Systemic inflammation and malaria co-infection in irritable bowel syndrome: a cross-sectional study of 142 Yemeni patients

Abstract

Objective

To determine the prevalence of malaria co-infection among patients with irritable bowel syndrome (IBS) in Yemen and to evaluate the association of systemic inflammatory biomarkers (neutrophil-to-lymphocyte ratio (NLR), mean platelet volume (MPV) and platelet-to-lymphocyte ratio (PLR)) with this co-infection.

Design

Multicentre, cross-sectional observational study conducted between April and December 2024.

Setting

Primary and secondary healthcare facilities across 21 governorates in Yemen.

Participants

142 consecutive adult patients (aged 18–70 years) diagnosed with IBS according to the Rome IV criteria.

Outcome measures

The primary outcome was the prevalence of malaria infection, confirmed by a rapid diagnostic test . Secondary outcomes included differences in NLR, MPV and PLR between groups, assessed using independent t-tests, and the diagnostic performance of these biomarkers evaluated by receiver operating characteristic curve analysis with AUC calculation. Multivariate binary logistic regression was used to identify independent predictors of malaria co-infection, adjusting for potential confounders.

Results

The mean age was 42.3 years (SD 11.7) with an equal gender distribution. The prevalence of malaria co-infection was 45.1% (64/142). Patients with malaria positivity had significantly higher NLR (mean difference 0.56, 95% CI 0.40 to 0.72; p<0.001) and PLR (mean difference 22.6, 95% CI 17.8 to 27.4; p<0.001) and lower MPV (mean difference −0.60 fL, 95% CI −0.77 to −0.43; p<0.001) compared with malaria-negative patients. In adjusted logistic regression, higher NLR (adjusted OR 3.21, 95% CI 1.89 to 5.45) and PLR (adjusted OR per 10-unit increase 4.02, 95% CI 2.24 to 7.18) were independently associated with malaria positivity.

Conclusions

Nearly half of the Yemeni patients with IBS in this study had malaria co-infection, with the highest burden in diarrhoea-predominant and mixed subtypes. Elevated NLR and PLR were strongly associated with co-infection, suggesting these readily available biomarkers could aid targeted screening in resource-limited, endemic settings.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This is the first multicentre cross-sectional study investigating malaria co-infection and associated inflammatory biomarkers among patients with irritable bowel syndrome (IBS) in a conflict-affected setting.

• Standardised diagnostic criteria (Rome IV for IBS) and WHO-endorsed rapid diagnostic tests, with microscopy confirmation for malaria, were strictly adhered to.

• The study design included participants from 21 governorates, enhancing geographical representativeness.

• The cross-sectional design limits causal inference regarding the relationship between IBS and malaria.

• The absence of comparator groups (healthy individuals and patients with malaria only) restricts the ability to attribute the observed inflammatory profile specifically to the co-infection state.

Introduction

Irritable bowel syndrome (IBS) represents one of the most prevalent functional gastrointestinal disorders worldwide, characterised by recurrent abdominal pain associated with altered bowel habits in the absence of organic pathology.¹ With a global prevalence estimated at 4.1%, IBS imposes a substantial burden on healthcare systems and significantly impairs patients’ quality of life.² The pathophysiology of IBS remains incompletely understood, though emerging evidence suggests a complex interplay between gut-brain axis dysfunction, visceral hypersensitivity, altered gut motility, low-grade mucosal inflammation and intestinal dysbiosis.^{3 4} Recent investigations have highlighted the potential role of systemic inflammatory processes in IBS symptomatology, as evidenced by alterations in haematological indices such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) and mean platelet volume (MPV).^{5 6}

Concurrently, malaria remains a formidable global health challenge, particularly in endemic regions such as sub-Saharan Africa and parts of Asia.⁷ Yemen, classified as a high-burden malaria country by the WHO, continues to face significant challenges in malaria control, exacerbated by ongoing conflict, humanitarian crises and healthcare system fragmentation.⁸ ‘Plasmodium falciparum’ and ‘P. vivax’ are the predominant species, with transmission occurring year-round in many governorates.⁹ Malaria infection triggers a profound systemic inflammatory response characterised by cytokine dysregulation, endothelial activation and haematological alterations, including thrombocytopenia, leucocyte redistribution and changes in platelet indices.^{10 11}

The intersection between chronic inflammatory conditions like IBS and acute infectious diseases like malaria represents an understudied yet clinically relevant area of investigation. Several pathophysiological mechanisms may connect these conditions: (1) malaria-induced systemic inflammation could exacerbate IBS symptoms through neuroimmune activation; (2) pre-existing gut barrier dysfunction in patients with IBS might increase susceptibility to parasitic infections; (3) shared inflammatory pathways involving cytokines such as tumour necrosis factor (TNF)-α, interleukin (IL)-6 and IL-8 could modulate both conditions.^{12 13} Furthermore, haematological inflammatory indices—readily available from routine complete blood counts—have emerged as potential biomarkers for various inflammatory and infectious conditions.^{14 15} However, their utility in characterising malaria co-infection in patients with IBS remains unexplored.

Despite the individual significance of IBS and malaria as public health concerns in Yemen, no previous study has investigated their potential association or the role of systemic inflammatory biomarkers in this context. This knowledge gap is particularly critical given the dual burden of communicable and non-communicable diseases in conflict-affected settings. Addressing this gap could inform integrated clinical management strategies and optimise resource allocation where both conditions are prevalent. Therefore, this study aimed to (1) determine the prevalence of malaria co-infection among Yemeni patients with IBS; (2) evaluate the association between IBS subtypes and malaria infection; (3) characterise systemic inflammatory biomarker profiles (NLR, MPV, PLR) in patients with IBS with and without malaria; and (4) assess the diagnostic and predictive utility of these biomarkers for malaria detection in this population.

