Morbidity Associated with Schistosomiasis Before and After Treatment in Young Children in Rusinga Island, Western Kenya

Abstract

Schistosoma mansoni infection is a major cause of organomegaly and ultimately liver fibrosis in adults. Morbidity in pre-school-aged children is less defined, and they are currently not included in mass drug administration (MDA) programs for schistosomiasis control. We report results of a study of the association of schistosomiasis with organomegaly in a convenience sample of 201 children under 7 years old in Rusinga, Kenya on two cross-sectional visits, before and after praziquantel treatment. Data included stool examination and serology for schistosomiasis, the Niamey ultrasound protocol to stage hepatosplenic morbidity including organomegaly, and potential confounders including malaria. Unadjusted and adjusted Poisson regressions were performed. The baseline prevalence of schistosomiasis by antibody and/or stool was 80.3%. Schistomiasis was associated with hepatomegaly (adjusted prevalence ratio [aPR] = 1.4; 95% confidence interval [CI]: 1.0–2.1) and splenomegaly (aPR = 2.1; 95% CI: 1.2–3.7). The association with hepatomegaly persisted posttreatment (aPR = 1.4; 95% CI: 1.1–1.6). Schistosomiasis was associated with morbidity in this cohort. Efforts to include young children in mass treatment campaigns should intensify.

Introduction

Schistosomiasis, a chronic tropical parasitic disease caused by infection with Schistosoma spp., is a major cause of global disability¹ concentrated in sub-Saharan Africa.² Pathology results from egg deposition in host tissues, particularly liver tissue in the case of S. mansoni, causing inflammation, organomegaly, and fibrosis. Infection prevalence peaks in 8- to 15-year-olds,³ but morbidity is best documented in older age groups in association with chronic infection. In addition to prevention through sanitation, the cornerstone of global schistosomiasis control is regular mass drug administration (MDA) using praziquantel,⁴ with school-aged children as the priority target group for treatment due to the focus on reducing morbidity from chronic infection.

Currently, mass distribution of praziquantel for children under 4 years or 94 cm in height is not recommended, and in practice, children not in school for any reason are rarely treated. However, children too young or small for MDA have considerable infection prevalences^5,6 in multiple African countries, where they can be exposed through bathing and other activities,^7,8 and they can be treated with no serious adverse events.^9–13 The morbidity pre-school-aged children (PSAC) experience from infection is not well defined, though: effects are known to include fecal occult bleeding,^14,15 hematuria, proteinuria,¹⁶ ultrasound abnormalities,^17,18 and possibly anemia,^19,20 but clinical implications remain unclear. Thus, although these findings have prompted calls for including PSAC in MDA,²¹ a 2010 WHO meeting concluded that more evidence was necessary.

The major tool for assessing S. mansoni infection–associated morbidity in older patients is ultrasound^22–28 using the WHO-recommended Niamey scoring system.^29,30 It includes assessment of expected morbidities such as splenomegaly and left lobe hepatomegaly, as well as successively more pronounced liver ultrasound changes scored as image patterns (IP) A (normal), B (“starry sky,” abnormal linear opacities of unclear significance), and C through F (progressive hepatic fibrosis). These characteristic, partially reversible^31,32 schistosomiasis-associated abnormalities in adults have also been found in school-aged children,^33,34 but morbidity in PSAC using this system or ultrasound more generally is again not well characterized.

Thus ultrasound investigations in PSAC have potential to provide clarity on morbidities associated with schistosomiasis, information that is needed to support decisions about inclusion in MDA. We investigated the association of S. mansoni infection in PSAC with Niamey liver texture pattern and organomegaly, as well as other indicators of morbidity including growth, anemia, and liver function tests (LFTs).

Methods

Ethics statement.

This study was approved by the Scientific Steering and Ethics Review Committees of the Kenya Medical Research Institute (KEMRI, SSC No. 2185) and of the Institutional Review Board of the U.S. Centers for Disease Control through a reliance agreement with KEMRI. Parents gave written informed consent for all participants. All children diagnosed with S. mansoni infection were treated with praziquantel (crushed, 40 mg/kg), those with soil-transmitted helminth (STH) infections with 400 mg albendazole, and those with malaria with Coartem® Dispersible. Any diagnosed with other abnormalities were referred for pediatric follow-up.

