A Conversation with Alexandre da Silva and Helen Murphy
In mid-July 2018, the FDA and the Centers for Disease Control and Prevention (CDC) learned of an outbreak of cyclosporasis that had sickened hundreds of people across 15 states. Cyclosporiasis is an intestinal infection caused by the Cyclospora parasite, which is often foodborne. FDA, using a new testing method developed and validated by agency scientists, subsequently confirmed the presence of Cyclospora cayetanensis in salad mix that had been sold at McDonald’s restaurants.
This is the first time FDA has been able to confirm the presence of Cyclospora in food since the late 1990s and early 2000s, when an earlier testing method was used that has not been available since then. But Cyclospora has continued to be a public health concern, and with outbreaks increasing in recent years it became apparent that the FDA needed a new way to detect this parasite.
In 2015, FDA’s Center for Food Safety and Applied Nutrition (CFSAN) established a Foodborne Parasitology Research Program tasked with creating a new, more sensitive method to detect this elusive parasite. The availability of this new method is a significant advancement in FDA's ability to investigate outbreaks of cyclosporiasis and identify the parasite in foods.
And recently, FDA included the use of this method in its surveillance sampling of fresh herbs, detecting Cyclospora in several samples of cilantro that were not associated with illnesses.
Alexandre da Silva, Ph.D, Senior Biomedical Research Service (SBRS ) Research Microbiologist and Lead Parasitologist at CFSAN’s Office of Applied Research and Safety Assessment (OARSA) and Research Microbiologist Helen Murphy, M.Sc, talk about the work that the FDA has done to develop and validate this new testing method and the work ahead to help prevent the contamination of food by Cyclospora.
Why did the FDA need a new way to test for Cyclospora? What happened to the old method?
Murphy: The old testing method was an effective way to identify Cyclospora. In 1999, the parasite was detected in a chicken pasta salad containing fresh basil that was linked to an outbreak of cyclosporiasis. In 2000, FDA in conjunction with CDC detected Cyclospora in the raspberry filling of a wedding cake after a number of guests became sick from cyclosporiasis. In both instances, FDA used the earlier version of the method to confirm the presence of Cyclospora. However, since that time FDA had been unable to detect Cyclospora in food because some of the essential equipment and supplies needed to conduct the earlier testing method were no longer being manufactured.
What impact has the development of this new testing method had on the Cyclospora outbreak of the summer of 2018?
Da Silva: In an outbreak, CDC looks for common factors among those who are ill to determine what made them sick. When an FDA-regulated food source appears to be involved, FDA uses that information to collect and test the implicated food. In this instance, CDC identified a number of people who had become ill after eating salad from McDonald’s. Once the salad was identified as a potential source, the next step was to test samples of the salad. This new method not only provided us a way to detect Cyclospora once again, but advanced the technology beyond the testing we used in previous years to provide better analytical tools with greater sensitivity to detect Cyclospora on fresh produce. This helped us connect the dots and supported CDC’s epidemiological findings about the likely source of illness. The laboratory results were consistent with the other available evidence, giving us a much stronger case for taking regulatory actions, if needed.
Once we confirmed that it was the salad that was contaminated, we needed to look at the supply chain of the product to figure out what happened to cause this outbreak and determine what actions were needed to be taken to ensure no other contaminated produce entered the marketplace. In this case, the detection of Cyclospora triggered further investigation at the processing facility and farms that supplied the processing facility.
To backtrack: What is Cyclospora and what makes it so different from the causes of other foodborne diseases?
Da Silva: Cyclospora is a parasite that is spread through the contamination of food or water with human feces. Investigators tend to be puzzled by the pattern of cyclosporasis outbreaks. They don’t follow the same patterns as pathogens like E. coli and Salmonella, but that makes sense. Cyclospora is a parasite, with a complicated lifecycle, not a bacterial pathogen, like those others. We tend to throw all microorganisms in the same basket, but we shouldn’t because parasites and bacteria are very different.
