Insomnia in patients on incident maintenance dialysis and the risk of major acute cardio-cerebrovascular events and all-cause mortality

ABSTRACT

Background

Insomnia is a known risk factor for cardio-cerebrovascular disease in the general population; however, its effect on cardio-cerebrovascular outcomes in end-stage kidney disease patients is unclear. Therefore, this study aimed to investigate the association between cardio-cerebrovascular outcomes and insomnia in patients who initiated maintenance dialysis.

Methods

This study used nationwide Korean health insurance claims data to analyze 79 420 patients who initiated maintenance dialysis from January 2009 to December 2017. Insomnia was defined using claim codes and sleep medication prescription data. Patients were categorized according to the presence of insomnia before and after dialysis initiation: (i) no insomnia, (ii) insomnia before dialysis only (improved insomnia), (iii) insomnia after dialysis only (developed insomnia) and (iv) insomnia in both periods (persistent insomnia). The primary and secondary outcomes were major adverse cardiac and cerebrovascular events (MACCE) and all-cause mortality, respectively. The outcome risks were estimated by Cox regression models with inverse probability of treatment weighting.

Results

The mean age was 61.4 ± 13.4 years, and 39.7% were women. During the transition period from pre-dialysis to maintenance dialysis, 13.2% experienced insomnia. The insomnia groups showed significantly higher risks for MACCE [weighted hazard ratios (95% confidence intervals): developed insomnia, 1.26 (1.25–1.28); improved insomnia, 1.31 (1.29–1.33); persistent insomnia, 1.39 (1.37–1.41)] and higher all-cause mortality risks than the no insomnia group. The insomnia-related cardio-cerebrovascular disease risk elevation was more prominent in younger and male patients.

Conclusions

Insomnia may increase cardio-cerebrovascular disease and all-cause mortality risk among end-stage kidney disease patients who initiate maintenance dialysis.

Graphical Abstract

Video

Watch the video of this contribution at https://academic.oup.com/ndt/pages/author_videos

cardio-cerebrovascular disease, dialysis, ESKD, insomnia, mortality

KEY LEARNING POINTS

What was known:

• •

  Insomnia is common in patients with end-stage kidney disease undergoing maintenance dialysis.

• •

  However, the impact of insomnia on cardio-cerebrovascular disease in patients with end-stage kidney disease is unclear.

This study adds:

• •

  In patients initiating maintenance dialysis, insomnia requiring medication increases the risk of cardiovascular disease and all-cause mortality.

Potential impact:

• •

  Close surveillance and active treatment for insomnia should be considered to improve cardio-cerebrovascular outcome in patients initiating maintenance dialysis.

INTRODUCTION

Insomnia is a common problem among patients with end-stage kidney disease (ESKD). In a study of a national cohort of incident dialysis patients in the USA, 14% of patients had poor sleep quality during the first year of dialysis [1]. In addition, the prevalence of sleep-related disorders, including restless legs syndrome and obstructive sleep apnea, in addition to insomnia, is reported to be higher in patients with ESKD than in the general population. Psychological causes, as well as factors including excessive accumulation of uremic toxins and chronic inflammation, have been identified as contributors to sleep problems in patients with ESKD [2].

In the general population, insomnia has been recognized to be related to poor clinical outcomes. A recent meta-analysis of 23 cohorts revealed that insomnia symptoms are significantly associated with the development of future cardio-cerebrovascular events [3]. Patients receiving dialysis are at an increased risk of developing cardio-cerebrovascular disease compared with the general population [4, 5]. Patient survival is also considerably poorer compared with those with preserved kidney function [6]. Uremic toxins and chronic inflammation, which are factors also implicated in insomnia development, are thought to be responsible for this unfavorable outcome [5, 7]. Despite the overlapping mechanistic features between insomnia and cardiovascular disease, whether the presence of insomnia is associated with an increased risk of cardio-cerebrovascular disease development in patients receiving maintenance dialysis is still not well elucidated.