Study site and population.

Rusinga is a Lake Victoria island in Mbita district in western Kenya, located in a region with known high prevalence of schistosomiasis among school-aged children.³⁵ Intestinal schistosomiasis attributable to S. mansoni is the main form of infection locally, with a few isolated foci of S. haematobium identified in adjacent districts.³⁶ Prior to this study, no MDA for schistosomiasis had been conducted in the area.

Study design, sample size, and recruitment.

We conducted a nested study within a larger ongoing community-randomized trial of the impact of integrated programs for control of neglected tropical diseases. Inclusion criteria for the parent study were residence in the district, an initial reported age of 1 year or more, and willingness to participate; the exclusion criterion was acute illness requiring immediate treatment. The eight villages with the highest community prevalences of S. mansoni infection as determined in the parent study were selected for our nested study: Utajo, Kamayoge, Kaktemo, Wariga “A,” Kamgere, Wakwala, Dier Aora, and Kakrigu (prevalence 34.0–92.5%) (Maurice Odiere, unpublished data). In each village, the community health worker and the study field team mobilized all willing mother-PSAC dyads enrolled in the parent study to participate in the nested study. The sample size of 201 in this exploratory study was limited by resources and willingness of enrolled dyads to participate.

We collected pre–post data on prevalence of schistosome infection and association with morbidities in PSAC, in the form of two serial cross-sectional studies in the same cohort before and after the administration of praziquantel. After the baseline visit, the praziquantel (40 mg/kg) was delivered to all children who could be located in one of two treatment campaigns conducted 3 or 5 months later. The interruption between treatment periods was due to local religious activities. The follow-up visit was then conducted, falling 1.5–2.0 months and 3.5–4.0 posttreatment of children treated in the first and second campaigns, respectively.

Data collection.

Baseline data, obtained at either the nested study baseline visit (May 2012) or parent study enrollment up to 2 weeks prior, included a parental questionnaire assessing symptoms consistent with acute schistosome infection (fever, rash or swelling, headache, muscle aches, dry cough or wheeze or trouble breathing, decreased activity level, malaise, diarrhea, loss of appetite, weight loss, and diaphoresis) in the past 2 months; 2 mL venous blood collected for serology and LFTs; one stool for ova of S. mansoni; fingerstick blood for hemoglobin and thick blood smears for malaria; height and weight to assess nutritional status; and ultrasound examination according to the Niamey protocol by a single experienced research ultrasonographer, with additional image capture (see below). Community interviewers visited enrolled families at home to obtain documentation of birthdates using birth certificates, baptism cards, or antenatal clinic visit cards.

Data obtained at the posttreatment visit (December 2012) included 2 mL venous blood collected for rapid diagnostic test (RDT) for malaria due to Plasmodium falciparum or other species (SD Bioline Malaria Ag P.f/Pan, Borhagal-ro, Giheung-gu, South Korea), hemoglobin, and repeat ultrasound examination according to the Niamey protocol. A second height and weight measure was obtained in February 2013.

All ultrasounds were performed by the same research ultrasonographer, using a generator-powered Aloka SSD-900V portable ultrasound machine with a UST-979-3.5 MHz convex transducer, with image capture using a Medicap USB-200. The standard Niamey protocol was observed, with measurements including spleen length, liver span, and liver IP, and with scoring done per the protocol by the ultrasonographer. Images of standard liver views were also captured, for expert review by an ultrasound radiologist for any novel findings.

Laboratory methods.

Duplicate slides were prepared by the Kato-Katz method to quantify ova of S. mansoni. The stool template held approximately 41.7 mg of feces. Slides were read within 12 hours of defecation. Thick blood smears for malaria were made on-site at the time of blood collection and stained with Giemsa stain in the laboratory.³⁷ Quality control for microscopy was performed by an independent, senior microscopist. This expert result was used in cases of conflict. Tests for antibodies to schistosomes were performed by enzyme-linked immunosorbent assay (ELISA) using soluble worm antigen preparation using previously described methods.³⁸ This test does not distinguish between Schistosoma spp.

