In 1983, Haitian patients with AIDS who had diarrhea were noted to have an acid-fast Cryptosporidium-like organism in their stool samples. The organism was intermediate in size between Cryptosporidium and I. belli, was morphologically similar to Cryptosporidium, and responded to therapy with trimethoprim-sulfamethoxazole. In 1985, a medical student from Cornell University Medical College (in New York City) working in Haiti developed severe intermittent diarrhea. Extensive studies in New York City showed an organism in feces from this student identical to the "Big Crypto" we had observed in Haitian patients with AIDS [5]. The diarrhea resolved without therapy, and the patient has remained well. Subsequent studies have shown prolonged but self-limited diarrheal illness in travelers to Mexico, Nepal, Morocco, Pakistan, and India; in U.S. medical personnel; and in infants and children from Peru [5, 7-10, 12, 13]. Our report describes the clinical manifestations, diagnosis, treatment, and secondary prophylaxis of Cyclospora infection in Haitian patients infected with human immunodeficiency virus (HIV).

Methods

Top Methods Results Discussion Author & Article Info References

Between June 1990 and February 1993, 4850 patients were referred to the Institut National de Laboratoire et de Recherches in Port-au-Prince, Haiti, for HIV testing. This unit is the national referral center for AIDS and provides free HIV testing. A total of 2400 adults had positive test results for HIV by enzyme immunoassay (Abbott Diagnostic, North Chicago, Illinois). Among 2400 adults who were seropositive for HIV, 840 had a history of chronic or intermittent diarrhea during the preceding 2 years and 502 currently had diarrhea. Diarrhea was defined as the passage of at least one liquid stool daily for 3 or more weeks. A total of 450 patients provided stool specimens on two different occasions within a 10-day period before the initiation of therapy.

Stool specimens were processed as previously described [3, 4]. Cultures for enteric bacterial pathogens were not routinely done because fewer than 1% of Haitian patients with AIDS who were previously studied had cultures that yielded Salmonella or Shigella species [4, 14]. Stool specimens were not examined for the presence of toxic Escherichia coli, Campylobacter species, or enteric viral pathogens. For examination, stool specimens were not concentrated and saline wet mounts were used. After formalin ethyl acetate concentration, coccidial oocysts were identified on slides stained by a variation of the modified Kinyoun acid-fast method. Specimens containing coccidia identified by wet mount or acid-fast staining were also stained with safranin, methylene blue, Giemsa, Gram, Grocott-Gomori silver stain, and auramine rhodamine [15]. The auramine rhodamine smears were examined by fluorescent microscopy. The criteria used for identification of Cyclospora species were presence of round acid-fast oocysts that were intermediate in size (8 to 9 microns) between Cryptosporidium (5 microns) and Isospora (25 x 15 microns) species.

Forty-three of 51 patients with chronic Cyclospora infection were enrolled in a prospective study of treatment and prophylaxis. Cyclospora was identified in two separate stool specimens from each patient. Patients were excluded if they were terminally ill, pregnant, or unable to be followed for a minimum of 1 month. Patients were treated with trimethoprim sulfamethoxazole (160 mg and 800 mg, respectively), given orally four times a day for 10 days. Stool examinations were repeated on the fifth day of treatment, on the tenth and final day of therapy, and weekly after therapy. Laboratory personnel were blinded as to which patients were receiving antibiotics. Patients with recurrent diarrhea and Cyclospora organisms in their stool were retreated with the same regimen and then placed on prophylaxis with trimethoprim-sulfamethoxazole (160 mg and 800 mg, respectively) orally, three times a week.

Results

Top Methods Results Discussion Author & Article Info References

During the period between June 1990 and February 1993, 51 of 450 patients seropositive for HIV who had chronic diarrhea for more than 3 weeks had Cyclospora species identified in their fecal specimens. Other protozoa identified included Cryptosporidium in 135 patients (30%); I. belli in 54 patients (12%); Giardia lamblia in 14 patients (3%); and Entamoeba histolytica in 4 patients (1%). Six patients had Cyclospora species identified in association with I. belli (2 patients), Cryptosporidium (2 patients), and G. lamblia (2 patients). A second stool examination was done only on patients with enteric pathogens noted on the initial test; 43 of 45 patients with Cyclospora as their sole pathogen had confirmation of the organism on the second stool examination. Cyclospora was not identified in a single stool examination of 50 adults with diarrhea who were seronegative for HIV and who were randomly selected from patients referred to the same facility for HIV testing. One of 50 persons without diarrhea who were seronegative for HIV had Cyclospora on a routine stool examination.

