The relationship between chronic kidney disease-associated pruritus and health-related quality of life: a systematic review.

BACKGROUND: Chronic kidney disease-associated pruritus (CKD-aP) is a common and burdensome condition for end-stage kidney disease (ESKD) patients, especially those receiving haemodialysis. High-quality evidence of the relationship between CKD-aP and health-related quality of life (HRQoL) can therefore inform clinicians and policymakers about treatment choice and reimbursement decisions.
METHODS: A systematic literature review and narrative synthesis stratified by study design and HRQoL instrument was conducted to evaluate in adult ESKD patients receiving in-centre haemodialysis the relationship between CKD-aP and HRQoL assessed using multi dimensional generic or condition-specific preference- or non-preference-based measures. MEDLINE, Embase, Web of Science, BIOSIS Citation Index, Cochrane Library and PsycINFO from inception to March 2020 were searched, with two reviewers extracting data independently.
RESULTS: Searches identified 2684 unique records, of which 20 papers relating to 18 unique studies [5 randomised controlled trials (RCTs) and 13 observational studies] were included. HRQoL was assessed using four generic and eight disease-specific measures. The impact of CKD-aP was assessed by comparison of means, linear regression and correlation. Observational studies employing comprehensively adjusted multivariable linear regression largely found associations between CKD-aP severities and HRQoL. Analyses suggest this relationship is partially mediated by the sleep disturbance caused by CKD-aP. RCTs showing improvements in CKD-aP severity were associated with clinically meaningful improvements in HRQoL. Compared with generic measures, disease-specific HRQoL instruments reported greater changes with reduced CKD-aP. Heterogeneity in study design and reporting precluded meta-analysis.
CONCLUSIONS: CKD-aP severity was found to be associated with a worsening of HRQoL in the majority of observational and RCT studies. Parallel improvements in CKD-aP and HRQoL with interventions may support their use (PROSPERO registration 175035).

INTRODUCTION

Uraemic or chronic kidney disease–associated pruritus (CKD-aP) is a common, burdensome and undertreated condition in end-stage kidney disease (ESKD) patients, particularly in those receiving haemodialysis (HD) [1]. There is increasing recognition that urea, other uraemic toxins and phosphate do not have a mechanistic role in the development of pruritus in this patient group, leading to the revised term CKD-aP. Existing evidence suggests that the severity of the pruritus has been associated with depression, poor sleep quality, increased mortality and reduced health-related quality of life (HRQoL) [2]. This was identified as an outer tier condition in stakeholder exercises by the Standardising Outcomes in Nephrology initiative [3]. A range of interventions have demonstrated varying success as therapies for CKD-aP [4].

Combining the improvements in HRQoL with any cost savings for pharmaco- and non-pharmacotherapy for CKD-aP allows for estimation of the health economic benefits of improving CKD-aP in ESKD patients, more specifically the estimation of the cost per quality-adjusted life year (QALY) [5]. Decision-makers use QALY cost thresholds to make reimbursement decisions [e.g. the UK's National Institute for Health and Care Excellence (NICE)], to set value-based pricing (the US Institute for Clinical and Economic Review) or to inform prioritization of research questions [6].

Current and best available evidence relating to the impact of CKD-aP on HRQoL is needed to support future research for optimal therapy and inform cost-effectiveness analysis. To support this, we conducted and report here a systematic review of the relationship between CKD-aP in HD patients and HRQoL.

MATERIALS AND METHODS

We conducted a systematic literature review to understand the relationship between CKD-aP and HRQoL in patients receiving HD for ESKD, following good practice and reporting guidelines recommended by the Centre for Reviews and Dissemination [7] and the NICE Decision Support Unit [8]. The protocol is registered on PROSPERO (175 035).

Literature searching

We conducted electronic literature searches in March 2020 to retrieve studies characterizing the impact of CKD-aP on HRQoL of patients on HD. We searched the following databases from the date of inception: MEDLINE, Embase, Web of Science: Science Citation Index Expanded and Conference Proceedings Citation Index, BIOSIS Citation Index, Cochrane Database of Systematic Reviews and Central Register of Controlled Trials and PsycINFO. A multidisciplinary team including an information specialist and a kidney clinical expert developed a search strategy (Supplementary data, File 1). We also searched several grey literature sources (ClinicalTrials.gov, National Institute for Health Research Journals Library, Kidney Care Research UK, Kidney Care UK, Open Grey and World Health Organization Global Index Medicus) and reference lists of included studies. All retrieved records were collected and transferred into a reference management database (Clarivate 2013, EndNote X9 Philadelphia, PA, USA).

Selection criteria for studies

Using pre-specified criteria, two reviewers independently examined the title, then the abstract and finally the full text of all retrieved records. Where there were disagreements on study selection, a third reviewer or clinical expert was consulted. Randomized controlled trials (RCTs), non-RCTs and observational studies were eligible for inclusion in the review if the studies reported the association between pruritus and HRQoL in a well-defined population or sub-group of people with ESKD who were undergoing in-centre HD; measured a patient's exposure to pruritus by symptom scores, visual analogue scale (VAS), itch intensity scores or any other investigator or patient-reported pruritus measure and included patient-reported HRQoL using multidimensional, generic, condition-specific or study-specific measures. To allow a robust investigation of the association between pruritus and HRQoL, we excluded studies only conducted in patients with coexisting comorbidities or conditions that may influence HRQoL, studies with <30 patients and conference abstracts.

Data extraction, methodological assessment and analytical approach

We extracted data using a pre-piloted data extraction form (Supplementary data, File 2). All data extractions were completed by one reviewer and checked by a second reviewer. As quality assessment of studies reporting HRQoL is not standardized, a bespoke assessment tool (Supplementary data, File 3) was developed following relevant recommendations [7, 8].

A meta-analysis with relevant subgroup analyses to assess the association between CKD-aP and HRQoL was planned but was not possible due to the considerable heterogeneity of the available literature. Therefore we summarized data and presented narrative syntheses, grouping studies according to study type (interventional or observational) and analysis performed (e.g. multivariable, comparison of means or correlation).