Methods

Study design and setting

Design and reporting

This is a multicentre, cross-sectional observational study. This report adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional studies. A completed STROBE checklist is provided as a supplementary file (online supplemental file S1).

We conducted this study across 15 primary and secondary healthcare facilities in Yemen between April 2024 and December 2024. This investigation represents an initial clinical study within a broader, structured research programme aimed at evaluating systemic inflammatory biomarkers in gastrointestinal disorders. Yemen represents a unique epidemiological setting characterised by a high burden of both communicable diseases, such as malaria, and functional gastrointestinal disorders, exacerbated by ongoing humanitarian challenges.

Study population and eligibility criteria

Consecutive adult patients (aged 18–70 years) presenting with chronic gastrointestinal symptoms to participating gastroenterology and internal medicine clinics were screened for eligibility. The inclusion criteria were (1) a confirmed diagnosis of IBS according to the Rome IV diagnostic criteria, established by a certified gastroenterologist; (2) residence in Yemen for at least the preceding 5 years. The exclusion criteria were (1) the presence of organic gastrointestinal disease (eg, inflammatory bowel disease, coeliac disease confirmed by serology/biopsy or colorectal cancer); (2) the use of antibiotics.

Immunosuppressants or anticoagulants within the past 4 weeks; (3) pregnancy or lactation; (4) known haematological disorders (eg, thalassemia, leukaemia); (5) acute febrile illness at the time of recruitment not attributed to malaria. A total of 142 eligible patients were enrolled. The data collection form used for all participants is available in online supplemental document S2, and participant information sheets are provided in online supplemental document S3.

Sample size considerations and power analysis

The target sample size was determined a priori using G*Power 3.1 software. Based on preliminary estimates suggesting a 30–40% prevalence of malaria in patients with chronic diseases in Yemen,^{8 16} and expecting a medium effect size (Cohen’s d=0.5) for biomarker differences between groups, with α=0.05 and β=0.20 (80% power), a minimum of 128 participants was required. Accounting for potential incomplete data, we aimed to recruit 140 participants. Our final sample of 142 patients provides a margin of error of ±8% for estimating the prevalence proportion, based on an assumed 50% prevalence and a 95% confidence level. This margin of error was considered acceptable for this study in a conflict-affected setting with limited prior data.

A post-hoc power analysis using the observed effect sizes from our study data confirmed adequate power (82–87%) for our primary comparisons. We acknowledge that post-hoc power calculations based on observed data are subject to circularity bias and should be interpreted with caution. These calculations are presented for descriptive purposes only and do not substitute for the a priori power analysis that determined the sample size. Detailed power analysis results are provided in online supplemental table S10, and sample size calculation details are available in online supplemental method S1.

Data collection and variables

Data were collected using a standardised, piloted questionnaire administered by trained research assistants through face-to-face interviews. We acknowledge the potential for self-reporting and recall bias inherent in face-to-face interviews, particularly for historical information such as symptom duration, previous malaria episodes and medication use. To minimise this bias, we used standardised questionnaires with clear, simple language, trained research assistants to probe consistently without leading questions and corroborated self-reported data with medical records where available. However, some degree of recall bias cannot be completely eliminated. The questionnaire captured the following domains:

Sociodemographic and clinical data

Age, gender, governorate of residence, detailed medical history and current symptoms.

IBS subtype classification

Patients were categorised into four subtypes based on the predominant bowel habit according to the Rome IV criteria: IBS with constipation (IBS-C), IBS with diarrhoea (IBS-D), IBS with mixed bowel habits (IBS-M) and unclassified IBS (IBS-U).

The data collection form used for all participants is available in online supplemental document S2, and participant information sheets are provided in online supplemental document S3.

Research staff training and quality assurance

All research assistants and laboratory technicians underwent a standardised 2-day training programme prior to the commencement of the study. The training covered (1) ethical conduct and obtaining informed consent, (2) administration of the clinical questionnaire and application of the Rome IV criteria, (3) standardised procedures for venous blood collection and handling, (4) performing and interpreting malaria rapid diagnostic test (RDTs) according to the manufacturer’s instructions and (5) principles of good clinical and laboratory practice. Competency was assessed via supervised practice and a written evaluation. Regular refresher sessions were held bi-monthly to ensure protocol adherence throughout the study period.

Assessment of IBS according to rome iv criteria

The diagnosis of IBS and subtype classification was established through a structured clinical assessment administered by certified gastroenterologists with specific training in functional gastrointestinal disorders. The assessment included:

1. A standardised Arabic-translated Rome IV diagnostic questionnaire covering:

   • Frequency and nature of abdominal pain (≥ 1 day/week in the last 3 months).

   • Association of pain with defecation.

   • Changes in stool frequency and consistency.

   • Symptom duration (≥ 6 months with onset ≥3 months prior).

2. Bristol Stool Form Scale assessment for subtype classification:

   • IBS-C: > 25% Bristol types 1–2 and<25% types 6–7.

   • IBS-D: > 25% Bristol types 6–7 and<25% types 1–2.

   • IBS-M: > 25% both types 1–2 and types 6–7.

   • IBS-U: insufficient stool consistency abnormality.

All gastroenterologists underwent a standardised training programme on the application of the Rome IV criteria, which included:

• A 4-hour workshop on the interpretation of the Rome IV criteria.

