Abstract

BackgroundStunted development and reduced fecundity of Schistosoma parasites in immunodeficient mice and the impaired ability of human immunodeficiency virus 1 (HIV-1)–infected humans to excrete schistosome eggs have been described. This study explores the effect that HIV-1–associated immunodeficiency has on the excretion of schistosome eggs in a large cohort of coinfected individuals

MethodsIn a cross-sectional survey, urine and stool samples were obtained from and HIV-1 status was determined for 1545 individuals. More extensive data, including quantitative measures of intensity of infection in schistosomiasis and immunodeficiency, were collected in the Mupfure schistosomiasis and HIV longitudinal cohort, composed of 379 participants of whom 154 were coinfected with HIV-1 and Schistosoma parasites

ResultsIn the cross-sectional survey, the overall prevalence of schistosomiasis was 43.4%, and 26.3% of the participants were infected with HIV-1. Schistosome infections were due to Schistosoma haematobium in 63.6% of cases, S. mansoni in 18.1% of cases, and dual infections in 18.4% of cases. Intensities of Schistosoma infections, measured by the number of eggs excreted and by the level of circulating anodic antigens, did not differ between HIV-1–negative and HIV-1–positive participants coinfected with S. haematobium, S. mansoni or both. CD4 cell counts were significantly lower in HIV-1–positive participants and in S. mansoni–infected HIV-1–negative participants than in other participants

ConclusionThe present study suggests that adult HIV-1–related immunodeficiency does not impair the ability to excrete eggs in low-intensity infection with S. haematobium, S. mansoni or both and that infection with HIV-1 may not have major implications for diagnosis and surveillance of schistosomiasis

If an interaction of importance exists between schistosomiasis and HIV-1 infection, 2 distinctly different diseases, this knowledge would be of most benefit to those living in the sub-Saharan region of Africa. Apart from hosting 85% of the world’s estimated 200 million people who have schistosomiasis [1, 2], this region also has the world’s highest prevalence as well as the highest total number of HIV-1–infected individuals [3]. Although antiretroviral therapy for HIV-1 infection is currently not widely available in this region, the proper diagnosis, treatment, and prevention of schistosomiasis are feasible [4–6 ], and such interventions could possibly moderate the HIV-1 epidemic

Despite separate research efforts into the pathogenesis of these diseases, few studies have explored possible interactions between them. It has been proposed that immune activation mediated through a shift in cytokine balance toward a Th2-type immune response caused by chronic infections constitutes a major factor in the pathogenesis of HIV-1 infection and AIDS in Africa [7–9 ]. Although such a connection has been demonstrated and accepted regarding the detrimental interactions between HIV-1 and leishmaniasis [10] and has been suggested for a range of soil-transmitted helminths [11], so far there are only indirect hints that schistosomiasis and HIV-1 infection may negatively affect one another [12]

Experimental murine models have shown that Schistosoma mansoni parasites are dependent on CD4 cell–mediated host immunocompetence for maturation and excretion of eggs [13–15 ]. Consequently, it has been hypothesized that Schistosoma parasites could fail to develop and be fruitful in humans who have impaired CD4 cell immunity because of concurrent infection with HIV-1

In a study of 16 HIV-1–positive and 37 HIV-1–negative participants infected with S. mansoni in Kenya, Karanja et al. found that the HIV-1–positive participants had significantly lower egg counts than did the HIV-1–negative participants, despite their having similar levels of intensity in their infections as measured by levels of circulating cathodic antigen (CCA) [16]. Karanja et al. concurrently demonstrated a positive relationship between egg counts and CD4 cell counts in HIV-1–positive participants. Another study of individuals coinfected with HIV-1 and S. mansoni also demonstrated a significantly lower egg count in coinfected participants [17]. Again, however, sample size limits the interpretation, because in that study of sugar-estate residents in Ethiopia, only 10 coinfected participants were compared with 348 HIV-1–negative participants infected with S. mansoni. In a prospective cohort study conducted in Zambia of 507 S. haematobium–infected participants of whom 73 were HIV-1 coinfected, Mwanakasale et al. found significantly lower egg counts in the coinfected participants, but unfortunately they were unable to quantify either the intensity of the S. haematobium infection or the level of immunodeficiency by other measures [18]

Because quantitation of eggs is the classic and simplest way to approximate the intensity of schistosomiasis [19], the hypothesis that HIV-1–coinfected individuals have defective excretion of schistosome eggs deserves to be assessed in a large study that includes complementary measures of schistosomiasis intensity and HIV-1–induced immunodeficiency. In the present study, we present the prevalence and intensity of active schistosomiasis, as determined by egg count and detection of schistosome circulating anodic antigen (CAA) in an HIV-1–infected population and describe the effect that HIV-1–associated immunodeficiency has on the excretion of schistosome eggs

Participants, Materials, and Methods

Setting and study populationThe study was conducted from October 2001 to June 2003 in Mupfure and adjacent areas in Shamva District, Mashonaland Central Province, Zimbabwe. This rural area is characterized by subsistence farming, and the main source of water for irrigation, bathing, and washing is the Mupfure River, which is infested mainly by Bulinus species snails but also by Biomphalaria species snails [20]. No schistosomiasis control program has previously targeted the adult population here

