点击显示 收起
Division of Parasitic Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention
Atlanta Research and Education Foundation, Atlanta, Georgia
Johns Hopkins University School of Hygiene and Public Health, Baltimore, Maryland
Hospital Arzobispo Loayza, Hospital Dos de Mayo
Asociacion Benefica Proyectos en Informática, Salud, Medecina, y Agricultura, Lima, Peru
We studied microsporidiosis in human immunodeficiency viruspositive patients in 2 Lima hospitals. Of 2652 patients, 66% were male, 6% received antiretroviral therapy (ART), and the median CD4 lymphocyte count was 131 cells/L. Sixty-seven patients (3%) had microsporidiosis; stool specimens from 56 were identified as having Enterocytozoon bieneusi of 10 different genotypes. The 2 most common genotypes, Peru-1 and Peru-2, were not associated with significant increases in chronic diarrhea; other genotypes were associated with a 4-fold increased risk. Risk factors for E. bieneusi infection segregated by genotype: contact with duck or chicken droppings and lack of running water, flush toilet, or garbage collection with genotype Peru-1 and watermelon consumption with other genotypes. Shortened survival was associated with low CD4 lymphocyte count (P < .0001), no ART (P < .0001), and cryptosporidiosis (P = .004) but not with microsporidiosis (P = .48). Our data suggest the possibility of zoonotic E. bieneusi transmission and an association with poor sanitary conditions.
Gastrointestinal infections, especially those that lead to chronic diarrhea, are among the most debilitating opportunistic infections that affect patients with AIDS in developing countries [1, 2]. Chronic diarrhea is associated with weight loss, decreased quality of life, and shortened survival among HIV-positive patients [3, 4]. In one study in Africa, chronic diarrhea was associated with a 72% mortality rate within 10 months of diagnosisa higher rate than that for any of the other specific symptom complexes examined [5]. Intestinal microsporidia, especially Enterocytozoon bieneusi, have been reported to be an important cause of chronic diarrhea in patients with AIDS [6, 7]. Because no antiparasitic agent has been approved for the treatment of E. bieneusi infection and antiretroviral therapy (ART) is still not widely available in developing countries, strategies to prevent infection would be useful. Prevention efforts have been hindered by a lack of understanding of the relative importance of waterborne, person-to-person, foodborne, and zoonotic transmission. We therefore evaluated the contribution of intestinal microsporidiosis to chronic diarrhea, risk factors associated with infection, and the influence of intestinal microsporidiosis on survival in a large group of HIV-positive patients in Lima, Peru.
SUBJECTS AND METHODS
Study population and sites.
HIV-positive patients were recruited through physician referral from the AIDS clinics of 2 public hospitals in central Lima during September 2000December 2002. Self-referral also occurred. The criteria for entry into the study were known HIV-seropositive status, the ability to provide informed consent, and collection of at least 1 stool specimen. The protocol was approved by the institutional review boards of the Centers for Disease Control and Prevention (CDC), Johns Hopkins University Bloomberg School of Public Health, Asociacion Benefica PRISMA, and the 2 study hospitals. All participants provided written, informed consent.
Each patient was asked to provide 3 stool specimens on 3 separate days and a 5-mL blood specimen for CD4 lymphocyte count. Using a structured questionnaire, the study nurse collected data on demographics, the presence and severity of gastrointestinal illness at the time of interview, treatment and disease prophylaxis, and potential exposures, including person-to-person (15 variables related to the presence of children in the home and contact with persons with diarrhea; 27 variables related to sexual practices), waterborne (12 variables related to drinking-water source, treatment, and storage; contact with surface water; and swimming), foodborne (33 variables related to the consumption of high-risk foods, such as fresh raw fruits and vegetables and food from street vendors), and zoonotic (presence of 15 specific domestic farm and pet animals in the home and contact with droppings from each of these animals) routes of transmission. The questionnaire collected risk-factor data for exposures during the month and year preceding the interview. Because the results were similar for both time periods but associations were stronger for the month than for the year preceding interview, only the data for the month preceding interview are reported here.
