Correspondence Address correspondence to: Ami D. Sperber, MD, MSPH, Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer-Sheva, Israel.
Department of Health Research Methods, Evidence and Impact, McMaster University, Hamilton, Ontario, CanadaPopulation Health Research Institute, McMaster University, Hamilton, Ontario, Canada
Center for Functional GI & Motility Disorders, University of North Carolina, Center for Education and Practice of Biopsychosocial Care, and Drossman Gastroenterology, Chapel Hill, North Carolina
Department of Gastroenterology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
Lynda K. and David M. Underwood Center for Digestive Disorder, Gastroenterology and Hepatology, Houston Methodist Hospital and Weill Cornell Medical College, Houston, Texas
Laboratory of Liver, Pancreas and Motility (HIPAM), Unit of Research in Experimental Medicine, Faculty of Medicine, Universidad Nacional Autónoma de Mexico (UNAM). Hospital General de México, Mexico City, Mexico
Integrative Functional Gastroenterology Research Center, Department of Internal Medicine, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran
Department of Internal Medicine, Universidade Federal do Rio Grande do Sul, Gastroenterology Division, Hospital de Clínicas de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brazil
Department of Gastroenterology, University Hospital Vall d’Hebron, Autonomous University of Barcelona & Neuro-Inmuno-Gastroenterology Lab, Vall d’Hebron Research Institute, Barcelona, Spain
Division of Gastroenterology, Department of Internal Medicine, Faculty of Medicine, Universitas Indonesia/Cipto Mangunkusumo Hospital, Jakarta, Indonesia
Department of Biochemical Markers of Chronic Non-Communicable Diseases Research National Medical Research Centre for Preventive Medicine, Moscow, Russia
Although functional gastrointestinal disorders (FGIDs), now called disorders of gut-brain interaction, have major economic effects on health care systems and adversely affect quality of life, little is known about their global prevalence and distribution. We investigated the prevalence of and factors associated with 22 FGIDs, in 33 countries on 6 continents.
Methods
Data were collected via the Internet in 24 countries, personal interviews in 7 countries, and both in 2 countries, using the Rome IV diagnostic questionnaire, Rome III irritable bowel syndrome questions, and 80 items to identify variables associated with FGIDs. Data collection methods differed for Internet and household groups, so data analyses were conducted and reported separately.
Results
Among the 73,076 adult respondents (49.5% women), diagnostic criteria were met for at least 1 FGID by 40.3% persons who completed the Internet surveys (95% confidence interval [CI], 39.9–40.7) and 20.7% of persons who completed the household surveys (95% CI, 20.2–21.3). FGIDs were more prevalent among women than men, based on responses to the Internet survey (odds ratio, 1.7; 95% CI, 1.6–1.7) and household survey (odds ratio, 1.3; 95% CI, 1.3–1.4). FGIDs were associated with lower quality of life and more frequent doctor visits. Proportions of subjects with irritable bowel syndrome were lower when the Rome IV criteria were used, compared with the Rome III criteria, in the Internet survey (4.1% vs 10.1%) and household survey (1.5% vs 3.5%).
Conclusions
In a large-scale multinational study, we found that more than 40% of persons worldwide have FGIDs, which affect quality of life and health care use. Although the absolute prevalence was higher among Internet respondents, similar trends and relative distributions were found in people who completed Internet vs personal interviews.
Functional gastrointestinal disorders (FGIDs, or disorders of gut–brain interaction) place an economic burden on healthcare systems and reduce quality of life, but little is known about their worldwide prevalence or distribution.
New Findings
In a large-scale multi-national study, the authors found that more than 40% of persons worldwide have FGIDs. Similar trends and relative distributions were found in people who completed internet vs personal interviews.
Limitations
Study participants completed questionnaires over the internet or by in-person interviews; further studies of the worldwide prevalence of FGIDs, where possible with confirmation, are indicated.
Impact
FGIDs are common in all regions of the world. Proportions of persons with irritable bowel syndrome are lower when the Rome IV criteria are used, compared with the Rome III criteria.
The functional gastrointestinal disorders (FGIDs), or disorders of gut-brain interaction (DGBIs), are gastrointestinal (GI) disorders related to any combination of motility disturbance, visceral hypersensitivity, altered mucosal and immune function, altered gut microbiota, and altered central nervous system processing.
However, population-based cross-sectional surveys have not satisfactorily delineated their actual prevalence. Published studies have involved highly variable diagnostic criteria, study populations, questionnaires, and data collection methods.
A global perspective on irritable bowel syndrome: a consensus statement of the World Gastroenterology Organisation Summit Task Force on irritable bowel syndrome.
For irritable bowel syndrome (IBS) and functional dyspepsia (FD), the 2 most researched disorders, reported prevalence estimates are very broad (1.1%–45.0% for IBS,
Thus, given the large methodological heterogeneity, it is inappropriate to pool individual prevalence rates, and we are left with an unanswered question as to whether the differences in prevalence rates seen among individual countries in prior surveys reflect genuine differences between populations or are due to methodological differences between studies.
The ideal global epidemiological study would use uniform methodology to assess nationally representative populations of sufficient size throughout the world, but this is not feasible. The present study, conducted in 33 countries at the same time, did use standardized methodology (although circumstances mandated 2 different data collection methods), with identical diagnostic questions to approximate to that ideal, assessed the global prevalence and burden of FGIDs, including sub-analyses by country, sex, and age groups. The results are summarized for all FGIDs, but the main focus is on 5 prevalent disorders because they are the most researched of the FGIDs and the most salient for clinicians: IBS, FD, functional constipation (FC), functional diarrhea, and functional bloating/distention.
