Alimentary Pharmacology and Therapeutics
Differentiation of functional constipation and constipation predominant irritable bowel syndrome based on Rome III criteria: a population-based study N. A. Koloski*, M. Jones†, M. Young* & N. J. Talley*
*Faculty of Health and Medicine, University of Newcastle, Callaghan, NSW, Australia. † Department of Psychology, Macquarie University, North Ryde, NSW, Australia.
Correspondence to: Professor N. J. Talley, Faculty of Health and Medicine, University of Newcastle, Callaghan, NSW 2308, Australia. E-mail: [email protected]. au
Publication data Submitted 21 September 2014 First decision 6 October 2014 Resubmitted 22 December 2014 Resubmitted 9 February 2015 Accepted 11 February 2015 EV Pub Online 4 March 2015 This article was accepted for publication after full peer-review.
SUMMARY Background While the Rome III classification recognises functional constipation (FC) and constipation predominant IBS (IBS-C) as distinct disorders, recent evidence has suggested that these disorders are difficult to separate in clinical practice. Aim To identify whether clinical and lifestyle factors differentiate Rome III-defined IBS-C from FC based on gastrointestinal symptoms and lifestyle characteristics. Method 3260 people randomly selected from the Australian population returned a postal survey. FC and IBS-C were defined according to Rome III. The first model used logistic regression to differentiate IBS-C from FC based on lifestyle, quality-of-life and psychological characteristics. The second approach was data-driven employing latent class analysis (LCA) to identify naturally occurring clusters in the data considering all symptoms involved in the Rome III criteria for IBS-C and FC. Results We found n = 206 (6.5%; 95% CI 5.7–7.4%) people met strict Rome III FC whereas n = 109 (3.5%; 95% CI 2.8–4.1%) met strict Rome III IBS-C. The case–control approach indicated that FC patients reported an older age at onset of constipation, were less likely to exercise, had higher mental QoL and less health care seeking than IBS-C. LCA yielded one latent class that was predominantly (75%) FC, while the other class was approximately half IBS-C and half FC. The FC-dominated latent class had clearly lower levels of symptoms used to classify IBS (pain-related symptoms) and was more likely to be male (P = 0.046) but was otherwise similar in distribution of lifestyle factors to the mixed class. Conclusion The latent class analysis approach suggests a differentiation based more on symptom severity rather than the Rome III view. Aliment Pharmacol Ther 2015; 41: 856–866
856
ª 2015 John Wiley & Sons Ltd doi:10.1111/apt.13149
Differentiation of functional constipation and IBS-C INTRODUCTION Chronic constipation is defined as persistent difficulty, infrequent or seemingly incomplete defecation.1, 2 Constipation-type symptoms are common, with reports that up to 27% of people in the community have constipation, depending on the demographic factors, sampling and definitions used.3–7 Most patients presenting with constipation do not have a structural or metabolic explanation for the symptoms; these patients are labelled as having functional constipation (FC)8 or constipation predominant – irritable bowel syndrome (IBSC) if abdominal pain is linked to the constipation.8 While the current Rome III criteria supports the view that IBS-C and FC are distinct disorders, this arbitrary classification is controversial, with evidence suggesting that patients with IBS-C and FC are similar in terms of symptomatology and pathophysiology.9, 10 To date, no studies have assessed whether distinct constipation groupings occur in the general population and whether these can be differentiated by clinical or lifestyle factors. The aetiology for constipation and its subtypes remains relatively obscure.11, 12 The evidence that lifestyle factors such as exercise and alcohol and caffeine intake are causally linked to constipation is weak,11–13 although nonsteroidal anti-inflammatory drug use, the use of other constipation-inducing medications and concurrent diseases are important risk factors for some people.13, 14 In IBS, increasing exercise was associated with symptom reduction.15 The Rome criteria considers the underlying pathophysiological mechanisms of FC and IBS-C to be different. The aetiology of FC is believed to be related to delayed transit and/or pelvic floor dysfunction compared with IBS-C which is conceptualised as dysfunction in the brain gut pathways.16, 17 Mearin and Rey, however, recently observed that physical activity and liquid and fibre intake were similar between people from the Spanish population with chronic constipation vs. IBS-C defined using the Rome III criteria.18 Others have shown that IBS-C is associated with greater psychological distress and health care seeking compared with chronic constipation.19 Whether IBSC and FC can be distinguished with respect to other factors is unknown. Thus, in a large population-based study, we aimed to identify clinical and lifestyle factors that differentiate Rome III-defined IBS-C from functional constipation (FC) based on gastrointestinal symptoms relevant to either IBS-C or FC and lifestyle characteristics.
Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
METHODS Participants A random sample of 8981 people aged 18 years and over with equal numbers of males and females were randomly selected from the Newcastle, Charlton, Shortland, Paterson, Hunter and New England electoral areas, New South Wales, Australia. Measures The Rome III diagnoses for functional constipation, IBS – constipation predominant and IBS8 were based on the Rome III Questionnaire.20 This is a self-report instrument that measures functional gastrointestinal symptoms experienced over the prior 3 months.20 Previous testing has shown this instrument to be reliable and valid.20 We also asked about the age when people first experienced constipation. Definitions of constipation8 Functional constipation (Rome III)8. Two or more of the following for the past 3 months with symptom onset at least 6 months before diagnosis: (i) Straining during at least 25% of bowel movements. (ii) Lumpy or hard stools for at least 25% of bowel movements. (iii) Feeling of incomplete evacuation for at least 25% of bowel movements. (iv) Feeling of anorectal blockage for at least 25% of bowel movements. (v) Use of manual manoeuvres to facilitate at least 25% of bowel movements. (vi) Less than three bowel movements/week. Additional features: (i) loose stool rarely present in the absence of laxative use. (ii) insufficient criteria met for irritable bowel syndrome. We also assessed less strict definitions of constipation including (i) Rome III Functional Constipation without the 6-month criteria and (ii) any constipation which included having the presence of two or more of the constipation symptoms listed above regardless of the frequency of symptoms.
857
N. A. Koloski et al. Irritable bowel syndrome8. Symptoms of recurrent abdominal pain or discomfort and a change in bowel habit for at least 6 months, with symptoms experienced on at least 3 days for at least 3 months. Two or more of the following must apply: (i) Pain is relieved by a bowel movement. (ii) Onset of pain is related to a change in frequency of stool. (iii) Onset of pain is related to a change in the appearance of stool.
Irritable bowel syndrome – constipation predominant (Rome III)8. People had to meet criteria for IBS as detailed above but also had to have the presence of hard stools but not loose stools over the past 3 months. Demographics We asked about age, gender and educational level attained (high school). Lifestyle factors We included questions on smoking, alcohol consumption and exercise based on the valid National Health Survey.21 Smoking was grouped into never, past or current smoker. Alcohol consumption was classified into nil, low (1–6 drinks), risky (7–8 drinks) and high (more than 9 drinks) a week. Exercise was classified by the presence (yes or no), number of times and amount of time spent walking in the last 2 weeks, doing any moderate exercise in the last 2 weeks and doing any vigorous exercise in the last 2 weeks. We also obtained information to calculate body mass index. Psychological distress Psychological distress was assessed with the reliable and valid Hospital Anxiety and Depression Scale (HAD).22 The questionnaire contains 14 items that assess anxiety (seven items) and depression (seven items).22 Each item is rated on a four-point scale ranging from 0–3, such that each sub-scale has a maximum score of 21. Higher scores on the HAD indicate depression or anxiety in a patient, with a score of 7 or less on either sub-scale indicating a ‘non case’, 8–10 as a ‘doubtful case’ and a score of 11 or more as a ‘case’ of anxiety or depression.22 Quality of life (SF-12) Health-related quality of life was assessed using the valid SF-12.23 This is a 12-item generic quality-of-life measure assessing mental and physical functioning over the past 858
4 weeks. Examples of questions assessing mental functioning included ‘Have you felt calm and peaceful?’ and ‘Have you felt down hearted and blue?’ Physical functioning was addressed with questions such as ‘During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?