LFTs were performed using the Cobas Integra 400 plus biochemistry analyzer (Roche, Berlin, Germany) with upper limit normal values of 37 U/L for aspartine aminotransferase (AST), 42 U/L for alanine aminotransferase (ALT), and 17.0 μmol/L for bilirubin. Detection of hepatitis B surface antigen (HBsAg) was performed using the KEMRI Hepcell kit, a reverse passive hemagglutination-based kit, with confirmation of positives via the Murex HBsAG Confirmatory Version 3 kit (DiaSorin,Vercelli, Italy).

Statistical analysis.

Analysis was performed using SAS version 9.3 (SAS Institute, Inc., Cary, NC). Each model was fit on the subsample of participants possessing complete data on all variables included in the model.

Cross-sectional analysis of baseline visit data.

Children were classified based on infection testing results as either uninfected (egg-negative and antibody-negative) or infected (egg-positive and/or antibody-positive). Egg-positive (egg+) children were further classified according to WHO standards³⁹ as having infection of light (1–99 eggs per gram [epg]), moderate (100–399 epg), or heavy (equal to or greater than 400 epg) intensity.

Organomegaly was diagnosed based on height at the baseline visit using the height-based cutoffs originally determined by Yazdanpanah and others,⁴⁰ with organ spans > 2 standard deviations above height-adjusted means considered organomegalic.

Initially, we examined prevalences of infection categories and outcome findings, and associations of key findings with infection category. Based on the high prevalence of IPB, we then analyzed associations of key potential predictive factors and morbidities with IPB. Poisson regression was used except for serum transaminase and bilirubin outcomes, for which linear regression was used (after log transformation in the case of bilirubin). Both unadjusted and adjusted analyses (corrected for sex and age) were performed, with clustering by village accounted for via generalized estimating equations⁴¹ and an exchangeable correlation structure. An exact model was fit for the analyses of malaria diagnosis by microscopy. A sub-analysis limited to only children ineligible for MDA by virtue of height < 94 cm was also performed.

Cross-sectional analysis of posttreatment visit data.

Because serology was not repeated at follow-up, individuals were classified only as egg-positive or egg-negative. Height at the time of follow-up was interpolated from height at the repeat measurement roughly 2 months later, assuming linear growth between the first and second anthropometry measures. Organometry outcomes were determined as before using this interpolated height.

Regressions were run using morbidity outcomes for schistosomiasis and those predictive factors for IPB that had shown significant associations at baseline, and with malaria determination by blood smear replaced with RDT data. An interaction term was included to test for interaction between schistosomiasis and malaria as predictors of organomegaly. The sub-analysis was performed as before.

Longitudinal analysis.

For the set of children who had all organometric and S. mansoni diagnostic data available for both visits and had received praziquantel treatment between visits, analysis of change over time in morbidity findings which had been associated with S. mansoni infection at baseline was performed. Analyses accounted for clustering by individual; correlation by village was negligible and was not included in the model. Multivariable analyses controlled for age, sex, and malarial infection based on slide result, since malaria RDT was obtained only at follow-up. A subset analysis of “responders” was performed by restricting the regression to the subset of children who were egg+ at baseline and were then either egg-negative or in a less intense infection category post treatment.

Multiple imputation.

Multiple imputation by chained equations⁴² using IVEware version 0.2⁴³ was performed to assess whether the missing data influenced the model inferences. Baseline, follow-up, and longitudinal models were included in this assessment and revealed that the missing data would not substantially change associations reported in this study.

Results

Cross-sectional analysis of baseline visit data.

Two hundred and one children were enrolled at baseline, representing a minimum of 74% of eligible children in each village. Missing data were primarily due to baseline anthropometry not being performed or missing LFT data. Four additional children were missing sex or age data and were omitted from multivariable analyses. Fifteen “PSAC” were found on review of age documentation to be 5 years of age or older.