RESULTS

Following removal of duplicates, we screened 2684 unique records. Of these, 20 full-text papers relating to 18 unique studies (studies with experimental design, n = 5 [9-13]; studies with non-experimental design, n = 13 [2, 14–29]) were included in the review (Figure 1).

Figure 1:

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) diagram of study selection.

Included studies

The study characteristics of included experimental and observational studies are presented in Table 1 and Table 2, respectively. Five studies with experimental study designs [9-13] included two international multicentre studies [9, 13], two single-country multicentre studies in Iran [10] and Turkey [12] and a single-centre study conducted in Brazil [11]. Three RCTs had only active comparator arms, with 30–128 patients per arm, and compared doxepin versus pregabalin [10], gabapentin versus dexchlorpheniramine [11] and nalbuphine 120 mg versus 60 mg [13]. Three studies compared an inactive treatment arm with an active control arm (35–189 patients per arm) and compared baby oil [12], difelikefalin [9] and nalbuphine (two different doses) [13] with an inactive control. Eligibility criteria (Table 1) varied according to the frequency and severity of pruritus.

Table 1.

Included studies with an experimental design: study characteristics

First author, year, study design, country	Comparisons	Inclusion criteria (brief description)	Exclusion criteria (brief description)	Patient demographics (age, gender, comorbidities	Pruritus measurement tool(s) and details	HRQoL instrument(s) and timing
Gobo-Oliveira 2018 [11]Randomized double-blind, controlled, parallelBrazil (1 dialysis unit, Clinical Hospital of the Botucatu Medical School)	Gabapentin, n = 30versusDexchlorpheniramine, n = 30	Age >18 years, HD for at least 3 months, pruritus at least thrice a week and lasting ≥30 days)Only patients who completed phase 1 (T0 to T15, i.e. 15 days of cold cream) and still complained of pruritus were randomized.	Chronic skin disease (allergic, parasitic, or infectious), internal malignancyUse of opioids or corticosteroids	Age mean 59 ± 12 yearsMale 19 (63%)	VAS scale:Cut-offs, not specifiedCharacterized period or duration of pruritus as: episodic during or shortly after HD (score of 1), day or night (score of 2), or all the time (score of 3)	DLQI questionnaire
Foroutan 2017 [10]Single blind RCTIran (6 dialysis centres of the Kerman University of Medical Sciences)	Doxepin, n = 44versusPregabalin, n = 46	Patients on haemodialysis aged 16 to 80 years experiencing pruritus	Patients with: major systemic conditions and other conditions associated with pruritus	Age mean 59.7 ± 15.8 yearsMale 54 (75%)	VAS scale:0 points, no itching0.1 to 3.9 points, mild4 to 6.9 points, moderate7 to 8.9 points, severe9 to 10 points, very severe	DLQI questionnaire
Karadag 2014 [12]Non-randomized controlled trial—pretest–post-test model with control groupsTurkey (1 state hospital in 2 provinces)	Baby oil, n = 36versusNo baby oil, n = 35	Intermittent pruritus for 6 months plus three episodes of VAS ≥ 5 within last 2 weeks	Oedema; open wound, cellulitis, infection, DVT, epilepsy, haemorrhage, paraplegia or pacemaker; Pruritus treatmentThrombocytosis,	Age mean 54.0 ± 16.0 yearsMale 41 (54.7%)	VAS scale:0–2 cm, no itching;3–4 cm, mild itching;5–6 cm, moderate itching;7–8 cm, severe itching;9–10 cm, unbearable itchingitching severity scale (ISS)(4-item scale)	SF-36 Quality of Life Scale
Fishbane 2020 [9]KALM-1 (NCT03422653)RCTUSA	Difelikefalin, n = 189Placebo, n = 189	Age >18 years, moderate-to-severe pruritus on HD at least three times weekly for at least 3 months	Pruritus due to other non-ESKD causes; scheduled kidney transplant; allergic to opiates	Difelikefalin: Age mean 58.2 ± 11.2 yearsMale 112 (59.3%)Placebo:Age mean 56.8 ± 13.9 yearsMale 118 (62.8%)	WI-NRS, 0–10 pointsHigher scores indicate worse itch	5-D itch scaleSkindex-10
Mathur 2017 [13]RCTUSA, Romania, Poland	Group 1: Nalbuphine 120 mg, n = 120Group 2: Nalbuphine 60 mg, n = 128Group 3: placebo, n = 123	HD for ≥ 3 months,Patient Assessed Disease Severity category of ‘B’ or ‘C, mean WI-NRS score = 6 during the week prior to randomization	Pruritus due to cholestasis, atopic dermatitis, lymphoma or any condition unrelated to end-stage renal disease		WI-NRS, 0–10 points(no itching to worst possible itching)Baseline: Daytime and night-time WI-NRS score before three dialysis visitsFollow-up: Daytime and night-time WI-NRS score weekly for 10 weeks	Skindex-10

ER, extended release.

Table 2.

Included studies with an observational design: study aims and population characteristics

