Limited Interchangeability of Shunt-Side SpO₂ Measurements During Hemodialysis

During hemodialysis, SpO₂ is measured on the non-shunt side, restricting movement. This single-center, prospective, observational study evaluated shunt-side interchangeability.

Forty-five patients underwent double-sided finger-sensor SpO₂ monitoring at five time points. Between-site agreement was assessed using Lin's concordance correlation coefficient (CCC), Deming regression, Bland-Altman analysis, and accuracy root mean square (Arms). Confidence intervals were calculated using subject-level cluster bootstrap resampling (2000 iterations; seed = 42). Using vector-type visualization, relationships between paired measurements were illustrated.

SpO₂ showed similar distributions between sites. The Spearman correlation was ρ = 0.705. CCC was 0.720 (95% CI: 0.590-0.811). Deming regression yielded y = 0.989x + 1.087. Bland-Altman bias was 0.060% (limits of agreement, ±3%). Arms were 1.41%, and the visualization showed minimal angular deviation (median |θ-45°| = 0.29°).

Shunt-side SpO₂ may be acceptable in selected clinical contexts, although interindividual variability warrants caution. These findings are not applicable to hypoxic states (SpO₂ < 90%).