Database search and selection criteria

The full study protocol was prospectively published on 26 February 2025 on the Open Science Framework website [19]. To ensure appropriate and transparent reporting, the PRISMA extension for Scoping Reviews (PRISMA-ScR) guidelines were followed (S1 File) [20]. A systematic search was performed on 3 March 2025 across five databases: PubMed (1946 – present), Embase (via Ovid) (1974 – present), Web of Science Core Collection (1964 – present), Global Index Medicus (1901 – present), and Global Health (via Ovid) (1973 – present). The search terms used (in the PubMed/MEDLINE format) were: (Schistosom* OR Bilharzia* OR snail* fever) AND (alcohol* OR drink*). This search string was adapted for each database, and full search strategies for all databases are provided in S2 File. Results were exported to Covidence [21], where duplicates were removed, and any missed duplicates from Covidence were removed manually.

Selection criteria

Two reviewers (BL and ML) independently screened all titles and abstracts for eligibility. Where results were eligible for full text screening, a full text of the study was obtained and all English full texts were screened by one reviewer (BL). A random 10% of the English full texts were screened by the second reviewer (ML). Disagreements at either of the screening stages were resolved through discussion between the two reviewers, and if unresolved then through a third reviewer (GFC). If a study was not in English but eligible for full text screening, a translation was sought. Reference lists of all the eligible full texts were searched for potentially eligible articles.

Studies conducting original research in any year up to the search date were eligible for inclusion. There were no restrictions on study participants or the species of schistosome. Included studies must have described both exposure to schistosome infection and alcohol use within the same individuals. Any description of alcohol use was included, including the frequency and type of alcohol consumed. Although this review was motivated by potential outcomes relating to liver disease, there was no restriction on the type or severity of clinical outcome studied to comprehensively capture all the reported associations for these exposures when they co-occur. Self-reported clinical outcomes and medical histories were included. To summarise effect sizes, studies were required to report statistical comparisons between alcohol use, schistosome infection, and clinical outcomes, or provide sufficient data to calculate these comparisons, for example through a 2 × 2 table. Studies were excluded if the case definition of the clinical outcome included schistosome infection or alcohol use as part of the diagnostic criteria, rather than treating them as independent exposures. Animal studies were not eligible, neither were ecological studies, case reports, reviews (systematic, narrative, meta-analyses), opinion articles or commentaries, or grey literature. Other study designs, including cross-sectional studies, cohorts, before-and-after studies, randomised controlled trials, and case-control studies were eligible. If a study was reported in multiple publications, the report with the most participants was considered if the two publications used the same study population, time period and geographical location. If the population, location, and number of participants was the same, the most recent publication was included.

Data extraction and qualitative synthesis

A data extraction form and data dictionary were adapted from Ockenden and colleagues [22], and piloted using Microsoft Excel to capture study characteristics and variables specific to answering the questions of the review. Three papers [23–25], which had been used to validate the search string, were used to pilot the extraction table, and any required changes were made before extracting from all the included studies. One reviewer (BL) extracted from all the included studies, and a second reviewer (ML) verified a random 10% of the extractions. The completed data extraction table is provided in S3 File. The extracted variables were related to study characteristics and study participants, exposures, outcomes, and statistical methods used. The study characteristics were author name(s), publication year, study year, country, study setting, locality, study inclusion and exclusion criteria, study aim, study design, time points, sample size, maximum and minimum age of participants, and sex of participants. Other extracted variables included schistosome species, schistosome diagnostic, method used to gather data on alcohol use, infection prevalence, alcohol use prevalence, co-occurrence prevalence, clinical outcome as defined by the study authors, statistical method used, measure of association reported, and, if possible, what relationship the authors suggested between co-occurring schistosome infection and alcohol use in relation to the clinical outcome. At the extraction stage, types of bias and confounding that were relevant to the study were considered, whether directly reported by the study authors or inferred in this analysis. Selection bias was identified in studies where the sample was not considered representative, for example from the use of non-random sampling methods or the recruitment of participants from settings like hospitals, where the sample would over-represent people who are already ill. Information bias was evaluated by assessing misclassification of the clinical outcome, such as when the outcome was not clearly defined or could be used inconsistently across participants. For each clinical outcome, outcome-specific confounders were identified both from the study reports and through assessment during data extraction; for example, hepatitis co-infection for liver fibrosis, or human immunodeficiency virus (HIV) viral load for renal impairment. For all studies, we also assessed whether common confounders were accounted for such as age and sex. The lack of consideration of confounding was particularly noted when studies reported unadjusted analyses.

