Clinical Trial: Effect of a 28‐Day Low FODMAP Diet on Gastrointestinal Symptoms Associated With Endometriosis (EndoFOD)—A Randomised, Controlled Crossover Feeding Study

ABSTRACT

Background

Gastrointestinal symptoms affect most women with endometriosis, but therapeutic interventions are poorly defined.

Aims

To compare the effects of a 28‐day low fermentable oligo‐, di‐ and mono‐saccharides and polyols (FODMAP) or control diet on gastrointestinal symptom severity in individuals with endometriosis and poorly controlled gastrointestinal symptoms.

Method

In this single‐blinded randomised, controlled cross‐over feeding trial, we randomised women aged ≥ 18 years to 28 days of either a low FODMAP (< 5 g/day FODMAPs) or control diet (20 g/day FODMAPs), both modelled on Australian Dietary Guidelines, before a ≥ 28‐day washout and crossover to the alternate diet. The primary outcome was the proportion of responders defined according to the response in overall gastrointestinal symptoms on a 100‐mm visual analogue scale. Secondary outcomes included gastrointestinal symptoms, quality of life and psychological status.

Results

Of 35 women randomised (mean age 31; 95% confidence interval 29, 33 years), 21 (60%) responded to the low FODMAP diet compared with 9 (26%) to the control diet (p = 0.008). In the 4th week of the dietary intervention, overall symptom scores were 35 (21, 42) mm on the low FODMAP diet and 58 (55, 65) mm on the control diet (p < 0.001). Abdominal pain, bloating, stool form and quality of life for both gastrointestinal and endometriosis‐associated scales were significantly better for the low FODMAP diet compared with the control diet, but not overall for perceived stress, anxiety or depression.

Conclusions

The low FODMAP diet ameliorates gastrointestinal symptoms related to endometriosis and improves quality of life. Confirmation of these findings in a real‐world setting is required.

Trial Registration

The trial is registered with the Australian New Zealand Clinical Trials Registry (ACTRN12621000153819).

In this randomised controlled crossover feeding study, 35 women with endometriosis and poorly controlled gastrointestinal symptoms were randomised to either a supplied low FODMAP diet for 4 weeks or a control diet. Participants washed out for 4 weeks, and crossed‐over to the alternate dietary intervention. Overall, 60% of participants responded to a low FODMAP diet, compared to 26% who responded to the control diet. Overall gastrointestinal symptom severity was lower on the low FODMAP diet, and these improvements were accompanied by improvements in abdominal pain, bloating and quality of life.

1. Introduction

Endometriosis is a common inflammatory gynaecological condition characterised by the presence of endometrial‐like tissue outside the uterus [1]. Endometriosis affects 1 in 7 Australian women aged up to 49 years [2] and has a profound negative impact on quality of life (QOL) [1]. While characteristic symptoms include dysmenorrhoea, pelvic pain and dyspareunia, gastrointestinal symptoms, such as abdominal pain, bloating, diarrhoea and dyschezia, are extremely common, affecting > 75% of individuals [3, 4, 5]. These symptoms typically intensify during menstruation [6] and occur irrespective of bowel infiltration of the disease [5]. Frustratingly for sufferers, most endometriosis clinical guidelines fail to mention gastrointestinal symptoms or recommend targeted treatments, instead focusing on the management of pain, treating infertility, slowing the growth of endometrial tissue and preventing recurrence [7, 8].

Patients with endometriosis frequently report receiving a label of irritable bowel syndrome (IBS) and are three times more likely to meet Rome criteria for IBS than the general population [9]. While this indicates an epidemiological association between functional gastrointestinal symptoms and endometriosis, it is important to note that Rome criteria have not been validated in women with endometriosis, and furthermore, that the aetiology of gastrointestinal symptoms associated with endometriosis remains unknown. These limitations cast doubt on the reliability of symptom‐based diagnostic tools, such as the Rome criteria, to identify IBS in women with endometriosis. Given the preponderance of gastrointestinal symptoms in endometriosis, attributing these to co‐morbid IBS oversimplifies their aetiology, which is more likely related to distinct but as yet poorly understood pathophysiological mechanisms.

Despite the high prevalence of reported gastrointestinal symptoms in endometriosis and considerable impact on QOL, a treatment gap exists. Few treatments specifically target gastrointestinal symptoms, and some medications, such as progesterone, opiates, non‐steroidal anti‐inflammatory drugs and tricyclic antidepressants, can make them worse [4]. Unfortunately, while pain and menstrual symptoms may respond to such therapies [1] gastrointestinal symptoms often remain refractory.

Not surprisingly, treatment dissatisfaction and discontinuation rates are high [10], and up to half of individuals with endometriosis adopt self‐management strategies such as diet [3, 11, 12]. However, the quality of evidence supporting any given dietary approach in endometriosis is extremely limited, with most dietary studies in women with endometriosis evaluating only dietary supplements, not whole‐diet interventions; few have examined the effect on gastrointestinal symptoms, and no previous dietary randomised controlled trials (RCTs) [13, 14, 15]. A diet therapy that has received some attention is the low fermentable oligo‐, di‐ and mono‐saccharide and polyol (FODMAP) diet, for which the initial phase of restricting all FODMAPs has support from multiple RCTs in patients with IBS [16, 17]. Observational studies have also suggested that restriction of dietary FODMAPs might benefit patients with endometriosis and gastrointestinal symptoms [6, 18]. However, this therapeutic diet is not without risk, and while it has been argued that data from IBS studies could be extrapolated to those suffering gastrointestinal symptoms associated with their endometriosis, as a different condition with distinct pathophysiological mechanisms, formal evaluation of efficacy should occur in this patient group. To date, the benefit of FODMAP restriction in endometriosis has not been subject to placebo‐controlled RCT evaluation.

Hence, the aim of the current study was to address the hypothesis that restriction of all FODMAPs, as applied in the first phase of the low FODMAP diet, ameliorates gastrointestinal symptoms in patients with endometriosis and poorly controlled gastrointestinal symptoms. A randomised, controlled crossover feeding study was performed in patients with endometriosis and poorly controlled gastrointestinal symptoms by comparing the effects of 28 days of a low FODMAP diet with a control diet, based on Australian Dietary Guidelines, on the severity of gastrointestinal symptoms. Effects on quality of life and psychological symptoms were also assessed.

