LAUSR

Dear Sir The recently updated MASCC/ESMO consensus recommendations for the management of nausea and vomiting in advanced cancer remain based on Blargely poor quality trials or uncontrolled trials and case studies^ [1]. However, there is a notable exception. Two placebo-controlled randomised trials with adequate power to discard the null hypothesis regarding the use of somatostatin analogues in malignant bowel obstruction (MBO), published since the 2010 recommendations, were included [2, 3]. Neither trial demonstrated a statistically significant difference in their primary outcome. We were therefore surprised to see octreotide recommended as first line for MBO, reported with high levels of consensus and supported by level II evidence, in the face of directly contradicting grade A data. The studies included in the MASCC/ESMO recommendations did not appear to be evaluated systematically as recommended by the Guidelines International Network (GIN) which aim to ensure consistent standards for the development and reporting of guidance documents [4]. Instead, the authors conjectured (no rationale given) that benefits seen in uncontrolled studies may not have been seen in the Currow trial [3] because trial participants also received treatments used in current practice and stated that the presence of carcinomatosis was unclear (the trial reports that all study participants had peritoneal carcinomatosis). For the Mariani trial, the comments focussed on the secondary analyses and positive secondary outcomes [2]. In contrast, a systematic review of the use of somatostatin analogues in MBO, published about the same time as the MASCC/ESMO recommendations, critically appraised the included trials using the Cochrane risk of bias tool. The two adequately powered RCTs [2, 3] both had low risk of bias. Therefore, their findings were given more weight than the findings of studies with high/ unclear risk of bias. The authors of this systematic review concluded that Bhigh-level evidence from trials with low risk of bias found no benefit of somatostatin analogues for their primary outcome^ [5]. The use of consensus methodology rather than critical appraisal of the evidence in developing guidelines risks selective exposure, or confirmation bias, whereby individuals favour data that reinforces their pre-existing views, whilst avoiding contradictory results. [6] This may encourage perpetuation of ineffective and potentially harmful practices (participants in the octreotide arm of the Currow trial required more antispasmodics [3]). The differences in interventions and outcome measures across the trials in the Obita et al. review [5] made it impossible to conduct a meta-analysis. As highlighted in the MASCC/ESMO guidelines, we do not have an agreed, patient-relevant primary outcome, or its minimal clinically important difference. Therefore, it is possible that somatostatin trials to date may have measured outcomes of little importance to the patient. We agree this is a significant concern; until we engage our MBO patients in identifying what symptoms distress them most, we are unlikely to move forward in deriving further evidence to inform management of this condition. * Miriam J. Johnson [email protected]