Small intestinal water

One of the two most established mechanisms by which FODMAPs are proposed to provoke symptoms in IBS is the augmentation of small intestinal water, which has been clearly demonstrated by both ileostomy recovery and MRI studies (table 1, figure 1). One randomised, single-blind, cross-over feeding study in 10 ileostomates showed effluent water increased by 20% after a 4 day very high FODMAP diet (112 g/day) compared with a with very low FODMAP diet (6 g/day).36 An even more pronounced effect on small intestinal water has been demonstrated in response to acute challenges using MRI. Healthy individuals exhibited a fourfold higher small intestinal water volume 60 min after consumption of a 17.5 g mannitol solution compared with an equimolar glucose solution.37 The same magnitude of effect was seen 60 min after administration of 40 g fructose,38 39 and this was partially resolved through contemporaneous ingestion of 40 g glucose,38 thought to be due to enhanced co-transport of fructose and glucose via the GLUT-2 transporter. Conversely, inulin (a high DP fructan) had no effect on small intestinal water in healthy individuals38 or in patients with IBS.39 Further study is needed on the effect on small intestinal water of smaller DP fructans that are more representative of those found in the diet. In addition, the effects of other oligosaccharides (e.g. galacto-oligosaccharides, GOS), the disaccharide lactose and other polyols (e.g. sorbitol) on small intestinal water are unknown.

The impact of increased small intestinal water on functional GI symptoms in IBS is unclear. First, the increase in luminal water may induce abdominal pain and bloating in those with visceral hypersensitivity, although recent research failed to demonstrate a correlation between peak small intestinal water and symptom exacerbation in IBS following blinded challenge with fructose (n=11) or inulin (n=13), perhaps related to the relatively small additional luminal volume (<100 mL).39 Second, the additional small intestinal water has been hypothesised to contribute to loose stool and diarrhoea; however, the maximal colonic water volume tolerated, although in healthy volunteers,40 has been shown to be much greater than the additional water induced by these FODMAPs.

Colonic gas production

The availability of non-digested and/or non-absorbed short-chain carbohydrates for colonic fermentation leads to accumulation of colonic gas including hydrogen and methane (figure 1). This is likely to lead to luminal distension, and therefore provocation of symptoms in IBS, specifically in those with visceral hypersensitivity. Table 1 summarises studies that have investigated the effect of FODMAPs on fermentation. A controlled, crossover feeding study demonstrated that a high FODMAP diet (50 g/day) led to a marked increase in 14-hours breath hydrogen production after two days compared with a low FODMAP diet (<10 g/day) in 15 patients with IBS and 15 healthy individuals,41 which was paralleled by higher symptoms scores in those with IBS. Furthermore, a recent crossover study in IBS showed that a 3-week low FODMAP diet (unknown total FODMAP dose) reduced 5-hour breath hydrogen following a lactulose challenge compared with a high FODMAP diet (unknown total FODMAP dose),42 suggesting the low FODMAP diet leads to a shift in colonic fermentation pattern independent of acute fermentable carbohydrate intake (i.e. lactulose challenge), which is likely mediated by an alteration in microbiota composition.

Quantification of the effect of fermentable carbohydrates on colonic fermentation has been elegantly demonstrated by MRI. For example, inulin challenge (40 g/day) leads to an approximate twofold greater colonic volume at 4 hours compared with glucose in healthy individuals and patients with IBS.38 39 It is also clear that there are distinct patterns of gas production elicited by individual FODMAPs. Specifically, inulin leads to a later and overall almost double the total gas production compared with fructose according to hydrogen breath testing in healthy individuals.38 This is likely due to differences in degree of absorption of individual carbohydrates and differences in GI transit time that leads to variable availability for fermentation in the proximal colon. Fermentation rates also vary between carbohydrates of different molecular geometry.32

However, recent research using MRI to investigate symptom induction in IBS challenges the assumption that people with IBS have elevated response to FODMAP ingestion (in terms of colonic gas production), as both breath hydrogen production and colonic volume kinetics were almost identical in patients compared with healthy individuals.39 Interestingly, this research also questions the extent to which increased colonic gas is responsible for symptoms induction in IBS. Patients with IBS who developed symptoms on FODMAP challenge did not, in fact, have greater colonic volume than those who do not report symptoms, suggesting that visceral hypersensitivity to luminal distension, rather than increased luminal distension per se, is key to symptom provocation during colonic fermentation.39 However, these associations were only measured in a limited subgroup of 12 patients and therefore replication of this work in a larger sample is required to verify these findings. Importantly, measurement of visceral hypersensitivity (e.g. by barostat) is also required to confirm its role in the causality of symptom provocation in response to FODMAP administration.