Therefore, this study examined the association between insomnia and the development of major adverse cardiac and cerebrovascular events (MACCE) in patients who had newly initiated maintenance dialysis by assessing a nationwide cohort of dialysis patients in Korea from a national health insurance database.

MATERIALS AND METHODS

Data source and study population

The Health Insurance Review and Assessment Service (HIRA) is a national organization that reviews and evaluates healthcare costs and quality of care. Healthcare providers in Korea are obligated to participate in this program. Information from the HIRA database, including data on patient demographics, prescriptions, treatments and diagnoses, was reviewed in this study. Study approval was obtained from the institutional review board of the Yonsei University Health System (IRB No. 4-2021-1514). De-identification was performed, and data usage was permitted by the HIRA's national health information data request review committee.

Due to the obligatory copayment assistance policy of the Korean National Health Insurance Service (NHIS), patients who receive maintenance dialysis are coded separately in the HIRA database, enabling accurate identification and selection for this study. Patients aged ≥19 and <85 years who initiated maintenance dialysis between 1 January 2009 and 31 December 2017 were initially screened. Maintenance dialysis was defined as hemodialysis (HD) or peritoneal dialysis (PD) administration lasting at least 90 days. Patients with a history of major psychiatric disorders, malignancy or dementia and those with <90 days of dialysis prior to the primary outcome were excluded. Of the screened patients, 79 420 were included in the final analysis (Fig. 1).

Figure 1:

Study flow.

Definition of insomnia

In order to screen patients with insomnia, patients with insomnia were defined as those who were diagnosed with an International Classification of Diseases, Tenth Revision (ICD-10) code of F51.0 (insomnia not due to a substance or known physiological condition) or G47.0 (insomnia) and who had been prescribed with sleep medication for 7 days or more during the 6 months before or 3 months after dialysis commencement. Sleep medications included doxepin hydrochloride, etizolam, flunitrazepam, flurazepam hydrochloride, lorazepam, quetiapine fumarate, trazodone hydrochloride, triazolam, zolpidem tartrate, alprazolam and eszopiclone (Supplementary data, Table S1). Patients were considered to have insomnia if they were prescribed hypnotics (doxepin hydrochloride, trazodone hydrochloride, zolpidem tartrate and eszopiclone) for 30 days or more regardless of whether they received an ICD-10 code (F51.0 or G47.0) diagnosis. Patients were divided into four groups according to the presence of insomnia in the periods before and after dialysis initiation: (i) no insomnia; (ii) insomnia only in the 6 months before dialysis initiation (improved insomnia); (iii) insomnia only in the 3 months after dialysis initiation (developed insomnia); and (iv) insomnia in both periods (persistent insomnia) (Fig. 1).

Outcomes and covariates

The primary outcome was MACCE, including all-cause mortality, stroke, nonfatal myocardial infarction (MI), acute coronary syndrome (including unstable angina) and heart failure requiring hospital admission. The secondary outcome was all-cause mortality. The observation period was defined as the time from the index date to the first occurrence of MACCE, death or 31 December 2018. The index date was defined as 3 months after dialysis initiation. Demographic data included age, sex, medical aid beneficiaries, comorbidities (diabetes, congestive heart failure, cerebrovascular disease, MI, chronic obstructive pulmonary disease, obstructive sleep apnea and restless legs syndrome) and medication use (any antihypertensive medication or statins). All diagnoses were determined using the ICD-10 diagnostic codes and related procedure and operation codes (Supplementary data, Tables S2 and S3). The study patients were followed from baseline until the date of outcome or end of the study period (31 December 2018). Patients who have undergone kidney transplantation after dialysis initiation were censored at the date of kidney transplantation.

