Epidemiology of constipation in Europe and Oceania: a systematic review

Background We aimed to review the literature regarding the epidemiology of constipation in Europe and Oceania and the associated prevalence/risk factors. Methods Two reviewers performed PubMed searches and a hand search of references. A study was considered eligible for inclusion if it reported data about the prevalence of constipation in any population, free of other gastrointestinal disorders, in Europe and Oceania. Studies were evaluated for quality. Data regarding the setting, type of study, definition of constipation, study population, prevalence of constipation, factors associated with increased odds for constipation, and the female to male ratio, were collected. Results The 21 reviewed studies depict prevalence rates in 34 different population groups ranging widely from a low 0.7% to a high 81%. In the general population of Europe the mean value of the reported constipation rates is 17,1 % and the median value 16.6%. Among the studies conducted in Oceania, the mean value of constipation prevalence was 15.3%. Female gender, age and socioeconomic and educational class seem to have major effect on constipation prevalence. A number of various other risk factors are, less clearly, associated with constipation. Conclusion This systematic review depicts the high prevalence and related risk factors of a disorder that decreases the health-related quality of life and has major economic consequences.

minority of patients. The majority of cases are attributed to functional disorders without a structural underlying cause that could explain symptoms. Risk factors, such as dietary habits, physical inactivity [8], socioeconomic level, psychological parameters, medications [9], age [10], gender [11], etc have been implicated in the development of chronic functional constipation. However, the specific etiology of this gastrointestinal disorder has not been elucidated.
Only a small proportion of patients suffering from constipation seek medical advice; thus, the exact prevalence of the disorder is difficult to be estimated. A systematic review of studies on the epidemiology of constipation in North America [12] recorded various prevalence rates from a low 1.9% to a high 27.2%. However, most of the included studies, report rates between 12% and 19%. To our knowledge, no previous report of the current evidence has systematically reviewed the epidemiology of constipation in Europe or Oceania.
We sought to review the literature regarding the epidemiology of functional constipation in Europe and Oceania. Specifically, our purpose was to identify the prevalence of constipation in the general population and in special population groups, identify the risk factors associated with this functional disorder and compare the findings with those of the systematic review of the epidemiology of constipation in North America [12].

Methods
We performed a systematic review of the current evidence regarding the prevalence of constipation in Europe and Oceania. Two reviewers (EGM and VGA) having a medical degree searched the PubMed database (until December 2006) to identify relevant studies for inclusion. The key word used in the literature search was "constipation" combined with "incidence, prevalence, rate, proportion, epidemiology, or morbidity". Studies were limited to those written in English, French and referring to humans. The titles and abstracts of all studies identified were further reviewed to meet the entry criteria. Moreover, we performed a hand search of the references of the selected studies to identify potentially overlooked articles.
A study was considered eligible for inclusion if it reported data about the prevalence of constipation in any population, free of other gastrointestinal disorders, in Europe and Oceania. All studies involving cases of functional constipation, regardless the diagnostic criteria used were considered for inclusion. The diagnostic criteria range widely from self-report and parent-report for children, to consensus criteria like Rome II criteria [13]; at least 12 weeks, which need not be consecutive, in the preceding 12 months, of two or more of: < 3 defecations per week (DPW), hard or lumpy stools in > 1/4 of defecations, straining in > 1/4, digital manipulation of stools in > 1/4, sensation of incomplete evacuation in > 1/4, sensation of anorectal obstruction or blockage in > 1/4. Two of the authors (EM and VGA) retrieved and studied in detail the complete texts of the original articles that were relevant to the focus of our review. Agreement between the investigators was essential for data inclusion and few minor disagreements were resolved with the consent of the senior author. Data regarding the first author, year of publication, setting, type of study, definition of constipation, study population, prevalence of constipation, factors associated with increased odds for constipation, and the female to male ratio, were extracted. The above were tabulated in detail without taking into consideration the differences in methodology among the reviewed studies. We calculated the mean values for prevalence of constipation in Europe and Oceania by averaging the prevalence obtained in each study.
We used the critical appraisal guidelines for research articles determining prevalence, developed by Loney et al [14], to evaluate the quality of studies included in our systematic review. The scoring system consists of 8 questions on the validity of the study design (appropriate methods and frame, adequate size, suitable criteria, outcome measurements, and response rate), the interpretation (prevalence given with confidence intervals) and the applicability of the results (detailed description of subjects and settings), weighted equality with one point [15]. Thus the maximum quality score that a study could achieve was 8.

