INTRODUCTION
In most areas of the developing world, anemia is exacerbated by helminthiasis1 and understanding the contribution of helminth infections to anemia is crucial for prioritization of scarce health care resources. Schistosoma japonicum has been related to anemia;2,3 however, the mechanisms mediating this relationship are unclear. The three main mechanisms proposed are extra-corporal iron loss due to fecal occult blood loss, anemia of inflammation/chronic disease, and autoimmune hemolytic anemia.4–6 Previous studies of occult blood loss in the context of schistosomiasis have provided varied results. One cross-sectional study in Anhui Province, China did not observe any relationship between occult blood and S. japonicum presence or intensity of infection. However, that study population harbored a low prevalence and intensity of S. japonicum infection.7 In cross-sectional and case-control studies in the context of S. mansoni, investigators found strong evidence for a relationship between stool occult blood loss and intensity of schistosomiasis infection.8–10 These studies, however, did not control for the presence of hookworm and Trichuris infections, which have been associated with occult blood loss at high intensities.1,11
To our knowledge, there have been no studies addressing fecal occult blood loss across a range of S. japonicum infection intensities. Moreover, no previous studies have examined the relationship between fecal occult blood loss and S. japonicum after controlling for potentially important mediating/confounding covariates. In addition, no large-scale, community studies have addressed the risk for fecal blood loss due to Trichuris across a range of infection intensities. The primary objective of this study was to assess the role of occult blood loss in mediating S. japonicum-associated anemia after adjusting for age, sex, socioeconomic status (SES), and other helminth infections. The secondary objective was to identify intensity categories of risk for occult blood loss for Trichuris and hookworm after adjustment for the presence of other helminth infections.
METHODS
Study area and population.
Cross-sectional data were collected from individuals 7–30 years old in three S. japonicum-endemic villages in Leyte, The Philippines. Approximately 90% of the individuals were screened for the presence of S. japonicum infection. Subjects were eligible if they provided a stool sample, were infected with S. japonicum, lived primarily in a study village, were between 7 and 30 years old, were not pregnant or lactating, and provided both child assent and parental consent. In addition, 111 individuals 7–18 years old not infected with S. japonicum were recruited as control subjects.
Stool examination.
For each participant, three stool specimens were examined in duplicate for S. japonicum, Ascaris lumbricoides, Trichuris trichuria, and hookworm by the Kato Katz method. For each of the stool specimens, the average eggs per gram (epg) of feces of the duplicate test was determined, and the overall mean epg was determined by averaging the three stool specimens. Intensity of infection for each helminth was determined using World Health Organization (WHO) criteria.1,12 Ten hookworm larvae obtained by culturing stool samples13 from 203 volunteers were speciated by a polymerase chain reaction; all were Necator americanus. One stool sample from each individual was evaluated with the Hemoccult® system (Beckman-Coulter, Fullerton, CA), a screening test for the presence of fecal occult blood.
Blood collection and processing.
Venipuncture was performed and blood was collected into Vacutainertubes® (Becton Dickinson and Company, Franklin Lakes, NJ). Complete hemograms were determined on a Serono Baker 9000 hematology analyzer (Serono Baker Diagnostics, Allentown, PA).
Socioeconomic status.
Socioeconomic status was based on questionnaire data, which had good internal consistency (Cronbach’s alpha = 82.4% for all questions). A summary SES score comprised of all questionnaire items was calculated using principal components analysis to appropriately weight questionnaire items as described by Filmer and Pritchett.14
Data management and statistical analyses.
Data forms collected in the field were bar coded and entered using Filemaker 5.5 software (Filemaker Inc., Santa Clara, CA). Normality diagnostics were first performed, and non-normally distributed variables, including epg, were loge transformed [ln(value + 1)]. For comparison of morbidity across infection intensities, e.g., heavy versus uninfected, low, or moderate, categorizations were based on the greatest differences in morbidity for each helminth infection as determined by analysis of variance.
Regression models were constructed to provide unconfounded estimates of the relationships of interest. The primary exposure of interest e.g., helminth infection intensity, was first entered into the model. Variables that were significantly related to the outcome in bivariate analyses (P < 0.10) were then sequentially entered into the model. These were retained if they were significantly related to the outcome of interest, or if their inclusion altered the beta coefficient for the primary exposure of interest by > 10%. Least squared means (LSM) analyses were conducted to determine the adjusted relationship between mean hemoglobin concentration and intensity of helminth infections as defined by WHO criteria.1,12 Logistic regression models were made comparing odds of occult blood loss and anemia (hemoglobin concentration < 11 g/dL) across infection categories, with a similar process for selecting and retaining covariates. All analyses were performed in JMP version 5.1 on PC computers (SAS Institute, Cary, NC).