The overall aims of this global study were to conduct an extensive multinational epidemiological study of all the FGIDs that are assessible by self-report, to obtain reliable and precise regional and local estimates of FGID prevalence, and to collect data on numerous potentially associated factors that might explain differences in FGIDs among populations and generate hypotheses to advance understanding of their pathophysiology.
Secondary aims included the development of a database that could serve as a source of data mining and be integrated with other similar databases in the future, and to establish a network of FGID experts with a track record of research collaboration on a global scale.
The present paper focuses on classic epidemiological findings: prevalence rates by country and geographical region, by age and sex, and preliminary indicators of burden of disease. As a descriptive study, there are no a priori hypotheses, so no hypothesis testing was conducted.
Methods
The study was conducted in 33 countries (Figure 1): Argentina, Australia, Bangladesh, Belgium, Brazil, Canada, China, Colombia, Egypt, France, Germany, Ghana, Holland, India, Indonesia, Iran, Israel, Italy, Japan, Malaysia, Mexico, Nigeria, Poland, Romania, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Turkey, the UK, and the US. This country selection, based on the availability of interested country principal investigators, provided a good global coverage, except for Africa (represented only by Egypt, South Africa, Ghana, and Nigeria) and the Middle East, especially Arab countries (Egypt only). As seen in Figure 1, data were collected by Internet survey only in 24 countries, by personal interview only in 7 countries and by both methods in 2 countries (see later in this article).
Figure 1Global map showing study countries, colored-coded by data collection method: Internet, household interviews, or both. The Internet survey was conducted in 26 countries and the household survey in 9 countries, 2 (Turkey and China) used both methods, totaling 33 countries in all.
A minimum of 2000 individuals were surveyed in each country, in both the Internet and household surveys. In India and China, the minimum number of individuals in the surveys was increased to allow for the size of the national populations. In Japan the sample size was raised to 2500 because the sex ratio among the first 2000 participants was higher among men in some age groups. We recruited an additional 500 participants, primarily women, to achieve a more balanced sex ratio. Thus, the final study population was larger than originally anticipated at 70,000 (33 countries with 2000 individuals each, plus double surveys in China and Turkey). The predefined demographic parameters for all countries were 50% female and 50% male individuals, and 40% for 18 to 39 years, 40% for 40 to 64 years, and 20% for 65+ years.
In countries where most adults use the Internet, a secured online survey (accessible only to preselected invited participants) was conducted using population samples provided by a professional company (Qualtrics, LLC, Provo, UT) who awarded participant points redeemable for gifts. These surveys were anonymous, nationwide, and had built-in quality-assurance measures to exclude poor-quality responders, including 2 attention-check questions, a completion-speed check, and repeat questions to detect inconsistent responders. The software ensured that there were no missing answers to compulsory questions, and had automated skip patterns, resulting in complete and accurate symptom pattern information.
In countries in which an Internet survey was unfeasible, usually because of poor Internet coverage, personal interviews were conducted in probability samples of individuals (1 per household) in selected villages and cities, without national representation. The household survey countries were Bangladesh, Ghana, India, Indonesia, Iran, Malaysia, and Nigeria. In the case of Iran, the Internet infrastructure was sufficient for an Internet survey, but Qualtrics, Inc. did not have access to a pool of potential subjects in that country as it did in the other countries where the survey was conducted by Internet. Residents of the participating villages were invited to meetings where the study was explained. They were encouraged by civic and religious community leaders to participate and were offered a 1-time free medical consultation in return. In China and Turkey, we collected data with the household methodology and the Internet survey, resulting in a household survey dataset from a total of 9 different countries. Unlike the other household surveys, the household study in Turkey was conducted nationwide after the Internet study had been completed, so we achieved a similar geographical, sex, and age distribution as the Internet survey, with interview responses captured directly into electronic devices, eliminating the problems with incomplete or missing responses found in other household surveys.
FGID Case Definitions
The survey included the complete Adult Rome IV Diagnostic Questionnaire
and a self-report checklist of organic diseases and surgeries that can cause gastrointestinal symptoms, to identify FGID cases. Twenty-two FGID diagnoses were assigned according to Rome IV criteria, based on responses to the Rome IV Diagnostic Questionnaire. Individuals who otherwise met Rome IV FGID criteria were excluded from FGID case definition if they self-reported a medical history that could represent organic or structural reasons for the symptoms. For example, subjects reporting celiac disease, GI cancer or inflammatory bowel disease (Crohn’s disease or ulcerative colitis) were excluded from all Rome IV FGID diagnoses. Subjects with a history of peptic ulcer disease were excluded from esophageal, gastroduodenal, and biliary diagnoses. Finally, subjects who reported diverticulitis or bowel resection were excluded from bowel and anorectal disorders. Because no independent medical evaluation was done, this exclusion method may have eliminated individuals who did not have a functional GI disorder.
For household countries, where a proportion of cases had missing responses to diagnostic questions on the 22 FGIDs, these cases were excluded from prevalence analysis for all the FGIDs (N = 4087) leaving a final total of 18,949. This was necessary because several Rome FGID diagnoses overlap and the determination of whether a person warrants a particular diagnosis may depend on whether criteria for one or more other FGIDs are met.
