Deaths after chiropractic: a comment

Abstract

Linked Comment: Ernst. Int J Clin Pract 2011; 65: 818. To the Editor: Ernst’s interesting article (1) discussed 26 published fatal cases following chiropractic treatment, emphasising vascular accidents from vertebral artery dissection (VAD) after cervical spinal manipulative therapy (cSMT), and he responded (2) to five letters (3-7). Ernst suggested that ‘causality is at least likely in such cases’ (1), but included my case where a coroner cleared cSMT from causing the stroke, which was considered more likely to be of carotid artery origin (8). This warrants consideration because only five published cases reported VAD (8-12), and in one of these (9), the practitioner involved was probably a non-chiropractor (13). This leaves only three relatively complete cases where a registered chiropractor had probably administered cSMT before a lethal stroke from VAD (10-12), and it seems impossible to determine precisely what contribution cSMT had made. The true association of cSMT with VAD stroke remains unknown because of limitations regarding the relevant epidemiological studies (14-16). It may be only a weak association for the commonly used cSMTs, and similar to that for neck mobilising. Two studies involving physiotherapists who used cSMT and neck mobilising found two stroke cases for both mobilising and cSMT (17, 18). Ernst stated that ‘...chiropractors routinely employ high velocity...thrusts on the upper cervical spine with a rotational element...’ but this is a likely over-generalisation. My Perth survey (56% response rate) found 90% of chiropractor respondents replying that they seldom or never used rotatory cSMT (19). Based on a 1994 Perth survey (20) and recent data (21, 22), an estimated half million Western Australian adults (30%) have attended a chiropractor. Apparently, no Western Australian chiropractor has been found responsible for a death. While under-reporting in journals is expected as a result of publishing bias, this is less likely in the media, which reports coroner inquests of these VAD deaths. This supports Wenban and Bennet’s suggestion of extremely low risk (3). When criticised for describing his review as ‘systematic’ (7), Ernst explained that systematic review criteria for case studies are unformulated (2). Perhaps this term should only be used when the criteria are determined and followed. Ernst (1) omitted discussing misclassification bias regarding ‘chiropractor’ and ‘chiropractic’ (8, 13, 23), and was dismissive of three such examples (2) in his review that Wenban and Bennett revealed (5). However, similarly biased cases could have been listed, but had missed author survey (13, 23). I agree with others (6, 7) that Ernst’s claim ‘...The risks of chiropractic neck manipulation by far outweigh the benefits’ is unsupported by his results partly because they lack accurate data on cSMT benefits. Listed were 26 published fatal cases that supposedly implicated chiropractic over 76 years internationally, so in this context are arguably not ‘numerous’. Ernst’s article, which emphasised cSMT, is potentially misleading because it also failed to disclose five non-cSMT cases from the nine internet sourced cases, as Perle et al. revealed (6), and omitted the coroner’s exoneration of chiropractic in the unpublished Mathiason case (24). Analysing the under-reporting of adverse events inaccurately reflects cSMT risk, which will need improved epidemiological studies to measure properly. Considering the limitations of the relevant studies, caution is required on both sides of the debate when discussing the risk/benefit of cSMT. I have no financial conflicts of interest and have had no funding regarding my submission. There are no acknowledgements to be made.