The included studies were grouped according to their primary aims and fell into six broad categories. Risk factor studies examined associations between various exposures, including alcohol use and schistosome infection, and disease markers or clinical outcomes such as PPF and impaired renal function. Prevalence and disease burden studies measured the occurrence of schistosome infection and alcohol use, as well as the related morbidities in various populations. Some targeted specific groups, such as people living with HIV or tuberculosis, or residents of endemic communities. Diagnostic studies were classified based on their focus of assessing diagnostic accuracy or validation of diagnostic methods and tools related to disease outcomes. Treatment studies evaluated interventions, for example evaluating the impact of chemotherapy on morbidity. Clinical consequence studies focused on the manifestations and complications of schistosomiasis or alcohol-related morbidity, as well as how schistosome infection or alcohol use influences the presentation of comorbid conditions. Finally, biomarker studies investigated various biological markers, such as those related to iron status, by exploring the influence of dietary and infectious factors on disease outcomes.

In accordance with guidelines from the Joanna Briggs Institute guidance for scoping reviews [26], data were synthesised using a basic qualitative content analysis. Key themes synthesised included the types of clinical outcomes studied (and their appropriateness, as assessed by biological plausibility), the statistical methods used for analysis, and the overall quality (assessed through bias and confounding) of the study. A key focus of the synthesis was on the identification of clinical relationships and potential disease pathways described in the literature, with the aim of understanding how outcomes may be influenced through potential mechanistic pathways involving alcohol use and schistosome infection and the temporality between exposures and exposures and outcomes.

Screening process

A flowchart of the scoping review process is presented in Fig 1. The initial database search yielded 2358 hits, of which 541 were duplicates. The number of hits retrieved from each database is provided in the S2 File. Following screening, 21 studies were extracted and included in this review. The most common reasons for exclusion when screening full texts were the absence of data on schistosome infection or alcohol consumption, lack of clinical outcome (mostly where studies sought to report the prevalence of schistosome infection), and the use of an ineligible diagnostic method for schistosome infection. The latter category comprised of studies relying on self-reported infection status as well as studies which used proxy indicators of infection, for example classifying participants as infected based on the presence of fibrosis. The full list of excluded full texts, and their reasons for exclusion, is provided in S4 File.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flowchart showing the study selection process.

https://doi.org/10.1371/journal.pntd.0014371.g001

Study characteristics

A summary of the key characteristics of the studies included in the review are presented in Table 1. The most frequently studied schistosome species was S. mansoni (80.0%, 16/20), followed by S. japonicum (15.0%, 3/20). Only one study focused on S. haematobium, and one study failed to report the species of schistosome being studied. The most commonly studied alcohol exposure was a binary variable which compared any alcohol consumption to none (42.9%, 9/21), followed by seven studies (33.3%) that compared participants with high alcohol consumption to those with lower or no alcohol intake.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Key characteristics of the 21 studies included in the synthesis. Percentages represent the proportion of studies that reported each characteristic.

https://doi.org/10.1371/journal.pntd.0014371.t001

Nine countries across four WHO regions were represented. The most common study countries were Brazil (33.3%, 7/21), Tanzania (19.0%, 4/21), and Uganda (14.3%, 3/21). Most studies (95.2%, 20/21), made no restrictions on participant sex, and only one study was female-only. Only 13 studies reported an age range of participants, with reported ages spanning from 5 to 90 years across all the studies. Nine studies reported a mean age, and six reported a median age. The total sample size across all the included studies was 11,794, with a range of 54 – 2834. The median number of participants was 257 (IQR 100 – 655). Cross-sectional studies were the most common (66.7%, 14/21), although cohort (23.8%, 5/21), case-control (4.8%, 1/21), and RCT (4.8%, 1/21) designs also were represented.