Other proposed mechanisms

There is preliminary evidence of mechanisms by which FODMAPs might induce symptoms in IBS, beyond small intestinal water volume and colonic fermentation. First, some FODMAPs increase GI motility. Small intestinal transit time is decreased following ingestion of a 30 g fructose-sorbitol mixture in healthy individuals,43 which further reduces opportunity for small intestinal absorption and increases availability for colonic fermentation. However, the evidence for the effect of a low FODMAP diet on transit time per se is lacking. One crossover RCT demonstrated no effect of a 3-week low FODMAP diet on whole gut transit time compared with a standard diet in patients with IBS.44 This is surprising considering the aforementioned effects of FODMAPs on small intestinal water; however, this study included both constipation and diarrhoea-predominant IBS subtypes, which may have masked potential differences between specific subtypes. The effect of FODMAP restriction on transit time, while controlling for fibre intake and other dietary factors that stimulate motility, requires clarification in large studies of specific IBS subtypes.

Second, there is now data to suggest that adherence to a low FODMAP diet is accompanied by changes in the GI microbiota and its metabolic output. The effect of a 3-week low FODMAP diet was compared with a high FODMAP diet (actual intakes unknown) in 37 patients with all IBS subtypes in a recent parallel design RCT.42 A higher abundance of the hydrogen-using genus Adlercreutzia was reported in the low FODMAP group compared with the high FODMAP group. This may have contributed to the reduction in symptoms and the diminished hydrogen response to lactulose challenge. Metabolomic analysis of urine was able to discriminate the low FODMAP and high FODMAP groups based on three key urinary metabolites including histamine, a modulator of inflammation and immune function. Several associations were also found between abundance of various taxa, the metabolome and clinical symptoms, suggesting that the observed diet-induced changes in the microbiota and metabolome may be in part responsible for clinical outcomes.

A major limitation of this study was the absence of dietary composition assessment and therefore only an arbitrary assessment of dietary adherence. However, this is the first evidence to suggest that reducing the availability of fermentable substrate in the colon impacts the metabolomic output of the microbiota, which may plausibly be involved in the generation of GI symptoms. Taxonomic correlations with clinical and metabolic markers are important in contributing to our understanding of the mechanisms and/or potential detrimental effects of the low FODMAP diet, but causality is more difficult to ascertain and longitudinal follow-up studies are needed to determine if diet-induced microbiota shifts lead to change in long-term health outcomes.

The low FODMAP diet may also reduce production of short-chain fatty acids (SCFAs). This is important, as there may be higher stool concentrations of SCFAs in IBS and an association with IBS symptomatology,45 with an animal study demonstrating SCFAs induce visceral hypersensitivity.46 Reducing luminal SCFAs could present another pathway by which the low FODMAP diet has its effect, although this has yet to be extensively studied, and findings are inconsistent. Reduction in total stool SCFAs,47 48 acetate47 49 and butyrate47 48 has been demonstrated in some studies in IBS, although no differences in SCFA between the low FODMAP diet and controls were reported in two small RCTs.44 50

Reduction in stool SCFA concentration in patients on a low FODMAP diet is likely the result of reduced availability of fermentable substrate and shifts in the abundance of taxa involved in SCFA production and/or cross-feeding reactions.51 The potential effect of diet-induced alteration in SCFAs, especially butyrate, on various factors including epithelial barrier function, risk of colorectal cancer is of significance particularly if dietary restriction is prolonged. However, current evidence regarding the effect of the low FODMAP diet on SCFA is limited by conflicting findings between studies,44 48–50 small patient numbers, differences in study design and SCFA quantification methodology. Furthermore, stool SCFA concentration is not an accurate measure of in vivo SCFA production due to the effect of colonic transit time on SCFA absorption52 and stool volume on SCFA dilution,53 leading to variations across IBS subtypes.52 Importantly, without assessment of SCFA concentration at the major site of production in the ascending colon, which requires technically demanding and invasive techniques,54 it will be difficult to confirm the interaction between the low FODMAP diet, SCFA production and symptom provocation in IBS.

Finally, others have suggested alternative mechanisms by which a low FODMAP diet improves symptoms. One RCT of patients provided low FODMAP advice in combination with other dietary advice (n=13) reported normalisation of colonic serotonin cell density after 3–9 months.55 It was suggested that this may have mediated symptom improvement by modulating transit time and visceral sensitivity. Furthermore, normalisation of stool lipopolysaccharide to levels comparable with healthy controls has been demonstrated in patients reporting a symptomatic response to a low FODMAP diet.56 A series of animal and in vitro studies using the stool supernatant of these patients showed beneficial effects on visceral sensitivity and colonic mucosal integrity.

