Introduction
Faecal incontinence and chronic constipation are common disorders of defaecation, with faecal incontinence affecting around 6% [
1], and chronic constipation affecting between 10% and 15% of the population, depending on diagnostic criteria used [
2]. Both conditions have impact on quality of life [
3], resulting in significant health burden [
3,
4]. While advances have been made in the assessment of anorectal function, which has improved the understanding of the pathophysiology underlying disorders of defaecation [
5], better understanding of risk factors may reveal further insights into pathophysiology, direct future research, and, most importantly, highlight targets for disease prevention or modification.
Several risk factors for developing faecal incontinence and chronic constipation have been identified, predominantly through cross-sectional studies which vary in the definitions used, methodology, study population, and potential risk factors assessed. These limitations prevent aggregation of results by meta-analysis [
6].
The Delphi technique is an effective method for arriving at consensus on broad and complex problems [
7,
8]. Using knowledge and experience of experts (primarily colorectal surgeons) in the field of pelvic floor disorders, in combination with contemporary published evidence, this Delphi study aimed to identify and evaluate risk factors for faecal incontinence and chronic constipation as a preliminary step toward building Bayesian models of disease prediction that combine data and expert knowledge.
Discussion
To our knowledge, this is the first study to comprehensively evaluate the importance of potential risk factors for benign disorders of defaecation using Delphi methodology. Consensus was achieved for classification (independent risk factor, co-factor, not a risk factor) of all potential risk factors evaluated (33 for faecal incontinence and 38 for chronic constipation). Mean importance scores were also produced for 19 of the 20 independent risk factors for faecal incontinence and all the 18 independent risk factors for chronic constipation.
Age and sex were the most evaluated risk factors for faecal incontinence and chronic constipation in the literature. While female sex was considered an independent risk factor for faecal incontinence and chronic constipation, increasing age was classified as an independent risk factor for faecal incontinence but a co-factor for chronic constipation. Meta-analyses of studies reporting the prevalence of chronic constipation in the general population have reported a significant association with female sex [
4,
16,
17]. Barberio et al. [
2] reported higher pooled prevalence of functional constipation in women compared with men, irrespective of the Rome definition used. The evidence for increasing age as a risk factor for chronic constipation is less consistent. While Suares and Ford [
15] reported a modest increase in the pooled prevalence (17% in the ≥ 60 years compared with 12% in the < 29 years) and risk of chronic constipation in the higher age group (OR of 1.41 in the ≥ 60 years compared with < 29 years as baseline), Barberio et al. [
2] found no statistical differences in the prevalence of functional constipation between the different age groups. A recent population survey even showed the highest prevalence of Rome IV functional constipation in the youngest age group (9.9% in those aged 18–29 years) [
17]. Evidence from population studies largely supports the association between faecal incontinence and age and female sex [
6,
18,
19].
The independent risk factors considered most important through Delphi approach for faecal incontinence were obstetric factors, including third- or fourth-degree tears (i.e. obstetric anal sphincter injury), instrumental delivery, and grand multiparity. Several systematic reviews have concluded that obstetric anal sphincter injury is significantly associated with an increased risk of anal [
20‐
22] and faecal incontinence [
22]. A meta-analysis by Cattani et al. [
22] demonstrated a significant risk of anal incontinence associated with forceps delivery (OR 1.35 [CI 1.12–1.63]) and vacuum extraction delivery (OR 1.17 [CI 1.04–1.31]). The evidence of association between multiparity and anal or faecal incontinence is equivocal, with several studies reporting a significant association [
23‐
25] and others suggesting the contrary [
19,
26].
The current study did not consider episiotomy and first- or second-degree tears to be a risk factor for faecal incontinence. Several systematic reviews and meta-analyses have been performed to assess the risk of anal or faecal incontinence associated with episiotomy. LaCross et al. [
20] suggested an increased risk of anal incontinence (OR 1.74 [CI 1.28–2.38]) in women who had an episiotomy; however, Bols et al. [
21] did not find any significant association between first- or second-degree tear with faecal incontinence. Cattani et al. [
22] described an increased risk of anal (OR 1.51 [CI 1.16–1.96],
p = 0.002) but not faecal (OR 1.11 [CI 0.36–3.41],
P = 0.85) incontinence when an episiotomy is performed. Interpretation of these findings is challenging because of heterogeneity in episiotomy practice (routine vs. selective) and type of episiotomy performed (median vs. mediolateral).
