Postprocedural Complications of External Ventricular Drains: A Meta-Analysis Evaluating the Absolute Risk of Hemorrhages, Infections, and Revisions.

BACKGROUND

External ventricular drain (EVD) insertion is often a lifesaving procedure frequently used in neurosurgical emergencies. It is routinely done at the bedside in the neurocritical care unit or in the emergency room. However, there are infectious and noninfectious complications associated with this procedure. This meta-analysis sought to evaluate the absolute risk associated with EVD hemorrhages, infections, and revisions. The secondary purpose was to identify and characterize risk factors for EVD complications.

METHODS

We searched the MEDLINE (PubMed) database for "external ventricular drain," "external ventricular drain" + "complications" or "Hemorrhage" or "Infection" or "Revision" irrespective of publication year. Estimates from individual studies were combined using a random effects model, and 95% confidence intervals (CIs) were calculated with maximum likelihood specification. To investigate heterogeneity, the t and I tests were utilized. To evaluate for publication bias, a funnel plot was developed.

RESULTS

There were 260 total studies screened from our PubMed literature database search, with 176 studies selected for full-text review, and all of these 176 studies were included in the meta-analysis as they met the inclusion criteria. A total of 132,128 EVD insertions were reported, with a total of 130,609 participants having at least one EVD inserted. The pooled absolute risk (risk difference) and percentage of the total variability due to true heterogeneity (I) for hemorrhagic complication was 1236/10,203 (risk difference: -0.63; 95% CI: -0.66 to -0.60; I: 97.8%), infectious complication was 7278/125,909 (risk difference: -0.65; 95% CI: -0.67 to -0.64; I: 99.7%), and EVD revision was 674/4416 (risk difference: -0.58; 95% CI: -0.65 to -0.51; I: 98.5%). On funnel plot analysis, we had a variety of symmetrical plots, and asymmetrical plots, suggesting no bias in larger studies, and the lack of positive effects/methodological quality in smaller studies.

CONCLUSIONS

In conclusion, these findings provide valuable information regarding the safety of one of the most important and most common neurosurgical procedures, EVD insertion. Implementing best-practice standards is recommended in order to reduce EVD-related complications. There is a need for more in-depth research into the independent risk factors associated with these complications, as well as confirmation of these findings by well-structured prospective studies.