Advertisement
Research Article

The Impact of Schistosoma japonicum Infection and Treatment on Ultrasound-Detectable Morbidity: A Five-Year Cohort Study in Southwest China

  • Elizabeth J. Carlton mail,

    ejcarlton@berkeley.edu

    Affiliation: Department of Environmental Health Sciences, University of California, Berkeley, California, United States of America

    X
  • Michelle Hsiang,

    Affiliation: Department of Pediatrics, University of California San Francisco, San Francisco, California, United States of America

    X
  • Yi Zhang,

    Affiliation: Institute for Parasitic Diseases, Sichuan Center for Disease Control and Prevention, Chengdu, China

    X
  • Sarah Johnson,

    Affiliation: Department of Environmental Health Sciences, University of California, Berkeley, California, United States of America

    Current address: New York City Department of Health and Mental Hygiene, New York, New York, United States of America

    X
  • Alan Hubbard,

    Affiliation: Department of Biostatistics, University of California, Berkeley, California, United States of America

    X
  • Robert C. Spear

    Affiliation: Department of Environmental Health Sciences, University of California, Berkeley, California, United States of America

    X
  • Published: May 18, 2010
  • DOI: 10.1371/journal.pntd.0000685

Abstract

Background

Ultrasonography allows for non-invasive examination of the liver and spleen and can further our understanding of schistosomiasis morbidity.

Methodology/Principal Findings

We followed 578 people in Southwest China for up to five years. Participants were tested for Schistosoma japonicum infection in stool and seven standard measures of the liver and spleen were obtained using ultrasound to evaluate the relationship between schistosomiasis infection and ultrasound-detectable pathology, and the impact of targeted treatment on morbidity. Parenchymal fibrosis, a network pattern of the liver unique to S. japonicum, was associated with infection at the time of ultrasound (OR 1.40, 95% CI: 1.03–1.90) and infection intensity (test for trend, p = 0.002), adjusting for age, sex and year, and more strongly associated with prior infection status and intensity (adjusted OR 1.84, 95% CI: 1.30–2.60; test for trend: p<0.001 respectively), despite prompt treatment of infections. While declines in parenchymal fibrosis over time were statistically significant, only 28% of individuals with severe parenchymal fibrosis (grades 2 or 3) at enrollment reversed to normal or grade 1 within five years. Other liver abnormalities were less consistently associated with S. japonicum infection.

Conclusions/Significance

Parenchymal fibrosis is an appropriate measure of S. japonicum morbidity and can document reductions in disease following control efforts. Other ultrasound measures may have limited epidemiological value in regions with similar infection levels. Because severe fibrosis may not reverse quickly following treatment, efforts to reduce exposure to S. japonicum should be considered in combination with treatment to prevent schistosomiasis morbidity.

Author Summary

Schistosomiasis is a water-borne parasite that infects approximately 200 million people worldwide. Schistosoma japonicum, found in Asia, causes disease by releasing eggs in the liver, leading to fibrosis, anemia, and, in children, impaired growth. Ultrasound can assess liver pathology from schistosomiasis; however more information is needed to evaluate the relevance of standard ultrasound measures. We followed 578 people for up to five years, testing for schistosomiasis infection and conducting ultrasound examinations to assess the relationship between infection and seven ultrasound measures and to evaluate the impact of treatment with anti-schistosomiasis chemotherapy (praziquantel) on morbidity. All infections were promptly treated. Fibrosis of the liver parenchyma, pathology unique to S. japonicum, was associated with schistosomiasis infection, and was most advanced in people with high worm burdens. Liver fibrosis declined significantly following treatment, but reversal of severe liver fibrosis was rare. Other ultrasound measures were not consistently related to schistosomiasis infection or treatment. These findings suggest parenchymal fibrosis can be used to measure morbidity attributable to S. japonicum and evaluate the impact of disease control efforts. Because reversal of severe fibrosis was limited, disease control efforts will be most effective if they can not only treat existing infections but also prevent new infections.

Introduction

Schistosomiasis causes morbidity in the human host through the schistosome egg, which triggers inflammation and fibrosis that can lead to anemia, impaired growth and in severe cases, gastrointestinal bleeding and death [1][4]. The major intestinal schistosomes, Schistosoma japonicum, found in Asia, and S. mansoni, found in Africa, the Americas and the Middle East, mature, mate and lay eggs in the portal and mesenteric blood vessels. Eggs are transported to the liver where they are encapsulated and the granulomas that form induce an inflammatory cascade that includes the deposition of collagen and extracellular matrix proteins, a normal liver repair process that can lead to fibrosis when fibrogenesis exceeds the replacement of scar tissue with healthy cells [5], [6]. The immune regulation of this process is currently being explored [7]. Approximately 700,000 people are infected with S. japonicum in China [8]. As in other parts of the world, schistosomiasis control efforts have focused primarily on the distribution of the antischistosomal drug, praziquantel [9], [10]. However the success of such efforts hinges on the ability to reduce not only schistosomiasis infections, but also morbidity. A means of documenting S. japonicum morbidity is essential to the evaluation of disease control efforts [11].