2. Methods

2.1. Study Design

This single‐blinded randomised, controlled cross‐over feeding trial was conducted at Monash University in Melbourne, Australia, and approved by the Monash University Human Research Ethics Committee (25358). The protocol was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12621000153819).

2.2. Participants

Between June 2020 and February 2023, women living in Melbourne (location restricted for logistical reasons related to food transport), aged 18 years to menopause, with a diagnosis of endometriosis confirmed via laparoscopy and/or ultrasound, were invited to participate via the websites and social media pages of endometriosis consumer organisations only. To ensure participants were blinded to the nature of the dietary interventions, the study was not advertised on any FODMAP‐related websites or social media pages, and study advertisements, information and consent materials stated the study was comparing two different diets but did not disclose the nature of either dietary intervention. Given the controversy as to the pathophysiology of gastrointestinal symptoms in endometriosis, patients were deliberately not required to meet Rome criteria for IBS. This ensured we were able to encompass the full spectrum of gastrointestinal complaints reported in endometriosis sufferers. Instead, potential participants with endometriosis who answered ‘no’ to the screening question—‘Over the past week, were your bowel and gut symptoms adequately controlled’?—were invited to a screening interview with the study gastroenterologist (REB) to confirm eligibility, exclude alternative pathologies and obtain written, informed consent. Participants were required to have a diagnosis of endometriosis confirmed by either ultrasound or laparoscopy. This was confirmed by the gastroenterologist in the screening interview although the original investigation reports were not sighted. Exclusion criteria were irregular periods, laparoscopic surgery within the past 3 months, other gastrointestinal disorders, food allergies and/or special diets, uncontrolled physical or mental illness and inability to read or understand English. The study gynaecologist (RR) was consulted with questions regarding eligibility as needed.

2.3. Randomisation and Masking

A randomisation sequence providing a 50:50 ratio without blocking was independently generated using www.randomizer.org. During the baseline, participants consumed their habitual diet before researchers assigned participants (using the randomisation sequence) to one of the two diets for 28 days. Participants and the statistician (but not researchers) were blinded to group assignment. Participants were also blinded to the nature of the dietary interventions.

2.4. Study Protocol

To control for the known effect of the menstrual cycle on gastrointestinal symptoms [6], baseline data were collected for 7 days preceding a period due date, and intervention diets were commenced on day 1 of a menstrual cycle, then followed for 28 days, unless this was hormonally suppressed (Figure 1). During a 7‐day baseline, participants completed daily food and symptom diaries to record food/beverage intake and gastrointestinal symptoms. On day 7 of the baseline, participants completed several questionnaires and rated overall gastrointestinal symptoms over the past 7 days on a 100 mm visual analogue scale (VAS). Those scoring > 30 mm were randomised to one of two dietary interventions for 28 days. After completing diet 1, participants washed out for ≥ 28 days (returning to their habitual diet), before crossover to the alternate 28‐day diet on the first day of a menstrual cycle (Figure 1). During each dietary intervention, food and symptom diaries were collected daily, and questionnaires were completed on day 28.

FIGURE 1.

Overview of the trial design. Gastrointestinal (GI). Visual analogue scale (VAS). Patient‐Reported Outcomes Measurement Information System (PROMIS). Quality of life (QOL). Gastrointestinal Quality of Life Index (GIQLI). Endometriosis Health Profile Questionnaire (EHPQ). Depression Anxiety Stress Scale‐21 (DASS 21).

2.5. Dietary Interventions

Both dietary interventions were modelled on the Australian Dietary Guidelines [7] to meet recommended food group servings. To aid blinding and adherence, most food was supplied to participants along with a detailed, 7‐day meal plan that repeated four times over the 28‐day dietary intervention. The 7‐day meal plans were matched between the two diets for meal type, pattern and nutritional content (energy, macronutrients and fibre); thus, the supplied diets were nutritionally similar, except for the content of FODMAPs (Method S1). The two diets were: (a) a ‘Control’ diet in which the FODMAP content was not manipulated; and (b) a ‘low FODMAP diet’ in which high‐FODMAP foods were swapped for nutritionally equivalent, low‐FODMAP alternatives to achieve a total FODMAP intake of < 5 g/day, consistent with previous studies [8, 9]. Both intervention diets contained gluten and were low in lactose. The nutritional composition of the diets was analysed using FoodWorks (Xyris Software Pty Ltd) and the FODMAP Calculator (www.monashfodmapcalculator.com.au).

Meals were cooked fresh, weighed, individually portioned, vacuum‐sealed and frozen in a commercial kitchen under the supervision of a research chef. To ensure blinding, all food was supplied in plain packaging. Meals were also supplied free of charge and delivered frozen to participants prior to each diet, along with storage, defrost, reheat and serving instructions. In addition to the supplied diet, participants were provided with a small, detailed list of low‐FODMAP staple foods and perishable items that they were required to self‐cater. Participants were discouraged from deviating from the supplied meal plan but were given a list of extra foods for hunger and dining out options if needed. The study dietitian (JV) was contactable for questions related to the study diets.

2.6. Dietary Adherence

Participants were classified as adherent/non‐adherent by comparing FODMAP intake each day with the dose of FODMAPs in the supplied diet on the corresponding day. To monitor adherence to the dietary intervention, participants were instructed to complete a daily food diary throughout the 28‐day dietary intervention, recording food and beverage intake (including details about recipes, brands, cooking methods and quantities consumed). Each day, participants were arbitrarily considered ‘adherent’ if they consumed at least 60% of the supplied FODMAPs and no more than 40% above the supplied amount of FODMAPs. Daily adherence was tallied over 28 days, and the proportion of days adherent was calculated.

2.7. Measures

Gastrointestinal symptoms were recorded in three ways.

• IBS‐VAS: This validated tool measures the severity of symptoms on a 100‐mm visual analogue score (VAS), where 0 represents no symptoms and 100 represents the worst possible gastrointestinal symptoms [1]. This was scored daily for overall and individual gastrointestinal symptoms and was also used to rate overall gastrointestinal symptoms over the past 7 days on the final day of the baseline and each dietary intervention. A change in VAS > 20 mm was arbitrarily considered clinically significant as previously applied [19, 20, 21].

