Post-infectious functional gastrointestinal disorders (FGID) are a relatively well-studied phenomenon among individuals who are resident in temperate and higher-income regions around the world. Less is known about the risk of FGID among residents in tropical and hyperendemic settings where acute gastroenteritis risk and exposure is high. This editorial summarizes the primary results from a recently published study focusing on this unique clinical scenario and considers the interpretation of the data as well as highlights additional research needs.
Drs. Ghoshal and Rahman and colleagues are commended for an important contribution to the epidemiology of post-infectious functional gastrointestinal disorders (FGID) where they describe the first rigorously designed observational study in an area with hyper-endemic enteric infection and advancing our knowledge-base which, to date, has been largely limited to populations of high-income countries . The gold standard study design in assessment of the association between acute enteric infection and the subsequent risk of chronic consequences is the prospective cohort, which if done rigorously, assures estimates of true incidence, absolute and relative risks, minimizes biases, and has the ability to consider confounding factors and high-risk sub-groups. Given the setting where individuals are likely to have had multiple pathogenic exposures early in life, higher frequency of chronic protozoal and helminth infections, and a high force of infection (often leading to asymptomatic infection and sometimes illness) throughout the life-span, the challenges in design and interpretation of such studies are significant.
In this study, the research team enrolled adult patients admitted for a short-stay encounter at the Dhaka Hospital in Bangladesh with the primary diagnosis of acute gastroenteritis (AG), and matched each of these by age and gender to a non-ill control from health attendants of patients and hospital staff of the hospital. Relevant eligibility criteria were applied to both AG case patients (exposed) and the non-ill (unexposed) population to exclude subjects with organic diseases and prevalent baseline FGID at enrollment to develop the “at-risk” cohort. Primary outcomes of post-infectious irritable bowel syndrome (PI-IBS), functional dyspepsia (PI-FD), and malabsorptive syndrome (PI-MAS) were ascertained at 6, 9, and 12 months post enrollment. Novel outcomes of chronic bowel dysfunction (CBD) and “dyspeptic symptoms” were ascertained and defined as symptoms meeting IBS or dyspepsia definitions, respectively, for more than 3 but less than 6 months. Additional diagnostic tests were performed on a subset of the individuals that developed a condition with persistent abdominal symptoms including breath testing, measurement of fecal fat, and duodenal biopsy to ascertain potential etiologies of new onset malabsorptive or other organic disease. Baseline testing for a limited number of potential enteropathogens was done at enrollment on patients with acute diarrhea, but not “unexposed” controls. Relative risk was estimated and effect modification and confounding assessed using multivariate methods for a number of demographic, symptom severity, and psychological covariates. The authors report both intention-to-treat and per-protocol analyses which are not clearly described or differentiated in the methods.
Among those with acute diarrhea, the absolute risk at 12 months post enrollment was high for PI-IBS (risk difference, 13.9%) and PI-FD (6.8%) compared to an incidence rate of 2.6% for IBS and 0.6% for functional dyspepsia in “unexposed” subjects. There was considerable overlap in meeting defined FGID diagnostic criteria. The relative risk for any FGID (IBS, FD, overlap) was 6.1 (95% CI: 3.2–11.2), a relatively higher effect estimate compared to those in recent non-endemic setting systematic reviews. “Chronic bowel dysfunction” was frequent and occurred in 18.2% of exposed and 3.2% of non-exposed cohort subjects. In analysis of risk factors for PI-FGID among those with acute diarrheal illness on enrollment, there was no association found with gender, age, or SES. Maximum frequency of diarrhea on acute illness trended toward a positive association with PI-FGID risk on univariate analysis, but fever trended toward being more prevalent in those not developing an FGID (39.1%) compared to those that did (28.8%). Weight loss associated with GE (though not defined as to whether this was acute weight loss or weight loss in the subsequent observation period) trended toward an association with increased PI-FGID risk. The presence of blood in the stools (e.g., dysentery) was overall rare and not differential between those with and without PI-FGID. Increased risk of “dyspeptic symptoms” and CBD was associated with an increased risk of subsequent PI-FGID; however, by definition of these symptoms, and in multivariate analyses there appears to be a high level of collinearity. This is supported by the instability of the model (Table 5) represented by enormous crude and adjusted odds ratios. The appropriateness of including these two “intermediate” outcomes in the FGID pathway as potentially concerning. Additionally, caution should be used when interpreting the odds ratios from logistic regression models for cohort-based studies with relatively common endpoints. The authors did not describe the overlap between those with dyspeptic symptoms and CBD, which by definition should be largely distinct (symptoms lasting between 3 and 6 months post infection) from the population that goes on to develop an FGID (symptoms present for at least 6 months). Bacterial etiologies were most common and occurred at relatively higher frequency in PI-IBS patients compared to PI-FD patients, though and the data are not presented in a way to differentiate the relative rates of infection among those who did or did not go on to develop an FGID. No microbial testing was performed in “unexposed” subjects so it is difficult to ascertain causality of a given pathogen. Oddly, no parasites were identified in this setting, and norovirus testing, which has previously been associated with increased risk in PI-FD, was not performed . Furthermore, Strongyloides stercoralis, an intestinal nematode infection, is of particular importance in the long-term traveler (as well as immigrants) and often attributed to non-diarrheal gastrointestinal syndromes. Given the infrequency of testing and diagnostic challenges of Strongyloides sp., the attributable burden of persistent abdominal symptoms in the high AG endemic setting due to Strongyloides sp. may be underestimated .