The survey also included an 80-item supplemental questionnaire on sociodemographic characteristics, medical and health history, comorbid symptoms and conditions, GI infections, health care utilization, medications, childhood and current living conditions, psychosocial variables, diet, QOL, and culture and religion (Supplemental Table 1). It incorporated validated questionnaires, such as the Patient Health Questionnaire-15,
were included in all 9 household survey countries and in 14 of the 26 Internet countries (Belgium, Brazil, Canada, China, Egypt, France, Germany, Holland, Israel, Japan, Mexico, Russia, Singapore, and Turkey) to compare IBS prevalence between Rome III and Rome IV criteria. The reason we did this for IBS and not for all FGIDs was that its criteria underwent the most substantial change between Rome III and IV and the length of the study questionnaires reached a limit that could not be expanded. Including all the Rome III questions for the other FGIDs would have increased the study questionnaire by about 50%.
We used 2 proxy variables to assess the burden of FGIDs: (1) health care utilization; ie, history of (a) physician consultation about bowel problems and (b) frequency of doctor visits per year for any health problems, and (2) QOL scores on the Patient-Reported Outcomes Measurement Information System (PROMIS) Global-10 questionnaire (range 4–20).
As part of the PROMIS, this questionnaire is a publicly available global health assessment tool that measures symptoms, functioning, and health care–related QOL for a wide variety of chronic diseases and conditions.
The study questionnaire underwent translatability assessment by a professional company (TransPerfect, Inc, New York, NY),
and was then translated by the same company into 21 languages with linguistic validation (cognitive debriefing). Each country principal investigator monitored this process to ensure that the translated questionnaires were linguistically valid and culturally adapted for their country. Where appropriate, the translations were “localized,” for example, the original English questionnaire was translated into Spanish for Mexico and then localized for Colombia, Argentina, and Spain.
Statistical Considerations
In a descriptive study, when estimating prevalence rates, sample size considerations are guided by the desired precision in the 95% confidence intervals. When estimating unknown prevalence rates, the most conservative approach (ie, the one that provides the largest variance and thus the widest confidence intervals), assumes a prevalence of 0.50. In this study, we chose the minimum sample size of 2000 participants per country to obtain high precision of within-country estimates of prevalence rates. Thus, 95% confidence intervals for prevalence rates as high as 0.50 would range within ± 0.022; and less common diagnoses (smaller prevalence rates) would have higher precision (narrower confidence intervals).
We calculated country-specific prevalence rates for all major FGID diagnoses by sex and age groups. Prevalence rates were pooled across countries using Yang’s meta-prevalence method,
which combines separate population survey prevalence estimates into an overall meta-prevalence estimate. Because of substantial differences in data collection methodology between the Internet and household methods, global pooling was done within survey type only.
Ethical review was completed for all countries. The study was approved or exempted from ethics board oversight (the latter for Internet survey countries, where subjects were anonymous to the investigators). All survey participants completed a written consent form, either electronically (Internet surveys) or on paper (household surveys).
Results
The survey was completed by 73,076 respondents; 36,148 women (49.5%) and 36,928 men (50.5%). The numbers of women by survey group were 26,576 respondents (49.1%) in the Internet countries and 18,949 (50.5%) in the household countries. We successfully achieved equal sex distribution and pre-planned age ranges in most countries with both surveying methods.
We do not have full data on response rates. In the Internet surveys, panels of registered country-specific survey-takers were contacted through e-mail until all quota categories were filled. There is no way to know how many saw those e-mails or how many were reached, so response rates are not available.
We have full response rates for Bangladesh, Malaysia, both surveyed regions in India, and Iran but the number of subjects approached in the other household sites was not tracked comprehensively. In Bangladesh the response rate was 99.5%, in Malaysia 92.8%, in northern India 99.2%, in southern India 99.0%, and in Iran 97.8%.
The sample demographics, by survey method, appear in Supplemental Table 2. All Internet survey countries met the minimum sample size (≥2000) and equal sex (50%:50%) parameters. In 6 Internet countries, the age group distribution was not fully met due to the inability to enroll sufficient numbers in the 65+ age group. In these countries, there is limited Internet access or lower usage. In Egypt, women were underrepresented, possibly due to lower Internet use or culture. Based on the US Census Bureau classification for rural communities (fewer than 2500 residents),
9.7% of the participants lived in rural communities in Internet countries and 43.3% in household countries.
The household surveys achieved the minimum target sample size of at least 2000 completed interviews, but the quality of the data were lower than in the Internet survey, particularly in Ghana (1190 records valid for analysis), Indonesia (1231), and Nigeria (1442). The total number of respondents who would have met the criteria for FGID diagnoses but were classified as non-FGID due to reporting organic diseases or a GI surgery was 4094 (7.6%) in Internet surveys and 748 (4.0%) in household surveys.
The prevalence results are presented below in accordance with GI tract anatomic regions, corresponding to the order of the questions in the Rome IV Diagnostic Questionnaire. The results for all the FGIDs appear in Table 1. These results are discussed in a more comprehensive and overlapping context in the Discussion section, later in this article. The country distribution of having any FGID is shown in Figure 2.
Table 1Pooled Prevalence Rates (% and 95% CI) for 22 Rome IV FGIDs, in a Combined Population-based Internet Survey Sample of 54,127 Individuals in 26 Countries and in a Combined Household Survey Sample of 18,949 Individuals in 9 Countries
The prevalence rates of 5 selected major FGIDs compared across all the countries surveyed are shown in Table 2 and Figure 3, to provide a more detailed view of the variance of these disorders globally.