Ultrasound is a non-invasive method that can be used to evaluate fibrosis resulting from schistosomiasis infection. Fibrosis along the portal vein and its branches produces a clay pipestem pattern, as the portal tracks are lined with fibrous tissue, and is observed following both S. mansoni and S. japonicum infection. This periportal fibrosis can be assessed qualitatively through image classification [12] or quantitatively by measuring the diameter of three secondary portal branches [13]. Unique to S. japonicum infection is parenchymal fibrosis, a network pattern that is often described as fish scale or tortoise shell-like. Parenchymal fibrosis is likely due to the smaller S. japonicum egg size which allows the parasite eggs to enter smaller portal veins and reach a greater portion of the liver [7]. S. japonicum adult females produce ten times more eggs per day than S. mansoni, and the eggs often are deposited in clusters, two additional factors that exacerbate the severity of S. japonicum morbidity relative to other schistosome species and may contribute to the unique fibrotic pattern [3]. Ultrasound can also be used to assess hepatomegaly, splenomegaly and dilation of the portal vein, all of which result from portal hypertension. Recognizing the potential of ultrasound to be used to evaluate the impact of disease control efforts, draft protocols for the use of ultrasound in assessing schistosomiasis morbidity were established for each of the three major schistosome species in 1990 [13], [14]. Follow-up meetings to refine the protocols for S. haematobium and S. mansoni were held in 1996 and 1997 [15]. Ultrasound is considered the gold standard for schistosomiasis morbidity assessment for these species [11]. However, due to insufficient evidence the protocol for S. japonicum has not been revised.

To date, a comprehensive evaluation of the S. japonicum ultrasound measures proposed in Cairo, including their relationship to S. japonicum infection and the way in which they change following treatment, is lacking. Li et al. [16], [17] offer the most complete examination of the proposed ultrasound measures to date, tracking a highly exposed cohort over five years, but as standard organ sizes have only recently been published [18], assessments of liver and spleen enlargement did not account for participant height as currently recommended [15]. Failure to account for height-specific variation in organ sizes can lead to underestimates of morbidity, particularly in pediatric populations [19]. Ideally, ultrasound measures are associated with S. japonicum infection and decline following treatment [13]. The available data is conflicting on both criteria. Treatment led to declines in parenchymal and periportal fibrosis in a highly exposed Chinese cohort; however 68% of participants experienced no change in parenchymal fibrosis over a five-year period [16], [17] and other studies have not found significant declines in periportal fibrosis following S. japonicum treatment [20], [21]. Further, a direct relationship between infection and fibrosis or measures of portal hypertension has not been demonstrated consistently [16], [17], [22].

We followed 578 individuals over five years in order to examine the relationship between S. japonicum infection and five liver ultrasound measurements recommended in the draft protocol as well as two spleen ultrasound measurements included in the standard Chinese examination. Specifically, we hypothesized that S. japonicum infection is associated with ultrasound-detectable measures of hepatic fibrosis and that treatment of infected individuals leads to declines in ultrasound-detected morbidity.

Methods

Participants were drawn from a cross-sectional survey in 2000 that revealed high infection prevalence (ranging from 3% to 73%) in 20 villages distributed in the hilly terrain of Sichuan Province in Southwest China where irrigated agriculture for the cultivation of rice, wheat, corn and tobacco is the primary source of S. japonicum infection [23]. As part of the survey, all residents aged 4 to 60 years in 20 villages Xichang County, Sichuan Province, China were invited to be tested for S. japonicum infection and to answer a brief questionnaire. We used a random number generator to select approximately 30% of the population, stratified by village and occupation, for ultrasound examination. Ten villages with high infection prevalence in 2000 were selected to be followed longitudinally using infection surveys and ultrasound examinations over the next five years. Participants in our cohort included individuals who had an ultrasound examination in 2000 and lived in one of the ten villages followed through 2005.

Ethics statement

All participants provided oral informed consent and were provided treatment following S. japonicum positive stool examinations. As the survey procedures used in this study are the same as those used by the Institute of Parasitic Diseases, Sichuan Center for Disease Control and Prevention (IPD) for schistosomiasis surveillance and given the high rates of illiteracy in the population, oral informed consent was obtained and documented by IPD staff. The research protocol and consent procedures were approved by the Sichuan Institutional Review Board and the University of California, Berkeley Committee for the Protection of Human Subjects.

Ultrasound examination

Ultrasound examinations were conducted in fall 2000, 2002 and 2005. All examinations were conducted by one trained ultrasonographer (YZ) using a single portable ultrasound machine (Hitachi EUB 405, Hitachi Corporation, Tokyo, Japan) and 3.5 MHz probe (Hitachi EUP-C314T, Hitachi Corporation, Tokyo, Japan) with participants in the supine position at a central location in each village. The ultrasonographer was blind to infection status.

Liver ultrasonography was conducted according to the 1990 draft guidelines [13], [14]. Liver parenchymal fibrosis was graded 1 through 3 based on observed lesions, or 0 if none were present. Periportal fibrosis was assessed by grading the average diameter, from outer wall to outer wall, of three peripheral branches of the portal vein between the first and third branching point (grade 0: <3mm; grade 1: 3 to 5 mm; grade 2: >5 to 7 mm; grade 3: >7 mm). As done previously, grades 0 and 1 were combined for analysis [16]. The internal diameter of the portal vein was measured at the entry point of the portal vein into the liver. The length of the left liver lobe was measured in a longitudinal section along the left parasternal line, and the length of the right liver lobe was measured as the maximum oblique diameter using a right anterior axillary view according to the revised guidelines established for S. mansoni [15].

