Carr CW, Veledar E, Chen SC. Factors Mediating the Impact of Chronic Pruritus on Quality of Life. JAMA Dermatol. 2014;150(6):613-620. doi:10.1001/jamadermatol.2013.7696
Copyright 2014 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Chronic pruritus is a common problem with a deleterious effect on quality of life. The factors mediating the relationship between chronic pruritus and quality of life are poorly understood.
To determine the factors that mediate the impact of chronic pruritus on quality of life.
Design, Setting, and Participants
Cross-sectional study based on a nationwide telephone survey of US military veterans sampled from the Veterans Hospital Patient Database. Participants were veterans who agreed to participate and were found to have chronic pruritus.
Main Outcomes and Measures
Statistical significance of predictor variables (characteristics of patients and their chronic pruritus) in multivariate analysis using ItchyQoL scores as response variables.
Of 6000 veterans telephoned, 1075 consented to participate; 405 of these had chronic pruritus. Factors that mediated the impact of chronic pruritus on quality of life were demographic characteristics (age [P = .007], race [P = .05], marital status [P = .04]), personality (extroversion [P = .03], neuroticism [P = .01]), pruritus characteristics (severity [P < .001], duration [P = .01], frequency [P < .001], location [P = .005]), and possible etiology (cutaneous vs systemic [P = .03]). Notable factors not found significant include sex (P = .98) and socioeconomic variables (education level [P > .99], employment status [P = .53], income [P = .62]).
Conclusions and Relevance
The multitude of factors that significantly influence the impact of chronic pruritus on quality of life demonstrates the complex nature of this relationship. A better appreciation of these factors will improve the clinical evaluation and treatment of chronic pruritus.
The scourge of chronic pruritus—pruritus lasting 6 or more weeks1—should not be underestimated. Recent studies have documented the extent of the problem, finding point prevalences ranging from 13% to 17%2- 4 and lifetime prevalences of 22% to 26%.2,4 Prior research on chronic pruritus has also found a substantial effect on quality of life (QoL). In several studies, Yosipovitch and colleagues5- 7 found pruritus to significantly influence QoL proxies including mood, concentration, eating habits, sexual function, and sleep. Wolkenstein et al8 found that chronic pruritus impaired QoL more than other skin problems. Kini et al9 found no significant difference between the impact of chronic pain and chronic itch on QoL.
Whereas the impact of chronic pruritus on QoL has been established, a second-order question—what factors mediate this impact—remains lightly studied. A negative correlation between pruritus severity and QoL has been documented,2,10- 12 and the importance of several other factors has been addressed with small studies. Desai et al13 found that pruritus etiology and sex influenced the impact of pruritus on QoL: urticaria engendered greater impact on functional aspects of QoL, and women were more negatively affected by chronic pruritus than men. Stander et al14 also found a greater impact of chronic pruritus on women. Kini et al9 found that unmarried persons were more adversely affected by chronic pruritus than their married counterparts. A study of patients with psoriasis found no relationship between age, sex, disease duration, sleep quality, and the impact of pruritus on QoL.12 Although these studies explored the relevance of some variables, the role of many patient and pruritus characteristics in determining the impact of pruritus on QoL remains a mystery. To address this knowledge void, we undertook a cross-sectional study of a large sample of US military veterans. We hypothesized that a multitude of factors would significantly affect how chronic pruritus influences QoL.
The Emory University institutional review board approved this cross-sectional study. This study consisted of 2 phases: (1) creating a panel of comprehensive surveys to capture information on patients and their pruritus and (2) administering the survey to a sample of patients from the Veterans Hospital Patient Database.
We designed a panel of surveys to capture patient demographic, medical history, therapeutic, and personality data. Questions from published surveys were assimilated, as follows:
ItchyQoL – ItchyQoL is a validated, pruritus-specific instrument that measures the degree to which pruritus affects QoL.13 This impact is quantified into 3 subscores (symptom, function, emotion) and an overall score. A higher score corresponds to a more adverse impact on QoL. We used the frequency version of this instrument.
Questionnaire for the Assessment of Pruritus – This questionnaire characterizes a patient’s pruritus.6
NEO 5-factor inventory – Developed by psychologists at the National Institutes of Health, this survey is predicated on the 5-factor personality model, which divides personality into 5 domains: openness, conscientiousness, extroversion, agreeableness, and neuroticism. Survey answers are used to score patients on these 5 domains; a higher score corresponds to a higher degree of that personality trait.15
12-Item Short Form Health Survey – This validated questionnaire measures generic health-related QoL. We do not report results of this questionnaire here.
