Prevalence and Outcomes of Chronic Kidney Disease-Associated Pruritus

Visual Abstract

Abstract

Key Points

• CKD-associated pruritus is highly prevalent among peritoneal dialysis patients.

• Pruritus is associated with reduced health-related quality of life, and the composite of mortality and transfer to hemodialysis for peritoneal dialysis patients.

• Efforts to better identify and manage pruritus in this population are needed.

Background

Pruritus is common in hemodialysis patients. Less is known about the prevalence and outcomes of pruritus among patients receiving peritoneal dialysis (PD). Herein, we describe the prevalence of pruritus and its associations with patient-reported outcomes (PROs) and mortality/transfer to hemodialysis.

Methods

We analyzed a multicenter, international cohort of PD patients enrolled in the PD Outcomes and Practice Patterns Study from 2014 to 2022. Pruritus was assessed at entry into the PD Outcomes and Practice Patterns Study with a single-question Likert Scale capturing the extent to which patients were bothered by itch ranging from 1: not at all to 5: extremely. The kidney disease quality of life-36 and the Center for Epidemiological Studies Depression Scale assessed various PROs. Moderate-to-extreme pruritus was defined as a Likert scale score ≥3. The associations of pruritus with PROs were assessed using linear/logistic regression where appropriate. Death or hemodialysis transfer was assessed using multivariable Cox regression models.

Results

Overall, 5535 patients from eight countries were included; 43% had moderate-to-extreme pruritus which was the highest in Thailand (50%) and the lowest in the United States (33%). The adjusted odds ratios of moderate-to-extreme pruritus were higher for diabetes, low albumin, and elevated phosphorus but lower for residual urine volume (adjusted odds ratio, 0.98 per 200 ml increase in 24-hour urine volume; 95% confidence interval, 0.96 to 1.00; P = 0.05). Patients with extreme pruritus had the lowest mental and physical health component scores and a higher burden of other PROs including restless legs and disturbed sleep. Overall, 921 patients died and 1150 were transferred to hemodialysis. Patients with moderate-to-extreme pruritus were at higher adjusted risk for death or hemodialysis transfer (adjusted hazard ratio, 1.12; 95% confidence interval, 1.02 to 1.23; P = 0.02) with similar point estimates for each subcomponent of the composite outcome.

Conclusions

Pruritus is highly prevalent in PD and associated with poor health outcomes. Efforts to better identify and manage pruritus should be considered in this population.

Introduction

Managing symptom burden is a major concern for individuals receiving maintenance dialysis,¹ and CKD-associated pruritus (the daily or near-daily occurrence of itch not because of another established cause)^2,3 is very common. In the Dialysis Outcomes and Practice Patterns Study (DOPPS), the prevalence of moderate-to-extreme pruritus ranged from 28%–48% across participating countries,⁴ and other studies have identified a similarly high prevalence.^2,3,5–9

Pruritus has been identified as the second most important research priority for patients,¹⁰ and in hemodialysis patients, pruritus has been associated with lower health-related quality of life (HRQOL), higher risk of death, more frequent hospitalizations, and other symptoms/outcomes prioritized by patients, such as sleep disturbance, restless legs, and poor hemodialysis recovery.^7,8

Less is known about pruritus prevalence and outcomes in peritoneal dialysis (PD). In several small PD cohort studies, pruritus was found to be highly prevalent (e.g., 63% and 74% in South Korea and the Netherlands, respectively^11,12) and was associated with lower residual kidney clearance, weekly kt/V, and serum albumin levels, higher serum phosphate,¹¹ and higher risks of treatment failure and mortality.^12–14 Although the results of these studies are informative, there are a paucity of PD studies that have explored the association between pruritus and HRQOL.

Therefore, in a large international cohort of patients receiving PD enrolled in the longitudinal PD Outcomes and Practice Patterns Study (PDOPPS), the objectives of this study were to (1) describe the prevalence of pruritus including differences across countries, (2) identify factors associated with moderate-to-extreme pruritus, and (3) determine the association of pruritus with HRQOL, patient survival, permanent hemodialysis transfer, and other patient-reported outcomes (PROs) including energy level, the sensation of feeling drained, depression, sleep quality, and self-reported symptom burden.

Methods

Population

We analyzed an international period-prevalent cohort of patients enrolled in PDOPPS from 2014 to 2022, with last follow-up to January 2023. PDOPPS is a prospective cohort study designed to advance the understanding of optimal practices for PD patients worldwide and includes demographics, comorbidities, blood chemistries, treatment, outcomes, and PROs on patients from Australia, Canada, Japan, New Zealand, South Korea, Thailand, the United Kingdom, and the United States.¹⁵ PDOPPS obtained institutional review board study approval, and patient consent was obtained to meet national and local ethics regulations.

Exposure

We included individuals who had an assessment of pruritus severity.¹⁵ Pruritus was assessed using a single item on the kidney disease quality of life (KDQOL)-36 survey¹⁶ answered by patients at the time of enrollment into the PDOPPS study. The KDQOL is administered using the same protocol for all prior DOPPS and PDOPPS studies,¹⁷ either face to face in the dialysis unit (using a study coordinator at each dialysis facility) or self-administration by completion of the survey and placement in a sealed envelope. The tool has been previously translated for each country, and translated questionnaires are reviewed by nephrologists to ensure face validity. Patients provided a response to the following Likert scale question: During the past 4 weeks to what extent were you bothered by itchy skin? Responses ranged from 1: not at all, 2: somewhat, 3: moderately, 4: very much, to 5: extremely. Pruritus severity was treated both as a continuous assessment and binary factor using a cut point score of ≥3 (corresponding to moderate to extreme pruritus).

