FODMAP diet efficacy and mechanistic plausibility

Systematic reviews and meta-analyses of randomised controlled trials (RCTs) conducted across different countries and research groups demonstrate that the low FODMAP diet (LFD) improves symptoms in people with irritable bowel syndrome (IBS)^{(Reference Cuffe, Staudacher and Aziz1)}. Consequently, the LFD is recommended in international clinical management guidelines as a first- or second-line treatment for IBS^{(Reference Lacy, Pimentel and Brenner2–Reference McKenzie, Bowyer and Leach4)}. However, despite this extensive evidence base, it remains unclear how the LFD produces symptom improvement^{(Reference Biesiekierski, Manning and Murray5)}. In particular, it is uncertain whether benefits are primarily explained by reductions in FODMAP intake itself or by other factors such as placebo effects, increased dietary attention, broader changes in nutrient intake (e.g. fat or fibre), or reduced food-related anxiety.

A substantial body of research supports the biological plausibility of FODMAP-related symptom generation in IBS. Poorly absorbed fermentable carbohydrates increase luminal water and gas production^{(Reference Barrett, Gearry and Muir6,Reference Ong, Mitchell and Barrett7)} , and together with visceral hypersensitivity provoke symptoms in susceptible individuals^{(Reference Murray, Wilkinson-Smith and Hoad8,Reference Major, Pritchard and Murray9)} . FODMAP restriction also alters gut microbiota composition and metabolic output^{(Reference Halmos, Christophersen and Bird10–Reference McIntosh, Reed and Schneider12)}, while emerging evidence implicates gut–brain axis dysregulation^{(Reference Wu, Masuy and Biesiekierski13)}. Rechallenge studies further demonstrate that reintroduction of specific FODMAP groups can trigger symptom recurrence in some individuals^{(Reference Eswaran, Jencks and Singh14,Reference Van den Houte, Colomier and Routhiaux15)} . Long-term follow-up studies suggest that symptom improvement can persist even as FODMAP intake is liberalised, highlighting complexity in the relationship between FODMAP intake and sustained symptom response^{(Reference McIntosh, Reed and Schneider12,Reference Rej, Sanders and Shaw16,Reference Harvie, Chisholm and Bisanz17)} .

While these findings support biological plausibility, they do not directly address a complementary question relevant to clinical trials: whether reductions in FODMAP intake explain the symptom improvements observed in LFD interventions. Demonstrating physiological effects of FODMAPs is not equivalent to showing that variation in FODMAP intake within trials accounts for therapeutic benefit. Addressing this question requires an analytical perspective that examines how changes in dietary intake relate to changes in symptoms. One approach is to apply principles from mediation analysis to distinguish intervention effects from the processes through which they occur.

Addressing this gap is central to the development of less restrictive, more targeted dietary strategies such as the ‘bottom-up’ or modified FODMAP approach^{(Reference Singh, Tuck and Gibson18)}. This strategy aims to restrict only those FODMAP groups that trigger symptoms, thereby reducing unnecessary dietary exclusion and improving long-term acceptability and sustainability^{(Reference O’Brien, Kasti and Halmos19,Reference Lomer20)} . However, implementing such an approach requires evidence identifying which FODMAP types matter most, how much reduction is required, and whether partial restriction is sufficient to achieve symptom control^{(Reference Chey, Keefer and Whelan21,Reference Rej, Shaw and Buckle22)} .

Conceptual framework: from efficacy to mechanism in low FODMAP diet research

Randomised trials can demonstrate that an intervention improves outcomes, but they do not necessarily explain how that improvement occurs. One way of exploring how treatments work is to examine intermediate processes that link an intervention to its effects on outcomes^{(Reference Schuler, Coffman and Stuart23,Reference Burton Murray and Ljótsson24)} . Such variables are often referred to as mediators, which are factors that may account for the relationship between an intervention and symptom change^{(Reference Kazdin25)}. Mediation analysis provides a formal statistical framework for examining whether an intervention improves outcomes because it changes a specific intermediate factor^{(Reference Schuler, Coffman and Stuart23)}. In the context of the LFD, this framework asks whether symptom improvement potentially occurs because FODMAP intake is reduced, or whether benefits arise through alternative mechanisms. In this review, mediation principles are applied conceptually to guide interpretation of existing LFD trials and to examine whether available evidence supports a mechanistic role for reductions in FODMAP intake; formal mediation effects are not estimated.

