Three sleep phenotypes emerged across 667 maintenance hemodialysis patients
Sleep in maintenance hemodialysis is not a single problem with a single face. In a multicenter cross-sectional study of 667 patients treated at seven hospitals in eastern China, latent class analysis separated sleep complaints into three distinct patterns rather than the usual good-versus-bad split used in many studies.
The largest group, accounting for 52.8%, fit a good overall sleep quality profile. A second, larger symptomatic cluster represented 36.7% of participants and was characterized as insomnia with daytime dysfunction. The smallest group, 10.5%, had poor overall sleep quality with low sleep efficiency. Those classes matter because dialysis patients do not simply sleep “well” or “poorly”; they may struggle with sleep initiation, fragmented nights, low restoration, or disabling daytime fatigue in different combinations.
A binary screen can miss that heterogeneity. In a population already burdened by uremia, anemia, fluid shifts, vascular disease, and repeated treatment sessions, the difference between poor sleep efficiency and daytime dysfunction has direct bedside consequences for cognition, adherence, and symptom reporting.
Mood symptoms and age clustered with the most burdensome sleep profiles
Investigators used the seven components of the Pittsburgh Sleep Quality Index to build the latent classes, then examined associated factors with multinomial logistic regression. The pattern that emerged was clinically familiar: older age, anxiety symptoms, and depressive symptoms tracked with the insomnia and daytime dysfunction class when compared with the good-sleep reference group. Older age, smoking, and depressive symptoms also tracked with the class defined by poor overall sleep quality and low sleep efficiency.
These associations fit the lived physiology of dialysis care. Older adults often carry more comorbidity, more medication burden, and more sleep fragmentation from pain, nocturia, or restless legs symptoms. Anxiety can prolong sleep latency and intensify nocturnal hyperarousal, while depression can blunt sleep continuity and deepen perceived fatigue during the day. Smoking adds another layer of arousal biology, with nicotine exposure disrupting sleep architecture and making consolidated rest less likely.
Income entered the model as well. Higher personal monthly income, defined as greater than 2000 China Yuan, was negatively associated with the poor overall sleep quality with low sleep efficiency class. In practical terms, financial security may ease access to transportation, home supports, better symptom management, or more stable routines, all of which can shape sleep continuity even in a high-burden chronic disease population.
Dialysis vintage signaled chronic strain, but the signal depended on the cutoff
Prolonged dialysis vintage stood out as one of the more striking findings. A dialysis duration of 15 years or longer was strongly associated with both the insomnia/daytime dysfunction class and the poor sleep efficiency class. The relationship, however, was fragile. When the cutoff was shifted to 10 years, the association disappeared.
That sensitivity is important because it warns against treating dialysis vintage as a fixed biological threshold. Long-term exposure to hemodialysis may still reflect cumulative symptom burden, vascular access challenges, inflammation, and repeated disruptions to daily life, but the effect was not stable across analytic definitions. Clinicians should read the result as a signal of chronic burden rather than a hard line separating risk from safety.
The abrupt disappearance of significance at a different cutoff also invites caution in bedside interpretation. A statistical association that changes with threshold choice can inform hypothesis generation, but it should not be promoted as a universal marker of sleep risk without replication in other cohorts.
Why sleep quality deserves phenotype-based screening in dialysis care
The physiologic importance of sleep in maintenance hemodialysis extends beyond rest. Poor sleep can worsen fatigue, mood instability, pain perception, and daytime functioning, and these effects can compound the already heavy symptom load of end-stage kidney disease. Sleep disruption also interacts bidirectionally with depression and anxiety, making the patient feel sicker, less active, and less able to tolerate treatment demands.
Phenotype-based screening offers a more useful clinical frame than a single total score. A patient with daytime dysfunction may need assessment for sleep apnea, nocturnal discomfort, circadian disruption, or depression-driven exhaustion. Another patient with low sleep efficiency may need closer attention to smoking, age-related fragility, or cumulative dialysis burden. The class structure in this study makes those distinctions visible.
The study’s cross-sectional design limits causal inference, so the data do not show whether anxiety and depression drive the sleep patterns or whether fragmented sleep aggravates mood symptoms first. Still, the associations are strong enough to justify targeted questioning at the chairside rather than a generic sleep query.
What the data can and cannot support
This cohort came from seven hospitals, which strengthens the generalizability within the sampled region, and the sample size was substantial for a latent class analysis. Sleep quality, depressive symptoms, and anxiety symptoms were measured with established instruments: the Pittsburgh Sleep Quality Index, the Patient Health Questionnaire-9, and the Generalized Anxiety Disorder-7. Those tools provide a structured snapshot of symptom burden, but they remain self-reported and cannot substitute for objective sleep testing when sleep apnea, periodic limb movement, or other primary sleep disorders are suspected.
The study also does not prove that intervening on smoking, mood, or social stress will move a patient from one sleep class to another. It does, however, sharpen the clinical map. Older patients, smokers, and those with anxiety or depressive symptoms deserve closer sleep surveillance, and dialysis teams should stop assuming that every sleep complaint in this setting is the same complaint.
For nephrology clinics, the most actionable message is simple: sleep assessment in hemodialysis should be pattern-based, not score-based alone. The published findings in PLOS One support a more granular approach to symptom review, especially in patients with long dialysis exposure and coexisting mood symptoms.
Citation
Sleep quality patterns and associated factors in patients undergoing maintenance hemodialysis: A latent class analysis by Ma R, Deng S, Zhou J, Liu Y, Xu Q, Feng X, and Hu H. PLoS One. 2026;21(11):e0362676. DOI: 10.1371/journal.pone.0362676
- • PLOS One Study DOI 10.1371/journal.pone.0362676: https://doi.org/10.1371/journal.pone.0362676