Sir,

We read with interest the paper by Spiteri Cornish and Reddy entitled ‘The use of propranolol in the management of periocular capillary haemangioma – a systematic review.’1 The authors provide an excellent review regarding the changing paradigms for the management of patients with infantile haemangiomas (IHs) and provide clear evidence as to the effectiveness of propanolol. The authors were unable to find an RCT comparing propanolol with a recognised standard or a placebo and concluded that an RCT would be beneficial but difficult to perform. The primary reason being that clinical equipoise cannot be met; given the weight of recent evidence the majority of clinicians would now be reluctant to revert to previous treatments for IHs, and therefore cannot ethically choose at random which treatment to provide.2 An RCT in this situation will have inherent flaws with confounding factors and bias.

Glasziou et al3 ask whether an RCT in this situation is actually required, illustrating similar historic milestones in medical practice where a new intervention has provided equally dramatic results. Occasionally, an RCT is not performed despite the intervention becoming mainstream practice, with examples including insulin to treat diabetes, defribrillation for ventricular fibrillation, and sutures to repair large wounds. They go on to describe the ‘rate ratio’ between the signal or intervention (propanolol) and the background noise (previous treatments). This is calculated as the rate of improvement of the lesion during the new treatment divided by the rate with previous treatments. Ideally, this ratio should be over 20, that is treatment with propanolol should increase the rate of improvement by over a factor of 20 when compared with a control that allows for bias and other confounding factors.

Clearly, the efficacy of propanolol is not disputed and issues regarding adverse effects should now be monitored using long-term data collection in the form of either a prospective database or registry. Such observational data will highlight rarer treatment complications and long-term effects. The optimal mode, dose, and duration of treatment of propanolol are yet to be defined and well-designed RCTs in this context would certainly be of benefit.