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

Food

  • Print
  • Share
  • E-mail

Draft Report of Quantitative Risk and Benefit Assessment of Consumption of Commercial Fish, Focusing on Fetal Neurodevelopmental Effects (Measured by Verbal Development in Children) and on Coronary Heart Disease and Stroke in the General Population: Appendix B, Modeling Results

January 15, 2009

This information is distributed solely for the purpose of pre-dissemination peer and public review under applicable information quality guidelines. It has not been formally disseminated by FDA. It does not represent and should not be construed to represent any agency determination or policy.

Return to Draft Risk and Benefit Report Tabe of Contents

(a) Baseline Estimates

The estimates generated by the model for the U.S. population are presented in this section. Results are presented for each of four population groups, Younger Women aged 15-45, Younger Men aged 15-45, Older Women aged 46 and older, and Older Men aged 46 and older. The results for the first group, which includes women of childbearing age, also address neurobehavioral effects on children resulting from maternal exposure.

Fish Consumption

The distributions for fish consumption of each subpopulation are presented in Table AB-1.

 Table AB-1: Estimates Average Daily Intake of Fish (in grams) for all Four Subpopulation Groups
Population PercentileWomen 15-45Women 46+Men 15-45Men 46+
Average13.1 (12.4, 13.8)14.9 (14.3, 15.7)17.9 (16.8, 19.0)18.8 (17.7, 20.0)
10th Percentile0.1 (0.0, 0.9)0.2 (0.0, 1.3)0.2 (0.0, 1.3)0.2 (0.0, 1.7)
25th Percentile2.8 (2.0, 3.4)3.4 (2.5, 4.3)3.7 (2.5, 4.6)4.5 (3.1, 5.6)
Median7.0 (6.1, 7.7)8.2 (7.2, 8.9)9.3 (8.1, 10.4)10.7 (9.1, 11.6)
75th Percentile16.0 (14.8, 17.3)18.0 (16.8, 19.2)21.3 (19.1, 23.1)22.3 (20.4, 24.1)
90th Percentile31.3 (29.2, 33.9)36.3 (33.3, 38.7)42.9 (39.2, 47.5)43.9 (39.8, 49.0)
95th Percentile45.4 (41.2, 51.3)53.3 (47.7, 59.8)63.9 (58.4, 73.2)64.8 (57.4, 77.5)
99th Percentile88.0 (73.6, 112.9)100.0 (84.9, 126.6)132.4 (109.7, 170.3)127.4 (100.8, 167.7)
NHANES Average10.314.216.820.8
All units are grams of fish consumed per day. The daily consumption was derived for each individual in the population by averaging daily consumption for one year. The central estimates are the median estimates of the uncertainty distribution. The 5th and 95th uncertainty percentiles are given in parentheses as confidence intervals.
Because our model is based in part on data from 1989-1991, the table includes average daily consumption taken from the 2003-2004 NHANES survey in order to verify that our results are consistent with current consumption patterns. The reference for this is the Documentation, Codebook, and Frequencies, Dietary Interview - Individual Foods (First Day), for the National Health and Nutrition Examination Survey for 2003-2004 (USDA/HHS 2007). This document is primarily involved with survey methodology. The data we used are in a file referenced in the document, i.e., SAS Transport File DR1IFF_C and DR2IFF_C.XTP.

 

Blood and Hair Mercury Levels

The distributions for blood and hair mercury levels in younger women are presented in Table AB-2. For purposes of comparison, the model estimates are compared to results from NHANES in Figures AB-1 and AB-2. It may be observed that the blood levels model are generally 20-25 percent lower than the levels observed in NHANES, while the hair levels are 30-40 percent lower. These discrepancies may be explained, at least in part, by two factors:

  • While the NHANES measurements are for total mercury in both blood and hair, the model is intended to estimate concentrations on just methylmercury.
  • The model has been updated with more recent (2005) market data indicating that the consumption patterns have changed somewhat so that the average commercial fish weighted by frequency of consumption, now has less methylmercury in it than when the most recently reported NHANES mercury biomarker survey data were conducted (2000 for hair and 2002 for blood).
 Table AB-2: Model Estimates of Blood and Hair Methylmercury levels in Women of Childbearing Age
Population PercentileDietary Hg from Fish
(µg/day)
Blood Hg
(µg/L):
Hair Hg
(ppm)
Average1.4 (1.3, 1.4)1.2 (1.2, 1.3)0.3 (0.2, 0.3)
10th Percentile0.0 (0.0, 0.1)0.1 (0.1, 0.1)0.0 (0.0, 0.0)
25th Percentile0.2 (0.1, 0.3)0.3 (0.2, 0.3)0.0 (0.0, 0.1)
Median0.7 (0.6, 0.7)0.6 (0.5, 0.7)0.1 (0.1, 0.1)
75th Percentile1.6 (1.5, 1.8)1.5 (1.3, 1.6)0.3 (0.2, 0.3)
90th Percentile3.4 (3.1, 3.6)2.9 (2.7, 3.2)0.6 (0.5, 0.7)
95th Percentile4.9 (4.5, 5.5)4.3 (3.8, 4.8)1.0 (0.8, 1.2)
99th Percentile10.3 (8.1, 12.8)8.8 (7.4, 10.7)2.2 (1.8, 2.7)

 

 Table AB-3: Model Estimates of Blood and Hair Methylmercury levels in Men Aged 16-45
Population PercentileDietary Hg from Fish
(µg/day)
Blood Hg
(µg/L)
Hair Hg
(ppm)
Average1.8 (1.7, 1.9)1.4 (1.3, 1.5)0.3 (0.3, 0.4)
10th Percentile0.0 (0.0, 0.1)0.1 (0.1, 0.1)0.0 (0.0, 0.0)
25th Percentile0.3 (0.2, 0.4)0.3 (0.2, 0.4)0.1 (0.0, 0.1)
Median0.9 (0.7, 1.0)0.7 (0.6, 0.8)0.1 (0.1, 0.2)
75th Percentile2.1 (1.9, 2.3)1.7 (1.5, 1.9)0.3 (0.3, 0.4)
90th Percentile4.3 (3.9, 4.7)3.3 (3.0, 3.7)0.7 (0.6, 0.8)
95th Percentile6.4 (5.6, 7.5)5.0 (4.4, 5.7)1.1 (0.9, 1.3)
99th Percentile13.4 (10.9, 17.3)10.4 (8.5, 12.8)2.4 (2.0, 3.2)

 

Figure AB-1: Quantile-Quantile Comparison of Model Estimates of Blood Mercury with Values from NHANES
Model Values µg/L versus Survey Values µg/L - link to description

NHANES survey data are taken from the 199-200, 2001-2002, and 2003-2004 surveys (CDC 2004). The following percentiles are plotted: 10th, 25th, 50th, 75th, 90th, 95th, 99th. 99.5th, and 99.9th. Percentiles were calculated with the demographic weights provided with the survey data.

Figure AB-2: Quantile-Quantile Comparison of Model Estimates of Hair Mercury with Values from NHANES
Model Values in ppm versus Survey Values in ppm - link to description

NHANES survey data are taken from CDC (2001), which reflects data collected from 1999-2000. The following percentiles are plotted: 10th, 25th, 50th, 75th, 90th, and 95th.

Neurobehavioral Effects Attributable to Methylmercury Exposure

The predicted neurobehavioral effects resulting from current levels of methylmercury on verbal performance in toddlers with and without taking into account potential offsetting effects from fish consumption are presented in Table AB-4. Since the endpoints or responses modeled are not identical, the results are all represented as a Z-score where the outcome is expressed relative to normal variation (i.e. each Z unit = one standard deviation). The Carrington and Bolger (2000) model for delayed talking was normalized using the standard deviation from Seychelles of 2.76 months. In order to investigate the possible consequences of beneficial effects of maternal consumption of fish, a simulation model was constructed that included both methylmercury effects characterized with the Carrington and Bolger (2000) delayed talking analysis, and an uncertain range of beneficial effects based on the verbal comprehension regression analyses from the Daniels et al. (2004) study. This analysis presumes that the contribution of fish to the slope for delayed talking (which was dominated by the data from Iraq, where methylmercury was ingested in bread) and the contribution of mercury to the apparent benefit observed in the Daniels study are both negligible.