Health care seeking We asked about how many times health care was sought for stomach and/or bowel problems over the past 12 months. Procedures. The study was approved by the Hunter New England Human Ethics Committee. Participants were sent a participant information sheet outlining the study, the survey and a reply paid envelope. We used the Dillman Total Design Method24 for the follow-up of nonresponders. The follow-up protocol included a reminder/ thank you letter sent to all participants at week 1, a replacement survey sent out to nonresponders at week 4 and a reminder/thank you letter sent at week 5 to those nonresponders identified at week 4. Additional measures to maximise response rates included a personalised covering letter personally signed by the chief investigator and an easy to understand attractive coloured questionnaire booklet. Participants did not receive any remuneration for their participation. Statistical analyses We adopted two separate approaches to provide two distinct views. The first was model-based using logistic regression to differentiate IBS-C from FC based on lifestyle, quality-of-life (QoL) and psychological characteristics. The extent to which subject characteristics discriminate IBS-C from FC was evaluated univariately via unconditional logistic regression. Odds ratios >1.0 are indicative that the discriminator presence or high values (as appropriate) are associated with higher odds of IBS-C rather than FC. Identification of risk factors which yield statistically significant incremental discrimination of IBS-C from FC has been via multiple unconditional logistic regression. This modelling was complicated by SF-36 and alcohol intake having substantial missing data (11% and 31% respectively). Model selection via backward elimination was initially undertaken using all potential discriminators except for alcohol intake and with missing values excluded. Potential bias in this model was considered through refitting the model using multiple imputation. Potential discriminators not Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
Differentiation of functional constipation and IBS-C selected, including alcohol intake, were also reconsidered in this step. No additional discriminators were selected and just one discriminator that was originally selected was omitted due to its P-value falling from P = 0.034 before multiple imputation to P = 0.075 after using multiple imputation. To allow comparability of the coefficients in the multiple variable model, all numeric discriminators have been standardised by dividing by their standard deviation. The second approach was data-driven employing latent class analysis (LCA) to identify naturally occurring clusters in the data considering all symptoms involved in the Rome III criteria for IBS-C and FC. A two-class solution was forced to determine whether the Rome III IBS-C and FC classifications were reproduced in a naturalistic approach. We then examined how the latent classes identified by this technique related to the individual symptoms, Rome III classification and lifestyle, qualityof-life and psychological characteristics. LCA is a statistical technique which is penetrating medical research publications although relatively recently.25 The technique adds new information to that provided by the regression approach by identifying clusters of individuals in the sample which are not directly observable (hence latent classes). An important point is that these classes are formed with respect to patient-reported symptoms only, that is, they are formed independently of the individual’s Rome classification. Hence, a comparison of the latent class membership with Rome category membership provides insight into how the expert-derived Rome criteria compare with how individuals cluster naturalistically. If the latent classes are each dominated by either IBS-C or FC, it would support the argument that these are two distinct disorders. The combination of these two quite different approaches yields insights in how these apparently distinct syndromes differ.
RESULTS Response rate A total of 3260 people returned a completed survey out of 8981, giving a response rate of 38%. The denominator was reduced by 13 people who had died and 466 return to senders. Responders were significantly more likely to be aged over 45 years compared with nonresponders (68.3% vs. 40.5%, P < 0.001). Responders were also more likely to be female vs. nonresponders, but although this was statistically significant, it was numerically very small due to the study power (53.4% vs. 48.0%, P < 0.001). There were no statistically significant differences between Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
responders and nonresponders with respect to socioeconomic status with the SIEFA Advantage–Disadvantage scores being similar for responders vs. for nonresponders (M = 965.8 vs. M = 965.5, P = 0.89).