First author, reported study design country study/trial name Relevant comparisons	Primary aims/objectives of the study	Inclusion and exclusion criteria (brief description)	Patient demographics (age, gender)	Pruritus measurement	HRQoL measurement	Timing/frequency of assessments (pruritus; HRQoL)
Afsar 2012 [29]Cross-sectional studyTurkeyDiabetics and non-diabetics with CKD-aP	To assess relationship between pruritus and HbA1c	Inclusion: HD patients receiving no systemic treatment for pruritus and neuropathy, includingantihistamines and gabapentin;no concomitant dermatological, liver or metabolic diseases-associated with pruritusExclusion: Not reported	Age 51.9 ± 13.5 yearsMales 41 (54.7%)	VAS10 cm horizontal line:0 (indicating no itch) to 10 (indicating worst possible imaginable itch)	SF-36	Baseline, only (cross-sectional study)
Curtin 2002 [14]Cross-sectional studyUSA	To examine the relationship between symptoms in HD patients and SF-36 PCS and MCS outcomes	Inclusion: Age, ≥18 yearsAble to read and write in EnglishExclusion: Not reported	Age 58.2 ± 15.10 yearsMales 147 (47.9%)	Likert, frequency (5 categories)Patients reported frequency of experiencing pruritus and 46 additional listed symptoms in the last 4 weeks, prior to the study.Responses0 = never1 = a little of the time2 = some of the time3 = most of the time4 = all of the time	SF-36	Baseline, only (cross-sectional study)
Tayebeh 2017 [25]Observational cohort studyIran	To examine predictors of pruritus and insomnia, the impact of pruritus on quality of health, sleeping problems, hospitalization and mortality in HD patient	Age ≥18 yearsMaintenance HD patients; receiving HD ≥ 2 weeksExclusion: Not reported	Age 57.2 ± 15.4 yearsMales 235 (56.5%)	NRSelf-reporting of severity of itching on a 5-point Likert scale	SF-36	NR followed-up for 28 months
Ibrahim 2016 [16]Cross-sectional case-control studyEgyptHD patients with UP and those without UP	To assess the influence of UP on QOL by comparing HD patients with UP to those without UP	Age ≥18 yearsUndergoing regular HDExclusion: Other possible causes of pruritus, such as: skin diseases or haematological diseases	Age 49.5 ± 11.5Males 37/100 (37%)	NRS, limited details about how comparator groups were identified	WHOQoL-BREFa	Baseline, only (cross-sectional study)
Rehman 2019 [21]Cross-sectional studyPakistan	To examine the relationship between CKD-aP and QOL in HD patients	Age ≥18 yearsboth genders, undergoingHD; proficient in the Urdu language and willing to participateExclusion: Not reported	Age median 42.0 (range NR, IQR 35.0–51.0)Male 176 (67.2%)	Urdu 5-D itch scale5 domains (duration, degree, direction,disability, distribution)Overall score—sum of scores for all 5 domains. A score <5 indicates no pruritus, a score of 25 indicates severe pruritus	Urdu FANLTC scaleb	Baseline, only (cross-sectional study)
Rehman 2020 [22] Cross-sectional studyMalaysia	To examine the association between CKD-aP and QOL in HD patients	Age ≥18 yearsboth genders, undergoingHD; Proficient in the Malay languageExclusion: None	Age: median 58.00 (IQR 47.00–67.00)Male 81 (40.7%)	Malay 5-D itch scale4 domains: duration, degree, direction, overall score—sum of scores for all 5 domains. A score <5 indicates no pruritus; a score of 25 indicates severe pruritus	Malay FANLTC scale	Baseline, only (cross-sectional study)
Satti 2019 [23]Cross-sectional studyPakistan	To report the prevalence of UP and identify a subset of patients at high risk for UP.To investigate the impact of UP on QOL in HD patient	Male patients with ESKD on HD ≥3 monthsExclusion: Mini-mental state examination score <9Other causes of pruritus, (psoriasis, eczema or dermatitis)	Age not reported for subgroup of n = 85 (with pruritus) who were analysedMale 173 (100%)	5-D itch	DLQIc	Baseline, only (cross-sectional study)
Susel 2014 [24]Cross-sectional studyPoland	To examine the effect of UP on depressive symptoms and QOL in patients with ESKD	Adults with ESKD undergoing regular HDExclusion: Other causes or pruritus	Age mean 59.05 years (range 22–88) male 124 (62%)	VAS0–10 scale4-item Itch questionnaire	SF-36 PCS and MCSDLQI	Baseline, only (cross-sectional study)
Tessari 2009 [26]Cohort studyItaly	To examine the impact of pruritus on QOL in patients receiving chronic dialysis	Consecutive patients treated with HD and PD with pruritus occurring regularly for 6 months or at least thrice in the 2 weeks or less of study entry.UP was accepted if pruritus was experienced after dialysis in the absence of an active condition-associated pruritusExclusion: Other causes or pruritus	HD + PD:Age 62.2 ± 13.7 years, (range 22.7–79.4)Male 111 (66%)d	VAS	SF-36Skindex 29	Baseline, only (cross-sectional study)
Lopes 2012 [18]Cross SectionalBrazil—(four dialysis units)Prospective Study of the Prognosis of Chronic Hemodialysis Patients (PROHEMO)	To investigate the impact of depressive symptoms, poor sleep and dry skin bother on the association between pruritus and kidney disease in HD patients.	Adult HD patients—no further informationExclusion: Not reported	No pruritus,Age 47.60 ± 14.3 yearsMale 323 (58.6%)Mild pruritus,Age 49.59 ± 14.14 yearsMale 150 (62.8%)Severe pruritus,Age 51.21 ± 13.52 yearsMale 112 (58.9%)	Kidney Disease Quality of Life Short Form (KDQOL-SF)—pruritus subscalePatients indicated the extent to which they were bothered by itchy skin (pruritus) and dry skin for the last 4 weeks, prior to data collection	Kidney Disease Quality of Life Short Form (KDQOL-SF)e—disease burden subscale	Baseline, only (cross-sectional study)
Kimata 2014 [17]Observational studyJapanDOPPS	To estimate the prevalence of pruritus, and examine its relationship to QOL, sleep quality, medication use, and mortality in HD patient	See Pisoni 2006 [2] for general DOPPS inclusion and exclusion criteria.	Moderate. to extreme bother by itch, JDOPPS 1 + 3Age 61.7 ± 12.6 yearsMale 1923 (67%)No bother to somewhat bother by itch, JDOPPS 1 + 3Age 59.38 ± 12.5 yearsMale 2093 (58%)	Study-specific assessmentPatients indicated the extent to which they were bothered by itchy skin during a 4-week period prior to data collectionResponses werenot at all bothered or somewhat, moderately, very much or extremely bothered	SF-36 or SF-12	Baseline, only (cross-sectional study)
Mathur 2010 [19]Cohort studyUSAITCH National Registry Study	To study the natural history of UP; to compare rating scales of itching intensity and investigate the assess utility of HRQOL instruments for CKD-aP	Age ≥18 years, had been receiving chronic HD ≥3 times a week. Pruritus (defined as a score of 10 mm on a 100-mm VAS). Able to understand and complete the patient questionnairesExclusion: Other causes or pruritus. Recent change in HD regime	Patient Type AAge 63 ± 11.7 yearsMale 20 (57%)Patient Type BAge 54 ± 15.3 yearsMale 21 (49%)Patient Type CAge 53 ± 3.2 yearsMale 13 (50%)	Study-specific UP intensity scales, adapted from the 100-mm VAS and the 11-point NRSAssessment of worse itching over the preceding 24 hours, with separate measurements for worst daytime and worst night-time itching	Skindex-10Brief Itching Inventory	Scheduled intervals (not specified) over 3.5 months
Pisoni 2006 [2] Prospective, longitudinal, observational studyMultiple countriesDOPPS	To estimate the prevalence of pruritus and examine its relationship to QoL, sleep quality, medication use and mortality in HD patient	A dialysis unit treating ≥25 HD patients within the unit to be eligible for study participation (participating facility eligibility criteria)Age ≥17 years (patient eligibility criterion)Exclusion: Other dialysis modalities or locations	Moderate to extreme itchAge 60.7 ± 14.49 years Male 3459 (60.3%)Mild itchAge 60.3 ± 14.64 years Male 4266 (56.4%)	Study-specific assessmentPatients indicated the extent to which they were bothered by itchy skin during a 4-week period prior to data collection	SF-36 or SF-12	Baseline, only (cross-sectional study)
Plewig 2019 [20]Cohort studyGermanyGEHIS	To examine the impact chronic itch in HD patients over time and its association with HRQoL, comorbidities and laboratory values.	All patients who had participated in GEHIS study in 2013 and had agreed to be contacted againExclusion: Not reported	All patients,age mean 60.8 ± 13.3 years(range 29.0–89.2)Male 57 (54.8%)fPatients with no current but previous chronic itchAge 58.9 ± 12.9 yearsMale 33 (63.5%)Patients with persistent chronic itchAge 62.7 ± 13.6 yearsMale 24 (46.2%)	VAS, 0 (no itch) to 10 (worst imaginable itch)IFSI classification based on dermatological examination	ItchyQoLg	2017 (4-year follow-up data of 2013 GEHIS study)
Weiss 2016 [27, 28]Cross-sectional studyGermanyGEHIS	To examine the prevalence and impact of chronic itch on QOL in patients with ESKD	Age ≥18 yearsDiagnosis of ESKD and on HD treatmentProficient in German languageNo current cognitive impairmentExclusion: Not reported	All participantsAge 68.2 ± 13.9 yearsMale 492 (57.2%)Patients with current chronic itchAge 64.3 ± 13.8 yearsPatients with no chronic itchAge 68.2 ± 13.2 years	VAS, 0 (no itch) to 10 (worst imaginable itch)	ItchyQoL SF-12	Baseline, only (cross-sectional study)