 Table AB-4: Baseline Estimates for Z-score Change Attributable to Methylmercury With and Without Offsetting Fish Benefits
Population PercentileCarrington (2000),
Age of Talking
Net Verbal ΔZ
Average-2.9E-3 (-4.5E-3, -6.1E-4)0.008 (0.000, 0.012)
1st Percentile-2.7E-2 (-4.2E-2, -8.5E-3)-0.001 (-0.020, 0.000)
5th Percentile-1.2E-2 (-1.8E-2, -3.3E-3)0.000 (-0.006, 0.001)
10th Percentile-7.3E-3 (-1.1E-2, -1.7E-3)0.001 (-0.003, 0.002)
25th Percentile-3.2E-3 (-4.8E-3, 2.2E-4)0.002 (0.000, 0.004)
Median-1.1E-3 (-1.8E-3, 0.0E0)0.005 (0.000, 0.008)
75th Percentile-2.9E-4 (-5.2E-4, 0.0E0)0.010 (0.001, 0.016)
90th Percentile0.0E0 (-6.8E-5, 0.0E0)0.020 (0.003, 0.030)
95th Percentile0.0E0 (6.8E0, 0.0E0)0.026 (0.005, 0.039)
99th Percentile0.0E0 (0.0E0, 0.0E0)0.035 (0.010, 0.048)
The values in the second column represent the net (added) change in Z-score from the Carrington (2000) delayed talking dose-response function and the Daniels et al. (2004) verbal test score dose-response function.

 

For comparison, predicted neurobehavioral effects for several other dose-response functions are presented in Table AB-5. Since IQ is defined as the Z score x 15, the Axelrad et al. (2007) and Cohen et al. (2005b) models were converted by dividing by 15. The Carrington (2000) model for delayed walking was normalized using the standard deviation from Seychelles of 1.96 months.

 Table AB-5: Baseline Estimates for Z-score Change Attributable to Fish Consumption With Benefits from Maternal Consumption of Fish
Population PercentileCarrington (2000),
Age of Walking
Cohen et al. (2005b)
IQ
Axelrad et al. (2007)
IQ
Average-6.0E-3 (-9.3E-3,-1.8E-3 (-1.3E-2, 1.3E-3)-3.0E-3 (-5.6E-3, -7.0E-4)
1st Percentile-6.0E-2 (-1.0E-1, 0.0E0)-1.6E-2 (-1.2E-1, 0.0E0)-2.6E-2 (-5.0E-2, -6.1E-3)
5th Percentile-2.4E-2 (-3.7E-2, 0.0E0)-7.1E-3 (-5.2E-2, 0.0E0)-1.2E-2 (-2.2E-2, -2.8E-3)
10th Percentile-1.4E-2 (-2.3E-2, 0.0E0)-4.5E-3 (-3.2E-2, 7.4E-6)-7.5E-3 (-1.4E-2, -1.6E-3)
25th Percentile-5.8E-3 (-1.0E-2, 0.0E0)-2.0E-3 (-1.5E-2, 1.5E-4)-3.4E-3 (-6.4E-3, -7.5E-4)
Median-1.8E-3 (-3.5E-3, 0.0E0)-7.4E-4 (-5.6E-3, 5.5E-4)-1.3E-3 (-2.4E-3, -2.7E-4)
75th Percentile-3.7E-4 (-9.3E-4, 0.0E0)-2.1E-4 (-1.7E-3, 1.4E-3)-3.6E-4 (-7.6E-4, -7.8E-5)
90th Percentile0.0E0 (-1.2E-4, 0.0E0)0.0E0 (-1.9E-4, 3.1E-3)-4.1E-6 (-1.0E-4, 0.0E0)

 

Cardiovascular Effects Attributable to Fish Consumption

Table AB-6 and Table AB-7 give the estimated annual reduction in the rate of occurrence of CHD Death and Stroke Death attributable to current levels of fish consumption.

