Kidney transplantation (KT) represents the optimal treatment for end-stage kidney disease (ESKD). Patients who undergo KT experience a markedly greater survival benefit compared to those who remain on long-term dialysis [1,2].

KT has been shown to reduce cardiovascular events compared to dialysis [3]. Due to the absence of the physical limitations that impact patients on hemodialysis (HD), recipients of KT tend to be more active in daily life [4]. Additionally, the socioeconomic costs associated with KT are lower, and the quality of life is generally better than that experienced by patients undergoing dialysis [5]. These advantages contribute to psychological improvements in patients who have received KT. A study by Alavi et al. [6] found that depression is more prevalent among patients on HD than among those who have undergone KT.

Despite the apparent positive psychological impact of KT in patients with ESKD, the literature includes few comparative studies of psychiatric disorders between patients treated with dialysis and those who have undergone KT. A study involving 100 patients in Brazil reported no significant differences in the prevalence of suicidal ideation and depression symptoms between HD and transplant recipients [7]. Until a sufficiently large sample of participants is examined, it is challenging to definitively conclude that no correlation exists between the type of kidney replacement therapy and the incidence of suicide or depression. A recent nationwide cohort study in Taiwan identified insomnia as an independent predictor of suicide attempts [8]. Furthermore, the prevalence of insomnia is markedly lower among transplant recipients compared to patients undergoing dialysis on the transplant waitlist [9].

Accordingly, we extracted data from the Korean National Health Insurance Service (NHIS) regarding a nationwide population of patients on dialysis and KT recipients. We then compared the rates of depression and suicide between these two groups.

The study received approval from the Institutional Review Board of Wonju Severance Christian Hospital (IRB No. CR321375).

Study Population

For this study, we employed a retrospective cohort design, drawing on data from the NHIS. As South Korea’s sole public health insurance system, the NHIS mandates that all disease diagnoses, as well as prescription information for drugs and procedures, be recorded in its database. This registration process allows patients to receive financial support from the government for their treatment costs. Disease diagnoses are entered into the NHIS database using the International Classification of Diseases, 10th Revision (ICD-10) codes. Additionally, we utilized the Statistics Korea database to obtain the date and primary cause of death up to December 31, 2018. In South Korea, physicians are legally required to register the primary disease or main cause of death in the Statistics Korea database. The NHIS database has been widely used in epidemiological studies, and its validity has been thoroughly examined in other publications [10]. For the present investigation, we first secured authorization from the NHIS to extract data on demographics, socioeconomic status, disease diagnoses, and treatment information from their database, which provided these details for the study population. We then received approval from Statistics Korea to access data on the dates and primary causes of death among the study participants. We linked the two databases using the resident registration number, a compulsory identification number assigned to all residents of South Korea, irrespective of nationality.

Data were obtained from the NHIS database for the years 2002 through 2018. We excluded 296,117 patients who lacked a specific dialysis code (V001 or V003) or kidney transplant code (V005 or R3280). Additionally, 2,698 patients who had died at baseline were excluded. A wash-out period from January 2002 to December 2009 resulted in the exclusion of 120,368 patients. We also excluded 169,519 patients who had been diagnosed with depression or insomnia prior to undergoing dialysis or KT. Furthermore, 3,166 individuals under the age of 19 were consecutively excluded. Ultimately, the final study population comprised 21,809 individuals with ESKD (Fig. 1).

Figure 1. Flow diagram showing selection of the study population from the Korean National Health Insurance Service (NHIS). KT, kidney transplantation; ESKD, end-stage kidney disease.