The most commonly reported aims of the included studies were to quantify disease prevalence and burden (28.6%, 6/21), investigate risk factors associated with the clinical outcome of interest (23.8%, 5/21), and evaluate treatment methods (19.0%, 4/21). Less frequent aims were those related to the study of biomarkers (14.3%, 3/21), clinical consequences (9.5%, 2/21), and diagnostics (4.8%, 1/21). The most frequently reported clinical outcome across the studies was PPF, addressed in 42.9% (9/21) of the studies. Biomarkers were the second most reported outcome (19.0%, 4/21), encompassing a study investigating serum endotoxin levels, as well as studies concerned with markers of iron deficiency. Portal hypertension was the outcome of interest in only two studies. Another two studies examined susceptibility to coinfections, namely HIV and tuberculosis. Additional outcomes, each represented in a single study were cirrhosis, bladder pathology, impaired renal function, and cancer.

Varied definitions of alcohol use, schistosome infection, and diseases

Across all the studies, 76.2% (16/21) defined alcohol use as a binary variable, although they used varied reference categories. Of these, nine (56.3%, 9/16) studies compared any alcohol consumption to none (drinkers versus non-drinkers). The other seven studies (43.7%, 7/16) compared high or author-defined problem alcohol use to lower or no use, such as alcoholics versus non-alcoholics, or alcohol abuse versus no alcohol abuse. Four studies (19.0%) treated alcohol use as a categorical variable; three studies classified participants as never, ever, or regular drinkers, while the fourth study included an additional category for ex-drinkers. Only 17 (81.0%) of the studies reported information on their alcohol data collection method. Of these, 64.7% (11/17) used questionnaires completed by the participant themselves. Three studies (17.6%) conducted interviews with the study participant directly or with the household head, and in three studies (17.6%), the method was simply referred to as “self-report” without clarification as to how the data were collected. Four studies (19.0%) provided no information as to how alcohol use data were collected. Only one study clearly reported using a standardised tool, which was the WHO STEPwise Approach to Surveillance (WHO STEPS). Definitions of alcoholism, problem drinking, or regular drinking were inconsistent across studies, with most failing to clearly describe their criteria. For the five studies that did report definitions, thresholds showed large variation. One study defined alcoholism as consumption of more than 20g or 30g of alcohol per day for men and women, respectively, and another defined it as consumption of more than 60g per day for all participants. Other definitions were based on frequency, such as consuming alcohol three or more times per week. The timeframe for alcohol consumption – whether it referred to lifetime use, recent use, or another specified period – was often unclear or not reported. One study did not clearly define how they collected and coded alcohol consumption data; however, it reported a correlation coefficient with “lifetime alcohol consumption”.

For schistosome infection, most studies (85.6%, 18/21) assessed infection status using microscopy. Of these, 17 (94.4%) analysed stool samples using Kato-Katz techniques and one analysed urine samples using urine filtration. All of the Kato-Katz microscopy studies, as well as one study which used tissue biopsy, defined infection as the presence of any eggs in the sample. Two studies used antigen-based diagnostics – one applied the circulating cathodic antigen (CCA) urine dipstick test, and one used a lateral flow assay to detect circulating anodic antigen (CAA). No studies assessed infection using antibody-based tests. Nearly all studies (95.2%, 20/21) treated infection as a binary variable (infected versus uninfected). Although some studies reported measures of infection intensity, only one study incorporated a categorical classification into its analysis, defining infection intensity based on S. haematobium egg counts from urine filtration as: no infection (0 eggs/ 10 mL), light infection (≤50 eggs/ 10 mL), heavy infection (>50 eggs/ 10 mL). Only 10 (47.6%) of the studies using microscopy reported the number of samples used to diagnose schistosome infection. Of these, six reported using one sample, two used two samples, and another two used three samples. Where multiple samples were collected, one sample was taken per day over consecutive days.