 Table AB-6: Baseline Estimates for CHD Death Benefit Attributable to Fish Consumption
Population GroupCHD Death Cases Per Person-Year
Pooled Analysis Model
CHD Death Cases Per Person-Year
Meta-Analysis Model
Women 15-45-1.1E-6 (-2.2E-5, 2.6E-6)-6.8E-7 (-1.3E-6, -1.3E-7)
Women 46+-2.8E-4 (-4.2E-3, 9.2E-4)-2.2E-4 (-4.3E-4, -4.0E-5)
Men 15-45-1.1E-5 (-1.2E-4, 3.4E-5)-8.9E-6 (-1.7E-5, -1.6E-6)
Men 46+-4.7E-4 (-8.8E-3, 6.5E-4)-3.7E-4 (-7.2E-4, -6.6E-5)
The values are median estimates of cases per person-year, with the 5th and 95th percentiles of the uncertainty distributions given as confidence intervals.

 

 Table AB-7: Baseline Estimates for Stroke Death Benefit Attributable to Fish Consumption
Population GroupStroke Death Cases Per Person-Year
Pooled Analysis Model
Stroke Death Cases Per Person-Year
Meta-Analysis Model
Women 15-45-3.7E-6 (-3.1E-5, 1.5E-6)-3.0E-6 (-5.6E-6, -9.1E-7)
Women 46+-2.2E-4 (-2.7E-3, 1.4E-4)-2.2E-4 (-4.1E-4, -7.1E-5)
Men 15-45-4.0E-6 (-3.1E-5, 3.1E-6)-3.2E-6 (-6.1E-6, -1.1E-6)
Men 46+-2.4E-4 (-2.0E-3, 9.2E-5)-1.8E-4 (-3.4E-4, -6.6E-5)
The values are median estimates of cases per person-year, with the 5th and 95th percentiles of the uncertainty distributions given as confidence intervals.

 

(b) Intervention Scenarios

Advisory Scenario: 12 oz Limits for Women of Childbearing Age

To examine the impact of fish consumption advisories, several scenarios were developed. The first scenario introduced a 12 oz. per week consumption limit for all women of childbearing age. This modification of the exposure model involved truncating consumption of fish at the specified limit. Whereas individuals coming more that 12 oz. /week are reduced to 12 oz., those individuals who are already consuming under that limit do not modify their consumption. The impact of the advisory on neurobehavioral outcome occurring as a result of the reduction in exposure to methylmercury, expressed as a Z-score, is presented in Table AB-8.

 Table AB-8: Z-score Change with a 12 oz Limit (vs. Baseline) for All Women of Childbearing Age
Population PercentileCarrington (2000),
Age of Talking
Axelrad (2007)
IQ
Net Verbal ΔZ
Average2.6E-4 (3.7E-5, 4.5E-4)2.2E-4 (4.6E-4, 7.1E-5)-7.3E-4 (-1.5E-3, -2.8E-5)
1st Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)-1.3E-1 (-3.0E-1, -3.0E-2)
5th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)-2.1E-2 (-4.6E-2, -2.7E-3)
10th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (-1.3E-3, 0.0E0)
25th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
Median0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
75th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
90th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
95th Percentile0.0E0 (0.0E0, 3.2E-4)0.0E0 (2.6E-4, 0.0E0)0.0E0 (0.0E0, 0.0E0)
99th Percentile7.8E-3 (1.8E-4, 1.3E-2)6.8E-3 (1.3E-2, 1.9E-3)0.0E0 (0.0E0, 0.0E0)

 

A second scenario used a 12 oz limit in conjunction with a stipulation that only fish species with average concentrations below 12 ppb are consumed by all women of childbearing age. This modification of the exposure model involved substituting fish species below the concentration limit for species above the limit. The impact of the advisory, relative to current levels of consumption, is presented in Table AB-9.