Covariates

Covariates were collected at entry into the PDOPPS cohort and included demographics (age, sex, race), country, 13 comorbidities (coronary artery diseases, congestive heart failure, other cardiovascular diseases, cancer, cerebrovascular diseases, peripheral vascular diseases, diabetes mellitus, prior gastrointestinal bleeding, hypertension, lung diseases, neurologic diseases, psychiatric disorders, and recurring cellulitis), PD vintage, PD modality (continuous ambulatory PD [CAPD] or automated PD [APD]), residual 24-hour urine volume, serum albumin (g/dl), potassium (mEq/L), and serum phosphorus (mg/dl). In addition, medications to treat the symptoms of pruritus during the same 4-month reporting round during which the pruritus question was answered were also collected, such that any use of a drug was counted as a yes. These medications included antihistamines, gabapentin or pregabalin, and nalfurafine.¹⁸

Outcomes

HRQOL was defined using the continuous assessment of the mental component summary (MCS) and physical component summary (PCS) scores of the KDQOL-36.¹⁶ Energy level was treated as a continuous assessment using a rescaled version of the six-point Likert scale in response to, how much of the time during the past 4 weeks did you have a lot of energy. Responses ranging from 0 to 6 were rescaled to 0: none of the time to 100: all of the time with sequential values of 0, 20, 40, 60, 80, and 100. Feeling drained was assessed as a five-point scale, ranging from 1 not at all bothered to 5 extremely bothered. The burden of depressive symptoms was captured using the ten-item Center for Epidemiological Studies Depression Scale (CESD) that ranges from 0 to 30, with depression defined as a score of ≥10. The single sleep question from the ten-item CESD was used to evaluate sleep quality. Respondents answered whether their sleep was restless during the past week and responses ranged from rarely or none of the time (<1 day), some or little of the time (1–2 days), occasionally or a moderate amount of the time (3–4 days), to most or all of the time (6–7 days). Responses were further grouped as restless sleep for ≥3 versus <3 days. Symptom burden included questions on how much kidney disease interferes with daily life, takes up time, causes frustration, or makes the respondent feel like a burden and ranged from a score of 0 to 100.¹⁹ Time to event outcomes included the composite of death or transfer to hemodialysis (noting that this unified outcome has been suggested for studies of PD patients²⁰) and the subcomponents of the composite outcome (death and transfer to hemodialysis). Death was all cause and reported as deaths/100 patient-years. Permanent hemodialysis transfer was defined as a permanent modality switch to hemodialysis or temporary transfer from PD to hemodialysis lasting at least 12 weeks.²¹

Analysis

Baseline characteristics were reported for the overall cohort and separately for each pruritus category ranging from 1 to 5 using univariate statistics. Differences in the prevalence of pruritus across each country were visually displayed. Associations with moderate-to-extreme pruritus (using scores of ≥3) were analyzed using adjusted logistic regression and reported using odds ratios with 95% confidence interval (CI), with facility-level clustering adjusted for using a repeated-measures approach with a compound symmetry covariance structure.²² In an a priori sensitivity analysis, adjusted associations with moderate-to-extreme pruritus were also reported for categories of phosphorus, potassium, albumin, 24-hour urine volume, age, and dialysis vintage. The MCS, PCS, symptom burden, and responses to questions on feeling drained and energy were analyzed using adjusted linear regression; the expected difference in each continuous assessed value between patients extremely bothered by pruritus versus each of the other responses (holding other covariables constant) was reported along with 95% CIs. Time to death was graphically displayed using Kaplan–Meier survival curves for categories of pruritus (none, somewhat/moderate, very much/extreme). Time to transfer to hemodialysis was graphically displayed using cumulative incidence curves for the same categories of pruritus. The adjusted association between moderate-to-extreme pruritus and each outcome (composite of death or transfer to hemodialysis and each subcomponent) was analyzed using a multivariable Cox survival analysis and reported using adjusted relative hazards (hazard ratio) with 95% CI. We presented the data using three sequential models (model 1: adjusted for demographics, model 2: adjusted for demographics, dialysis vintage, dialysis modality, and comorbidities, and model 3: adjusted for all baseline characteristics). Time at risk commenced from the date when patients answered the pruritus question of KDQOL-36 or 60 days after the start of PD, whichever date was later. Follow-up was censored at transplant, recovery of kidney function, death, end of study period, or 60 days after departure if the patient transferred to another PD facility or changed modality to hemodialysis. In a secondary analysis, we also evaluated the association between pruritus severity and transfer to hemodialysis using a Fine and Gray²³ competing risk regression approach with mortality as the competing event. This approach was used to evaluate for indirect effects of pruritus severity on treatment failure (appreciating that the Fine and Gray approach can result in biased results for etiologic questions).^23,24 Facility-level clustering was accounted for using a robust sandwich estimator. Missing data were handled using the Statistical Analysis System Enterprise version 8.3 proc multiple imputation procedure with 25 imputations, combined using the Statistical Analysis System MIANALYZE procedure. The Adapted Strengthening the Reporting of Observational Studies Epidemiology checklist for Observational Studies was used to guide the reporting of this study (Supplemental Table 1).²⁵

Results

The cohort consisted of 5535 patients who completed the pruritus question on the KDQOL-36 out of a period-prevalent sample of 5652 incident and prevalent patients who completed at least some part of the questionnaire; 117 patients did not respond to the item on pruritus. Overall, 43% of patients had moderate-to-extreme pruritus (scores of ≥3). Baseline characteristics are noted in Table 1. Patients who were extremely bothered by pruritus tended to have lower serum albumin levels (3.3 g/dl), lower residual urine volume (270 versus 500 ml for those without pruritus), and a higher burden of most comorbid conditions. Patients with extreme pruritus tended to use a higher proportion of antihistamines (19% versus 9% comparing extreme with no pruritus). Nalfurafine was rarely used, primarily among those with pruritus scores of 4 and 5 who resided in Japan. Gabapentin/pregabalin use was similar across all pruritus categories.