• An assessment of 10 standardised case vignettes (achieving >90% diagnostic concordance).

• Monthly calibration sessions to maintain diagnostic consistency.

Details of the validated instruments used, including the Rome IV criteria and RDT specifications, are provided in online supplemental resource S1.

This structured approach ensured the uniform application of diagnostic criteria across all study sites.

Laboratory analysis and biomarker calculation

Malaria diagnosis

Malaria infection was diagnosed using a WHO-prequalified RDT (SD BIOLINE Malaria Ag Pf/Pan, Abbott). The test detects Plasmodium falciparum-specific histidine-rich protein 2 and pan-malarial Plasmodium lactate dehydrogenase. The procedure was performed according to the manufacturer’s instructions: 5 µL of whole blood was added to the sample well, followed by two drops of assay buffer. Results were read at 20 min and interpreted as follows: (1) positive: presence of both the control line and test line(s); (2) negative: presence of the control line only; (3) invalid: absence of the control line. All positive RDT results were confirmed by microscopic examination of Giemsa-stained thick and thin blood films. Discordant results were resolved by a third senior microscopist. Quality control procedures included daily monitoring of test kit storage temperature (2–30°C) and verification of lot numbers and expiry dates.

Complete blood count analysis

Complete blood count (CBC) analysis was performed within 2 hours of sample collection using a fully automated, calibrated haematology analyser (Sysmex XN-1000, Sysmex Corporation). The analyser provided direct measurements of absolute neutrophil count (ANC), absolute lymphocyte count (ALC), platelet count (PLT) and MPV. The following systemic inflammatory biomarkers were calculated:

• NLR: Calculated as ANC divided by ALC.

• PLR: Calculated as PLT divided by ALC.

The mathematical derivations and normal ranges for these biomarkers are detailed in online supplemental resource S4. All laboratory procedures followed strict quality control protocols, including daily calibration and the use of internal controls. Complete laboratory protocols are detailed in online supplemental document S4. Laboratory procedures were conducted according to standardised operating protocols with daily quality control checks. A detailed laboratory QC checklist is provided in the online supplemental materials.

Statistical analysis

Data were analysed using SPSS software V.28.0 (IBM Corp) and R V.4.2.1 (R Foundation for Statistical Computing). All statistical tests were two-tailed, with a p value of <0.05 considered statistically significant. Graphical representations, including box plots, scatter plots and receiver operating characteristic (ROC) curves, were generated to visualise key findings.

Descriptive statistics and normality assessment

Descriptive statistics were presented as means±SD for normally distributed continuous variables, medians with IQRs for non-normally distributed variables, and frequencies with percentages for categorical variables. The normality of data distribution was assessed using the Shapiro-Wilk test. Given our moderate sample size (n=142), we employed robust statistical approaches to ensure validity: (1) non-parametric tests (Mann-Whitney U, Kruskal-Wallis) were used when normality assumptions were questionable; (2) exact tests (Fisher’s exact) were applied for categorical variables with small expected cell counts (< 5); (3) effect sizes (Cohen’s d for continuous variables, φ coefficient for categorical variables) were reported alongside p values to provide magnitude estimates; and (4) bootstrapping with 1000 iterations was used to generate CIs for regression coefficients, enhancing robustness to distributional assumptions.

Missing data

There were no missing data for the primary outcome variables (malaria status, NLR, MPV, PLR) or for key demographic variables (age, gender, IBS subtype). Three participants (2.1%) had incomplete data for education level and were excluded from analyses involving this variable. A complete-case analysis was used for all primary comparisons, and sensitivity analyses with multiple imputation (m=5) confirmed that the exclusion of these cases did not alter the findings (online supplemental table S8).

Comparative analyses

To compare baseline characteristics and biomarker levels between groups, we employed independent samples t-tests (for normally distributed continuous variables) or Mann-Whitney U tests (for non-normally distributed variables). For categorical variables, χ² tests were used, with Fisher’s exact test applied when appropriate. A one-way analysis of variance (ANOVA) with a post-hoc Tukey test was used for comparisons across the four IBS subtypes. The Bonferroni correction was applied to adjust for multiple comparisons when conducting multiple statistical tests on related outcomes.

Correlation and diagnostic performance analysis

Pearson’s correlation coefficients were calculated for normally distributed continuous variables, while Spearman’s rank correlation coefficients were used for non-normally distributed variables to examine bivariate relationships between inflammatory biomarkers and other continuous measures. ROC curve analysis was performed to evaluate the diagnostic performance of NLR, MPV and PLR in identifying malaria co-infection. Area under the curve (AUC) values were calculated with 95% CIs, and optimal cut-off values were determined using the Youden index (J=sensitivity + specificity − 1). Statistical comparison of ROC curves was performed using the DeLong test.

Multivariate modelling

To identify independent predictors of malaria co-infection, variables demonstrating associations with a p value <0.10 in univariate analyses were entered into a multivariate binary logistic regression model using forward stepwise selection. The multivariate logistic regression model adjusted for the following covariates: IBS subtype (dichotomised as IBS-D/M vs IBS-C/U), NLR (per 1-unit increase), MPV (per 1 fL increase) and PLR (per 10-unit increase). These variables were selected based on their association with malaria status in univariate analyses (p<0.10). The model did not include age or gender, as they showed no significant association in univariate analyses, and their inclusion did not improve model fit (likelihood ratio test, p>0.05). Model results were expressed as adjusted odds ratios (aOR) with 95% CIs. The goodness-of-fit of the final model was assessed using the Hosmer-Lemeshow test, with p>0.05 indicating an adequate fit. Multicollinearity was evaluated using variance inflation factors, with values<5 considered acceptable. The discriminative ability of the model was quantified using the Nagelkerke R² statistic.