 Table AB-9: Z-score Change with a 12 oz Limit and Low-Mercury Fish (vs. Baseline) for All Women of Childbearing Age
Population PercentileCarrington (2000),
Age of Talking
Axelrad (2007)
IQ
Net Verbal ΔZ
Average1.3E-3 (1.6E-4, 1.9E-3)1.2E-3 (4.3E-4, 2.0E-3)5.2E-4 (-8.7E-4, 1.4E-3)
1st Percentile0.0E0 (0.0E0, 0.0E0)-4.1E-5 (-2.4E-4, -6.4E-6)-3.4E-2 (-6.8E-2, -9.3E-3)
5th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)-1.3E-2 (-2.3E-2, -3.1E-3)
10th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)-6.5E-3 (-1.2E-2, - 1.5E-3)
25th Percentile2.5E-5 (0.0E0, 7.1E-5)2.9E-5 (2.3E-6, 8.4E-5)-1.5E-3 (-3.2E-3, -1.3E-4)
Median3.1E-4 (0.0E0, 4.6E-4)3.1E-4 (1.0E-4, 5.5E-4)3.7E-5 (0.0E0, 4.9E-4)
75th Percentile1.1E-3 (0.0E0, 1.6E-3)1.0E-3 (3.5E-4, 1.8E-3)2.9E-3 (1.3E-3, 4.9E-3)
90th Percentile3.0E-3 (7.5E-5, 4.3E-3)2.6E-3 (9.3E-4, 4.5E-3)9.2E-3 (4.1E-3, 1.5E-2)
95th Percentile5.3E-3 (3.8E-4, 7.8E-3)4.6E-3 (1.7E-3, 7.7E-3)1.6E-2 (7.0E-3, 2.3E-2)
99th Percentile1.7E-2 (3.0E-3, 2.5E-2)1.4E-2 (5.1E-3, 2.5E-2)2.8E-2 (1.6E-2, 4.2E-2)

 

The change in cardiovascular risk for both these scenarios (fish consumption is identical) are presented in Tables AB-10 and AB-11.

 Table AB-10: CHD Death Risk with a 12 oz Limit (vs. Baseline) in Women of Childbearing Age
Population GroupCHD Death Cases Per Person-Year
Pooled Analysis Model
CHD Death Cases Per Person-Year
Meta-Analysis Model
Women 15-450.0E0 (-3.7E-8, 1.3E-7)7.1E-8 (1.3E-8, 1.6E-7)
The values are median estimates of cases per person-year, with the 5th and 95th percentiles of the uncertainty distributions given as confidence intervals.

 

 Table AB-11: Stroke Death Risk with a 12 oz Limit (vs. Baseline) in Women of Childbearing Age
Population GroupStroke Death Cases Per Person-Year
Pooled Analysis Model
Stroke Death Per Person-Year
Meta-Analysis Model
Women 15-450.0E0 (-3.1E-8, 8.0E-8)6.3E-8 (-1.3E-7, 2.4E-7)
The values are median estimates of cases per person-year, with the 5th and 95th percentiles of the uncertainty distributions given as confidence intervals.

 

As a third variation on the 12 oz limitation, a scenario where all women of childbearing age consume exactly 12 oz of fish was examined. For this scenario, the methylmercury concentration was presumed to correspond to the market average of 0.086 ppm, resulting in an exposure of 4.3 µg/day of methylmercury. The impact of the advisory, relative to current levels of consumption, is presented in Table AB-12.