Questions regarding risk factors for pruritus were developed through a literature review. Economic and family impact questionnaires were developed by the adaptation of existing validated measures. Phase 1 of the study involved testing the panel of surveys with veterans for content and face validity, clarity, reliability, and feasibility. Cognitive interviewing and psychometric analyses were performed. The results of this testing have been reported elsewhere.16 The final comprehensive survey consisted of more than 300 questions.
Patients for phase 2 of this study were drawn from the Veterans Hospital Patient Database. The database was stratified using the US Census Bureau’s 6 geographic areas to ensure diverse geographic representation. From this stratified database, 6000 veterans were sampled at random. Interviewers at the University Center for Social and Urban Research at the University of Pittsburgh attempted to call these 6000 veterans at least 3 times each. Exclusionary criteria were inability to speak English or inability to answer questions due to disability. Ultimately, 1075 veterans agreed to participate; all participants provided verbal informed consent. At the time of the telephone call, 405 of the veterans had chronic pruritus. The entire survey was administered to these 405 veterans; the remaining 670 veterans answered fewer questions because the itch-related questions were not applicable. Because our focus is the impact of chronic pruritus on QoL, our analysis uses only the replies from the 405 veterans with chronic pruritus.
Survey responses were translated into both categorical and continuous variables.
Many survey questions featured multiple discrete answer choices (more than 12 choices in some cases); we transformed these questions into categorical variables. To reduce the number of categories, answers were combined in a logical fashion. For example, hand, forearm, and upper arm were separate answer choices for a question asking the anatomic location of pruritus. For analysis, these 3 answers were combined to form a single upper extremity category. Answers to some survey questions that were initially continuous (eg, income, body mass index [BMI, calculated as weight in kilograms divided by height in meters squared], frequency of itch, duration of pruritic condition) were translated into categorical variables when the responses did not show a normal distribution. The categorical variables used in our analysis are listed in Table 1.
To derive the “possible etiology” variable, patients were asked a series of questions regarding whether they had any of a number of cutaneous and systemic diseases that can cause pruritus. On the basis of their answers to these questions, patients were then grouped into the cutaneous only, systemic only, cutaneous and systemic, and idiopathic categories. However, we cannot say for certain that a particular disease affirmed by a patient is definitively causing the patient’s pruritus, hence the modifier “possible” prepended to “etiology.”
Continuous variables are pruritus severity, personality domain scores, and age. Patients scored pruritus severity on a scale of 1 to 10. The NEO 5-factor inventory consists of 60 items. The interviewer read each item to a patient (eg, “I really enjoy talking to people”), who then responded on a scale of 1 (strongly disagree) to 5 (strongly agree). Personality scores were calculated using the NEO inventory’s prescribed addition and subtraction of various responses. These personality scores are interval variables. The means and standard deviations of the 5 personality domains depend on the population to which the inventory is applied, so no inferences can be gleaned from the absolute level of these statistics.
The frequency version of the ItchyQoL instrument consists of 22 items. These items are divided into 3 constructs: symptomatic impact (6 items), functional limitations (7 items), and emotional impact (9 items). The interviewer read each item to a patient (eg, “my itchy skin often makes it difficult to concentrate”), who then responded on a scale of 1 (never) to 5 (always). Subscores for each construct are summations of the numerical responses. The ItchyQoL total score for a patient is the sum of the symptom, function, and emotion subscores. Some veterans responded to ItchyQoL items with “I don’t know.” For a given subscore, if a patient responded “I don’t know” to 1 or 2 items, these responses were replaced with the mean numerical response of that subscore’s other items. In total, 42 items were answered “I don’t know,” necessitating adjustment to 40 subscores (3% of the total subscores). One of the 405 patients answered 3 emotion items with “I don’t know;” his data point was removed entirely for the analyses involving the ItchyQoL total score and emotion subscore.
As a first step, we undertook univariate analysis to screen for relationships between the total ItchyQoL score and other variables. Relationships between continuous variables and the total ItchyQoL score were determined using analysis of variance and t testing. Relationships between categorical variables and ItchyQoL scores were determined using χ2 testing. A priori, we had decided to use all variables from the univariate analysis with P values ≤.20 as predictor variables in multivariate analysis. All of the aforementioned variables met this threshold.