The study population was composed of adults (⩾18 years old) residing in the area who were willing to submit urine, stool, and blood samples and to be tested for HIV-1. Recruitment of participants was achieved through community meetings and was facilitated by local village health workers

The Medical Research Council of Zimbabwe and the Central Medical Scientific Ethics Committee of Denmark approved the present study, and informed consent was obtained from all participants. In addition, permission was given by the provincial medical director of Mashonaland Central Province, the district medical officer of Shamva District, and the village leaders

Screening procedure for HIV-1 serological and parasitological testingScreening procedures were performed on 2281 participants. HIV-1 testing was performed confidentially, and pretest and posttest counseling was provided in the participants’ native language (Shona) by qualified personnel. In the field, a rapid HIV-1/2 test kit was used on a dry blood spot (Determine; Abbott Laboratories). All individuals who were initially found to be HIV-1 positive were retrospectively retested using a different rapid test kit (Oraquick by Orasure Technologies, Serodia by Fujirebio, or Capillus by Trinity Biotech). For participants subsequently included in the cohort, 2 ELISAs were performed on serum samples; 1 test was performed in Harare (Recombigen; Cambridge Biotech), and the other was performed in Copenhagen (Ortho). No discrepancies were found between the results of the initial rapid HIV-1/2 test and those of the 2 subsequent ELISAs

Microscopic examination of fixed-volume urine samples filtered on Nytrel filters (VesterGaard Frandsen) was used to identify and quantitate eggs of S. haematobium by the syringe urine filtration technique [21]. Because of the diurnal and day-to-day variation in egg output, the urine samples were collected on 3 consecutive days [22, 23]. The modified formol-ether concentration technique was used on 1 stool sample from each participant to detect eggs of S. mansoni and other helminth eggs or parasites [24]. The use of exactly 1 g of stool for the procedure rendered this technique appropriate for quantitative diagnosis of S. mansoni infection and was particularly useful in individuals with low-intensity infections [25]. At all stages, when performing parasitological examinations, the technician was blinded to clinical and serological information

Establishment of cohortAfter the screening procedure, HIV-1–infected individuals who were coinfected with Schistosoma parasites were included in a prospective cohort. Simultaneously, a number of HIV-1–negative but schistosomiasis-positive individuals were included as controls, as were individuals infected only with HIV-1 and individuals with neither infection. Allocation of participants to these 3 control groups was done randomly: 1 participant with schistosomiasis was selected for every 2 coinfected participants, and 1 HIV-1–positive participant or 1 healthy participant was selected for every 4 coinfected participants. The 379 participants included in the cohort were interviewed to obtain sociodemographic data and medical history, and a clinical examination was performed

Exclusion criteria were applied to participants presenting with clinical signs/symptoms of tuberculosis, terminal stages of schistosomiasis, or severe anemia, but no participants were excluded for these reasons. Pregnant women were excluded from the study but were diagnosed and were offered praziquantel as treatment for schistosomiasis after delivery and termination of breast-feeding

Forty milliliters of blood was drawn from each participant. CD4 cell counts were measured by flow cytometry (FacsCalibur; Becton Dickinson) at the Department of Hematology, Parirenyatwa Hospital, Harare, where full blood counts were also performed (Hematology Analyzer SF 3000; Sysmex) within 36 h of collection of the samples. Levels of CAA, which originates from the parasite gut and is a unique marker of an active schistosome infection, were measured in serum samples by an ELISA performed at the Department of Parasitology in Leiden [26–28 ]

Statistical analysesAll statistical analyses were performed using SAS software (version 8.2; SAS Institute). Egg counts and CAA levels were log-transformed to approximate normal distribution. Because detection of CAA does not allow for species to be distinguished, results from egg counts were used to stratify the schistosomiasis status of the participants into 4 subgroups (no schistosomiasis, infection with S. haematobium only, infection with S. mansoni only, and infection with both species). A 2-way analysis of variance (ANOVA), with HIV-1 status and schistosomiasis status as classifying variables, was used to identify differences between groups with respect to egg counts, age, CD4 cell counts, and blood subsets. When this test revealed an interaction, a subsequent 2-way ANOVA, with the 2 schistosome species as classifying variables, was applied within each HIV-1 stratum to explore the magnitude of this effect. If there was no interaction, the ANOVA was performed on the combined HIV-1 strata. A t test was performed to evaluate differences in egg counts between HIV-1 groups and was complemented with an analysis of covariance (ANCOVA) to allow for adjustments according to age, sex, CAA levels, and schistosome species. The magnitude of effects was evaluated by back transformation of the log-transformed difference in means and 95% confidence intervals (CIs) between groups. Hence, the difference is presented as a ratio of means and 95% CIs. Multiple regression models of egg counts with the covariates CD4 cell counts, CAA levels, age, sex, schistosomiasis status, and HIV-1 status were performed, and results are presented as regression coefficients (RCs) and 95% CIs. Normality and independence of residuals were checked graphically. Regarding the ANCOVA, slopes did not differ. P<.05 was considered to be significant

Results

Cross-sectional surveyFrom the initially screened 2281 individuals, we obtained 3 consecutive urine samples, 1 stool sample, and information on HIV-1 status from 1545 individuals. Overall, 26.3% were HIV-1 positive, 43.4% had schistosomiasis, and there were no differences in distribution of schistosome infections according to HIV-1 status (table 1)