 Table AB-12: Z-score Change with Fixed 12 oz Consumption (vs. Baseline) for Women of Childbearing Age
Population PercentileCarrington (2000),
Age of Talking
Cohen (2005b)
IQ
Axelrad (2007)
IQ
Average-5.2E-3 (-7.2E-3, -4.0E-4)-3.1E-3 (-2.3E-2, 2.3E-3)-4.6E-3 (-8.1E-3, -1.6E-3)
1st Percentile-2.1E-2 (-3.4E-2, -7.4E-3)-2.3E-2 (-1.8E-1, -1.5E-3)-1.5E-2 (-3.2E-2, -5.4E-3)
5th Percentile-1.5E-2 (-2.1E-2, -3.6E-3)-1.1E-2 (-8.4E-2, -8.3E-4)-1.1E-2 (-2.0E-2, -4.1E-3)
10th Percentile-1.2E-2 (-1.7E-2, -2.1E-3)-8.1E-3 (-6.0E-2, -1.7E-4)-9.6E-3 (-1.7E-2, -3.4E-3)
25th Percentile-7.9E-3 (-1.1E-2, 0.0E0)-6.8E-3 (-5.0E-2, 2.6E-4)-7.0E-3 (-1.2E-2, -2.4E-3)
Median-4.8E-3 (-6.7E-3, 0.0E0)-4.7E-3 (-3.6E-2, 1.3E-3)-4.5E-3 (-7.7E-3, -1.5E-3)
75th Percentile-2.5E-3 (-3.6E-3, 0.0E0)-3.0E-3 (-2.4E-2, 2.2E-3)-2.6E-3 (-4.6E-3, -8.3E-4)
90th Percentile-4.5E-4 (-1.1E-3, 0.0E0)-1.7E-3 (-1.3E-2, 3.5E-3)-5.3E-4 (-1.3E-3, 0.0E0)
95th Percentile1.8E-3 (0.0E0, 3.2E-3)-2.0E-4 (-3.7E-3, 4.7E-3)1.7E-3 (4.7E-4, 3.3E-3)
99th Percentile1.4E-2 (1.7E-3, 2.1E-2)1.9E-3 (1.2E-4, 9.5E-3)1.2E-2 (4.4E-3, 2.2E-2)

 

The change in cardiovascular and stroke risk for the fixed 12 oz scenario are presented in Table AB-13 and Table AB-14, respectively.

 Table AB-13: CHD Death Risk with 12 oz Consumption (vs. Baseline) in Women of Childbearing Age
Population GroupCHD Death Cases Per Person-Year
Pooled Analysis Model
CHD Death Cases Per Person-Year
Meta-Analysis Model
Women 15-45-1.4E-6 (-9.7E-6, 2.9E-6)-1.7E-6 (-3.3E-6, -1.8E-7)
The primary values are the median cases per person-year of the uncertainty distribution, the 5th and 95th percentiles given as confidence intervals.

 

  Table AB-14: Stroke Death Risk with 12 oz Consumption (vs. Baseline) in Women of Childbearing Age
Population GroupStroke Death Cases Per Person-Year
Pooled Analysis Model
Stroke Death Cases Per Person-Year
Meta-Analysis Model
Women 15-45-3.9E-6 (-1.3E-5, 4.0E-6)-2.2E-6 (-6.5E-6, 2.0E-6)
The primary values are the median cases per person-year of the uncertainty distribution, the 5th and 95th percentiles given as confidence intervals.

 

As an exploration of potential unintended consequences of an advisory, the impact of reductions in the fish consumption in other population groups was modeled. Tables AB-15 through AB-18 present the estimate changes in CHD and cardiovascular death rates resulting from either a 10 percent reduction in the amount of fish consumed by all consumers or a one percent increase in the number of consumers who do not eat fish at all.

 Table AB-15: CHD Death Risk from a One Percent Reduction in Persons Consuming Fish in Men and Older Women
Population GroupCHD Death Cases Per Person-Year
Pooled Analysis Model
CHD Death Cases Per Person-Year
Meta-Analysis Model
Women 46+2.9E-6 (-9.9E-6, 4.7E-5)2.3E-6 (3.5E-7, 5.1E-6)
Men 15-451.2E-7 (-3.5E-7, 1.3E-6)9.2E-8 (1.4E-8, 2.1E-7)
Men 46+5.4E-6 (-7.9E-6, 1.1E-4)3.9E-6 (6.0E-7, 8.7E-6)
Estimates of increased rates of CHD Death resulting from decreased number of fish consumers in three population groups. The primary values are the median cases per person-year of the uncertainty distribution, the 5th and 95th percentiles given as confidence intervals.