Ethical clearance.
This study was reviewed and approved by Brown University and The Philippines Research Institute of Tropical Medicine Institutional Review Boards. Written informed consent was obtained from each adult participant or the parents of assenting children.
RESULTS
The study sample consisted of 729 individuals, of which, 449 were male and 280 were female. The mean age (95% confidence interval [CI]) among the volunteers was 14.9 (14.5–15.3) years. Infection with more than one helminth was common, since the prevalence of each of four helminth infections assessed was greater than 50%.
Hemoglobin and helminths.
There was no significant relationship between Ascaris and hemoglobin concentration, or the outcomes listed below. The adjusted mean hemoglobin concentration significantly decreased with increasing infection intensity for all three other helminths (LSM P values for S. japonicum, hookworm, and T. trichuria = < 0.0031, < 0.0001, and < 0.0001, respectively) (Figure 1).
Fecal occult blood loss and helminths.
We evaluated the role of blood loss as assessed by fecal occult blood status across different intensities of infection. Participants with heavy intensity S. japonicum infection were 3.5 times more likely to be fecal occult blood positive than the rest of the cohort (P = 0.018) (Figure 2a). After adjusting for SES, S. japonicum infection, and T. trichuria infection, we did not observe a significant relationship between hookworm infection and occult blood loss (odds ratio = 1.70, 95% CI = 0.25, 6.58, P = 0.50) (Figure 2b). After adjusting for SES, participants with moderate or heavy intensity T. trichuria infections were 2.68 times more likely to be fecal occult blood positive than individuals without Trichuris or with low intensity infections (P = 0.013) (Figure 2c).
Anemia and helminths.
Participants with heavy intensity S. japonicum infection were 3.7 times more likely to be anemic the rest of the cohort (P = 0.0002). After adjusting for SES, S. japonicum infection, and T. trichuria infection, participants with moderate and heavy intensity hookworm infections were 5.3 times more likely to be anemic than individuals without hookworm or with low intensity infection (P = 0.0003). After adjusting for SES, participants with moderate or heavy intensity T. trichuria infection were 1.6 times more likely to be anemic than individuals without Trichuris or with low intensity infection (P = 0.021).
DISCUSSION
This is the first study to our knowledge to adjust for coinfection with other helminths while assessing the independent risk of occult blood loss from each parasite; our analyses provide adjusted estimates of risk unconfounded by the presence of other helminths. This is also the first study to examine risk for occult blood loss in S. japonicum infection with a sample size that has a sufficient number of heavy intensity infections.
S. japonicum, T. trichuria, and hookworms were inversely related to hemoglobin concentration. This supports findings from previous studies where S. japonicum,2 moderate and heavy hookworm,15 and moderate and heavy T. trichuria infections were associated with a decreased hemoglobin concentration.16 Importantly, this study observed these relationships independently, after adjustment for SES and the presence of other helminth infections. The relationship between S. japonicum and hemoglobin concentration was noted even when the analysis excluded those with heavy intensity infection, suggesting a generalized phenomenon not limited to those most intensely infected. Notably, the relationship between S. japonicum infection and hemoglobin concentration remained even when occult blood loss was retained in the multivariate model (heavy S. japonicum infection LSM = 10.95; uninfected, and low and moderate S. japonicum infections LSM = 11.63, P = 0.01). This suggests that some other mechanism, such as anemia of inflammation/chronic disease, may be contributing to S. japonicum-associated anemia. Fecal occult blood loss was associated with heavy S. japonicum infection. The only other study that has evaluated occult blood loss in the context of S. japonicum infection did not observe differences across intensity groups. Of note, that study was limited by a low number (n = 38 of 2,310) of heavy intensity infections.7 Thus, it is likely that occult blood loss plays a role only at heavier intensity infections.
Fecal occult blood loss was also associated with moderate and heavy T. trichuria infections. One study did not observe any relationship between fecal occult blood loss and T. trichuria infection in the absence of dysentery syndrome. The limitation of that sample to those without dysentery led to a very low prevalence of occult positivity (n = 1 of 104).17 Other studies have demonstrated fecal blood loss in hospital-based samples of children with severe trichuriasis.11,18,19 Our community-based findings of increased risk for occult blood loss in the presence of moderate and heavy Trichuris infections may reflect variability in predisposition for mucosal inflammation, colonic ulcers, and blood loss at different intensities of infection.
