| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Kaposi sarcoma (KS) is endemic in equatorial Africa, but its incidence has increased substantially with the advent of the AIDS epidemic,1 calling attention to immunosuppression as a co-factor. Human herpesvirus 8 (HHV8, also known as KS-associated herpesvirus or KSHV), is necessary for KS to develop.2 However, HHV8 pathogenicity is low in immunocompetent persons,3 suggesting contribution from co-factors. In Uganda, before the onset of AIDS, endemic KS was unevenly distributed, accounting for ~5% of cancers diagnosed in the central districts and 18% of cancers diagnosed in the West Nile districts.4 Correspondingly, adult non–HIV-related HHV8 seroprevalence seems lower in the central districts than in other districts,5,6 including the West Nile districts,7 but the differences do not sufficiently account for the geographic variation in endemic KS.5
Risk factors for KS include rural residence, walking barefoot, peasant farming, contact with surface water, and male sex.8 These risk factors point to environmental or outdoor exposure as playing an important role in KS risk. However, they are also risk factors for infection with intestinal parasites.9 Interestingly, co-occurrence of KS in patients with helminthic parasites has been noted,10 but whether KS patients are more likely to be infected with intestinal parasites has not been assessed in analytic studies. To test this hypothesis, we examined the frequencies of intestinal parasite carriage among KS patients and other cancer patients in Uganda.
We studied 943 KS patients and 1,042 other cancer patients admitted to the Uganda Cancer Institute (UCI) between 1994 and 2004. The UCI provides medical oncology services to histologically confirmed cancer patients. Before treatment, all patients are screened by stool microscopy to determine presence of abnormal cells and specific intestinal parasites. We abstracted basic demographic information, including admission date, age, sex, tribe, region of residence, cancer diagnosis, and stool microscopy results from laboratory accession books. We examined associations between KS and selected patient characteristics with odds ratios (ORs) and 95% confidence intervals (CIs) from logistic regression models. We examined the independent association of variables with KS using multivariable logistic regression models by simultaneously including all variables. Two-sided P < 0.05 was considered statistically significant.
Table 1
shows selected patient characteristics. The most frequent cancers other than KS included 261 (25%) non-Hodgkin lymphomas (NHL), 246 (24%) breast cancers, 106 (10%) leukemias, 77 (7%) other sarcomas, and 69 (7%) Hodgkin lymphomas. Compared with other cancer patients, KS patients were more frequently men and significantly younger. Residences of KS patients were overrepresented in the central districts and underrepresented in the northern districts. KS patients had been admitted from 1994 to 1999, whereas other cancer patients had been admitted from 1994 to 2004.
|
). Compared with other cancer patients, KS patients were more likely to carry Strongyloides stercoralis larvae (6.0% versus 3.1%; OR, 2.0; 95% CI, 1.3–3.2). Furthermore, ova or cysts of Schistosoma mansoni, Giardia lamblia, Trichomonas hominis, and Entamoeba histolytica were detected more frequently in KS patients, but the differences were not statistically significant. In contrast, hookworm ova and Entamoeba coli ova or cysts were detected less frequently in KS patients than in other cancer patients (OR, 0.7; 95% CI, 0.6–1.0 and OR, 0.8; 95% CI, 0.6–1.1, respectively). In analyses simultaneously adjusting for all covariates (sex, age, admission year, region of residence, and presence of S. stercoralis, hookworm, and E. coli), KS was positively associated with male sex (OR, 3.0; 95% CI, 2.4–3.9), age < 35 years (OR, 3.6; 95% CI, 2.8–4.6), and carrying S. stercoralis (OR, 2.1; 95% CI, 1.2–3.7) and inversely associated with hookworm (OR, 0.6; 95% CI, 0.4–1.0) and E. coli (OR, 0.7; 95% CI, 0.5–1.0).