aWHOQoL-BREF: 26 items; four domains: physical health (four items), psychological health (six items), social relationships (three items) and environmental health (eight items); it also contains QOL and general health items.

bFANLTC: 26- item scale with 4 subscales (physical well-being, social/family well-being, emotional well-being and functional well-being). In each subscale, each item is scored from 0 (‘not at all’) to 4 (‘very much’). The scoring of positively stated items is 4, 3, 2, 1 and 0, while the negatively stated items are reverse scored. Subscale scores equal the sum of item scores multiplied by the number of items in each subscale and divided by the number of items answered. Overall FANLTC score equals the sum of the 4 subscale scores (range 0–104, with a higher score indicating better quality of life.

cScore of 0–1 = no effect of pruritus, 2–5 = small effect, 6–10 = moderate, 11–20 = large effect and 21–30 = very large and severe limiting effects on the patient's life.

dData for HD and PD patients.

eAssessment of perceived disease burden related to impact of patient's life (including time commitment for care) emotions, specifically frustration and family life. Responses (five options) ranging from definitely true to definitely false.

fGender data relate to 2013 data.

g22 items regarding symptoms, functions, emotions and self-perception.

BII, Brief Itch Inventory; FANLTC, Functional Assessment for Non-Life-Threatening Conditions; PROHEMO, Prospective Study of the Prognosis of Chronic Haemodialysis Patients.

Thirteen unique observational studies reported in 16 papers and a parallel publication [2, 14–29] were included in the review (Table 2). Eligible studies included five cohort studies: the Dialysis Outcomes and Practice Patterns Study (DOPPS) (an international multicentre study) [2, 17], the German Epidemiological Haemodialysis Itch Study (GEHIS) [20, 28], the ITCH National Registry Study [19] and two studies conducted in Iran and Italy [25, 26]; and eight cross-sectional studies from the USA [14], Egypt [16], Pakistan [21, 23], Malaysia [22], Poland [24], Brazil [18, 27, 28] and Turkey [29]. Evaluation of the impact of pruritus on HRQoL in HD patients was reported as one of several study outcomes in 12 observational papers [2, 14, 16–24, 26, 28]; however, it was the primary aim of only 3 papers [14, 26, 28] to estimate the impact of pruritus on HRQoL.

Methodological quality of included studies

Based on the bespoke assessment tool developed following accepted recommendations [7, 8], all studies fell into the ‘good’ or ‘fair’ category, although none scored well across all items (Table 3). Issues for both study types included representativeness of the sampled patients and lack of sample size calculations relating to the assessment of the association between CKD-aP and HRQoL. Methods for handling missing data were not often presented. Few RCT studies defined and measured confounders and observational studies did not always adjust for all important potential confounders relating to patient characteristics, comorbidities and biochemical variables. The severity of pruritus for patients to be included was not always well specified, while HRQoL was more consistently well-defined across studies.

Table 3.