Table 1.

Patient characteristics stratified by pruritus score

Characteristic	Overall	Degree to Which Bothered by Pruritus					Missing (%)
		1: Not at All	2: Somewhat	3: Moderately	4: Very Much	5: Extremely
No.	5535	1406	1740	1115	804	470
Demographics
Patient age (yr), mean±SD	59 (14)	58 (14)	59 (14)	60 (14)	59 (14)	60 (15)	0
Black racea, No. (%)	340 (6)	131 (9)	104 (6)	51 (5)	35 (4)	19 (4)	0
Male, No. (%)	3195 (58)	785 (56)	989 (57)	676 (61)	463 (58)	282 (60)	0
Comorbidities, No. (%)
Coronary artery diseases	890 (16)	234 (17)	257 (15)	175 (16)	131 (16)	93 (20)	0
Other cardiovascular diseases	576 (10)	154 (11)	167 (10)	119 (11)	75 (9)	61 (13)	0
Heart failure	689 (13)	169 (12)	213 (12)	136 (12)	106 (13)	65 (14)	1
Cerebrovascular diseases	438 (9)	108 (9)	124 (8)	87 (8)	66 (9)	53 (12)	7
Peripheral vascular diseases	439 (9)	117 (9)	131 (8)	76 (7)	62 (8)	53 (12)	7
Diabetes mellitus	2480 (45)	594 (43)	744 (43)	510 (46)	390 (49)	242 (52)	0
Gastrointestinal bleeding	80 (2)	20 (1)	17 (1)	20 (2)	14 (2)	9 (2)	1
Cancer	417 (8)	123 (9)	140 (8)	80 (7)	42 (5)	32 (7)	1
Hypertension	4831 (88)	1198 (86)	1525 (89)	969 (88)	719 (90)	420 (90)	1
Lung diseases	182 (3)	47 (3)	51 (3)	36 (3)	24 (3)	24 (5)	0
Neurologic diseases	170 (3)	47 (3)	47 (3)	34 (3)	19 (2)	23 (5)	0
Psychiatric diseases	433 (8)	131 (9)	128 (7)	71 (6)	51 (6)	52 (11)	7
Recurrent cellulitis	58 (1)	13 (1)	15 (1)	11 (1)	10 (1)	9 (2)	2
Dialysis parameters
PD vintage (yr), mean±SD	1.9 (2.6)	1.9 (2.7)	1.9 (2.5)	1.9 (2.5)	2.0 (2.7)	1.9 (2.4)	0
PD vintage categories, No. (%)
<3 mo	1760 (32)	444 (32)	571 (33)	329 (30)	263 (33)	153 (33)	0
3 mo to 0.9 yr	1051 (19)	296 (21)	303 (18)	220 (20)	155 (19)	77 (16)	0
1–2.9 yr	1485 (27)	347 (25)	469 (27)	327 (30)	207 (26)	135 (29)	0
3+ yr	1217 (22)	309 (22)	390 (23)	234 (21)	179 (22)	105 (22)	0
Urine volume (L/24 h), mean±SD	0.40 (0.67)	0.50 (0.75)	0.40 (0.67)	0.38 (0.64)	0.33 (0.60)	0.27 (0.52)	0
APD (versus CAPD), No. (%)	2328 (44)	679 (50)	702 (42)	448 (42)	328 (42)	171 (38)	4
Blood chemistries, mean±SD
Albumin (g/dl)	3.4 (0.5)	3.5 (0.5)	3.5 (0.5)	3.4 (0.5)	3.4 (0.5)	3.3 (0.6)	6
Potassium (mEq/L)	4.1 (0.7)	4.1 (0.7)	4.1 (0.7)	4.1 (0.8)	4.1 (0.8)	4.1 (0.8)	2
Phosphorus (mg/dl)	4.8 (1.5)	4.7 (1.4)	4.9 (1.5)	4.9 (1.6)	5.0 (1.6)	4.9 (1.7)	6
Medications, No. (%)
Antihistamine	570 (12)	116 (9)	150 (10)	116 (12)	111 (16)	77 (19)	13
Gabapentin	218 (5)	66 (5)	52 (3)	37 (4)	38 (5)	25 (6)	13
Pregabalin	81 (2)	26 (2)	13 (1)	15 (2)	16 (2)	11 (3)	13
Nalfurafineb	21 (0)	1 (0)	3 (0)	2 (0)	4 (1)	11 (3)	13

Mean±SD or median were used for continuous variables as appropriate. ADL; activities of daily living; APD, automated peritoneal dialysis; CAPD, continuous ambulatory peritoneal dialysis; PD, peritoneal dialysis.

^aBlack race varied by country. Among US patients, 18% were Black and 4% in Canada and 4% in the United Kingdom; no patients were coded as Black in the other countries.

^bNalfurafine was only used for patients in Japan at proportions of 0.5%, 1%, 1.2%, 2.9%, and 10.1% for those in categories 1–5 of pruritus severity, respectively.

The distribution of pruritus scores stratified by country location is noted in Figure 1. The highest prevalence of moderate-to-extreme pruritus was noted in Thailand (50%), followed by the United Kingdom (49%) while the prevalence was the lowest in the United States (33%).

Figure 1.

Proportion of patients with each pruritus score stratified by country location.