Two measurements of the spleen that are part of the standard Chinese examination were also included: spleen thickness, measured from the hilum to the opposite section, and the internal diameter of the spleen vein, measured at the entry point to the spleen [24].

Because organ and vein sizes vary with height, left and right liver lobe length, portal vein diameter and spleen thickness were evaluated using height-specific standard sizes drawn from a Chinese population where schistosomiasis is not endemic [18]. Measurements greater than two standard deviations above the mean size for height were classified as abnormal. Height was measured in 2002 in 440 of the 578 cohort members. The 343 adults (≥18 years old in 2000) with height measurements were assigned their 2002 height throughout the study. Children (<18 years old in 2000) were measured again in 2007 as part of a study described elsewhere [19]. The height measurements from 2002 and 2007 from 119 children (60 with 1 height measurement, 59 with two height measurements) were used to generate an age- and sex-dependent random intercept model in order to impute heights during the years children weren't measured. Height for child i at time j was calculated using the following equation:
where Aij represents a child's age at time j, Si represents his or her sex (S = 1 for males) and represents his or her random intercept. The model was fit using xtmixed in Stata 10.1 (StataCorp, College Station, TX, USA) and was predicted using empirical Bayes [25]. As random intercept models assume parametric distribution of residuals, first and second order residuals were examined and were normally distributed. The selected model was superior to a model that did not include a random intercept (likelihood ratio test, χ2 = 38.20, 1 d.f., p<0.001). Organ and vessel measures could not be height adjusted for the 86 adults and 26 children without height measurements. Participants with any versus no height data did not differ significantly by any of the predictors examined in the analyses including sex, mean age or infection status. No height standardized values were available for spleen vein diameter, so it was evaluated by the standard threshold used in China: >8 mm [24].

Infection and treatment

Participants were tested for infection with S. japonicum in the fall of 2000 and 2002 using the miracidial hatch test [26] and the Kato-Katz thick smear procedure [27]. For each hatch test, a stool sample weighing at least 30 g was suspended in aqueous solution, filtered using copper mesh to remove large particles (40–60 mesh) followed by nylon gauze (260 mesh) to concentrate schistosome eggs. This sediment was re-suspended with distilled water in a 250 ml Erlenmeyer flask. Flasks were examined for miracidia 30–60 minutes, 4 hours and 8 hours after suspension if temperatures were above 30 degrees Celsius, or at 6, 12 and 18 hours at lower ambient temperatures. In 2000, three miracidial hatch tests were conducted per person using stool samples from three distinct days. In 2002, due to logistical constraints, one miracidial hatch test was conducted per person; however the Kato-Katz protocol was identical both years. The Kato-Katz procedure involved the preparation of three 41.5 mg slides from one homogenized stool sample in 2000 and 2002. Infection intensity, in eggs per gram of stool (EPG), was calculated as the total number of S. japonicum eggs present on the slides divided by the total sample weight. Participants were classified as infected if at least one test was positive. Everyone testing positive for S. japonicum was provided praziquantel treatment by the county Anti-Schistosomiasis Control Station. In addition, praziquantel was administered to all residents in the study villages in 2003, as part of a nation-wide effort to control infectious diseases following the outbreak of severe acute respiratory syndrome.

Participant age, sex, occupation and highest level of schooling were obtained by interview in fall 2000.

Statistical analysis

In order to assess whether the participants in the ultrasound cohort were representative of the village populations from which they were selected, cohort participants were compared to individuals who participated only in the cross-sectional demographic and infection surveys in terms of age, sex, occupation, educational attainment and 2000 S. japonicum infection status. Similarly, cohort members with complete vs. incomplete follow-up were compared in terms of age, sex, baseline morbidity and infection in order to assess non-differential loss to follow-up. In both cases, comparisons were conducted using the χ2 and t-test.

Analyses that included multiple observations from individuals accounted for within subject correlation of outcomes. Liver parenchymal fibrosis grade, an ordinal measure, was examined using ordinal logistic regression, a population averaged model using a sandwich type estimator for inference accounting for within-subject residual correlation [28]. Because ordinal logistic regression assumes the effect of a predictor on an outcome is constant for each stepwise increase in the outcome, the Brant test [29] was used to check that this parallel regression assumption was not violated. For all other liver and spleen measures and for predictors of S. japonicum infection status, generalized estimating equation (GEE) logistic regression with exchangeable correlation was conducted [30]. The Huber/White/sandwich estimator of variance was used which is robust to misspecification of the outcome distribution [31], [32].

The relationship between S. japonicum infection and ultrasound-detected abnormalities was first assessed by examining the impact of current infection status and, separately, infection intensity, on current ultrasound measures. Because infection intensity was highly right skewed, it was categorized into approximate quartiles among those who were infected. In order to examine the role of past infection on current morbidity, we also examined the relationship between infection status and intensity two to three years prior to the ultrasound examination and ultrasound-detected abnormalities (for example: 2000 infection status as a predictor of ultrasound-detected morbidity in 2002, and 2002 infection status as a predictor of ultrasound-detected morbidity in 2005). We hypothesized that age, sex and year of examination could modify the effect of infection on morbidity, so each model was run including all possible first-order interaction terms. The Wald test was used to test for significant interactions and if present, terms were removed from the model step-wise until only interaction terms significant at p-values <0.05 remained. In the absence of effect modification, these same variables could confound the relationship between infection and ultrasound-detected morbidity. While occupation, village and educational status were considered potential predictors of S. japonicum infection, they are unlikely to effect ultrasound morbidity independent of infection status. As they do not fit the definition of confounders [33], they were not controlled for in models examining the relationship between infection and ultrasound-detected abnormalities. The change in ultrasound-detected abnormalities over time was examined adjusting for age and sex.