Epidemiology is very complicated for foodborne parasites. One thing that makes it complicated is the incubation period. The incubation period for Cyclospora may take several days, sometimes even more than 10 days, which means that a person who eats food contaminated by Cyclospora may not develop symptoms for about a week or more. That’s one thing that’s very different from bacterial foodborne pathogens, which often make people sick more rapidly. Plus, patients may not immediately seek medical care, be diagnosed, and have their illness reported to public health authorities, meaning they may not be asked about the foods they ate until weeks after they became ill. It’s very difficult for patients to recall what they ate weeks earlier or where they ate it. This all makes trying to find the source of an outbreak for a parasite like Cyclospora challenging for epidemiologists.
What else makes Cyclospora hard to detect?
Da Silva: We often find it in very small amounts. Because it’s small and there may only be a few parasites in any positive food sample, we would never find it by looking under a microscope, which is the method traditionally used to find parasites. It would be like finding a needle in a haystack.
With Cyclospora, you might have, say, five or so parasites in a half a cup of shredded lettuce and it’s possible that you could eat the part of the lettuce without the parasite and be fine, and someone else could eat the other part with the parasite and get sick. We saw this in 2000, at the time I was working at CDC and we were investigating an outbreak associated with wedding cake. The cake was sampled several times and we would see certain areas of the cake were positive and others were negative, meaning that the parasite wasn’t distributed evenly throughout the cake. This means labs often have to test several samples, and even break those samples into smaller samples. Even then you may only get a few positives from many samples, like one or two. That’s typical.
What prompted FDA to invest in the development of this method? Why hadn’t we focused on developing it earlier than 2013?
Da Silva: Before 2013, the only large multi-state cyclosporiasis outbreaks that were reported were those in the mid-1990s and early 2000s. All the other outbreaks seemed to be isolated or small outbreaks that didn’t cross state lines. If you check the CDC historic outbreak data for cyclosporiasis you will see this trend from 2000 to 2012; with one exception these outbreaks were relatively small. We didn’t have the data to suggest that this was a much larger problem. But around 2013, we started seeing more large, complex, and multi-state outbreaks. That year in particular, we experienced a very large outbreak in which 631 people became sick. Around the same time surveillance and diagnostic testing methods for human illness were improving, and we began to realize that Cyclospora might be a bigger problem than we previously thought.
Leadership was also a big help. FDA leadership recognized the public health need and they responded by establishing and fully supporting the Parasitology Program and the research that led to the development and validation of this method. And they acted at just the right time. In every year since 2013 we have experienced major cyclosporiasis outbreaks.
Can you walk us through the process of developing the testing method?
Murphy: Well, to give you the highlights, the process actually started before 2013. Before 2012, I had been working on approaches to improve the detection method. There was an old method that we could no longer use. Then the outbreaks in 2013 happened, and we really ramped up the efforts to get the method in place and it was around that time that we realized we needed a parasitology team. And so, we were very fortunate to get Alex, a parasitologist with a lot of experience working with Cyclospora.
We carried out many preliminary studies to identify and optimize procedures for detection of Cyclospora in food samples. The first thing we looked at was how best to recover Cyclospora from fresh produce. There had been studies in the scientific literature indicating that there were produce wash solutions that worked better than water in recovering single-cell parasites and so we did our own studies and identified one that enhanced the recovery of Cyclospora. After that, we identified an improved technique to perform extraction of genetic material (DNA) from the parasite which could be used in the regulatory labs.
Then we considered molecular detection by using real-time polymerase chain reaction (PCR), which replicates segments of DNA, and decided to adapt and further improve the PCR method that CDC was using for the detection of Cyclospora in human stool samples. There wasn’t anything else I tested that was more sensitive than the PCR method used by CDC.
Once you developed the method, what came next?
Da Silva: Once you have a method, you need to validate it to ensure that the method works as advertised. So, you give the method to a number of laboratories and give them samples and they need to use the method and get the same results. That demanded significant efforts involving several offices within the FDA, including the Office of Food and Veterinary Medicine (OFVM), CFSAN, and the Office of Regulatory Affairs (ORA).
Murphy: We encountered some unique circumstances in validating this method. It required specific lab infrastructure, a large amount of purified Cyclospora, and a lot of time and resources to prepare the number of samples needed to fulfill the requirements for validation. This limited the number of labs involved. In the end, five labs and multiple analysts created 10 different data sets that were used to validate the method. It took about one year and was completed in 2016.
Then in 2017, the results were published in the BAM, which is FDA’s Microbiological Methods & Bacteriological Analytical Manual. BAM lists the agency’s preferred laboratory techniques for the detection of pathogens and microbial toxins in food and cosmetic products.