• NIH PROMIS Gastrointestinal Symptom Scale: This was completed on the final day of the baseline and each diet. These validated symptom scales collect data on the severity, frequency and impact of specific gastrointestinal symptoms, and responses are used to calculate a T‐score, which is a standardised score from 0 to 100, in which 50 is the mean of a relevant reference population (general US population who reported at least 1 gastrointestinal symptom) and 10 is the standard deviation (SD) of that population. Higher scores indicate more of the parameter being measured (e.g., diarrhoea, constipation or bloating). A clinically meaningful change in T score is considered 0.5–0.6 SD or 5–6 points on the T‐score scale. T‐score cut points have also been assigned to the scales to indicate severity, whereby scores between 50 and 60 are considered mild, 60 and 70 are moderate and scores > 70 are severe [4].

• Stool frequency and consistency: All bowel actions were recorded and rated for stool consistency using the Bristol Stool Chart (BSC), which is a validated 7‐point scale that participants use to categorise stools passed according to a pictorial stool chart. Although not diagnostic, generally stool types 1–2 indicate constipation, types 3–5 indicate normal stool and types 6–7 indicate diarrhoea [2].

Participants completed the Women's Health Symptom Survey Questionnaire at baseline only in order to characterise the study population. This is a non‐validated instrument that collects information about medical and menstrual history, use of contraception, pregnancy history, family history and symptoms of endometriosis.

Quality of life was assessed using two instruments:

• The Gastrointestinal Quality of Life Index (GIQLI) questionnaire: This assessed quality of life in relation to gastrointestinal health. This validated questionnaire includes 36 items belonging to 1 of 5 dimensions—symptoms (19 items), emotions (5 items), physical function (7 items), medical treatment (1 item) and social function (4 items). The symptom dimension relates to the impact of unpleasant gastrointestinal symptoms such as gas/wind, bloating and abdominal pain; the physical dimension relates to the impact of tiredness, disrupted sleep and physical strength; the psychological dimension relates to feelings of stress, anxiety and frustration; the social dimension relates to impacts on social life and relationships; and the medical treatment dimension focuses on the impact of medical treatment. The questionnaire produces a summary score ranging from 0 to 144, with higher scores indicating better health‐related quality of life (HRQOL). This summary score has undergone validation and is reliable, whereas the individual dimensions have not undergone validation [5].

• The Endometriosis Health Profile Questionnaire (EHPQ): This is a validated, disease‐specific tool that measures HRQOL. The questionnaire includes a core component that consists of 30 items applicable to all women with endometriosis, covering dimensions such as pain, control and powerlessness, emotional wellbeing, social support and self‐image. The modular questionnaire includes 23 items applicable to some women, including dimensions such as work, relationships with children, sexual relationships, feelings about the medical profession, treatment for endometriosis and problems conceiving. Questions are asked about the impact of endometriosis over the past 4 weeks and are answered on a 5‐point Likert scale, from 0 (never) to 4 (always). Raw scores for each dimension are summed and transformed to a 0–100 scale, with higher scores indicating worse HRQOL [3].

Mental health was assessed with the Depression Anxiety Stress Scale‐21 (DASS 21), which is a validated, self‐report questionnaire that measures the severity of symptoms related to depression, anxiety and stress. The DASS‐21 consists of 21 items that belong to one of 3 dimensions—depression, anxiety and stress. Items are measured on a 4‐point Likert scale, from 0 = did not apply to me at all to 3 = applied to me very much or most of the time. Participants were asked to rate how much each item related to them over the past 7 days. Raw scores for each dimension are summed, with higher scores reflecting more severe symptoms for the respective dimension. The maximum possible score for each dimension is 21. Recommended cut‐off values to define those with abnormal scores are 9 for depression, 7 for anxiety and 14 for stress [6].

2.8. Outcomes

The primary outcome was the proportion of responders at the conclusion of the intervention, defined as those reporting a decrease of > 20 mm VAS‐IBS from baseline to the end of each dietary intervention or, if the baseline score was ≤ 50 mm, VAS‐IBS < 30 mm. The VAS used for the primary outcome measured overall gastrointestinal symptom severity over the past 7 days and was assessed on the final day of the baseline and each diet. Secondary outcomes included comparisons of dietary intake across the intervention diets, adherence to the dietary interventions and changes from baseline to the end of each diet for overall and individual gastrointestinal symptoms, stool consistency, the time course of symptom change, QOL and psychological symptoms (Table S2).

2.9. Statistical Analysis

Power calculation was based on quantitative data from studies in IBS in which a 100‐mm VAS had been applied [20]. The standard deviation in overall symptoms was conservatively estimated to be 25 mm [20]. Assuming this, 27 participants were needed to detect a change in overall gastrointestinal symptoms with 90% power and a significance level (adjusted for sidedness) of 0.05 if the true difference on the VAS was 20 mm. With a ~30% drop‐out rate, 35 participants were required.

The primary endpoint was analysed as intention‐to‐treat (ITT). Comparison of end‐of‐intervention endpoints between the interventional dietary periods or between baseline and end‐of‐intervention was analysed per protocol (PP) without data interpolation. Summary data are presented as mean or median with 95% confidence intervals according to normality of distribution. Analyses of continuous endpoints (EHPQ, PROMIS‐GI, GIQLI and daily logs) were performed using a linear mixed‐effects model. Fit was restricted by maximum likelihood using the lme4 package (version 1.1–30) across intervention periods, which were modelled as fixed effects, with individual participants and arm order modelled as random effects. Where model residuals were not normally distributed, data were normalised via log transformation, but were presented as non‐transformed values. Multiple comparisons between the baseline, control and low FODMAP diet were made using the multcomp package (version 1.4–15) with no post hoc corrections made. Analyses of categorical endpoints were performed using the chi‐square test, with the risk ratio calculated using Fisher's exact test. Average symptoms over time during dietary interventions were compared by a 2‐way analysis of variance. Scores from the DASS‐21 questionnaire were examined by Friedman's test across diets, and pairwise comparisons were made using the Wilcoxon signed rank test. A p‐value ≤ 0.05 was considered statistically significant, except where a Bonferroni correction was made for multiple comparisons. R statistical software (version 4.0.2; R Foundation for Statistical Computing, Vienna, Austria) and GraphPad Prism (version 10.0.2; GraphPad Software, San Diego) were used for analyses.