The mean age of enrolled children was 3.5 years (1.2–7.0 years); mean height was 94.6 cm. At baseline, 95/201 (47.3%) of those enrolled were male. Infection and morbidity findings are shown in Table 1. Notable findings included high proportions infected with Schistosoma spp., with 80.3% of children S. mansoni egg+ and/or antibody positive, and high intensity of infection, with over half of egg+ children having moderate or heavy infection. Proportion with schistosomiasis increased with age, from 62.5% (20/32) among children < 2.5 years to 80.9% (72/89) among children 2.5–< 4 years and 88.4% (61/69) among children ≥ 4 years. Corresponding proportions with egg+ infection were 25.0% (8/32), 40.4% (36/89), and 60.9% (42/69). Hepatitis B, malaria by smear, and hepatic right lobe atrophy were rare at baseline; other morbidities and findings, including anemia and organomegaly, were common (hepatomegaly 50.0%, splenomegaly 29.2%). Modest AST elevations were nearly universal. Morbidity complaints ranged in prevalence from 24.4% for decreased activity to 84.1% for fever (not shown).

Table 1.

Baseline and posttreatment prevalences of demographic traits and findings

Type	Trait	Baseline (N = 201)			Posttreatment (N = 180)
		n	N	%	n	N	%
Age	< 2.5 years	32	198	16.2	7	180	3.9
	2.5 years ≤ age < 4 years	89	198	44.9	79	180	43.9
	Age ≥ 4 years	69	198	34.8	88	180	48.9
	Unable to confirm	8	198	4.0	6	180	3.3
	Schistosomiasis-infected	159	198	80.3	−	−	−
Infection status	Egg−/Ab+	70	198	35.4	−	−	−
	Egg+	89	198	44.9	62	180	34.4
	Egg+, Ab+	80	198	40.4	−	−	−
	Egg+, Ab–	9	198	4.5	−	−	−
	Light intensity (< 100 EPG*)	41	198	20.7	34	180	18.9
	Medium intensity (100–400 EPG)	28	198	14.1	19	180	10.6
	Heavy intensity (> 400 EPG)	20	198	10.1	9	180	5.0
	Malaria	(Smear) 7	198	3.5	(Smear) 30	180	16.7
					(RDT) 70	180	38.9
	Hepatitis B (surface antigen)	2	190	1.1	−	−	−
Liver pattern	Liver IPB	30	200	15.0	36	180	20.0
Morbidities	Anemia	161	196	82.1	106	173	61.3
	Undernutrition*	44	182	24.2	39	166	23.5
	Left lobe hepatomegaly†	91	182	50.0	99	152	65.1
	Right-lobe atrophy†	3	177	1.7	1	149	0.7
	Splenomegaly†	54	185	29.2	66	158	41.8
Serum findings	Elevated AST (> 37 units/L)	155	180	86.1	−	−	−
	Mild (< 148 units/L)	154	180	85.6	−	−	−
	Moderate/severe (≥ 148 units/L)	1	180	0.1	−	−	−
	Elevated ALT (> 42 units/L)	3	180	1.7	−	−	−
	Elevated bilirubin (> 17 μmol/L)	4	180	2.2	−	−	−

ALT = alanine aminotransferase; AST = aspartine aminotransferase; EPG = eggs per gram; IPB = image pattern B.

^*Undernutrition = weight for age, height for age, or weight for height < −2 standard deviations below mean (WHO).

^†Organomegaly = span > 2 standard deviations above mean for height as determined by Yazadanpanah and others.

Baseline associations between schistosome infection and IPB as predictive factors for the morbidity/finding outcomes are shown in Table 2 (see Supplemental Table 2 for P-values). No associations between infection status and morbidity complaints by parent report were significant (not shown). S. mansoni infection was associated with left lobe hepatomegaly (aPR = 1.4, 95% CI: 1.0–2.1) and splenomegaly (aPR = 2.1, 95% CI: 1.2–3.7). In the < 94 cm subset, the association between schistosomiasis and splenomegaly remained significant (Supplemental Table 4: aPR = 5.2; 95% CI: 1.001–27.3) and a univariable association between schistosomiasis and anemia was seen (PR = 1.2; 95% CI: 1.0, 1.4).

Table 2.