Within this mediation-informed framework, two pathways are central: whether the intervention alters FODMAP intake (a path), and whether variation in FODMAP intake is associated with symptom outcomes, controlling for treatment allocation (b path). The distinction between the a and b paths is shown in Figure 1 ^{(Reference VanderWeele26,Reference Fairchild and McDaniel27)} . Mediation is formally tested by estimating the ab product, commonly referred to as the indirect or mediated effect^{(Reference Preacher and Hayes28)}. This effect reflects the extent to which the intervention influences symptoms through changes in FODMAP intake and is evaluated statistically in formal mediation analysis. Although mediation effects are not estimated here, this framework provides a basis for evaluating whether existing trials assess the elements necessary to support such a pathway.

Figure 1.

Mediation-informed framework applied in this review. The a path represents whether the intervention alters FODMAP intake. The b path represents whether differences in intake are associated with symptom outcomes independent of intervention group. That is, the FODMAP diet is presumed to change FODMAP intake (a path); and in turn the change in FODMAP intake affects change in symptoms (b path). Symptom improvement may also occur through pathways not explained (mediated) by FODMAP intake. This framework is used to evaluate whether existing trials provide evidence relevant to these pathways; mediation effects are not formally estimated.

Figure 1. Long description

The flowchart depicts a mediation-informed framework used to evaluate the impact of FODMAP intake on gut symptoms. The process begins with treatment allocation, which influences the mediator, FODMAP intake. The a path shows how the intervention alters FODMAP intake. The b path illustrates the association between changes in FODMAP intake and symptom outcomes, independent of the intervention group. The framework suggests that the FODMAP diet changes FODMAP intake, which in turn affects symptom improvement. Additional pathways not explained by FODMAP intake may also contribute to symptom improvement. This framework is used to assess whether existing trials provide evidence relevant to these pathways, though mediation effects are not formally estimated.

Although multiple biological pathways may contribute to symptom improvement, including alterations in gut microbiota composition and metabolite production, this review focuses specifically on FODMAP intake because it is the dietary exposure directly targeted by the intervention. Downstream physiological processes (e.g., fermentation, microbial metabolism, gut–brain signalling) may represent subsequent mediators of symptom improvement, but evaluation of these biological pathways is beyond the scope of this review.

Mediation-informed framework for evaluating low FODMAP trials

Building on the conceptual framework outlined above, we evaluated how published LFD RCTs were designed, measured, and analysed. Rather than estimating mediation effects, we examined whether key elements needed to support a mechanistic role for FODMAP intake were present in existing studies^{(Reference Cuffe, Staudacher and Aziz1)}. Evaluating whether reductions in FODMAP intake plausibly explain symptom improvement requires study designs that allow intervention effects on intake to be distinguished from the relationship between intake and symptoms.

Randomised trials provide an important advantage when examining the first component of this pathway. Because participants are randomly assigned to dietary interventions, differences in FODMAP intake between groups can be attributed to the intervention rather than to pre-existing differences between participants. Randomisation therefore allows causal inference regarding whether the intervention altered FODMAP intake (the a path). However, the relationship between FODMAP intake and symptom outcomes (the b path) remains observational even within randomised trials, because intake itself is not randomised. Demonstrating that individuals with lower intake experience greater symptom improvement therefore requires analyses that directly relate intake differences to symptom outcomes while accounting for diet assignment.

Analyses that measure intake and symptoms at a single time point, such as cross-sectional associations measured at the end of an intervention, provide limited insight into this pathway^{(Reference Schuler, Coffman and Stuart23,Reference Kazdin25)} . Because both intake and symptoms may be influenced by the intervention, such analyses cannot determine whether intake differences explain symptom improvement or simply reflect differences between intervention groups^{(Reference Fairchild and McDaniel27,Reference Alaiti, Saragiotto and Fukusawa29)} . As a result, cross-sectional analyses may suggest plausibility but cannot test whether changes in FODMAP intake explain symptom change.

Many LFD trials use single-group pre–post designs, in which participants are assessed before and after dietary treatment without a comparator group. These studies are valuable for describing how dietary intake and symptoms change over time but have limited capacity to inform causal mechanisms. Because all participants receive the same intervention, it is not possible to distinguish intervention effects from other influences such as natural variation in symptoms, placebo effects, or general time-related change such as seasonal dietary variation^{(Reference Kazdin25,Reference Fairchild and McDaniel27)} . By contrast, studies that include a comparator group provide the variation needed to distinguish intervention effects from background change. This allows assessment of whether a LFD reduces FODMAP intake relative to an alternative approach, and whether differences in intake are accompanied by differences in symptoms. For this reason, the present review focuses on randomised trials when considering mechanistic pathways, while recognising the descriptive and hypothesis-generating value of non-randomised studies^{(Reference Lee, Cashin and Lamb30)}.

In summary, this review examines randomised trials to assess what evidence exists for the two component pathways required to support a mechanistic role for FODMAP intake: whether interventions altered intake (a path), and whether variation in intake was associated with symptom outcomes (b path).