The sensitivity of other diagnostic methods for Cyclospora was evaluated using fecal specimens that gave positive results with the modified acid-fast stain. The diagnostic methods gave sensitivities of 75% for wet mount, of 30% for safranin stain, and of 23% for auramine rhodamine. Cyclospora was not identified by any of the other six staining techniques used, including Gram, Giemsa, methylene blue, Grocott-Gomori silver, Lugol iodine, and hematoxylin-eosin. Acid-fast stained organisms appeared red or pink with variable staining intensity and had 10- to 20-dot granular inclusions that tended to coalesce and to clump as crescent shape structures (Figure 1). Cyclospora organisms were nonrefractile and measured 8 to 9 microns in diameter in wet mount preparations. Safranin staining more clearly outlined the membrane, but internal structure definition was poor. Fluorescence with auramine rhodamine staining was weak and irregular, but the internal structure with granules and network was more readily visualized.

View larger version (104K): [in this window] [in a new window]   Figure 1. Concentrated fresh stool sample from a patient with AIDS, stained with modified Kinyoun stain and showing oocysts indicative of Cyclospora species. Note the variable intensity of staining and the size of the oocysts (9 microns in diameter).

The mean age of patients was 24 years (range, 25 to 28 years); there were 30 men and 13 women. The patients had a history of chronic or intermittent diarrhea for a mean of 4 months (range, 1 to 24 months). Therefore, all patients fulfilled the current World Health Organization (WHO) definition of chronic diarrhea (duration 1 month). At the time of their entry into the study, all patients had had a weight loss of 10% or more during the previous 2 months. Diffuse, crampy abdominal pain was reported in 56% of patients, and a temperature of 38.5 °C or more was documented in 35% of patients. No patients had vomiting that interfered with administration of oral medications or rehydration. Severe dehydration required using intravenous fluid in 3 patients (7%).

All 43 patients completed the 10-day course of treatment with trimethoprim-sulfamethoxazole; no adverse drug reactions were observed during primary therapy. One patient developed pruritus while receiving trimethoprim-sulfamethoxazole and isoniazid as prophylaxis. The trimethoprim-sulfamethoxazole was withdrawn and then restarted without incident. Diarrhea and abdominal pain stopped in all patients after a mean of 2.5 days (range, 1 to 5 days). Stool samples were examined for the presence of parasites on the fifth day of treatment in 23 patients and on the last day of treatment in all 43 patients. All test results for Cyclospora species were negative, including repeat examinations done on all patients 1 week after discontinuation of therapy. No other protozoa or helminths were found either during or immediately after therapy.

Twenty-eight patients provided weekly stool specimens for a minimum period of 1 month after completing therapy (mean, 9 months; range, l to 28 months). Twelve of 28 patients (44%) had recurrent diarrhea associated with the presence of Cyclospora in stool at a mean time of 2.5 months after therapy (range, 1 to 3 months). The recurrent episodes of cyclosporiasis were clinically indistinguishable from the initial episode and also responded after 2 days of trimethoprim-sulfamethoxazole therapy. After retreatment, these 12 patients were placed on trimethoprim-sulfamethoxazole prophylaxis orally three times a week. After a mean follow-up period of 7 months (range, 2 to 27 months), 1 of 12 patients had a third episode of cyclosporiasis that again responded promptly to trimethoprim-sulfamethoxazole therapy.

Sixteen of 43 patients (37%) seropositive for HIV who had cyclosporiasis had previously met criteria from the Centers for Disease Control and Prevention (CDC) for a diagnosis of AIDS (wasting syndrome, 10 patients; pulmonary tuberculosis, 5 patients; and cryptosporidiosis, 1 patient). Seven patients had diagnoses concurrently established that fulfilled criteria for AIDS (Candida esophagitis, 3 patients; isosporiasis, 3 patients; and cryptosporidiosis, 1 patient). Sixteen of the remaining 20 patients fulfilled CDC criteria for AIDS within a mean period of 4 months of their diagnosis of cyclosporiasis. The opportunistic infections identified included Candida esophagitis (3 patients); cryptosporidiosis (2 patients); tuberculosis (2 patients); anogenital herpes infection of 3 months or more (2 patients); isosporiasis (1 patient); Pneumocystis carinii pneumonia (1 patient); central nervous system toxoplasmosis (1 patient); and chronic salmonellosis (1 patient). The remaining 4 patients had not fulfilled AIDS criteria after a mean follow-up of 7.5 months (range, 5 to 11 months).