As the second step, we undertook multivariate analysis using a generalized linear model with ItchyQoL scores (total and subscores) as the response variables for a total of 4 different models. By convention, we deemed P ≤ .05 to be significant.
While undertaking the survey, occasionally a veteran replied “I don’t know” or declined to answer a survey question. We have outlined how we treated such responses to ItchyQoL items. When veterans responded “I don’t know” or declined to answer other questions, predictor variables derived from those questions could not be calculated and/or categorized. Ergo, such a patient was excluded from univariate and multivariate analyses involving that variable. Fortunately, the “I don’t know” and declined responses were infrequent, with most involving the question on income. The sample size for multivariate analysis only decreased from 404 to 384 individuals. All analyses were performed with SAS software, version 9.3 (SAS Institute).
Table 1 and Table 2 summarize the frequencies of the categorical variables and the means and standard deviations of the continuous variables, respectively. Because the means of the personality scores are not informative (as explained in the Methods section), these variables are omitted from Table 2. Our pruritic veteran population consisted primarily of white men; the mean age was 60.7 years. The majority were married, had at least some college education, and had both cutaneous and systemic conditions that could engender pruritus. Most veterans reported pruritus affecting multiple locations and lasting more than 2 years. Region of residence, income, and employment status were fairly evenly distributed.
The multivariate model exhibited statistical significance (P < .001) and good explanatory power (R2 = 52%). Table 3 summarizes the results of the multivariate analysis, with the continuous and categorical variables at the top and bottom of the table, respectively.
Itch severity significantly (P < .001) directly correlates with the total ItchyQoL score and its 3 constitutive constructs. Itch frequency, duration, and anatomic location are all significantly related to the impact of pruritus on QoL. Pairwise comparison generally supports that a higher itch frequency significantly (P < .01) correlates with a greater impact on QoL. Likewise, pairwise comparison shows that pruritus of longer duration exerts a greater impact on QoL. Time of day when itch is worst is not significantly related to ItchyQoL scores.
Marital status, age, race, and possible etiology are all significantly related to the impact of pruritus on QoL. Marriage is associated with a lower ItchyQoL total score and symptom subscore. Age exhibits significant, negative correlation with the total ItchyQoL score and all 3 subscores. With respect to possible etiology, the ordinal ranking of the categories from greatest to least impact on QoL is as follows: (1) cutaneous and systemic reason to itch, (2) systemic reason, (3) cutaneous reason, (4) neither. However, pairwise comparisons of these categories revealed no significant differences. Race is significant, with nonwhites experiencing a significantly greater impact on QoL than whites. Pairwise comparisons failed to find a significant difference between the race categories.
Multivariate analysis revealed a significant negative coefficient for patients’ extroversion score with respect to ItchyQoL total and function scores. Conversely, analysis found a significant positive coefficient for patients’ neuroticism score with respect to ItchyQoL total, symptom, and emotion scores. Three personality domains failed to show significance: openness, conscientiousness, and agreeableness.
Sex, region of residence, education level, employment status, income, and BMI did not exhibit significant association with ItchyQoL scores.
Medical research has established that chronic disease substantially affects QoL.17 Similarly, dermatologic investigation has found a deleterious effect of chronic pruritus on QoL.8,9 For some chronic diseases (eg, diabetes mellitus, congestive heart failure, end-stage renal disease), investigation has revealed the factors that mediate the impact of the disease on patients’ QoL. For chronic pruritus, the identity of these factors is largely unknown. To remedy this knowledge void, we collected extensive demographic, medical, and QoL data from a large, diverse sample of pruritic patients. Multivariate analysis revealed that an array of variables influences the impact of chronic pruritus on QoL. Important insights can be gleaned from the results.
Not surprisingly, itch severity was found to significantly influence the ItchyQoL total score and all 3 subscores. The correlation is positive, with greater pruritus severity yielding a greater impact on QoL. At the same time, itch frequency significantly correlates with total ItchyQoL. Taken in isolation, the significance of itch frequency is not surprising. However, it is interesting that the impact of itch frequency on QoL is not fully captured by the itch severity measure. One could reasonably expect that itch frequency is a prominent determinant of itch severity and as such would not be independently significant. However, our results suggest otherwise.
