Guidelines. Algorithms. Evidence based medicine. These all play a significant part in the safe and effective management of the majority of our patients. As a result there is a danger that treatment pathways are followed blindly without critiquing their use and there is real risk we can loose sight of what’s best for the patient in front of us. Guidelines encourage inflexible decision making, which creates further challenge when we are met by patients who do not fit standard treatment pathways. If this is the case then the management of cardiac arrest, which is taught and delivered in a didactic and protocol driven fashion, is surely the pinnacle of the problem. Standard Advanced Life Support (ALS) is totally appropriate for the majority of cardiac arrests, but what happens when it fails our patients? Refractory ventricular fibrillation (rVF) is, by its very nature, defined by the failure of ALS, but frustratingly there is very little evidence, or guidance, surrounding how to manage this patient group. I was faced with this situation when called to support an ambulance crew who were resuscitating an out-of-hospital VF arrest. When benchmarked against the ALS guidelines their management had been exemplary, but the patient remained in VF after eight shocks. So what now? The UK Resuscitation Council doesn’t specifically discuss rVF, but offers the advice that it is “usually worthwhile continuing” if the patient remains in VF. Not a particularly controversial statement, but not much help either. They do discuss the potential for ECMO use, but this is currently a very rare option in UK practice, or thrombolysis for known or suspected pulmonary embolism. Other potential interventions not in the guidelines include IV magnesium and placing defibrillator pads in the anterior-posterior orientation. PCI can also be considered if there is a suitable receiving centre available and the patient can be delivered in a safe and timely fashion. A final option is the use of double sequential defibrillation (DSD), using two defibrillators, charged to their maximum energy setting, to deliver two shocks in an almost simultaneous fashion.DSD was first described in human subjects in 1994 when it was used to successfully defibrillate five patients who entered rVF during routine electrophysiologic testing. These patients were otherwise refractory to between seven and twenty single shocks. Looking at the available literature there has been little interest in DSD since then, until the last two to three years when it appears to have undergone a small revival. Sadly, there is no evidence for its use beyond case reports and small case series. The case reports appear to show good results with four in the last two years reporting survival to discharge with good neurological outcome. There are also a handful of other cases discussed in online blogs and articles with good outcomes. But these case reports and articles almost certainly represent an excellent example of publication bias. The most recent case series reviewed the use of DSD by an American ambulance service over a period of four years. During this time DSD was written into their refractory VF protocol, with rVF defined as failure of five single shocks. The study included twelve patients, three of whom survived to discharge with two of these demonstrating a cerebral performance score of 1. Despite appearing to demonstrate reasonable outcomes for DSD, sadly this study has a number of significant limitations. One important point the authors fail to discuss is that the two neurologically intact survivors received their DSD shocks after two and three single shocks respectively, not after five shocks which would have been per-protocol and consistent with the authors definition of refractory VF. This highlights a further problem with analysing the use of DSD. Not only is there a dearth of high-level evidence, but the literature that is available is highly inconsistent. There are a range of definitions for re