Relative to the United States, the highest odds for moderate-to-extreme pruritus was noted among patients from Thailand (adjusted odds ratio [aOR], 2.39; 95% CI, 1.89 to 3.03; P < 0.001; Figure 2, and Table 2). Age (aOR, 1.08; 95% CI, 1.03 to 1.13 for every 10 years; P < 0.001), male sex (aOR, 1.12; 95% CI, 1.00 to 1.26; P = 0.04), diabetes (aOR, 1.23; 95% CI, 1.10 to 1.38; P < 0.001), and serum phosphate concentration (aOR, 1.14; 95% CI, 1.09 to 1.19 for each mg/dl; P < 0.001) were associated with moderate-to-extreme pruritus. By contrast, greater daily residual urine volume was associated with a lower odds of moderate-to-severe pruritus (aOR, 0.98 per 200 ml greater 24-hour urine volume; 95% CI, 0.96 to 1.00; P = 0.05). Similar associations with moderate-to-extreme pruritus were found when treating continuous variables as categorical (Table 2).

Figure 2.

Adjusted associations between patient characteristics, country, and moderate-to-severe pruritus (score ≥3) versus being somewhat or not bothered by pruritus. APD, automated peritoneal dialysis; CAPD, continuous ambulatory peritoneal dialysis; CI, confidence interval; PD, peritoneal dialysis.

Table 2.

Unadjusted and adjusted^a associations between patient characteristics and moderate-to-extreme pruritus (score of ≥3) with both categorical and continuous variables for age, vintage, potassium, phosphorus, albumin, and 24-hour urine volume

Characteristic	Unadjusted OR (95% CI)	aOR (95% CI)
Patient age, yr
<55	0.76 (0.64 to 0.90) P < 0.001b	0.71 (0.59 to 0.86) P < 0.001b
55–64	0.89 (0.75 to 1.06) P = 0.18	0.84 (0.69 to 1.01) P = 0.06b
65–74	0.86 (0.72 to 1.04) P = 0.12	0.82 (0.68 to 0.99) P = 0.04b
75+	Reference	Reference
Patient age, per 10 yr	1.06 (1.02 to 1.10) P = 0.001b	1.08 (1.03 to 1.13) P < 0.001b
PD vintage, yr
0–3 mo	0.99 (0.85 to 1.15) P = 0.90	0.94 (0.80 to 1.10) P = 0.41
3 mo to 1 yr	1.02 (0.86 to 1.21) P = 0.81	1.15 (0.96 to 1.37) P = 0.14
1–2.9 yr	1.11 (0.95 to 1.29) P = 0.20	1.16 (0.99 to 1.36) P = 0.08
3+ yr	Reference	Reference
PD vintage, per year	1.00 (0.98 to 1.02) P = 0.76	1.00 (0.98 to 1.03) P = 0.75
Dialysis modalityc
APD	0.84 (0.76 to 0.94) P = 0.003b	1.13 (0.98 to 1.30) P = 0.09
CAPD	Reference	Reference
Racec
Black	0.57 (0.45 to 0.72) P < 0.001b	0.86 (0.66 to 1.12) P = 0.27
Non-Black	Reference	Reference
Sexc
Female	Reference	Reference
Male	1.14 (1.02 to 1.26) P = 0.02b	1.12 (1.00 to 1.26) P = 0.04b
Coronary artery diseases	1.08 (0.94 to 1.25) P = 0.28	1.04 (0.88 to 1.22) P = 0.69
Other cardiovascular diseases	1.05 (0.88 to 1.25) P = 0.59	0.98 (0.81 to 1.19) P = 0.85
Heart failure	1.06 (0.91 to 1.25) P = 0.44	0.97 (0.81 to 1.16) P = 0.81
Cerebrovascular diseases	1.14 (0.94 to 1.38) P = 0.20	1.11 (0.90 to 1.37) P = 0.39
Peripheral vascular diseases	0.99 (0.81 to 1.21) P = 0.93	0.98 (0.79 to 1.21) P = 0.81
Diabetes	1.23 (1.11 to 1.37) P < 0.001b	1.23 (1.10 to 1.38) P < 0.001b
Gastrointestinal bleeding	1.54 (0.99 to 2.41) P = 0.06	1.46 (0.92 to 2.30) P = 0.10
Serum potassium, mEq/L
<3.5	1.22 (1.00 to 1.48) P = 0.05b	1.16 (0.94 to 1.45) P = 0.19
3.5–4.9	1.03 (0.87 to 1.21) P = 0.75	1.10 (0.93 to 1.31) P = 0.28
5+	Reference	Reference
Serum potassium, per mEq/L	0.92 (0.86 to 0.99) P = 0.03b	0.94 (0.87 to 1.02) P = 0.15
Serum phosphorus, mg/dl
<3	1.20 (0.97 to 1.49) P = 0.10	0.94 (0.75 to 1.19) P = 0.64
3–4.4	Reference	Reference
4.5–4.9	0.94 (0.82 to 1.08) P = 0.40	1.04 (0.90 to 1.20) P = 0.56
5–6.9	1.16 (1.00 to 1.35) P = 0.05b	1.36 (1.16 to 1.60) P < 0.001b
7+	1.60 (1.29 to 1.98) P < 0.001b	1.98 (1.57 to 2.50) P < 0.001b
Serum phosphorus, per mg/dl	1.06 (1.02 to 1.10) P = 0.002b	1.14 (1.09 to 1.19) P < 0.001b
Serum albumin, g/dl
<3	1.62 (1.34 to 1.97) P < 0.001b	1.33 (1.07 to 1.65) P = 0.01b
3–3.4	1.38 (1.15 to 1.64) P < 0.001b	1.28 (1.06 to 1.55) P = 0.01b
3.5–3.9	1.17 (0.99 to 1.39) P = 0.06	1.15 (0.96 to 1.38) P = 0.11
4+	Reference	Reference
Serum albumin, per g/dl	0.71 (0.64 to 0.79) P < 0.001b	0.79 (0.71 to 0.89) P < 0.001b
24-hour urine volume, L
0	1.53 (1.32 to 1.77) P < 0.001b	1.18 (1.00 to 1.39) P = 0.05b
0.1–0.4	1.25 (1.00 to 1.56) P = 0.05b	1.11 (0.88 to 1.40) P = 0.41
0.5–0.9	1.23 (1.00 to 1.50) P = 0.05b	1.17 (0.95 to 1.44) P = 0.14
1+	Reference	Reference
24-h urine volume, per 200 ml greater 24-h urine volume	0.95 (0.94 to 0.97) P < 0.001b	0.98 (0.96 to 1.00) P = 0.05b
Country
The United States	Reference	Reference
Australia/New Zealand	1.73 (1.28 to 2.34) P < 0.001b	1.51 (1.10 to 2.09) P = 0.01b
Canada	1.69 (1.36 to 2.09) P < 0.001b	1.69 (1.34 to 2.12) P < 0.001b
South Korea	1.47 (1.21 to 1.79) P < 0.001b	1.66 (1.28 to 2.15) P < 0.001b
Japan	1.72 (1.45 to 2.03) P < 0.001b	1.69 (1.37 to 2.08) P < 0.001b
The United Kingdom	2.08 (1.78 to 2.42) P < 0.001b	2.13 (1.60 to 2.83) P < 0.001b
Thailand	2.05 (1.58 to 2.67) P < 0.001b	2.39 (1.89 to 3.03) P < 0.001b