Sample characteristics Among our final sample (n = 3260), we found there were slightly more females (53.8%) compared with males (46.4%). The mean age for our sample was 54.3 years (standard deviation 15.6 years). More than half of the sample had completed greater than high school level education (54.6%). Prevalence of different types of constipation. We found having any constipation to be very common with just over a third of the general population (34.4%; 95% CI 32.8–36.1%) reporting more than two constipation symptoms in the past 3 months. When functional constipation was defined according to the strict Rome III criteria, we found 6.5% (95% CI 5.7–7.4%) of the population met criteria for this condition. This was increased to 11.0% (95% CI 9.9–12.1%) when the 6-month criteria was removed. The prevalence of constipation predominant IBS was much lower at 3.5% (95% CI 2.8–4.31%). The remaining results pertain to the subsample who met Rome III criteria for either IBS-C (n = 109) or functional constipation (FC, n = 206) for a total sample of 315 individuals. Logistic regression approach: Functional constipation vs. IBS-C. In terms of demographic features, we found that people with FC were significantly older compared with people with IBS-C, albeit this was borderline significant (Table 1). People with FC were also significantly older when they first experienced constipation compared with IBS-C (Table 1). There were no significant differences between the two groups in terms of gender or educational level (Table 1). With respect to lifestyle factors, people with FC and IBS-C were not significantly different in terms of smoking status or alcohol consumption (Table 1). However, people with IBS-C engaged in significantly more exercise in terms of walking including the number of times and amount of time spent walking in the past 2 weeks compared with people with FC. There were no differences between FC and IBS-C with respect to moderate or vigorous exercise, although there was a trend for higher levels of these types of exercise with people with IBS-C vs. FC (Table 1). Nor were there any differences between the two groups in terms of body mass index (Table 1). 859
N. A. Koloski et al. Table 1 | Demographic, lifestyle and health-related impact variables associated with FC vs. IBS-C Variable Demographics Age -Mean (s.d.) in years Gender (% females) Education level % (n) HSC Age first constipation – Mean (s.d.) in years Lifestyle Smoking – % (n) Never Past Current Alcohol consumption – % (n) Nil Low Risky High Exercise – % (n) Walking Presence – yes % (n) Number of times – mean (s.d.) Amount of time – mean in min (s.d.) Moderate Presence – yes % (n) Number of times – mean (s.d.) Amount of time – mean in min (s.d.) Vigorous Presence – yes % (n) Number of times – mean (s.d.) Amount of time- mean in min (s.d.) Body mass index – mean (s.d.) Quality of life – SF-12 Mental functioning – mean (s.d.) Physical functioning – mean (s.d.) Psychological distress – HADS Anxiety – Mean (s.d.) Depression – Mean (s.d.) Health care seeking for stomach/bowels % (n) Nil 1–4 times 5–8 times 9–12 times >12 times
FC (n = 206)
IBS-C (n=109)
56.3 (15.7) 76.2 (157)
52.9 (15.1) 83.3 (90)
P = 0.05 P = 0.1
41.8 6.8 51.5 33.7
38.5 6.4 55.1 27.7
P = 0.8
(86) (14) (106) (21.3)
(42) (7) (60) (19.8)
FC vs. IBS-C
P = 0.03
56.2 (11.3) 33.3 (67) 10.5 (21)
49.5 (51) 39.8 (41) 10.7 (11)
P = 0.5
18.8 70.1 7.6 3.5
10.8 73.0 12.2 4.1
P = 0.4
(27) (101) (11) (5)
(8) (54) (9) (3)
58.5 (120) 4.1 (5.3) 195.7 (348.5)
70.6 (77) 4.8 (4.5) 248.3 (415.1)
P = 0.04 P = 0.05 P = 0.07
52.2 (107) 2.8 (4.0) 170.5 (426.3)
58.7 (64) 3.3 (4.0) 195.5 (303.8)
P = 0.3 P = 0.2 P = 0.2
21.3 1.0 46.6 27.2
P P P P
16.7 0.6 42.7 27.9
(34) (1.7) (190.2) (6.2)
(23) (2.5) (121.6) (5.8)
= = = =
0.3 0.4 0.3 0.6
45.6 (8.4) 46.5 (10.4)
41.5 (9.7) 44.4 (11.9)
P = 0.003 P = 0.3
10.1 (1.6) 5.7 (2.8)
10.5 (1.8) 6.1 (2.9)
P = 0.2 P = 0.2
49.5 39.3 5.6 3.7 1.9
P = 0.003
64.6 33.3 0.5 0.5 1.0
(126) (65) (1) (1) (2)
(53) (42) (6) (4) (2)
HSC - Higher School Certificate HADS- Hospital Anxiety and Depression Scale S.d. -Standard Deviation
In terms of health-related impact, we found that people with IBS-C had significantly lower scores (indicating greater impairment) on the mental functioning sub-scale of the SF-12 compared with FC (Table 1). The two groups were similar with respect to physical functioning. 860
The two groups were also similar in terms of psychological status (Table 1). However, significantly more people with IBS-C visited health care in the past 12 months for stomach and bowel problems compared with FC (Table 1). Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
Differentiation of functional constipation and IBS-C Multivariate analysis. We found being younger, having poorer mental and physical functioning and engaging in walking exercise over the past 2 weeks to be independent predictors of meeting criteria for IBS-C vs. FC (Table 2).
Latent class analysis approach: Latent class analysis yielded two classes. One latent class that was predominantly (75%, n = 164) FC compared with IBS-C (25%, n = 54); the other class was approximately half IBS-C (56.7%, n = 55) and half FC (43.3%, n = 42; Figure 1). Interestingly, if the solution is relaxed to allow three classes, one class (total n = 132) is comprised of 97% FC, while the remaining two classes are both 43% FC (data not shown). The latter classes are differentiated in degree of both IBS and constipation symptoms. The FC-dominated latent class had clearly lower levels of symptoms used to classify IBS (pain-related symptoms), whereas the latent class that was equally IBS-C and FC, in addition, reported higher levels of constipation symptoms compared with the FC-dominated class (Table 3). The latent class dominated by FC was more likely to be male (P = 0.046) but was otherwise similar
Table 2 | Multivariable discriminators of IBS-C from FC Variables
Odds ratio (95% CI)
P value
Age Mental functioning Physical functioning Walking over the past 2 weeks
0.75 0.54 0.63 2.48
0.045
Differentiation of functional constipation and constipation predominant irritable bowel syndrome based on Rome III criteria: a population-based study N. A. Koloski*, M. Jones†, M. Young* & N. J. Talley*
*Faculty of Health and Medicine, University of Newcastle, Callaghan, NSW, Australia. † Department of Psychology, Macquarie University, North Ryde, NSW, Australia.