 

 Table AB-16: Stroke Death Risk from a One Percent Reduction in Persons Consuming Fish in Men and Older Women
Population GroupStroke Death Cases Per Person-Year
Pooled Analysis Model
Stroke Death Cases Per Person-Year
Meta-Analysis Model
Women 46+2.5E-6 (-1.5E-6, 3.2E-5)2.4E-6 (7.1E-7, 4.6E-6)
Men 15-454.3E-8 (-2.8E-8, 3.4E-7)3.5E-8 (1.1E-8, 6.8E-8)
Men 46+2.8E-6 (-9.1E-6, 2.1E-5)1.9E-6 (6.0E-7, 4.0E-6)
Estimates of increased rates of Stroke Death resulting from decreased number of fish consumers in each of three population groups. The primary values are the median cases per person-year, with the 5th and 95th percentiles given as confidence intervals.

 

 Table AB-17: Comparison of CHD Death Risk from a 10 Percent Reduction in Persons Consuming Fish in Men and Older Women
Population GroupCHD Death Cases Per Person-Year
Pooled Analysis Model
CHD Death Cases Per Person-Year
Meta-Analysis Model
Women 46+1.2E-5 (-3.4E-5, 1.4E-4)2.2E-5 (4.0E-6, 4.3E-5)
Men 15-454.5E-7 (-1.5E-6, 3.7E-6)8.9E-7 (1.6E-7, 1.7E-6)
Men 46+2.9E-5 (-3.5E-5, 2.2E-4)3.7E-5 (6.6E-6, 7.2E-5)
Estimates of increased rates of CHD Death resulting from decreased fish consumption in three population groups. The primary values are the median cases per person-year of the uncertainty distribution, the 5th and 95th percentiles given as confidence intervals.

 

 Table AB-18: Stroke Death Risk from a 10 Percent Reduction in Persons Consuming Fish in Men and Older Women
Population GroupStroke Death Cases Per Person-Year
Pooled Analysis Model
Stroke Death Cases Per Person-Year
Meta-Analysis Model
Women 46+1.3E-5 (-1.2E-5, 5.7E-5)9.6E-6 (-2.2E-6, 2.2E-5)
Men 15-452.1E-7 (-2.0E-7, 7.2E-7)1.4E-7 (-5.6E-8, 3.4E-7)
Men 46+1.2E-6 (-8.6E-5, 3.8E-5)7.3E-6 (-4.4E-6, 1.9E-5)
Estimates of increased rates of Stroke Death resulting from decreased fish consumption in each of three population groups. The primary values are the median cases per person-year, with the 5th and 95th percentiles given as confidence intervals.

 

Advisory Scenario: Limiting Species Without Limiting Amount Consumed

This scenario limited the species consumed without altering the amount consumed. Specifically, this putative advisory stipulates that only fish species with average concentrations below 12 ppb (see Table AA-2 in "methodology" Appendix A) are consumed by all women of childbearing age. This modification of the exposure model involved substituting fish species below the average concentration limit for species above the limit. The impact of the advisory, relative to current levels of consumption, is presented in Table AB-19. The presence of the negative values at the first percentile is a result of the possibility of substituting an individual fish from a species with a lower average concentration with a mercury level that higher than the individual fish that would have been consumed from the species with a higher average concentration. Since the amount of fish consumed is unchanged in this scenario, there is no impact on neurobehavioral benefits from fish or on cardiovascular endpoints, and as a result, the predicted effects with and without the inclusion of the Daniels benefit dose-response function are identical.

 Table AB-19: Z-score Change with Low-Methylmercury Fish and No Consumption Limit (vs. Baseline) for All Women of Childbearing Age
Population PercentileCarrington (2000),
Age of Talking
Axelrad (2007)
IQ
Net Verbal ΔZ
Average1.1E-3 (0.0E0, 1.7E-3)1.0E-3 (0.0E0, 1.7E-3)1.1E-3 (0.0E0, 1.7E-3)
1st Percentile-3.9E-5 (0.0E0, 0.0E0)-4.2E-5 (0.0E0, -6.8E-6)-3.9E-5 (0.0E0, 0.0E0)
5th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
10th Percentile0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)0.0E0 (0.0E0, 0.0E0)
25th Percentile2.5E-5 (0.0E0, 7.0E-5)2.9E-5 (0.0E0, 8.3E-5)2.5E-5 (0.0E0, 7.0E-5)
Median3.0E-4 (0.0E0, 4.6E-4)3.0E-4 (0.0E0, 5.4E-4)3.0E-4 (0.0E0, 4.6E-4)
75th Percentile1.1E-3 (0.0E0, 1.5E-3)1.0E-3 (0.0E0, 1.7E-3)1.1E-3 (0.0E0, 1.5E-3)
90th Percentile2.8E-3 (0.0E0, 4.1E-3)2.5E-3 (0.0E0, 4.3E-3)2.8E-3 (0.0E0, 4.1E-3)
95th Percentile4.8E-3 (0.0E0, 6.9E-3)4.1E-3 (0.0E0, 6.9E-3)4.8E-3 (0.0E0, 6.9E-3)
99th Percentile1.3E-2 (0.0E0, 1.9E-2)1.1E-2 (0.0E0, 1.9E-2)1.3E-2 (0.0E0, 1.9E-2)