aOR, adjusted odds ratio; APD, automated peritoneal dialysis; CAPD, continuous ambulatory peritoneal dialysis; CI, confidence interval; OR, odds ratio; PD, peritoneal dialysis.

^aAdjusted for categorical age, categorical peritoneal dialysis vintage, race, sex, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding), categorical blood chemistries (potassium, phosphorus, and albumin), categorical 24-hour urine volume, peritoneal dialysis modality, and country. Results for continuous versions of categorical variables each came from separate models, adjusting for categorical versions of the remaining categorical variables.

^bStatistically significant, P < 0.05.

^cFor binary variables, the null state is the reference group.

MCS and PCS scores were the lowest for those with extreme pruritus (41±12 and 34±10, Table 3). Linear regression results for each PRO are noted in Figures 3 and 4 and Table 3. Relative to those without pruritus, patients with extreme pruritus had the largest differences in MCS score (−7.9; 95% CI, −9.0 to −6.8; P < 0.001) and PCS score (−7.4; 95% CI, −9.0 to −6.8; P < 0.001; Figure 3). Symptom burden and perceptions of low energy or feeling drained followed a similar pattern to the MCS and PCS results (Figure 4 and Table 3). Relative to those without pruritus, the aOR for depression (CESD ≥10) was the highest for those with extreme pruritus (aOR, 5.19; 95% CI, 4.06 to 6.63; P < 0.001; Table 4). Restless sleep followed a similar pattern to depression symptoms (Table 4).

Table 3.

Crude and adjusted associations between categories of pruritus and patient-reported outcomes^a,b

During the Past 4 wk, To What Extent Were You Bothered By Itchy Skin?	Mental Component Score		Physical Component Score		Symptom Burden		Feeling Drained		Energy
	Mean±SD	b (95% CI)	Mean±SD	b (95% CI)	Mean±SD	b (95% CI)	Mean±SD	b (95% CI)	Mean±SD	b (95% CI)
1: Not at all	50 (10)	Reference	42 (10)	Reference	54 (27)	Reference	2 (1)	Reference	48 (28)	Reference
2: Somewhat	47 (10)	−1.9 (−2.7 to −1.2) P < 0.001	40 (10)	−1.8 (−2.4 to −1.1) P < 0.001	49 (27)	−3.9 (−5.7 to −2.1) P < 0.001	2 (1)	0.2 (0.2 to 0.3) P < 0.001	44 (27)	−3.7 (−5.6 to −1.9) P < 0.001
3: Moderately	45 (10)	−3.7 (−4.6 to −2.9) P < 0.001	38 (9)	−3.7 (−4.5 to −3.0) P < 0.001	42 (26)	−10.0 (−12.0 to −7.9) P < 0.001	3 (1)	0.6 (0.5 to 0.6) P < 0.001	38 (25)	−8.0 (−10.0 to −5.9) P < 0.001
4: Very much	44 (10)	−4.7 (−5.6 to −3.8) P < 0.001	36 (10)	−5.2 (−6.0 to −4.4) P < 0.001	38 (24)	−13.4 (−15.7 to −11.2) P < 0.001	3 (1)	0.7 (0.6 to 0.8) P < 0.001	33 (24)	−13.9 (−16.2 to −11.6) P < 0.001
5: Extremely	41 (12)	−7.9 (−9.0 to −6.8) P < 0.001	34 (10)	−7.4 (−8.4 to −6.5) P < 0.001	31 (25)	−20.3 (−23.0 to −17.6) P < 0.001	3 (1)	1.2 (1.1 to 1.3) P < 0.001	26 (25)	−19.5 (−22.3 to −16.7) P < 0.001

CI, confidence interval.

^aRange of scores: mental/physical component scores (0–100, with increased scores representing better health status), symptom burden (0–100, with increased scores representing lower symptom burden), feeling drained (1–5, with lower scores representing not feeling drained), energy level (0–100, with increased scores representing higher energy).

^bAdjusted for age, peritoneal dialysis vintage, race, sex, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding), blood chemistries (potassium, phosphorus, and albumin), 24-hour urine volume, peritoneal dialysis modality, and country.

Figure 3.

Adjusted associations between categories of pruritus. (A) mental health component summary score and (B) physical health component summary score. Adjusted for age, PD vintage, race, sex, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding), blood chemistries (potassium, phosphorus, and albumin), 24-hour urine volume, PD modality, and country. MCS, mental component summary; PCS, physical component summary.

Figure 4.