Correspondence to: Professor N. J. Talley, Faculty of Health and Medicine, University of Newcastle, Callaghan, NSW 2308, Australia. E-mail: [email protected]. au
Publication data Submitted 21 September 2014 First decision 6 October 2014 Resubmitted 22 December 2014 Resubmitted 9 February 2015 Accepted 11 February 2015 EV Pub Online 4 March 2015 This article was accepted for publication after full peer-review.
SUMMARY Background While the Rome III classification recognises functional constipation (FC) and constipation predominant IBS (IBS-C) as distinct disorders, recent evidence has suggested that these disorders are difficult to separate in clinical practice. Aim To identify whether clinical and lifestyle factors differentiate Rome III-defined IBS-C from FC based on gastrointestinal symptoms and lifestyle characteristics. Method 3260 people randomly selected from the Australian population returned a postal survey. FC and IBS-C were defined according to Rome III. The first model used logistic regression to differentiate IBS-C from FC based on lifestyle, quality-of-life and psychological characteristics. The second approach was data-driven employing latent class analysis (LCA) to identify naturally occurring clusters in the data considering all symptoms involved in the Rome III criteria for IBS-C and FC. Results We found n = 206 (6.5%; 95% CI 5.7–7.4%) people met strict Rome III FC whereas n = 109 (3.5%; 95% CI 2.8–4.1%) met strict Rome III IBS-C. The case–control approach indicated that FC patients reported an older age at onset of constipation, were less likely to exercise, had higher mental QoL and less health care seeking than IBS-C. LCA yielded one latent class that was predominantly (75%) FC, while the other class was approximately half IBS-C and half FC. The FC-dominated latent class had clearly lower levels of symptoms used to classify IBS (pain-related symptoms) and was more likely to be male (P = 0.046) but was otherwise similar in distribution of lifestyle factors to the mixed class. Conclusion The latent class analysis approach suggests a differentiation based more on symptom severity rather than the Rome III view. Aliment Pharmacol Ther 2015; 41: 856–866
856
ª 2015 John Wiley & Sons Ltd doi:10.1111/apt.13149
Differentiation of functional constipation and IBS-C INTRODUCTION Chronic constipation is defined as persistent difficulty, infrequent or seemingly incomplete defecation.1, 2 Constipation-type symptoms are common, with reports that up to 27% of people in the community have constipation, depending on the demographic factors, sampling and definitions used.3–7 Most patients presenting with constipation do not have a structural or metabolic explanation for the symptoms; these patients are labelled as having functional constipation (FC)8 or constipation predominant – irritable bowel syndrome (IBSC) if abdominal pain is linked to the constipation.8 While the current Rome III criteria supports the view that IBS-C and FC are distinct disorders, this arbitrary classification is controversial, with evidence suggesting that patients with IBS-C and FC are similar in terms of symptomatology and pathophysiology.9, 10 To date, no studies have assessed whether distinct constipation groupings occur in the general population and whether these can be differentiated by clinical or lifestyle factors. The aetiology for constipation and its subtypes remains relatively obscure.11, 12 The evidence that lifestyle factors such as exercise and alcohol and caffeine intake are causally linked to constipation is weak,11–13 although nonsteroidal anti-inflammatory drug use, the use of other constipation-inducing medications and concurrent diseases are important risk factors for some people.13, 14 In IBS, increasing exercise was associated with symptom reduction.15 The Rome criteria considers the underlying pathophysiological mechanisms of FC and IBS-C to be different. The aetiology of FC is believed to be related to delayed transit and/or pelvic floor dysfunction compared with IBS-C which is conceptualised as dysfunction in the brain gut pathways.16, 17 Mearin and Rey, however, recently observed that physical activity and liquid and fibre intake were similar between people from the Spanish population with chronic constipation vs. IBS-C defined using the Rome III criteria.18 Others have shown that IBS-C is associated with greater psychological distress and health care seeking compared with chronic constipation.19 Whether IBSC and FC can be distinguished with respect to other factors is unknown. Thus, in a large population-based study, we aimed to identify clinical and lifestyle factors that differentiate Rome III-defined IBS-C from functional constipation (FC) based on gastrointestinal symptoms relevant to either IBS-C or FC and lifestyle characteristics.
Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
METHODS Participants A random sample of 8981 people aged 18 years and over with equal numbers of males and females were randomly selected from the Newcastle, Charlton, Shortland, Paterson, Hunter and New England electoral areas, New South Wales, Australia. Measures The Rome III diagnoses for functional constipation, IBS – constipation predominant and IBS8 were based on the Rome III Questionnaire.20 This is a self-report instrument that measures functional gastrointestinal symptoms experienced over the prior 3 months.20 Previous testing has shown this instrument to be reliable and valid.20 We also asked about the age when people first experienced constipation. Definitions of constipation8 Functional constipation (Rome III)8. Two or more of the following for the past 3 months with symptom onset at least 6 months before diagnosis: (i) Straining during at least 25% of bowel movements. (ii) Lumpy or hard stools for at least 25% of bowel movements. (iii) Feeling of incomplete evacuation for at least 25% of bowel movements. (iv) Feeling of anorectal blockage for at least 25% of bowel movements. (v) Use of manual manoeuvres to facilitate at least 25% of bowel movements. (vi) Less than three bowel movements/week. Additional features: (i) loose stool rarely present in the absence of laxative use. (ii) insufficient criteria met for irritable bowel syndrome. We also assessed less strict definitions of constipation including (i) Rome III Functional Constipation without the 6-month criteria and (ii) any constipation which included having the presence of two or more of the constipation symptoms listed above regardless of the frequency of symptoms.
857
N. A. Koloski et al. Irritable bowel syndrome8. Symptoms of recurrent abdominal pain or discomfort and a change in bowel habit for at least 6 months, with symptoms experienced on at least 3 days for at least 3 months. Two or more of the following must apply: (i) Pain is relieved by a bowel movement. (ii) Onset of pain is related to a change in frequency of stool. (iii) Onset of pain is related to a change in the appearance of stool.
Irritable bowel syndrome – constipation predominant (Rome III)8. People had to meet criteria for IBS as detailed above but also had to have the presence of hard stools but not loose stools over the past 3 months. Demographics We asked about age, gender and educational level attained (high school). Lifestyle factors We included questions on smoking, alcohol consumption and exercise based on the valid National Health Survey.21 Smoking was grouped into never, past or current smoker. Alcohol consumption was classified into nil, low (1–6 drinks), risky (7–8 drinks) and high (more than 9 drinks) a week. Exercise was classified by the presence (yes or no), number of times and amount of time spent walking in the last 2 weeks, doing any moderate exercise in the last 2 weeks and doing any vigorous exercise in the last 2 weeks. We also obtained information to calculate body mass index. Psychological distress Psychological distress was assessed with the reliable and valid Hospital Anxiety and Depression Scale (HAD).22 The questionnaire contains 14 items that assess anxiety (seven items) and depression (seven items).22 Each item is rated on a four-point scale ranging from 0–3, such that each sub-scale has a maximum score of 21. Higher scores on the HAD indicate depression or anxiety in a patient, with a score of 7 or less on either sub-scale indicating a ‘non case’, 8–10 as a ‘doubtful case’ and a score of 11 or more as a ‘case’ of anxiety or depression.22 Quality of life (SF-12) Health-related quality of life was assessed using the valid SF-12.23 This is a 12-item generic quality-of-life measure assessing mental and physical functioning over the past 858
4 weeks. Examples of questions assessing mental functioning included ‘Have you felt calm and peaceful?’ and ‘Have you felt down hearted and blue?’ Physical functioning was addressed with questions such as ‘During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?