These latter human studies are limited by very small patient numbers and/or substantial attrition rate,55 56 and the absence of the assessment of dietary composition or dietary control during the intervention period.55 56 Crucially, it is unknown whether these findings are causal or merely epiphenomena. However, they present preliminary findings suggesting that the low FODMAP diet may have effects over and above reducing luminal distension, through effecting changes in epithelial integrity and inflammatory processes that have been implicated in IBS symptomatology.57 Much research is required in this area, including in vitro studies that aim to identify discrete effects of individual FODMAPs on intestinal physiology, larger studies of homogenous IBS cohorts that identify key microbiome and metabolomic effects linked with response to the low FODMAP diet, and how these changes alter gut function.

Short-term clinical endpoints

The last decade has seen publication of numerous trials of the efficacy of the low FODMAP diet in reducing IBS symptoms. Publication of two recent systematic reviews, although with different conclusions, confirms the growing interest in the area.58 59 Many studies of the low FODMAP diet in IBS are limited due to their retrospective design,60–63 lack of control or comparator groups,64–66 and where these existed a lack of randomisation,58 the use of low FODMAP diet in conjunction with other dietary advice67 68 and the widespread lack of dietary assessment to confirm adherence. However some better-designed RCTs have been undertaken, with promising findings, many which have only recently been published, and only these will be reviewed in detail here.

At least 10 RCTs have been undertaken investigating the efficacy of a low FODMAP diet in adults with IBS (table 2). Five studies were unblinded and involved patients being randomised to low FODMAP dietary advice from a dietitian for 4–6 weeks, with all five studies reporting improvements in functional GI symptoms using various symptom scoring tools60 68–71 as well as improvement in quality of life.69 71 There are inevitable limitations with unblinded intervention studies, particularly in a disorder with a reportedly high placebo response.72

However, recently two blinded placebo-controlled studies have been performed evaluating the effect of the low FODMAP diet on IBS symptoms. In the first, a 21-day crossover feeding study that compared the low FODMAP diet with a ‘typical diet’ showed that the low FODMAP diet led to an improvement in overall GI symptoms, and resulted in a response (defined as 10 mm reduction on a visual analogue scale) in 70% of patients (response in ‘typical diet’ group not reported).73 An advantage of a feeding study such as this is the ability to carefully control dietary intake; however, controlled feeding does not mimic the real-life challenges experienced by free-living individuals undergoing dietary change. Furthermore, the crossover nature of the study raises questions regarding the minimum washout period required between interventions. Furthermore, symptoms worsened during the ‘typical diet’ control, which may have been the result of increased FODMAP intake compared with habitual diet, leading to an artificially greater difference in symptoms between groups. Nevertheless, this was the first placebo-controlled low FODMAP intervention trial and a vital contribution to the current evidence base for the low FODMAP diet in the management of functional GI symptoms.

The second placebo-controlled RCT used dietary advice to evaluate the effect of the low FODMAP diet as it would be used in clinical practice.49 In total, 104 patients with IBS were randomised to either low FODMAP dietary advice or to sham (placebo) dietary advice for 4 weeks, with lower IBS Severity Scoring Scale (IBS-SSS) scores and greater numbers experiencing a response (≥50 point reduction in IBS-SSS) in the low FODMAP group (73%) compared with sham dietary advice (42%, p=0.005) although adequate relief was reported by similar numbers in the low FODMAP (57%) group compared with sham (38%, p=0.051). This trial presents the first evidence that the low FODMAP diet is superior to placebo when administered as dietary advice.

Five RCTs have compared the low FODMAP diet with active interventions. Two randomised comparative trials have been undertaken comparing low FODMAP dietary advice with standard dietary guidelines for IBS (e.g. small frequent meals, limit caffeine and alcohol, restrict intake of fatty or spicy food).74 One study randomised 38 patients to low FODMAP advice and 37 to standard dietary advice for 4 weeks with a clinical response (≥50 point reduction in IBS-SSS) reported in 50% in the low FODMAP group and 46% in the standard advice group, with no difference between groups.75 In the second RCT, although there was no difference in proportion of patients achieving the adequate relief endpoint, there were a greater number of pain responders in the low FODMAP group (51%) compared with standard advice (23%, p=0.008), as well as a greater magnitude of response of multiple individual symptoms (abdominal pain, bloating, stool consistency frequency and urgency).70 These data on individual symptom response are consistent with previous findings of a non-randomised comparative trial that reported a significant difference in the numbers responding to low FODMAP dietary advice (76%) compared with standard National institute for Health and Care Excellence (NICE) guideline dietary advice (54%).63 One recent trial also compared the low FODMAP diet with hypnotherapy in IBS.71 Both interventions improved symptoms, although there was no difference in symptom severity between groups at 6 weeks.