The complete statistical analysis code (R, SPSS and Python) is available in onlinesupplemental code S1,S2 S3undefined, respectively. Comprehensive quality control procedures are detailed in online supplemental method S2.

Data management

Data management procedures followed standardised checklists for data collection, entry, validation, storage and sharing. These checklists ensured data integrity, security and reproducibility throughout the study (available in the online supplemental material). The detailed data management plan, including procedures for collection, entry, validation, storage and sharing, is available in online supplemental method S3.

Quality assurance

Multiple quality assurance checklists were implemented covering: (1) participant recruitment and enrolment, (2) data collection procedures, (3) laboratory sample processing, (4) statistical analysis and (5) manuscript preparation. These checklists are provided in the online supplemental materials.

Patient and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of this research. This study was conducted in a conflict-affected setting with limited infrastructure for patient and public engagement in research design. However, the research questions were developed based on clinical observations of the dual burden of IBS and malaria in Yemeni patients, and the findings will be disseminated to local health authorities and participating communities through summary reports and community meetings where feasible.

Results

Demographic and clinical characteristics of the study cohort

A total of 142 patients diagnosed with IBS were enrolled in this study, constituting our final analytical cohort. The demographic and baseline clinical characteristics are summarised in table 1. Comprehensive demographic and clinical characteristics stratified by malaria status are provided in online supplemental table S1. The study flow diagram is shown in online supplemental figure S1.

Table 1. Baseline characteristics of the study participants (n=142).

Characteristic	Total (n=142)	Malaria positive (n=64)	Malaria negative (n=78)	P value	Test statistic
Age (years), mean±SD	42.3±11.7	43.1±10.9	41.7±12.3	0.451	t=0.76
Age group, n (%)				0.872	χ²=1.23
25–35	38 (26.8)	16 (25.0)	22 (28.2)
35–45	39 (27.5)	19 (29.7)	20 (25.6)
45–55	38 (26.8)	18 (28.1)	20 (25.6)
>55	27 (19.0)	11 (17.2)	16 (20.5)
Gender, n (%)				1.000	χ²=0.00
Male	71 (50.0)	32 (50.0)	39 (50.0)
Female	71 (50.0)	32 (50.0)	39 (50.0)
Residence, n (%)				0.235	χ²=1.41
Urban	85 (59.9)	35 (54.7)	50 (64.1)
Rural	57 (40.1)	29 (45.3)	28 (35.9)
Education, n (%)				0.189	χ²=4.78
Illiterate	22 (15.5)	12 (18.8)	10 (12.8)
Primary	45 (31.7)	23 (35.9)	22 (28.2)
Secondary	48 (33.8)	18 (28.1)	30 (38.5)
University	27 (19.0)	11 (17.2)	16 (20.5)
IBS duration, n (%)				0.312	χ²=2.33
<2 years	38 (26.8)	15 (23.4)	23 (29.5)
2–5 years	56 (39.4)	28 (43.8)	28 (35.9)
>5 years	48 (33.8)	21 (32.8)	27 (34.6)
IBS subtype, n (%)				0.021*	χ²=9.87
IBS-D (Diarrhoea)	48 (33.8)	25 (39.1)	23 (29.5)
IBS-M (Mixed)	36 (25.4)	18 (28.1)	18 (23.1)
IBS-C (Constipation)	34 (23.9)	10 (15.6)	24 (30.8)
IBS-U (Unclassified)	24 (16.9)	11 (17.2)	13 (16.7)
P value for trend (IBS-D/M vs IBS-C/U)				0.008	χ²=7.02
Current medications, n (%)
Antispasmodics	68 (47.9)	32 (50.0)	36 (46.2)	0.743	χ²=0.11
Antidiarrheals	45 (31.7)	25 (39.1)	20 (25.6)	0.098	χ²=2.74
Laxatives	32 (22.5)	12 (18.8)	20 (25.6)	0.398	χ²=0.71
Antidepressants	18 (12.7)	8 (12.5)	10 (12.8)	1.000	Fisher’s exact
Comorbidities, n (%)
Hypertension	22 (15.5)	10 (15.6)	12 (15.4)	1.000	χ²=0.00
Diabetes	18 (12.7)	8 (12.5)	10 (12.8)	1.000	χ²=0.00
Anaemia (Hb<12)	35 (24.6)	22 (34.4)	13 (16.7)	0.018	χ²=5.62
Vitamin D deficiency	42 (29.6)	20 (31.3)	22 (28.2)	0.712	χ²=0.14
Previous Malaria, n (%)
Within 1 year	28 (19.7)	18 (28.1)	10 (12.8)	0.027	χ²=4.89
Within 5 years	45 (31.7)	26 (40.6)	19 (24.4)	0.045	χ²=4.02

P value from χ2 test; t-test for continuous variables; Fisher’s exact test where expected cell count <5.

IBS, irritable bowel syndrome; IBS-C, IBS with constipation; IBS-D, IBS with diarrhoea; IBS-M, IBS with mixed bowel habits; IBS-U, unclassified IBS.