3. Results

3.1. Participants

Between 1 December 2020 and 28 February 2023, 233 participants were screened, 193 were excluded for ineligibility, 35 were randomised and 25 completed the study. Dropouts occurred in diet 1 (n = 2), washout (n = 2) and diet 2 (n = 6). Participant flow is described in Figure 2. Table 1 describes demographic and clinical data for the randomised participants.

FIGURE 2.

Participant flow through the study. Intention to treat (ITT). Per protocol (PP).

TABLE 1.

Demographic and clinical characteristics of the randomised participants at baseline.

Characteristic			N a
Demographics	Age, years, mean (95% CI)		35	31 (29, 33)
	Body mass index, kg/m2, mean (95% CI)		35	27 (24, 29)
	Current smoker, n (%)		35	2 (6%)
Endometriosis‐associated features, n (%)	Contraception used in the past 3 months		35	25 (71%)
	Heavy periods		25	12 (48%)
	Severe period pain		25	16 (64%)
	Pelvic pain (other than with periods or intercourse)		35	30 (86%)
	Pain during or after sexual intercourse		35	23 (66%)
	Pain improved or stopped following bowel movement		25	19 (76%)
	Previous hospital admissions for pain		30	10 (30%)
Gastrointestinal symptoms (visual analogue scale), mm	Overall symptoms		35	51.10 (46.8, 55.4)
	Abdominal pain		35	52.3 (48.2, 56.3)
	Bloating		35	55.6 (51.5, 59.6)
Days with satisfactory control of gastrointestinal symptoms in 1 week, %, mean (95% CI)			35	58 (48, 69)
Mental health (DASS‐21)	Depression	Score, median (IQR) Stress	35	10 (11) 21 (60%)
		n (%) abnormal (> 9)		21 (60%)
	Anxiety	Score, median (IQR) Stress	35	14 (13)
		n (%) abnormal (> 7)		20 (57%)
	Stress	Score, median (IQR) Stress	35	14 (14) 17 (49%)
		n (%) abnormal (> 14)		17 (49%)
Quality of life	GIQLI scores–overall		35	77.8 (73.0, 82.7)
	EHPQ‐30–overall		35	55.1 (49.6, 60.6)

Abbreviations: DASS, Depression Anxiety Stress Scale‐21; EHPQ, Endometriosis Health Profile Questionnaire; GIQLI, Gastrointestinal Quality of Life Index.

^a N represents participants who provided answers to the relevant question.

3.2. Dietary Intake

There were no statistically significant differences in the macronutrient and fibre intake between the two interventional diets (Table 2). Total FODMAP intake was similar between baseline and control diets, although individual FODMAPs did differ. Intake of all FODMAPs (excluding lactose) was lower on the low FODMAP diet compared to the control diet. Lactose intake was minimal on both diets.

TABLE 2.

Comparison of daily nutritional intake across the study (n = 35).

Nutrient	Diet a			p (linear mixed models)
	Baseline	Control b	Low FODMAP	Baseline vs. control	Baseline vs. low FODMAP	Control vs. low FODMAP
Energy, kJ/day	8363 (7292, 8829)	8547 (8141, 8711)	8268 (7873, 8554)	0.62	0.79	0.47
Protein, g/day	89 (78, 98.7)	67 (64, 70)	70 (67, 73)	< 0.001	< 0.001	0.96
Fat, g/day	82 (71, 89.8)	78 (74, 79)	80 (72, 83)	0.38	0.38	0.98
Saturated fat	33 (26, 35.2)	32 (30, 34)	30 (28, 31)	0.50	0.44	0.17
Carbohydrate a g/day	193 (175, 218.0)	239 (227, 257)	239 (219, 255)	< 0.001	0.002	0.38
Dietary fibre a g/day	23 (20, 25.9)	26 (25, 29)	24 (22, 28)	0.037	0.81	0.072
FODMAPs (total) g/day	12.0 (10.1, 16.9)	19.9 (18.3, 20.3)	4.0 (3.3, 4.6)	0.047	< 0.001	< 0.001
Excess fructose b	1.5 (0.8, 2.3)	10.4 (9.8, 11.2)	1.7 (1.1, 2.0)	< 0.001	0.56	< 0.001
Lactose	5.1 (3.1, 9.8)	0.4 (0.4, 0.7)	0.4 (0.4, 0.5)	< 0.001	< 0.001	0.58
Polyols	1.1 (0.8, 1.5)	4.4 (3.7, 5.1)	0.6 (0.5, 0.6)	< 0.001	< 0.001	< 0.001
Sorbitol	0.5 (0.3, 0.9)	3.6 (3.0, 4.4)	0.4 (0.3, 0.4)	< 0.001	0.007	< 0.001
Mannitol	0.3 (0.2, 0.7)	0.8 (0.6, 0.9)	0.2 (0.1, 0.2)	< 0.001	< 0.001	< 0.001
Oligosaccharides a	3.6 (3.2, 3.7)	4.1 (3.7, 4.4)	1.3 (1.2, 1.5)	0.008	< 0.001	< 0.001
Fructans a	2.9 (2.2, 3.2)	3.2 (2.9, 3.6)	1.1 (0.9, 1.2)	0.009	< 0.001	< 0.001
Galacto‐oligosaccharides	0.7 (0.6, 0.8)	0.8 (0.7, 0.9)	0.3 (0.3, 0.3)	0.33	< 0.001	< 0.001

Note: Data are shown as median (95% confidence intervals) and compared by linear mixed models. Differences were considered statistically significant where p ≤ 0.006 after correction for multiple comparisons.

^a Data log‐transformed for analysis unless otherwise indicated.

^b Excess fructose = fructose–glucose.

Participants adhered to the low FODMAP diet on 75% (95% confidence intervals 67%, 83%) of days compared with 71% (59, 83) of days on the control diet. In the absence of a gold‐standard method of measuring or interpreting dietary adherence, these rates were considered acceptable.

