Intestinal parasites in patients with diarrhea and human immunodeficiency virus infection in Zimbabwe

Objectives: To determine the prevalence of intestinal parasites and risk factors for infection associated with diarrhea in HIV-infected patients in Harare, Zimbabwe.

Design: Prospective observational study.

Methods: Single stool samples were collected from 88 HIV-infected individuals presenting with diarrhea of greater than 1 week duration. Stools were examined for intestinal parasites using modified acid fast stain, fluorescence- labeled monoclonal antibody for Cryptosporidium parvum, as well as a modified trichrome stain and a PCR-based protocol for Enterocytozoon bieneusi.

Results: C. parvum was detected in 9% (seven out of 82) of samples evaluated, but no Cyclospora was detected. E. bieneusi was detected in 18% (10 out of 55) of stool by trichrome staining and in 51% (28 out of 55) of stool examined by PCR. Risk factors for E. bieneusi infection were: living in rural areas, consumption of nonpiped water, contact with cow dung and household contact with an individual with diarrhea.

Conclusion: E. bieneusi infection was common in HIV-infected patients with diarrhea in Zimbabwe and may be acquired through person-to-person and fecal-oral transmission.

Introduction

Chronic diarrhea occurs in up to 70% of patients infected with HIV in Africa [¹]. Previous studies have shown that Cryptosporidium parvum and Isospora belli are frequently associated with diarrhea in HIV-infected patients in Africa [^2,3]. However, the role of emerging pathogens such as microsporidia and Cyclospora in HIV-associated diarrhea in Africa has not been extensively examined. In the present study, we examined the prevalence of intestinal parasites and risk factors for infection in HIV-infected patients with diarrhea in Zimbabwe using standard parasitological techniques as well as a newly-described PCR-based protocol to detect microsporidia in stool samples [⁴].

Methods

The study was performed at Harare Hospital, Zimbabwe, between July and December 1995. The study was approved by the Superintendent of Harare Hospital, the Medical Research Council of Zimbabwe and the Institutional Review Board at MetroHealth Medical Center, Cleveland, Ohio, USA. Informed consent was obtained from each patient.

HIV-infected individuals 18 years or older who presented with diarrhea were enrolled. Diarrhea was defined as two loose or watery bowel movements per day, for at least 1 week. A questionnaire regarding risk factors for intestinal infection was administered to each patient. Enzyme-linked immunosorbent assay for antibody to HIV was performed twice on each patient. CD4 cell counts and HIV viral load were not determined.

A single stool sample was obtained from each patient. Stool was smeared on to slides, air dried, and stored at 4°C. Stool was also fixed in 10% formalin, concentrated using formyl/ethyl acetate and examined by direct observation using a phase contrast microscopic. Stool smears and concentrated samples were also examined by modified acid-fast staining, fluorescence-labeled monoclonal antibody (Merflour, Meridian Diagnostics, Cincinnati, Ohio, USA) for C. parvum, and the modified trichrome technique of Weber [⁵]. PCR for Enterocytozoon bieneusi was performed on formalin-fixed stool samples as described previously [⁴]. In some cases there was insufficient stool samples to perform all tests.

Statistical analysis was performed using the Epi info 6 computer software package [Centers for Disease Control (CDC), Atlanta, Georgia, USA, 1995]. Analysis was performed by a Fisher‚s exact test for qualitative variables.

Results

Eighty-eight patients (39 female, 49 male) with a mean age of 33 years (range, 20-59 years) were enrolled. An intestinal parasite was identified in 62% (53 out of 88) of stools examined. Blastocystis hominis was identified in 12% (10 out of 85), C. parvum in 9% (seven out of 82), whereas I. belli and Giardia lamblia were each detected in 2% (two out of 77) samples examined. Ascaris lumbricoides, Schistosoma mansoni, Chylomastix mesnilii, Entamoeba coli, Ancylostoma duodenale and Endolimax nana were each found in 1% (one out of 77) of specimens examined. Cyclospora was not detected in 82 samples examined.

Using the trichrome technique, E. bieneusi was identified in 18% (10 out of 55) of samples. All 10 patients infected with E. bieneusi were classified as category C HIV disease by CDC classification [⁶]. In contrast, when samples were examined using PCR, 51% (28 out of 55) of stools were positive for E. bieneusi. No patients were classified as category A, whereas 14% (four out of 28) were classified as category B, and 86% (24 out of 28) were category C HIV disease. All patients infected with E. bieneusi, as demonstrated by the modified trichrome technique, were positive by PCR. Four patients were co-infected with C. parvum.

E. bieneusi infection, as determined by PCR, was associated significantly with rural residence, consumption of nonpiped water, contact with chickens, contact with cow dung and household contact with a patient with diarrhea (Table 1). When stratified for rural areas, contact with cow dung [relative risk (RR), 1.6; 95% confidence interval (CI), 1.2-2.1], and household contact with a diarrhea patient (RR, 1.6; 95% CI, 1.2-2.3) were still significantly associated with infection. All patients in rural areas and two individuals in urban areas consumed nonpiped water. No significant association was found between E. bieneusi infection and season of the year, sex or prior history of diarrhea. No significant risk factors were identified for infection with E. bieneusi, as determined by Weber‚s technique. Analysis of risk factors for organisms other than E. bieneusi was not performed because the number of individuals infected with these parasites was too low.

Table 1

Discussion

Diarrhea is a significant problem in HIV-infected patients in Africa [¹]. Previous studies have demonstrated that C. parvum is frequently associated with diarrhea in AIDS patients in Uganda and Zaire where the prevalence of infection ranges from 22 to 48% [^2,3]. In our study, C. parvum infection was detected in only 9% of individuals with diarrhea. Similar results have been reported previously from Harare, Zimbabwe [⁷]. The prevalence of infection with other intestinal protozoa and helminth parasites was also low. Moreover, Cyclospora, which was associated with diarrhea in Egypt and South Africa [^8,9] was not demonstrated in our study. This low incidence of enteric parasites has been documented before in Zimbabwe but the reasons are unknown [⁵].

The most striking finding of our study was the high prevalence of E. bieneusi infection. Microscopic stool examination demonstrated that 18% of individuals were infected with this parasite and PCR analysis showed that 52% were infected. A previous study, using only light microscopy, demonstrated that the prevalence of E. bieneusi infection in HIV-infected patients with diarrhea in Zimbabwe was 11% [⁷]. The use of PCR in our study suggests that the prevalence of E. bieneusi infection in HIV-infected patients may be significantly greater than reported previously. Underestimation of the prevalence of infection by light microscopy may reflect the difficulty in parasite detection using this technique as well as significant day-to-day variation in parasite number below the level of detection.

Our study also suggested a significant association between E. bieneusi infection as detected by PCR, and residence in rural areas, use of nonpurified drinking water, contact with cow dung, as well as household contact with individuals with diarrhea. No associations were found using trichrome staining. Further studies are needed to examine the relationship between these risk factors and E. bieneusi infection.

Our study has several limitations. Because only a single stool specimen was examined, the prevalence of infection may have been underestimated [¹⁰]. Furthermore, the role of bacterial and viral pathogens was not addressed. Finally, CD4 cell counts and HIV viral load were not determined and therefore the true level of immunosuppression is unknown.

In summary, our study suggests that E. bieneusi is an important cause of chronic diarrhea in AIDS patients in Zimbabwe. Infection is associated with consumption of nonpurified water, contact with cow dung, household contact with a chronic diarrhea patient and living in rural areas suggesting person-to-person and fecal-oral routes of transmission.

Acknowledegments

The authors thank J. Johnson and R. Maier for technical support.