A full summary of the definitions used in each study is presented in Table 2. Diagnostic methods and case definitions used to assess clinical outcomes varied across studies. For the outcome of PPF, all studies used abdominal ultrasonography as the diagnostic method. Only one study explicitly described using point-of-care ultrasound, although the other studies did note that imaging was performed in hospitals or health clinics by an experienced operator. The WHO Niamey protocol was the most commonly used ultrasound grading method (66.7%, 6/9). Among these studies, four classified liver image patterns C to F as indicative of PPF. The two other PPF studies used a severity grouping that combined diffuse general fibrosis with PPF, classifying liver image patterns A-D as no or mild fibrosis, and liver image patterns E or F as indicating severe fibrosis. Two studies were published before the Niamey protocol - one study used the Cairo protocol, and the other study applied the Homeida classification. One study, published in 2006, did not specify a grading system but defined PPF as moderate to severe wall thickening of the portal branches of the liver visible on ultrasound. Portal hypertension was identified through endoscopy based on the presence of the clinical complication of gastro-oesophageal varices. In studies assessing anaemia as the main outcome, anaemia was diagnosed using complete blood count testing on venous blood samples, with thresholds defined according to WHO guidelines as haemoglobin <120 g/L for women and <130 g/L for men (studies did not include children). Serum ferritin, treated as a continuous variable and measured in μg/L, was also measured through blood tests. Cirrhosis was diagnosed by liver biopsy and scored using the METAVIR system. In the study examining colorectal cancer, the diagnostic modality was not specified.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Summary of the definitions of schistosome infection, alcohol use, and outcome used in each study. Also included are the study population and age range (if reported), study exclusion criteria, number of participants, the statistical methods used in the analysis, and the association observed.

https://doi.org/10.1371/journal.pntd.0014371.t002

Associations of alcohol use and schistosome infection to disease outcomes

The associations between schistosome infection, alcohol use, and clinical outcomes were investigated across multiple disease domains, with considerable heterogeneity in how both exposures were defined. Statistical methods used and the associations observed are summarised in Table 2.

Hepatic outcomes were most frequently studied, with PPF being the most common outcome (9/21, 42.9%). The evidence was inconsistent and came from predominantly cross-sectional studies. Effect sizes ranged from negligible to strong for both exposures, with no consistent pattern. For example, one study reported higher odds of PPF with alcohol use but not schistosome infection [34], while another study found PPF severity to be associated with schistosome infection intensity but not alcohol use [24]. Other studies reported no association for either exposure [23,36,42,44]. A cohort study which evaluated treatment found that moderate alcohol consumption was associated with lower odds of PPF improvement following praziquantel, suggesting a potential exacerbating effect [30]. For the outcome of cirrhosis, both alcoholism and presence of PPF were found to be strong independent predictors [37].

Three studies (14.3%) investigated iron deficiency through markers including haemoglobin and serum ferritin. In these studies, alcohol use was consistently associated with higher haemoglobin and serum ferritin levels [27,29], whereas schistosome infection was linked to reduced ferritin in only one study [27], and showed no association in another [29].

Two studies (9.5%) examined infectious comorbidities. One reported that HIV acquisition was significantly associated with alcohol use but not schistosome infection [31]. A cohort study on tuberculosis found no association between alcohol use and S. mansoni-tuberculosis coinfection, though coinfection itself predicted higher mortality [35].

For the remaining studies, a distinct outcome was examined. These included bladder pathology (associated with schistosome infection intensity but not alcohol [28]), impaired renal function (associated with unhealthy drinking but not infection [25]), and colorectal cancer (associated with infection [40]). One study also found no significant effect of alcohol consumption on diagnostic accuracy of biomarkers for portal hypertension [41].

Temporal relationships between exposures and diseases

Temporality, both in terms of prospective associations between the exposures and outcomes, and in sequential relationships between schistosomiasis and alcohol use, was rarely studied, as most studies (66.7%, 14/21) used cross-sectional designs (Table 1). Even among the studies which were not cross-sectional in design, the frequency of alcohol use assessments was either unclear or limited to a single baseline assessment. This approach rendered the analysis of alcohol use again as cross-sectional.

Five (23.8%) studies reported information which could be used to understand temporality either for individual exposures or for the sequential order of the two exposures. Andrade et al. modelled cirrhosis as the outcome to understand the role of schistosome infection and its related pathology [37]. Their multivariate regression revealed that the presence of PPF was the strongest predictor of cirrhosis (odds ratio [OR] 4.56, 95% CI 2.10 – 9.91), and that alcoholism was also a significant predictor (OR 2.46, 95% CI 1.16 – 5.19). A separate calculation using descriptive data showed that current schistosome infection was a weaker predictor of cirrhosis (OR 2.28, 95% CI 1.46 – 3.55). This suggests that the long-term damage associated with chronic infection is a more important indicator for cirrhosis than current infection status. A study by Marinho et al. considered PPF as the outcome, and reported an OR of 3.54 (95% CI 1.4 – 8.8) comparing drinkers to non-drinkers [34]. When stratified by age, the OR among individuals aged 30 years and older increased to 9.28 (95% CI 2.60 – 33.06) compared to non-drinkers aged 19 and younger.