3.3. Gastrointestinal Symptoms

Of the 35 randomised patients, 21 (60%) responded to the low FODMAP diet compared with 9 (26%) on the control diet (p = 0.008), with participants more likely to respond to the low FODMAP diet than the control diet (RR 0.26, 0.09, 0.71; p = 0.015; Figure 3). The PP analysis (n = 25) revealed consistent results: 18 (72%) responded to the low FODMAP diet compared with 8 of 25 (32%) who responded to the control diet (p = 0.01) with a higher likelihood of responding to the low FODMAP diet (RR 0.18, 95% CI: 0.05, 0.62; p = 0.01).

FIGURE 3.

Proportion of participants who responded to the control diet versus low FODMAP diet. Responders are defined as participants reporting a decrease of > 20 mm VAS‐IBS from baseline to the end of each dietary intervention and/or VAS‐IBS < 30 mm at the conclusion of the intervention.

At the conclusion of the intervention, overall gastrointestinal symptoms were 35 (21, 42) mm on the low FODMAP diet and 58 (55, 65) mm on the control diet (p < 0.001) (Figure 4A). Compared with overall gastrointestinal symptoms at a baseline of 64 (61, 67) mm, participants improved overall scores by 26 (19, 34) mm on the low FODMAP diet and 3 (0, 15) mm on the control diet (Figure 4B). Overall gastrointestinal symptoms averaged over each week of the intervention diets improved over the 4 weeks to a greater extent on the low FODMAP diet than the control diet (n = 27; p = 0.034; 2‐way ANOVA), and the plots diverged within 2 weeks (Figure 4C).

FIGURE 4.

The effect of dietary periods on gastrointestinal symptoms. (A) Overall gastrointestinal (GI) symptom severity, as assessed via a 100‐mm visual analogue scale (VAS) on day 7 baseline versus day 28 of the control diet or low FODMAP diet arms. Bars represent median values per intervention period, dots represent individual values per participant, and lines chart the difference in individual scores. Differences were considered significant where p ≤ 0.017. (B) Change in overall gastrointestinal symptom severity on day 28 of each dietary intervention compared to the day 7 baseline. Differences were considered significant where p ≤ 0.05. (C) Comparison of the mean of weekly averages of overall gastrointestinal symptoms as assessed daily via a 100‐mm visual analogue scale (VAS). Results across the time points for the low FODMAP diet and control diet were statistically different (p = 0.034; 2‐way analysis of variance). NS, not significant.

Carry‐over (order) effects were absent, with no differences in overall gastrointestinal symptoms between the first and second dietary interventions, regardless of allocation (data not shown). When only data from the first‐consumed diet were compared, 9/16 (56%) responded to the low FODMAP diet compared with 5/19 (26%) in the control (p = 0.09) in the ITT analysis. End‐of‐study overall symptoms were 40 (28, 53) mm compared with 57 (47, 67) mm (p = 0.035), and scores reduced by 24% (11, 38) compared with 6% (−4, 16) on the low FODMAP diet and control diets, respectively (p = 0.020).

For individual gastrointestinal symptoms, participants reported less severe abdominal pain/discomfort and bloating on the low FODMAP diet compared with the baseline and the control diet, while wind and satisfaction with stool consistency improved on the low FODMAP diet only relative to the baseline (Figure 5). No improvements in tiredness/lethargy or nausea were seen.

FIGURE 5.

Gastrointestinal symptoms were assessed via visual analogue scale (VAS) daily and consolidated over the 7 days of the baseline and the final 7 days of the control diet and the low FODMAP diet (n = 25) for (A) abdominal pain/discomfort; (B) abdominal bloating; (C) wind; (D) tiredness and lethargy; (E) nausea; (F) satisfaction with stool consistency. Bars represent median values per intervention period; dots represent individual values per participant. Differences were considered significant where p < 0.006. NS, Not significant.

NIH PROMIS Gastrointestinal Symptom Scales showed no changes in gastrointestinal symptoms on the control diet compared with baseline. However, on the low FODMAP diet, there were clinically meaningful improvements in diarrhoea, abdominal pain and gastro‐oesophageal reflux (Table 3, Figure S3).

TABLE 3.

Comparison of gastrointestinal symptoms measured by the NIH PROMIS gastrointestinal symptom scales across the study (n = 25).

Symptom	T‐score, median (95% CI)			p
	Baseline	Control	Low FODMAP	Baseline vs. control	Baseline vs. low FODMAP	Control vs. low FODMAP
Diarrhoea	55.1 (53.4, 57.0)	54.3 (51.7, 57.6)	45.3 (39.9, 53.4)	0.923	0.028	0.001
Swallowing	45.4 (40.3, 49.5)	44.4 (40.3, 51.5)	40.3 (40.3, 47.5)	0.976	0.143	0.025
Bloating	61.8 (60.9, 64.2)	61.0 (58.4, 64.4)	57.6 (55.8, 59.4)	0.416	< 0.001	< 0.001
Nausea	57.7 (52.9, 59.4)	52.4 (50.6, 56.6)	53.8 (49.7, 55.9)	0.214	0.209	0.966
Constipation a	58.1 (56.2, 60.1)	54.9 (52.8, 58.0)	53.5 (51.2, 55.5)	0.099	< 0.001	0.035
Reflux	51.7 (42.9, 54.2)	44.6 (40.3, 51.6)	40.4 (37.0, 47.0)	0.385	< 0.001	0.028
Belly pain	64.3 (61.5, 68.4)	62.4 (58.7, 66.7)	59.3 (53.1, 60.9)	0.090	< 0.001	0.006

Note: Data are presented as T‐scores, where 50 is the mean of the reference population and 10 is the standard deviation (SD) of that population. A clinically meaningful change in T score is considered 0.5–0.6 SD, or 5–6 points on the T‐score scale. Data are shown as median (95% confidence interval) and were analysed via a linear mixed model. Differences were considered significant where p ≤ 0.007 after correction for multiple comparisons. N represents participants who provided data for the specific question. Patient‐Reported Outcomes Measurement Information System (PROMIS).

^a Data log‐transformed for analysis unless otherwise indicated.