Table 1

Sex and age distribution and prevalence of schistosome and HIV-1 infections in the cross-sectional survey

Table 1

Sex and age distribution and prevalence of schistosome and HIV-1 infections in the cross-sectional survey

The prevalence of other helminth infections was negligible: very few individuals were diagnosed with Taenia saginata (n = 2), Strongyloides stercoralis (n = 2), or Trichuris trichiura (n = 1). Other identified intestinal parasites were the protozoans Entamoeba histolytica (n=54; 47 [4.7%] were HIV-1–negative participants and 7 [1.7%] were HIV-1–positive participants; P<.05) and Giardia lamblia (n=9; 8 were HIV-1–negative participants and 1 was an HIV-1–positive participant). Table 1 also presents the sex and age distribution of the screened population. Predominantly women participated in the present study (75%), and there were no differences in sex distribution across the subgroups according to HIV-1 status or schistosomiasis status

In a 2-way ANOVA of age by HIV-1 status and schistosomiasis status, there was no effect attributable to the interaction (P=.17), and there was no effect attributable to HIV-1 status (P=.30; mean difference in uninfected vs. infected participants, 1.0 years; 95% CI, −0.9 to 2.8 years), but a significant difference was attributable to the effect of schistosomiasis (P < .001). A subsequent 2-way ANOVA of S. haematobium infection and S. mansoni infection on the combined HIV-1 strata revealed no interaction (P=.22), a weak tendency toward lower age in S. mansoni–infected participants (P=.13; mean difference in uninfected vs. infected participants, 1.3 years; 95% CI, −0.4 to 3.0 years), and a significantly lower age in S. haematobium–infected participants (P<.0001; mean difference in uninfected vs. infected participants, 4.8 years; 95% CI, 3.1–6.5 years)

Egg counts are shown in figure 1A and 1B. In a 2-way ANOVA of egg counts in urine and stool by HIV-1 status and schistosomiasis status, there was no interaction (urine egg counts, P = .45; fecal egg counts, P=.22), and no difference was attributable to HIV-1 status (urine egg counts: P=.30; mean HIV-1–positive participants:HIV-1–negative participants, 0.83; 95% CI, 0.59–1.18; fecal egg counts: P=.65; mean HIV-1–positive participants:HIV-1–negative participants, 0.93; 95% CI, 0.68–1.27). There was, however, an effect of schistosomiasis status (urine egg counts: P<.01; mean S. mansoni–positive participants:S. mansoni–negative participants, 1.64; 95% CI, 1.16–2.33; fecal egg counts: P=.01; mean S. haematobium–positive participants:S. haematobium–negative participants, 1.50; 95% CI, 1.10–2.04). Both tests showed increased egg counts when coinfection was present

Figure 1

A Mean urine egg count by HIV-1 status. No difference was found between the HIV-1 groups. Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). Egg counts have been adjusted to median age and female sex. B Mean fecal egg count by HIV-1 status. No difference was found between HIV-1 groups. Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). Egg counts have been adjusted to median age and female sex

Figure 1

A Mean urine egg count by HIV-1 status. No difference was found between the HIV-1 groups. Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). Egg counts have been adjusted to median age and female sex. B Mean fecal egg count by HIV-1 status. No difference was found between HIV-1 groups. Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). Egg counts have been adjusted to median age and female sex

Because the ANOVA revealed no significant interaction between HIV-1 status and schistosomiasis status, the direct effect of HIV-1 status was further analyzed by t tests and ANCOVA. The t test comparing urine egg counts confirmed no difference according to HIV-1 status for S. haematobium–infected participants (P=.51; mean HIV-1–positive participants:HIV-1–negative participants, 0.91; 95% CI, 0.67–1.22). An ANCOVA that tested the same relationship but was adjusted for age and sex gave similar results (P=.68; mean ratio, 0.94; 95% CI, 0.70–1.26), with a significant effect according to age (P<.0001; see regression analyses) and a tendency for men to have higher egg counts (P=.06; RC, 0.12 log10 egg count in men/log10 egg count in women; 95% CI, −0.005 to 0.247 log10 egg count in men/log10 egg count in women). Further adjustments for S. mansoni status did not alter the results. Also, application of the t test to fecal egg counts in S. mansoni–infected individuals revealed no difference according to HIV-1 status (P=.55; mean HIV-1–positive participants:HIV-1–negative participants, 0.91; 95% CI, 0.67–1.24), and the equivalent ANCOVA adjusted for age, sex, and S. haematobium infection gave similar results

Cohort studyOf the 379 participants included in the prospective cohort, complete baseline information was available on 356 participants. The characteristics of the cohort population and the distribution of the schistosome and HIV-1 infections are presented in table 2. Mean body mass index (BMI) of the HIV-1–positive participants was lower than that of the HIV-1–negative participants (P<.001), but there were no differences in BMI between the schistosomiasis subgroups within the HIV-1 strata

Table 2

Sex, age, body mass index (BMI), and distribution of schistosome and HIV-1 infections and circulating anodic antigen (CAA) levels in the cohort study

Table 2

Sex, age, body mass index (BMI), and distribution of schistosome and HIV-1 infections and circulating anodic antigen (CAA) levels in the cohort study