Assessment of the a path (Intervention → FODMAP intake)

Applying this framework, we first examined whether LFD trials demonstrated that the intervention altered FODMAP intake (a path). To support the a path, studies must demonstrate that the LFD intervention reduces FODMAP intake to a greater extent than the comparator diet. Analyses that examine changes within a single group before and after the intervention cannot determine whether intake changed because of the diet itself or for other reasons. In contrast, comparing intake changes between intervention and comparator groups allows estimation of the intervention’s effect on FODMAP intake.

The strength and precision of this evidence depend on how intake is analysed. Post-intervention comparisons between groups provide an estimate of intake differences but may be influenced by baseline variation. Analyses that account for baseline intake, such as baseline-adjusted regression models (ANCOVA), generally provide more precise estimates of intervention effects by reducing between-subject variability. Analyses based on change scores or two-timepoint mixed models similarly evaluate between-group differences in intake change over time but may be somewhat less efficient in the presence of measurement error in baseline values.

Some studies also report physiological markers, such as breath hydrogen responses, alongside dietary intake changes. While aligned changes in dietary intake and physiological markers support the plausibility that the intervention altered intake, these measures do not replace the need for appropriate analysis of dietary intake data and should be viewed as supportive rather than definitive evidence for the a path.

To provide a consistent basis for evaluating whether low FODMAP trials altered FODMAP intake, studies were classified according to the analytic approaches outlined in Box 1.

Assessment of the b path (FODMAP Intake → symptom outcomes)

We next examined whether trials evaluated the second component of the proposed mechanistic pathways: whether variation in FODMAP intake was associated with symptom outcomes (b path). To support the b path, studies need to examine whether individuals with lower intake (or larger reductions in intake) also experience greater symptom improvement. Many trials interpret symptom improvement as reflecting FODMAP reduction, although the intake–symptom pathway is rarely formally examined. Analyses must take diet assignment into account, as allocation to a LFD influences both intake and symptoms.

Post-intervention comparisons or simple intake–symptom associations that do not adjust for diet group and baseline symptoms provide limited evidence for the b path. These approaches cannot determine whether intake itself explains symptom improvement or whether the observed association reflects the overall effect of the intervention.

Comparisons of responders and non-responders based on baseline FODMAP intake do not test the intake–symptom pathway. Baseline intake precedes the intervention and therefore cannot explain symptom changes that occur during treatment.

Some studies report associations between physiological markers (e.g., breath hydrogen or gut microbiota changes) and symptoms. While these findings support biological plausibility, they do not demonstrate that variation in FODMAP intake explains symptom improvement.

Clearer evidence for the b path comes from analyses that directly relate differences in FODMAP intake (or changes in intake) to symptom outcomes while accounting for diet assignment and baseline symptom severity. We used the framework outlined in Box 2 to classify how existing trials addressed this relationship.

Assessment of FODMAP intake measurement and its implications for mechanistic interpretation

Beyond evaluating the a path and b path, mechanistic interpretation also depends on how FODMAP intake itself is measured, analysed, and interpreted. The following sections consider dietary assessment, dietary collinearity, adherence, and the practical challenges of mediation analysis in dietary trials.

Dietary assessment considerations

Understanding whether changes in FODMAP intake leads to symptom improvement depends on how clearly intake is measured, reported, and analysed. While analytic issues have been discussed above, limitations in dietary assessment constrain what can be inferred about mechanism.

Measuring dietary intake is challenging due to reporting error, behaviour changes during recording, coding inaccuracies, and limitations of food composition databases^{(Reference Staudacher, Yao and Chey48)}. Diet intake also varies substantially within individuals from day to day, particularly for FODMAPs^{(Reference Nybacka, Störsrud and Liljebo49)}. While two to three days of recording may rank individuals by total FODMAP intake, substantially more days are required to estimate usual intake accurately, especially for FODMAP subgroups such as excess fructose and polyols^{(Reference Nybacka, Störsrud and Liljebo49)}.

Studies that summarise FODMAP intake using ranks, scores, or categories indicate qualitative changes but provide limited information about the magnitude of intake change. While such measures can demonstrate differences between intervention groups, they are less suited to examining whether symptom improvement varies according to the degree of FODMAP reduction. Measuring FODMAP intake in grams allows evaluation of whether symptom improvement increases with greater reductions in intake, whether a threshold level is required, and whether specific FODMAP subgroups contribute more strongly to symptom change; questions that are important to develop more targeted dietary strategies. Expressing intake in grams also enables comparison across studies using a common scale.

These considerations provide context for interpreting how FODMAP intake was assessed and analysed across randomised trials (Table 3), and how methodological variability limits conclusions about how FODMAP intake changes relate to symptom improvement.

Table 3.

