imaida K, Ishiwata H, Toyoda K, Kawanishi T, Uneyama C, Hayashi S, Takahashi M, Hayashi Y. Urinary bladder carcinogenesis induced by melamine in F344 male rats: Correlation between carcinogenicity and urolith formation. Carcinogenesis 1995; 16: 2773-2777.

    Okumura M, Hasegawa R, Shirai T, Ito M, Yamada S, Fukushima S. Relationship between calculus formation and carcinogenesis in the urinary bladder of rats administered the non-genotoxic agents, thymine or melamine. Carcinogenesis 1992; 1043-1045.

    Polacek E, Vocel J, Neugebauerova L, Sebkova M, Vechetova E. The osmotic concentrating ability in healthy infants and children. Arch Dis Child 1965; 40: 291-295.

    Polin RA, Fox WW, Abman SH (Eds). 2004. Fetal and Neonatal Physiology. Section XVI. The Kidney, Vol. 2, 3rd ed., Sauders, Philadelphia, PA.

    Thompson ME, Lewin-Smith MR, Kalasinsky VF, Pizzolato KM, Fleetwood ML, McElhaney MR, Johnson TO. Characterization of melamine-containing and calcium oxalate crystals in three dogs with suspected pet food-induced nephrotoxicosis. Vet Pathol 2008; 45: 417.

FDA Response: Reviewer #1 indicated that the renal function of both pre-term and term infants is immature and it may take as long as three years before similar renal function as adults is achieved. FDA is aware that renal function, uric acid levels and glomerular filtration rates are different in infants and adults. The rats used in the NTP study were weanling rats.

In terms of the reviewer's comment that the S/RA is based more on assumptions than being derived from real data, FDA partially agrees with this point. It is more common than not that the agency is required to make regulatory and safety decisions on less than complete datasets. As more data are collected from the ongoing and planned studies by the FDA and others, they will be integrated into the present S/RA procedure. FDA anticipates that future test results will replace some, it not of all, of the assumptions and provide an even stronger and defensible scientific basis for decision making by the FDA.

The reviewer believes that the FDA did not acknowledge the possibility of accumulation of melamine alone or in combination with its analogues (or other environmental factors) in the immature infant system and how that possibility should be considered in the estimation of risk. FDA agrees that there are other unknown host and environmental factors that may significantly affect the S/RA. FDA directs the reviewer to the responses it has made in this document to the numerous questions and suggestions by the Peer Review Committee. Furthermore, FDA cannot characterize the risk from unrecognized factors that might make a contribution to risk. This is the reason that most regulatory authorities, including the FDA, apply safety or uncertainty factors to compensate for missing data/information. There are some studies that describe mitigation of the carcinogenicity of melamine with respect to cancer formation (Cremonezzi et al 2001), however there is no information regarding potentiators of the risk for melamine or its analogues.

Cremonezzi DC, Silva RA, . del Pilar D|¨az RA, Valentich MA, Eynard AR. Dietary polyunsatured fatty acids (PUFA) differentially modulate melamine-induced preneoplastic urothelial proliferation and apoptosis in mice. Prostaglandins, Leukotrienes and Essential Fatty Acids (2001) 64(3),151-159.

In response to the comment by Reviewer #1 that the phrase "does not raise public health concerns" is not defined in the document, FDA offers the following. The phrase is used to convey that the circumstance under discussion is not expected to require immediate corrective Agency action to protect the health of consumers.

FDA agrees with Reviewer #3 that data from these proposed and ongoing studies are pivotal in evaluating the risk from this combined exposure. Also, see our responses to Question #10. In our communications with the Chinese pediatricians, the primary effect of the milk and formula contamination has been stone formation (radiolucent stones, melamine - uric acid composition) and not renal crystals as occurred in pets in 2007. In the 2008 episode, the primary adulterant was melamine alone.

Although renal anatomy does not vary significantly between people and domestic animals, serum uric acid concentrations are greater in humans, especially human infants. Children have higher urine uric acid clearance (fractional excretion rates of 38-61% vs. the adult rate of approximately 10%). Thus, children may be more susceptible to developing hyperuricosuria, which may make them more likely to develop urinary uric acid precipitates. The neonatal kidney is, however, less efficient at concentrating and acidifying urine during the first weeks of life, which may help protect the kidney from potentially harmful effects of excreting the relatively large quantities of uric acid. However, increased uric acid load from surgery, drugs or asphyxia could overwhelm the neonatal kidney and pose a potential risk for developing acute urate nephropathy (Stapleton 1983). Elevated uric acid levels can also result in nephrolith formation. In fact, uric acid stones are the most common radiolucent kidney stone of children (Fathallah-Shaykh and Neiberger 2008). Thus, it is prudent to keep potential differences between humans and animals in mind due to the lack of uric acid oxidase in the former. These differences could also impact potential treatment regimes.

Fathallah-Shaykh F, Neiberger R. Uric Acid Stones E-medicine. http://www.emedicine.com/ped/topic2361 Updated 8 July 2008, Accessed 2 Dec 2008

Reviewer #4 pointed out the potential confusion in the definitions and use of the analytical terms ND, LOD (MDL) and LOQ in the appendices of the earlier 5/24/07 Interim Melamine and Analogues Safety/Risk Assessment. These analytical terms were not used in the 10/3/08 or 11/28/08 Interim and Update to the Interim Safety and Risk Assessment of Melamine and its Analogues in Food for Humans and we will be sure to clarify and be consistent with these terms in our future updates of the safety and risk assessments of melamine and its analogues.

In regards to Reviewer #5's suggested revision of the S/RA to more rigorously support the statements and conclusions of the risk assessment, FDA plans on adding numerous references in support of the S/RA and its conclusions in its revision of the S/RA. FDA thanks this reviewer for this additional list and discussion of these references. They will be carefully reviewed and evaluated for relevance and support for the next revision of the S/RA.


 1 Principles for the Safety Assessment of Food Additives and Contaminants in Food (Environmental Health Criteria 70), World Health Organization, Geneva, 1987, p. 111.

 2 See the FDA guidance for industry regarding the "Preparation of Premarket Submissions for Food Contact Substances:Chemistry Recommendations;" http://www.cfsan.fda.gov/~dms/opa3pmnc.html.

3 Average body weight determined by dividing average food intake in g/p/d by the average food intake in g/kg bw/d.

4 The portion of fluid milk as a percentage of the total average amount of milk and dairy products consumed was determined by dividing the average amount of fluid milk consumed by a population subgroup by the average amount of milk and dairy products consumed by the population subgroup as determined using eaters-only 2003-2004 NHANES survey data.

5 Fraction of the TDI/10 determined by diving the average melamine intake assuming that milk and dairy products contain 2.5 mg/kg melamine by the TDI/10 of 0.063 mg/kg bw/d. For example, for the 2-5 year old population subgroup: (0.068 mg/kg bw/d) / (0.063 mg/kg bw/d) = 1.1.