There were no differences in egg counts attributable to either the interaction between HIV-1 and schistosomiasis or to HIV-1 status or schistosomiasis status separately. Performance of the ANCOVA on urine egg count as in the cross-sectional survey and adjustment by CAA level, age, sex, and S. mansoni infection still did not reveal an effect attributable to HIV-1 status (P = .28; mean HIV-1–positive participants:HIV-1–negative participants, 0.87; 95% CI, 0.61–1.24), nor did the equivalent ANCOVA performed on fecal egg count change conclusions from those found in the cross-sectional survey when adjustments were made for CAA level, age, sex, and S. haematobium infection (P = .98; mean HIV-1–positive participants:HIV-1–negative participants, 0.99; 95% CI, 0.49–2.27)

As expected, there were significantly lower CD4 cell counts in the HIV-1–positive participants than in the HIV-1–negative participants (P<.0001; mean difference, 552 cells/μL; 95% CI, 482–622 cells/μL) (figure 2). The 2-way ANOVA showed an effect of the interaction between HIV-1 status and schistosomiasis status (P=.02). The subsequent 2-way ANOVA of S. haematobium and S. mansoni in HIV-1–negative participants revealed no interaction (P=.20) and no effect of S. haematobium infection (P=.89; mean difference in uninfected vs. infected participants, 13 cells/μL; 95% CI, −176 to 202 cells/μL) but a significant effect of S. mansoni infection (P=.01; mean difference in uninfected vs. infected participants, 239 cells/μL; 95% CI, 52–428 cells/μL). In the HIV-1–positive participants, no effects attributable to schistosome species were found (figure 2)

Figure 2

CD4 cell count by Schistosoma infection and HIV-1 status. The marked difference in CD4 cell count is attributable to HIV-1 status (P < .0001). Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). HIV-1–positive participants are indicated by dark box plots

Figure 2

CD4 cell count by Schistosoma infection and HIV-1 status. The marked difference in CD4 cell count is attributable to HIV-1 status (P < .0001). Data are presented as group means (+), medians (center lines) 25th–75th percentiles (boxes) and extremes (whiskers). HIV-1–positive participants are indicated by dark box plots

Regression Analyses

Cross-sectional surveyIn a multiple linear regression performed on urine egg count in S. haematobium–infected participants with the covariates age and HIV-1 status, there was a tendency toward the interaction of HIV-1 status and age (P = .07; RC, 0.12 log10 egg count/1 unit increase in HIV-1 status × 10 years; 95% CI, −0.01 to 0.25 log10 egg count/1 unit increase in HIV-1 status × 10 years). When HIV-1 strata were combined, lower egg count was associated with higher age (P < .001; RC, −0.13 log10 egg count/10 years; 95% CI, −0.18 to −0.08 log10 egg count/10 years; R2=0.05). However, because of the tendency toward interaction, separate regressions were performed on each HIV-1 stratum. In HIV-1–positive participants, age had no evident effect (P=.68; RC, −0.03 log10 egg count/10 years; 95% CI, −0.15 to 0.10 log10 egg count/10 years; R2<0.01), whereas in HIV-1–negative participants, higher age was associated with lower egg count (P<.0001; RC, −0.15 log10 egg count/10 years; 95% CI, −0.20 to −0.10 log10 egg count/10 years; R2=0.08). Regressions were performed on both HIV-1 strata, with adjustments for sex, and both had similar results (data not shown). We cannot conclude that there is a difference between the HIV-1 strata in these correlations, because the 95% CIs of slopes just meet, but we note that the negative correlation between age and urine egg count is found only in HIV-1–negative participants

In a multiple linear regression performed on fecal egg count in S. mansoni–infected participants, no interaction between age and HIV-1 status was found (P=.70; RC, −0.03 log10 egg count/10 years; 95% CI, −0.17 to 0.11 log10 egg count/10 years; R2<0.01). When HIV-1 strata were combined, age had no effect (P=.87; RC, −0.005 log10 egg count/10 years; 95% CI, −0.061 to 0.051 log10 egg count/10 years)

CohortIn a multiple linear regression performed on urine egg count, there was no interaction between CAA level and HIV-1 status (P=.48; RC, −0.031 log10 egg count/1 unit increase in HIV-1 status × 10-fold increase in CAA level; 95% CI, −0.118 to 0.056 log10 egg count/1 unit increase in HIV-1 status × 10-fold increase in CAA level). When HIV-1 strata were combined, we found the expected effect of CAA level (P<.0001; RC, 0.114 log10 egg count/10-fold increase in CAA level; 95% CI, 0.071–0.158 log10 egg count/10-fold increase in CAA level; R2=0.11). There was an interaction between CAA level and HIV-1 status on fecal egg counts (P<.05; RC, 0.166 log10 egg count/1 unit increase in HIV-1 status × 10-fold increase in CAA level; 95% CI, 0.030–0.302 log10 egg count/1 unit increase in HIV-1 status × 10-fold increase in CAA level). Separate modeling of HIV-1–positive participants demonstrated an effect (P<.01; RC, 0.146 log10 egg count/10-fold increase in CAA level; 95% CI, 0.041–0.251 log10 egg count/10-fold increase in CAA level; R2=0.18) not seen in HIV-1–negative participants (P=.59; RC, −0.020 log10 egg count/10-fold increase in CAA level; 95% CI, −0.099 to 0.059 log10 egg count/10-fold increase in CAA level; R2=0.02). We note here that the 95% CI overlap between the HIV-1 groups and adjustments by age, sex, and Schistosoma species did not alter these results