We modeled age as a categorical variable when examining the relationship between age and ultrasound-detected morbidity to allow a non-linear relationship. Because liver abnormalities increase with age, age was treated as a continuous variable when included as a confounder in models testing the relationship between infection and morbidity, or changes in morbidity over time. Periportal fibrosis grade was modeled as a binary variable, as grades 0 and 1 were combined and no grade 3 fibrosis was detected. Tests for trend were calculated by treating categorical variables as ordinal. All results were assessed for statistical significance setting α = 0.05. Statistical analyses were conducted using Stata 10.1 (StataCorp, College Station, TX, USA).

Results

In 2000, 578 people from ten villages were examined using ultrasound. The mean age of participants was 29.8 years (range 4–61 years). Half (51%) were female (Table 1). Most adults (≥18 years) were farmers (91%) and had no formal schooling beyond elementary school (59%). Participants were similar in terms of occupation, educational attainment and 2000 infection status to the 1,333 residents in these villages who participated in infection and demographic surveys only, although cohort members were slightly older than the rest of the population (mean age 29.8 vs. 28.2, p = 0.037).

thumbnail

Table 1. Characteristics of the 578 participants in the ultrasound cohort in 2000.

doi:10.1371/journal.pntd.0000685.t001

We conducted ultrasound examinations with 444 people in 2002 and 321 people in 2005. The 320 people with complete ultrasound follow-up were no more likely to be infected with S. japonicum at enrollment than those who missed at least one follow-up examination, but they were older (mean age 33.0 vs. 25.6, p<0.001) and more likely to have at least one liver abnormality in 2000 (57% vs. 37%, p<0.001).

Nearly half (47%) of participants tested positive for S. japonicum in 2000. Mean infection intensity was 53.4 EPG. In 2002, infection prevalence declined to 32% and intensity to 9.4 EPG. As shown in Table 2, adults aged 50 and older were less likely to be infected than younger participants. Neither sex nor occupation was associated with infection prevalence, but among adults, higher educational attainment was protective. Infection prevalence in 2002 was significantly higher among those who were infected vs. uninfected in 2000 despite the distribution of treatment to everyone who tested positive (41% vs. 23%, p<0.001). Infection prevalence varied significantly by village.

thumbnail

Table 2. The distribution of S. japonicum infection by age, sex, occupation, educational attainment, year and village.

doi:10.1371/journal.pntd.0000685.t002

Table 3 describes the prevalence of liver parenchymal fibrosis, periportal fibrosis and abnormal liver and spleen measurements from 2000 to 2005. Following the initiation of schistosomiasis testing and treatment of all infections in 2000, parenchymal fibrosis, periportal fibrosis and right liver lobe enlargement decreased significantly through 2005, controlling for age and sex (Table 4). Decreases in spleen enlargement were also observed, although the trend was of marginal significance (p = 0.091).

thumbnail

Table 3. Prevalence of abnormal liver and spleen measurements from 2000 to 2005.

doi:10.1371/journal.pntd.0000685.t003
thumbnail

Table 4. The relationship between ultrasound-detected abnormalities and age, sex and year.

doi:10.1371/journal.pntd.0000685.t004

Liver abnormalities increased significantly with age, most notably for parenchymal fibrosis (Table 4). Spleen enlargement was not associated with age. Individuals aged 18 to 29 years had the highest odds of spleen vein dilation. The relationship between liver abnormalities and sex varied: men were more likely to have periportal fibrosis and enlarged right liver lobes; women were more likely to have enlarged left liver lobes.

Schistosomiasis infection at the time of ultrasound was associated with an increase in liver parenchymal fibrosis grade (OR 1.40, 95% CI: 1.03–1.90) adjusting for age, sex and year of ultrasound (Table 5). Infection intensity at the time of ultrasound was also associated with an increase in liver parenchymal fibrosis grade (test for trend: p = 0.002). Individuals with greater than 50 EPG had 2.10 times greater odds of more advanced fibrosis than those not excreting eggs (95% CI: 1.33–3.32).

thumbnail

Table 5. The impact S. japonicum infection on ultrasound-detected liver abnormalities.

doi:10.1371/journal.pntd.0000685.t005

Schistosomiasis infection two to three years prior to ultrasound was associated with an increase in liver parenchymal fibrosis grade, and the association was stronger than that of current infection (OR 1.84, 95% CI: 1.30–2.60). Prior infection intensity was also associated with liver parenchymal fibrosis (test for trend, p<0.001). Individuals with greater than 50 EPG two to three years prior to the ultrasound examination had 2.84 times greater odds of advanced fibrosis than those not excreting eggs two to three years prior to ultrasound (95% CI: 1.71–4.73).

