The Return of Cyclospora in 1997: Another Outbreak of Cyclosporiasis in North America Associated with Imported Raspberries

In the spring of 1996, a multistate outbreak of cyclosporiasis linked to fresh Guatemalan raspberries occurred in the United States and Canada (1-3). Although the mode of contamination of the raspberries was not identified, the outbreak was probably caused by an attribute or practice shared by multiple, but possibly relatively few, Guatemalan farms (1, 2). During the 1996 fall and winter export season for Guatemalan raspberries, no outbreaks of cyclosporiasis were noted in North America. In anticipation of the 1997 spring export season, the Guatemalan Berry Commission voluntarily instituted various control measures on farms; these measures focused on improving sanitation, employee hygiene, and the quality of water used in agriculture (1, 4). The Commission stipulated that only farms that the Commission classified as low risk in such regards could export fresh raspberries to the United States in the spring of 1997 (beginning on 22 April 1997). Nonetheless, another multistate, multicluster outbreak of cyclosporiasis linked to Guatemalan raspberries occurred in the spring of 1997 (4-6). The Centers for Disease Control and Prevention (CDC) learned of this outbreak in early May, when several health departments reported clusters of cases of cyclosporiasis associated with events (such as banquets) held in April. The outbreak, which is described here, ended shortly after the Guatemalan government and the Guatemalan Berry Commission voluntarily suspended exports of fresh raspberries to the United States. No shipments occurred from 29 May through 14 August (4). The occurrence of raspberry-associated outbreaks of cyclosporiasis in consecutive years, as well as other outbreaks of cyclosporiasis in 1997 that were linked to mesclun lettuce and basil that did not come from Guatemala (4, 7), highlights the need for improved understanding of the biology and epidemiology of Cyclospora cayetanensis. Although much has been learned in recent years about this coccidian parasite (8-10), unresolved issues that are relevant to foodborne outbreaks include whether animal reservoirs of infection exist (11-18); the maximum rate at which excreted, noninfectious oocysts can sporulate (that is, develop two internal sporocysts, each with two internal sporozoites) and become infectious; the effects of environmental conditions on the rate of sporulation; the median infective dose of oocysts; the factors that make cyclosporiasis seasonal and influence whether infection is symptomatic (for example, causing protracted, relapsing gastroenteritis) (8, 19-21); and the therapeutic alternatives to trimethoprim-sulfamethoxazole (22) for persons intolerant toward sulfa drugs. Methods Epidemiologic Investigation A cluster of cases of cyclosporiasis was defined as two or more cases occurring among persons who 1) ate at an event during the period from 1 April through 15 June 1997, 2) developed at least one gastrointestinal symptom [such as loose stools] 12 hours to 14 days after the event, and 3) were not known to be associated with the outbreaks linked to mesclun lettuce (in March and early April in Florida [4]) or basil (from mid-June through mid-July in the metropolitan area including northern Virginia, the District of Columbia, and Baltimore [7]). At least one case per cluster had to be confirmed by laboratory testing (for example, with modified acid-fast or hot safranin staining, examination of wet mounts, or demonstration of autofluorescence or oocyst sporulation) (8-10, 23-25). Clinical case definitions for probable cases varied among clusters (Table 1). Health departments conducted retrospective cohort studies by using structured questionnaires about symptoms and event-related exposures. Table 1. Clusters of Cases of Cyclosporiasis and Laboratory-Confirmed and Probable Cases of Cyclosporiasis in the Raspberry-Associated Outbreak, by Likely Site of Acquisition of Infection, in the United States and Canada in 1997 Case-patients with sporadic cases had laboratory-confirmed cyclosporiasis, developed gastrointestinal symptoms from 1 March through 31 August, were not included in clusters or known to be associated with other outbreaks, and had not traveled outside of the United States or Canada during the 2 weeks before they became ill. The cases of persons who became ill during the period from 1 April through 15 June were defined as having occurred during the period of this outbreak. Traceback Investigation To identify sources of the implicated berries, case-patients and investigators identified establishments where the berries were bought or eaten and the dates of purchase or consumption. These establishments identified suppliers (such as distributors) and delivery dates, and suppliers provided shipping documents. A branch of a traceback corresponding to a particular supplier was considered well documented if each step from the consumers back to the country or state of origin was confirmed