Mixture model analysis identifies irritable bowel syndrome subgroups characterised by specific profiles of gastrointestinal, extraintestinal somatic and psychological symptoms.
Polster A, Van Oudenhove L, Jones M et al. Aliment Pharmacol Ther 2017; 46: 529-539
Irritable bowel syndrome (IBS) is a functional bowel disorder affecting 10-15% of the population. It has a significant impact on quality of life (QoL) and is also an important cause of healthcare utilisation and absenteeism from work. There is significant overlap with other functional gastrointestinal (GI) disorders and these can be viewed as parts of a functional disorders continuum.
Irritable bowel syndrome (IBS) is a functional bowel disorder affecting 10-15% of the population. It has a significant impact on quality of life (QoL) and is also an important cause of healthcare utilisation and absenteeism from work. There is significant overlap with other functional gastrointestinal (GI) disorders and these can be viewed as parts of a functional disorders continuum.
Current subtyping of IBS is exclusively based on the predominant stool consistency according to Rome criteria: IBS with constipation (IBS-C), IBS with diarrhoea (IBS-D), IBS with mixed bowel habits (IBS-M) and unsubtyped IBS (IBS-U). However, other relevant GI, extraintestinal somatic or psychological symptoms are not considered despite being important for decisions about management and the presence of extraintestinal or psychological symptoms affecting outcomes. This study aimed to identify subgroups of IBS patients based on a combination of GI and extraintestinal somatic symptoms, as well as psychological features.
Adult patients (≥18 years) meeting Rome III criteria for IBS were recruited at a Swedish hospital outpatient clinic. Clinical history was taken and diagnosis was based on clinical presentation and additional investigations if considered necessary. A two-week stool diary (using the Bristol Stool Form scale) was used to determine IBS subtypes based on predominant stool consistency. All patients avoided medication affecting pain, motility or stool form as well as psychotropic medication for the study period. Exclusion criteria included abnormal results on standard screening tests, severe psychiatric, systemic or other GI diseases, history of drug or alcohol abuse and an inability to respond to the questionnaires in Swedish.
The Gastrointestinal Symptom Rating Scale (GSRS-IBS), Bristol stool form diary, Patient Health Questionnaire (PHQ), Hospital Anxiety and Depression (HAD) scale and IBS severity scoring system (IBS-SSS) were used to collect the relevant data. Mixture modelling was then used to identify naturally occurring subgroups within the study population.
In all, data from 172 patients with IBS was analysed; this consisted of 119 women (69%) and the mean age was 33.7 (range: 18-60) years. Based on Rome III subtyping, the group consisted of 40 patients with IBS-C (23%), 64 patients with IBS-D (37%), 27 patients with IBS-M (16%) and 41 patients with IBS-U (24%). Approximately half of the patients (52%) had severe IBS based on the IBS-SSS results and an IBS duration of >10 years (47%). A third of patients (36%) had scores showing clinically relevant anxiety and only a small proportion showed scores indicating clinically-relevant depression.
This study identified six subgroups naturally occurring in the study population based on a combination of symptoms. Two groups showed constipation-predominant GI symptoms and two groups diarrhoea-predominant symptoms. Further distinction between these groups was through the presence or absence of an additional profile of extraintestinal somatic and psychological symptoms. Of the two remaining groups, one was characterised by a heterogeneous mix of mostly severe GI, extraintestinal somatic and psychological symptoms whilst the other showed a profile of overall low symptom severity.
The attempt to expand the current means of subtyping IBS patients to include a wider range of symptoms has allowed identification of subgroups with separate symptom profiles. Whilst a partial overlap exists with the Rome III subtypes, the addition of a wider range of symptoms has resulted in a different characterisation of patients. Whilst the clinical implications of these findings are unclear at this stage, this study offers an initial step to a multi-level approach which attempts to identify clinically meaningful subgroups and can be followed by further studies investigating factors such as response to therapies, pathophysiological or genotypic characteristics.