Results
PubMed database returned a list of 2060 articles for the selected search string. From these, 2039 reports were excluded, after having reviewed their titles and abstracts, based on the criteria outlined in the Methods section. The full texts of the remaining 21 studies  were read and finally included in our systematic review. One more study [37] was identified following the hand search of references. However, Galvez et al evaluated a population already described by Garrigues et al study (included in this review). Reviewers disagreed on the inclusion of three studies [25,26,33]. However, they were included with the consent of the senior author.
In Tables 1, 2 and 3, we summarize the characteristics (author, year, type of study, depicted epidemiological data and quality score) of the reviewed studies regarding the prevalence of constipation in the general population of Europe, in special population groups of Europe and in all population groups of Oceania respectively. Overall the reviewed studies depict prevalence rates in 34 different  Abbreviations DPW: defecations per week, F/M: female/male ratio, RCT: randomized controlled trial, Rome criteria for children: in infants and preschool children, at least 2 wk of pebble-like, hard stools for a majority of stools, firm stools 2 or fewer times per week, and no evidence of structural, endocrine, or metabolic disease, *: at least one of less than three bowel movements weekly, hard or lumpy stools, straining on defecation, incomplete evacuation, in the preceding 12 months + : critical appraisal guidelines for research articles determining prevalence, developed by Loney et al 14 population groups ranging widely from a low 0.7% in a pediatric population in Italy [30] to a high 81% in a hospitalized, elderly, male population [24]. The mean value of the reported constipation rates is 22.3 and the median value 19.8. More than half of the reported constipation rates range from 8% to 26%.
Seven studies have data on the prevalence of constipation in the general population in Europe ( Table 1). The reported constipation rates range from a low 5% in a prospective study using the diagnostic criterion of < 3 defecations per week (DPW) [22] to a high 35% in an interview questionnaire survey of self-reported constipation [20]. The mean value of the reported constipation rates is 17.1 % and the median value 16.6%. Nine studies depict constipation prevalence rates in special population groups in Europe: infants, children, elderly population, women and participants living in long-term care facilities (Table 2). Finally, 5 studies regarding the epidemiology of constipation in Oceania are tabulated in Table 3. The lowest prevalence was observed among an elderly population in New Zealand (4.3%) [33]. Howel et al recorded the highest prevalence rate of constipation in the general population of Sydney Australia (30.7%). Among the studies conducted in Oceania, the mean value of constipation prevalence is 15.3%.
A secondary purpose of this review was to identify risk factors related with this functional constipation disorder. In particular, the predominance of females in the constipation prevalence is documented in most of the reviewed studies. In order to depict the exact magnitude of this association we present in Tables 1, 2 and 3 the female/ male ratio, in the studies where available. The mean and median value of the above ratio is calculated to 1.78 and 1.58 respectively. There are differences of the female/male ratio estimates by definition of constipation. However, the mean value remains above 1 for all case criteria, 1.7 for Rome I, 1.8 for Rome II, and 2.3 for self-report of constipation.
The effect of ageing on constipation prevalence has been recorded by 3 of the reviewed studies [24,27,33]. Socioeconomic and educational class influence on the constipation prevalence is also depicted in 3 studies [28,32,34]. Various other risk factors are associated with constipation by at least one research paper included in our systematic review of the current evidence; dietary habits like the consumption of olive oil and meat, life-style factors like living in a large community or old peoples home and immobility, frequency of breastfeeding, waist/hip ratio, anxiety and depression, co morbidities like hemorrhoids, other bowel disorders, previous hysterectomy and posterior colpocele and the use of medications like laxatives.