Study limitations include the lack of characterization of the Trichuris dysentery syndrome, such that we cannot comment on whether blood loss is occurring only among those with true dysentery syndrome or more generally. Additionally, only one fecal occult blood examination was performed. It is possible that occult blood loss from S. japonicum egg translocation across the large bowel and/or polyps is too sporadic to sensitively capture with a single test. Thus, occult blood loss may still play a role in the decreased hemoglobin concentration observed even in those without heavy intensity infections.
In conclusion, this study suggests that mechanisms other than extra-corporeal iron loss may be contributing to S. japonicum-associated anemia, possibly anemia of inflammation. This has important public health implications since iron repletion in the context of inflammation will have markedly dampened clinical efficacy. Ongoing studies of the relationship between S. japonicum-associated inflammation and anemia in this cohort should shed light on this question.
Adjusted mean hemoglobin (Hgb) concentrations by intensity of helminth infection with 95% confidence intervals with least squared means analysis adjusted for the covariates specified. A, Hgb by intensity of Schistosoma japonicum infection adjusted for socioeconomic status (SES), moderate and heavy Trichuris trichuria infections, and moderate and heavy hookworm infections. B, Hgb by intensity of hookworm infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy T. trichuria infections. C, Hgb by intensity of T. trichuria infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy hookworm infections.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Adjusted mean hemoglobin (Hgb) concentrations by intensity of helminth infection with 95% confidence intervals with least squared means analysis adjusted for the covariates specified. A, Hgb by intensity of Schistosoma japonicum infection adjusted for socioeconomic status (SES), moderate and heavy Trichuris trichuria infections, and moderate and heavy hookworm infections. B, Hgb by intensity of hookworm infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy T. trichuria infections. C, Hgb by intensity of T. trichuria infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy hookworm infections.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Adjusted mean hemoglobin (Hgb) concentrations by intensity of helminth infection with 95% confidence intervals with least squared means analysis adjusted for the covariates specified. A, Hgb by intensity of Schistosoma japonicum infection adjusted for socioeconomic status (SES), moderate and heavy Trichuris trichuria infections, and moderate and heavy hookworm infections. B, Hgb by intensity of hookworm infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy T. trichuria infections. C, Hgb by intensity of T. trichuria infection adjusted for SES, heavy S. japonicum infection, and moderate and heavy hookworm infections.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Prevalence of fecal occult blood positivity by helminth intensity. a, Prevalence of occult blood by Schistosoma japonicum infection. OR = odds ratio. b, Prevalence of occult blood by hookworm infection. The OR and P value were adjusted for socioeconomic status (SES), heavy S. japonicum infection, and moderate and heavy Trichuris trichuria infections. c, Prevalence of occult blood by T. trichuria infection. The OR and P value were adjusted for SES.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Prevalence of fecal occult blood positivity by helminth intensity. a, Prevalence of occult blood by Schistosoma japonicum infection. OR = odds ratio. b, Prevalence of occult blood by hookworm infection. The OR and P value were adjusted for socioeconomic status (SES), heavy S. japonicum infection, and moderate and heavy Trichuris trichuria infections. c, Prevalence of occult blood by T. trichuria infection. The OR and P value were adjusted for SES.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Prevalence of fecal occult blood positivity by helminth intensity. a, Prevalence of occult blood by Schistosoma japonicum infection. OR = odds ratio. b, Prevalence of occult blood by hookworm infection. The OR and P value were adjusted for socioeconomic status (SES), heavy S. japonicum infection, and moderate and heavy Trichuris trichuria infections. c, Prevalence of occult blood by T. trichuria infection. The OR and P value were adjusted for SES.
Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 72, 2; 10.4269/ajtmh.2005.72.115
Authors’ addresses: Hemal K. Kanzaria, Gretchen C. Langdon, Stephen T. McGarvey, Jonathan D. Kurtis, and Jennifer F. Friedman, International Health Institute, Brown University, Box G-B495, Providence, RI 02912, E-mails: Hemal.Kanzaria@alumni.brown.edu, Gretchen_Langdon@Brown.edu, Stephen_McGarvey@Brown.edu, Jonathan_Kurtis@Brown.edu, and Jennifer_Friedman@Brown.edu. Luz P. Acosta, Daria L. Manalo, and Remigio M. Olveda, Research Institute of Tropical Medicine, FICC, Alabang, Muntinlupa City 1770, Metro Manila, The Philippines, E-mails: Lacosta@ritm.gov.ph and dmanalo@ritm.gov.ph
Acknowledgments: We thank our field staff (Blanca Jarilla, Mario Jiz, Archie Pablo, Raquel Pacheco, Patrick Sebial, Mary Paz Urbina, and Jemaima Yu) for their diligence and energy. We also thank the study participants from Macanip, Buri, and Pitogo in Leyte, The Philippines.