Previous studies have presented the hypothesis of dual aetiology based on the assumption that symptoms of a subgroup of IBS patients result from peripheral (‘biological’) mechanisms and those of another subgroup result from central (‘psychological’) aetiology. The findings of this study support dual or multi-aetiological hypotheses suggesting peripheral mechanisms such as local hypersensitivity, motility abnormalities or other changes to the GI tract are the primary process driving symptoms in the subgroups without psychological or extraintestinal somatic symptoms. In comparison, central mechanisms or disturbances of the of the gut-brain axis interactions may be driving symptoms in subgroups with a combination of symptoms bearing in mind the importance of bidirectional interaction between the central nervous system and the gut in development of GI and psychological symptoms has been demonstrated in earlier work. In addition, the patients within the low symptom severity subgroup could be representing a state of remission, although it is important to remember this is a cross-sectional study and data was only collected at one time point.
The recent publication of the revised Rome IV criteria also saw a Multi-Dimensional Clinical Profile (MDCP) system introduced in recognition of the variability of clinical presentation in IBS patients and with the aim of incorporating many additional criteria in the clinical management of patients. This study supports this approach and demonstrates different underlying multifaceted profiles within the IBS cohort.
These findings need to be validated in further studies to determine if these profiles can be reproduced and whether different pathophysiological or treatment responses can be attributed to the different subgroups. The clinical application of these findings is promising and may support more targeted and personalised therapies in the future. However, it is known that predominant bowel habits in patients with IBS tend to fluctuate over time and there are no long-term studies looking at how extraintestinal symptoms change over time meaning that there is a question around stability of the identified subgroups within this study. It also remains to be shown whether symptom profiles incorporating a wider range of symptoms have an improved predictive potential regarding treatment outcomes.
Whilst based on a large, well-phenotyped patient cohort, there are limitations to this study aside from its cross-sectional nature that need to be considered when interpreting the results. All study participants were recruited from a secondary/tertiary referral centre, which may result in recruitment bias towards more severe symptoms. However, only about half the sample were classified as having severe IBS, clinically relevant psychological comorbidities were only present in a subset of the sample and the symptom duration varied substantially. When these factors are considered alongside the fact that most participants were referred from a primary care physician, it suggests that the study sample is fairly representative of the general, heterogeneous IBS population.
The study concludes that it is possible to assign patients with IBS into subgroups based on a combination of GI, psychological and extraintestinal somatic symptoms, that are common in this patient group. This may be relevant for the clinical management of patients as patients with comorbidities may benefit from more targeted central mechanism interventions whilst those with solely GI symptoms may benefit from interventions targeting these more specifically. The predictive abilities regarding treatment outcome and long-term stability of these subgroups remains to be supported by other studies.
Link to full paper
Adult patients (≥18 years) meeting Rome III criteria for IBS were recruited at a Swedish hospital outpatient clinic. Clinical history was taken and diagnosis was based on clinical presentation and additional investigations if considered necessary. A two-week stool diary (using the Bristol Stool Form scale) was used to determine IBS subtypes based on predominant stool consistency. All patients avoided medication affecting pain, motility or stool form as well as psychotropic medication for the study period. Exclusion criteria included abnormal results on standard screening tests, severe psychiatric, systemic or other GI diseases, history of drug or alcohol abuse and an inability to respond to the questionnaires in Swedish.
The Gastrointestinal Symptom Rating Scale (GSRS-IBS), Bristol stool form diary, Patient Health Questionnaire (PHQ), Hospital Anxiety and Depression (HAD) scale and IBS severity scoring system (IBS-SSS) were used to collect the relevant data. Mixture modelling was then used to identify naturally occurring subgroups within the study population.
In all, data from 172 patients with IBS was analysed; this consisted of 119 women (69%) and the mean age was 33.7 (range: 18-60) years. Based on Rome III subtyping, the group consisted of 40 patients with IBS-C (23%), 64 patients with IBS-D (37%), 27 patients with IBS-M (16%) and 41 patients with IBS-U (24%). Approximately half of the patients (52%) had severe IBS based on the IBS-SSS results and an IBS duration of >10 years (47%). A third of patients (36%) had scores showing clinically relevant anxiety and only a small proportion showed scores indicating clinically-relevant depression.