Participants reported normal stool form (BSC 3–5) on 43% of days at baseline compared with 56% of days on the control diet (p = 0.181) and 71% of days on the low FODMAP diet (p < 0.001; Figure 5). Participants also reported loose stool (BSC 6–7) less often on the low FODMAP diet compared with the baseline (p = 0.002) and control diet (p = 0.01). No differences in the number of bowel movements per day were observed (Figure 6).

FIGURE 6.

Comparison of stool characteristics according to the Bristol Stool Chart (BSC) between baseline, the control diet and the low FODMAP diet (n = 25). Data collected daily at baseline were consolidated over the 7 days, and comparisons were made with daily data collected over the final 7 days of each dietary intervention. Bars represent median proportions and were analysed via linear mixed models, visually shown relative to 100%. Differences were considered significant where p ≤ 0.017. NS, not significant.

3.4. Psychological Symptoms

Abnormal baseline scores for depression (> 9), anxiety (> 7) and stress (> 14) were observed in 49%–60% of participants at baseline (Table 1). When all participants were considered in the ITT analysis, no differences across baseline, control and low FODMAP diet were seen for depression (p = 0.14; Friedman's test) or anxiety scores (p = 0.20). However, pairwise comparisons revealed that depression scores tended to reduce on the low FODMAP diet compared with baseline (Figure 7). When only patients with abnormal baseline scores were evaluated, a statistically significant reduction of depression (p = 0.009) and a trend for reduced anxiety (p = 0.029) were seen with the low FODMAP diet. Stress scores differed across the diets (p = 0.004), with pairwise comparisons showing these were lower on the low FODMAP diet compared with the baseline diet for all participants (Figure 7). When only those with abnormal stress scores at baseline were considered (n = 13), stress scores were lower on both the low FODMAP (p = 0.005) and control diets (p = 0.01) than at baseline, but no differences were observed between diets (p = 0.24).

FIGURE 7.

Paired ratings and their differences for depression, anxiety and stress, measured via the Depression, Anxiety and Stress Scale (DASS‐21) for the intention‐to‐treat cohort during baseline, low FODMAP and control diets. Values considered within normal limits are denoted by the green shading. The paired comparisons were performed with the Wilcoxon signed‐rank test, and p‐values are shown if ≤ 0.05. Using the Bonferroni correction for comparisons across 3 dietary groups, p‐values < 0.017 were considered statistically significant.

3.5. Quality of Life

Overall GIQLI scores were higher (indicating better QOL) on the low FODMAP diet compared with the baseline and control diet, with no changes observed on the control diet (Figure 8). Individual dimensions for symptoms and physical function improved on the low FODMAP diet, while other dimensions were unaffected by either dietary intervention (Table 4).

FIGURE 8.

Comparison of Gastrointestinal Quality of Life Index (GIQLI) scores on day 7 of the baseline and day 28 of the control diet and the low FODMAP diet. Bars represent median scores per intervention period; dots represent individual values per participant. Differences were considered significant where p ≤ 0.01. NS, not significant.

TABLE 4.

Comparison of Gastrointestinal Quality of Life Index (GIQLI) scores across the study (n = 25).

	Median (95% CI)			p
	Baseline	Control	Low FODMAP	Baseline vs. control	Baseline vs. low FODMAP	Control vs. low FODMAP
Symptoms	48.0 (45.0, 51.0)	51.0 (48.0, 54.0)	56.0 (53.0, 59.0)	0.091	< 0.001	0.003
Emotions	9.0 (8.0, 10.0)	10.0 (9.0, 12.0)	10.0 (9.0, 11.0)	0.149	0.044	0.498
Physical function	10.0 (8.0, 12.0)	10.0 (8.5, 11.5)	14.0 (10.0, 15.0)	0.817	0.002	0.009
Medical treatment	3.0 (2.0, 3.0)	3.0 (2.0, 3.0)	3.0 (2.5, 4.0)	0.616	0.204	0.094
Social function	7.0 (6.0, 7.0)	7.0 (6.0, 7.5)	7.0 (6.0, 8.0)	0.301	0.306	0.947
Overall HRQOL a	75.0 (72.0, 83.0)	81.0 (77.0, 86.0)	88.0 (78.0, 96.0)	0.100	< 0.001	0.003

Note: Data collected daily at baseline were consolidated over the 7 days, and comparisons were made with daily data collected over the final 7 days of participation across each arm. Data are shown as median (95% confidence interval) and were analysed via a linear mixed model. Differences were considered significant where p ≤ 0.01.

^a Health‐related quality of life; data log‐transformed for analysis.

EHPQ‐30 scores improved on the low FODMAP diet compared to baseline. Improvements in overall QOL compared to control were driven primarily by improvements in the impact of pain (Table 5, Figure S4). EHPQ‐23 scores for the impact of work improved on the low FODMAP diet compared with baseline, but not compared to the control diet (Table 6, Figure S5).

TABLE 5.

Comparison of Endometriosis Health Profile questionnaire core questionnaire (EHPQ‐30) scores across the study (n = 25).

	Diet			p
	Baseline	Control	Low FODMAP	Baseline vs. control	Baseline vs. low FODMAP	Control vs. low FODMAP
Pain	47.7 (45.5, 54.5)	40.9 (25.0, 47.7)	22.7 (18.2, 36.4)	0.001	< 0.001	0.005
Control & powerlessness	66.7 (58.3, 79.2)	50.0 (45.8, 68.8)	41.7 (29.0, 54.2)	< 0.001	< 0.001	0.045
Emotional well‐being	50.0 (41.7, 58.3)	37.5 (29.2, 47.9)	33.3 (25.0, 41.7)	0.026	< 0.001	0.078
Social support	56.2 (43.8, 68.8)	53.1 (43.8, 56.2)	43.8 (31.2, 56.2)	0.12	0.001	0.13
Self‐image	58.3 (50.0, 75.0)	54.2 (45.8, 70.8)	50.0 (50.0, 58.3)	0.080	< 0.001	0.12
Total a	55.6 (48.0, 62.0)	47.1 (37.8, 56.7)	45.3 (29.6, 50.9)	0.027	< 0.001	0.028

Note: Data collected daily at baseline were consolidated over the 7 days, and comparisons were made with daily data collected over the final 7 days of participation across each arm. Data are shown as median (95% confidence interval) and were analysed via linear mixed models. Differences were considered significant where p ≤ 0.01.