Adjusted associations between categories of pruritus. (A) symptom burden and energy, (B) depression symptom and restless sleep, and (C) sensation of feeling drained. Adjusted for age, PD vintage, race, sex, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding), blood chemistries (potassium, phosphorus, and albumin), 24-hour urine volume, PD modality, and country. CESD, Center for Epidemiological Studies Depression Scale.

Table 4.

Adjusted Associations between pruritus and (A) depression symptoms (Center for Epidemiological Studies Depression Scale ≥10) and (B) restless sleep

During the Past 4 wk, To What Extent Were You Bothered By Itchy Skin?	OR for CESD Score ≥10 (95% CI)a	OR for Restless Sleep (95% CI)b
1: Not at all	1 (reference)	1 (reference)
2: Somewhat	1.55 (1.31 to 1.84), P < 0.001	1.22 (1.04 to 1.43), P = 0.01
3: Moderately	2.37 (1.98 to 2.85), P < 0.001	1.99 (1.68 to 2.37), P < 0.001
4: Very much	3.00 (2.45 to 3.67), P < 0.001	2.93 (2.43 to 3.54), P < 0.001
5: Extremely	5.19 (4.06 to 6.63), P < 0.001	5.25 (4.15 to 6.64), P < 0.001

Logistic model results, adjusted for age, peritoneal dialysis vintage, race, sex, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding), blood chemistries (potassium, phosphorus, and albumin), 24-hour urine volume, peritoneal dialysis modality, and country. CESD, Center for Epidemiological Studies Depression Scale; CI, confidence interval; OR, odds ratio.

^aThreshold for depression symptom using the ten-item Center for Epidemiological Studies Depression Scale.

^bDefined as restless sleep for three or more nights/week.

Overall, 921 patients died and 1150 permanently switched to hemodialysis during follow-up. With each incrementally higher pruritus score, there was a progressively higher rate of death (Table 5). Unadjusted associations between categories of pruritus severity and mortality are shown Supplemental Figure 1. Cumulative incidence curves for transfer to hemodialysis by pruritus category are shown in Supplemental Figure 2. Adjusted associations between pruritus severity and outcomes of death, transfer to hemodialysis, and the composite of death/transfer to hemodialysis are noted in Table 6. Generally, these associations were stronger in the models adjusting only for patient demographics (model 1) and were slightly weaker in the fully adjusted model (model 3). Patients with moderate-to-extreme pruritus were at higher risk for the composite of death or hemodialysis transfer compared with those without pruritus (adjusted hazard ratio [aHR], 1.12; 95% CI, 1.02 to 1.23; P = 0.02). The point estimates for the association between moderate-to-extreme pruritus and death (aHR, 1.14; 95% CI, 0.99 to 1.30; P = 0.06) and between moderate-to-extreme pruritus and transfer to hemodialysis alone (aHR, 1.10; 95% CI, 0.97 to 1.25; P =0.12) were similar. The point estimate for transfer to hemodialysis while treating death as a competing event was comparable with the competing risk analysis (aHR for hemodialysis transfer, 1.13; 95% CI, 0.99 to 1.29; P = 0.06).

Table 5.

Crude associations between categories of pruritus and death, transfer to hemodialysis, or the composite of death/hemodialysis transfer

During the Past 4 wk, To What Extent Were You Bothered By Itchy Skin?	Deaths/100 Patient-Years	Hemodialysis Transfer/100 Patient-Years	Death or Hemodialysis Transfer/100 Patient-Years
1: Not at all	9.0	12.0	21.0
2: Somewhat	9.5	12.9	22.4
3: Moderately	12.3	13.3	25.6
4: Very much	13.1	15.2	28.4
5: Extremely	13.6	17.2	30.8

Table 6.

Crude and adjusted associations for those with moderate-to-extreme pruritus (scores ≥3) and death, transfer to hemodialysis, or the composite of death/transfer to hemodialysis

Outcome	Events (No.)	Event Rate (per 100 Patient-Year)	Model 1	Model 2	Model 3
			HR (95% CI)	HR (95% CI)	HR (95% CI)
Death or transfer to hemodialysis	2071	23.8
No pruritis	1091	21.4	Reference	Reference	Reference
Moderate-to-extreme pruritis	980	27.0	1.20 (1.10 to 1.31), P < 0.001	1.18 (1.07 to 1.29), P < 0.001	1.12 (1.02 to 1.23) P = 0.02
Death	921	10.6
No pruritis	464	9.1	Reference	Reference	Reference
Moderate-to-extreme pruritis	457	12.6	1.23 (1.07 to 1.41), P = 0.003	1.20 (1.05 to 1.38), P = 0.01	1.14 (0.99 to 1.30) P = 0.06
Transfer to hemodialysis	1150	13.2
No pruritis	627	12.3	Reference	Reference	Reference
Moderate-to-extreme pruritis	523	14.4	1.17 (1.04 to 1.32), P = 0.01	1.15 (1.02 to 1.29), P = 0.02	1.10 (0.97 to 1.25) P = 0.12

Model 1: adjusted for age, race, sex; model 2 adjusted for variables in model 1, peritoneal dialysis vintage, peritoneal dialysis modality, comorbidities (coronary artery diseases, other cardiovascular diseases, heart failure, cerebrovascular diseases, peripheral vascular diseases, diabetes, and gastrointestinal bleeding); model 3: adjusted for variables in model 1, blood chemistries (potassium, phosphorus, and albumin), 24-hour urine volume, and country. CI, confidence interval; HR, hazard ratio.

Discussion

In this study, moderate-to-extreme pruritus was most common among older patients with diabetes, with variation in its prevalence between different countries. Several factors were associated with moderate-to-severe pruritus, whereas patients with greater residual urine volume had a lower risk. Finally, moderate-to-extreme pruritus was associated with mortality, hemodialysis transfer, reduced HRQOL, poor sleep, low energy, and depressive symptoms.

