Hemoglobin Measurement by Point-of-Care Blood Gas Analysis Versus Central Laboratory in Hemodialysis Patients

Background: In hemodialysis patients, precise hemoglobin (Hb) monitoring is essential for anemia management. Point-of-care blood gas analyzers (BGAs), such as the ABL800 Flex, offer rapid Hb determinations, but their accordance and comparability with central laboratory measurements remains to be assessed in the hemodialysis setting. Methods: We performed a retrospective analysis (April 2017-February 2024) of 10,802 paired Hb measurements from 291 hemodialysis patients. BGA and laboratory values within 90 min were compared using paired t-tests, non-inferiority testing (margin 0.5 g/dL), a Bland-Altman analysis, and linear regression. Results: The mean ± standard deviation Hb (g/dL) values were 10.14 ± 1.64 (BGA) versus 9.90 ± 1.55 (laboratory). The overall mean difference (BGA-laboratory) was 0.24 ± 0.49 g/dL (95% CI: 0.23-0.25), demonstrating non-inferiority (p < 0.0001). Measurement delay correlated with increasing analysis discrepancies (mean difference in g/dL: 0.22 at <30 min vs. 0.27 at 60-90 min; p < 0.001). We derived the equation of laboratory Hb = 0.90 × BGA Hb + 0.72; a simplified correction (BGA-0.3 g/dL) produced a mean absolute error (MAE) of 0.30 g/dL and root mean square error (RMSE) of 0.50 g/dL, and patient-level 10-fold cross-validation yielded MAE ≈ 0.30 and RMSE ≈ 0.49 g/dL. The Bland-Altman analysis confirmed a small systematic bias of 0.24 g/dL with 95% limits of agreement ranging from -0.73 to +1.21 g/dL. Conclusions: BGA Hb measurements via the ABL800 Flex are non-inferior to central laboratory values across clinical scenarios, with minimal bias. After regression correction, the estimated total error was ≈0.78 g/dL. If hemodialysis centers accept this level of total error and apply confirmatory testing near decision points, BGA could be used to guide anemia management.