Craniectomy – a no-brainer?

Raised pressure in and around the brain is associated with (notice I didn’t say “causes”) bad outcomes in patients with traumatic brain injury. Management of such patients centres around reducing this pressure, either by managing their breathing and giving drugs, or by surgical decompression of the brain, usually achieved by removing a piece of skull (craniectomy). Craniectomy is common practice, and it has been around for over 100 years. This recent comparative trial showed that craniectomy was successful in reducing the pressure around the brain, but caused (notice how I didn’t say “was associated with”) more harm than good. A case of “the operation was a success, but the patient died”.

Craniectomy is a classic example of treating the numbers, not the patient and our obsessions with normalising things that we can measure that has fascinated doctors since the first measurements could be taken – temperature, pulse and blood pressure. Instead of getting the patient better (improving real clinical outcomes), we focus on improving the numbers, and assume that this will do the trick.

With all trials, there are questions about the generalisability of a study that only included patients with a certain diagnosis and certain pressures and certain treatments, but this will always be the case. My response to critics of trials who quote the problem of generalisability is to suggest that they do another trial – one that includes the exact patients that they think should be included, rather than using a lack of generalisability in the current evidence as an excuse to persist with untested treatments.

Another similar trial is underway (the RESCUEicp trial). The inclusion criteria are slightly different (higher threshold pressure for example), but the question is the same: is doing a decompressive craniectomy better than the not doing one? The study has nearly finished recruiting and I will include a link to the results when they are available.

The bottom line

Many medical treatments persist for decades, based on theory, observation and tradition. Often, comparative randomised trials are never done, but when they are, and when they show the long-standing practice to be ineffective (or, as is the case here, harmful), they strengthen the need for such trials and cause us to question other practices that are based on non-experimental evidence.