Concerning sequential effects of schistosomiasis and alcohol use, Anjorin et al. proposed a lagged effect in the correlation between alcohol use and PPF, hypothesising that alcohol consumption during adolescence may increase the likelihood of developing PPF in adulthood, whereas schistosome exposure occurs before alcohol use in early childhood [42]. However, this study did not present data to assess this association quantitatively. Further supporting the sequence of schistosome exposure first then alcohol use later in life, Li et al. demonstrated that moderate to heavy alcohol intake exacerbates PPF following schistosome infection [30]. This conclusion was drawn from observations that participants who reported moderate alcohol consumption were less likely to show an improvement in PPF following treatment than those who did not (OR 0.29, 95% CI 0.08 - 0.91). Additionally, apart from PPF, Wang et al. observed higher alcohol consumption among patients with schistosome-associated appendicitis, implying either a sequential relationship where infection precedes alcohol-related pathology, or confounding [40].

Statistical methods

Several statistical methods were used across the included studies, with some studies using more than one method. Seven studies (33.3%) presented only adjusted analyses, nine (42.9%) reported only unadjusted analyses, four (19.0%) presented both, and two (9.5%) studies provided sufficient descriptive data for odds ratios to be calculated using a 2 × 2 table. Regression analyses were used in 15 (71.4%) of the studies – 11 (73.3%, 11/15) employed logistic regression, three (20%, 3/15) used linear regression, and one used both. Other unadjusted methods included chi-squared tests, used in four studies, analysis of variance (ANOVA, one study) and receiver operating characteristic (ROC) curve analysis (one study). No studies included interaction terms. Two (9.5%) studies presented stratified analyses. Malenganisho et al. stratified by age and sex, and found that PPF was more prevalent among males compared to females within the same age groups, with this disparity becoming more pronounced with increasing age [24]. Marinho et al. stratified their analysis by age and alcohol consumption, finding that drinkers aged 20–29 had 2.5 times the odds of PPF compared to non-drinkers under 19 (95% CI 0.48 – 12.88), while drinkers aged 30 and older had 9.28 times the odds of PPF compared to the same reference group (95% CI 2.60 – 33.06) [34].

The most frequently included covariates in adjusted regression models were age (10/11), sex (7/10 as one study was female only), smoking (4/11), HIV status (3/11), and occupation (3/11). Table 3 presents the associations between common confounders and the two most common study outcomes – PPF and iron deficiency – stratified by schistosome infection diagnostic method and adjustment status. This stratification highlights that the evidence is mostly from studies using Kato-Katz microscopy, limiting comparisons across diagnostic techniques.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Direction of associations between common confounders and study outcomes. Associations are stratified by schistosome infection diagnostic modality and by whether the analyses were adjusted or unadjusted.

https://doi.org/10.1371/journal.pntd.0014371.t003

Age and sex (male versus female) were consistent covariates, with higher age and male sex frequently associated with increased odds of PPF. The influence of smoking and HIV status is less clear as they were reported in fewer studies. The variability in the reported associations shows how the unadjusted estimates for key exposures may be susceptible to confounding. Only four studies in this synthesis presented adjusted and unadjusted analyses for various outcomes, allowing for the assessment of confounding. Anjorin et al. observed that alcohol use was associated with increased odds of PPF in the unadjusted analysis (OR 3.63, 95% CI 2.66 – 4.91), but this association was not present in the adjusted model (OR 1.09, 95% CI 0.70 – 1.68), indicating confounding by factors like age and sex [42]. In contrast, Ssetaala et al. found that alcohol use remained associated with higher odds of HIV acquisition even after adjustment [31]. The OR for regular alcohol use compared to never drinkers increased from 3.49 (95% CI 1.6 – 7.8) in the unadjusted model to 6.85 (95% CI 1.3 – 37.4) in the adjusted model, and the OR for rare alcohol use rose from 1.92 (95% CI 0.8 – 4.8) to 9.21 (95% CI 1.4 – 61.8) - perhaps all signifying positive confounding. Baya et al. reported that none of age, gender, smoking, or alcohol use were associated with schistosome-tuberculosis coinfection in either the unadjusted or adjusted analyses [35]. Finally, Silva et al. found no significant associations between alcohol use or age and the odds of severe PPF in either adjusted or unadjusted models [32]. Although sex was not included in their adjusted analysis, they reported an unadjusted OR comparing males to females of 2.85 (95% CI 1.0 – 8.28).

