Combining Rome III criteria with alarm symptoms ...

1 downloads 0 Views 167KB Size Report
Questionnaire on Pediatric Gastrointestinal Symptoms Rome III, including questions on alarm symptoms, was used. A diagnostic review of their medical records ...
Acta Pædiatrica ISSN 0803-5253

REGULAR ARTICLE

Combining Rome III criteria with alarm symptoms provides high specificity but low sensitivity for functional gastrointestinal disorders in children Agneta Uusij€arvi ([email protected])1,2

, Ola Olen2,3,4, Petter Malmborg3,5, Martina Eriksson6, Peter Grimheden7, Henrik Arnell1,8

1.Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden €dersjukhuset, Karolinska Institutet, Stockholm, Sweden 2.Department of Clinical Science and Education, So 3.Sachs’ Children’s Hospital, Stockholm, Sweden 4.Clinical Epidemiology Unit, Department of Medicine Solna, Karolinska Institutet, Stockholm, Sweden 5.Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden 6.Diagnostic Radiology, Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden €dert€alje, Sweden 7.Sodertalje Hospital, So 8.Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden

Keywords Alarm symptoms, Functional dyspepsia, Irritable bowel syndrome, Questionnaire validation Correspondence Agneta Uusij€arvi, Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Pediatric Digestive Diseases, SE-141 86 Stockholm, Sweden. Tel: +46858587329 | Fax: +46 (0) 858 58 77 80 | Email: [email protected] Received 11 November 2017; revised 18 January 2018; accepted 21 February 2018. DOI:10.1111/apa.14297

ABSTRACT Aim: This study aimed to validate the Rome III criteria and alarm symptoms with regard to their ability to discriminate between organic and functional diagnoses in children with gastrointestinal complaints. Methods: We recruited 258 children aged four years to 17 years who consulted a paediatrician in secondary or tertiary care in Stockholm from January 2013 to May 2014 due to gastrointestinal complaints. A symptom questionnaire based on the official Questionnaire on Pediatric Gastrointestinal Symptoms Rome III, including questions on alarm symptoms, was used. A diagnostic review of their medical records was also carried out. Results: The reference diagnoses were organic (16%), pain-predominant functional gastrointestinal disorders (54%) and other functional diseases (30%). When the reported symptoms that fulfilled the Rome III criteria for pain-predominant functional gastrointestinal disorders were combined with an absence of alarm symptoms, they had a high specificity (0.90) for a functional diagnosis, but a low sensitivity (0.15). Alarm symptoms were equally common in patients with organic (83%) and functional diseases (80%, p = 0.66). Conclusions: Combining the Rome III criteria and an absence of alarm symptoms from patient questionnaires had high specificity but low sensitivity when diagnosing painpredominant functional gastrointestinal disorders in children seeking medical care for gastrointestinal complaints.

INTRODUCTION Abdominal pain in children is frequent in paediatric care and general practice, with an overall prevalence of as much as 30% reported by some cross-sectional studies (1,2). These patients often undergo a large number of investigations and test treatments with minimal or no yield (3–8). The aetiology of functional gastrointestinal disorders (FGIDs) is multifactorial (2,9–11). Paediatric diagnostic criteria for FGIDs have been developed so that studies can be compared, but, ideally, they should also be used to clinically establish a positive diagnosis of FGID, rather than relying on a diagnosis of exclusion. The international

Abbreviations FGID, Functional gastrointestinal disorder; IBD, Inflammatory bowel disease; IgA-TGA ab, Immunoglobulin-A-transglutaminase antibodies; OGID, Organic gastrointestinal disease; pFGID, Pain-predominant functional gastrointestinal disorder.

©2018 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

paediatric diagnostic Rome criteria for FGIDs were updated from version III to version IV in 2016 and are compared in Figure 1 (12–14). To fulfil the Rome III criteria for functional abdominal pain, functional dyspepsia or irritable

Key notes 





This study validated the ability of the Rome III criteria and alarm symptoms to discriminate between organic and functional diagnoses in children with gastrointestinal complaints. We recruited 258 children aged four years to 17 years who consulted a paediatrician in secondary or tertiary care due to gastrointestinal complaints. Combining the Rome III criteria and an absence of alarm symptoms from patient questionnaires had high specificity but low sensitivity when diagnosing painpredominant functional gastrointestinal disorders.

1

Symptoms criteria for functional gastrointestinal disorders

bowel syndrome, hereafter referred to as pain-predominant functional gastrointestinal disorder (pFGID), patients need to have suffered from abdominal pain or discomfort at least once a week in the past two months (13). Furthermore, the Rome criteria state that an appropriate medical evaluation must show that the symptoms cannot be attributed to another medical condition. There is expert consensus and a long tradition on the use of alarm symptoms to rule out organic disease (15). Studies on the paediatric Rome III criteria in combination with alarm symptoms are scarce (15–19) and, to our knowledge, there has only been one validation study (20). We therefore conducted this study to evaluate the clinical usefulness of the Rome III criteria. Our hypothesis was that the Rome III criteria, in combination with selected alarm symptoms, would be useful in clinical practice with regard to discriminating between functional and organic disease and could save afflicted children from unnecessary investigations and test treatments.