In a multiple linear regression performed on urine egg count in S. haematobium–infected individuals with the covariates age, sex, CD4 cell count, HIV-1 status, and the interaction of CD4 cell count and HIV-1 status, we found significance toward the interaction (P<.01; RC, 0.08 log10 egg count/1 unit increase in HIV-1 status × increase of 100 cells in CD4 cell count; 95% CI, 0.03–0.13 log10 egg count/1 unit increase in HIV-1 status × increase of 100 cells in CD4 cell count; R2=0.05). A significant effect of CD4 cell count on urine egg count was found in HIV-1–positive participants (P<.01; RC, 0.06 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, 0.02–0.10 log10 egg count/increase of 100 cells in CD4 cell count; R2=0.07), and a tendency toward interaction in HIV-1–negative participants was also found (P=.08; RC, −0.04 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.07 to 0.002 log10 egg count/increase of 100 cells in CD4 cell count; R2=0.08) in this age-adjusted and sex-adjusted model (figure 3A). Adjusting for CAA level and S. mansoni infection did not alter these results

Figure 3

A Scatterplot of mean urine egg counts and CD4 cell count for HIV-1–positive participants (× and solid line) and HIV-1–negative participants (circles and dotted line). Egg counts were adjusted for age, sex, Schistosoma mansoni infection, and circulating anodic antigen (CAA) level. Multiple regression, HIV-1–negative participants: P=.08; regression coefficient (RC), −0.029 log10 egg count/increase of 100 cells in CD4 cell count; 95% confidence interval (CI), −0.062 to 0.003 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.22; multiple regression, HIV-1–positive participants: P<.05; RC, 0.049 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.007 to 0.091 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.13. B Scatterplot of fecal egg count and CD4 cell count for HIV-1–positive participants (× and solid line) and HIV-1–negative participants (circles and dotted line). Egg counts were adjusted for age, sex, S. haematobium infection, and CAA level. Multiple regression, HIV-1–negative participants: P=.75; RC, 0.01 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.04 to 0.05 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.13; multiple regression, HIV-1–positive participants: P=.31; RC, 0.02 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.02 to 0.06 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.26

Figure 3

A Scatterplot of mean urine egg counts and CD4 cell count for HIV-1–positive participants (× and solid line) and HIV-1–negative participants (circles and dotted line). Egg counts were adjusted for age, sex, Schistosoma mansoni infection, and circulating anodic antigen (CAA) level. Multiple regression, HIV-1–negative participants: P=.08; regression coefficient (RC), −0.029 log10 egg count/increase of 100 cells in CD4 cell count; 95% confidence interval (CI), −0.062 to 0.003 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.22; multiple regression, HIV-1–positive participants: P<.05; RC, 0.049 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.007 to 0.091 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.13. B Scatterplot of fecal egg count and CD4 cell count for HIV-1–positive participants (× and solid line) and HIV-1–negative participants (circles and dotted line). Egg counts were adjusted for age, sex, S. haematobium infection, and CAA level. Multiple regression, HIV-1–negative participants: P=.75; RC, 0.01 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.04 to 0.05 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.13; multiple regression, HIV-1–positive participants: P=.31; RC, 0.02 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.02 to 0.06 log10 egg count/increase of 100 cells in CD4 cell count; R2 = 0.26

A similar multiple linear regression was performed on fecal egg count. No interaction between CD4 cell count and HIV-1 status was found (P=.99; RC, 0.000 log10 egg count/1 unit increase in HIV-1 status × increase of 100 cells in CD4 cell count; 95% CI, −0.059 to 0.060 log10 egg count/1 unit increase in HIV-1 status × increase of 100 cells in CD4 cell count; R2=0.16), and no effect of CD4 cell count was found (P=.20; RC, 0.030 log10 egg count/increase of 100 cells in CD4 cell count; 95% CI, −0.017 to 0.077 log10 egg count/increase of 100 cells in CD4 cell count) (figure 3B)

Discussion

The most important conclusion from the present study is that no differences in egg counts were found between HIV-1–positive participants and HIV-1–negative participants infected with S. haematobium, S. mansoni or both. This finding was the same even after adjustments were made for differences in age, sex, and dual infection in the large survey population and in the cohort, in which further adjustment for CAA level was possible

This conclusion is in contrast to those of previous studies by Karanja et al. [16] and Mwanakasale et al. [18], who found a significantly lower egg count in S. mansoni– and S. haematobium–infected participants, respectively, who were coinfected with HIV-1, compared with that in HIV-1–negative participants. These findings—and in particular those of Karanja et al., which are supported by the demonstration of similar levels of CCA (another schistosome gut–associated antigen) in the 2 groups and correlations between egg counts and CD4 cell counts—have led to speculation of an immunomodulatory inhibition of the human host’s ability to excrete Schistosoma eggs when immunodeficient because of HIV-1 coinfection. Complementary to these suggestions are the findings of Doenhoff et al., Davies et al., and other researchers who demonstrated attenuated development of S. mansoni—characterized by overall reduction in body size, delayed development, and a reduction in fecundity—in immunocompromised mice [13–15 , 29, 30]. Our data set is larger than that of both Karanja et al. and Mwanakasale et al. [16, 18], in terms of both the combined schistosomiasis subgroups and each subgroup considered separately. Using our ANCOVA models, we estimate that HIV-1–positive participants excrete 94% and 91% of the eggs in urine and feces, respectively, that their HIV-1–negative counterparts excrete. Because with lower CIs, HIV-1–positive participants excrete at least 70% and 67% of the eggs in urine and feces, respectively, that their HIV-1–negative counterparts excrete, these data argue against any clinical relevant difference between egg counts in HIV-1–positive participants and HIV-1–negative participants