Finally, two RCTs have compared symptom response to the low FODMAP diet with interventions not intended to improve symptoms. The comparator group in one study was a high FODMAP diet42 and in the other was a low FODMAP diet supplemented with fructo-oligosaccharide (fructans) to achieve a ‘normal’ FODMAP intake.48 While the primary aim of these studies was not to investigate symptom response, both found that the low FODMAP diet led to symptomatic benefit compared with the comparator interventions, both of which were high in fermentable carbohydrates.

In summary, evidence to date suggests that 50%–80% of patients with IBS report symptomatic benefit on a low FODMAP diet in the short term, and data from comparative trials suggest that it is at least as effective as general dietary and lifestyle interventions. Preconceived expectations about a treatment may prime patients to sense and record symptom outcomes differently, and this is a particular problem in IBS where outcomes are subjectively assessed, highly sensitive to participant behaviour, and where placebo effect is considerable (20%–40%).72 Therefore, recent placebo-controlled trials provide robust evidence for its clinical efficacy over placebo, and the first meta-analysis of low FODMAP RCTs reports a greater odds of reduction of abdominal pain (OR 1.81), abdominal bloating (OR 1.75) and overall GI symptoms (OR 1.81) compared with controls.59 In line with these findings, national guidelines for the dietary management of IBS in the UK now advise consideration of a low FODMAP diet where basic diet and lifestyle measures (eg, NICE guideline advice) have been unsuccessful.17 74

Long-term clinical endpoints

The majority of clinical trials of the low FODMAP diet evaluate short-term clinical endpoints (≤12 weeks). The durability of the diet over the longer term is arguably more meaningful given the chronicity of symptoms in IBS, but there is limited data available. This is particularly important given that standard practice is to systematically reintroduce FODMAPs to tolerance once symptoms are controlled. A retrospective study suggests the continued effectiveness of the diet in the longer term. Follow-up of patients consuming a FODMAP-modified diet suggests clinical benefit in 57%–74% of patients at 14–16 months,62 although the results are likely subject to significant recall bias.

The RCT previously described comparing low FODMAP advice with hypnotherapy examined the long-term symptomatic outcomes of patients.71 Lower severity scores were evident for overall symptoms in 24 patients in the low FODMAP group at 6 months compared with baseline, with an impressive 82% categorised as responders. However, the improvement in symptoms in patients was not different to 25 patients who had originally been randomised to receive hypnotherapy or combined low FODMAP diet-hypnotherapy treatment (p=0.32). Further high-quality long-term RCTs are required to clarify the longevity of symptom response in patients who receive low FODMAP dietary advice, especially following reintroduction of FODMAPs to individual tolerance. Furthermore, considering the chronicity of patients’ symptoms, the impact of long-term dietary manipulation on nutrient intake requires evaluation.

GI microbiota

Despite the clinical efficacy of a low FODMAP diet in IBS, some potentially unfavourable consequences have been reported. In particular, the low FODMAP diet leads to a considerable reduction in intake of prebiotic fructans and GOS,50 75 and therefore a sizeable reduction in substrate available for colonic fermentation. Significant dietary restriction will alter the composition and functioning of the GI microbiota. Six studies, including a number already reviewed in the clinical section, have investigated the effect of the low FODMAP diet on the gut microbiota in IBS (table 3).

Two studies, undertaken by our group, have investigated the effect of a 4-week low FODMAP diet on stool microbiota in patients with IBS. Using fluorescence in situ hybridisation, the first demonstrated that a 50% reduction in FODMAP intake led to a marked sixfold reduction in relative abundance of Bifidobacteria compared with controls who maintained their habitual diet.50 There were no differences in total bacteria or other bacterial groups such as Lactobacillus or Faecalibacterium prausnitzii, or fermentation by-products such as stool SCFA concentration or pH between groups. These findings were corroborated in a recent large, placebo-controlled RCT where low FODMAP dietary advice led to a reduction in Bifidobacteria concentration measured using quantitative PCR.49

The third study was the previously described 3-week feeding study that used quantitative PCR to show lower absolute Bifidobacteria concentration, F. prausnitzii and Clostridium cluster IV accompanied by a substantially lower total bacterial load of 47% during the low FODMAP diet compared with habitual diet.44 Diversity of Clostridium cluster XIV was greater after low FODMAP intervention compared with habitual diet, which was postulated to be due to species adaptation to altered substrate availability. This was a crossover study, and therefore the potential of carryover effects cannot be ruled out and the 80% reduction in FODMAP intake, which was achieved through total food provision, is unlikely to be reflective of the reductions achieved in dietary advice in routine practice.