Observational studies such as the one described by Ghoshal et al. are challenging and, given the non-controlled nature, design and execution can yield a number of biases and limitations, which necessitate the results to be interpreted cautiously. It is interesting to note that at baseline, 3.8% of those with AG and 1.9% of unexposed “controls” had prior IBS. Overall these rates appear relatively low from a general population perspective among countries in the region . Additionally, the two-fold higher baseline prevalence of IBS among AG subjects supports the notion that AG may be increased among those with pre-existing IBS, as previously observed . The relatively low IBS prevalence at screening could represent some uncontrolled misclassification bias of an “at risk” for FGID sample in this study. For example, it was not described as to whether cohort participants at baseline were asked only about a prevalent condition of IBS (or FD) or if they had ever had it in their life-times. Certainly FGID symptoms can wax and wane over a person’s life and those with “inactive” FGID could be more susceptible to AG as well as “recurrence” of FGID following infection.
The authors report on the importance of overlap between PI-IBS and PI-MAS in the high AG endemic setting. In the present study, 23 of the 57 (40%) of patients fulfilling the Rome III criteria for IBS underwent testing for evidence of malabsorption. Of these, 2 (8.7%) were considered to meet criteria based on breath, fecal fat, and duodenal biopsy for malabsorption. It is unclear if those who underwent these tests are generalizable to those who did not, and it is unclear how many of these individuals (or those in the “unexposed” group) had baseline abnormalities. Interestingly, these two PI-MAS subjects improved with antibiotic regimens containing metronidazole, which is consistent with a recently described non-randomized controlled study of 102 Israeli adults with persistent abdominal symptoms attributed to travel and a negative work-up for infectious etiology who were treated with a combination therapy of oral tinidazole and albendazole . Baseline persistent abdominal symptoms of diarrhea, abdominal pain, flatulence, and bloating were common and present for an average of 16 months prior to therapy with an additional two-thirds with extreme fatigue. When followed up 6 months after treatment, 69% reported an improvement in GI symptoms and over a third reported full recovery within a few weeks post treatment. Future studies among individuals with persistent abdominal symptoms after infection in these hyper-endemic settings should continue to undergo testing for underlying pathogenic mechanisms for which directed treatments may be available, including malabsorption. However, at present, given the aforementioned limitations of our understanding, the need for exclusion of PI-MAS in this setting before a diagnosis of IBS cannot be yet suggested.
There are always questions of residual confounding regarding what might not have been measured or analyzed in understanding risk factors. A observational study from Malaysia describes persistent abdominal pain and IBS after a major flooding event in which there were major populations effected by significant break downs in sanitation and hygiene with attendant AG and environmental contaminant exposures . Interestingly, poor WaSH practices and impaired quality of life during flood were significantly associated with IBS and, at least in part, could be explained by disturbed microbiota resulting from these extraordinary exposures. In the present study there was no comparison of AG “exposed” and “unexposed” with relevant factors having to do with access to hygiene or SES, though there were no differences in these measures among those who developed PI-FGD and those who did not. It would be interesting to know whether the “unexposed” controls that were selected were different than the “exposed” AG cases in relation to these important factors of general sanitation hygiene levels and acute/chronic life-stress.
As stated in the title, the results from this study are an important contribution to the literature and should lead to further investigation to confirm reported results and clarify some of the questions that have emerged from the reported observations. Whether our understanding of post-infectious FGD risk among high AG endemic settings is clearer, or rather more questions have been raised which cloud our interpretation and ability to translate these findings is an open question. In fact, both may be true. What is clear is that future studies are needed to understand the risk and pathological basis of the spectrum of functional and potentially treatable persistent abdominal symptoms in this population.
The contents of this publication are the sole responsibility of the authors (or speaker) and do not necessarily reflect the views, assertions, opinions or policies of the Uniformed Services University of the Health Sciences (USUHS), the Department of Defense (DoD), or the Departments of the Army, Navy, or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government.