The cohort was equally distributed by gender (71 males, 71 females), with a mean age of 42.3±11.7 years. Patients originated from 21 different governorates, reflecting a broad geographical distribution across Yemen. The most prevalent IBS subtype was IBS-D (33.8%), followed by IBS-M (25.4%), IBS-C (23.9%) and IBS-U (16.9%). No statistically significant differences were observed between patients with and without malaria in terms of age, gender or governorate distribution. Detailed demographic and epidemiological background information on these governorates is available in online supplemental resource S3.

Additional demographic characteristics

Data on ethnicity were not collected, as the Yemeni population is predominantly Arab with minimal ethnic diversity. Proximity to healthcare facilities was assessed based on the governorate of residence: 62.7% of participants resided in governorates with functional secondary or tertiary healthcare facilities within 50 km, while 37.3% resided in areas with limited healthcare access (primarily rural and conflict-affected regions). Socioeconomic indicators were not systematically collected due to the challenges of accurate assessment in a conflict-affected setting with economic instability. Body mass index (BMI) was calculated for 138 participants (97.2%): the mean BMI was 26.4±4.8 kg/m², with 28.3% classified as overweight (BMI 25–29.9) and 15.2% as obese (BMI≥30). Lifestyle factors, including smoking (current smokers: 18.3%) and dietary patterns, were recorded but not analysed in depth due to incomplete data. These variables were not included in the primary analysis, as they showed no significant association with malaria status in exploratory univariate analyses.

The cohort included patients from 21 of Yemen’s 22 governorates, encompassing diverse geographic regions (coastal, mountain, highland and island) and both urban (59.9%) and rural (40.1%) populations. The age and gender distribution of the study sample (mean age 42.3 years, 50% female) is comparable to previously reported demographic characteristics of patients with IBS in the Middle East and North Africa region,² suggesting that our cohort is reasonably representative of the broader IBS population in Yemen. However, given the conflict-affected context and challenges in healthcare access, some caution is warranted in generalising these findings to the entire Yemeni population with IBS.

Data completeness and quality were assessed using standardised checklists, with >98% data completeness achieved for all primary variables. Medication use and comorbid conditions were also recorded; these did not differ significantly between groups, except for a higher prevalence of anaemia in patients with malaria positivity (online supplemental table S11).

Prevalence of malaria co-infection and association with IBS subtype

The overall prevalence of malaria co-infection among patients with IBS was 45.1% (64/142). A significant association was found between IBS subtype and malaria status (χ²=9.87, p=0.021). The prevalence was highest in the IBS-D subgroup (52.1%, 25/48), followed by IBS-M (50.0%, 18/36), IBS-U (45.8%, 11/24), and was lowest in IBS-C (29.4%, 10/34). A post-hoc analysis comparing diarrhoeal/mixed subtypes (IBS-D and IBS-M) vs constipation/unclassified subtypes (IBS-C and IBS-U) revealed a significantly higher malaria prevalence in the former group (51.2% vs 36.2%, p=0.008). Detailed geographic distribution data are available in online supplemental table S6 and a geographic distribution map is provided in online supplemental figure S2. Analysis of enrolment timing showed no significant seasonal variation in malaria prevalence or biomarker levels among the study cohort (onlinesupplemental table S12 figure S8). The distribution of these biomarkers is visualised in figure 1.

Figure 1. Distribution of inflammatory biomarkers by malaria status. Box plots show the median, IQR and range for (A) neutrophil-to-lymphocyte ratio (NLR), (B) mean platelet volume (MPV) and (C) platelet-to-lymphocyte ratio (PLR). The group with malaria positivity exhibited significantly elevated NLR and PLR and reduced MPV compared with the group with malaria negativity (***p<0.001 for all comparisons).

Comparison of systemic inflammatory biomarkers

The levels of the NLR, MPV and PLR were compared between groups, as detailed in table 2. Complete haematological parameters with reference ranges are provided in online supplemental table S2.

Table 2. Comparison of haematological inflammatory biomarkers.

Biomarker	Total (n=142)	Malaria positive (n=64)	Malaria negative (n=78)	Mean difference (95% CI)	t-value	P value
NLR, Mean±SD	3.52±0.61	3.82±0.59	3.26±0.48	0.56 (0.40 to 0.72)	6.82	<0.001
MPV (fL), Mean±SD	7.88±0.68	7.56±0.52	8.16±0.61	−0.60 (−0.77 to −0.43)	−6.45	<0.001
PLR, Mean±SD	161.2±18.7	173.4±16.2	150.8±14.3	22.6 (17.8 to 27.4)	8.91	<0.001

MPV, mean platelet volume; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio.

Patients with malaria positivity exhibited a profoundly distinct inflammatory profile. NLR values were significantly elevated in the malaria-positive group (median (IQR): 3.80 (3.40–4.25) vs 3.25 (2.90–3.60), p<0.001). Conversely, MPV was significantly lower in patients with malaria co-infection (median (IQR): 7.55 (7.20–7.95) fL vs 8.15 (7.70–8.65) fL, p<0.001). The PLR demonstrated the most marked difference, with values of 172.5 (162.0–185.0) in the positive group compared with 150.0 (140.0–162.0) in the negative group (p<0.001). Figure 2 illustrates the prevalence of malaria co-infection across the different IBS subtypes.

Figure 2. Malaria prevalence across IBS subtypes. Bar chart demonstrating the percentage of malaria co-infection in each IBS subtype. IBS-D (diarrhoea-predominant) and IBS-M (mixed) subtypes showed the highest prevalence rates (52.1% and 50.0%, respectively). Error bars represent 95% CIs. IBS-C, IBS with constipation; IBS-U, unclassified IBS.