Exposure interactions and biological mechanisms

Potential mechanisms linking alcohol use and schistosome infection to the clinical outcomes of interest were infrequently discussed by study authors or were not clearly specified in the model assumptions and motivation. For the few studies which discussed mechanisms, alcohol use and infection were presented individually rather than as co-occurring and potentially interacting exposures. Andrade et al. suggested that both schistosome infection and alcoholism contribute independently to liver cirrhosis progression, which may include synergistic effects of concurrent injuries, though interaction effects were not tested [37]. Similarly, Li et al. directly interpreted their findings as evidence that alcohol consumption “exacerbates schistosome-induced parenchymal morbidity”, also suggesting a potential synergistic relationship, though again not formally tested [30]. Malenganisho et al. reported a higher proportion of PPF among alcohol drinkers compared to non-drinkers [24]. Although the difference was not statistically significant, they posited that alcohol may be a contributing factor to schistosome-induced liver morbidities and should be studied further. They also reported qualitatively that excessive alcohol consumption was predominantly observed in men, who also exhibited higher rates of PPF, suggesting a gender-dependent relationship.

Biological mechanisms were typically discussed for the exposures separately. Friis et al. discussed how alcohol consumption may enhance iron absorption due to the iron content of alcoholic drinks, or through increased iron solubilisation [27]. In contrast, schistosome infection was linked to intestinal bleeding, resulting in iron loss and reduced serum ferritin. Similarly, Malenganisho et al. found that alcohol use was positively associated with serum ferritin, likely reflecting increased iron absorption, while schistosome infection was linked to lower ferritin levels due to internal bleeding caused by infection [29].

Bias and confounding

Reporting of bias and confounding varied across the studies. The most frequent source of bias reported by study authors was selection bias (28.6%, 6/21), which arose from either non-random sampling methods or non-random loss to follow-up. Several studies that did not explicitly mention selection bias also used sampling methods that could limit generalisability. For example, multiple studies recruited participants from hospitals or clinical settings, potentially causing people who were already unwell and seeking healthcare to be over-represented. In contrast, one study excluded people classified as drunkards. Survivor bias was reported in two studies. Misclassification bias was discussed in one study, where variability in egg output throughout the day was noted as a potential cause of misclassifying light infections as negative. Measurement error may have been present in studies which did not explicitly define their outcome measures, as variability in the definitions used by study staff could lead to inconsistencies and potential misclassification. However, it is also possible that standardised definitions were used and not reported. Recall bias was explicitly acknowledged in another two of the studies; however, all studies relied on self-reported alcohol consumption data, collected through interviews, questionnaires, or in a few cases, described simply as self-reported. The use of self-reported data introduces a potential for recall bias across all the studies, whether or not it was acknowledged by the study authors. The extent of social desirability bias may vary geographically, as the stigma surrounding alcohol use varies by region. In study settings where alcohol use is more stigmatised, under-reporting may be more likely, which could lead to smaller effect sizes being observed. This underreporting may be evident in one study of individuals aged five years and older in rural Uganda where the household head and spouse were asked to report alcohol use on behalf of all household members.

Study designs varied in representativeness of relevant populations for shared disease outcomes. The included studies covered a wide age range, although some explicitly excluded children. Several studies also lacked clear age reporting, limiting the assessment of age representativeness. Study sizes were generally small, with only five (23.8%) including more than 500 participants. Some schistosomiasis-endemic countries such as Brazil, Tanzania, and Uganda were well-represented, however, others were underrepresented, such as China which only had two studies.