The other hepatosplenic ultrasound measures were not associated with current infection status or intensity. Several measures were associated with prior infection, although the relationships were not as consistent as those observed for periportal fibrosis. Prior infection appeared to elevate the probability of left liver lobe enlargement (OR 1.45, 95% CI: 0.97–2.17). Prior infection intensity but not prior infection status, was associated with increased odds of portal vein dilation (test for trend, p = 0.051). The impact of prior infection status on right liver lobe enlargement varied by the year of examination and the sex of the participant. Prior infection was associated with increased odds of right liver lobe enlargement among males examined in 2005 (OR 3.95, 95% CI: 1.82–8.57) and decreased odds of right liver lobe enlargement among females examined in 2002 (OR 0.33, 95% CI: 0.13–0.86). Prior infection intensity was not associated with right or left liver lobe enlargement. Due to the limited number of individuals with periportal fibrosis, models were unable to yield stable estimates of the effect of prior infection on this measure Spleen enlargement and spleen vein dilation were not associated with current or prior infection.

Most people with liver and spleen enlargement, portal vein dilation, periportal fibrosis or spleen vein dilation in 2000 had normal pathology by 2002 (Table 6). However, this was not the case for parenchymal fibrosis: 67% of people with grade 2 fibrosis and 100% with grade 3 fibrosis in 2000 remained at or above grade 2 throughout the five-year follow-up.

thumbnail

Table 6. The reversal of ultrasound-detected liver and spleen abnormalities among individuals with abnormal measures in 2000.

doi:10.1371/journal.pntd.0000685.t006

Discussion

We found evidence of a direct, exposure-response relationship between S. japonicum infection and parenchymal fibrosis. While there has been suggestive evidence of an association between infection and parenchymal fibrosis, including an association between progression of parenchymal fibrosis and current infection [16], this is the first study to show the risk of parenchymal fibrosis is higher in people who are infected vs. uninfected, and highest in individuals with the greatest infection intensities. Parenchymal fibrosis declined significantly following treatment however, improvements were limited among individuals with advanced fibrosis: 72% of people with severe fibrosis at enrollment (grades 2 or 3) had not resolved below grade 2 by the end of the five-year study. This suggests parenchymal fibrosis is an appropriate measure of S. japonicum morbidity and can document improvements in morbidity following treatment, although little improvement may be observed among those with advanced fibrosis.

In contrast, the remaining measures were less consistently associated with S. japonicum infection and are of questionable epidemiological use in regions with similar infection levels. Periportal fibrosis was rare in this population and could not be associated with S. japonicum infection, although it did decline significantly following the initiation of targeted treatment. Others have used image-based classification to assess periportal fibrosis [34], [35], which has been shown to have better reproducibility for S. mansoni-related fibrosis, perhaps because it does not require the ultrasonographer to measure narrow vessel widths [36]. Our measures of periportal fibrosis were not height adjusted, as Chinese standards for the diameter of portal vein branches as measured here have not been published. The lack of height adjustment may have led to underestimates of the prevalence of periportal fibrosis and may explain the higher prevalence of periportal fibrosis in men, however higher prevalence of fibrosis has also been observed in males using image-based classification [35].

The remaining three hepatic measures, left and right liver lobe enlargement and portal vein dilation, were associated with S. japonicum infection or declined following chemotherapy. However, the relationships between S. japonicum infection and these morbidity measures were less consistent than those observed with parenchymal fibrosis and several associations were of marginal statistical significance. The spleen measures were neither associated with infection nor did they decline with treatment. All five measures require the ultrasonographer to measure organ or vessel sizes and participant's height. Like the measures of periportal fibrosis used here, these methods present opportunities for measurement error which may bias estimates of exposure-disease relationships toward the null, producing attenuation and exacerbating non-linearity [37]. The reproducibility of some of these measures has been shown to be poor in a pediatric population, and further efforts to evaluate the accuracy of organ and vein size measurements are needed [19].

Overall, the prevalence of morbidity in this cohort was lower than has been observed elsewhere in China as was infection intensity [16], [17], [34], [38]. Measures of liver and spleen enlargement, portal vein dilation and spleen vein dilation were less informative than measures of parenchymal fibrosis but may be appropriate in areas where infection intensity and associated morbidities are higher. Declines in hepatomegaly and splenomegaly have been demonstrated following treatment in regions with higher average worm burdens [16], [17]. The strong associations between age and all liver measures highlight the importance of considering age as a potential confounder when examining the relationship between S. japonicum infection and morbidity and declines in morbidity over time. For example, the unadjusted prevalence of parenchymal fibrosis did not change over time, however as loss to follow-up was highest in younger populations which were least likely to have fibrosis, the age and sex adjusted prevalence did decline.

Our findings shed light on the development of hepatic fibrosis following S. japonicum infection. The lack of reversal of grade 3 parenchymal fibrosis to grade 1 or normal and the association between past infection and current fibrosis suggest severe fibrosis may persist or even progress following treatment. Prior studies have also found minimal reversal of severe fibrosis following treatment [16], [39] and hepatic fibrosis has been documented in people living in areas where the parasite, and therefore exposure, was eliminated 20 years previously [40]. While treatment with praziquantel can reduce infection prevalence, intensity and fibrosis, this analysis provides further evidence that severe hepatic fibrosis may be unlikely to reverse quickly. Minimal declines in severe hepatic fibrosis associated with S. mansoni infection have also been detected [41] but other studies have noted complete reversal of severe morbidity [42].