PATIENTS AND METHODS Children aged four years to 17 years of age who were referred to, or seeking, secondary or tertiary care because of gastrointestinal complaints were invited to participate in the study at their initial visit. From January 2013 to May 2014, we invited 310 children attending seven paediatric outpatient centres to take part in the study. The clinics were based at four secondary centres, in Huddinge, Nacka, € derta €lje, and at three tertiary centres at Liljeholmen and So Solna University Hospital, Huddinge University Hospital and Sachs’ Children’s Hospital. Signed consent was required from both parents and was missing for 52 children, meaning that the main analysis comprised 258 participants, which was 83% of those initially recruited. The sources of data for this study consisted of prospectively collected information from questionnaires at inclusion, data from patient records and diagnostic reviews. Questionnaires All participants answered a questionnaire at the inclusion visit. Parents filled in the questionnaire for children up to the age of nine years and children who were 10 years or more answered it themselves. The treating physician was blinded to the results of the questionnaires. The first part of the questionnaire included questions on eight gastrointestinal alarm symptoms during the two months preceding the visit: heredity for inflammatory bowel disease (IBD),

Frequency

Duration

Rome III (2006)

once per week

two months

Rome IV (2016)

four mes per month

two months

Figure 1 Comparison of Rome III and Rome IV criteria for pain-predominant functional gastrointestinal disorders.

2

Uusij€arvi et al.

coeliac disease or peptic ulcer disease, defecation at night, bloody stools, waking up at night because of abdominal pain, right-sided abdominal pain, unintentional weight loss, painful swallowing and unexplained fever. Alarm symptoms were chosen according to the Rome III documents, clinical experience and tradition (13). The symptoms of heredity, defecation at night and bloody stools were associated with organic gastrointestinal disease (OGID) in our study and therefore a missing answer on these items excluded the child from any further analysis. Because the other five alarm symptoms did not show any significant association with OGID, we choose to regard a missing answer as negative. The second part of the questionnaire was based on a validated Swedish version of the Questionnaire on Pediatric Gastrointestinal Symptoms Rome III, translated in adherence with the translation protocol suggested by the Rome Foundation (Appendix S1). Data from patient records Data were extracted by one of the authors (ME) from the electronic patient records from the patient’s first visit due to gastrointestinal complaints until six months to 12 months after inclusion in the study. The following data were obtained: number of previous visits for gastrointestinal complaints, specified laboratory tests, diagnostic radiology and endoscopies, bowel biopsy analyses, set diagnoses and prescribed treatments. The median period of electronic patient records scrutinised for each participant was 2.7 years (range: 0.6–15.3 years). Diagnostic review At least 12 months after the study inclusion, two experienced paediatric gastroenterologists (PM and HA) blinded to the answers to the questionnaire, reviewed all electronic patient records independently from each other to set the most likely clinical diagnosis. In 249 of 258 cases (97%), there was initial agreement between the two reviewers on whether the cause of abdominal discomfort had an organic or functional origin. Of the remaining nine cases, there was initial discordance between the two reviewers. These patient records were rereviewed by the two gastroenterologists together until a common diagnosis was agreed upon in all these patients. Five of the nine children with initial diagnostic discordance received a functional diagnosis and four an organic diagnosis. The expert diagnosis was then regarded as the reference when assessing the ability for Rome III symptom profiles, alarm symptoms and laboratory tests to discriminate between functional and organic causes of gastrointestinal complaints. The validity of the Rome III symptom criteria for pFGID, alone or in combination with negative alarm symptoms or normal laboratory tests, was calculated in four-field tables. The questionnaire answers provided information to classify whether children fulfilled the Rome III criteria for pFGID, but not for other functional diagnosis, for example constipation, vomiting and nausea. Accordingly, the children with a reference functional diagnosis of other than pFGID (n = 80) were excluded from the main analysis.

©2018 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Uusij€arvi et al.

Symptoms criteria for functional gastrointestinal disorders

Table 1 Children four years to 17 years seeking treatment for gastrointestinal complaints

Male Age in years at inclusion, median (25–75 percentile) QUESTIONNAIRE DATA: Symptom profiles of pFGID according to Rome III criteria§ ≥one alarm symptom present¶ pFGID Rome III and negative alarm symptom††

Total n = 258 n (%)

Reference organic diagnosis† n = 42 n (%)

116/258 (45.0) 10.8 (7.8–13.8)

19/42 (45.2) 8.5 (6.2–13.4)

182/253 (71.9) 207/256 (80.9) 32/253 (12.6)

27/40 (67.5) 35/42 (83.3) 4/40 (10.0)

Reference functional diagnosis† n = 216 n (%) 97/216 (44.9) 11.4 (8.2–13.9) 155/213 (72.6) 172/214 (80.4) 28/213 (13.1)

p‡ n.s. 0.029 n.s. n.s. n.s.



Children with a clinical diagnosis of organic gastrointestinal disease or functional gastrointestinal disorder. Chi-square test for categorical variables and the Kruskal–Wallis test for continuous variable (age). § Based on child/parental questionnaire response: children fulfilling symptom criteria for pain-predominant functional gastrointestinal disorder according to the paediatric Rome III criteria on duration and frequency. ¶ The alarm symptoms were as follows: heredity for inflammatory bowel disease, coeliac disease or peptic ulcer disease, defecation at night, bloody stools, waking up at night because of abdominal pain, right-sided abdominal pain, unintentional weight loss, painful swallowing, unexplained fever. †† As footnote ¶combined with absence of alarm symptoms. ‡

The remaining patients were categorised according to three different variables: positive paediatric Rome III symptom criteria, negative alarm symptoms and normal laboratory tests regarding faecal calprotectin (