Other explanations may contribute to our contrasting findings. In particular, the average intensities of both S. haematobium and S. mansoni infections in the present study groups were considerably lower than those in the groups studied by Karanja et al. and Mwanakasale et al. It is plausible that the ability to excrete eggs is more disturbed by reduced immunocompetence when the worm burden is larger. The observed differences in intensities may also be the result of differences in age between the study populations. In the study by Karanja et al., the mean±SD age was 27.7±9.6 years, whereas in the study by Mwanakasale et al., the mean±SD age of HIV-1–positive participants was 23.0±11.1 years, and the mean ± SD age of HIV-1–negative participants was 18.8±9.5 years. In contrast to this, in the present study population, the mean ± SD age of HIV-1–negative participants was 37.1±13.4 years, and the mean±SD age of HIV-1–positive participants was 35.4 ± 9.7 years. Mwanakasale et al. reported that the mean age of the HIV-1–positive participants in their study was significantly higher than that of the HIV-1–negative participants. In the present study, the age did not differ significantly between groups, and the conclusions did not change when adjustments were made for age. Furthermore, our results are supported by CAA measurements indicating similar levels of intensity in active schistosome infections and allowing for CAA adjustments

In accordance with the findings of Karanja et. al., we also found lower egg counts in both urine and feces with lower CD4 cell counts in HIV-1–positive participants. However, the association was weak and it was evident that even HIV-1–positive participants with very low CD4 cell counts could have egg counts close to those observed in HIV-1–negative participants. It is notable that the HIV-1–positive participants had very low CD4 cell counts, indicating that we primarily studied HIV-1–positive participants with advanced infections. This might restrict our conclusions to refer to only moderately to severely immunocompromised individuals

The observation of measurable CAA levels in individuals not excreting eggs (table 2) is probably the result of false negative egg counts, because the ELISA for CAA is highly specific. It is known that the egg count can vary greatly, so that even after examination of repeated stool or urine samples, not all schistosome-positive cases will be found [31]. In the presentation of our data, however, the egg count was used as the primary diagnostic criteria for schistosomiasis, primarily because that was the only tool available in the field work of the study (this formed the basis for treatment decisions), but also because use of the CAA level would not permit a distinction between schistosome species. In the subsequent data analyses, adjustment for CAA level did not alter the conclusions

The present results differ from the findings of the murine experimental models, because we did not find lower egg counts despite marked immunodeficiency in these participants and because the CAA levels did not suggest reduced intensities of Schistosoma infection because of immunodeficiency. Apart from the evident differences between an observational study with humans and an experimental animal model, another important difference may be that, in the murine model, the mice are initially immunocompromised and subsequently exposed to schistosomes. In the present study, the sequence is likely the reverse: the participants probably had been exposed to schistosomes as children, before they acquired their HIV-1–related immunodeficiency. It is possible that, under these circumstances, an adaptation of the host-parasite relationship can develop that leads to other outcomes, including perhaps both normal parasite development and excretion of eggs

To our knowledge, this is the first report on findings of interactions between HIV-1 and concurrent infections with 2 schistosome species, and it is interesting to observe that higher egg counts are the apparent outcome of simultaneous schistosomiasis. A behavioral explanation might be that an individual dually infected with Schistosoma parasites has been more heavily or more frequently exposed and therefore harbors an infection of higher intensity

Interestingly, our cohort data showed decreased CD4 cell counts in S. mansoni–infected HIV-1–negative participants. This finding is in agreement with other reports of an association between CD4 lymphocytopenia and helminthic and other infections [32, 33]. These differences in CD4 cell counts apparently disappear in S. mansoni–infected HIV-1–positive participants, indicating that any possible subtle effect of schistosomiasis on CD4 cell counts seems to be masked by the dramatic HIV-1–related decline in CD4 cell counts

We did find that HIV-1–positive participants had a lower BMI than did HIV-1–negative participants. However, within their respective HIV-1 groups, there were no differences in BMI between subgroups of participants with schistosomiasis. Surprisingly, even in the HIV-1–negative participants, there were no differences in BMI between the participants without schistosomiasis and the other subgroups. This similarity may reflect that the intensities of schistosomiasis in the population in the present study were moderate to low and therefore did not give rise to physiological implications that affect BMI

In summary, despite the difference between the intensities of the Schistosoma infections in the population in the present study and those in other studies, it appears that HIV-1–induced immunodeficiency does not impair the ability of participants with low-intensity schistosomiasis to excrete either S. haematobium or S. mansoni eggs. Our results further question the applicability of murine studies that show a dependency of adequate CD4 cell immunity for S. mansoni development and fecundity to human conditions. One may assume that, in our cohort, some antischistosome immune responses were already established when HIV-1 was encountered. This may imply that the CD4 cells acquired the capacity to react with the parasite so that it could undergo a normal life cycle. However, after this process is established, a secondary CD4 cell immunodeficiency, such as the one that occurs in HIV-1–positive participants, does not appear to dramatically change the development of S. haematobium or S. mansoni or the ability of the human host to excrete their eggs