Statistical analysis

Continuous variables are presented as the mean and standard deviation, and categorical variables are expressed as the number and percentage. For the main analysis, inverse probability of treatment weighting (IPTW) based on propensity scores was used to build four weighted groups with different insomnia patterns. The propensity scores were derived from baseline covariates, including age, sex, medical aid beneficiaries, comorbidities, medication use and dialysis modality. A maximum of 10 000 iterations were performed with an iteration stopping point that minimized the absolute standardized mean difference of the effect size (effect size mean). A maximum absolute standardized difference <0.1 was allowed in the baseline covariates between the study groups [8, 9]. After weighting, the weighted incidence rates of MACCE and all-cause mortality were calculated by dividing each weighted event number by the total follow-up duration and presented per 1000 person-years (PY). The risks of each outcome were computed using weighted Cox proportional hazards models with IPTW. All statistical analyses were performed using R (version 3.5.1; https://www.r-project.org/ R Foundation for Statistical Computing, Vienna, Austria) and SAS Enterprise Guide (version 6.1; SAS Institute). The level of significance was set at P < .05.

Subgroup and sensitivity analyses

Subgroup analyses were conducted by stratifying the participants by sex and age (<65 and ≥65 years). According to the stratification, IPTWs were recalculated within the subgroups, and maximum absolute standardized differences were considered as in the main analysis. Additional sensitivity analyses were performed. First, the minimal days of sleep medication use in the definition of insomnia was increased from 7 days to 30 days. Second, analyses were conducted for patients without cardio-cerebrovascular events or cerebrovascular disease within 1 year before dialysis initiation. Third, to minimize the effect of the dialysis modality, analyses were also performed for patients who primarily underwent HD or PD. Those who were prescribed PD after 3 months of dialysis initiation were defined as PD patients. Lastly, additional analyses were performed, wherein individuals with weights exceeding the 97.5th percentile were excluded from the IPTW process (Supplementary data, Table S4). This was undertaken to minimize the potential impact of extreme weights on the analysis [10, 11].

RESULTS

Patient characteristics

The baseline characteristics of the patients are shown in Table 1. Of the 79 420 patients on maintenance dialysis, the mean age was 61.4 ± 13.4 years, 31 542 (39.7%) patients were women and 12 197 (15.4%) patients were on PD. A total of 10 459 (13.2%) patients experienced insomnia during the transition period from pre-dialysis to maintenance dialysis. These patients were classified into four groups: no insomnia (n = 68 961, 86.8%), developed insomnia (n = 5298, 6.7%), improved insomnia (n = 2388, 3.0%) and persistent insomnia (n = 2773, 3.5%). Compared with the no insomnia group, the patients who experienced insomnia at any time before or after dialysis commencement were more likely to be women, medical aid beneficiaries and older. In addition, these patients had more comorbidities and were more frequently prescribed medications, including antihypertensive medication and statins. After IPTW adjustment, the maximum pairwise standardized difference showed that all the variables were well balanced (maximum absolute standardized difference <0.1).

Table 1:

Baseline characteristics of the total study population according to insomnia status.

		Insomnia				Maximum pairwise standardized difference
Characteristics	Total	No insomnia	Developed insomnia	Improved insomnia	Persistent insomnia	Before IPTW	After IPTW
Patients, n (%)	79 420 (100)	68 961 (86.8)	5298 (6.7)	2388 (3.0)	2773 (3.5)
Mean age (SD), years	61.4 (13.4)	60.9 (13.4)	64.3 (12.2)	65.7 (12.3)	66.4 (11.8)	0.411	0.057
Women, n (%)	31 542 (39.7)	27 083 (39.3)	2196 (41.4)	1120 (46.9)	1143 (41.2)	0.156	0.027
Medical aid beneficiaries, n (%)	13 834 (17.4)	11 725 (17.0)	952 (18.0)	533 (22.3)	624 (22.5)	0.145	0.015
Medical conditions, n (%)
Diabetes mellitus	53 754 (67.7)	46 009 (66.7)	3797 (71.7)	1827 (76.5)	2121 (76.5)	0.209	0.030
Congestive heart failure	10 045 (12.6)	8363 (12.1)	787 (14.9)	428 (17.9)	467 (16.8)	0.175	0.008
Cerebrovascular disease	9483 (11.9)	7812 (11.3)	724 (13.7)	454 (19.0)	493 (17.8)	0.238	0.007
Myocardial infarction	2152 (2.7)	1755 (2.5)	191 (3.6)	84 (3.5)	122 (4.4)	0.116	0.017
Peripheral vascular disease	7005 (8.8)	5811 (8.4)	561 (10.6)	309 (12.9)	324 (11.7)	0.160	0.004
Chronic obstructive pulmonary disease	14 618 (18.4)	11 974 (17.4)	1214 (22.9)	619 (25.9)	811 (29.2)	0.309	0.012
Obstructive sleep apnea	381 (0.5)	319 (0.5)	23 (0.4)	18 (0.8)	21 (0.8)	0.047	0.025
Restless legs syndrome	626 (0.8)	509 (0.7)	47 (0.9)	24 (1.0)	46 (1.7)	0.105	0.020
Medication use, n (%)
Antihypertensive medication	76 439 (96.2)	66 261 (96.1)	5126 (96.8)	2333 (97.7)	2719 (98.1)	0.103	0.030
Statin	53 934 (67.9)	46 366 (67.2)	3716 (70.1)	1735 (72.7)	2117 (76.3)	0.195	0.020
PD, n (%)	12 197 (15.4)	11 095 (16.1)	536 (10.1)	278 (11.6)	288 (10.4)	0.164	0.019