Variables

The patient variables included in the analyses were sex, age, area of residence, insomnia, depression, suicide, and medications. Depression was defined as having been prescribed antidepressant drugs at least once or having ICD-10 codes F32 (depressive episode) or F33 (recurrent depressive disorder) recorded as either a principal or additional diagnosis at least once during the study period [11]. The antidepressants considered encompassed all those available on the South Korean market during the study interval: selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, norepinephrine-dopamine reuptake inhibitors, tricyclic antidepressants, and monoamine oxidase inhibitors [11]. Insomnia was defined as having been prescribed hypnotic drugs at least once or having ICD-10 codes F51.0 (nonorganic insomnia) or G47.0 (disorders of initiating and maintaining sleep, insomnias) recorded as either the principal or an additional diagnosis at least once during the study period [12]. Suicide was defined as death by self-harm, as indicated by ICD-10 codes X60-X84, R458, and Z915. All diagnostic codes were in accordance with the ICD-10 classification system [13]. The diagnostic codes and medications are detailed in Supplementary Tables 1 and 2. Each participant’s area of residence was classified as urban (including Seoul, Sejong, Busan, Incheon, Daegu, Gwangju, Daejeon, and Ulsan) or rural (encompassing Gyeonggi, Gangwon, Gyeongsangbuk, Gyeongsangnam, Chungcheongbuk, Chungcheongnam, Jeollabuk, Jeollanam, and Jeju Provinces).

Propensity Score Matching

To minimize the influence of selection bias and confounders, propensity score matching (PSM) was employed at a 1:2 ratio using greedy (nearest neighbor) matching techniques with a caliper set to 0.1 standard deviations. This was performed to match individuals in the dialysis group with those in the KT group. Adjustment was applied for variables including age, sex, hypertension, diabetes mellitus, dyslipidemia, insomnia, area of residence, and the Charlson comorbidity index score, based on their significant differences between groups as identified in the univariate analysis [14]. After PSM, the balance of covariates was assessed by calculating the standardized mean differences between the groups. These differences were less than 0.1 for all covariates, indicating an adequate balance between groups (Supplementary Fig. 1).

Statistical Analysis

Continuous variables were expressed as mean±standard deviation, while categorical variables were presented as number (percentage). The baseline characteristics of patients on dialysis and KT recipients were compared using the independent t-test for continuous variables and the chi-square test for categorical variables. A Cox proportional hazards model was employed to assess the hazard ratios (HRs) for suicide and depression risks among patients on dialysis compared to KT recipients. To mitigate confounding bias in the analysis of HR for suicide and depression, we adjusted for variables that remained significantly different after matching: age, myocardial infarction, heart failure, cerebrovascular diseases, dementia, chronic lung diseases, peptic ulcer disease, mild liver disease, hemiplegia, and malignancy without metastasis. P-values less than 0.05 were considered to indicate statistical significance. The study utilized SAS Enterprise Guide 7.1 (SAS Institute), provided by the NHIS.

Data on 21,809 patients with ESKD were extracted from the Korean NHIS database, covering the period from January 2002 to December 2018. These patients exhibited no history of depression or insomnia prior to initiating renal replacement therapy.

Table 1 presents the baseline characteristics of the participants before and after PSM. Prior to PSM, the proportion of male patients among those undergoing HD/peritoneal dialysis (PD; 12,879, 73.0%) was significantly higher than the percentage of male patients among KT recipients (2,876, 69.1%). The distribution of individuals across urban and rural areas was relatively even, regardless of whether they were receiving HD/PD or had undergone KT. Notably, insomnia was found in 45.0% (7,949) of those on dialysis, compared to 25.7% (1,070) of KT recipients (P<0.001). Furthermore, patients on dialysis (22.8%, 4,019) were more likely to experience depression than KT recipients (8.6%, 358; P<0.001). Additionally, the likelihood of suicide was higher in those on dialysis (0.2%, 33) compared to KT recipients (0.1%, 5; P=0.047). After PSM, no significant differences were found in the matching variables, except for age, myocardial infarction, heart failure, cerebrovascular diseases, dementia, chronic lung diseases, peptic ulcer disease, mild liver disease, hemiplegia, and malignancy without metastasis. Kaplan-Meier curves and log-rank tests revealed significant differences in the incidence of suicide and depression between the dialysis and KT groups during the 10-year follow-up period (Fig. 2).

