Pre-operative stress testing in the evaluation of patients undergoing non-cardiac surgery: A systematic review and meta-analysis

Abstract

Background: Pre-operative stress testing is widely used to evaluate patients for non-cardiac surgeries. However, its value in predicting peri-operative mortality is uncertain. The objective of this study is to assess the type and quality of available evidence in a comprehensive and statistically rigorous evaluation regarding the effectiveness of pre-operative stress testing in reducing 30-day post -operative mortality following non -cardiac surgery.

Methods: The databases of MEDLINE, EMBASE, and CENTRAL databases (from inception to January 27, 2016) were searched for all studies in English. We included studies with pre-operative stress testing prior to 10 different non-cardiac surgery among adults and excluded studies with sample size<15. The data on study characteristics, methodology and outcomes were extracted independently by two observers and checked by two other observers. The primary outcome was 30-day mortality. We performed random effects meta-analysis to estimate relative risk (RR) and 95% confidence intervals (95% CI) in two-group comparison and pooled the rates for stress test alone. Heterogeneity was assessed using I2 and methodological quality of studies using Newcastle-Ottawa Quality Assessment Scale. The predefined protocol was registered in PROSPERO #CRD42016049212.

Results: From 1807 abstracts, 79 studies were eligible (297,534 patients): 40 had information on 30-day mortality, of which 6 studies compared stress test versus no stress test. The risk of 30-day mortality was not significant in the comparison of stress testing versus none (RR: 0.79, 95% CI = 0.35-1.80) along with weak evidence for heterogeneity. For the studies that evaluated stress testing without a comparison group, the pooled rates are 1.98% (95% CI = 1.25-2.85) with a high heterogeneity. There was evidence of potential publication bias and small study effects.

Conclusions: Despite substantial interest and research over the past 40 years to predict 30-day mortality risk among patients undergoing non-cardiac surgery, the current body of evidence is insufficient to derive a definitive conclusion as to whether stress testing leads to reduced peri-operative mortality.

Fig 1. Flow chart.

Adapted from: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Iterns for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): el000097. doi:10.1371/journal.pmedl000097. For more information, visit www.prisma-statement.org.

Fig 2. Meta-analysis of 30-day mortality in a comparison of stress test versus no stress test among non-cardiac surgery patients, N = 6 studies.

There are only 6 studies which had both groups (had stress test versus no stress test) among 79 studies.

Fig 3. Meta-analysis of 30-day mortality among non-cardiac surgery patients who received stress test using procedures specific to binomial data and exact methods, N = 40.

ES is effect size- here it is %. Here we calculate the pooled estimate after Freeman-Tukey Double Arcsine Transformation (Freeman, M. F., and Tukey, J. W. 1950) to stabilize the variances.