SD, standard deviation.

Insomnia and the risk of MACCE

During a median follow-up of 2.9 years (interquartile range 1.5–5.2 years), 39 543 MACCE and 27 920 cases of all-cause mortality occurred. Compared with the no insomnia group, the weighted incidence rates for MACCE were higher for patients who experienced insomnia at any time before or after initiating dialysis [crude event number (weighted rate per 1000 PY): no insomnia, 33 223 (134.6); developed insomnia, 3079 (168.5); improved insomnia, 1504 (175.0); persistent insomnia, 1737 (184.9)] (Table 2). Insomnia experienced at any time before or after initiating dialysis was related to higher risks of MACCE: the weighted hazard ratios (HRs) were 1.26 [95% confidence interval (CI) 1.25–1.28] for the developed insomnia group, 1.31 (95% CI 1.29–1.33) for the improved insomnia group and 1.39 (95% CI 1.37–1.41) for the persistent insomnia group (Table 3). Similarly, insomnia was associated with higher risks of all-cause mortality in all insomnia groups compared with the no insomnia group: the weighted HRs were 1.31 (95% CI 1.29–1.33) for the developed insomnia group, 1.41 (95% CI 1.40–1.43) for the improved insomnia group and 1.47 (95% CI 1.44–1.49) for the persistent insomnia group (Table 3).

Table 2:

Overall event rates in the study cohort.

Outcome	Number of individuals, n	Number of events, n	Number of events after IPTW	Crude IR (per 1000 PY)	Weighted IR (per 1000 PY)
MACCE
No insomnia	68 961	33 223	38 975.6	131.2	134.6
Developed insomnia	5298	3079	42 536.5	183.1	168.5
Improved insomnia	2388	1504	42 372.3	206.1	175.0
Persistent insomnia	2773	1737	43 077.9	219.2	184.9
All-cause mortality
No insomnia	68 961	23 138	27 272.8	85.5	88.0
Developed insomnia	5298	2292	31 391.3	126.1	115.6
Improved insomnia	2388	1170	32 251.9	147.9	124.9
Persistent insomnia	2773	1320	32 171.3	155.4	130.0

IR, incidence rate.

Table 3:

MACCE and all-cause mortality after IPTW.

Outcomes	HR (95% CI)	P-value
MACCE
Developed insomnia vs no insomnia	1.26 (1.25–1.28)	<.001
Improved insomnia vs no insomnia	1.31 (1.29–1.33)	<.001
Persistent insomnia vs no insomnia	1.39 (1.37–1.41)	<.001
All-cause mortality
Developed insomnia vs no insomnia	1.31 (1.29–1.33)	<.001
Improved insomnia vs no insomnia	1.41 (1.40–1.43)	<.001
Persistent insomnia vs no insomnia	1.47 (1.44–1.49)	<.001