In a contemporary analysis of the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD), 74% of PD patients had pruritus and 35% rated it as moderate to extreme.¹¹ Similarly, we identified that 75% of the cohort had some degree of pruritus, and 43% rated it as moderate to extreme. There were differences between our study (which included prevalent patients) and the NECOSAD study (which comprised incident patients). However, 32% of our cohort was within 0–3 months of dialysis initiation, and the proportion of patients within 3 months of dialysis initiation did not change across categories of itch. Other studies have identified a lower prevalence of itch,^13,26–28 which may relate to the tool being used or timing relative to dialysis initiation. For example, pruritus may improve shortly after initiating PD and then stabilize.²⁹ Other studies have shown that loss of residual kidney clearance may be associated with lower itch severity,¹¹ and this may relate to clearance of uremic toxins implicated in CKD-associated pruritus. Not surprisingly, longer dialysis vintage has been shown to be associated with pruritus.^27,30 We found that patients having lower residual kidney function were more likely to be moderately to extremely bothered by pruritus but did not see any meaningful relationship between PD vintage and pruritus severity in our analysis.

Similar to other studies, we found that the prevalence of itch varied across countries. In a DOPPS study of hemodialysis patients, the United Kingdom was identified as having the highest prevalence, whereas the United States and Germany were the lowest.⁴ We found that both the United Kingdom and Thailand had the highest pruritus prevalence. There are several potential explanations that may underpin this finding. Variability in prevalence may result from differences in the proportion of nephrologists inquiring about itch for PD patients between countries, although to our knowledge, this has not been studied in PD. In addition to asking about itch, there may be cultural differences in the perception and self-report of symptoms across countries. For example, it has been shown that there are differences in the proportion of patients with elevated 5-D itch scores across different race groups.³¹ This may emphasize the need for cultural adaptation of cutoff values for itch assessment tools. Finally, differences in the practice of PD may have influenced the observed variability across countries. For example, although we did not explore characteristic differences in PD delivery across PDOPPS countries and the relationship with itch severity, previous studies have shown that patients in Thailand have the lowest residual urine volumes.³² In our study, the association between receiving dialysis in Thailand and moderate-to-severe itch was partially explained by residual urine volume (data not shown), although other factors also contributed to between-country differences.

There were many other factors associated with pruritus in this study. We found a significant association between hypoalbuminemia and moderate-to-extreme pruritus in keeping with the theory that inflammation may contribute to pruritus.³ A similar relationship between hypoalbuminemia and moderate-to-extreme pruritus has also been seen in patients receiving hemodialysis in the DOPPS.⁴ We also found that hyperphosphatemia was associated with pruritus (similar to other studies)¹¹ although the association between measures of bone mineral metabolism and pruritus in dialysis is inconsistent.^4,27 Similar to a previous study,¹¹ we found that reduced residual urine volume was associated with higher odds of moderate-to-extreme itch, particularly for those that were anuric. Furthermore, the point estimate for APD versus CAPD and moderate-to-extreme itch was 1.14 (95% CI, 0.99 to 1.31). This result is comparable with the recent NECOSAD follow-up study¹¹ in which there was a positive association between dialysis clearance and pruritus in longitudinal analysis. We hypothesize that APD may be preferentially used to maximize dialysis in the face of a loss of residual urine volume, leading to the paradoxical finding of increased itch severity in those using APD versus CAPD.

Although the prevalence of pruritus was high, only a small proportion of patients were on medications for relieving pruritus. Gabapentin/pregabalin was only used in a small minority of patients, despite some evidence that it may be efficacious³³ Similar to hemodialysis,⁴ antihistamines were the most commonly prescribed medication in the 4 months before measuring pruritus, despite no evidence for benefit.³³ Some of the limited use may relate to the limited efficacy of therapies to treat pruritus in dialysis. Furthermore, although effective therapies exist for pruritus (i.e., opioid pathway modifiers including nalfurafine and difelikefalin) many of these treatments are not readily available or not approved for use in PD.^34,35 Finally, the cause of pruritus is multifactorial, and an exact unifying mechanism has not been found and may not exist.³⁶ Therefore, given the multitude of factors that may influence pruritus severity in a dialysis patient, existing treatments may not equally benefit all patients.

We found a strong association between pruritus and reduced HRQOL. In a systematic review of dialysis patients, there was an association between pruritus and both generic and disease-specific HRQOL.³⁷ Our study adds to the literature suggesting the detrimental effects of pruritus on patient-important outcomes^37–40 and that the association between pruritus and HRQOL is not unique to hemodialysis.⁴¹ Furthermore, we did identify associations between pruritus and other determinants of HRQOL including symptom burden, low energy, poor sleep quality, and depressive symptoms, suggesting that pruritus may be associated with reductions in numerous domains of a patient's well-being and which has been expansively documented recently by Menzaghi et al.⁴² based on qualitative interviews with patients on hemodialysis. The association between HRQOL and pruritus cannot be overstated. In our study, the mean MCS and PCS scores were 9 and 8 points higher comparing those without pruritus with those with extreme pruritus and the adjusted difference in expected MCS and PCS of HRQOL (comparing those without pruritus with those with moderate-to-severe pruritus) was 7.5 and 7.3, respectively. This is above the minimal clinically important difference in HRQOL of 0.5 SD (approximately equivalent to 5 points) in both scales.³⁷

Similar to previous studies in both hemodialysis and PD,^7,8 we observed that pruritus was associated with a higher point estimate for the cause-specific outcomes of mortality and transfer to hemodialysis (although CIs crossed one) and that there was a statistically significant association between pruritus and the composite outcome of death or transfer to hemodialysis, even after multivariable adjustment.¹³ We hypothesize that some patients with pruritus are transferred to hemodialysis potentially in an effort to treat itch or alternatively, other concurrent symptoms or clinical findings suggestive of suboptimal dialytic clearance.