Correlation analysis of study variables

Bivariate correlation analysis revealed significant interrelationships between the biomarkers and their association with malaria status. A strong positive correlation was observed between NLR and PLR (r=0.79, p<0.001). Both NLR and PLR showed strong positive point-biserial correlations with malaria status (r_pb=0.72 and 0.81, respectively, p<0.001), while MPV showed a strong negative correlation (r_pb=−0.65, p<0.001). The comprehensive matrix of correlation coefficients for all study variables is provided in online supplemental table S7. Correlation networks are visualised in online supplemental figure S9.

Diagnostic performance of inflammatory biomarkers

ROC curve analysis was performed to evaluate the diagnostic potential of these biomarkers for identifying malaria co-infection. The PLR demonstrated the highest discriminatory power, with an AUC of 0.89 (95% CI 0.83 to 0.94). The NLR also showed excellent discrimination (AUC=0.85, 95% CI 0.79 to 0.91), while the MPV had a lower but significant AUC of 0.78 (95% CI 0.70 to 0.85). An optimal PLR cut-off value of 162.5 yielded a sensitivity of 84.4% and a specificity of 82.1%. Comprehensive diagnostic metrics, including positive predictive value, negative predictive value and accuracy, are presented in online supplemental table S4, and extended ROC analyses are shown in online supplemental figure S7 and S10.

Statistical comparison of the ROC curves using the DeLong test revealed that PLR demonstrated significantly better discriminatory performance than NLR (difference in AUC=0.04, 95% CI 0.01 to 0.07, p=0.023) and MPV (difference in AUC=0.11, 95% CI 0.06 to 0.16, p<0.001). NLR also performed significantly better than MPV (difference in AUC=0.07, 95% CI 0.03 to 0.11, p=0.002). The combined model (NLR+PLR) showed improved discrimination compared with PLR alone, but this difference did not reach statistical significance (difference in AUC=0.03, 95% CI −0.01 to 0.07, p=0.089).

The results of this analysis are shown in figure 3.

Figure 3. Receiver Operating Characteristic (ROC) curves for inflammatory biomarkers. Comparison of the diagnostic performance of PLR (AUC=0.89), NLR (AUC=0.85) and MPV (AUC=0.78) in predicting malaria co-infection among patients with IBS. The diagonal line represents random chance (AUC=0.50). The inset table shows the optimal cut-off values with corresponding sensitivity and specificity. IBS, irritable bowel syndrome; MPV, mean platelet volume; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio.

Statistical power and sample size considerations

To objectively assess the adequacy of our sample size, we conducted comprehensive post-hoc power analyses for all primary comparisons. As detailed in online supplemental table S14, our study demonstrated robust statistical power (>80%) for detecting the observed effects, with particularly strong power for biomarker comparisons and ROC analyses. Complete power analysis details are provided in online supplemental table S10.

All statistical assumptions were verified using appropriate diagnostic checks, as outlined in our statistical analysis checklist (online supplemental materials). Sensitivity analyses confirmed the robustness of these primary findings, demonstrating consistent effect sizes across multiple scenarios (online supplemental figure S8).

These analyses confirm that our sample size of 142 patients provided sufficient statistical power to detect clinically meaningful effects for our primary hypotheses, though we acknowledge the potential for limited power in exploratory subgroup analyses.

Multivariate logistic regression analysis for factors associated with malaria co-infection

The multivariate logistic regression model was adjusted for the following covariates: IBS subtype (dichotomised as IBS-D/M vs IBS-C/U), NLR (per 1-unit increase), MPV (per 1 fL increase) and PLR (per 10-unit increase). These variables were selected based on their association with malaria status in univariate analyses (p<0.10). The model did not include age or gender, as they showed no significant association in univariate analyses, and their inclusion did not improve model fit (likelihood ratio test, p>0.05).

To identify independent factors associated with malaria co-infection, variables significant in univariate analysis (IBS subtype (dichotomised as IBS-D/M vs IBS-C/U), NLR, MPV, PLR) were entered into a multivariate binary logistic regression model. The model was statistically significant (χ²=68.4, p<0.001) and explained 52% (Nagelkerke R²) of the variance in malaria status. The Hosmer-Lemeshow test indicated a good fit (p=0.42). The results are presented in table 3. A comparison of multiple logistic regression models and comprehensive model diagnostics is provided in online supplemental table S5.

Table 3. Univariate and multivariate logistic regression analysis for factors associated with malaria co-infection.

Predictor	Univariate analysis		Multivariate analysis
	OR (95% CI)	P value	aOR (95% CI)	P value
Age (per 10 year increase)	1.12 (0.85 to 1.48)	0.421	–	–
Gender (male vs female)	1.00 (0.52 to 1.92)	1.000	–	–
Residence (rural vs urban)	1.48 (0.78 to 2.81)	0.236	–	–
IBS duration (>5 vs ≤5 years)	0.92 (0.47 to 1.80)	0.812	–	–
IBS subtype (IBS-D/M vs IBS-C/U)	2.12 (1.08 to 4.16)	0.029	1.98 (0.92 to 4.28)	0.082
NLR (per 1-unit increase)	4.52 (2.78 to 7.35)	<0.001	3.21 (1.89 to 5.45)	<0.001
MPV (per 1 fL increase)	0.24 (0.14 to 0.41)	<0.001	0.32 (0.18 to 0.58)	<0.001
PLR (per 10-unit increase)	5.18 (3.12 to 8.59)	<0.001	4.02 (2.24 to 7.18)	<0.001
Anaemia (Hb<12 g/dL)	2.62 (1.17 to 5.87)	0.019	Not included*
Previous malaria (within 1 year)	2.67 (1.12 to 6.37)	0.027	Not included*

Model fit statistics for multivariate model: χ²=68.4, p<0.001; Nagelkerke R²=0.52; Hosmer-Lemeshow test p=0.42; AUC=0.92 (95% CI 0.87 to 0.96).