Summary of risk of bias of included studies

	Patient selection and allocation		Outcome reporting					Analysis
Author, date	Representative sample	Groups balanced at baseline	Pruritus definition and assessment	Pruritus measured same way in all	HRQoL definition and assessment	HRQoL measured in same way in all	Follow up period	Sample size	Groups comparable at analysis	Complete outcome data for all patients	Reported loss to follow up with reasons	Accounted for missing data	Confounders defined and measured	Appropriate accounting for confounding was undertaken	Overall rating
Experimental studies
Fishbane 2020 [9]	N	Y	Y	Y	Y	Y	CT	CT	Y	N	Y	Y	N	P	Good
Foroutan 2017 [10]	N	Y	P	Y	Y	Y	CT	CT	Y	N	Y	N	N	P	Good
Gobo-Oliveira 2018 [11]	N	Y	Y	Y	P	Y	CT	CT	P	N	Y	Y	P	P	Good
Karadag 2014 [12]	N	CT	P	Y	Y	Y	CT	CT	Y	N	Y	CT	N	N	Fair
Mathur 2017 [13]	N	Y	P	Y	Y	Y	CT	CT	P	N	P	Y	N	P	Good
Observational studies
Afsar 2012 [29]	CT	NA	Y	Y	Y	Y	NA	CT	CT	CT	CT	CT	P	N	Fair
Curtin 2002 [14]	Y	NA	Y	Y	Y	Y	NA	CT	CT	N	N	N	P	P	Good
Ibrahim 2016 [16]	CT	NA	N	Y	Y	Y	NA	Y	N	Y	Y	Y	N	N	Good
Kimata 2014 [17]	Y	NA	P	Y	Y	N	NA	P	N	N	N	N	P	P	Good
Lopes 2012 [18]	CT	NA	Y	Y	Y	Y	NA	CT	P	CT	CT	Y	P	P	Good
Mathur 2010 [19]	CT	NA	P	Y	Y	Y	CT	CT	N	CT	CT	CT	P	N	Fair
Pisoni 2006 [2]	Y	NA	Y	Y	P	N	NA	P	N	N	N	N	P	P	Good
Plewig 2019 [20]	CT	NA	CT	Y	Y	Y	Y	CT	P	N	N	N	N	N	Fair
Rehman 2019 [21]	CT	NA	Y	Y	Y	Y	NA	Y	N	Y	Y	NA	P	P	Good
Rehman 2020 [22]	CT	NA	Y	Y	Y	Y	NA	Y	N	Y	Y	NA	P	P	Good
Satti 2019 [23]	P	NA	Y	Y	Y	Y	NA	Y	N	Y	Y	NA	N	N	Good
Susel 2014 [24]	CT	NA	Y	Y	Y	Y	NA	Y	CT	N	P	N	N	N	Fair
Tayebeh 2017 [25]	CT	NA	Y	Y	Y	Y	NA	CT	N	CT	CT	CT	P	P	Fair
Tessari 2009 [26]	Y	NA	Y	Y	Y	Y	NA	CT	P	N	N	CT	P	P	Good
Weiss 2016 [28]	P	NA	CT	Y	Y	Y	NA	CT	CT	N	N	N	P	P	Fair

N, no; Y, yes; CT, can't tell; P, partial; NA, not applicable.

Measurements of pruritus and HRQoL

There was extensive clinical and methodological heterogeneity in assessment and reporting of the impact of CKD-aP on HRQoL in experimental and observational studies (Table 1 and Table 2). Pruritus instruments included study-specific tools, generic scales assessing intensity of itchiness [VAS and Worst Itch Numeric Rating Scale (WI-NRS)] and condition-specific measures comprising items assessing duration, degree, distribution and disability related to itchiness (e.g. 5-D itch scale). HRQoL instruments also varied across studies. These included four generic and eight disease-specific measures, which are listed with the domains assessed in Table 4. Follow-up periods varied from 1 to 12 weeks in experimental studies, while the majority of epidemiological studies [2, 17, 20, 27, 28] were cross-sectional analyses.

Table 4.

Summary of domains assessed by identified QoL instruments

Instruments	High-level QoLa	Pain	Energy	Appearance	Mobility	Sleep	Anxiety/depression	Intrusion/achievement	Social	Work	Itch severity
Generic measures
SF-12 (PCS and MCS)	Y	Y	Y	N	Y	N	Y	Y	Y	Y	N
SF-36 total score and domain scores	Y	Y	Y	N	Y	N	Y	Y	Y	Y	N
WHOQOL-BREF	Y	Y	Y	Y	Y	Y	Y	N	Y	N	N
FANLTC	N	Y	Y	N	N	Y	Y	N	Y	Y	N
Disease-specific measures
Brief Itching Inventory	N	N	N	N	N	Y	Y	Y	Y	Y
DLQI	N	N	N	N	N	N	N	Y	Y	Y	Y
KDQOL-SF36	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y
Skindex-10	N	N	N	Y	N	N	Y	N	Y	Y	Y
Skindex-29	N	Y	Y	N	N	Y	Y	Y	Y	Y	Y
ItchyQoL	N	Y	N	Y	N	Y	N	N	N	Y	N

Y: yes, the domain is assessed by instrument; N: no, the domain is not assessed by instrument

aThe patient is asked ‘What is your quality of life today?’

FANLTC, Functional Assessment for Non-Life-Threatening Conditions.

In interventional studies, pruritus was assessed using the VAS [10-12], WI-NRS, [9, 13], Itch Severity Scale (ISS) [12] and 5-D itch scale [10]. Four observational studies utilised the VAS [20, 24, 26, 29], four used ordinal severity (none to extreme, five levels) [17, 2, 16, 18], three observational studies used the multi dimensional 5-D itch scale [21-23] and one study asked about the presence or absence of pruritus [20].