Further, parenchymal fibrosis was associated with current infection status and intensity but more strongly associated with infection status and intensity two to three years prior to ultrasound, despite timely treatment of infections. The surprising association between past infection and fibrosis raises questions about the progression of fibrosis following infection and treatment. Praziquantel kills the adult worm but schistosome eggs can remain trapped in host tissues decades after exposure [43]. While schistosome eggs are thought to disintegrate within weeks of granuloma formation [7], it is possible, due to an inflammatory cascade, fibrogenesis continues over a longer time period. Collagen deposition has been shown to occur following treatment with praziquantel in S. japonicum infected mice, suggesting fibrosis may continue despite removal of adult worms [44]. In humans, progression of fibrosis has been observed following treatment and was not explained by reinfection [21], [45]. Alternatively, it is possible that the relationship between past infection and fibrosis is due to reinfection. Past infection predicted subsequent infection. However, if reinfection rather than past infection determined fibrosis, one would expect stronger relationships between current infection and fibrosis than between past infection and fibrosis, which was not observed. It is also possible that some individuals who were treated were not cured. A single dose of praziquantel cures 90% of S. japonicum infections [46], suggesting 10% of those treated may continue to harbor adult worms and face continued egg production. In our work in this region, we have also found individuals who decline to take praziquantel even after a positive infection test. Non-compliance and treatment failure are realities of any chemotherapy-based control program, however the marked declines in infection prevalence, intensity and morbidity suggest the number of individuals who were not cured was limited.

This study examined one of the largest populations to be followed over five years in order to assess ultrasound-detectable morbidity and S. japonicum infection. Participants were randomly sampled from a comprehensive cross-sectional survey in order to minimize selection bias. Loss to follow-up was greater among those without ultrasound-detectable morbidity, which suggests the prevalence of morbidity in 2002 and 2005 may be overestimated, and therefore the true declines in morbidity over time may be greater than observed. Retention was independent of infection status, minimizing bias in estimates of the impact of infection on pathology. No information was available on alcohol consumption or infection with hepatitis B virus (HBV), two factors that can induce liver pathology and may exacerbate schistosomiasis morbidity. The parenchymal network pattern due to S. japonicum is distinct from the lesions produced by HBV, as HBV produces a finer, meshwork texture [47], suggesting the observed prevalence of parenchymal fibrosis is specific to schistosomiasis. While HBV has been shown to hinder regression of periportal fibrosis following treatment of S. mansoni infections [41], the extent to which HBV exacerbates morbidity due to S. japonicum remains unclear and warrants further study. Reductions in parenchymal fibrosis may be impaired by alcohol consumption, which in our study areas is confined almost exclusively to males [17]. Unless alcohol consumption or HBV impacts a person's probability of infection, they are unlikely to confound the relationships between infection and ultrasound-detected morbidity.

Community-wide testing and treatment of all infections with praziquantel yielded marked declines in infection prevalence and intensity. However, reinfection was high: 32% of people were infected two years after the first round of targeted treatment. High rates of infection following treatment are not uncommon and underscore the challenges of sustainably reducing human schistosomiasis [48]. Adults aged 50 and over were less likely to be infected than younger populations, possibly corresponding to a decline in water contact later in life or acquired immunity [49].

In summary, we present evidence that ultrasound-detectable liver fibrosis is associated with S. japonicum infection status and intensity, controlling for age, sex and year of ultrasound examination, and this measure can be used to monitor S. japonicum induced morbidity. Other ultrasound measures including hepatomegaly and splenomegaly were less clearly associated with infection. Our findings also suggest that some morbidity may not reverse within five years of treatment and may even progress despite treatment. Praziquantel has yielded remarkable declines in schistosomiasis morbidity in China and throughout the globe but reinfection following treatment is common as observed in this study population. S. japonicum infection, particularly high egg loads, may lead to fibrosis that is not rapidly reversed by treatment, underscoring the importance of measures to prevent new infections as well as treating current disease.

Supporting Information

Checklist S1.

STROBE checklist.

doi:10.1371/journal.pntd.0000685.s001

(0.09 MB DOC)

Alternative Language Abstract S1.

Translation of abstract into Chinese by YZ.

doi:10.1371/journal.pntd.0000685.s002

(0.03 MB DOC)

Acknowledgments

We thank Michael Bates, John Swartzberg, Lei Xuezhong and several anonymous reviewers for their thoughtful comments on earlier drafts of this manuscript. We thank Zhao Lianguo, the director of the Xichang County Anti-Schistosomiasis Control Station, and his staff for their help with data collection over the five years of this study and the participants in Xichang County who made this research possible.

Author Contributions

Conceived and designed the experiments: YZ RCS. Performed the experiments: YZ. Analyzed the data: EJC MH SJ AH. Wrote the paper: EJC. Edited the paper: EJC MH YZ AH RCS.