^a Data log‐transformed for analysis.

TABLE 6.

Comparison of Endometriosis Health Profile questionnaire modular questionnaire (EHPQ‐23) scores across the study.

	N	Diet			p
		Baseline	Control	Low FODMAP diet	Baseline vs. control	Baseline vs. lowFODMAP diet	Control vs. low FODMAP diet
Work	29	50.0 (35.0, 60.0)	40.0 (25.0, 50.0)	25.0 (10.0, 42.5)	0.023	< 0.001	0.032
Relationships with children a	7	50.0 (25.0, 62.5)	50.0 (0.0, 50.0)	25.0 (0.0, 43.8)	0.098	0.070	0.958
Sexual intercourse	34	61.2 (50.0, 80.0)	52.5 (44.4, 75.0)	60.0 (47.5, 70.0)	0.936	0.373	0.469
Medical profession	25	43.8 (25.0, 56.2)	56.2 (37.5, 68.8)	43.8 (31.2, 50.0)	0.551	0.556	0.219
Treatment for endometriosis	24	58.3 (50.0, 75.0)	54.2 (25.0, 58.3)	50.0 (8.3, 66.7)	0.034	0.012	0.975
Problems conceiving	8	78.1 (37.5, 100.0)	25.0 (18.8, 93.8)	81.2 (75.0, 93.8)	0.506	0.912	0.258

Note: Data collected daily at baseline were consolidated over the 7 days, and comparisons were made with daily data collected over the final 7 days of participation across each arm. Data are shown as median (95% confidence interval) and were analysed via linear mixed models. Differences were considered significant where p ≤ 0.008. N represents participants who provided data for the specific question. Note that participants were requested only to answer questions that were relevant to them.

^a Data log‐transformed for analysis.

3.6. Adverse Events

One serious adverse event was reported. This participant developed appendicitis during washout (after the control diet) requiring hospitalisation and laparoscopic removal of the appendix, which was spotted with endometrial tissue. Researchers subsequently withdrew this participant from the study, although the event was considered unlikely to be related to the dietary intervention. Two participants developed gastroenteritis‐like symptoms whilst consuming the control diet. After consultation with the study gastroenterologist (RB), both agreed to recommence the protocol following a ≥ 2‐week break and upon recovery, but 1 subsequently withdrew citing life stress.

4. Discussion

Numerous dietary approaches are used to manage symptoms of endometriosis, but until now, efficacy data were lacking. This single‐blinded RCT provides the first high‐quality data to show that a low FODMAP diet reduces gastrointestinal symptoms associated with endometriosis. Compared with a control diet modelled on healthy eating guidelines, over twice as many responded to the low FODMAP diet compared with the control diet, and the low FODMAP diet was associated with clinically significant improvements in overall gastrointestinal symptoms in 6 out of 10 participants. Abdominal pain and bloating improved, loose stool consistency normalised, QOL improved when measured in relation to gastrointestinal symptoms and the impact of endometriosis, and depression improved in those with clinically significant depression scores prior to dietary intervention.

On most measures, participants were markedly symptomatic on their habitual diet at baseline, with VAS scores for overall gastrointestinal symptoms of > 50 mm in nearly all participants. Improvements in gastrointestinal symptom severity occurred by week 2, at which point a waning of the placebo effect was observed on the control diet. Symptom improvements continued on the low FODMAP diet, albeit more slowly, and by the end of the intervention, clinically meaningful differences in overall and individual gastrointestinal symptom severity were seen compared with both paired baseline symptoms and symptoms on the control diet. These responses mimicked those observed during the pivotal feeding study in IBS [20]. Abdominal pain and bloating also improved, even when the relatively harsh Bonferroni correction was applied to avoid errors due to multiple comparisons. Furthermore, the NIH PROMIS scales confirmed the clinically meaningful changes in multiple gastrointestinal symptoms on the low FODMAP diet.

The low FODMAP diet had a normalising effect on stool form in participants with loose stool. This effect is consistent with findings from a meta‐analysis that examined the effect of a low FODMAP diet in diarrhoea‐predominant IBS [17]. Having controlled for dietary fibre intake, it is plausible that these changes were driven by reductions in osmotic load and/or a slowing of colonic transit time on the low FODMAP diet—both of which factors are known to influence stool form [22, 23].

The low FODMAP diet also had positive effects on measures of QOL when measured in relation to gastrointestinal symptoms and symptoms of endometriosis. These observations are consistent with IBS studies showing improvements in QOL when gastrointestinal symptoms improve on a low FODMAP diet [24], and endometriosis studies showing that self‐management strategies are associated with greater personal agency, improved health outcomes and better QOL [25, 26]. It is plausible that enabling participants to improve their gastrointestinal symptoms through diet provided a sense of empowerment and control, decreasing feelings of helplessness and increasing QOL [27].

About half of the participants had abnormal depression, anxiety and stress scores at baseline, and in those participants, intervention effects were more pronounced. On the low FODMAP diet, feelings of depression reduced and anxiety tended to reduce among participants with abnormal scores at baseline. Stress was impacted the most, but no diet‐specific difference was noted, suggesting that stress is affected by the provision of food and/or changes in diet quality but not by alterations in FODMAP intake. This is in contrast to studies in IBS, where psychological distress reduces when gastrointestinal symptoms improve on a low FODMAP diet. Perhaps the lack of a clear effect in this study indicates that in endometriosis, psychological symptoms are more affected by the vast array of other symptoms potentially present [28].

Given the preponderance of inadequately controlled gastrointestinal symptoms among women with endometriosis [3] and the absence of targeted treatments [1, 7], findings from the current study have the potential to close a treatment gap and positively impact the lives of individuals living with this condition. Gastrointestinal symptoms are over‐represented in women with endometriosis compared to control populations, a phenomenon frequently attributed to comorbid IBS. However, this observation is invalid since the criteria used to diagnose IBS were generated from populations who did not have endometriosis. Investigations such as motility testing and bowel imaging will not validate such a concept. It was for this reason that the frequency of Rome IV‐defined IBS was not calculated. The fact that the low FODMAP diet was effective in reducing many, but not all, abdominal and pelvic symptoms is suggestive that IBS and chronic abdominal symptoms in patients with endometriosis might share common pathogenic pathways. With these issues in mind, future studies are warranted: (1) to evaluate the real‐world application and longer‐term effectiveness of the low FODMAP diet in ameliorating gastrointestinal symptoms, (2) to identify factors that might influence the response to the intervention (such as treatment use and disease stage/severity or gastrointestinal microbiome alterations), (3) to determine the impact on other pain‐related endometriosis symptoms and (4) to clarify mechanisms of action. The effectiveness of the diet in the longer term would require assessment after phase 2 (reintroduction) and phase 3 (personalisation) of the diet.