This study identified six subgroups naturally occurring in the study population based on a combination of symptoms. Two groups showed constipation-predominant GI symptoms and two groups diarrhoea-predominant symptoms. Further distinction between these groups was through the presence or absence of an additional profile of extraintestinal somatic and psychological symptoms. Of the two remaining groups, one was characterised by a heterogeneous mix of mostly severe GI, extraintestinal somatic and psychological symptoms whilst the other showed a profile of overall low symptom severity.
The attempt to expand the current means of subtyping IBS patients to include a wider range of symptoms has allowed identification of subgroups with separate symptom profiles. Whilst a partial overlap exists with the Rome III subtypes, the addition of a wider range of symptoms has resulted in a different characterisation of patients. Whilst the clinical implications of these findings are unclear at this stage, this study offers an initial step to a multi-level approach which attempts to identify clinically meaningful subgroups and can be followed by further studies investigating factors such as response to therapies, pathophysiological or genotypic characteristics.
Previous studies have presented the hypothesis of dual aetiology based on the assumption that symptoms of a subgroup of IBS patients result from peripheral (‘biological’) mechanisms and those of another subgroup result from central (‘psychological’) aetiology. The findings of this study support dual or multi-aetiological hypotheses suggesting peripheral mechanisms such as local hypersensitivity, motility abnormalities or other changes to the GI tract are the primary process driving symptoms in the subgroups without psychological or extraintestinal somatic symptoms. In comparison, central mechanisms or disturbances of the of the gut-brain axis interactions may be driving symptoms in subgroups with a combination of symptoms bearing in mind the importance of bidirectional interaction between the central nervous system and the gut in development of GI and psychological symptoms has been demonstrated in earlier work. In addition, the patients within the low symptom severity subgroup could be representing a state of remission, although it is important to remember this is a cross-sectional study and data was only collected at one time point.
The recent publication of the revised Rome IV criteria also saw a Multi-Dimensional Clinical Profile (MDCP) system introduced in recognition of the variability of clinical presentation in IBS patients and with the aim of incorporating many additional criteria in the clinical management of patients. This study supports this approach and demonstrates different underlying multifaceted profiles within the IBS cohort.
These findings need to be validated in further studies to determine if these profiles can be reproduced and whether different pathophysiological or treatment responses can be attributed to the different subgroups. The clinical application of these findings is promising and may support more targeted and personalised therapies in the future. However, it is known that predominant bowel habits in patients with IBS tend to fluctuate over time and there are no long-term studies looking at how extraintestinal symptoms change over time meaning that there is a question around stability of the identified subgroups within this study. It also remains to be shown whether symptom profiles incorporating a wider range of symptoms have an improved predictive potential regarding treatment outcomes.
Whilst based on a large, well-phenotyped patient cohort, there are limitations to this study aside from its cross-sectional nature that need to be considered when interpreting the results. All study participants were recruited from a secondary/tertiary referral centre, which may result in recruitment bias towards more severe symptoms. However, only about half the sample were classified as having severe IBS, clinically relevant psychological comorbidities were only present in a subset of the sample and the symptom duration varied substantially. When these factors are considered alongside the fact that most participants were referred from a primary care physician, it suggests that the study sample is fairly representative of the general, heterogeneous IBS population.
The study concludes that it is possible to assign patients with IBS into subgroups based on a combination of GI, psychological and extraintestinal somatic symptoms, that are common in this patient group. This may be relevant for the clinical management of patients as patients with comorbidities may benefit from more targeted central mechanism interventions whilst those with solely GI symptoms may benefit from interventions targeting these more specifically. The predictive abilities regarding treatment outcome and long-term stability of these subgroups remains to be supported by other studies.
Link to full paper
www.drschaer-institute.com