|
KS patients had a higher prevalence of S. stercoralis carriage than patients with other cancers. This observation echoes a previous report that noted co-occurrence of onchocerciasis with KS.10 The correlation between KS and luminal or tissue parasites may be caused by common risk factors for parasites and KS, or it may indicate an etiologic association. For example, S. stercoralis is capable of multiplying even in the immunocompetent host, and filariform larvae in the upper small bowel can re-invade through the colonic mucosa or the perianal areas, creating a long lasting infection and immunologic modulation. After infection, intestinal parasites downregulate T helper (Th)-1 cellular responses and upregulate Th-2 humoral responses.16 These immunologic changes may render the infected host unable to effectively control viral infection. Plausibly, infection with S. stercoralis may be associated with poor control of lytically replicating HHV8, expansion of HHV8-infected B cells in the peripheral blood, and a heightened risk for KS. Although biologically plausible, reports of disseminated strongyloidiasis as a feature of KS distinctively rare.17,18
Confounding by HIV infection, which is a risk factor for KS19 and may increase the risk for S. stercoralis,20 is possible. To address this question, we performed a sensitivity analysis to assess to what extent HIV may explain our findings. Because we lacked individual data on HIV status, we used published data on HIV prevalence among cancer patients treated at the UCI during the 1990s.19,21 The impact of HIV infection on carriage of S. stercoralis is not well understood.18 In Uganda, S. stercoralis larvae are detected frequently in HIV-infected patients22; however, disseminated strongyloidiasis is rare, and the risk of carriage of S. stercoralis among HIV-infected versus uninfected is unknown. Among HIV-infected persons, lower CD4 lymphocyte counts was associated with indirect, rather than direct, development of S. stercoralis.23 Because direct development is needed for hyperinfection associated with disseminated strongyloidiasis, these results suggest an explanation for the infrequency of disseminated strongyloidiasis in persons with HIV infection in countries such as Uganda despite relatively common carriage. Thus, for our sensitivity analyses, we assumed that HIV-infected persons are 2-fold more likely to carry S. stercoralis larvae compared with uninfected persons; therefore, the postulated effect size was similar to that observed between S. stercoralis and KS. Thus, assuming that 71% of KS patients,19 62% of NHL patients, and 21% of other cancer patients21 were HIV-infected and that HIV infection doubles the risk for S. stercoralis detection, one would expect a 1.3-fold association of KS with S. stercoralis. Based on these assumptions, we estimate that ~60% of the observed association could be explained by confounding with HIV infection.
Our findings of inverse associations between KS and hookworm or E. coli were unexpected. Because the probability of infection with specific parasite species may vary by geography, we performed analyses stratified by region and tribe and obtained essentially similar results. Possibly, patients with KS may be more likely to receive anti-helminthic drugs than other cancer patients. This difference could be caused by Ugandan physicians considering the risk factors for intestinal or tissue parasite infestation, i.e., rural residence, contact with surface water, and walking barefoot, which coincidentally are risk factors for KS,8 and therefore prescribing anti-helminthics to KS patients. HIV infection may confound the associations because immunosuppression has been reported to decrease egg production by hookworms, which could result in lower frequency of detection in stool.24 Alternatively, epidemic KS is associated with higher socioeconomic indicators, possibly lowering exposure to ovacontaminated soils.25
The strengths of our pilot study include its relatively large size, detailed data on specific stool parasites measured on admission, the novelty of the associations examined, and the hypotheses it explores. Nonetheless, several limitations remain. First, the study is cross-sectional; thus, the temporality of the associations cannot be determined. Our results are based on examining only one stool sample, which has less optimal sensitivity and would attenuate our findings, especially for species that are hard to detect like S. stercoralis. Second, we did not have HIV status for the patients to more precisely ascertain the degree of confounding. Third, referral bias to the UCI limits generalizability of our findings. Finally, we lacked information on prior anti-helminthic treatment and other exposure data, which might explain the inverse associations with hookworm. Nonetheless, our results provide additional reasons to study relationships between KS and infection with intestinal parasites.
To summarize, the novel associations are positive association of KS with carriage of S. stercoralis larvae and inverse associations with hookworm ova and E. coli cysts. The significance of these associations is uncertain, but they raise questions about a possible etiologic relationship between KS and intestinal parasites that should be addressed in studies using a prospective design, controlling for HIV, socioeconomic indicators, and other confounders.
Received November 8, 2007. Accepted for publication December 15, 2007.
Financial support: This study was funded, in part, through the Intramural Program of the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, contract N02-CO-12400.
* Address correspondence to Sam M. Mbulaiteye, 6120 Executive Boulevard, Executive Plaza South Room 7080, Rockville, MD 20852. E-mail: mbulaits{at}mail.nih.gov 
Authors’ addresses: Cynthia J. Lin, Charles S. Rabkin, and Sam M. Mbulaiteye, 6120 Executive Boulevard, Executive Plaza South Room 7080, Rockville, MD 20852, Tel: 301-496-8115, Fax: 301-402-0817, E-mail: mbulaits{at}mail.nih.gov. Edward Katongole-Mbidde, Tadeos Byekwaso, and Jackson Orem, Uganda Cancer Institute, PO Box 3935, Kampala, Uganda, Tel: 256-414-540410.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