Table 2Prevalence Rates (% and 95% CI) for 5 Selected Major Functional Gastrointestinal Diagnoses (Rome IV)– for Any FGID (26 Countries) and Rome III IBS (14 Countries) in the Internet Survey and for All 9 Countries in the Household Survey
Figure 3Distribution of country-specific (circles) and pooled (boxes) prevalence rates for 5 selected major FGIDs in the countries surveyed by Internet (N = 26) and household interviews (N = 9) with Rome IV criteria.
The most prevalent esophageal disorder in Internet and household surveys was functional dysphagia, with pooled prevalence rates of 3.2% (3.0%–3.3%) and 1.2% (1.0%–1.3%), respectively. The rates for functional heartburn, reflux hypersensitivity, and esophageal chest pain were substantially lower. All esophageal disorders were more prevalent among women in both survey methods. However, there was a divergence in results in terms of age, with decreasing rates in the older age groups in the Internet countries but increasing rates with age in the household countries.
Gastroduodenal Disorders
FD was the most prevalent gastroduodenal disorder, with a pooled prevalence rate of 7.2% (7.1%–7.4%) for Internet and 4.8% (4.5%–5.1%) for household surveys. In the Internet surveys, the subtype distribution was 66.6% postprandial distress syndrome (PDS), 15.3% epigastric pain syndrome (EPS), and 18.1% overlapping PDS/EPS. In the household countries, the subtype distribution was 59.5% PDS, 28.1% EPS and 12.4% overlapping PDS/EPS. FD rates varied widely between countries, from 2.2% in Japan to 12.3% in Egypt in the Internet surveys and from 0.7 (0.5–1.0) in India to 19.4 (17.7–21.2) in Bangladesh in the household surveys.
Women had higher mean FD rates in the Internet surveys than men, with an odds ratio (OR) of 1.6 (1.5–1.7) for overall FD, 1.6 for PDS (1.5–1.7), and 1.4 (1.3–1.6) for EPS. FD and its 2 subtypes were most common among young adults and decreased steadily in prevalence across the adult life span.
Functional Bowel Disorders
The most prevalent bowel disorder in both survey types was FC, with pooled rates of 11.7% (11.4%–12.0%) and 6.6% (6.3%–6.9%) for Internet and household surveys, respectively. Other prevalent disorders were functional diarrhea at 4.7% (4.5%–4.9%) and 1.2% (1.0%–1.3%), IBS at 4.1% (3.9%–4.2%) and 1.5% (1.3%–1.7%), and functional abdominal bloating/distention at 3.5% (3.3%–3.6%) and 1.2% (1.0%–1.3%), respectively.
The prevalence rates of IBS among Internet survey countries ranged from a low of 1.3% (0.8%–1.8%) in Singapore to 7.6% (6.4%–8.7%) in Egypt (Table 2 and Figure 3). However, most of the countries (19 of 26) had IBS rates between 3% and 5%. The outliers besides Singapore and Egypt were Japan (2.2%) China (2.3%), Russia (5.9%), South Africa (5.9%), and the United States (5.3%). Twenty-four of the 26 countries had prevalence rates between 2% and 6%, with Singapore and Egypt as outliers. In the household countries, IBS prevalence ranged from 0.2% (0.1%–0.3%) in India to 4.6% (3.7%–5.5%) in Bangladesh, and the variance was greater than in the Internet countries (Table 2 and Figure 3). The pooled prevalence rates for IBS were substantially higher among women in both survey methods, with a female-to-male OR of 1.8 (1.7–2.0) for the Internet and 2.0 (1.5–2.5) for the household countries. IBS prevalence decreased with age in the Internet surveys, from 5.3% (5.0%–5.6%) to 3.7% (3.5%–4.0%) to 1.7% (1.4%–1.9%), whereas it increased with age in the household group from 1.4% (1.1%–1.7%) to 1.5% (1.2%–1.7%) to 1.9% (1.4%–2.4%).
As a group, the functional bowel disorders were the most prevalent of all GI regions, with 33.4% (33.0%–33.8%) of the 54,127 Internet participants and 16.0% (15.5%–16.5%) of the 18,949 household participants having at least 1 of those 6 disorders.
Centrally Mediated Abdominal Pain Syndrome and Biliary Pain
There were almost no cases of either of these diagnostic entities. The rate for centrally mediated abdominal pain syndrome was 0.02% (n = 9) for the Internet survey and 0.05% (n = 9) for the household survey. The corresponding rates for biliary pain were 0.08% (n = 44) and 0.03% (n = 5), respectively.
Anorectal Disorders
In the Internet surveys, 7.7% (7.5%–8.0%) of subjects met criteria for at least one anorectal disorder, compared with 2.6% (2.3%–2.8%) in the household surveys. In both cases, the most prevalent disorder was proctalgia fugax at 5.6% (5.4%–5.8%) in the Internet surveys and 1.5% (1.4%–1.7%) in the household surveys.
Comparison of IBS Prevalence by Rome IV and Rome III Diagnostic Criteria
In the 14 Internet countries where Rome III questions were included, the overall IBS prevalence was 3.8% (3.6%–4.0%) by Rome IV criteria and 10.1% (9.8%–10.5%) by Rome III criteria (Table 3). Rome IV IBS rates were substantially lower than Rome III in all countries, ranging from 24% to 57% of Rome III IBS prevalence rates. In the 9 household countries, the pooled IBS prevalence rates were 1.5% (1.3%–1.7%) using Rome IV and 3.5% (3.3%–3.8%) using Rome III, with Rome IV prevalence rates ranging from 18% to 75% of Rome III IBS prevalence rates. In this group of countries, the prevalence increased for both criteria with increasing age.