Da Silva: After that it was official and then it went into the ORA Field labs. That is where we are now—implementation. And I told Helen when we started the validation process that this was just the beginning.
What do you mean by that?
Da Silva: Well, first you need to implement the method in the labs that are testing food. You wouldn’t build a beautiful car and leave it in your garage. It’s the same idea; you need to get it out there. We set up a training workshop which was held in late February/early March of this year. During that workshop we trained 10 analysts from all of the ORA field labs. However, only three labs so far have the capacity, meaning the tools and equipment, in place to use the method. FDA is expanding this capacity in the coming months.
What we need to do now is train more analysts to further increase ORA laboratory capacity to detect Cyclospora. We have plans to train a total of 30 analysts this fiscal year. We also need to expand the capacity to use the method across all ORA labs. With only three labs we are already getting good results, like we saw with the McDonald’s salads and our surveillance sampling.
This past summer, the FDA’s surveillance sampling of fresh herbs detected Cyclospora in cilantro at a port of entry at the U.S.-Mexico border, as well as in cilantro grown domestically. The agency refused entry of the cilantro in lots associated with the import samples that tested positive. With respect to the domestic sample that tested positive, we worked with the responsible firm to carry out a recall, thus removing contaminated product from the marketplace. We continue to investigate how the cilantro became contaminated in each case.
We will be able to expand on this kind of work, both investigative and surveillance sampling, as more analysts are trained and field labs gain the capacity to test for Cyclospora. I don’t think we can stop all outbreaks, but our new efforts are a game changer.
We also now have industry contacting us to learn about the method because they probably want to implement this method as part of their food safety controls. This is a good thing, but there is still a lot of work to be done.
So, what’s next?
Da Silva: Well for one thing, we know that Cyclospora has been an issue when it comes to imported foods and so we want to expand the use of this method to our international partners. Already we have trained two analysts from two regulatory agencies in Mexico, COFEPRIS and SENASICA. We also want to share this method and train our regulatory partners in states.
Another thing we are working on is a method that can be used to detect Cyclospora in water. This is a parasite that can live in water, and it becomes infectious and active when the water reaches certain temperatures. When it comes to outbreak investigations, water is often an important consideration because it’s used in many ways when growing and processing crops. That method has now been single-lab validated and we are planning on multi-lab validation.
We are also working with CDC on determining the genetic makeup of Cyclospora and building a network similar to the GenomeTrakr, which focuses on the whole genome sequencing of foodborne pathogens. Until recently, we didn’t know how to sequence the genome of the parasite. Now we do, but it’s difficult. This is a parasite that comes from human fecal matter and so first you need to access samples from infected people and extract the Cyclospora DNA material. In order for the genomic research to be completed you need to have a sufficient amount of the parasite and it needs to be purified, or cleaned. This limits our research capabilities when compared with the genomic work we are seeing with bacterial pathogens. I mentioned earlier that when you find a potential source during a cyclosporasis outbreak you may only get one or two positive samples so that makes this kind of work really difficult. Cyclospora isn’t something that we can grow like you can with bacterial pathogens. Instead, we really just have to work with what we have. But we are meeting these challenges and ultimately what we would like to see happen is when we have illnesses from Cyclospora, CDC and FDA will be able to look at the genetic fingerprints and tell when cases or clusters appear to be part of a common source outbreak.
We have people here working on all of this. We are a small but very functional team. We have a few people working on genomics, a few people working on developmental detection in food and water; and we are also working alongside OFVM and ORA to implement the methods in the ORA labs.
Murphy: It is multidisciplinary in that there are moving parts, but they need to be connected or else the program is not going to have the impact that it needs to have. So, while the development and validation of this method was underway, we also had a couple people who joined who were developing approaches for trying to sequence the Cyclospora genome. At the same time, we also brought someone onboard with experience in water who’s working on the water method. There are a lot of exciting things to come.
And thinking about what’s next, I just want to add that with the implementation of the FDA Food Safety Modernization Act (FSMA), people are thinking more about how to put controls in place to prevent Cyclospora contamination. FSMA also gave us more authority over food safety. Generally, there is just more awareness about Cyclospora. It makes a huge difference the more it’s out there and people talk about it.