Association between pruritus and HRQoL

Eligible outcomes were reported as mean for the entire study population or different pruritus categories at baseline or specified time points using different scales, instruments, recall periods and time points (Supplementary data, File 4 and Supplementary data, File 5). Overall, studies reporting condition-specific and itch-specific HRQoL outcomes [e.g. Brief Itch Inventory, Dermatology Life Quality Index (DLQI) questionnaire, ItchyQoL, Skindex-10 and Skindex-29] showed a more consistent relationship between worsening pruritus and worse perceived HRQoL compared with generic tools such as the 12- and 36-item Short Form Health Surveys (SF-12 and SF-36).

Non-experimental studies

Two studies [2, 17] reported SF-36 or SF12 scores using DOPPS data (n = 9659 patients) [2] and a subgroup from Japan [Japanese DOPPS (JDOPPS), n = 3755 patients] [17]. The prevalence of moderate to extreme pruritus over a 4-week recall period ranged between 36% and 50%. Following adjustment for patient and dialysis characteristics, comorbidities and laboratory variables, relative to no itching, there was a linear reduction in SF-12 mental component summary (MCS) and physical component summary (PCS) scores of between 1.5 and 8.6 points as pruritus worsened from ‘bothered somewhat’ to ‘extremely bothered’ in both studies [2, 17]. Following the inclusion of sleep disturbance in the regression, the effect size of the association was reduced to between 0.4 for ‘somewhat bothered’ to 4.4 points for ‘extremely bothered’, suggesting that some of the impact on HRQoL by CKD-aP is mediated by sleep quality [17].

Frequently experienced symptoms and their association with SF-36 MCS and PCS scores were examined in 307 HD patients randomly selected from 14 dialysis centres in the USA, in whom 73.2% experienced pruritus at least a little of the time [14]. In a multiple linear regression model including age, gender, race, diabetes, difficulty with sleep, mobility, sexual concerns and the symptoms of dry mouth, restless legs and lack of appetite, increasing pruritus severity (e.g. from some to most of the time) was associated with a 1.43-point reduction in the SF-36 PCS [95% confidence interval (CI) −0.69 to −2.17], but no reduction in MCS [0.25 points (95% CI −0.61–1.11)] [14].

Moderate to extreme pruritus was experienced by 38% of the 532 HD patients sampled from nine facilities to identify predictors of pruritus and insomnia and the impact of pruritus on HRQoL, sleep, hospitalization and mortality [25]. A significant inverse correlation between SF-36 PCS but not MCS scores and pruritus severity was reported alongside progressive reductions in total and component scores from 56 to 59 in the ‘no pruritus’ group to between 35 and 39 for the ‘extreme pruritus’ group. The authors reported that using multivariate linear regression ‘20% variation of SF36 score was respectively explained by insomnia and pruritus’; however, the adjusted effect sizes for this model were not reported [25].

The impact of pruritus (VAS) on HRQoL (SF-36 and Skindex-29) was assessed in 139 HD and 30 peritoneal dialysis (PD) patients [26]. The prevalence of pruritus in HD and PD patients was similar (52%), with one-third of these patients reporting pruritus as continuous. The dialysis modality (HD or PD) was a covariate in a multiple linear regression on all 168 dialysis patients. The regression did not find a significant relationship between a unit change in pruritus VAS score and SF-36 PCS and MCS scores (PCS: 0.07, P = 0.535; MCS: 0.01, P = 0.483). However, an association between pruritus scores and the Skindex-29 subscales of symptoms (0.34, P = 0.012), social functioning (0.29, P = 0.01) and emotions (0.22, P = 0.039) were noted [26].

Three publications [20, 27, 28] based on the GEHIS (n = 860 patients) reported SF-12 and ItchyQoL outcomes. Comparing patients with and without pruritus, the prevalence of pruritus (not considering severity) at the time of questioning was 25.2%, with 35.2% reporting having chronic pruritus previously. Statistically significantly worse SF-12 PCS (34.6 versus 37.1) and MCS (52.0 versus 54.3) scores were reported in those with pruritus [27, 28]. A multivariable linear regression adjusting for various patient characteristics including pain, sleep impairment, anxiety and depression reported a significant correlation between mean chronic itch severity and ItchyQoL total score (ß = 0.40), with higher ItchyQoL scores indicative of the greatest impairments in the emotions and self-efficacy subscales (0.46 and 0.34, respectively; P < 0.001) [28]. These effect sizes were greater when evaluating the worst severity of chronic itch in the last 6 weeks compared with severity at the time the instruments were being completed. There was no significant relationship between ItchyQoL scores and sleep quality, although pruritus severity was not specifically evaluated. In a separate univariate analysis comparing patients with persistent chronic itch from study entry to the GEHIS (n = 52) and those whose itching at baseline had resolved between 2013 and 2017 (n = 52) [13], those with resolved chronic itch had statistically better ItchyQoL scores (1.7 versus 2.1; P < 0.05). However, no significant difference in SF-12 MCS and PCS scores were observed for the relevant subgroups.

One study used the Brazilian version of the Kidney Disease Quality of Life Short Form (KDQOL-SF) to assess the extent of bother due to pruritus and dry skin (i.e. perceived disease burden) over the previous 4-week period in 980 patients recruited from four dialysis units in Brazil [18]. Mild pruritus or worse was experienced by 43.8% of patients. The five-level 36-item KDQOL-SF pruritus category was recoded into three categories in a multivariable logistic regression analysis predicting the KDQOL-SF disease burden (total score) with sequential adjustment for sociodemographic, HD and laboratory variables and comorbidities. Although severity of pruritus had a negative impact on the perceived burden of disease in univariate analyses (−5.11 and −11.2 for mild and severe categories, respectively, compared to none), no statistically significant relationship was observed following adjustment for depression, sleep and bother with dry skin.

One case–control study [16] assessed pruritus using the numerical rating scale (NRS) and HRQoL with the 26-item World Health Organization Quality of Life Brief (WHOQoL-BREF) instrument, which assesses the following: physical health, social relationships, psychological health and environmental health. Overall, a significant difference in HRQOL in all four domains was noted between patients with CKD-aP compared with patients without pruritus (Table 5).