^*Variables not included in final multivariate model due to lack of independent association after adjustment for biomarkers.

aOR, adjusted odds ratios; IBS, irritable bowel syndrome; IBS-C, IBS with constipation; IBS-D, IBS with diarrhoea; IBS-M, IBS with mixed bowel habits; IBS-U, unclassified IBS; MPV, mean platelet volume; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio.

In the final adjusted model, elevated NLR (aOR=3.21, 95% CI 1.89 to 5.45) and elevated PLR (aOR=4.02 per 10-unit increase, 95% CI 2.24 to 7.18) were strong, independent factors associated with malaria positivity. Conversely, a higher MPV was associated with a significantly lower odds of malaria infection (aOR=0.32, 95% CI 0.18 to 0.58). The association of IBS subtype with malaria status was attenuated and lost statistical independence after adjustment for the inflammatory biomarkers (aOR=1.98, p=0.082).

Participants were recruited from 21 governorates across Yemen, reflecting a broad national distribution. Malaria prevalence among patients with IBS showed geographical variation, with the highest rates observed in the coastal governorates of Hodeidah (62.5%) and Al-Hudaydah (57.1%), as well as the highland governorate of Al-Mahwit (60.0%). Detailed geographic distribution data are available in online supplemental table S6.

Analysis of biomarker profiles across ibs subtypes and their interrelationships

No statistically significant differences were observed in NLR (p=0.12) or PLR (p=0.09) across the four IBS subtypes in a one-way ANOVA. However, when comparing the biomarker levels between the IBS-D/M and IBS-C/U subtypes, a significantly higher PLR was observed in the former group (165.5±19.1 vs 155.2±16.8, p=0.019). A detailed analysis of IBS subtypes is provided in online supplemental table S3), and subtype-specific profiles are visualised in online supplemental figure S6). The strong synergistic relationship between PLR and NLR in predicting malaria status is visually demonstrated in figure 4. The raw values for these biomarkers across all individual IBS subtypes are presented in online supplemental table S15.

Figure 4. Scatter plot of PLR vs NLR stratified by malaria status. Visualisation of the strong positive correlation between PLR and NLR (r=0.79, p<0.001) among 142 patients with IBS. Each point represents one patient, colour-coded by malaria status (blue: negative, red: positive). Dashed lines represent linear regression fits for each group. The shaded areas indicate 95% CIs around the regression lines. The clear separation between groups along both axes underscores the combined utility of these biomarkers in identifying malaria co-infection. IBS, irritable bowel syndrome; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio.

The scatter plot reveals distinct clustering patterns, with patients with malaria positivity predominantly occupying the upper right quadrant (high PLR and high NLR), while patients with malaria negativity cluster in the lower left quadrant (low PLR and low NLR). This visual separation corresponds with the statistical findings from our regression analysis and supports the clinical utility of using both biomarkers in tandem for risk stratification.

Interaction analysis

An exploratory two-way ANOVA was conducted to test for an interaction between IBS subtype (D/M vs C/U) and malaria status on PLR levels. The interaction term was not statistically significant (F(1138)=2.12, p=0.148), suggesting that the effect of malaria co-infection on elevating PLR is consistent across IBS subtypes. The main effect of malaria status remained significant (F(1138)=89.45, p<0.001), while the main effect of IBS subtype was not significant (F(1138)=2.89, p=0.091). The sample size (n=142) provides 78% power to detect a medium interaction effect (f=0.25) at α=0.05, but limited power for smaller interaction effects.

Discussion

Principal findings

This cross-sectional study provides novel insights into the intersection of IBS and malaria in Yemen, revealing several significant findings. To our knowledge, this represents the first investigation of malaria co-infection prevalence and systemic inflammatory biomarker profiles among patients with IBS in a high-burden malaria setting. The key findings are (1) a high prevalence (45.1%) of malaria co-infection among patients with IBS, substantially exceeding the estimated national malaria prevalence; (2) subtype-specific variation, with the highest rates in IBS-D (52.1%) and IBS-M (50.0%); (3) a distinct inflammatory profile in co-infected patients characterised by elevated NLR and PLR and reduced MPV; and (4) excellent diagnostic performance of PLR (AUC=0.89) for identifying malaria co-infection (figure 3, online supplemental table S4).

Comparison with existing evidence and potential mechanisms

The observed malaria prevalence of 45.1% far exceeds the estimated national malaria prevalence in Yemen (approximately 12–15% according to recent WHO reports),¹⁶ suggesting that patients with IBS may represent a vulnerable subgroup. This elevated prevalence aligns with studies suggesting that chronic inflammatory conditions may increase susceptibility to infectious diseases.^{12 13} The biomarker values in our co-infected patients (mean NLR 3.82, PLR 173.4, MPV 7.56 fL) are consistent with those reported in patients with malaria without IBS in Yemen⁹ and other malaria-endemic regions,^{17 18} but are notably higher than values reported in patients with IBS alone in other populations (NLR 2.8, PLR 145.2),⁶ suggesting an additive or synergistic inflammatory effect. This is further supported by the particularly high co-infection prevalence in IBS-D and IBS-M subtypes, which are known to exhibit more pronounced inflammatory profiles compared with constipation-predominant IBS.¹⁹ The elevated NLR and PLR in co-infected patients reflect a characteristic inflammatory response to malaria, consistent with previous studies documenting NLR and PLR elevation in various infectious and inflammatory conditions,^{20 21} with neutrophilia occurring in response to parasite clearance and lymphopenia resulting from lymphocyte sequestration and apoptosis.¹⁷ The significantly reduced MPV aligns with established haematological changes in malaria, where platelet consumption and activation lead to smaller, more active platelets.¹⁸