Associations at baseline visit of morbidities and findings with S. mansoni infection status by stool and serum testing, and IPB

Finding	Schistosomiasis (Egg+ or Ab+)				IPB
	Unadjusted		Adjusted*		Unadjusted		Adjusted*
	PR or coefficient (95% CI)	N	PR or coefficient (95% CI)	N	PR or coefficient (95% CI)	N	PR or coefficient (95% CI)	N
Anemia	1.1 (0.9, 1.3)	193	1.1 (0.9, 1.4)	186	0.9 (0.7, 1.2)	195	0.9 (0.6, 1.2)	188
Undernutrition	0.8 (0.4, 1.5)	171	0.8 (0.4, 1.4)	167	1.6 (1.0, 2.7)	181	1.8 (1.1, 2.9)	176
Left lobe hepatomegaly	1.5 (1.0, 2.2)	179	1.4 (1.0, 2.1)	172	0.6 (0.3, 1.4)	181	0.7 (0.3, 1.5)	174
Splenomegaly	2.4 (1.4, 4.2)	182	2.1 (1.2, 3.7)	174	1.5 (0.9, 2.6)	184	1.7 (0.9, 3.0)	176
AST (units/L)†	−5.1 (−10.8, 0.5)	177	−5.4 (−10.3, −0.6)	171	1.5 (−6.6, 9.7)	179	2.2 (−5.7, 10.2)	173
ALT (units/L)†	0.2 (−3.7, 4.1)	177	−1.0 (−4.3, 2.3)	171	0.1 (−4.5, 4.7)	179	0.9 (−2.9, 4.8)	173
Log bilirubin (μmol/L)‡	−0.2 (−0.3, −0.1)	177	−0.2 (−0.3, −0.1)	171	−0.8 (−1.8, 0.9)	179	−1.3 (−1.8, 1.0)	173

ALT = alanine aminotransferase; AST = aspartine aminotransferase; CI = confidence interval; PR = prevalence ratio. Bolded values are statistically significant at P < 0.05.

^*Adjusted for sex and age.

^†Increase (units/L) with presence of predictor.

^‡Ratio to value in uninfected.

Small, significant negative associations were seen between schistosomiasis and AST and bilirubin levels. The negative association with bilirubin persisted in the < 94 cm subset.

Associations between potential etiologies of IPB as predictive factors and IPB itself as an outcome were as follows (see Supplemental Table 1). At baseline, heavy intensity as compared with light intensity infection among egg+ children was associated with increased risk for IPB (aPR = 11.4, 95% CI: 2.7–48.3). The association with heavy intensity infection remained strongly significant in the < 94 cm subset of children (aPR = 7.9, 95% CI: 2.3–27.8). Hepatitis B could not be modeled as a predictive factor due to rarity, but neither of the two children positive for hepatitis B surface antigen had IPB. In the < 94 cm subset, IPB was associated with undernutrition (aPR = 1.3; 95% CI: 1.03–1.7) and splenomegaly (aPR = 3.8; 95% CI: 2.3–6.4).

No previously undescribed ultrasound findings associated with schistosomiasis in young children were found. Radiologist characterizations of liver texture and echogenicity were also not associated with liver pattern as designated by the sonographer.

Cross-sectional analysis of posttreatment visit data.

Of 159 children with schistosomiasis at baseline, 152 were locatable and received praziquantel. 187 of the children examined in the baseline visit (93.0%) participated in follow-up at the posttreatment visit. Two children with AST and ALT > 300 U/L were excluded as having probable acute liver disease. Among included children, mean age was 4.2 years, 46.7% (84/180) were male, and mean height was 98.8 cm. The only outcome with missing data was organometry at 17%. The only covariable with missing data was age at 3%.

Prevalences of outcomes and covariables are shown in Table 1. Notably, at the posttreatment visit, prevalence of egg+ S. mansoni infection decreased from 44.9% to 34.4% (28.6% in children < 2.5 years, 36.7% in children 2.5–< 4 years, 33.0% in children > 4 years) whereas heavy-intensity infection decreased from 10.1% to 5.0%. Malaria prevalence increased from 3.5% by smear at baseline to 16.7% by smear and 38.9% by RDT at follow-up. Prevalence of IPB increased from 15.0% to 20.0%; left lobe hepatomegaly, 50.0% to 65.1%; and splenomegaly, 29.2% to 41.8%.