In addition to these studies, two others have evaluated microbiota composition in response to FODMAP restriction in IBS using metagenomic sequencing methods.42 76 Deep analysis of microbiota composition and structure revealed no change in α-diversity (number of operational taxonomic units (OTU), ie, number of species) in children after a 1-week low FODMAP diet in a small uncontrolled study.76 Furthermore, the previously described 3-week study that reported low FODMAP-induced metabolomic alterations in adults demonstrated no effect on α-diversity or β-diversity (differences in species composition between samples at baseline and follow-up) compared with baseline.42 In the latter study, the abundance of several taxa increased (Clostridiales XIII Incertae sedis spp. and Porphyromonas spp.) and decreased (Propionibacteriaceae, Bifidobacteria) in the low FODMAP group.

Whether quantitative analysis of the microbiota could be harnessed as a predictor of response to the low FODMAP diet is an exciting proposition, and has been evaluated in a limited number of studies. In one study, increased baseline abundance of taxa such as Bacteroides, Ruminococcaceae and F. prausnizii predicted response to a 2-day low FODMAP diet in children, where response was based on pain frequency.77 This was a crossover feeding study, however, and symptom response occurred in only 24% of patients, but it did suggest patients with a higher abundance of saccharolytic microbiota may benefit the most from a reduction in dietary fermentable substrates. No such association has been demonstrated in adult patients,44 and more data are required from large parallel-arm trials.

Clearly, there is still much to understand regarding the impact of the low FODMAP diet on the GI microbiota. Whether the effect of the low FODMAP diet on the specific bacterial groups (e.g. Bifidobacteria) or overall microbiota community has identifiable downstream effects on their metabolic output or has detrimental effects on colonic health is unknown. Furthermore, it is not yet known whether the mucosal compartment is affected, whether there is a definitive effect on SCFA concentration and luminal pH in the proximal colon, or if there is a critical time point at which microbiota alterations might have functional consequences and whether these changes alter short or long-term health outcomes. The vast majority of studies thus far are descriptive observations of the microbiome and metabolome in clinical trials of the low FODMAP diet. In vitro studies will be important in guiding investigations of low FODMAP-induced alterations in the microbiome and its output in humans. Reintroduction of FODMAPs to tolerance may attenuate some of these changes, and studies of FODMAP restriction combined with probiotic or prebiotic supplementation are underway.

Nutrient intake

Most patients with IBS meet nutrient requirements,14 78 and their nutrient intake does not differ from that of healthy controls.79 The clinical effectiveness of a therapeutic diet must always be weighed against the impact it has on maintaining appropriate nutrient intake, as well as the difficulties of following the diet. In relation to macronutrients, three dietary advice studies have reported that the low FODMAP diet leads to a lower total carbohydrate intake compared with habitual diet,50 70 75 although the effect sizes are small and results in carbohydrate intakes reflect those of healthy individuals (150–200 g/day).80 One previously described study reports a considerable reduction in energy intake in patients following low FODMAP advice,75 which may be of concern particularly if the diet is followed in the longer term. However, the same change was seen in patients following standard IBS advice, and therefore it is unlikely this effect is specific to the low FODMAP diet, and may instead be due to dietary vigilance.

The 4-week study evaluating the effect of low FODMAP dietary advice compared with controls following habitual diet is the only RCT to date that has examined micronutrient intake.50 Iron intake was not different to controls, suggesting that iron-rich foods restricted on the low FODMAP diet (eg, iron-fortified cereal, pulses, nuts) were adequately substituted when advice was given by a registered dietitian. However, a lower calcium intake was reported (600 mg/day in the low FODMAP group vs 730 mg/day in controls, p=0.016), and it is likely that this was due to restriction of lactose-containing dairy products with insufficient replacement of high calcium alternatives. Measurement error due to lack of low lactose food composition data may also be a contributing factor. Nevertheless, further larger studies are required to confirm whether the intake of calcium and/or other micronutrients is compromised when patients with IBS follow a low FODMAP diet. Importantly, evaluation of nutrient intake and dietary diversification at sequential periods following FODMAP reintroduction will be important to establish whether the low FODMAP diet poses any nutritional risk in IBS in the long term.