Additional confounders were rarely commented on by study authors, potentially due to the majority of analyses already being adjusted for key variables. The most commonly identified confounders by study authors were socioeconomic status (9.5%, 2/21) and infections (19.0%, 4/21), particularly HIV and hepatitis, or referred to more broadly as just “infections”. An assessment of confounding was possible in the four studies mentioned previously that presented both unadjusted and adjusted analyses. The results from these studies demonstrate that, for the studies presenting only unadjusted analyses, confounding by factors such as age, sex, and socioeconomic status may be impacting the observed effect sizes. Other potential confounders, including occupation, water contact behaviour, and smoking status, were identified through the data extraction where confounding was not reported on by study authors, depending on the population and clinical outcome being studied.

Disease outcomes and hypothesised pathways

The relationship between alcohol use and schistosome infection has been investigated primarily within the context of hepatic morbidity, namely PPF, since both exposures are well-established aetiological agents of liver disease. Schistosome infection was reported to be the main driver of PPF, with alcohol use presented as a significant effect modifier which could exacerbate pathology in certain contexts [34,37]. However, there is a need for future studies to more clearly define chronic or heavy alcohol use. The dominance of schistosome-specific pathologies in the literature means that outcomes associated with alcohol use may have been missed. For instance, hepatocellular carcinoma is considered one of the final manifestations of ALD, and infection with S. japonicum can also independently cause hepatocellular carcinoma [46]. Similarly, steatosis (fatty liver) is a common and reversible early-stage manifestation of ALD, but it is not causally attributable to schistosome infection alone [47]. It therefore remains unclear whether schistosome infection modifies the risk or progression of alcohol-specific outcomes.

A synergistic interaction between schistosome infection and alcohol consumption, leading to accelerated liver fibrosis, is mechanistically plausible. Egg deposition during schistosome infection initially causes focal injury through granuloma formation and inflammation in the portal veins, and then alcohol metabolism induces diffuse hepatocellular damage and oxidative stress [48–51]. In co-exposed individuals, the toxicity of alcohol metabolism in hepatocytes near and around the portal vein could exacerbate schistosome-induced damage and cause more severe manifestations of PPF. Findings from murine models also provide evidence that supports this synergistic pathway. Experimental co-exposure in mice has resulted in more severe liver pathology, including larger granulomas and increased fibrosis, than would be expected from each exposure alone [16,17].

There is also experimental evidence to support an antagonistic pathway between schistosome infection and alcohol use through immune modulation via alcohol-induced immune suppression. Chronic alcohol intake has been reported to have significantly reduced the size of schistosome-induced granulomas in mice [52]. The authors of this study suggest that the impairment of cell-mediated immunity due to alcohol use dampens the host inflammatory response to schistosome eggs. An antagonistic pathway is also possible in the development of steatosis, although this was not an outcome explored in any of the included studies. There is preliminary mechanistic evidence from mice that schistosomes rely on host high-density lipoproteins for egg development and continuation of the parasite life cycle [53]. Experimental data from mice also found reduced steatosis in schistosome-infected mice, even following exposure to alcohol [16]. In human studies, schistosome infection has been associated with dyslipidaemia, reflecting disturbances in host lipid metabolism and transport [54,55]. Therefore, an antagonistic interaction between alcohol use and schistosome infection mediated through regulation of hepatic lipid metabolism is plausible. Another antagonistic relationship between schistosome infection and alcohol use may be possible for iron-related outcomes. A potential pathway for this is through regulation of hepcidin – a hormone that is primarily produced in the liver and plays a key role in the regulation of blood iron levels [56]. Alcohol consumption typically suppresses hepcidin synthesis, which increases iron absorption into the blood, whereas the immune response to schistosome infection has been shown to increase hepcidin synthesis, resulting in iron absorption becoming more tightly regulated [57,58]. This pathway may explain the differences in ferritin patterns observed in the included studies [27,29].

Although both synergistic and antagonistic relationships have support from murine studies, their translational relevance to human disease is limited. Firstly, experimental models typically administer the exposures in close temporal proximity in adult animals [16,17,52], which is not an accurate reflection of exposure in humans – schistosome infection typically begins in early childhood, before alcohol use is initiated later in life. Furthermore, the controlled dose of each exposure in mice does not reflect the heterogeneity in timing and dose that occurs in human populations, for example when considering behaviours such as binge drinking. Therefore, murine models alone would not be able to resolve which kind of interaction prevails for humans. Instead, prospective human studies over much longer periods of time, such as those underway in SchistoTrack [59], are needed.