Table 3A Comparison of Pooled Prevalence Rates (% and 95% CI) for Rome III and Rome IV Diagnostic Criteria for IBS in 14 Internet Countries (N = 29,606) and 9 Household Countries (N = 18,949)
In contrast to the household surveys, by both criteria, IBS rates were lower on average in older individuals in the Internet surveys. Women had substantially higher IBS rates than men in all age groups by both criteria: Rome III OR 1.7 (1.6–1.9) and Rome IV OR 1.7 (1.5–1.9).
The overall Rome IV IBS subtype distribution was 28.7% IBS-D, 32.4% IBS-C, 32.4% IBS-M, and 6.5% IBS-U in the 26 Internet countries, and 28.8% IBS-D, 37.9% IBS-C, 17.2% IBS-M, and 16.1% IBS-U in the 9 household countries.
Individuals with Rome IV IBS had higher mean IBS-SSS severity scores, 250 (244–256) vs Rome III IBS 191 (187–194), in the Internet countries. In the household countries, individuals with Rome IV IBS had an IBS-SSS severity score of 174 (158–190) vs Rome III IBS 134 (124–144).
Burden of FGIDs
In the Internet surveys, individuals with FGIDs were more likely than others to be high-frequency medical consulters, with one or more doctor visits per month for any health problem [(OR=1.7 (1.6–1.8)], but this difference was not seen in the household surveys [(OR=1.0 (0.9–1.2)] (Table 4). Also, individuals who met Rome IV criteria for any FGID were more likely to have visited doctors at any time in the past because of bowel problems than those with no FGID: 46.5% vs 27.1% in the Internet, and 26.1% vs 12.0% in the household survey. The same applied for each of the 5 selected major FGIDs.
Table 4Comparison of PROMIS-10 Quality of Life Scores (Physical And Mental) for Patients With and Without at Least One FGID– and Rates of Doctor Visits for Bowel and Any Other Health Problems in the 26 Internet Survey Countries and in the 9 Household Survey Countries
N
PROMIS-10 Physical (mean, 95% CI)
PROMIS-10 Mental (mean, 95% CI)
Has visited a doctor for bowel problem (%, 95% CI)
One or more doctor visits/month for any health problem (%, 95% CI)
Health-related QOL was lower on the PROMIS Global-10 questionnaire for individuals with any FGID compared with subjects with no FGID for global mental and global physical scores, in both Internet and household surveys. Most participants reported QOL scores in the middle of the possible range of scores, with little variability.
Discussion
This is the first global study of the epidemiology and impact of the FGIDs (DGBIs). By assessing large population samples from 33 globally distributed countries using the same survey instruments and statistical analyses, we can provide a meaningful picture of FGIDs around the world. The study methodology was rigorous, especially for the 26 countries surveyed via the Internet, where we not only achieved predetermined parameters for sample size, sex, and age distribution, but also a national distribution that reflected closely the actual geographical population distributions.
Several findings are noteworthy from the results presented previously: (1) the overall rate for meeting at least 1 FGID diagnosis was generally consistent between countries within each sampling method, with a pooled mean of 40.3 (39.9–40.7) in Internet and 20.7 (20.2–21.3) in household countries. The rate of having any FGID was exceptionally low in the household surveys in Turkey and India, at less than 10%. (2) The prevalence of having any FGID was higher among women than men, with an OR of 1.7 (1.6–1.7) in Internet and 1.3 (1.3–1.4) in household countries. When surveyed via the Internet, 46.5% of the entire adult female population across the 6 continents surveyed met the diagnostic criteria for 1 or more of the FGIDs, supporting previous findings that FGIDs are more prevalent in women than in men. The corresponding figure for the household surveys was 23.1%. Our data show that the female predominance of FGIDs is present for FGIDs in all regions of the GI tract, from the esophagus to the rectum, and with both Internet and household survey methodologies. (3) IBS prevalence rates by Rome IV were lower than in most studies using previous versions of the Rome criteria
and generally half or less of Rome III prevalence rates in the same countries. This is in line with a recently published study in the United States, Canada, and the United Kingdom, that used a similar study methodology.
(4) Rome IV IBS rates were similar among most of the Internet countries, with 19 of the 26 having prevalence rates between 3% and 5%. Singapore and Egypt were clear outliers at 1.3% and 7.6%, respectively. In the household countries, the prevalence was more variable, ranging from 0.2% in India to 4.6% in Bangladesh. As has been reported previously, we found women to have higher rates of IBS than men. We also found the sexes to have a different IBS subtype pattern: among women the rate of IBS-C is higher than IBS-D, whereas among men this is reversed. (5) In the Internet countries, FGID prevalence decreased with age, but there was an opposite trend seen in the household countries. (6) FGID prevalence rates for the household countries were consistently lower than the Internet countries. This also holds true for Turkey, even though their household survey methodology was much more similar to Internet surveys than in other household survey countries. Notably, the pattern of relative prevalence among the various disorders was consistent among the various FGIDs among all countries. (7) A few disorders, such as functional dysphagia, rumination, and proctalgia fugax, have prevalence rates that are higher than might be expected in light of clinical experience. It is possible that as we deepen our analyses of the entire database, some patterns may evolve that we are unaware of at the present.