Subgroup and sensitivity analyses

Subgroup analyses were performed by sex and age (<65 or ≥65 years). The weighted incidence rates and weighted HRs show that the associations between insomnia and MACCE and all-cause mortality risk were maintained in all subgroups (Supplementary data, Table S5). However, the increased risk was more prominent in younger patients (<65 years). The sensitivity analysis, which utilized a more stringent definition of insomnia (sleep medication for 30 days or more), consistently yielded similar results (Supplementary data, Table S6). The analysis, which included patients who did not experience any cardio-cerebrovascular events in the 1 year before dialysis commencement, also showed comparable findings (Supplementary data, Table S7). Similarly, analyses excluding individuals with weights surpassing the 97.5th percentile in the IPTW process demonstrated consistent outcomes (Supplementary data, Table S8). When evaluating risk based on dialysis modality, prevalence of insomnia was higher among HD patients compared with PD patients. Nevertheless, insomnia retained its significance as a risk factor for MACCE and all-cause mortality, irrespective of dialysis modality (Table 4).

Table 4:

MACCE and all-cause mortality after IPTW in patients according to the dialysis modalities.

Outcomes		Number of individuals, n (%)	Number of events, n (%)	Weighted IR (per 1000 PY)	HR (95% CI)
MACCE
HD	No insomnia	57 866 (86.1)	28 283 (48.9)	139.0	1.00 (Reference)
	Developed insomnia	4762 (7.1)	2761 (58.0)	170.7	1.24 (1.22–1.26)
	Improved insomnia	2110 (3.1)	1328 (62.9)	181.3	1.32 (1.30–1.34)
	Persistent insomnia	2485 (3.7)	1542 (62.1)	194.3	1.41 (1.39–1.43)
PD	No insomnia	11 095 (91.0)	4940 (44.5)	112.8	1.00 (Reference)
	Developed insomnia	536 (4.4)	318 (59.3)	159.2	1.44 (1.39–1.49)
	Improved insomnia	278 (2.3)	176 (63.3)	145.8	1.30 (1.26–1.35)
	Persistent insomnia	288 (2.4)	195 (67.7)	158.0	1.41 (1.36–1.47)
All-cause mortality
HD	No insomnia	57 866 (86.1)	19 668 (34.0)	90.6	1.00 (Reference)
	Developed insomnia	4762 (7.1)	2038 (42.8)	116.3	1.28 (1.26–1.30)
	Improved insomnia	2110 (3.1)	1032 (48.9)	130.6	1.43 (1.41–1.46)
	Persistent insomnia	2485 (3.7)	1151 (46.3)	133.2	1.45 (1.43–1.48)
PD	No insomnia	11 095 (91.0)	3470 (31.3)	75.3	1.00 (Reference)
	Developed insomnia	536 (4.4)	254 (47.4)	112.9	1.51 (1.44–1.57)
	Improved insomnia	278 (2.3)	138 (49.6)	101.0	1.33 (1.28–1.39)
	Persistent insomnia	288 (2.4)	169 (58.7)	126.9	1.70 (1.63–1.78)

IR, incidence rate.

DISCUSSION

In this nationwide retrospective cohort study of 79 420 incident dialysis patients, insomnia was associated with an increased risk of cardio-cerebrovascular disease and all-cause mortality. Patients who experienced insomnia at any time before or after the initiation of dialysis showed a higher risk of cardio-cerebrovascular disease and all-cause mortality compared with those without insomnia. This finding was still significant after adjusting for confounding variables, including demographic characteristics, medical conditions and medication use.

The risk of cardio-cerebrovascular disease in ESKD patients is associated with older age, diabetes, hypertension, dyslipidemia and dialysis-related factors, including anemia, volume overload, mineral metabolism, vascular calcification, systemic inflammation and infection [12, 13]. The effect of insomnia on cardio-cerebrovascular disease outcomes in patients with ESKD has not been studied. Previous studies have suggested that sleep problems may affect cardio-cerebrovascular risk in patients with ESKD [1, 14, 15]; however, these studies only examined the effect of poor sleep on overall mortality and did not investigate insomnia separately from other sleep disorders, for which treatment may be different. Furthermore, no study has researched the effect of insomnia on cardio-cerebrovascular risk according to the duration of insomnia.

