Adverse Events After Cervical Spinal Manipulation - A Systematic Review and Meta-Analysis of Randomized Clinical...

Cervical manipulations are widely used by physiotherapists, chiropractors, osteopaths, and medical doctors for musculoskeletal dysfunctions like neck pain and cervicogenic headache. The use of cervical manipulation remains controversial, since it is often considered to pose a risk for not only benign adverse events (AEs), such as aggravation of pain or muscle soreness, but also severe AEs such as strokes in the vertebrobasilar or carotid artery following dissections. Studies finding an association between cervical manipulation and serious AEs such as artery dissections are mainly case control studies or case reports. These study designs are not appropriate for investigating incidences and therefore do not imply causal relationships. Randomized controlled trials (RCTs) are considered the gold standard study designs for assessing the unconfounded effects of benefits and harms, such as AEs, associated with therapies.

Due to the unclear risk level of AEs associated with high-velocity, low-amplitude (HVLA) cervical manipulation, the aim of this study was to extract available information from RCTs and thereby synthesize the comparative risk of AEs following cervical manipulation to that of various control interventions.

Systematic review and meta-analysis.

A systematic literature search was conducted in the PubMed and Cochrane databases. This search included RCTs in which cervical HVLA manipulations were applied and AEs were reported. Two independent reviewers performed the study selection, the methodological quality assessment, and the GRADE approach. Incidence rate ratios (IRR) were calculated. The study quality was assessed by using the risk of bias 2 (RoB-2) tool, and the certainty of evidence was determined by using the GRADE approach.

Fourteen articles were included in the systematic review and meta-analysis. The pooled IRR indicates no statistically significant differences between the manipulation and control groups. All the reported AEs were classified as mild, and none of the AEs reported were serious or moderate.

The search strategy was limited to literature in English or German. Furthermore, selection bias may have occurred, since only PubMed and Cochrane were used as databases, and searching was done by hand. RCTs had to be excluded if the results did not indicate the group in which the AEs occurred. A mandatory criterion for inclusion in the meta-analysis was a quantitative reproduction of the frequencies of AEs that could be attributed to specific interventions.

In summary, HVLA manipulation does not impose an increased risk of mild or moderate AEs compared to various control interventions. However, these results must be interpreted with caution, since RCTs are not appropriate for detecting the rare serious AEs. In addition, future RCTs should follow a standardized protocol for reporting AEs in clinical trials.