Our findings on the cumulative presence of FGIDs (meeting diagnostic criteria for at least 1 FGID) are congruent with a previous study of the US population with the original Rome criteria (when there were 20 FGIDs) published in 1993,
which found 69% of US adults to have any FGID. The high prevalence in both cases is simply the result of evaluating simultaneously the presence of many disorders in the same population samples, yielding a comprehensive picture of the vast scope of FGIDs as a societal health problem.
Among the Internet-surveyed countries, the prevalence rates for many of the FGIDs, and IBS in particular, were quite similar and the variance was low. This illustrates that IBS and the other FGIDs are truly worldwide disorders. It also shows that the Rome IV diagnostic questionnaire can identify these disorders across geographic regions and in numerous translations. Yet, as has been reported previoiusly,
there is variance among the countries in the prevalence of these disorders. There are several potential explanations for this variability, including cultural differences, social reporting sensitivity, ethnic diversity, genetics, and dietary habits. We are confident that the differences found in this study are not due to differences in study methodology, which was uniform within Internet and household surveys. Thus, in light of the rigorous and uniform research methodology we applied, we believe that the observed prevalence rates do accurately reflect differences among countries and are variable enough to warrant further investigation into reasons for differences between countries and regions,
and their association with potential predictive factors covered in our supplemental questionnaire. Such analyses can provide insights into more subtle aspects of the FGIDs and generate hypotheses for future research but are beyond the scope of the present article.
One of the more notable findings of our study is that IBS was less than half as prevalent using Rome IV compared with Rome III. This comparison was included in the study to assess whether worldwide regional prevalence differed according to the criteria used.
The current Rome IV criteria are more stringent, requiring at least weekly abdominal pain (discomfort was not included), whereas Rome III required abdominal pain or discomfort at least 2 to 3 times monthly.
A comparative reappraisal of the Rome II and Rome III diagnostic criteria: are we getting closer to the 'true' prevalence of irritable bowel syndrome?.
Epidemiological, clinical, and psychological characteristics of individuals with self-reported irritable bowel syndrome based on the Rome IV vs Rome III criteria.
in a shift in prevalence from IBS-C to FC and from IBS with diarrhea (IBS-D) to functional diarrhea, because the increased pain frequency threshold required for IBS was not reached or the subjects suffered from discomfort rather than pain. This is consistent with the approach of viewing individual patients with constipation on a pain frequency spectrum
where differences in pain occurrence determine shifts from IBS to FC or functional diarrhea.
The Rome IV IBS criteria, being more restrictive than Rome III, lead to more similar diagnostic groups for clinical research and drug trials. Consequently, the Rome IV criteria are identifying more severe cases of IBS, as also reflected in the IBS-SSS scores, rather than the totality of the condition as seen by clinicians. However, clinicians may not use such stringent criteria in practice, as treatment is likely to be the same even for “sub-threshold” patients with slightly less frequent abdominal pain. Clinicians tend to rely more on symptom presentation and clusters. Thus, the relative prevalence “shift” away from IBS may have more implications for recruitment into research studies, especially clinical trials, where the Rome IV criteria define a more severe or specific population than those seen in clinical practice. Because FC and functional diarrhea and their corresponding IBS subtypes often respond to the same therapies,
the shift in diagnosis may have less impact on treatment.
Unspecified functional bowel disorder was the most prevalent bowel diagnosis in our study. As it is the default diagnosis for people who have significant bowel symptoms but fail to qualify for another bowel disorder, the diagnostic criteria for functional bowel disorders, especially IBS, may be too restrictive. In contrast, the least prevalent subtype of IBS (Table 1) was IBS-U, so the addition of the Bristol Stool Form Scale as a discriminator for IBS subtypes may have facilitated the classification of the 3 specific subtypes (IBS-C, IBS-D, and IBS-M), reducing the number of nonspecific cases. Diagnostic criteria for cannabinoid hyperemesis syndrome, central abdominal pain syndrome, and functional biliary pain, where hardly any cases were identified, may also be restrictive. These disorders may be particularly difficult to identify in studies based on questionnaires. Previous studies have shown somewhat higher prevalence rates, especially for chronic abdominal pain syndrome, the more studied of these relatively rare disorders.
Functional gastrointestinal disorders in Canada: first population-based survey using Rome II criteria with suggestions for improving the questionnaire.
Although the prevalence rates for many of the individual FGIDs were low, a large proportion of individuals met diagnostic criteria for at least 1 FGID. Combined with the findings that individuals meeting FGID criteria were twice as likely to consult doctors for bowel problems and had significantly lower general QOL than others, the collective burden of these disorders is substantial. The results of our study confirm that FGIDs are more prevalent among women than men. This is consistent with previous reports over the years.
However, in the household surveys, prevalence rates tended to rise with increasing age, as discussed further as follows.
A key strength of this study that has implications for future research was the effectiveness of Internet surveys. Not only is this now becoming the default option in most countries because telephone and mail surveys are not feasible ways to reach the general population, but it provided reliable, quality-controlled data with a nationally representative distribution. This could not be accomplished with household surveys.
The most important limitations of our study relate to the lack of national representation and missing data in the household surveys (excepting Turkey). Furthermore, the necessity of relying on 2 different survey methodologies precluded calculation of pooled global prevalence rates for all 33 countries together.