References

  1. 1. Leenstra T, Coutinho HM, Acosta LP, Langdon GC, Su L, et al. (2006) Schistosoma japonicum reinfection after praziquantel treatment causes anemia associated with inflammation. Infect Immun 74: 6398–6407.
  2. 2. King CH, Dickman K, Tisch DJ (2005) Reassessment of the cost of chronic helmintic infection: a meta-analysis of disability-related outcomes in endemic schistosomiasis. Lancet 365: 1561–1569.
  3. 3. Ross AG, Sleigh AC, Li Y, Davis GM, Williams GM, et al. (2001) Schistosomiasis in the People's Republic of China: prospects and challenges for the 21st century. Clin Microbiol Rev 14: 270–295.
  4. 4. Friedman JF, Kanzaria HK, Acosta LP, Langdon GC, Manalo DL, et al. (2005) Relationship between Schistosoma japonicum and nutritional status among children and young adults in Leyte, the Philippines. Am J Trop Med Hyg 72: 527–533.
  5. 5. Dessein A, Kouriba B, Eboumbou C, Dessein H, Argiro L, et al. (2004) Interleukin-13 in the skin and interferon-gamma in the liver are key players in immune protection in human schistosomiasis. Immunol Rev 201: 180–190.
  6. 6. Olaso E, Friedman SL (1998) Molecular regulation of hepatic fibrogenesis. J Hepatol 29: 836–847.
  7. 7. Burke ML, Jones MK, Gobert GN, Li YS, Ellis MK, et al. (2009) Immunopathogenesis of human schistosomiasis. Parasite Immunol 31: 163–176.
  8. 8. Zhou XN, Guo JG, Wu XH, Jiang QW, Zheng J, et al. (2007) Epidemiology of schistosomiasis in the People's Republic of China, 2004. Emerg Infect Dis 13: 1470–1476.
  9. 9. Hotez PJ, Molyneux DH, Fenwick A, Kumaresan J, Sachs SE, et al. (2007) Control of neglected tropical diseases. N Engl J Med 357: 1018–1027.
  10. 10. Utzinger J, Zhou XN, Chen MG, Bergquist R (2005) Conquering schistosomiasis in China: the long march. Acta Trop 96: 69–96.
  11. 11. Webster JP, Koukounari A, Lamberton PH, Stothard JR, Fenwick A (2009) Evaluation and application of potential schistosome-associated morbidity markers within large-scale mass chemotherapy programmes. Parasitology 1–11.
  12. 12. Doehring-Schwerdtfeger E, Mohamed-Ali G, Abdel-Rahim IM, Kardorff R, Franke D, et al. (1989) Sonomorphological abnormalities in Sudanese children with Schistosoma mansoni infection: a proposed staging-system for field diagnosis of periportal fibrosis. Am J Trop Med Hyg 41: 63–69.
  13. 13. Jenkins JM, Hatz C (1992) The use of diagnostic ultrasound in schistosomiasis–attempts at standardization of methodology. Cairo Working Group. Acta Trop 51: 45–63.
  14. 14. WHO (1991) Meeting on Ultrasonography in Schistosomiasis: Proposal for a Practical Guide to the Standardized use of Ultrasound in the Assessment of Pathological Changes. Cairo, Egypt.
  15. 15. Richter J, Hatz C, Campagne G, Bergquist NR, Jenkins JM (2000) Ultrasound in Schistosomiasis: A practical guide to the standard use of ultrasonography for the assessment of schistosomiasis-related morbidity. Geneva: WHO.
  16. 16. Li YS, Sleigh AC, Li Y, Tanner M, Dessein A, et al. (2002) Five-year impact of repeated praziquantel treatment on subclinical morbidity due to Schistosoma japonicum in China. Trans R Soc Trop Med Hyg 96: 438–443.
  17. 17. Li YS, Sleigh AC, Ross AG, Li Y, Williams GM, et al. (2000) Two-year impact of praziquantel treatment for Schistosoma japonicum infection in China: re-infection, subclinical disease and fibrosis marker measurements. Trans R Soc Trop Med Hyg 94: 191–197.
  18. 18. Li YS, Kardorff R, Richter J, Sun KY, Zhou H, et al. (2004) Ultrasound organometry: the importance of body height adjusted normal ranges in assessing liver and spleen parameters among Chinese subjects with Schistosoma japonicum infection. Acta Trop 92: 133–138.
  19. 19. Hsiang MS, Carlton EJ, Zhang Y, Zhong B, Qiu D, et al. (2010) Use of Ultrasonography to Evaluate Schistosoma japonicum-related Morbidity in Children, Sichuan Province, China, 2000–2007. Am J Trop Med Hyg 82: 103–111.
  20. 20. Olds GR, Olveda R, Wu G, Wiest P, McGarvey S, et al. (1996) Immunity and morbidity in schistosomiasis japonicum infection. Am J Trop Med Hyg 55: 121–126.
  21. 21. Coutinho HM, Acosta LP, Wu HW, McGarvey ST, Su L, et al. (2007) Th2 cytokines are associated with persistent hepatic fibrosis in human Schistosoma japonicum infection. J Infect Dis 195: 288–295.
  22. 22. Wiest PM, Wu G, Zhong S, McGarvey ST, Yuan J, et al. (1994) Impact of annual screening and chemotherapy with praziquantel on schistosomiasis japonica on Jishan Island, People's Republic of China. Am J Trop Med Hyg 51: 162–169.
  23. 23. Spear RC, Seto E, Liang S, Birkner M, Hubbard A, et al. (2004) Factors influencing the transmission of Schistosoma japonicum in the mountains of Sichuan Province of China. Am J Trop Med Hyg 70: 48–56.