Associations between follow-up S. mansoni infection, malaria or IPB, and follow-up morbidity and other outcomes are shown in Table 3 (see Supplemental Table 3 for P-values). Right liver lobe atrophy was again too rare (N = 2) to analyze as an outcome. S. mansoni infection remained associated with left lobe hepatomegaly on univariable and adjusted analyses (aPR = 1.4, 95% CI: 1.1–1.6), but it was not associated with splenomegaly. With RDT as the mode of malaria testing, malaria was associated with both hepatomegaly (aPR = 1.3, 95% CI: 1.0–1.7) and splenomegaly (aPR = 2.8, 95% CI: 1.9–4.0). In the under-94 cm subset, egg-positive schistosomiasis remained associated with left lobe hepatomegaly (aPR = 1.8, 95% CI: 1.3–2.6), and malaria by RDT remained associated with both left lobe hepatomegaly (aPR = 1.8, 95% CI: 1.3–2.6) and splenomegaly (aPR = 3.0, 95% CI: 1.6–5.8). (Supplemental Table 5).The interaction term between malaria by RDT and schistosomiasis was not significant.

Table 3.

Unadjusted and adjusted associations at posttreatment visit of morbidities and findings with S. mansoni infection status by stool testing alone, IPB, and malaria

Finding	Schistosomiasis (egg+)				IPB				Malaria (RDT positive)
	Unadjusted		Adjusted*		Unadjusted		Adjusted†		Unadjusted		Adjusted*
	PR (95% CI)	N	PR (95% CI)	N	PR (95% CI)	N	PR (95% CI)	N	PR (95% CI)	N	PR (95% CI)	N
Left lobe hepatomegaly	1.3 (1.1, 1.6)	152	1.4 (1.1, 1.6)	147	0.7 (0.4, 1.0)	152	0.6 (0.4, 1.0)	147	1.3 (1.0, 1.7)	152	1.3 (1.0, 1.7)	147
Splenomegaly	1.0 (0.6, 1.6)	158	0.9 (0.6, 1.6)	153	1.2 (0.8, 1.7)	158	1.2 (0.8, 1.7)	153	2.8 (2.0, 4.0)	158	2.7 (1.9, 4.0)	153

CI = confidence interval; PR = prevalence ratio. Bolded values are statistically significant at P < 0.05.

^*Values drawn from the same model (including sex, age, malaria on RDT, and current S. mansoni infection by stool egg status).

^†Sex and age.

Associations between potential IPB etiologies as predictive factors and IPB as an outcome were as follows. Malaria, by smear (aPR = 1.6, 95% CI: 1.2–2.1) and RDT (aPR = 2.1, 95% CI: 1.3–3.2), was strongly associated with IPB. S. mansoni infection again was not associated with IPB. In contrast to the baseline visit, moderate infection intensity was negatively associated with IPB (aPR = 0.15; 95% CI: 0.03–0.75) and heavy intensity was not associated.

Longitudinal analysis of data from both visits.

Of the 201 children examined in the baseline visit, 135 (67.2%) met analytic inclusion criteria for the longitudinal analysis. Covariables with missing data were age at 3.7% and malaria slide at 1.5%. Of 58 included children who were egg+ at baseline, 46 (79.3%) demonstrated a decrease in infection intensity (responders).

On univariable analyses, there were statistically significant increases between visits in the prevalence of hepatomegaly (PR = 1.3, 95% CI: 1.1–1.5) and splenomegaly (PR = 1.4, 95% CI: 1.1–1.8). On multivariable analyses, including malaria status by RDT, this increase remained significant for hepatomegaly (aPR = 1.2, 95% CI: 1.0–1.5) but not for splenomegaly. When children were stratified by initial infection status, increases in hepatomegaly prevalence remained significant on univariable analysis only for the group that was uninfected at baseline (PR = 1.2; 95% CI: 1.0–1.5); and in splenomegaly only in the group that was schistosomiasis infected at baseline (PR = 2.8; 95% CI: 1.1–6.7); increases were not significant on multivariable analysis in these strata. On analyses restricted to responders, increases in hepatomegaly and splenomegaly prevalence between visits were no longer statistically significant.

