Nephrology ·
CRRT dose estimates should not be pooled across different models, authors say
In a published response, authors of a retrospective cohort study corrected two marginal structural model estimates and explained why these should not be pooled with conventional Cox estimates. They said the corrected estimates were not statistically significant and that the 13 mL/kg/h threshold should be considered exploratory, not evidence of a minimum safe dose.
- Corrected dose-threshold estimates were not statistically significant.
- Dose categories often changed during successive three-day blocks.
- Different Cox and marginal structural estimates should not be pooled.
- Authors say 13 mL/kg/h is not a proven minimum safe dose.
The authors of a retrospective cohort study of continuous renal replacement therapy (CRRT) dosage and mortality responded to letters about their analysis and a related systematic review. They corrected a transposition in the original Results text: for dose thresholds of 13 and 20 mL/kg/h, the marginal structural model hazard ratios were 1.08 (95% CI 0.74–1.58) and 0.87 (95% CI 0.66–1.14), respectively. Neither estimate was statistically significant, and the authors said the correction does not change the study’s conclusion.
They argued that CRRT dose is adjusted over time in response to changing clinical factors that also predict mortality and may be affected by earlier treatment. They reported that only about half of patients remained in the same dose category across consecutive three-day blocks; patients starting below 13 or 20 mL/kg/h usually had their dose increased above 20 mL/kg/h within the first week. The authors therefore cautioned that whole-course mean dose with baseline-only adjustment is vulnerable to time-dependent confounding.
The authors agreed that marginal structural and conventional Cox hazard ratios estimate different effects and should not be pooled. They also noted that a pooled estimate of 1.72 from two studies using the same baseline-adjusted approach, with I²=0%, indicates similarity but does not rule out shared bias. They said the 13 mL/kg/h threshold should be treated as exploratory, not as evidence of a minimum safe CRRT dose. Three ongoing randomized trials are expected to inform the question.
In practice
For clinicians interpreting CRRT dose studies, this response cautions against treating 13 mL/kg/h as an established minimum safe dose: the corrected marginal structural estimate at that threshold was not statistically significant, and the authors describe substantial dose changes over time alongside time-dependent confounding. Marginal structural and conventional Cox estimates target different estimands and should not be pooled. This is a methodological clarification of observational evidence, not a randomized comparison establishing a dose target; the authors point to three ongoing randomized trials.
AI-generated clinical interpretation · Not editor-reviewed
Is this summary clinically accurate?
Help fellow clinicians: your rating sends inaccurate summaries straight to our editors.
Sign in to rate this summary →Source
Critical Care: Time-varying dose decisions in CRRT: why marginal structural and baseline Cox estimates should not be pooled ↗This is an automated AI-condensed summary that has not yet been reviewed by an editor. Always consult the full item at the original source.