  24. 24. Zhou YC, Guo WX (2006) Ultrasonographic Medicine, 4th Edition [in Mandarin]. Beijing: Scientific and Technical Literature Publishing House.
  25. 25. Rabe-Hesketh S, Skrondal A (2005) Multilevel and Longitudinal Modeling Using Stata. College Station, TX: Stata Press.
  26. 26. The Office of Endemic Disease Control MoH (2000) Handbook of Schistosomiasis Control. Shanghai: Shanghai Science & Technology Press.
  27. 27. Katz N, Chaves A, Pellegrino J (1972) A simple device for quantitative stool thick-smear technique in Schistosomiasis mansoni. Rev Inst Med Trop Sao Paulo 14: 397–400.
  28. 28. Williams RL (2000) A note on robust variance estimation for cluster-correlated data. Biometrics 56: 645–646.
  29. 29. Long JS, Freese J (2006) Regression Models for Categorical Dependent Variables Using Stata, Second Edition. College Station, Texas: Stata Press.
  30. 30. Liang K, Zeger SL (1986) Longitudinal data analysis using generalized linear models. Biometrika 73: 13–22.
  31. 31. Huber PJ (1967) The behavior of maximum likelihood estimates under nonstandard conditions. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, CA: University of California Press. pp. 221–233.
  32. 32. White H (1980) A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48: 817–838.
  33. 33. Kleinbaum DG, Kupper LL, Morgenstern H (1982) Epidemiologic Research, Principles and quantitative methods. New York: John Wiley & Songs, Inc.
  34. 34. Wiest PM, Wu G, Zhong S, McGarvey ST, Tan E, et al. (1993) Schistosomiasis japonica on Jishan Island, Jiangxi Province, People's Republic of China: persistence of hepatic fibrosis after reduction of the prevalence of infection with age. Trans R Soc Trop Med Hyg 87: 290–294.
  35. 35. Coutinho HM, McGarvey ST, Acosta LP, Manalo DL, Langdon GC, et al. (2005) Nutritional status and serum cytokine profiles in children, adolescents, and young adults with Schistosoma japonicum-associated hepatic fibrosis, in Leyte, Philippines. J Infect Dis 192: 528–536.
  36. 36. Berhe N, Geitung JT, Medhin G, Gundersen SG (2006) Large scale evaluation of WHO's ultrasonographic staging system of schistosomal periportal fibrosis in Ethiopia. Trop Med Int Health 11: 1286–1294.
  37. 37. Carroll RJ, Ruppert D, Stefanski LA, Crainiceanu CM (2006) Measurement Error in Nonlinear Models: A Modern Perspective, Second Edition. Boca Raton: Chapman & Hall/CRC.
  38. 38. Li YS, He YK, Zeng QR, McManus DP (2003) Epidemiological and morbidity assessment of Schistosoma japonicum infection in a migrant fisherman community, the Dongting Lake region, China. Trans R Soc Trop Med Hyg 97: 177–181.
  39. 39. Ohmae H, Tanaka M, Hayashi M, Matsuzaki Y, Kurosaki Y, et al. (1992) Improvement of ultrasonographic and serologic changes in Schistosoma japonicum-infected patients after treatment with praziquantel. Am J Trop Med Hyg 46: 99–104.
  40. 40. Cai WM, Qiu DC, Hatz C (1992) Studies on ultrasonographic diagnosis of schistosomiasis japonica in China–a review of selected Chinese studies. Acta Trop 51: 37–43.
  41. 41. Berhe N, Myrvang B, Gundersen SG (2008) Reversibility of schistosomal periportal thickening/fibrosis after praziquantel therapy: a twenty-six month follow-up study in Ethiopia. Am J Trop Med Hyg 78: 228–234.
  42. 42. Martins-Leite P, Gazzinelli G, Alves-Oliveira LF, Gazzinelli A, Malaquias LC, et al. (2008) Effect of chemotherapy with praziquantel on the production of cytokines and morbidity associated with schistosomiasis mansoni. Antimicrob Agents Chemother 52: 2780–2786.
  43. 43. Chou YH, Chiou HJ, Tiu CM, Chiou SY, Lee SD, et al. (2003) Duplex Doppler ultrasound of hepatic Schistosomiasis japonica: a study of 47 patients. Am J Trop Med Hyg 68: 18–23.
  44. 44. Kresina TF, He Q, Degli Esposti S, Zern MA (1993) Hepatic fibrosis and gene expression changes induced by praziquantel treatment during immune modulation of Schistosoma japonicum infection. Parasitology 107 (Pt 4): 397–404.
  45. 45. Chigusa Y, Otake H, Ohmae H, Kirinoki M, Ilagan EJ, et al. (2006) Determination of the period for establishment of a liver network echogenic pattern in Schistosoma japonicum infection. Parasitol Int 55: 33–37.
  46. 46. Yu DB, Li Y, Sleigh AC, Yu XL, Li YS, et al. (2001) Efficacy of praziquantel against Schistosoma japonicum: field evaluation in an area with repeated chemotherapy compared with a newly identified endemic focus in Hunan, China. Trans R Soc Trop Med Hyg 95: 537–541.
  47. 47. Hatz CF (2001) The use of ultrasound in schistosomiasis. Adv Parasitol 48: 225–284.
  48. 48. King CH (2009) Toward the elimination of schistosomiasis. N Engl J Med 360: 106–109.
  49. 49. Ross AG, Sleigh AC, Li YS, Williams GM, Aligui GD, et al. (2000) Is there immunity to Schistosoma japonicum? Parasitol Today 16: 159–164.