Advisory Scenario: 50 Percent Increased Fish Consumption

 

This scenario examines the potential impact of an advisory that recommends increasing fish consumption. The model was modified by presuming an across the board increase in fish consumption of 50 percent. Table AB-20 and Table AB-21 present the estimated changes in CHD and Stroke Death relative to current levels of exposure, respectively.

 Table AB-20: CHD Death Benefit from a 50 Percent Increase in the Amount of Fish Consumed, Relative to Baseline
Population GroupCHD Death
Pooled Analysis Model
CHD Death
Meta-Analysis Model
Women 15-45-1.7E-7 (-2.7E-6, 4.2E-7)-3.4E-7 (-6.7E-7, -6.7E-8)
Women 46+-4.1E-5 (-5.3E-4, 1.3E-4)-1.1E-4 (-2.1E-4, -2.0E-5)
Men 15-45-1.9E-6 (-1.4E-5, 5.8E-6)-4.4E-6 (-8.6E-6, -8.0E-7)
Men 46+-1.1E-4 (-1.0E-3, 1.4E-4)-1.9E-4 (-3.6E-4, -3.3E-5)
The negative values represent a decrease in estimated mortality.

 

 Table AB-21: Comparison of Stroke Death Risk from a 50 Percent Increase in the Amount of Fish Consumed, Relative to Baseline
Population GroupStroke Death
Pooled Analysis Model
Stroke Death
Meta-Analysis Model
Women 15-45-8.1E-7 (-3.0E-6, 6.0E-7)-5.4E-7 (-1.3E-6, 2.2E-7)
Women 46+-5.1E-5 (-1.9E-4, 5.5E-5)-3.9E-5 (-1.0E-4, 2.5E-5)
Men 15-45-8.0E-7 (-2.5E-6, 8.6E-7)-6.0E-7 (-1.6E-6, 5.7E-7)
Men 46+-4.4E-5 (-1.3E-4, 3.7E-5)-3.1E-5 (-8.4E-5, 3.5E-5)
The negative values represent a decrease in estimated mortality.


 


Descriptions of Figures

  •  Figure AB-1: Quantile-Quantile Comparison of Model Estimates of Blood Mercury with Values from NHANES
    Figure AB-1 is a quantile-quantile comparison for survey values from NHANES and model estimates for mercury in blood in women of childbearing age. Confidence intervals are given for the model estimates only. A line equality demarcates where the plotted values would if they were equal. The values plotted range from 0 to 25 micrograms per liter. The values below 5 micrograms per liter are all very close to the line of equality, indicating that the model estimates are very close to the survey values. At higher levels, above 5 micrograms per liter, the model estimates appear to be about 20% lower than the survey values.
  •  Figure AB-2: Quantile-Quantile Comparison of Model Estimates of Hair Mercury with Values from NHANES
    Figure AB-2 is a quantile-quantile comparison for survey values from NHANES and model estimates for mercury in hair in women of childbearing age. Confidence intervals are given for both the model estimates and the survey values. A line equality demarcates where the plotted values would if they were equal. The values plotted range from 0 to ppm. The model values are all lower than the survey values, with greater discrepancies at the higher percentiles. At levels below 1 ppm, the model estimates are less than 20% below the survey values. Above 1 ppm, the model estimates are about 40% lower than the survey values.