Several patterns for the influence of chronic disease duration on QoL have been hypothesized: no effect, adaptation, and chronic exasperation.18 Studies of congestive heart failure, chronic prostatitis, and diabetes mellitus have shown that disease duration has no effect on QoL.19- 21 Chronic pruritus does not behave in the same fashion, as our analysis provides evidence for chronic exasperation: increasing duration significantly correlates with greater negative QoL impact. The significance of all 3 ItchyQoL subscores indicates that symptoms, functional status, and emotional health deteriorate as chronic pruritus persists.
The impact of chronic pain on QoL depends in part on its anatomic location.22 We included itch location in our analyses to investigate for a similar relationship, and location did prove significant. Interestingly, pruritus afflicting the upper extremities has the most adverse impact on QoL, although pairwise comparison revealed no significant differences vs other anatomic locations. In the multivariate analyses that used the 3 ItchyQoL subscores as dependent variables, anatomic location has the lowest P value (P < .001) when modeled with the function subscore. This finding suggests that the higher degree of impact experienced by patients with upper extremity pruritus may be due to functional impairment.
Studies of the impact of age on QoL in chronic disease have produced dissimilar results. Some researchers found increasing age to diminish the negative impact on QoL20; some found no relationship between age and QoL19,23; others found that increasing age correlates with declining QoL.21 In our analysis, age has a significant, negative coefficient when modeled against the total ItchyQoL score and the 3 subscores. The negative coefficient indicates that older age correlates with less adverse impact on the QoL, and the pansignificance of all 3 ItchyQoL constructs demonstrates that the mitigating influence of age pervades the spectrum of morbidity. The statistical significance of age lies in contrast to the findings of a study of QoL in pruritic patients with psoriasis.12
Research on the relationship between race and QoL in chronic disease has yielded heterogeneous results. Some research has found no correlation between race and QoL measures.20,24 End-stage renal disease has been shown to more negatively affect the QoL of whites25 and nonwhites.26 Kington et al27 found that greater functional impairment of nonwhites with chronic disease was attributable to socioeconomic status and not race. Our analysis has found a significant association of race with the impact of chronic pruritus on QoL, even after we controlled for socioeconomic variables. The total ItchyQoL scores for both the African American and other race categories are approximately equal and 9% higher than the score for whites, although no comparisonwise error rate was significant. With respect to ItchyQoL subscores, race was significant (P = .02) only when modeled with the function subscore, suggesting that race’s influence on QoL is mediated through functional impairment.
Marital status significantly influences the impact of pruritus on QoL, with chronic pruritus having less effect on married persons. The mean ItchyQoL score for an unmarried person is 7% worse. Interestingly, the ItchyQoL symptom subscore was the only significant construct, intimating that marriage ameliorates the impact of pruritus more through symptom relief than through the function and emotion constructs. The beneficial impact of marriage on QoL is consistent with prior QoL research21,28 and the aforementioned finding of Kini et al.9 We hypothesize that the social support available in marriage is responsible for this finding.
The presence of a disease that can cause pruritus (ie, the possible etiology variable) exhibits significant correlation with the impact of pruritus on QoL. This finding suggests that the impact of pruritus on QoL is determined not just by the characteristics of the pruritus itself; the etiology matters. However, we could not through a telephone survey determine with sufficient certainty whether the medical condition(s) affirmed by the patient are actually causing the patient’s pruritus. As such, although the data support a relationship between possible etiology and QoL, the evidence is not as conclusive as for other variables.
To capture the influence of personality on QoL, we incorporated personality domains into our analysis. Extroversion and neuroticism proved statistically significant. This finding is consistent with the psychology literature, which has shown that extroversion and neuroticism are important factors in determining well-being.29,30 Psychologists define neuroticism with the modifiers anxious, self-pitying, tense, touchy, unstable, and worrying.31 An association of neuroticism with lower QoL has been shown in patients with inflammatory bowel disease,32 Parkinson disease,33 chronic headache,34 cancer,35 breast cancer specifically,36 and chronic disease in older patients.37 In our analysis, neuroticism featured a positive coefficient with respect to total ItchyQoL; both the symptom and emotion constructs were significant. Ergo, we conclude that chronic pruritus has a more detrimental QoL impact in more neurotic patients, manifesting as an exacerbation of symptoms and decline in emotional well-being.
