Low-quality evidence: CBT and hypnosis may be effective for pain relief in children with functional abdominal pain

Référence

Gordon M, Sinopoulou V, Tabbers M, et al. Psychosocial interventions for the treatment of functional abdominal pain disorders in children: a systematic review and meta-analysis. JAMA Pediatr 2022;176(6):560-568.

Sommaire

These authors searched several databases and registries to identify published and unpublished randomized trials of psychosocial interventions for children with functional abdominal pain disorders. The studies had to enroll children 4 to 18 years of age with pain disorders diagnosed via explicit standardized diagnostic criteria. Ultimately, they included 33 trials with 2657 children (67% were girls). The follow-up ranged from 5 days to 4 months. The studies compared several different interventions with no intervention or with other therapies: cognitive behavioral therapy (CBT), educational support, different types of hypnotherapies, guided imagery, and relaxation. The following interventions were not capable of being meta-analyzed: different types of osteopathy, fiber intake, biofeedback, and journal keeping. Although the risk of bias tool evaluates several different domains, only one study was at low risk of bias for masking (since it used a sham control), and the remainder were of mixed risk of this and several other types of bias. Additionally, the authors found variability in outcome measures. Overall, they found that CBT was more likely than no intervention to “successfully treat” children (only 334 children; 38% vs 15%; number needed to treat [NNT] = 5; 95% CI 4 - 8) and to decrease pain frequency (446 children; standard mean difference [SMD] -0.36; modest effect) and pain intensity (332 children; SMD -0.58; modest effect). Additionally, hypnotherapy was more effective to “successfully treat” children (91 children; 56% vs 19%; NNT = 3; 2 - 6). However, it is unclear if CBT was more effective than educational support, if yoga was more effective than no intervention, or if any of the other previously mentioned interventions were effective. Finally, although the authors made great efforts to identify all the relevant studies, I suspect significant publication bias exists.