Financial support: This work was funded by National Institutes of Health grants RO1AI48123 and K23AI52125.
REFERENCES
- 1↑
WHO, 1987. Prevention and control of intestinal parasitic infections. World Health Organ Tech Rep Ser 749 :8–28.
- 2↑
McGarvey ST, Aligui G, Graham KK, Peters P, Olds GR, Olveda R, 1996. Schistosomiasis japonica and childhood nutritional status in northeastern Leyte, the Philippines: a randomized trial of praziquantel versus placebo. Am J Trop Med Hyg 54 :498–502.
- 3↑
Yuesgeng L, Dongbao Y, Yi L, Yang X, Meng X, 1992. Morbidity and Health Impact of Schistosomiasis japonica in the Dongting Lake Area. A Final Report: WHO Collaborating Centre for Research and Control on Schistosomiasis in the Lake Region. Geneva: World Health Organization.
- 4↑
Robinson A, Lewert RM, 1980. The production and nature of anemia in Schistosoma japonicum infections. Am J Trop Med Hyg 29 :1301–1306.
- 5
Mahmoud AA, Woodruff AW, 1972. Mechanisms involved in the anaemia of schistosomiasis. Trans R Soc Trop Med Hyg 66 :75–84.
- 6↑
Stephenson LS, 1987. The Impact of Helminth Infections on Human Nutrition. London: Taylor & Francis.
- 7↑
Warren KS, Su DL, Xu ZY, Yuan HC, Peters PA, Cook JA, Mott KE, Houser HB, 1983. Morbidity in schistosomiasis japonica in relation to intensity of infection. A study of two rural brigades in Anhui Province, China. N Engl J Med 309 :1533–1539.
- 8↑
Cook JA, Baker ST, Warren KS, Jordan P, 1974. A controlled study of morbidity of schistosomiasis mansoni in St. Lucian children, based on quantitative egg excretion. Am J Trop Med Hyg 23 :625–633.
- 9
Lehman JS Jr, Mott KE, Morrow RH Jr, Muniz TM, Boyer MH, 1976. The intensity and effects of infection with Schistosoma mansoni in a rural community in northeast Brazil. Am J Trop Med Hyg 25 :285–294.
- 10↑
Ndamba J, Makaza N, Kaondera KC, Munjoma M, 1991. Morbidity due to Schistosoma mansoni among sugarcane cutters in Zimbabwe. Int J Epidemiol 20 :787–795.
- 11↑
Layrisse M, Aparcedo L, Martinez-Torres C, Roche M, 1967. Blood loss due to infection with Trichuris trichiura. Am J Trop Med Hyg 16 :613–619.
- 12↑
Stephenson LS, Holland CV, Cooper ES, 2000. The public health significance of Trichuris trichiura. Parasitology 121 (Suppl):S73–S95.
- 13↑
Widjana DP, Sutisna P, 2000. Prevalence of soil-transmitted helminth infections in the rural population of Bali, Indonesia. Southeast Asian J Trop Med Public Health 31 :454–459.
- 14↑
Filmer D, Pritchett LH, 2001. Estimating wealth effects without expenditure data–or tears: an application to educational enrollments in states of India. Demography 38 :115–132.
- 15↑
Albonico M, Stoltzfus RJ, Savioli L, Tielsch JM, Chwaya HM, Ercole E, Cancrini G, 1998. Epidemiological evidence for a differential effect of hookworm species, Ancylostoma duodenale or Necator americanus, on iron status of children. Int J Epidemiol 27 :530–537.
- 16↑
Ramdath DD, Simeon DT, Wong MS, Grantham-McGregor SM, 1995. Iron status of schoolchildren with varying intensities of Trichuris trichiura infection. Parasitology 110 :347–351.
- 17↑
Raj SM, 1999. Fecal occult blood testing on Trichuris-infected primary school children in northeastern peninsular Malaysia. Am J Trop Med Hyg 60 :165–166.
- 19↑
Gilman RH, Chong YH, Davis C, Greenberg B, Virik HK, Dixon HB, 1983. The adverse consequences of heavy Trichuris infection. Trans R Soc Trop Med Hyg 77 :432–438.