Health care seeking We asked about how many times health care was sought for stomach and/or bowel problems over the past 12 months. Procedures. The study was approved by the Hunter New England Human Ethics Committee. Participants were sent a participant information sheet outlining the study, the survey and a reply paid envelope. We used the Dillman Total Design Method24 for the follow-up of nonresponders. The follow-up protocol included a reminder/ thank you letter sent to all participants at week 1, a replacement survey sent out to nonresponders at week 4 and a reminder/thank you letter sent at week 5 to those nonresponders identified at week 4. Additional measures to maximise response rates included a personalised covering letter personally signed by the chief investigator and an easy to understand attractive coloured questionnaire booklet. Participants did not receive any remuneration for their participation. Statistical analyses We adopted two separate approaches to provide two distinct views. The first was model-based using logistic regression to differentiate IBS-C from FC based on lifestyle, quality-of-life (QoL) and psychological characteristics. The extent to which subject characteristics discriminate IBS-C from FC was evaluated univariately via unconditional logistic regression. Odds ratios >1.0 are indicative that the discriminator presence or high values (as appropriate) are associated with higher odds of IBS-C rather than FC. Identification of risk factors which yield statistically significant incremental discrimination of IBS-C from FC has been via multiple unconditional logistic regression. This modelling was complicated by SF-36 and alcohol intake having substantial missing data (11% and 31% respectively). Model selection via backward elimination was initially undertaken using all potential discriminators except for alcohol intake and with missing values excluded. Potential bias in this model was considered through refitting the model using multiple imputation. Potential discriminators not Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
Differentiation of functional constipation and IBS-C selected, including alcohol intake, were also reconsidered in this step. No additional discriminators were selected and just one discriminator that was originally selected was omitted due to its P-value falling from P = 0.034 before multiple imputation to P = 0.075 after using multiple imputation. To allow comparability of the coefficients in the multiple variable model, all numeric discriminators have been standardised by dividing by their standard deviation. The second approach was data-driven employing latent class analysis (LCA) to identify naturally occurring clusters in the data considering all symptoms involved in the Rome III criteria for IBS-C and FC. A two-class solution was forced to determine whether the Rome III IBS-C and FC classifications were reproduced in a naturalistic approach. We then examined how the latent classes identified by this technique related to the individual symptoms, Rome III classification and lifestyle, qualityof-life and psychological characteristics. LCA is a statistical technique which is penetrating medical research publications although relatively recently.25 The technique adds new information to that provided by the regression approach by identifying clusters of individuals in the sample which are not directly observable (hence latent classes). An important point is that these classes are formed with respect to patient-reported symptoms only, that is, they are formed independently of the individual’s Rome classification. Hence, a comparison of the latent class membership with Rome category membership provides insight into how the expert-derived Rome criteria compare with how individuals cluster naturalistically. If the latent classes are each dominated by either IBS-C or FC, it would support the argument that these are two distinct disorders. The combination of these two quite different approaches yields insights in how these apparently distinct syndromes differ.
RESULTS Response rate A total of 3260 people returned a completed survey out of 8981, giving a response rate of 38%. The denominator was reduced by 13 people who had died and 466 return to senders. Responders were significantly more likely to be aged over 45 years compared with nonresponders (68.3% vs. 40.5%, P < 0.001). Responders were also more likely to be female vs. nonresponders, but although this was statistically significant, it was numerically very small due to the study power (53.4% vs. 48.0%, P < 0.001). There were no statistically significant differences between Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
responders and nonresponders with respect to socioeconomic status with the SIEFA Advantage–Disadvantage scores being similar for responders vs. for nonresponders (M = 965.8 vs. M = 965.5, P = 0.89).
