Comment on “the impact of sleep duration on the incidence of new-onset chronic kidney disease in community-dwelling adults: a nationwide cohort study”

Comment on "the impact of sleep duration on the incidence of new-onset chronic kidney disease in community-dwelling adults: a nationwide cohort study"To the Editor I read with interest the KoGES-based analysis reporting a U-shaped association between self-reported sleep duration and incident CKD [1].The topic is clinically relevant, but several issues warrant caution.First, even with multivariable adjustment, sleep duration is strongly intertwined with socioeconomic status, occupational factors (e.g.shift work), mental health, obstructive sleep apnea, frailty, and medication use-determinants that also shape CKD risk and may not be adequately captured by standard covariates.Importantly, "long sleep" can represent underlying illness, depression, inactivity, or inflammatory burden rather than a causal exposure.Excluding CKD events in the first 2 years helps, but does not fully address reverse causation when subclinical kidney dysfunction, anemia, uremic symptoms, or cardiometabolic disease precede overt CKD by many years.This concern is consistent with the broader literature in which both short and long sleep track with adverse cardiometabolic profiles and mortality, often interpreted as partially non-causal signals [2].Second, the analysis appears to rely on baseline sleep only.In long follow-up, sleep duration frequently changes with retirement, incident diabetes/hypertension, depression, and aging.Treating sleep as fixed risks exposure misclassification and may bias hazard ratios toward or away from the null depending on how sleep evolves relative to emerging CKD risk.Repeated-measures approaches (time-updated Cox models, joint modeling, or marginal structural models) could better align etiologic inference with a dynamic exposure and help distinguish "persistent" short/long sleep from transient patterns.Given evidence that sleep behaviors are modifiable and fluctuate over time, capturing trajectories may be more clinically meaningful than a single baseline report [3].Third, the fully adjusted model includes diabetes, hypertension, BMI, and CRP.If inadequate sleep contributes to CKD partly through cardiometabolic pathways and inflammation, these variables may lie on the causal pathway.Adjusting for intermediates can attenuate the total effect and complicate interpretation (estimating something closer to a controlled direct effect).Presenting a prespecified causal diagram and reporting effect estimates under sequential adjustment (confounders-only vs confounders + potential mediators) would clarify what estimand is being targeted.This is not a statistical quibble; it directly affects how clinicians interpret whether "improving sleep" might plausibly reduce CKD risk versus merely serving as a risk marker [4].Finally, the subgroup patterns by age and sex are provocative, yet interaction testing, multiplicity control, and power in extreme sleep strata are critical.Without formal interaction terms and transparent reporting of the number of subgroup comparisons, there is a risk of overinterpreting chance findings.Given prior reports linking sleep duration with kidney outcomes, a focused set of a priori interactions, rather than extensive stratification, would strengthen credibility and reduce false-positive concerns [5].In summary, the study adds valuable long-term data, but causal interpretation is limited by residual confounding, time-varying exposure concerns, and estimand ambiguity from adjusting for likely mediators.Future analyses incorporating repeated sleep assessments, explicit causal modeling, and disciplined interaction testing would improve clinical interpretability and guide whether sleep duration should be targeted as an intervention point or used primarily for risk stratification.