Given the high prevalence, efforts to systematically capture the burden of pruritus using established tools is of paramount importance. The limited use of therapies to manage pruritus in this population is concerning and emphasizes the need for algorithmic treatment approaches,^43,44 while avoiding treatments that have little evidence of benefit.³³ Furthermore, there is a need to address pruritus with novel treatments that have not yet been tested in PD. For example, difelikefalin has shown evidence of benefit in hemodialysis,^45,46 but studies in PD are lacking.

To our knowledge, this is the largest study of PD patients and patients with pruritus. In addition, PDOPPS has used the same instrument to quantify itch across all countries. This ensures that our results are reproducible and generalizable across countries. This study does have limitations. Unfortunately, owing to the limitations of survey availability, we did not have more extensive tools to evaluate some PROs, for example, sleep/sleep quality. Other tools to assess pruritus severity (i.e., the worst-intensity numerical rating scale)⁴⁷ or the effect of pruritus on function and HRQOL (i.e., the Skindex-10 or 5-D tool)^6,48 were not part of PDOPPS data collection. Nonetheless, the single question screening tool used in PDOPPS has been previously studied¹¹ and in hemodialysis has predictive validity.⁷ Furthermore, the single-question screening tool has shown moderate correlation with the worst-intensity numerical rating scale in previous study.⁴⁹ Because patients may develop pruritis after a long period on dialysis, we displayed pruritis prevalence within vintage categories to provide more specific information for patients at different points in time. We acknowledge that the prevalences we report could change within vintage categories because of incident pruritis and/or survivor bias. Although we were able to evaluate for associations with pruritus, there were a number of factors that we could not account for or for which there was a lack of robust data to provide insights into their contribution to pruritus. These include treatment-related factors such as diuretic use, which has been shown to have a negative associate with pruritus in previous study.⁵⁰ We acknowledge that detailed skin examinations may not have been performed in patients before administering the pruritus question. Thus, some patients may have another condition associated with pruritus. Finally, longitudinal changes in pruritus severity were not available which may be important considering the fluctuating presentation; in the KALM-1 and KALM-2 pooled results of difelikefalin, 35% of patients in the placebo arm had resolution of pruritus at the end of follow-up.⁴⁶

In conclusion, in this international study of PD patients, we identified that pruritus is common and associated with poor health outcomes including reduced HRQOL and mortality or transfer to hemodialysis. Efforts to better capture pruritus in PD and to identify treatments that may improve pruritus severity are needed in this population.

Supplementary Material

cjasn-19-1622-s001.pdf ^{(1.4MB, pdf)}

cjasn-19-1622-s002.pdf ^{(316.3KB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/CJN/C65.

Funding

This work was supported by CSL Behring, Amgen, Akebia Therapeutics, Astellas Pharma US, Bard Peripheral Vascular, Inc., Baxter Healthcare Corporation, Bayer Yakuhin, Bayer, Cara, Chugai Pharmaceutical, GlaxoSmithKline, Japanese Society for Peritoneal Dialysis, JMS Co., Ltd., Kidney Foundation, Japan, Kissei, Kyowa Hakko Kirin, Merck Sharp and Dohme K.K., Nikkiso Co,. Ltd., Ono Pharmaceutical, Terumo, Torii Pharmaceutical, Vifor Pharma, and National Institutes of Health (Agency for Healthcare Research and Quality)

Author Contributions

Conceptualization: Brian Bieber, Yeoungjee Cho, David W. Johnson, Talerngsak Kanjanabuch, Hideki Kawanishi, Yong-Lim Kim, Mark Lambie, Keith McCullough, Jeffrey Perl, Ronald L. Pisoni, Claudio Rigatto, Martin Schreiber, Jenny Shen, Karthik K. Tennankore.

Data curation: Yeoungjee Cho, David W. Johnson, Talerngsak Kanjanabuch, Hideki Kawanishi, Yong-Lim Kim, Mark Lambie, Keith McCullough, Jeffrey Perl, Ronald L. Pisoni, Claudio Rigatto, Martin Schreiber, Jenny Shen.

Formal analysis: Keith McCullough, Karthik K. Tennankore.

Funding acquisition: Jeffrey Perl.

Investigation: Yeoungjee Cho, David W. Johnson, Keith McCullough, Jeffrey Perl, Ronald L. Pisoni, Karthik K. Tennankore.

Methodology: Keith McCullough, Jeffrey Perl, Ronald L. Pisoni, Karthik K. Tennankore.

Project administration: Ronald L. Pisoni, Karthik K. Tennankore.

Resources: Ronald L. Pisoni.

Validation: Keith McCullough.

Visualization: Keith McCullough.

Writing – original draft: David W. Johnson, Keith McCullough, Ronald L. Pisoni, Karthik K. Tennankore.

Writing – review & editing: Brian Bieber, Yeoungjee Cho, David W. Johnson, Talerngsak Kanjanabuch, Hideki Kawanishi, Yong-Lim Kim, Mark Lambie, Keith McCullough, Jeffrey Perl, Ronald L. Pisoni, Claudio Rigatto, Martin Schreiber, Jenny Shen, Karthik K. Tennankore.

Data Sharing Statement

For investigators interested in using PDOPPS data for a scientific investigation, a request in the form of a brief research proposal can be submitted at the DOPPS website (see details at https://www.dopps.org/PartnerwithUs.aspx).

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/CJN/C64.

Supplemental Table 1. STROBE checklist.

Supplemental Figure 1. Kaplan–Meier survival curves for mortality stratified by categories of pruritus severity (none, somewhat moderate, very much extreme).

Supplemental Figure 2. Cumulative incidence curves for transfer to hemodialysis stratified by categories of pruritus severity (none, somewhat moderate, very much extreme).