Discussion
The prevalence rates of constipation in Europe and Oceania as depicted by this review are consistent with the epidemiology of the disorder in North America [12]. Prevalence rates of constipation recorded in other developed countries are also within the same range; 14.3% in the general population in Hong-Kong [38], 16.5% in the general population in Korea [39], 24.5% in women in Taiwan [40], 26% in a population of young women in Japan [41], 29.6% in young children in Hong Kong [42], and 11.6% in an elderly Asian population [43]. The similarity of constipation rates recorded in the developed countries may partially be due to the common dietary habits of the studied populations. Constipation in individuals with previously normal bowel function is associated with specific dietary patterns; low fiber intake diet is involved with the pathogenesis of the disorder. Europe and Oceania are similar to North America in terms of health system and population's dietary habits, life style, physical activity and socioeconomic level.
Differing case definitions of constipation used, provide researchers with diverse prevalence rates. This is evident in the Garrigues et al study, where different diagnostic criteria were compared in the same individuals, resulting in significant differences in the depicted prevalence rates; 29.5% for self-report, 19.2% for Rome I criteria and 14% for Rome II criteria. We identified 3 studies using the Rome II criteria; the mean prevalence rate among these reports is 16.5%. The relevant mean prevalence rates for Rome I criteria and self-report are 19.2% and 25.6% respectively. It should be mentioned that the Rome III consensus provides the most up to date definition and categorization of functional gastrointestinal disorders in children/adolescent [44] and neonate/toddler [45]. None of the reviewed studies implemented the Rome III criteria. Future studies on the epidemiology of constipation should rather be based on consensus diagnostic criteria.
Female gender is associated with elevated constipation prevalence rates [17,18,21,25,32,33,36]. In North America, females are 2.2 times more likely to report constipation than males [12]. This predominance of females has been attributed to hormonal factors, inflicting a higher risk of constipation during the luteal phase of the menstrual cycle, under the effect of progesterone, and damage to the pelvic floor muscles, which may occur in women during childbirth or gynecological surgery [35]. Posterior colpocele is an independent risk factor for constipation [31]; however, causal relationship between these disorders is difficult to be assumed.
In general, individuals of lower social, economic and educational level have a tendency towards higher constipation rates. Bytzer et al divided the sample of their questionnaire survey into five socioeconomic classes from 1 st (highest) to 5 th (lowest). They showed that the constipation prevalence rates among those five groups followed the pattern of their socioeconomic class, ranging from a low 6.3% for the 1 st class to a considerably higher 10.2% for the 5 th class. Of interest, according to another study [28], low maternal educational level is considered as a factor significantly associated with increased odds for constipation of the newborn.
Chiarelli et al questionnaire survey compares constipation prevalence among three age groups, 18-23 years, 45-50 years and 70-75 years and depicts considerable differences in prevalence rates; 14.1%, 26.6% and 27.7% respectively. A study focusing on infants followed up for 6 months, showed a prevalence rate of 17.6% [29]. This is surprisingly high when compared with the constipation prevalence in a pediatric population aged from 0-12 years [30]; Miele et al used the Rome criteria in 9960 children and report a prevalence rate of 0.7%.
Psychological factors and in particular anxiety and depression, are considered predisposing to constipation [19]. These psychological disorders, as well as obsessive compulsion and social dysfunction, have been implicated in the pathogenesis of constipation, are believed to slow down colonic transit, have greater impact on women, and are associated with less frequent use of coping strategies [38]. Behavioral treatment, specifically biofeedback, is alleged to ameliorate the course of the disorder and is frequently used as alternative therapy for persons with symptoms unresponsive to traditional medical treatment [46,47].
Assessing quality in clinical trials is well described but much less attention has been given to similar strategies for observational epidemiological studies. We believe that the quality assessment tool used [14,15] addresses efficiently the need for critical appraisal of prevalence studies. There were significant differences in the quality and methodology among the reviewed studies. The best score achieved was 7/8 and the worst 3/8 (Tables 1, 2 and 3).
It should be acknowledge that when interpreting the findings of this systematic review one should take under consideration various limitations and shortcomings. First, the included studies use different definitions and diagnostic criteria of constipation, which could lead to inconsistencies regarding the prevalence estimates. No doubt, there is a discrepancy between self-reported constipation and the condition diagnosed based on the established Rome criteria. The rates are considerably higher when based on the definition of self-reported constipation [26]. This discrepancy highlighted by other investigators too [39], may be due to differences of personal perception regarding the problem, or could be involved with the validity of criteria. In addition, the use of questionnaires depends on the ability of the patient to recall symptoms, whereas prospective studies with the administration of diary cards are certainly more credible. Furthermore, the range of age groups studied is wide, and includes pediatric, as well as elderly study populations. Finally, it should be acknowledged that our search of the literature was limited to journals included in the PubMed database. However, the further hand search of references reduces possibly omitted studies. The language limit (articles published in English and French only) should be attributed to the available human resources. Despite this limitation, studies from 7 different European countries were included in our review (Spain, France, Italy, Netherlands, Norway, Sweden and Finland).

Conclusion
Because of the decrement in the health-related quality of life induced by constipation, and the economic consequences, physicians should be aware of the magnitude of the problem, and be able to give efficient instructions to patients for management, and particularly, prevention. We believe that our systematic review of current evidence on constipation provides a useful tool to physicians dealing with this high prevalence clinical problem, and a reference for future development of studies on the epidemiology and etiology of the disorder.