Several interconnected pathways may explain this association: gut barrier dysfunction may facilitate the systemic dissemination of malaria antigens22,24; malaria-induced inflammation could further disrupt gut barrier function through cytokine-mediated tight junction modulation (particularly TNF-α and IL-6)^{25 26}; mast cell activation may create a vicious cycle of inflammation^{27 28}; cytokine cross-talk (TNF-α, IL-6, interferon (IFN)-γ in malaria¹⁷; IL-8, IL-10 in IBS²⁹) can modulate the gut-brain axis, affecting visceral sensitivity and motility³⁰; and microbiome alterations may exacerbate symptoms or impair immune responses31,33 (online supplemental figure S3). Age-stratified biomarker profiles are presented in online supplemental figure S4.

A detailed comparison of our biomarker values with those reported in other populations is provided in online supplemental table S9.

Diagnostic and predictive utility of inflammatory biomarkers

The excellent discriminatory performance of PLR (AUC=0.89) and NLR (AUC=0.85) for identifying malaria co-infection suggests that these readily available, cost-effective biomarkers could serve as useful screening tools in resource-limited settings. In multivariate analysis, both biomarkers were strongly associated with malaria positivity, with PLR demonstrating a particularly robust association (aOR=4.02 per 10-unit increase). These findings extend previous research on inflammatory biomarkers in infectious diseases^{34 35} to the specific context of IBS-malaria co-infection. The optimal PLR cut-off value of 162.5 (sensitivity 84.4%, specificity 82.1%) provides a clinically applicable threshold that could guide targeted malaria testing. The direct clinical implication is straightforward: in a gastroenterology clinic in a malaria-endemic region like Yemen, a patient with IBS—especially one with IBS-D or IBS-M—who is found to have an elevated PLR (>162.5) or NLR (>3.65) on routine CBC should be prioritised for malaria testing, even in the absence of classic febrile symptoms. This targeted, biomarker-guided approach could significantly improve case detection and resource efficiency, aligning with WHO recommendations for rational diagnostic testing in resource-limited settings.³⁶

Strengths and limitations

This study possesses several notable methodological strengths, including its pioneering nature, multicentre design across 21 governorates, rigorous diagnostic criteria (Rome IV for IBS, WHO-endorsed RDTs for malaria), robust statistical modelling and the identification of clinically applicable biomarker cut-offs (online supplemental figure S5). However, several limitations should be acknowledged. The cross-sectional design precludes causal inference regarding the temporal relationship between IBS and malaria; we cannot determine whether IBS predisposes individuals to malaria infection, vice versa, or if shared environmental or host factors explain the association, as gut-brain axis dysregulation may play a role.³⁷ Longitudinal studies are needed to elucidate these temporal dynamics. The absence of control groups (healthy individuals and patients with malaria only) limits our ability to attribute the observed inflammatory profiles specifically to the co-infection state (online supplemental table S13). Residual confounding by unmeasured factors cannot be excluded, and the sample size, while adequate for primary analyses, limits robust subgroup analyses. Furthermore, the lack of PCR confirmation and parasite density quantification precludes the analysis of species-specific associations or dose-response relationships. Finally, while our findings are most directly applicable to Yemen’s unique context—a high-burden malaria setting affected by ongoing conflict and healthcare system fragmentation—the specific biomarker cut-offs require validation in other geographical settings.

Unanswered questions and future research

Building on our findings, several critical research avenues emerge. Longitudinal cohort studies are needed to establish the temporal relationship and directionality between IBS and malaria. Mechanistic investigations should explore gut barrier integrity markers (eg, zonulin), detailed cytokine profiling (TNF-α, IL-6, IL-8, IL-10),^{17 29} microbiome analysis31,33 and platelet function studies^{5 23} in co-infected patients. Interventional trials examining whether anti-malarial treatment in co-infected patients improves IBS symptoms would provide direct clinical evidence for integrated management. Diagnostic validation studies are needed to confirm our biomarker cut-offs in independent populations and settings, including comparative studies with other Plasmodium species, given Yemen’s mixed P. falciparum/P. vivax transmission.⁹ The registered PROSPERO protocol (CRD420251069483) provides a framework to formally test our findings against the global literature and examine the interaction between infectious burden and IBS pathophysiology on a larger scale.

Conclusion

This study establishes a significant association between IBS and malaria co-infection in Yemen, with nearly half of the patients with IBS testing positive for malaria. The distinct systemic inflammatory profile characterised by elevated NLR and PLR, along with reduced MPV, provides biological plausibility for this association and offers practical biomarkers for risk stratification. The excellent diagnostic performance of PLR and NLR suggests that these inexpensive, readily available indices could enhance malaria detection in patients with IBS, particularly in resource-limited settings where diagnostic testing capacity may be constrained. These findings underscore the importance of integrated approaches to communicable and non-communicable disease management in endemic regions and highlight the need for increased clinical awareness of potential malaria co-infection in patients with IBS who present with gastrointestinal symptoms.

Data availability statement

Data are available in a public, open access repository.