The results of this study suggest that the effect of insomnia on cardio-cerebrovascular risk in ESKD patients is dose-dependent. Studies of community-based cohorts have shown a dose-dependent relationship between the duration of insomnia and general mortality as well as neurological or psychiatric morbidity [16–18]. Insomnia contributes to inflammatory response activation via the interaction between the central nervous system and the immune system [19]. Therefore, chronic insomnia may contribute to an increased risk of cardio-cerebrovascular disease and general mortality by activating the sympathetic nervous system and pro-inflammatory cytokines. Insomnia also increases the risk of cardio-cerebrovascular disease through various risk factors intermediated by sympathetic nervous system overactivity and hypothalamic–pituitary–adrenal dysregulation [20–23]. However, in the present study, individuals with improved insomnia had a similar risk of MACCE or even a higher risk of all-cause mortality compared with those with persistent insomnia. Although the exact mechanism of this phenomenon is difficult to explain, a study using claims data in Taiwan showed no difference in the risk of stroke between improved insomnia and persistent insomnia [16]. A well-designed study is needed to determine whether insomnia treatment can reduce the risk of cardio-cerebrovascular disease or death.

This study found that the effect of insomnia on the risk of cardio-cerebrovascular disease and general mortality was more prominent among under 65 years of age than among older patients. Similar findings were observed in previous studies. A prospective cohort study of approximately 0.5 million Chinese adults showed a prominent association of insomnia with cardiovascular disease incidence in younger adults [24]. Another cohort study of 85 752 adults also showed that the insomniacs/non-insomniacs incidence rate ratio for stroke was higher among young adults than among older adults [16]. Older adult patients tend to have more comorbidities (e.g. diabetes, heart failure, atherosclerotic heart disease, peripheral vascular disease, stroke and other cardiac diseases) than younger patients at the time of ESKD onset [25], which might attenuate the effect of insomnia on MACCE in older patients.

In this study, the prevalence of insomnia during the transition period from pre-dialysis to maintenance dialysis was found to be 13.2%, which is somewhat lower than the prevalence reported in previous studies involving patients receiving maintenance dialysis. A recent systematic review indicated that more than 20% of ESKD patients experience insomnia [26]. However, it is important to note that the definition used to define insomnia varied for each study included in the meta-analysis. In this study, insomnia was defined using a rather stringent criterion. Individuals who had symptoms severe enough to warrant a clinical diagnosis and who required medication for their condition were categorized as having insomnia. Furthermore, the studies included in that systematic review involved prevalent ESKD patients rather than incident patients who had recently initiated dialysis. A recent report that assessed sleep quality in ESKD patients within a year after dialysis initiation found the prevalence of insomnia to be 14% [1], which is comparable to the prevalence observed in the current study.

Differences in the prevalence of insomnia by dialysis method are controversial. A meta-analysis of 93 studies on the prevalence of insomnia in patients with chronic kidney disease suggested no significant difference in the prevalence of insomnia by dialysis method [27]. However, of the studies included in the meta-analysis, only one compared HD patients with PD patients, and most studies were conducted only with HD patients. A study of 154 patients on dialysis revealed that insomnia was more prevalent in PD patients than in HD patients [28]. Thus, it may be premature to conclude from the meta-analysis alone that there is no difference in the prevalence of insomnia according to the method of dialysis.

Considering the risk factors for insomnia in dialysis patients, including pain associated with dialysis, uremic toxins, fluid accumulation, restless legs syndrome, obstructive sleep apnea, reduced melatonin levels, change in circadian rhythm and comorbidities [2], ESKD patients are a high-risk group for the insomnia-mediated cardio-cerebrovascular disease. Guidelines for insomnia treatment recommend cognitive behavioral therapy as the most effective strategy, but few studies have investigated the effect of cognitive behavioral therapy on insomnia in patients with ESKD [29]. Two studies in which cognitive behavioral therapy for insomnia was administered for 3–4 weeks in patients with ESKD demonstrated improvements in insomnia symptoms but not in inflammatory markers [30]. Nonetheless, the short duration of the study and the fact that sleep quality improvements after treatment was associated with changes in brain connectivity and decreased interleukin-1β suggest the need for future studies with a larger number of patients over a sufficient study period.