Strengths of this study included the carefully designed, supplied diets that were nutritionally matched except for FODMAPs. Meal plans were also matched to aid blinding, and confounding was limited by controlling for the known effect of menstrual cycles on gastrointestinal symptoms. Dietary adherence was optimised by the provision of most food and confirmed in our analysis, which showed adequate rates of adherence. The crossover design minimised potential confounders such as the inter‐individual heterogeneity of symptom patterns within the menstrual cycle and of potential pathogenic influences by, for example, the gut microbiota. Carry‐over effects of the first‐received diet were minimised by the long washout period. Furthermore, order effects were not evident in the results, and analysis of the effects after only the first‐received diet showed similar efficacy of the low FODMAP diet as observed in the crossover analysis.

However, there were six key limitations with the study. First, the dose of FODMAPs supplied to participants on the control diet may have been too high. While the control diet was designed to represent a ‘typical’ FODMAP intake, limited data describe habitual FODMAP intake in Australia [29]. Despite this, there was little evidence that symptoms worsened or that placebo responses were lost with the control diet. Second, while the study gastroenterologist confirmed verbally with participants that they had previously been diagnosed with endometriosis via ultrasound or laparoscopy, investigation reports were not sighted and the date of diagnosis was not routinely collected. Nevertheless, previous studies have shown that self‐reporting a prior diagnosis of endometriosis is extremely accurate in women with endometriosis, and hence this methodology is not expected to significantly affect outcomes [30]. Third, the issue of whether two separate conditions—endometriosis and IBS—were indeed present was not contemplated for the reasons already outlined. It is poor science to assume that IBS treatments are effective in this distinct population. Instead, research that specifically evaluates IBS treatments in women with endometriosis is badly needed if we are to close the treatment gap and facilitate evidence‐based recommendations in clinical guidelines regarding the management of gastrointestinal symptoms associated with endometriosis. Fourth, despite considerable effort and attention paid to ensuring that participant information and the design of the dietary interventions provided few clues as to the nature of the nutritional difference between the dietary interventions, the effectiveness of blinding was not assessed. Of course, asking participants to correctly identify the nature of the dietary interventions would be biased when the major outcome is symptom severity. Fifth, the effects on endometriosis‐specific symptoms (such as pelvic pain, dysmenorrhea and dyspareunia) were not specifically measured but should be considered in future research. Finally, as a dietary feeding study that recruited participants from metro‐Melbourne, the generalisability of findings to other settings and real‐world scenarios might be questioned. In practice, people would implement a low FODMAP diet either unsupervised or guided by a dietitian, under which circumstances dietary adherence would likely be lower, as individuals face real‐world challenges that compromise adherence, such as motivation, understanding the prescribed intervention, food availability and convenience [31].

5. Conclusion

Gastrointestinal symptoms affect the vast majority of patients with endometriosis and negatively impact QOL but are poorly targeted by existing pharmacological and surgical treatments. As such, many individuals with endometriosis complement standard therapies with non‐evidence‐based dietary approaches in an attempt to manage symptoms and take charge of their illness [11]. This single‐blinded RCT provides high‐quality data supporting the use of a low FODMAP diet to improve gastrointestinal symptoms and QOL in patients with endometriosis. Further research is required to assess the impact of the reintroduction and personalisation phases of the FODMAP diet on long‐term outcomes.

Author Contributions

Jane E. Varney: investigation, funding acquisition, conceptualization, writing – original draft, methodology, visualization, writing – review and editing, project administration, data curation. Daniel So: formal analysis, writing – review and editing, visualization. Peter R. Gibson: writing – review and editing, formal analysis, visualization. Dakota Rhys‐Jones: project administration. Yuet Sang Jimmy Lee: project administration. Jane Fisher: funding acquisition, conceptualization, writing – review and editing. Judith S. Moore: conceptualization, investigation, funding acquisition, project administration. Roni Ratner: writing – review and editing, project administration. Mark Morrison: writing – review and editing, funding acquisition. Rebecca E. Burgell: conceptualization, investigation, funding acquisition, writing – review and editing, project administration. Jane G. Muir: conceptualization, investigation, funding acquisition, writing – review and editing, resources, supervision.

Ethics Statement

This study was approved by the Monash University Human Research Ethics Committee (ID 25358).

Consent

Written informed consent was obtained from all participants, including consent for the publication of anonymised data.

Conflicts of Interest

J.G.M., J.E.V. and P.R.G. work in a department that financially benefits from the sales of a digital application (Monash University FODMAP diet app), patient booklets, cookbooks and online courses, all of which relate to the low FODMAP diet therapy. J.G.M. and J.E.V.'s salaries are paid by the above‐stated FODMAP commercial activities. P.R.G. and R.E.B. are consultants or advisory board members for Anatara and Atmo Biosciences and have received speaker honoraria from Dr Falk Pharma. R.E.B. has also received speaker honoraria from Bayer. P.R.G. is also a consultant or advisory board member for Topas and Comvita. P.R.G. has received research grants for investigator‐driven studies and/or speaker honoraria from Atmo Biosciences and Mindset Health. P.R.G. and D.S. are shareholders in Atmo Biosciences. D.S. has received consulting and speaking honoraria from Procter & Gamble and is an employee of Atmo Biosciences. Research funding for this project came from a competitively awarded grant from the Australian Government's Medical Research Future Fund. The funder had no role in study design, data collection, analysis, interpretation or report writing. No other relationships or activities could appear to have influenced the submitted work.

Supporting information

Appendix S1.

Data Availability Statement

Deidentified data will be shared on application and consideration by the authors.