For the study population as a whole, follow-up was optional, but many patients chose to return and submitted stool specimens on subsequent visits as well. The cross-sectional analysis classified patients as to their parasitic infection status based on the 3 stool samples collected at the time of enrollment. Thus, patients who tested negative for microsporidia at the time of enrollment and were later diagnosed with microsporidiosis on the basis of stool samples collected >1 month after the time of enrollment were classified as microsporidia negative for the analysis.
Nested cohort study.
In February 2002, we initiated a nested cohort study to test the hypothesis that microsporidiosis was independently associated with shortened survival. We selected all known patients with microsporidiosis as of that date and recruited 3 control subjects per patient who were matched by age (±5 years), CD4 lymphocyte count (100 or >100 cells/L), and recruitment hospital. For each patient with microsporidia, the next 3 microsporidia-negative patients at the same hospital who met the matching criteria were recruited as control subjects. Patients whose microsporidiosis was detected on visits subsequent to enrollment (classified as microsporidia negative in the cross-sectional analysis) were eligible to enter the nested cohort as case patients; control subjects for these case patients were recruited in the manner described above. Patients or their family members were contacted by telephone, if possible, to determine whether each patient was alive, and, if not, the date and circumstances of death. If telephone contact was unsuccessful, a member of the study staff visited the patient at home. Cohort entry date was the date of first contact (stool specimen or interview). Patients were counted as microsporidia positive from the date of their first positive stool specimen. Cohort exit date was the date of death or of the last contact confirming that the patient was alive. Follow-up continued until April 2003.
Laboratory methods.
Stool specimens were examined for microsporidia spores by microscopy. Thin fecal smears were made, fixed with ethanol, and stained by the Weber's modified chromotrope stain [8]. Slides were examined under ×1000 magnification. Positive samples were stored in cryovials at -70°C and were shipped frozen to the laboratory at the CDC in Atlanta, Georgia. For specimens that tested positive by microscopy, molecular analysis (by polymerase chain reaction ) was conducted as described elsewhere [9]. Stool specimens were also screened for Cryptosporidium species by acid-fast smear and for Isospora and Cyclospora species by light microscopy and UV epifluorescence. Standard methods were used to screen for other ova and parasites. CD4 lymphocyte counts were determined by use of the Manual CD4 Cell Count Kit (Coulter).
Definitions and data analysis.
Two analyses were performedthe cross-sectional analysis of data from all patients at the time of enrollment and the survival analysis of the nested cohort data. The enrollment date was the date when the first stool specimen was provided. The criteria for inclusion in the cross-sectional analysis were availability of results of microscopic parasite assay for at least 1 stool specimen, epidemiologic data within 35 days of the date that the stool specimen was provided, and CD4 lymphocyte count determined within 90 days of the date that the stool specimen was provided. If a patient had 1 stool specimen that tested positive for an organism during the month after the enrollment date, the patient was considered to have infection. We defined diarrhea on the basis of the patient's report of 3 loose or liquid stools within a 24-h period. Chronic diarrhea was defined as diarrhea that lasted at least 28 days. A diarrheal episode was considered to end when the patient had 7 diarrhea-free days.
Each potential exposure variable was examined in a multivariable model that included CD4 lymphocyte count category and source population. The source population was categorized as hospital 1, hospital 2, and "referred." The referred category was composed of patients who were not regular attendees of the AIDS clinic at hospital 1 or 2. Because the study offered CD4 lymphocyte count data free of charge, many physicians from local clinics or hospitals referred their patients to us. Variables that were significant in these trivariate models (at the P = .10 level) were tested in a series of models that included CD4 lymphocyte count category, source population, and 2 exposure variables. We tested associations with different genotypes of E. bieneusi separately, when the number in each category was sufficient, and then grouped genotypes when the results were consistent. For rare genotypes (<5 per category), we examined cross-tabulations, to assess consistency.
The nested cohort data were analyzed in a left-truncated Cox proportional hazards model. Microsporidia and Cryptosporidium infections were counted from the date of the first positive stool specimen, and survival was counted from the date of enrollment until the date of death or the date of the last contact confirming that the patient was alive. Variables examined in the model included CD4 lymphocyte count at the time of enrollment and receipt of ART at any time during the illness. Some patients who were not receiving ART at the time of enrollment started a regimen during the cohort follow-up period and thus were classified as not receiving ART for the cross-sectional analysis but as receiving it in the survival analysis. Patients missing data on CD4 lymphocyte count, ART use, or vital status were excluded. Analysis was conducted by use of SAS software (version 8.0; SAS Institute).