verbally and in writing (for example, by invoices). An entire traceback was considered well documented if all branches of the traceback were well documented. For Guatemalan berries, airway bill and invoice numbers were used to identify farms that contributed to shipments. The initials previously used to designate Guatemalan exporters (that is, A-G [1]) are used here. The U.S. Department of Agriculture's Agricultural Marketing Service supplied weekly data for sources and amounts of domestic and imported raspberries shipped in the United States. Statistical Analysis We used Epi-Info, version 6.04a (CDC, Atlanta, Georgia), for analyses. Univariate relative risks were calculated for exposure variables. Two-tailed P values were computed by using the chi-square test or, if appropriate, the Fisher exact test. A P value less than 0.05 was considered to indicate statistical significance. A relative risk was defined as infinite if the attack rate was greater than 0 among exposed persons but was 0 among the unexposed and if no row or column total in the two-by-two table was 0. A relative risk or P value was considered undefined if a row or column total was 0. Results Epidemiologic Investigation Clusters Forty-one clusters of cases of cyclosporiasis were reported (Tables 1 and 2) in association with events held during the period from 1 April through 26 May (Figure 1, top). Of the 41 events, 20 (48.8%) occurred in private residences; 16 (39.0%) occurred in restaurants, clubs, hotels, inns, and resorts; and 5 (12.2%) occurred in other locations. An estimated total of 2541 persons attended the 41 events. Information was available for 1572 attendees (61.9%); 848 had event-associated illnesses and 762 (48.5% of interviewees) were classified as case-patients. Of these 762 persons, 192 (25.2%) had laboratory-confirmed cyclosporiasis. Table 2. Comparison of Clusters of Cases of Cyclosporiasis in the Berry-Associated Outbreaks in 1996 and 1997 Figure 1. Top. Dates of 41 events associated with clusters of cases of cyclosporiasis ( n =762 cases) in the United States and Canada in April and May 1997. Bottom. The incubation periods for the cases ranged from 1 to 14 days (by definition, the incubation period was<15 days); two cases with incubation periods of 1 day and one case with an incubation period of 2 days were laboratory confirmed. The median of the event-specific median incubation periods was 7 days (based on 38 events with available data). The median interval from symptom onset (the date associated with the event's median incubation period) to notification of public health personnel of event-associated illness was 17 days (range, 5 to 59 days; based on 37 events). The index cases of clusters were brought to the attention of public health authorities in various ways. Index cases for at least 19 clusters (46.3%) were reported by laboratories. Ten of these 19 cases were reported because the laboratory was in a FoodNet site (26) and thus had enhanced laboratory surveillance for various emerging gastrointestinal pathogens (5 cases), cases of cyclosporiasis were officially reportable in the locale (3 cases), or both (2 cases). At least 6 clusters (14.6% of 41) were reported by an event attendee who learned of cyclosporiasis through the media, the Internet, or a medical journal. Fresh raspberries were the only food common to all 41 events (at 1 of these events, they quite possibly were served). For 28 (71.8%) of 39 events, the raspberries had reportedly been rinsed in water. Estimates of the numbers of raspberries per serving were available for 4 events; these estimates were 1, 4, 5, and 12 raspberries. At 16 events (39.0%), raspberries were the only type of berry served (9 events [22.0% of 41]) or were served separately from other berries (7 events [17.1%]). Fresh strawberries, blackberries, and blueberries were served at no more than 30 (73.2%), 22 (53.7%), and 20 (48.8%) events, respectively. At some events, mesclun lettuce (7 events), fresh basil (3 events), or both (2 events) probably were served; none of these events occurred in states known to have had outbreaks linked to these foods (4, 7). Consumption of raspberries was strongly associated with cyclosporiasis. For the 37 events for which information was available about more than 10% of attendees, the median event-specific attack rate, regardless of exposures, was 66.3% (range, 13.0% to 100%). The median attack rate was higher (91.7% [range, 32.5% to 100%]) among persons who ate items that contained raspberries, with or without other types of berries. Event-specific relative risks for the associations between raspberry-containing items and cyclosporiasis were elevated (>3.0 for 9 events [24.3% of 37 events]; median, 6.8 [range, 3.5 to 10.1]), infinite (16 events [43.2%]), or undefined (11 events [29.7%]) for all but 1 event (relative risk, 1.1). P values for the associations between raspberry-containing items and cyclosporiasis were undefined for 11 events (29.7%) and statistically significant for 15 events (40.5%), including 11 even