Perioperative mortality in older patients: a systematic review with a meta-regression analysis and meta-analysis of observational studies

Abstract

Study objective: Older patients have a higher probability of developing major complications during the perioperative period than other adult patients. Perioperative mortality depends on not only on a patient condition but also on the quality of perioperative care provided. We tested the hypothesis that the perioperative mortality rate among older patients has decreased over time and is related to a country's Human Development Index (HDI) status. Design: A systematic review with a meta-regression and meta-analysis of observational studies that reported perioperative mortality rates in patients aged ≥60 years was performed. We searched the PubMed, EMBASE, LILACS and SciELO databases from inception to December 30, 2019. Setting: Mortality rates up to the seventh postoperative day were evaluated. Measurements: We evaluated the quality of the included studies. Perioperative mortality rates were analysed by time, country HDI status and baseline American Society of Anesthesiologists (ASA) physical status using meta-regression. Perioperative mortality and ASA status were analysed in low- and high-HDI countries during two time periods using proportion meta-analysis. Main results: We included 25 studies, which reported 4,412,100 anaesthesia procedures and 3568 perioperative deaths from 12 countries. Perioperative mortality rates in high-HDI countries decreased over time (P = 0.042). When comparing pre-1990 to 1990–2019, in high-HDI countries, the perioperative mortality rates per 10,000 anaesthesia procedures decreased 7.8-fold from 100.85 (95% CI 43.36 to 181.72) in pre-1990 to 12.98 (95% CI 6.47 to 21.70) in 1990–2019 (P < 0.0001). There were no studies from low-HDI countries pre-1990. In the period from 1990 to 2019, perioperative mortality rates did not differ between low- and high-HDI countries (P = 0.395) but the limited number of patients in low-HDI countries impaired the result. Perioperative mortality rates increased with increasing ASA status (P < 0.0001). There were more ASA III-V patients in high-HDI countries than in low-HDI countries (P < 0.0001), and the perioperative mortality rate increased 24-fold in ASA III-V patients compared with ASA I-II patients (P < 0.0001). Conclusion: The perioperative mortality rates in older patients have declined over the past 60 years in high-DHI countries, highlighting that perioperative safety in this population is increasing in these countries. Since data prior to 1990 were lacking in low-HDI countries, the evolution of their mortality rates could not be analysed. The perioperative mortality rate was similar in low- and high-HDI countries in the post-1990 period, but the low number of patients in the low-HDI countries does not allow a definitive conclusion.