RESULTS
A total of 2652 HIV-seropositive patients entered the study588 from hospital 1, 1191 from hospital 2, and 873 referred from other facilities. The mean age was 31 years (range, 1973 years), and 66% were men. The median CD4 lymphocyte count was 131 cells/L; only 6% of the study population reported receiving ART. Six hundred eighty men and 893 women reported that they were heterosexual, and 1068 men reported that they were homosexual or bisexual. Only 14 (<1%) participants reported ever using injection drugs. Diarrhea was reported at the time of interview by 28% of the study participants; 10% had chronic diarrhea (lasting 28 days). Of 2652 patients, 75 (3%) had microsporidiosis detected by light microscopy at the time of enrollment, and 30 additional patients had microsporidiosis detected on a subsequent visit. Of these 105 patients, 89 had microsporidia isolates amplified and genotyped; all were E. bieneusi, and 11 genotypes were differentiated [9] (table 1). For the other 16 patients with microsporidiosis detected by light microscopy, no specimens were available for molecular testing.
The cross-sectional analysis was composed of the 2506 patients who had a CD4 lymphocyte count done within 90 days of enrollment; of these, 67 had microsporidiosis, and 56 had the E. bieneusi genotype identified (table 2). Other enteric protozoa detected included Cryptosporidium species in 234 (9%) patients, Isospora belli in 64 (3%) patients, and Cyclospora cayetanensis in 34 (1%) patients. Microsporidiosis was associated with chronic diarrhea in models adjusted for CD4 lymphocyte count and other enteric protozoal infections (table 2). However, the 2 most common E. bieneusi genotypes, Peru-1 and Peru-2, were not associated with a statistically significant increase in the risk of chronic diarrhea, whereas E. bieneusi genotypes Peru-311 were associated with a 4-fold increased risk, compared with patients without these parasites.
A number of factors altered the risk of microsporidiosis in the study population (table 3). The strongest risk factor was low CD4 lymphocyte count: having a count <100 cells/L was associated with a 12-fold increased risk of microsporidiosis; 90% of patients with microsporidiosis had a CD4 lymphocyte count <100, compared with 42% of the study population. In addition, an increased risk of microsporidiosis was associated with the presence of animals in the household, especially ducks, chickens, rabbits, and sheep. Contact with duck and/or chicken droppings in the month before interview was also associated with an increased risk of microsporidiosis. The presence of dogs and cats and contact with their droppings were not associated with an increased risk. A number of factors related to poor sanitary conditions led to an increased risk of microsporidiosis, including a lack of municipal garbage collection, running water, or a flush toilet. Of the 33 variables related to the consumption of fresh fruits and vegetables and food or beverages from street vendors, only the consumption of watermelon in the month before interview was associated with an increased risk of microsporidiosis. Watermelon consumption was highly seasonal, but E. bieneusi detections showed no significant association with season, and the inclusion of season in the model did not alter the estimates for association of E. bieneusi infection with watermelon consumption.
These factors were examined separately according to genotype of E. bieneusi, initially by individual genotype, and then as Peru-1 versus all other genotypes (Peru-211). All of the variables related to animals, contact with animal droppings, and poor sanitary conditions were found to be associated with a risk of infection with E. bieneusi genotype Peru-1 only, whereas consumption of watermelon was associated with a risk of infection with E. bieneusi genotypes Peru-211 (table 4). Adjusting the model for season did not alter the association of genotypes Peru-211 with watermelon consumption. We found no increased risk of microsporidiosis (all or by genotype) associated with sexual orientation, specific sexual behaviors, contact with diaper-age children with or without diarrhea, swimming, or other water contact, with the exception of that described above.