One of the most important independent risk factors for chronic constipation was history of abuse, but no large population study has examined this risk factor, most likely because of its sensitive nature. Several small observational studies have consistently reported a significant association between history of abuse and constipation [
27], functional evacuation disorder [
28,
29], symptoms of incomplete evacuation [
30], or multiple pelvic floor complaints [
31].
Potentially modifiable risk factors for faecal incontinence and chronic constipation included dietary factors and obesity. Dietary factors were considered a co-factor for faecal incontinence and an independent risk factor for chronic constipation. A systematic review by Colavita and Andy [
32] found very limited data on the role of diet in the pathogenesis of faecal incontinence. Only one out of four studies found an association between low dietary fibre and faecal incontinence, but all five studies which assessed the effectiveness of diet as a treatment for faecal incontinence showed that fibre supplement improved faecal incontinence symptoms. Studies that have evaluated the dietary differences between individuals with and without chronic constipation have found a significant association with fluid intake [
16,
33,
34], but evidence of association between low dietary fibre and chronic constipation is equivocal [
35]; however, several systematic reviews have surmised that fibre supplementation is an effective treatment for chronic constipation [
36,
37]. Obesity was considered a co-factor for both faecal incontinence and chronic constipation. Evidence from several observational studies have found a significant association between obesity and faecal [
19,
26,
38‐
40] or flatus incontinence [
38,
41], which may be due to an increased risk of loose stools [
39] or the use of medications for weight loss [
42] or diabetes [
43]. Though some studies have reported no significant relationship [
44] or an inverse relationship [
16] between obesity and chronic constipation when defined as hard or infrequent stools, several studies have suggested an association between obesity and difficulty in rectal evacuation [
39,
41]. Other modifiable risk factors for faecal incontinence included diarrhoea, evacuation disorder, diabetes, medications, and excessive alcohol consumption. Other modifiable risk factors for chronic constipation included eating disorders, diabetes, hypothyroidism, lack of exercise, and medications.
Diarrhoea (or loose stools) was considered one of the most important risk factors for faecal incontinence by several observational studies [
18,
19,
26] but our experts assigned less importance to this risk factor. This discrepancy likely reflects the clinical practice of our experts, who were predominantly British colorectal surgeons, and may differ from opinions of patients and other specialists (general practitioner and gastroenterologists). This is a significant limitation on the generalisability of our findings. Further, although our results may be relevant to other developed countries with similar populations and risk factors, they may not be applicable to countries with less economic and healthcare resources. Acknowledging that the Delphi technique has been criticised for the quality of scientific evidence, the validity of the results, and the inconsistency of the study design [
9,
10], we generated the questionnaire from scientific evidence published in peer-reviewed journals and were rigorous in pre-defining and adhering to the core elements of the Delphi process. Though the first round of this study was a structured round, which is a deviation from the ‘classical’ Delphi approach [
10], this is common practice within clinical Delphi studies [
9] and is consistent with recent Delphi publications within the field of coloproctology (Supplement A). The opportunity for ‘experts’ to freely express their opinion was maintained by the provision of free-text suggestions. The method enabled the importance of risk factors with low volume of evidence due to sensitive nature, e.g. history of abuse, or low prevalence within the general population, e.g. anal trauma, to be evaluated in the same manner as well-established risk factors such as age. There was an attrition rate between rounds (41% from round 1 to round 2, and 71% from round 2 to round 3), which is not unexpected in a Delphi study [
45]. Despite this, we were able to maintain a multidisciplinary representation of experienced practitioners, which included colorectal surgeons, clinical nurse specialists, gastrointestinal physiologists, and a radiologist, from the first to the final round. Finally, this study did not include patients in the participants as it would be not have been possible to ensure unbiased and up to date clinical knowledge across all participants.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.