Discussion

Top Methods Results Discussion Author & Article Info References

The enteric coccidia are an important cause of chronic and intermittent diarrhea in infants, travelers, and immunocompromised patients, including those with AIDS. During the past decade, studies in Haiti have documented the occurrence of chronic diarrhea in most patients with AIDS, usually associated with either Cryptosporidium, I. belli, and now Cyclospora species [14, 16]. The organism we called "Big Crypto" was noted initially in Haitian patients with AIDS who had diarrhea in 1983. It was characterized by its size (8 to 9 microns), which was intermediate between Cryptosporidium and I. belli in diameter, and was further distinguished by its prompt response to trimethoprim-sulfamethoxazole therapy. In 1986, Soave and colleagues [5] reported finding a similar organism thought to be either an unsporulated coccidia or a fungal spore in a fecal specimen from a former Cornell medical student who developed intermittent diarrhea after working in Haiti. The morphologic characteristics of the organism were identical to what we previously observed in patients with AIDS. Ortega and colleagues [13] reported that this organism belonged to the genus Cyclospora, as determined by sporulation and excystation studies and by appearance using a transmission electron microscope.

The clinical syndrome observed in immunocompetent persons and patients with AIDS is virtually indistinguishable from that produced by I. belli and Cryptosporidium. In the largest study [9] of cyclosporiasis in immunocompetent patients, the illness was associated with a prolonged but self-limiting diarrhea lasting a mean of 43 days. The limited epidemiologic and environmental data suggested that the organism was waterborne. A case–control study [12] of travelers and foreign residents in Nepal identified Cyclospora in 11% of 964 persons with gastrointestinal symptoms but in only 1 of 96 symptom-free controls. Seven percent of residents in the U.S. Embassy community had the infection. Patients with Cyclospora infection were more likely to have consumed untreated water, and coccidial organisms of the same appearance were identified in an epidemiologically implicated water sample. Epidemiologic data from a Chicago outbreak in hospital personnel also suggested a waterborne source [8]. Several studies [13] of infants and children living in the slums of Lima, Peru, documented a high prevalence (6% to 18%) of Cyclospora infection but noted symptomatic disease in only 11% to 28% of those infected.

Cyclospora was not observed in 50 controls who were seronegative for HIV with diarrhea whose fecal specimens were tested at our clinic. We have not identified Cyclospora species from fecal specimens of more than 2000 infants hospitalized for diarrhea and rehydration at the Hospital of the University of Haiti, despite finding Cryptosporidium in up to 16% of these infants [3]. Age may be an important factor in determining Cyclospora prevalence and disease occurrence because the children from the Peruvian slum were 1 to 2 years of age, in contrast to a mean age of less than 6 months in our hospitalized infant population.

There have been few anecdotal reports of an association between Cyclospora species and prolonged diarrhea in patients with AIDS. In the present study, the prevalence of Cyclospora species in Haitian patients with AIDS was similar to that of I. belli. There are several reasons why Cyclospora has not been previously recognized as an important opportunistic infection in patients with AIDS in the United States. These include the following: 1) Routine screening for acid-fast organisms in stool is not uniformly done; 2) without precise measurements of fecal oocysts, an inexperienced observer may confuse Cyclospora with Cryptosporidium; 3) the frequent use of trimethoprim-sulfamethoxazole for therapy and prophylaxis of toxoplasmosis and P. carinii infection may be treating or preventing Cyclospora infection; and 4) the prevalence of Cyclospora may be low in countries with less environmental contamination.

The response to trimethoprim-sulfamethoxazole therapy was identical to that observed in isosporiasis with cessation of symptoms in a mean of 2.5 days. Secondary prophylaxis was successful with only a single relapse in a group of 12 patients taking prophylaxis for a mean of 7 months. Cyclospora may be considered an opportunistic infection in the presence of HIV infection because it preceded the development of AIDS in 37% of patients. The temporal relations among initiation of therapy, clearance of Cyclospora species, and cessation of symptoms also suggest that Cyclospora is a causative agent of the diarrhea rather than some other unrecognizable enteric pathogen.