Acknowledgments

We thank the Mupfure community and in particular the village health workers and the environmental health technician, for the willing participation and contribution to our study; Mupfure Secondary School, for accommodations; the technical team and in particular its core members—E. N. Kurewa, N. Taremeredzwa, W. Mashange, C. Mukahiwa, S. Nyandoro, W. Soko, B. Mugwagwa, and E. Mashiri—for tireless, hard work under difficult circumstances; D. Kornelis of the Department of Parasitology at the Leiden University Medical Center, the Netherlands, for diligent and accurate testing of the serum samples for circulating anodic antigen; and the Department of Hematology of Parirenyatwa Hospital, Harare, and in particular B. Mudenge, for continual laboratory support

References

1.
Chitsulo
L
Engels
D
Montresor
A
Savioli
L
The global status of schistosomiasis and its control
Acta Trop
 , 
2000
, vol. 
77
 (pg. 
41
-
51
)
2.
Engels
D
Chitsulo
L
Montresor
A
Savioli
L
The global epidemiological situation of schistosomiasis and new approaches to control and research
Acta Trop
 , 
2002
, vol. 
82
 (pg. 
139
-
46
)
3.
Joint United Nations Progamme on HIV/AIDS (UNAIDS)
Report on the global HIV/AIDS epidemic 2002
 , 
2002
Geneva
UNAIDS
4.
Prevention and control of schistosomiasis and soil‐transmitted helminthiasis
World Health Organ Tech Rep Ser
 , 
2002
, vol. 
912
 pg. 
i‐57,
  