Temporality

The current literature is majorly limited in its ability to establish temporal relationships between schistosome infection, alcohol use, and clinical outcomes. The predominance of cross-sectional studies, alongside the common assessment of alcohol use at a single timepoint, assumes that exposures and outcomes are both static and concurrent. This approach fails to capture possible changes in exposure intensity, duration, and sequence that can occur over the life-course and alter their interactions. Additionally, the reliance on binary measures of current schistosome infection limits the ability to assess the impact of chronic and cumulative infection on outcomes such as cirrhosis. However, the review did identify some studies which can be used to comment on temporality and identify future directions. The stratified analysis by Marinho et al., which found greatly higher odds of PPF among older drinkers, could reflect increased vulnerability to liver damage as enzyme activity decreases with age [60]. Another possible explanation is that drinking patterns change over the life of an individual – older people might drink more than younger people in the study population. This finding aligns with the observations reported by Li et al., which suggests that alcohol consumption can alter the trajectory of disease progression and response to treatment [30]. Together, these observations highlight the necessity for longitudinal studies which are designed to elucidate the complexity of changing exposure dynamics over time.

A thorough understanding of temporality requires a consideration of the life-course patterns of both schistosome infection and alcohol use. Schistosome infection prevalence and intensity peaks among young children, however the severe pathology associated with chronic infections, such as PPF, tends to manifest in adulthood [61]. Previous work by Ewuzie et al. reported that current schistosome infection is only weakly linked to PPF, and may not be the best proxy indicator of cumulative infection burden or morbidity risk [12]. Similarly, it has been found that current water contact, the key exposure for infection, is poorly correlated with infection status [62], which reflects the complex immune response to infection and the resistance to reinfection that can build over time. These findings highlight how morbidity results from the cumulative, chronic nature of infection rather than from a single point of current exposure.

Alcohol consumption also follows its own temporal trajectory. While initiation of alcohol use tends to occur during adolescence [5,63], patterns of hazardous drinking and alcohol use disorders frequently peak in middle age [64,65], years or even decades after initial schistosome infection was established. This staggered exposure timing may suggest that alcohol-induced liver injury would develop years after schistosome infection, and alcohol use may compound already established schistosome-induced liver damage. Given the timing of schistosome exposure and alcohol use in later life, future studies should consider monitoring populations starting relatively young and first tracking schistosome infection history, then the development of alcohol use behaviours to understand shared disease outcomes.

Limitations

There are some limitations to our study. As with all scoping reviews, the findings are constrained by the quality and completeness of reporting in the included studies. In particular, reporting of alcohol use varied markedly in quality, although alcohol use was often not the primary focus of the study, so this may be an unfair expectation. Future studies should employ standardised, clinically meaningful definitions of alcohol use. Previous research has suggested that categorisations of alcohol use should be based on biological consequence in relation to the disease outcome – for example, combining lifetime abstainers with very low-level alcohol users, rather than a broad drinker versus non-drinker divide, which risks diluting the effect estimates for higher alcohol use groups [10]. A further limitation relates to the interpretation of joint effects between alcohol use and schistosome infection. None of the studies reported formally testing interaction terms – instead, alcohol use and schistosome infection were measured independently using exposure classifications that are subject to measurement error. In the context of joint exposure assessment, correlated misclassification, arising from underreporting of alcohol consumption or false-negative schistosome infection classification when single-sample microscopy is used, may have resulted in individuals being misclassified into lower exposure categories for both factors. Consequently, some observed associations may reflect measurement error rather than true joint biological effects. Future studies could mitigate these limitations by using repeated measures of alcohol consumption in a subsample of participants, and by applying more reliable schistosomiasis exposure assessment strategies, such as repeated stool sampling or the use of more sensitive antigen-based diagnostics. Additionally, as the majority of the studies examined S. mansoni, there are not enough studies involving other schistosome species to allow for an assessment of species-specific effects on clinical outcomes. Selection bias was also evident, with several studies recruiting from clinical settings and one study excluding heavy drinkers, resulting in study populations that are not likely to be representative. As mentioned previously, the analytical approaches used in the included studies also limit the review. Only four studies reported both adjusted and unadjusted analyses which allowed for the assessment of confounding. This suggests that confounding will impact the associations observed in other studies, especially those reporting only unadjusted analyses.