Table 1. Baseline characteristics of patients on dialysis versus KT recipients in 2002–2018, before and after PSM

Variable	Before PSM				After PSM
	Dialysis (n=17,649)	KT (n=4,160)	P-value		Dialysis (n=8,320)	KT (n=4,160)	P-value
Age (yr)	58.32±14.80	47.43±12.87	<0.001		48.96±12.18	47.43±12.87	<0.001
Male sex	12,879 (73.0)	2,876 (69.1)	<0.001		5,829 (70.1)	3,326 (79.9)	0.289
Hypertension	17,367 (98.4)	4,060 (97.6)	<0.001		8,143 (97.9)	4,060 (97.6)	0.323
Diabetes mellitus	15,296 (86.7)	3,524 (84.7)	0.001		7,076 (85.1)	3,524 (84.7)	0.620
Dyslipidemia	16,184 (91.7)	3,991 (95.9)	<0.001		8,000 (96.2)	3,991 (95.9)	0.557
Insomnia	7,949 (45.0)	1,070 (25.7)	<0.001		2,244 (27.0)	1,070 (25.7)	0.136
Area of residence			0.915				0.142
Urban	9,255 (52.4)	2,188 (52.6)			4,485 (53.9)	2,188 (52.6)
Rural	8,394 (47.6)	1,963 (47.2)			3,835 (46.1)	1,972 (47.4)
Charlson comorbidity index	4.93±2.21	4.51±2.14	<0.001		4.53±2.08	4.51±2.14	0.561
Myocardial infarction	691 (3.9)	74 (1.8)	<0.001		211 (2.5)	74 (1.8)	0.008
Heart failure	4,016 (22.8)	785 (18.9)	<0.001		1,719 (20.7)	785 (18.9)	0.019
Peripheral artery disease	2,912 (16.5)	561 (13.5)	<0.001		1,182 (14.2)	561 (13.5)	0.273
Cerebrovascular diseases	2,249 (12.7)	275 (6.6)	<0.001		691 (8.3)	275 (6.6)	0.001
Dementia	499 (2.8)	18 (0.4)	<0.001		61 (0.7)	18 (0.4)	0.046
Chronic lung diseases	3,465 (19.6)	872 (21.0)	0.053		1,385 (16.7)	872 (21.0)	<0.001
Connective tissue diseases	548 (3.1)	157 (3.8)	0.028		288 (3.5)	157 (3.8)	0.375
Peptic ulcer disease	2,971 (16.8)	965 (23.2)	<0.001		1,277 (15.4)	965 (23.2)	<0.001
Mild liver disease	4,906 (27.8)	1,535 (36.9)	<0.001		2,352 (28.3)	1,535 (36.9)	<0.001
Hemiplegia	273 (1.5)	18 (0.4)	<0.001		86 (1.0)	18 (0.4)	0.001
Malignancy without metastasis	1,508 (8.5)	377 (9.1)	0.285		513 (6.2)	377 (9.1)	<0.001
Moderate to severe liver disease	343 (1.9)	75 (1.8)	0.285		120 (1.4)	75 (1.8)	0.126
Metastatic solid tumor	152 (0.9)	10 (0.2)	<0.001		38 (0.5)	10 (0.2)	0.066
Acquired immunodeficiency syndrome	6 (0)	3 (0.1)	0.276		4 (0.1)	3 (0.1)	0.593
Outcomes
Depression	4,019 (22.8)	358 (8.6)	<0.001		1,341 (16.1)	358 (8.6)	<0.001
Suicide	33 (0.2)	5 (0.1)	0.047		12 (0.1)	5 (0.1)	0.731

Values are presented as mean±standard deviation or number (%).

KT, kidney transplantation; PSM, propensity score matching.

Figure 2. Kaplan-Meier analysis of the probabilities of (A) suicide and (B) depression, comparing dialysis and kidney transplantation (KT) .

Tables 2 and 3 present the univariate and multivariate HRs with 95% confidence intervals (CIs) for suicide and depression, respectively, comparing dialysis with KT. As shown in Table 2, the HR for suicide among dialysis recipients was 2.85 (95% CI, 0.83–9.84). Notably, among patients without depression, those on dialysis were particularly likely to complete suicide relative to KT recipients, with an HR of 4.48 (95% CI, 1.01–19.82). As Table 3 illustrates, the HR for depression in patients on dialysis was 1.76 (95% CI, 1.56–1.99). Based on subgroup analysis, the relatively high risk of depression among patients on dialysis was particularly notable for individuals with insomnia (HR, 2.10; 95% CI, 1.80–2.44), residents of rural areas (HR, 1.95; 95% CI, 1.64–2.32), male patients (HR, 1.76; 95% CI, 1.51–2.04), and individuals under 65 years of age (HR, 1.84; 95% CI, 1.60–2.09).