back cover
5.
Crompton
DW
Engels
D
Montresor
A
Neira
MP
Savioli
L
Action starts now to control disease due to schistosomiasis and soil‐transmitted helminthiasis
Acta Trop
 , 
2003
, vol. 
86
 (pg. 
121
-
4
)
6.
Montresor
A
Crompton
DW
Gyorkos
TW
Savioli
L
Helminth control in school‐age children: a guide for managers of control programmes
 , 
2002
Geneva
World Health Organization
7.
Bentwich
Z
Kalinkovich
A
Weisman
Z
Immune activation is a dominant factor in the pathogenesis of African AIDS
Immunol Today
 , 
1995
, vol. 
16
 (pg. 
187
-
91
)
8.
Bentwich
Z
Kalinkovich
A
Weisman
Z
Grossman
Z
Immune activation in the context of HIV infection
Clin Exp Immunol
 , 
1998
, vol. 
111
 (pg. 
1
-
2
)
9.
Bentwich
Z
Maartens
G
Torten
D
Lal
AA
Lal
RB
Concurrent infections and HIV pathogenesis
AIDS
 , 
2000
, vol. 
14
 (pg. 
2071
-
81
)
10.
Harms
G
Feldmeier
H
HIV infection and tropical parasitic diseases—deleterious interactions in both directions?
Trop Med Int Health
 , 
2002
, vol. 
7
 (pg. 
479
-
88
)
11.
Fincham
JE
Markus
MB
Adams
VJ
Could control of soil‐transmitted helminthic infection influence the HIV/AIDS pandemic
Acta Trop
 , 
2003
, vol. 
86
 (pg. 
315
-
33
)
12.
Mwinzi
PN
Karanja
DM
Colley
DG
Orago
AS
Secor
WE
Cellular immune responses of schistosomiasis patients are altered by human immunodeficiency virus type 1 coinfection
J Infect Dis
 , 
2001
, vol. 
184
 (pg. 
488
-
96
)
13.
Doenhoff
MJ
Pearson
S
Dunne
DW
, et al.  . 
Immunological control of hepatotoxicity and parasite egg excretion in Schistosoma mansoni infections: stage specificity of the reactivity of immune serum in T‐cell deprived mice
Trans R Soc Trop Med Hyg
 , 
1981
, vol. 
75
 (pg. 
41
-
53
)
14.
Doenhoff
MJ
Hassounah
O
Murare
H
Bain
J
Lucas
S
The schistosome egg granuloma: immunopathology in the cause of host protection or parasite survival?
Trans R Soc Trop Med Hyg
 , 
1986
, vol. 
80
 (pg. 
503
-
14
)
15.
Davies
SJ
Grogan
JL
Blank
RB
Lim
KC
Locksley
RM
McKerrow
JH
Modulation of blood fluke development in the liver by hepatic CD4+ lymphocytes
Science
 , 
2001
, vol. 
294
 (pg. 
1358
-
61
)
16.
Karanja
DM
Colley
DG
Nahlen
BL
Ouma
JH
Secor
WE
Studies on schistosomiasis in western Kenya: I. Evidence for immune‐facilitated excretion of schistosome eggs from participants with Schistosoma mansoni and human immunodeficiency virus coinfections
Am J Trop Med Hyg
 , 
1997
, vol. 
56
 (pg. 
515
-
21
)
17.
Fontanet
AL
Woldemichael
T
Sahlu
T
, et al.  . 
Epidemiology of HIV and Schistosoma mansoni infections among sugar‐estate residents in Ethiopia
Ann Trop Med Parasitol
 , 
2000
, vol. 
94
 (pg. 
145
-
55
)
18.
Mwanakasale
V
Vounatsou
P
Sukwa
TY
Ziba
M
Ernest
A
Tanner
M
Interactions between Schistosoma haematobium and human immunodeficiency virus type 1: the effects of coinfection on treatment outcomes in rural Zambia
Am J Trop Med Hyg
 , 
2003
, vol. 
69
 (pg. 
420
-
8
)
19.
Cheever
AW
A quantitative post‐mortem study of Schistosomiasis mansoni in man
Am J Trop Med Hyg
 , 
1968
, vol. 
17
 (pg. 
38
-
64
)
20.
Taylor
P
Makura
O
Prevalence and distribution of schistosomiasis in Zimbabwe
Ann Trop Med Parasitol
 , 
1985
, vol. 
79
 (pg. 
287
-
99
)
21.
Mott
KE
Baltes
R
Bambagha
J
Baldassini
B
Field studies of a reusable polyamide filter for detection of Schistosoma haematobium eggs by urine filtration
Tropenmed Parasitol
 , 
1982
, vol. 
33
 (pg. 
227
-
8
)
22.
Doehring
E
Vester
U
Ehrich
JH
Feldmeier
H
Circadian variation of ova excretion, proteinuria, hematuria, and leukocyturia in urinary schistosomiasis
Kidney Int
 , 
1985
, vol. 
27
 (pg. 
667
-
71
)
23.
Weber
MC
Blair
DM
de Clarke
VV
The distribution of viable and non‐viable eggs of Schistosoma haematobium in the urine
Cent Afr J Med
 , 
1969
, vol. 
15
 (pg. 
27
-
30
)
24.
Knight
WB
Hiatt
RA
Cline
BL
Ritchie
LS
A modification of the formol‐ether concentration technique for increased sensitivity in detecting Schistosoma mansoni eggs
Am J Trop Med Hyg
 , 
1976
, vol. 
25
 (pg. 
818
-
23
)
25.
Ebrahim
A
El Morshedy
H
Omer
E
El Daly
S
Barakat
R
Evaluation of the Kato‐Katz thick smear and formol ether sedimentation techniques for quantitative diagnosis of Schistosoma mansoni infection
Am J Trop Med Hyg
 , 
1997
, vol. 
57
 (pg. 
706
-
8
)
26.
Deelder
AM
Qian
ZL
Kremsner
PG
, et al.  . 
Quantitative diagnosis of Schistosoma infections by measurement of circulating antigens in serum and urine
Trop Geogr Med
 , 
1994
, vol. 
46
 (pg. 
233
-
8
)
27.
Polman
K
Diakhate
MM
Engels
D
, et al.  . 
Specificity of circulating antigen detection for Schistosomiasis mansoni in Senegal and Burundi
Trop Med Int Health
 , 
2000
, vol. 
5
 (pg. 
534
-
7
)
28.
Polman
K
Stelma
FF
Le Cessie
S
, et al.  . 
Evaluation of the patterns of Schistosoma mansoni infection and re‐infection in Senegal, from faecal egg counts and serum concentrations of circulating anodic antigen
Ann Trop Med Parasitol
 , 
2002
, vol. 
96
 (pg. 
679
-
89
)
29.
Dunne
DW
Hassounah
O
Musallam
R
, et al.  . 
Mechanisms of Schistosoma mansoni egg excretion: parasitological observations in immunosuppressed mice reconstituted with immune serum
Parasite Immunol
 , 
1983
, vol. 
5
 (pg. 
47
-
60
)
30.
Harrison
RA
Doenhoff
MJ
Retarded development of Schistosoma mansoni in immunosuppressed mice
Parasitology
 , 
1983
, vol. 
86
 (pg. 
429
-
38
)
31.
De Vlas
SJ
Engels
D
Rabello
AL
, et al.  . 
Validation of a chart to estimate true Schistosoma mansoni prevalences from simple egg counts
Parasitology
 , 
1997
, vol. 
114
 
Pt 2
(pg. 
113
-
21
)
32.
Kalinkovich
A
Weisman
Z
Greenberg
Z
, et al.  . 
Decreased CD4 and increased CD8 counts with T cell activation is associated with chronic helminth infection
Clin Exp Immunol
 , 
1998
, vol. 
114
 (pg. 
414
-
21
)
33.
Martin
DJ
Sim
JG
Sole
GJ
, et al.  . 
CD4+ lymphocyte count in African participants co‐infected with HIV and tuberculosis
J Acquir Immune Defic Syndr Hum Retrovirol
 , 
1995
, vol. 
8
 (pg. 
386
-
91
)
Presented in part: 14th International AIDS Conference, Barcelona, Spain, 7–12 July 2002 (abstract WeOrC1376); Fifth Federation of African Immunological Societies Conference, Victoria Falls, Zimbabwe, 28 April–1 May 2003
Financial support: Danish AIDS Foundation (grants F01-18 and F01-19); Essential National Health Research Fund of the Ministry of Health and Child Welfare of Zimbabwe (grant P355); Danish Embassy in Zimbabwe; Danish International Development Agency Health Program in Zimbabwe; US Centers for Disease Control and Prevention Program in Zimbabwe; Faculty of Health Sciences of the University of Copenhagen, Denmark (grant to P.K.)