Discussion

A high proportion of sampled young children were infected with S. mansoni in Rusinga Island, a previously treatment-naïve area. Our results showed an association between schistosomiasis and both hepatomegaly and splenomegaly in young children, which persisted for hepatomegaly after treatment with praziquantel. This association was still seen when analysis was restricted to children currently ineligible for MDA by virtue of height. Despite treatment of infected children, the overall prevalence of organomegaly increased between visits; we attribute this primarily to the increase in malaria prevalence, which was consistent with malaria seasonality in western Kenya.

The association of schistosomiasis with organomegaly is consistent with others' findings in older children.^29,33 Health implications of chronic hepatosplenomegaly in African children have been discussed recently by Wilson and others⁴⁴ and can include possible increased risk for stunting and wasting and, in extreme cases, esophageal varices. More broadly, an inflammatory mechanism is postulated in the genesis of hepatomegaly,⁴⁴ and long-term effects of this potential chronic inflammation are unknown.

In contrast to others,⁴⁵ we found no evidence for rapid improvement in organomegaly after praziquantel treatment, including among children whose stool egg densities decreased. The rising prevalence of malaria between visits is a likely cause; misclassification of children as responders due to low sensitivity of Kato-Katz on a single stool sample is another.

IPB, considered in adults to represent a possible intermediate stage in development of schistosomiasis-associated hepatic fibrosis, was not primarily attributable to schistosomiasis in our sample—consistent with recent findings in slightly older children⁴⁶—and in fact appears more closely related to malaria infection. The biological significance of IPB is not known; it may represent a common pathway for multiple abnormal hepatic processes. Regardless, our findings suggest that this element of the Niamey protocol is not a measure of schistosomiasis-associated pathology at least in young children.

The high prevalence of modest AST elevations may reflect the need for a region-specific reference range, such as that recently proposed based on population biochemical data from western Kenya, giving 50.4 units/L as the upper limit of normal.⁴⁷ The reason for the negative association between S. mansoni infection and AST and bilirubin at baseline is unclear.

The low apparent curative efficacy of 40 mg/kg praziquantel against S. mansoni in our sample is striking, agreeing with some previous findings in PSAC¹¹ and contrasting with others.^10,48 Given the high transmission pressure in this setting, treatment failure is difficult to distinguish from reinfection, and others have noted substantial reinfection after treatment, though over somewhat longer intervals.⁴⁹ Concern has been raised about whether the extension of current height-based dosing calculations to young children would result in their undertreatment, and our results may further substantiate those concerns.

Several limitations of this study are worth noting. A convenience sample was used, selected from the highest-prevalence villages to ensure data were balanced by infection status. A few children initially identified as PSAC were found in home documentation to be school-aged. Further, unmeasured confounders may have affected the observed associations, given the large number of possible causes for organomegaly and the unknown causes and clinical importance of IPB. The sample size for this study was small, which may have resulted in missed associations, and larger studies are needed to confirm statistically significant associations. Water contact data, which might have been helpful in distinguishing nonresponders from reinfected children, was not collected. It is also possible that the multiple tests performed in this study at baseline might have led to appearance of spurious associations; however, at the posttreatment visit only associations found to be significant at the baseline visit were tested, and the persistence of the observed relationships suggests the key findings are robust. Finally, the lack of correlation between on-site liver pattern determination by sonographer and off-site echotexture and echogenicity determination by radiologist review may reflect loss of resolution in image transfer, but most importantly reflects the incompleteness of characterizing IPB in terms of echotexture and echogenicity. It became apparent in exploring this lack of correlation that echotexture and echogenicity did not reliably predict IPB even using selected teaching images.

The associations found here, particularly if confirmed in larger population-based studies, add to the growing body of evidence for measurable sequelae associated with schistosomiasis in children currently excluded from the global schistosomiasis control program by virtue of their young age. In addition to scaling up sanitation measures that benefit all age groups, this supports the continued pursuit of appropriate formulations and dosing standards for MDA in this population.

Disclaimer: The findings and conclusions in this report are the findings and conclusions of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.