Table 2. Risk of suicide among patients on dialysis compared to kidney transplant recipients after propensity score matching

Variable	HR (95% CI)
	Univariate	Multivariatea)
Kidney transplantation	1	1
Dialysis (HD/PD)	3.30 (0.98–11.16)	2.85 (0.83–9.84)
Subgroup analysis
With depression	0.81 (0.09–7.83)	0.88 (0.09–8.62)
Without depression	4.63 (1.06–20.12)	4.48 (1.01–19.82)
Rural area of residence	1.27 (0.33–4.92)	1.34 (0.35–5.15)
Urban area of residence	NA	NA
With insomnia	2.26 (0.26–19.33)	1.50 (0.15–14.97)
Without insomnia	3.83 (0.87–16.85)	4.04 (0.91–17.94)
Male sex	4.65 (1.08–20.04)	4.34 (0.99–19.03)
Female sex	0.53 (0.03–8.54)	0.71 (0.04–14.11)
Age >65 yr	NA	NA
Age ≤65 yr	3.33 (0.99–11.24)	2.63 (0.76–9.15)

HR, hazard ratio; CI, confidence interval; HD, hemodialysis; PD, peritoneal dialysis; NA, not applicable.

^a)Adjusted for age, myocardial infarction, heart failure, peripheral artery disease, cerebrovascular diseases, dementia, chronic lung diseases, connective tissue diseases, peptic ulcer disease, malignancy without metastasis, and metastatic solid tumor.

Table 3. Risk of depression among patients on dialysis compared to kidney transplant recipients after propensity score matching

Variable	HR (95% CI)
	Univariate	Multivariatea)
Kidney transplantation	1	1
Dialysis (HD/PD)	1.85 (1.65–2.08)	1.76 (1.56–1.99)
Subgroup analysis
Urban area of residence	2.16 (1.84–2.53)	2.10 (1.80–2.44)
Rural area of residence	1.31 (1.05–1.64)	1.20 (0.97–1.48)
With insomnia	1.61 (1.34–1.94)	1.59 (1.34–1.89)
Without insomnia	1.92 (1.60–2.31)	1.95 (1.64–2.32)
Male sex	1.75 (1.49–2.06)	1.76 (1.51–2.04)
Female sex	1.80 (1.44–2.24)	1.77 (1.43–2.18)
Age >65 yr	1.13 (0.73–1.74)	1.18 (0.79–1.76)
Age ≤65 yr	1.84 (1.60–2.10)	1.84 (1.62–2.09)

HR, hazard ratio; CI, confidence interval; HD, hemodialysis; PD, peritoneal dialysis.

^a)Adjusted for age, myocardial infarction, heart failure, peripheral artery disease, cerebrovascular diseases, dementia, chronic lung diseases, connective tissue diseases, peptic ulcer disease, malignancy without metastasis, and metastatic solid tumor.

The findings of this study indicate that KT recipients experience lower rates of suicide and depression compared to patients on dialysis. Furthermore, our research demonstrates that KT is associated with a lower prevalence of depression among patients with ESKD in Korea, particularly for individuals with insomnia, residents of rural areas, male patients, and those under 65 years of age.

It is well established that patients with chronic kidney disease (CKD) exhibit elevated circulating concentrations of uremic toxins. Studies have demonstrated that levels of uremic toxins are directly associated with mental health issues, including depression [15]. Recent research indicates that the prevalence of depression and suicidal ideation significantly increases as kidney function deteriorates in patients across all stages of CKD [16]. The etiology of depression in patients with CKD undergoing dialysis is thought to be multifactorial, influenced by biological, psychological, and social variables [17]. Potential biological factors include increased cytokine levels, vitamin B12 deficiency, anemia, genetic predisposition, and neurotransmitter alterations due to uremia [17,18]. Psychological and social factors involve experiences of hopelessness, perceptions of loss and diminished control, job loss, and shifts in family and social dynamics [17,19,20]. Moreover, research conducted in uremic rats indicated a disrupted balance of amino acid neurotransmitters, suggesting that uremic toxins may exert a central effect on mental health and behavior [21]. Additionally, systemic inflammation and oxidative stress, which can directly affect the central nervous system, are significantly associated with depression and suicide [22,23]. Even mild reductions in renal function can exacerbate systemic inflammation and oxidative stress [24].