The anonymous Internet survey methodology constitutes a very different survey experience for subjects than the face-to-face household survey methodology, and cultural sensitivities around reporting of FGID symptoms may have led to the large differences in prevalence rates observed between the 2 survey methods. Our assessment is that the Internet surveys provided more reliable estimates of prevalence rates because (1) we achieved national representation, and (2) we achieved more complete, accurate and reliable data collection because no question that required an answer could be skipped, and questions that should have been skipped, based on responses to key trigger questions, were always skipped and could not be answered. In addition, quality control measures including maximum speed of questionnaire completion, repeat questions for response consistency assessment, and attention-check questions were included. Finally, data were automatically and accurately entered into the study database, eliminating manual entry errors. Another strength of the study was the uniform translation methodology with linguistic validation and cultural adaptation, which also generated a repository of translated study questionnaires for future global research in FGIDs.
The pattern of relative prevalence among the various FGIDs was consistent in spite of the substantial differences in the absolute prevalence rates between the Internet and household surveys. Although the reasons for the lower mean FGID prevalence in the household surveys compared with the Internet surveys are unclear, there are several possible explanations. It could be a consequence of reluctance to report sensitive or intimate digestive tract symptoms in face-to-face interviews. This could be more salient in younger respondents, consistent with the unique finding in household countries that rates increased with age. Moreover, because we included only 1 subject per household and FGIDs often cluster in families, this method might have underestimated the prevalence.
There was a much higher percentage of household than Internet survey participants living in rural communities (43.3% vs 9.7%, respectively). We explored if this disparity could explain the difference in prevalence rates between the two survey types. We found that FGIDs were reported slightly more frequently in rural than urban areas in household survey countries OR 1.1 (1.0–1.2), while slightly less frequently in rural than urban areas in Internet survey countries OR 0.9 (0.8–0.9), so this factor does not explain the difference, even in part. Because the prevalence rates were particularly low in India and Turkey (household), we rechecked the data entry process and the diagnostic scoring syntax, but no mistakes were found to explain this. The 2 geographically and linguistically separate sites surveyed in India had similarly low prevalence rates. To date, we have no definitive explanation for these unusually low rates, especially in light of the very different results from Bangladesh, a country with much in common with India.
Another limitation of the study is the lack of response rates for the Internet surveys and limited response rate information for the household surveys. However, in those countries for which we have precise data, Bangladesh, Malaysia, both regions in India, and Iran, the response rates were all more than 90%. Because similar participation encouragement methods were used in all household countries, we feel confident that the response rates were high in the others as well, although we do not have the exact figures.
An additional limitation of our study is that because it was a nonclinical questionnaire study of the general population, participants were not evaluated with procedures such as endoscopy or manometry, so some of the participants could have had an “organic” cause of their digestive symptoms. However, we believe that our inclusion of a checklist of organic diagnoses that might account for GI symptoms, and our exclusion of such cases from FGID prevalence counts (7.6% in Internet countries and 4.0% in household countries) compensated at least partially for this. A further limitation was that we attained less than satisfactory coverage of Africa and the Middle East. Our attempts to rectify this proved futile due to difficulties in recruiting interested investigators in the relevant countries.
The data collected with the supplemental questionnaire used in this study may yield findings on a range of variables with possible associations with FGIDs. Future analyses using those data will enable us to look at the relevance of differences in factors such as diet, hygiene, economic status, level of education, previous GI infections, and psychological comorbidity, with regard to FGID prevalence. These are likely to generate observations and hypotheses for further work that eventually may produce new insights into the pathophysiological mechanisms of FGIDs.
In conclusion, this article represents the first report documenting the global prevalence of FGIDs assessed with a uniform diagnostic questionnaire and research methodology. The results may influence substantially future planning of health care resources and clinical trials. Funding for research in the FGIDs is universally low, and they are viewed as a nonpriority. The data highlight a strong need and rationale for this to change. They should be of interest to multiple medical disciplines in addition to gastroenterologists, including general practitioners, family physicians, internists, nurses, dieticians, epidemiologists, public health experts, as well as other allied health care providers. We expect that the results presented here, and those to be reported from our future analyses, will serve as essential reference data for years to come.
Acknowledgments
The authors thank Mr Jerry Schoendorf for his contribution to the graphic design of Fig. 1.
