Accuracy of Point-of-Care Blood Glucose Meters Compared With Central Laboratory Reference Methods in Critically Ill AdultsA Systematic Review

Abstract

Background: Point-of-care (POC) blood glucose meters shorten turnaround time for bedside glycemic management, but critical illness introduces physiologic and analytical conditions that can degrade agreement with central laboratory measurements. Objective: To synthesize evidence on the analytical and clinical accuracy of POC blood glucose meters versus central laboratory reference methods in critically ill adults and identify conditions associated with disagreement. Methods: Searches of PubMed, Google Scholar, and SciSpace identified 1,003 records published from 2010 through September 13, 2026. After removal of 280 duplicates, 723 unique records remained. An initial screening limit left 223 deduplicated records without title/abstract assessment; a subsequent completeness audit applied the same eligibility criteria to these records and re-evaluated full-text decisions. Thirty studies were included. Risk of bias was assessed with an adapted QUADAS-2 framework. Results: POC performance was strongly device-, specimen-, and context-dependent. Several modern systems showed small average bias and high standards compliance in selected ICU cohorts, whereas capillary measurements deteriorated substantially in shock, hypotension, catecholamine/vasopressor exposure, anemia or hematocrit extremes, acidemia, hypoglycemia, and selected drug/interference states. Reported ISO compliance ranged from very high in stable or validated device-specific conditions to markedly poor in shock subgroups; these percentages were not directly pooled because studies used different devices, specimens, standards, thresholds, and outcome metrics. Conclusion: POC glucose meters cannot be treated as analytically interchangeable with central laboratory testing across all critically ill adults. Risk-stratified confirmation with laboratory testing is warranted when POC results are clinically discordant or obtained under physiologic conditions known to impair meter performance.