Beyond biological factors, relative to KT recipients, patients on dialysis may be more prone to depression and suicidal thoughts due to differences in lifestyle. Dialysis requires the use of an HD or a PD machine, which physically confines patients for several hours during treatment, whereas KT recipients only need to take immunosuppressants [20]. Recent PSM analysis revealed that the cumulative risk of incident depression among patients on dialysis was almost three times higher than among KT recipients [25]. Our results, along with previous research, suggest that the development of depression is influenced by socioeconomic factors related to the patient’s place of residence, such as education level, employment status, and family income [20]. These social factors contribute to varying rates of depression among younger patients, particularly those under 50 years of age [19,20,25]. However, our study excluded patients diagnosed with depression before starting renal replacement therapy, focusing instead on cases of depression that arose after the initiation of this therapy. Regrettably, we are not aware of any existing research on the prevalence of depression specifically following renal replacement therapy. Consequently, the finding that depression occurs more frequently in patients under the age of 65 years after renal replacement therapy requires further examination, as it contrasts with previous findings [25,26].

Research on suicide presents substantial challenges due to its relatively low prevalence. The United Network for Organ Sharing (UNOS) reported a total of 230 suicides among 941,252 KT recipients over a nearly 30-year study period, which equates to 0.0244%—a rate higher than that of the general United States population at 0.0139% [27]. In a Taiwanese cohort study, the suicide rate among patients with end-stage renal disease was found to be 2.4 times greater than that of the general population [28]. According to data from the World Health Organization, the incidence of suicide in the Korean general population (0.0286%) exceeds those of both UNOS and Taiwanese KT recipients [29]. However, our study indicates that within South Korea, the suicide rate among KT recipients (0.12%) is lower than that of patients on dialysis (0.19%). Compared to the general population, male patients have a higher likelihood of committing suicide [27,30]. Additionally, an age-related protective effect against suicide was noted in both the UNOS data and the Taiwanese study [27,28]. While the present investigation was a nationwide cohort study, the extremely low incidence of suicide prevented us from calculating the HR for patients over 65 years old and those residing in urban areas. To further investigate these aspects, a multinational cohort study is warranted.

This study has several limitations. It utilized a retrospective cohort study design, relying on diagnosis and prescription histories, and included only patients covered by national insurance. Depression and suicide were identified through medical records, such as ICD-10 codes and medication data. Alternative methods, such as questionnaires and psychological evaluations, were not employed. Moreover, we were unable to ascertain either the severity of depression or the occurrence of suicide attempts, as these details were not recorded. In certain instances, accurately determining the disease or cause of death can be extremely challenging. Antidepressant medications, such as selective serotonin reuptake inhibitors, may also be prescribed for conditions like anxiety, panic disorders, or obsessive-compulsive disorder, but this study did not account for such varied uses. Lastly, our analysis of the NHIS data had limitations in its capacity to examine all types of psychiatric disorders.

The current evidence strongly suggests that KT is more advantageous than dialysis for patients with ESKD in terms of minimizing the risks of depression and suicide. This finding highlights the importance of considering these risks when evaluating treatment options for ESKD.

Conflict of Interest

No potential conflict of interest relevant to this article was reported.

Funding/Support

This study was supported by research grant from the Korean Society for Transplantation (2024-00-01001-003).

Author Contributions

Conceptualization: JL, JYL. Data curation: MSK, DRK, SHK, JL, JYL. Formal analysis: MSK, DYK, DRK, SHK, JL, JYL. Funding acquisition: JYL. Supervision: JL, JYL. Validation: all authors. Visualization: MSK, DRK, SHK, JL, JYL. Writing–original draft: all authors. Writing–review & editing: MSK, DRK, SHK, JL, JYL. All authors read and approved the final manuscript.

Supplementary Materials

Supplementary materials can be found via https://doi.org/10.4285/ctr.24.0004.