Abstract
One potentially fatal complication of spine surgery is myocardial infarction (MI). There is still uncertainty of the true incidence of MI within subsets of spine surgeries. The aim of this study was to survey the contemporary spine literature and ascertain the true incidence of MI after lumbar spine surgery, as well as to provide commentary on the inherent assumptions made when interpreting cohort versus database studies on this topic.
A systematic search of 4 electronic databases from inception to November 2020 was conducted following PRISMA guidelines. Articles were screened against prespecified criteria. MI incidence was then estimated by random-effects meta-analyses of proportions based on cohort versus database studies.
A total of 34 cohort studies and 32 database studies describing 767,326 lumbar procedures satisfied all criteria for selection. There were 12,170 (2%) cases from cohort studies and 755,156 (98%) cases from database studies. Cohort studies reported a significantly older patient cohort (P < 0.01) and longer follow-up period than did database studies (P < 0.03). Using cohort studies only, the incidence of MI was 0.44% (P heterogeneity < 0.01), whereas using database studies only, the incidence of MI was 0.41% (P heterogeneity < 0.01). These 2 incidences were statistically different (P interaction = 0.01). Bias analysis indicated that cohort studies were more vulnerable to small-study biases than were database studies.
Although infrequent, the incidence of MI after lumbar spine surgery is unequivocally nonzero. Furthermore, the literature on this topic remains skewed based on study type, and translation of academic findings into practice should be wary of this.