To define extroversion, psychologists use the adjectives active, assertive, energetic, enthusiastic, outgoing, and talkative.31 Unlike neuroticism, evidence for the influence of extroversion on QoL in chronic disease is scant. Among patients with Parkinson disease, Dubayova et al33 found extroversion to predict less impact on QoL in men only. Similarly, our analysis found a significant negative association of extroversion with total ItchyQoL and the function construct. These findings indicate that a greater degree of extroversion reduces the impact of pruritus on QoL, specifically on functional capacity. Note that this stands in contrast to neuroticism, in which a greater degree of neuroticism increases the symptomatic and emotional impact.
A number of variables did not exhibit significance: sex, region of residence, education level, employment status, income, BMI, and time of day when itch is worst. Several interesting observations can be made. First, socioeconomic status—as represented by education level, employment status, and income—does not mediate the impact of chronic pruritus on QoL. Second, pruritus that is worst after 8 pm (ie, pruritus that affects sleep) does not produce a more negative impact on QoL. Not only is the worst-itch–time variable not significant, but the after–8 pm category exhibited the second lowest impact on QoL of the 7 time groups (although pairwise comparisons revealed no significant differences). This finding challenges the widely held belief that pruritus (at least generalized pruritus) is worst at night.38- 40 Third, sex was not found significant. This discovery lies in contrast to previous studies that found chronic pruritus to more severely affect the QoL of female patients.13,14 Women constitute only 7% of our sample, but the difference between the estimates for men and women was miniscule (P = .98). There is no evidence to suggest that sex would become significant with a multiplicative increase in the sample size.
Our sample features racial, socioeconomic, and regional diversity but is constituted entirely of US military veterans. As such, generalization of results to nonveteran populations merits caution. Furthermore, caution must also be exercised when extrapolating our findings to younger populations because only 8% of our sample was 40 years or younger. Conclusions drawn from our sample may not be applicable to other populations.
Theoretically, one could argue that the problem of causality complicates the interpretation of the personality results. Our analysis presumes that personality affects a patient’s response to chronic pruritus, but the possibility exists that chronic pruritus may influence personality. This chicken-and-egg issue (ie, how personality and events influence each other) has been studied extensively in the psychology literature, where it is generally accepted that adult personality is stable.41,42 As such, the presence of chronic pruritus would not be expected to alter a person’s neuroticism and extroversion quotients, and the use of personality factors as independent variables would not be complicated by bidirectional causality.
This study represents the next epidemiologic step in chronic pruritus research: understanding the factors that mediate the impact of chronic pruritus on QoL. Using data collected from a large, geographically diverse sample of the US veteran population, we found that multiple patient and pruritus characteristics significantly influence this relationship: demographic characteristics (age, race, marital status), personality (extroversion, neuroticism), pruritus parameters (severity, duration, frequency, anatomic location), and etiology (cutaneous, systemic). The significance of some of these variables (marital status, neuroticism, anatomic location) mirrors what has been found in other chronic diseases. Notable factors not found significant include sex and socioeconomic variables (education level, employment status, income). The multitude of significant factors demonstrates the complex relationship between chronic pruritus and QoL. We hope that better appreciation of these factors will improve the clinical evaluation and treatment of chronic pruritus. Future directions of research include investigating why race influences the impact of chronic pruritus on QoL and exploring whether support groups or personality-specific interventions may help mitigate the QoL impact of chronic pruritus.
Accepted for Publication: August 22, 2013.
Corresponding Author: Suephy C. Chen, MD, MS, Department of Dermatology, Emory University, 1525 Clifton Rd, First Floor, Atlanta, GA 30305 (firstname.lastname@example.org).
Published Online: March 26, 2014. doi:10.1001/jamadermatol.2013.7696.
Author Contributions: Drs Chen and Veledar had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Chen.
Acquisition of data: Chen.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Carr.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Veledar.
Obtained funding: Chen.
Administrative, technical, and material support: Carr.
Study supervision: Chen.
Conflict of Interest Disclosures: Dr Chen has received royalties from Eli Lilly (2007) and Pierre Fabre (2011) for licensing of the ItchyQoL instrument and has been an outcomes research consultant for Biogen and Astellas and a principal investigator for studies sponsored by Hoffman-LaRoche and Galderma. No other disclosures are reported.
Funding/Support: The Veterans Health Administration (VHA) provided (1) financial support and (2) access to the Veterans Hospital Patient Database for sampling purposes.
Role of the Sponsors: The VHA had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank the VHA for its support of this research.