The nested cohort study included 373 patients, of whom 85 (23%) had microsporidiosis and 104 (28%) had cryptosporidiosis. The median CD4 lymphocyte count at the time of enrollment was 37 cells/L. The mean length of follow-up was 9.7 months; by the end of follow-up, 183 (49%) patients had died. Low CD4 lymphocyte count and cryptosporidiosis were significantly associated with shortened survival, whereas receiving ART was protective (table 5). Microsporidiosis as a whole was not associated with shortened survival; analysis by genotype (categorized as Peru-1, Peru-2, and Peru-311) also did not show any significant association with shortened survival.
DISCUSSION
Since the onset of the AIDS pandemic, enteric microsporidia, especially E. bieneusi, have been reported to be human pathogens associated with chronic diarrhea and wasting in severely immunocompromised patients [10, 11]. However, some authors, having found no association between intensity of microsporidia infection and clinical symptoms, have expressed doubts about the pathogenicity of microsporidia [12, 13]. Although 1 study documented significant morbidity and high mortality among patients with intestinal microsporidiosis, the lack of a control group made it impossible to assess the disease burden attributable to microsporidiosis itself, as opposed to that of severe immunosuppression [14]. Data concerning risk factors for intestinal microsporidiosis are sparse. Previous studies have suggested that sexual and waterborne transmission of microsporidia may occur [15, 16]. Recent studies based on sequencing of the internal transcribed spacer of the rRNA gene have demonstrated that identical genotypes of E. bieneusi can infect humans and a number of mammalian and nonmammalian hosts, which suggests the additional possibility of zoonotic transmission [1719]. Our study enabled us to examine these issues and to give a comprehensive description of intestinal microsporidiosis in HIV-positive patients in Lima, Peru.
In our study population, all of the intestinal microsporidia infections examined by PCR were identified as E. bieneusi. Although the prevalence of microsporidiosis was low, our data demonstrate a strong association with chronic diarrhea in models adjusted for CD4 lymphocyte count and infection with other protozoa, which confirms that, in this patient population, E. bieneusi does indeed cause significant morbidity. In addition, we were able to identify several significant risk factors for infection: the household presence of and contact with droppings of domestic animals (especially ducks and chickens); lack of a flush toilet, municipal garbage collection, and running water; and consumption of watermelon in the month before interview. We may have missed some mild infections because we used microscopy as our primary screening tool; PCR would have been more sensitive [20] but was impractical for screening the large number of specimens collected in the study. Nevertheless, if we missed some infections, this would bias our analysis toward the null by misclassifying patients with microsporidiosis as negative; we therefore are confident that our epidemiologic results are reliable.
Additional context is provided by preliminary data from an ongoing cohort study of 273 children in Lima (V.C., unpublished data). In this cohort, 21 microsporidia infections were detected microscopically in the first 12 months of follow-up. In the several specimens tested to date by PCR, the organism was identified as E. bieneusi. Only 2 (9.5%) of the microsporidiosis episodes were associated with diarrhea; both diarrheal episodes resolved without treatment in <2 weeks.
Interestingly, specific findings in the patients with HIV were clearly associated with some genotypes of E. bieneusi and not with others, providing the first indications that there may be differences in biologic characteristics and transmission routes by genotype. Chronic diarrhea was associated with genotypes other than genotypes Peru-1 (human/A/Germany/AF101197) and Peru-2 (human/type IV/France/AF242278, cat/K/Germany/AF267141, or cattle/BEB5/USA), which suggests that there may be differences in pathogenic potential. The animal and sanitation findings held only for genotype Peru-1 and watermelon consumption for genotypes other than Peru-1. Our data demonstrate a strong association between contact with ducks and chickens and a risk of infection with E. bieneusi genotype Peru-1. To date, this E. bieneusi genotype has been detected only in humans [21, 22], and the known zoonotic E. bieneusi genotypes have been reported predominantly in mammals [17, 18]. Nevertheless, a recent study confirmed that chickens can harbor E. bieneusi [23], and, in our study, contact with poultry and their droppings remained significant risk factors for microsporidiosis in models adjusted for poor sanitary conditions. Further molecular studies of microsporidiosis in poultry and humans may help to clarify whether our finding indicates zoonotic transmission or whether the presence of poultry is a marker for other conditions that facilitate infection.