The present study has several strengths. First, this study evaluated a large population from a nationwide dialysis cohort. More than 98% of the Korean population is enrolled in the mandatory NHIS program, and the remaining people receive government benefits as they are included in the lowest income bracket (medical aid beneficiaries). Due to the obligatory copayment assistance policy of the Korean NHIS, dialysis patients are coded separately in the HIRA database, which enabled the detection of a nationwide population of patients undergoing maintenance dialysis. Second, to minimize potential bias due to nonrandomized group allocation, this study used IPTW based on propensity scores. Because insomnia and its pharmacological treatment are dependent on underlying comorbidities in ESKD patients, such a strategy further strengthens the possibility that there is an independent association between insomnia and the risk of MACCE.

The findings of this study should be interpreted considering the following limitations. First, this study was observational; thus, although the results revealed a significant association between insomnia and the risk of MACCE, future prospective evaluations should further assess the cause–effect relationship. Second, the characteristics of insomnia, including sleep duration, sleep depth and timing, could not be evaluated because the patients with insomnia were not subject to questionnaire-based assessments. Third, it was difficult to follow the progress of insomnia after the initiation of pharmacological treatment due to the limitations of the data used in this study. Fourth, as this study evaluated the association between insomnia and MACCE using claims data from a national insurance service database, potential confounding variables—including lifestyle factors, anthropometric factors and laboratory information—could not be examined. Fifth, the HIRA database only provides access to claimed reimbursable medical expenditures due to the policy of the Korean NHIS. Therefore, over-the-counter sleep medication (diphenhydramine) and non-reimbursable sleep medication (melatonin) could not be considered in this study, and the incidence of insomnia may have been underestimated. Lastly, dialysis-related factors, including the type of vascular access, dialysis time, electrolyte homeostasis, acidosis correction, residual renal function, ultrafiltration, dialysis adequacy and method of peritoneal dialysis, were not available in this study due to the nature of the claims database. Further investigation of the factors affecting insomnia in ESKD patients is warranted.

In conclusion, this nationwide observational study showed that insomnia was associated with an increased risk of MACCE and all-cause mortality in patients receiving maintenance dialysis. Because patients with persistent insomnia before and after dialysis commencement were at the highest risk of MACCE and all-cause mortality, active insomnia management in patients who recently started or will soon start dialysis may improve outcomes. Further well-designed studies that can more accurately assess insomnia are needed to verify the association between insomnia and cardio-cerebrovascular risk among patients receiving maintenance dialysis.

ACKNOWLEDGEMENTS

This study used research data (M20210923520) from the Health Insurance Review and Assessment Service (HIRA). The views expressed are those of the authors and not necessarily those of the HIRA or the Korean Ministry of Health and Welfare.

FUNDING

The National Research Foundation of Korea, funded by the Ministry of Science, ICT & Future Planning, Republic of Korea, supported the present study (grant numbers: 2017R1A2B3008214, 2022R1A2B5B03002611). The funding source was not involved in the study design, data collection or writing of the report.

AUTHORS’ CONTRIBUTIONS

Conceptualization: E.L., J.T.P.; data curation, formal analysis: H.W.K.; investigation: H.W.K., G.Y.H., H.J.K., S.-W.K., J.T.P. and E.L.; writing–original draft: H.W.K.; writing–review and editing: H.W.K., G.Y.H., H.J.K., S.-W.K., J.T.P. and E.L.; all authors read and approved the final manuscript.

DATA AVAILABILITY STATEMENT

The datasets generated during and/or analyzed during the current study are not available. However, the original data that support the findings of this study are available from HIRA. Data are available from the authors upon request and with permission of HIRA.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no competing interests.