CRediT Authorship Contributions
Ami D. Sperber, MD, MSPH (Conceptualization: Lead; Data curation: Equal; Formal
Supplementary Table 1Item Content of the Entire Global Study Questionnaire
Question content
No. of questions
Demographic questions:
Age
1
Sex
1
Years of education
1
Relationship status
1
Size of local community where respondent lives
1
Region (eg, state, province) of residence
1
Race/Ethnicity (not administered in all survey countries)
1
Religious/spiritual self-identification
1
Personal Health Questionnaire-15 (PHQ-15)
15
Rome IV Diagnostic Questionnaire for Adult FGIDs
89
IBS-SSS
7
Current living conditions: Whether respondent lives on a farm, has running water and electricity, daily access to telephone and internet, number of people in the household, number of bedrooms and toilets·
4
Childhood living conditions up to age 7: Whether respondent lived on a farm, had running water and electricity, number of people in the household, number of bedrooms and toilets·
4
Childhood country of residence
1
Childhood size of local community
1
Access to medical care if needed
1
Type of medical care that would be sought if needed (Western style medicine and/or traditional or folk healer
1
Frequency of doctor visits
1
Who pays for medical expenses
1
History of medical diagnoses (checklist of 12 GI diagnoses and conditions that may affect FGIDs)
1
History of GI and abdominal surgeries· Checklist of 5 surgery types
1
Medications taken regularly (at least once a week)· Yes/no list of 10 types of medications
1
Bowel infection history: Whether current bothersome symptoms first started immediately after bowel infection
1
Symptoms, conditions and treatment of bowel infection preceding first onset of current bothersome bowel symptoms
2
History of visiting doctor because of a bowel problem (yes/no)
1
Type of doctor seen for bowel problems
1
Concern about own bowel functioning (yes/no)
1
Embarrassment about bowel functioning (yes/no)
1
Impact of stress, pressure or tension on bowel functioning (yes/no)
1
Diet: Days per week of consumption of 10 food types
1
PROMIS Global-10 quality of life questionnaire
10
Personal Health Questionnaire – 4 (PHQ-4): Anxiety and depression screening measure
4
Height and weight
2
Rome III diagnostic questions for IBS (not administered in all countries)
Supplementary Table 2Countries, Language, and Distribution by Sex and Age for the Internet and Household Surveys
Country
Languages
N
Sex distribution (%)
Age distribution (%)
Male (50%)
Female (50%)
18–39 (40%)
40–64 (40%)
65+ (20%)
Internet
Argentina
Spanish (L)
2,058
50.6
49.4
39.5
40.2
20.4
Australia
English (L)
2,037
50.2
49.8
39.8
40.2
20.0
Belgium
French (L), Dutch (L)
2,021
50.1
49.9
40.2
40.0
19.8
Brazil
Portuguese (O)
2,000
50.0
50.0
39.8
40.0
20.2
Canada
English (L), French (L)
2,029
50.1
49.9
39.9
40.0
20.1
China
Chinese (O)
3,013
50.2
49.8
40.1
40.1
19.8
Colombia
Spanish (L)
2,088
49.9
50.1
41.1
43.4
15.4
France
French (O)
2,043
49.9
50.1
40.2
40.8
19.0
Germany
German (O)
2,042
49.8
50.2
40.3
40.2
19.5
Holland
Dutch (O)
2,008
50.0
50.0
39.8
40.1
20.1
Israel
Hebrew (O), Arabic (L), Russian (L), English (L)
2,014
50.0
50.0
40.3
40.4
19.3
Italy
Italian (O)
2,073
50.3
49.7
39.9
40.2
19.9
Japan
Japanese (O)
2,504
48.6
51.4
40.5
39.8
19.7
Mexico
Spanish (O)
2,001
50.4
49.6
40.4
40.4
19.2
Poland
Polish (O)
2,057
49.9
50.1
40.2
40.1
19.7
Romania
Romanian (O)
2,049
50.1
49.9
40.5
54.9
4.6
Russia
Russian (O)
2,000
50.7
49.3
40.0
40.3
19.7
Singapore
English (L), Chinese (L), Bahasa Malay (L)
2,047
50.0
50.0
47.7
48.6
3.7
South Africa
English (L)
2,023
50.3
49.7
53.1
40.6
6.3
South Korea
Korean (O)
2,085
50.4
49.6
38.6
51.0
10.3
Spain
Spanish (L)
2,071
50.2
49.8
39.8
40.3
19.9
Sweden
Swedish (O)
2,088
50.2
49.8
39.8
40.1
20.1
Turkey
Turkish (O)
2,010
49.8
50.2
50.1
44.7
5.2
United Kingdom
English (L)
2.027
50.1
49.9
39.8
40.1
20.1
United States
English (O)
2,026
50.1
49.9
39.8
40.0
20.2
Household
Bangladesh
Bengali (L)
2018
49.01
50.99
39.69
40.39
19.92
China
Chinese (O)
2710
47.42
52.58
33.32
47.27
19.41
Ghana
English (L)
1190
51.09
48.91
40.92
40.34
18.74
India
Hindi (O), Telugu (O), Bengali (O)
4592
50.20
49.80
42.09
41.70
16.20
Indonesia
Bahasa (L)
1231
48.90
51.10
39.32
40.70
19.98
Iran
Farsi (O)
1840
49.84
50.16
40.16
40.00
19.84
Malaysia
Bahasa-Malay (O)
1976
47.67
52.33
46.51
40.13
13.36
Nigeria
English (L)
1442
51.53
48.47
39.67
41.68
18.65
Turkey
Turkish (O)
1950
50.67
49.33
53.69
43.49
2.82
NOTE. The planned sex distribution was 50/50 and the planned age distribution was 40% (18–39), 40% (40–64), and 20% (65+). L, localized translation; O, original translation.
A global perspective on irritable bowel syndrome: a consensus statement of the World Gastroenterology Organisation Summit Task Force on irritable bowel syndrome.
A comparative reappraisal of the Rome II and Rome III diagnostic criteria: are we getting closer to the 'true' prevalence of irritable bowel syndrome?.
Epidemiological, clinical, and psychological characteristics of individuals with self-reported irritable bowel syndrome based on the Rome IV vs Rome III criteria.
Functional gastrointestinal disorders in Canada: first population-based survey using Rome II criteria with suggestions for improving the questionnaire.
Conflict of interest The authors disclose no conflicts.
Funding The study was funded, in part, by research grants from Ironwood , Shire , Allergan , and Takeda . The study in Malaysia was funded by the Fundamental Research Grant Scheme (FRGS) of the Ministry of Education of Malaysia (Reference: 203.PPSP.6171192). The study in Israel was funded by Takeda-Israel . The study in Romania was funded by the Romanian Society of Neurogastroenterology . None of the funders was involved in the planning, design, implementation, statistical analyses or any other aspect of the study including preparation of the paper or knowledge of its contents.