Methods used to assess and report FODMAP intake in low FODMAP RCTs

Table 3. Long description

A table comparing methods used to assess and report FODMAP intake in low FODMAP randomized controlled trials. The table has 18 rows and 4 columns. The columns are labeled Study, Assessment method, FODMAPs reported, Intake metric (scale & summary), and Database analysis. Each row provides detailed information on the study, the assessment method used, the specific FODMAPs reported, the intake metric and summary, and the database analysis method. The table highlights the variability in methods and reporting across different studies, which limits the conclusions about how FODMAP intake changes relate to symptom improvement.

RCTs, randomised controlled trials. g/day, grams per day. mean (SD), mean (standard deviation). FOS, Fructo-oligosaccharides. GOS, galacto-oligosaccharides. NSP, non-starch polysaccharides. CNAQ, Comprehensive Nutrition Assessment Questionnaire. HPLC, high-performance liquid chromatography.

^a Reporting format (mean vs median) was recorded to illustrate comparability limitations, not to assess study quality.

^b Fermentable carbohydrates do not sum to total fermentable carbohydrate value due to transformation of data^{(Reference Staudacher, Lomer and Anderson31)}.

^c Post-intervention intake not measured; intake inferred from controlled diet composition^{(Reference Halmos, Power and Shepherd32)}

^d Fructose in excess of glucose cannot be estimated from the FoodWorks program^{(Reference Halmos, Power and Shepherd32)}.

^e Excess fructose calculated based on assumption glucose and fructose main monosaccharides in food^{(Reference Böhn, Störsrud and Liljebo33)}.

^f 14-point scoring system designed for this study^{(Reference McIntosh, Reed and Schneider12)}.

^g Patients with IBD and IBS-like symptoms^{(Reference Pedersen, Ankersen and Felding35)}.

^h Dietary intake was recorded digitally using a national dietary assessment platform- Riksmaten Flex 2021 of the Swedish Food Agency^{(Reference Roth, Nseir and Jeppsson45)}.

ⁱ Dietary variables were reported, but none were explicitly defined or quantified as FODMAP intake^{(Reference Roth, Nseir and Jeppsson45)}.

^j Gluten intake was used as a surrogate measure for fructan intake^{(Reference Singh, Chey and Nee47)}.

Summary and implications for mechanistic understanding of the LFD

The purpose of this review was to examine whether existing LFD trials provide evidence that reductions in FODMAP intake explain symptom improvement in IBS. Using a mediation-informed framework, we evaluated whether published LFD trials assessed the key elements needed to support a mechanistic role for FODMAP intake. Overall, while the LFD is clearly effective, there is limited evidence linking the degree of FODMAP reduction to symptom outcomes. Addressing this gap through mediation-informed trial design is essential for informing less restrictive dietary strategies and for advancing personalised dietary management of IBS.

Although a substantial body of RCTs demonstrates consistent symptom improvement with the LFD, evidence for efficacy has not been matched by equivalent evidence explaining how the diet exerts its effects. Many trials show that LFD interventions reduce FODMAP intake, but few assess whether differences in intake change are associated with differences in symptom improvement between individuals – this largely reflects the design and analytic priorities of existing trials. Most studies were structured to demonstrate clinical benefit rather than test mechanistic hypotheses, and therefore rarely examined intake–symptom relationships directly. Limitations in dietary assessment precision, reliance on proxy adherence measures, co-occurring changes in other dietary components, and the absence of intake-symptom modelling all constrain interpretation of whether reduced FODMAP intake itself explains symptom improvement.

Reductions in FODMAP intake may influence symptoms through downstream biological processes in the gut, including altered fermentation, shifts in microbiota composition, and changes in metabolite production. These processes represent plausible biological mediators linking dietary exposure to symptom outcomes. However, establishing whether such mechanisms operate within clinical trials first requires evidence that variation in FODMAP intake is associated with variation in symptoms. Evaluating intake–symptom relationships is therefore a necessary step in determining whether downstream biological changes form part of the causal pathway linking FODMAP reduction to symptom improvement.

Advancing understanding of how the LFD works will require trials that align dietary intake measurement and analysis with mechanistic aims. Even when the primary goal is to establish efficacy, studies can be designed and analysed in ways that strengthen mechanistic insight, including accurate quantitative assessment of intake and analyses examining intake-symptom relationships. More precise dietary assessment and adequately powered analyses may require additional resources, but these can be incorporated into trial design and funding where mechanistic understanding is a key objective. Such approaches would help identify the FODMAP components and degree of restriction most strongly linked to symptom improvement, supporting the development of less restrictive, targeted, and personalised dietary strategies. Mechanistic studies conducted earlier in intervention development could also help refine dietary approaches before large-scale efficacy trials, improving both intervention design and research efficacy. Moving beyond demonstrating that the diet works toward understanding how and for whom it works represents the next phase of LFD research.