The association of consumption of watermelon, but not of other fresh fruits or vegetables, with infection with E. bieneusi genotypes Peru-211 is intriguing but difficult to explain. Watermelon, already cut and often doused with water of questionable cleanliness, is often sold on the street in Lima; however, a number of other fruits that we examined (e.g., mango and cucumber) are sold the same way and showed no association with risk of E. bieneusi infection. Outbreaks of salmonellosis associated with melons and subsequent microbiologic studies have suggested contamination in the field and the adherence of bacteria to the rind [24]; further study will be needed to elucidate the meaning of this association for transmission of E. bieneusi.
Finally, our data indicate that, although some genotypes of E. bieneusi cause chronic diarrhea, microsporidiosis does not appear to have a major impact on survival. Our nested cohort was designed specifically to address the effect of microsporidiosis on survival. Nevertheless, our analysis was able to show the negative impact of cryptosporidiosis on survival, as is consistent with findings of previous studies [25, 26]. This suggests that our statistical power would have been sufficient to detect an impact from microsporidiosis, if the effect were of a similar magnitude as for cryptosporidiosis. These data suggest that the best way to manage HIV-positive patients with symptomatic microsporidiosis is by immune reconstitution through an adequate ART regimen [27]. The real public health challenge is to increase the availability of ART in settings like Lima, where too few patients with AIDS can hope at present to receive effective treatment for their disease.
Acknowledgments
We thank our study nurses, Yrma Chuquiruna, Eleana Sanchez, Fanny Garcia, Sonia Lopez, and Nurys Cabanillas, for their dedication and hard work; Lilia Cabrera, for helping to train the nurses; Carmen Taquiri, Jacqueline Balqui, Juan Jimenez, and Manuela Verastegui, for laboratory diagnostics and specimen handling; Marco Varela, for data management; Paula Maguia, Ana Rosa Contreras, and Paola Maurtua, for administrative support; Tom Navin, Anne Moore, Jeff Jones, Tom Handzel, and James Maguire, for scientific input; and J. B. Phu and D. Sara, for technical assistance.
References
1. Colebunders R, Francis H, Mann JM, et al. Persistent diarrhea, strongly associated with HIV infection in Kinshasa, Zaire. Am J Gastroenterol 1987; 82:85964. First citation in article
2. Mwachari C, Batchelor BI, Paul J, Waiyaki PG, Gilks CF. Chronic diarrhoea among HIV-infected adult patients in Nairobi, Kenya. J Infect 1998; 37:4853. First citation in article
3. Beaugerie L, Carbonnel F, Carrat F, et al. Factors of weight loss in patients with HIV and chronic diarrhea. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 19:349. First citation in article
4. Lubeck DP, Bennett CL, Mazonson PD, Fifer SK, Fries JF. Quality of life and health service use among HIV-infected patients with chronic diarrhea. J Acquir Immune Defic Syndr 1993; 6:47884. First citation in article
5. Sewankambo NK, Gray RH, Ahmad S, et al. Mortality associated with HIV infection in rural Rakai District, Uganda. AIDS 2000; 14:2391400. First citation in article
6. Sobottka I, Schwartz DA, Schottelius J, et al. Prevalence and clinical significance of intestinal microsporidiosis in human immunodeficiency virusinfected patients with and without diarrhea in Germany: a prospective coprodiagnostic study. Clin Infect Dis 1998; 26:47580. First citation in article
7. van Gool T, Dankert J. Human microsporidiosis: clinical, diagnostic and therapeutic aspects of an increasing infection. Clin Microbiol Infect 1995; 1:7585. First citation in article
8. Weber R, Bryan RT, Owen RL, Wilcox CM, Gorelkin L, Visvesvara GS. Improved light-microscopical detection of microsporidia spores in stool and duodenal aspirates: the Enteric Opportunistic Infections Working Group. N Engl J Med 1992; 326:1616. First citation in article