Table 5.

Summary of findings: impact of pruritus on HRQoL

Author, year	Analysed, n	HRQoL measure	Pruritus measure	Study design	Analysisa	Statistical relationship summary	Overall assessment of relationship
BII
Mathur 2010 [19]	103	BII	VAS/NRS adaptation	Observational	Linear regression	P < 0.001, effect size not reported	✓
DLQI
Foroutan 2017 [10]	72	DLQI	VAS (0–5)	RCT	Mean at different time points	Statistical significance NR, data largely consistent with P-HRQoL inverse relationship	✓
Gobo-Oliveira 2018 [11]	60	DLQI	VAS (0–10)	RCT	Spearman's rank correlation	r = 0.76, P < 0.01	✓
Susel 2014 [24]	76	DLQI	VAS (0–10)	Observational	Spearman's rank-order correlation	r = 0.56, P < 0.0001	✓
Satti 2019 [23]	173	DLQI	5-D Itch	Observational	Spearman's rank-order correlation	r = 0.78, P < 0.000	✓
Susel 2014 [24]	76	DLQI	4-Item Itch	Observational	Spearman's rank-order correlation	r = 0.48, P < 0.0001	✓
FANLTC
Rehman 2019 [21]	262	FANLTC	5-D itch (Urdu)	Observational	Multivariate linear regression	Statistically significant association between pruritus severity and HRQoL [ß = −0.949 (95% CI −1.450 to −0.449)]	✓
Rehman 2020 [22]	205	FANLTC	5-D itch (Malay)	Observational	Bivariate analysis with Pearson's correlation coefficient	r = −0.282, P ≤ 0.001	✓
ItchyQoL
Plewig 2019 [20]Weiss 2016 [28](GEHIS cohort)	83 [20]	ItchyQoL	VAS (0–10)	Observational	Mean HRQoL in different pruritus categories, t-test/ANOVA	Statistically significantly worse HRQoL by pruritus severity for most subscales and total score	✓
	189 [28]	ItchyQoL	VAS (0–10)	Observational	Multivariable linear regression	Mean severity, worst severity in past 6 weeks and at-the-time severity statistically significant Beta (range 0.23–0.46) for total score, and symptom, functional and emotion domains for most analysesb
KDQOL-SF
Lopes 2012 [18]	980	KDQOL-SF	KDQOL-SF	Observational	Linear regression (various adjustments)	No statistically significant association after adjustment for depression, sleep and bother with dry skin	✗
SF-36
Karadag 2014 [12]	70	SF-36	VAS (0–10)	RCT	Mean at different time points	Statistical significance NR, data largely consistent with P-HRQoL inverse relationship	✓
Afsar 2012 [29]	75	SF-36	VAS (0–10)	Observational	Linear regression analysis	Not statistically significant	✗
Susel 2014 [24]	76	SF-36	VAS (0–10)	Observational	Spearman's rank-order correlation	r = −0.35, P = 0.002	✓
Tessari 2009 [26]	169c	SF-36	VAS (0–10)	Observational	Multiple linear regression	PCS: 0.07, P = 0.535c MCS: 0.01, P = 0.483c	✗
Karadag 2014 [12]	70	SF-36	ISS	RCT	Mean at different time points	Statistical significance NR, data largely consistent with P-HRQoL inverse relationship	✓
Curtin 2002 [14]	306	SF-36	Likert, frequency (5 categories)	Observational	Linear multiple regression (adjusted)	PCS: β = –1.43, P < 0.000 MCS: β = –0.25, P = 0.564	PCS: ✓ MCS: ✗
Susel 2014 [24]	76	SF-36	4-Item Itch	Observational	Spearman's rank-order correlation	r = –0.43, P ≤ 0.001	✓
Tayebeh 2017 [25]	532	SF-36	Likert intensity (5 categories)	Observational	Multivariate regression analysis	20% of variation SF-36 score explained by pruritus (statistical significance NR)	✓
SF-12
Plewig 2019 [20]	87 [20]	SF-12	VAS (0–10)	Observational	Mean HRQoL in different pruritus categories, t-test/ANOVA	No statistically significant association	✗
SF-36 or SF-12
Kimata 2014 [17] (JDOPPS) Pisoni 2006 [2] (DOPPS)	3755	SF-36 or SF-12	Likert intensity (5 categories)	Observational	Mean HRQoL in different pruritus categories (adjusted)	MCS and PCS in pruritus patients, 2.3–6.7 points lower (P < 0.0001)	✓
	9659	SF-36 or SF-12	Likert intensity (5 categories)	Observational	Mean HRQoL in different pruritus categories	MCS and PCS 3.1–8.6 points lower (P < 0.0001)
Skindex-10
Fishbane 2020 [9]	189	Skindex-10	WI-NRS (0–10), 5D-itch	RCT	Mean change over time	Statistical significance NR, data largely consistent with P-HRQoL inverse relationship	✓
Mathur 2017 [13]	128	Skindex-10	WI-NRS (0–10)	RCT	Mean at different time points	Statistical significance NR, Some evidence of correlation from post-hoc analysis of most severe pruritus patients	?/✗
Mathur 2010 [19]	103	Skindex-10	VAS/NRS adaptation	Observational	Linear regression	P < 0.001	✓
Skindex-29
Tessari 2009 [26]	169c	Skindex-29	VAS (0–10)	Observational	Multiple linear regression	Symptoms score: β = 0.34, P = 0.012c Social functioning: β = 0.29, P = 0.01c Emotions: β = 0.22, P = 0.039c	✓
WHOQoL-BREF
Ibrahim 2016 [16]	100	WHOQoL-BREF	NRS (no details)	Observational	Mean in pruritus cases versus controls, Chi-squared.	Statistically significantly worse HRQoL in pruritus cases versus controls for most physical, social, psychological and environmental aspects of WHOQoL-BREF	✓

Note: Italics indicate adjusted analyses

✓: presence of a significant relationship between pruritus and HRQoL; ?: absence of a significant relationship between pruritus and HRQoL; χ: data inconclusive.

aWhere adjusted and unadjusted analyses reported, only the adjusted analysis was included in this table.

bExcept between ItchyQoL symptoms domain and VAS pruritus severity at the time of investigation (β = 0.16, P = 0.05); ItchyQoL functional domain and VAS worst severity in the past 6 weeks (β = 0.19, P = 0.02) and ItchyQoL pruritus functional domain and severity at the time (β = 0.19, P = 0.02). Significance level at P < 0.01.

cThe adjusted analysis included both HD and PD patients, but type of HD was a covariate in the multiple linear regression.