Sample characteristics Among our final sample (n = 3260), we found there were slightly more females (53.8%) compared with males (46.4%). The mean age for our sample was 54.3 years (standard deviation 15.6 years). More than half of the sample had completed greater than high school level education (54.6%). Prevalence of different types of constipation. We found having any constipation to be very common with just over a third of the general population (34.4%; 95% CI 32.8–36.1%) reporting more than two constipation symptoms in the past 3 months. When functional constipation was defined according to the strict Rome III criteria, we found 6.5% (95% CI 5.7–7.4%) of the population met criteria for this condition. This was increased to 11.0% (95% CI 9.9–12.1%) when the 6-month criteria was removed. The prevalence of constipation predominant IBS was much lower at 3.5% (95% CI 2.8–4.31%). The remaining results pertain to the subsample who met Rome III criteria for either IBS-C (n = 109) or functional constipation (FC, n = 206) for a total sample of 315 individuals. Logistic regression approach: Functional constipation vs. IBS-C. In terms of demographic features, we found that people with FC were significantly older compared with people with IBS-C, albeit this was borderline significant (Table 1). People with FC were also significantly older when they first experienced constipation compared with IBS-C (Table 1). There were no significant differences between the two groups in terms of gender or educational level (Table 1). With respect to lifestyle factors, people with FC and IBS-C were not significantly different in terms of smoking status or alcohol consumption (Table 1). However, people with IBS-C engaged in significantly more exercise in terms of walking including the number of times and amount of time spent walking in the past 2 weeks compared with people with FC. There were no differences between FC and IBS-C with respect to moderate or vigorous exercise, although there was a trend for higher levels of these types of exercise with people with IBS-C vs. FC (Table 1). Nor were there any differences between the two groups in terms of body mass index (Table 1). 859
N. A. Koloski et al. Table 1 | Demographic, lifestyle and health-related impact variables associated with FC vs. IBS-C Variable Demographics Age -Mean (s.d.) in years Gender (% females) Education level % (n) HSC Age first constipation – Mean (s.d.) in years Lifestyle Smoking – % (n) Never Past Current Alcohol consumption – % (n) Nil Low Risky High Exercise – % (n) Walking Presence – yes % (n) Number of times – mean (s.d.) Amount of time – mean in min (s.d.) Moderate Presence – yes % (n) Number of times – mean (s.d.) Amount of time – mean in min (s.d.) Vigorous Presence – yes % (n) Number of times – mean (s.d.) Amount of time- mean in min (s.d.) Body mass index – mean (s.d.) Quality of life – SF-12 Mental functioning – mean (s.d.) Physical functioning – mean (s.d.) Psychological distress – HADS Anxiety – Mean (s.d.) Depression – Mean (s.d.) Health care seeking for stomach/bowels % (n) Nil 1–4 times 5–8 times 9–12 times >12 times
FC (n = 206)
IBS-C (n=109)
56.3 (15.7) 76.2 (157)
52.9 (15.1) 83.3 (90)
P = 0.05 P = 0.1
41.8 6.8 51.5 33.7
38.5 6.4 55.1 27.7
P = 0.8
(86) (14) (106) (21.3)
(42) (7) (60) (19.8)
FC vs. IBS-C
P = 0.03
56.2 (11.3) 33.3 (67) 10.5 (21)
49.5 (51) 39.8 (41) 10.7 (11)
P = 0.5
18.8 70.1 7.6 3.5
10.8 73.0 12.2 4.1
P = 0.4
(27) (101) (11) (5)
(8) (54) (9) (3)
58.5 (120) 4.1 (5.3) 195.7 (348.5)
70.6 (77) 4.8 (4.5) 248.3 (415.1)
P = 0.04 P = 0.05 P = 0.07
52.2 (107) 2.8 (4.0) 170.5 (426.3)
58.7 (64) 3.3 (4.0) 195.5 (303.8)
P = 0.3 P = 0.2 P = 0.2
21.3 1.0 46.6 27.2
P P P P
16.7 0.6 42.7 27.9
(34) (1.7) (190.2) (6.2)
(23) (2.5) (121.6) (5.8)
= = = =
0.3 0.4 0.3 0.6
45.6 (8.4) 46.5 (10.4)
41.5 (9.7) 44.4 (11.9)
P = 0.003 P = 0.3
10.1 (1.6) 5.7 (2.8)
10.5 (1.8) 6.1 (2.9)
P = 0.2 P = 0.2
49.5 39.3 5.6 3.7 1.9
P = 0.003
64.6 33.3 0.5 0.5 1.0
(126) (65) (1) (1) (2)
(53) (42) (6) (4) (2)
HSC - Higher School Certificate HADS- Hospital Anxiety and Depression Scale S.d. -Standard Deviation
In terms of health-related impact, we found that people with IBS-C had significantly lower scores (indicating greater impairment) on the mental functioning sub-scale of the SF-12 compared with FC (Table 1). The two groups were similar with respect to physical functioning. 860
The two groups were also similar in terms of psychological status (Table 1). However, significantly more people with IBS-C visited health care in the past 12 months for stomach and bowel problems compared with FC (Table 1). Aliment Pharmacol Ther 2015; 41: 856–866 ª 2015 John Wiley & Sons Ltd
Differentiation of functional constipation and IBS-C Multivariate analysis. We found being younger, having poorer mental and physical functioning and engaging in walking exercise over the past 2 weeks to be independent predictors of meeting criteria for IBS-C vs. FC (Table 2).
Latent class analysis approach: Latent class analysis yielded two classes. One latent class that was predominantly (75%, n = 164) FC compared with IBS-C (25%, n = 54); the other class was approximately half IBS-C (56.7%, n = 55) and half FC (43.3%, n = 42; Figure 1). Interestingly, if the solution is relaxed to allow three classes, one class (total n = 132) is comprised of 97% FC, while the remaining two classes are both 43% FC (data not shown). The latter classes are differentiated in degree of both IBS and constipation symptoms. The FC-dominated latent class had clearly lower levels of symptoms used to classify IBS (pain-related symptoms), whereas the latent class that was equally IBS-C and FC, in addition, reported higher levels of constipation symptoms compared with the FC-dominated class (Table 3). The latent class dominated by FC was more likely to be male (P = 0.046) but was otherwise similar
Table 2 | Multivariable discriminators of IBS-C from FC Variables
Odds ratio (95% CI)
P value
Age Mental functioning Physical functioning Walking over the past 2 weeks
0.75 0.54 0.63 2.48
0.045