9. Sulaiman IM, Bern C, Gilman R, et al. A molecular biologic study of Enterocytozoon bieneusi in HIV-infected patients in Lima, Peru. J Eukaryot Microbiol 2003; 50(Suppl):5916. First citation in article
10. Lambl BB, Federman M, Pleskow D, Wanke CA. Malabsorption and wasting in AIDS patients with microsporidia and pathogen-negative diarrhea. AIDS 1996; 10:73944. First citation in article
11. Canning EU, Hollister WS. Enterocytozoon bieneusi (microspora): prevalence and pathogenicity in AIDS patients. Trans R Soc Trop Med Hyg 1990; 84:1816. First citation in article
12. Clarridge JE III, Karkhanis S, Rabeneck L, Marino B, Foote LW. Quantitative light microscopic detection of Enterocytozoon bieneusi in stool specimens: a longitudinal study of human immunodeficiency virus-infected microsporidiosis patients. J Clin Microbiol 1996; 34:5203. First citation in article
13. Rabeneck L, Gyorkey F, Genta RM, Gyorkey P, Foote LW, Risser JM. The role of microsporidia in the pathogenesis of HIV-related chronic diarrhea. Ann Intern Med 1993; 119:8959. First citation in article
14. Dascomb K, Clark R, Aberg J, et al. Natural history of intestinal microsporidiosis among patients infected with human immunodeficiency virus. J Clin Microbiol 1999; 37:34212. First citation in article
15. Hutin YJ, Sombardier MN, Liguory O, et al. Risk factors for intestinal microsporidiosis in patients with human immunodeficiency virus infection: a case-control study. J Infect Dis 1998; 178:9047. First citation in article
16. Cotte L, Rabodonirina M, Chapuis F, et al. Waterborne outbreak of intestinal microsporidiosis in persons with and without human immunodeficiency virus infection. J Infect Dis 1999; 180:20038. First citation in article
17. Dengjel B, Zahler M, Hermanns W, et al. Zoonotic potential of Enterocytozoon bieneusi. J Clin Microbiol 2001; 39:44959. First citation in article
18. Sulaiman IM, Fayer R, Lal AA, Trout JM, Schaefer FW III, Xiao L. Molecular characterization of microsporidia indicates that wild mammals harbor host-adapted Enterocytozoon spp. as well as human-pathogenic Enterocytozoon bieneusi. Appl Environ Microbiol 2003; 69:4495501. First citation in article
19. Sulaiman IM, Fayer R, Yang C, Santin M, Matos O, Xiao L. Molecular characterization of Enterocytozoon bieneusi in cattle indicates that only some isolates have zoonotic potential. Parasitol Res 2004; 92:32834. First citation in article
20. Tumwine JK, Kekitiinwa A, Nabukeera N, Akiyoshi DE, Buckholt MA, Tzipori S. Enterocytozoon bieneusi among children with diarrhea attending Mulago Hospital in Uganda. Am J Trop Med Hyg 2002; 67:299303. First citation in article
21. Breitenmoser AC, Mathis A, Burgi E, Weber R, Deplazes P. High prevalence of Enterocytozoon bieneusi in swine with four genotypes that differ from those identified in humans. Parasitology 1999; 118:44753. First citation in article
22. Rinder H, Katzwinkel-Wladarsch S, Loscher T. Evidence for the existence of genetically distinct strains of Enterocytozoon bieneusi. Parasitol Res 1997; 83:6702. First citation in article
23. Reetz J, Rinder H, Thomschke A, Manke H, Schwebs M, Bruderek A. First detection of the microsporidium Enterocytozoon bieneusi in non-mammalian hosts (chickens). Int J Parasitol 2002; 32:7857. First citation in article
24. Centers for Disease Control and Prevention. Multistate outbreaks of Salmonella serotype Poona infections associated with eating cantaloupe from MexicoUnited States and Canada, 20002002. MMWR Morb Mortal Wkly Rep 2002; 51:10447. First citation in article
25. Manabe YC, Clark DP, Moore RD, et al. Cryptosporidiosis in patients with AIDS: correlates of disease and survival. Clin Infect Dis 1998; 27:53642. First citation in article
26. Colford JM Jr, Tager IB, Hirozawa AM, Lemp GF, Aragon T, Petersen C. Cryptosporidiosis among patients infected with human immunodeficiency virus: factors related to symptomatic infection and survival. Am J Epidemiol 1996; 144:80716. First citation in article
27. Carr A, Marriott D, Field A, Vasak E, Cooper DA. Treatment of HIV-1-associated microsporidiosis and cryptosporidiosis with combination antiretroviral therapy. Lancet 1998; 351:25661. First citation in article