ANOVA, analysis of variance; BII, Brief Itch Inventory; FANLTC, Functional Assessment for Non-Life-Threatening Conditions; IFSI, International Forum for the Study of Itch; NR, not reported.

Experimental studies

Five RCTs evaluated treatments with itching severity instruments (WI-NRS, VAS) alongside changes in the SF-36 and disease-specific HRQoL measures. HD patients (n = 378) with a WI-NRS of >4 points were randomized to difelikefalin or placebo in a US study. [9] After 12 weeks, 51.9% patients experienced a ≥3-point improvement in the WI-NRS, compared to 30.9% in the placebo group, while the least squares mean change in the 5-D itch and Skindex-10 was −5.0 and −17.2 in the difelikefalin arm compared with −3.7 and −12.0 in the control arm (all endpoints <0.001). A single-blind study conducted in Iran randomized 72 HD patients to pregabalin or doxepin [10]. Although the level of pruritus severity at baseline was not reported, after 4 weeks of treatment the VAS improved from 7.5 to 2.1 in the pregabalin group compared with an improvement from 7.1 to 4.2 in the doxepine group. In parallel, greater improvements in the 5-D itch (19.2 to 8.5 versus 17.0 to 12.7) and the DLQI score (3.8 to 1.2 versus 3.6 to 2.2) were observed in the pregabalin group.

One study showed that gabapentin was comparable to dexchlorpheniramine after 21 days in 60 HD patients experiencing pruritus at least three times a week for a minimum of 30 days [11]. Patients were randomized following a 15-day run-in period with cold cream. Overall, median VAS scores significantly improved from 8 to 5 in the run-in period and 5 to 1 in the intervention period, while the median DLQI improved from 4 to 2 and from 2 to 1 during the respective periods [11].

Extended release nalbuphine at doses of 120 or 60 mg or placebo for 8 weeks were randomly assigned to 373 HD patients who during the preceding week had a mean NRS score ≥4.5, with two scores >5 [13]. From an overall mean of 6.9, the NRS improved by 3.5, 3.1 and 2.8, while the Skindex-10 improved by 17.0, 13.8 and 15.0 in the respective groups. Seventy Turkish HD patients with three episodes of itching lasting 5 minutes or longer during the preceding 2 weeks were randomized to cooled baby oil three times a week for 1 month or no intervention [12]. The VAS significantly improved by 2.51 in the baby oil group compared with 0.05 in the control, while the SF-36 PCS improved by 9.87 and 0.28 and the MCS improved by 8.45 and 0.82, respectively. Both of these studies reported improvements in sleep disruption where pruritus severity improved.

DISCUSSION

This systematic review examined the impact of CKD-aP on HRQoL in HD patients. Inclusion criteria, assessment methods, duration of follow-up and statistical techniques varied across included studies, precluding a meta-analysis. A number of studies reported the presence of a relationship between increased reductions in HRQoL as pruritus severity worsened. Persistent pruritus resulted in worsening HRQoL over time. Overall, condition-specific pruritus tools were more sensitive to changes in HRQoL. Improvements in sleep disruption were observed with improvement in pruritus severity as a result of interventions [12, 13]. Multivariable regressions on observational data suggested that sleep disruption partially mediates the relationship between pruritus and HRQoL [17].

This review highlighted the challenges in the conduct of research in this area. Swarna et al. [30] conducted a review of CKD-aP but did not focus on HRQoL. The authors also stated that ‘most studies are not comparable given their small group samples, study designs, and lack of standardized study measures’ [30]. However, this review has yielded new insights: the available longitudinal data on the condition of CKD-aP informs the natural history and placebo effect in this condition. Although the experimental studies [9-13] were unlikely to be adequately powered to assess these questions, a placebo effect for both pruritus and HRQoL was observed in blinded [11, 13] but not unblinded [12] randomized studies. This could also represent regression to the mean, as these studies included a severity threshold for inclusion; however, no studies reported data that could be used to assess for the presence of this phenomena.

The strengths of this systematic review include adherence to recommended standards and searches conducted on a broad selection of databases. Limitations include that 50% of included studies were not primarily designed to characterize the impact of pruritus on HRQoL and planned subgroup analyses were not feasible due to a lack of available evidence.

The available data could be used to consider if meaningful gains in HRQoL would be obtained through treatment of pruritus. The associated improvements in HRQoL exceeded 0.5 SD of the baseline value where reported, generally recognized as a clinically meaningful difference [31]. The placebo effect combined with the observation around the impact of persistent pruritus on HRQoL might argue for the serial measurement of pruritus in clinical practice. Generally, serial measurement of symptoms and responses to them are the subject of ongoing trials in nephrology, having shown promising results in other chronic diseases such as cancer [32, 33]. Further research could include assessing the appropriateness of generic HRQoL instruments [34] (e.g. European Quality of Life 5-Dimensions questionnaire) that are used to evaluate health technologies in Europe or to generate value-based prices in the USA. As HD patients prioritize fatigue as a core outcome measure in clinical trials [35] and sleep disruption tends to modify the relationship between HRQoL and pruritus, instruments that capture energy or fatigue should be prioritized.

In conclusion, this systematic review has demonstrated that CKD-aP severity was associated with HRQoL in the majority of observational and RCT studies. Parallel improvements in CKD-aP and HRQoL with interventions may support the clinical and health-economic argument for their use in clinical practice.

Click here for additional data file.