Validation of the labeled magnitude scale for the assessment of itch intensity in patients with chronic pruritus

Introduction: The visual analogue scale (VAS), the numerical rating scale (NRS), and the verbal rating scale (VRS) are routinely used to assess pruritus intensity. However, these scales have limitations, especially the ceiling effect of linear scales (VAS, NRS) and the reduced sensitivity to change of the VRS. In the labeled magnitude scale (LMS) consisting of a vertical line with verbal anchors distributed quasi-logarithmically, the ceiling effect is reduced, allowing a higher discriminative assessment at the higher end of the scale. Methods: A total of 172 patients completed the LMS twice within 30–60 minutes to assess the reliability of the scale, as well as the NRS and VAS for analyses of convergent validity. Sensitivity to change of the LMS was investigated by analyzing changes in scores after a twice-daily application of an emollient containing menthoxypropanediol for 14 days. Results: Test-retest reliability was excellent for the LMS assessing average (Cronbach α: 0.955) and worst (Cronbach α: 0.945) pruritus intensity in the last 24 hours, while strong to very strong correlations were observed between the LMS and NRS assessing the worst (r=0.783) and average (r=0.808) pruritus intensity, respectively. Treatment with an emollient lead to a significant decrease in LMS scores (P<0.01). In patients with severe pruritus at baseline (VRS ≥3), we recorded a 30.2% improvement of average pruritus intensity using the LMS compared with a 25.0% improvement using the NRS and a 27.8% improvement of worst pruritus intensity using the LMS compared with an 11.1% improvement with the NRS. Most patients considered the LMS an appropriate instrument to assess pruritus intensity (89.6%) and would use it again (91.9%). Discussion: The LMS is an appropriate well-accepted instrument to assess CP. It is especially useful to detect variations in pruritus intensity in patients with severe CP.

Chronic pruritus (CP), defined as pruritus lasting for 6 weeks or longer, is a subjective multidimensional symptom, which is complex to measure [1] . In routine dermatological care and clinical trials the use of tools measuring the intensity of pruritus has been prioritized [2,3] . Linear instruments, such as the numerical rating scale (NRS; range: 0-10) and the visual analogue scale (VAS; range 0-10), and the ordinal verbal rating scale (VRS; range 0-4) have been validated to assess CP and are widely used [4,5] . However, these scales show important shortcomings. One limitation of the linear scales (NRS, VAS) is the ceiling effect, in which distinct intensities at the higher end of the scale cannot be adequately discriminated [6] , resulting in an inaccurate assessment of severe and very severe pruritus. As for the VRS, the small number of categories of this scale limits its sensitivity to change [7] , especially when comparing to the 11-point NRS [8] .
To address these issues, we adapted the labeled magnitude scale (LMS) for the assessment of CP (Fig. 1). LMS is a tool developed for assessing sensory symptoms consisting of a 10 cm long, vertical line with verbal anchors distributed in a quasilogarithmic manner [9] . It contains thus elements of the VRS (the presence of verbal anchors) while allowing patients to mark the perceived pruritus intensity on a continuous line, as in the VAS. Importantly, due to its architecture, the ceiling effect is reduced in the LMS compared with the NRS or VAS. The LMS has been validated for the assessment of acute and chronic pain [10] . In addition it has been used in acute pruritus [11] ; however, the validation for the measurement of CP is still lacking.
Aim of this study was to validate the LMS in German for the assessment of CP in regard to its test-retest reliability and convergent validity, as well as assess the sensitivity to change of this novel instrument in CP.
Münster (Germany) were consecutively enrolled. Only patients showing dry skin and reporting in the previous 24 hours a maximal itch intensity of 3 or more on the NRS were considered. Exclusion criteria included acute exacerbation of a skin condition, cutaneous infection, pregnancy or breastfeeding, and drug abuse. All study procedures were performed respecting the Helsinki declaration and later revisions. The study was approved by the local Ethics Committee (Medical Faculty of the University of Münster, 2017-415-f-S) and was registered at the German Registry of Clinical Trials (DRKSDRKS00013074).

Study design
This study consisted of 2 parts. In the first part, the feasibility of the use of the LMS in the clinical routine was assessed. Patients with CP (n = 120) filled out the scale before a routine appointment. Generally, the scale was understood and well accepted by the patients (data not shown), so that validation of the scale was sought in the subsequent phase of this study. The second part of the study consisted of 2 visits. In the first visit, CP patients were asked to complete the LMS and the NRS twice in an interval of 30-60 minutes, as well as the other pruritus intensity scales (VAS, VRS) once. Afterward, patients were instructed to use an emollient containing menthoxypropanediol twice daily until the next study visit 2 weeks later. In the second visit, patients completed once more the LMS, NRS, VAS and VRS, and a feasibility questionnaire. Both the average and the worst pruritus intensity of the previous 24 hours were assessed with the LMS. Table 1 summarizes the instruments used in this study for the assessment of pruritus intensity, including recall periods.

Reliability and validation
To assess the test-retest reliability of the LMS, patients were asked to complete this scale twice within an interval of 30-60 minutes (study visit 1). Convergent validity was evaluated by correlating LMS scores to pruritus intensity scores provided by external instruments, namely the NRS and VAS.

Sensitivity to change
In order to assess the sensitivity to change, LMS scores before and after treatment with an emollient containing menthoxypropanediol were compared in patients reporting an improvement of the pruritus. Change in LMS scores was compared with change in the NRS. To check whether using the LMS can diminish the ceiling effect, we analyze the course of the scores of LMS and NRS of those patients who indicated their pruritus intensity as severe or very severe on the VRS before treatment begins.

Feasibility questionnaire
After completing the LMS, patients were asked to fill out a feasibility questionnaire. Here it was questioned whether patients considered the LMS to be clearly formulated, if it was difficult to decide on a score, whether the verbal anchors were adequately distributed and whether the scale was too short. Finally, patient perspective on the suitability of the LMS for the assessment of CP and whether they would use this instrument again was assessed.

Statistics
We used SPSS v. 26.0 for Windows (IBM Corporation, Armonk, NY) for statistical analyses. Descriptive data are shown as number of patients (percentage), median [interquartile range (IQR)] and range. To assess test-retest reliability, intraclass correlation coefficients (ICC) and Cronbach α were used. Correlation coefficients between LMS and the external scales to assess convergent validity were calculated with the Spearman rank test. We performed the Wilcoxon test for analyses of sensitivity to change. The level of significance was set at P ≤ 0.05.

Patients
A total of 172 patients with CP completed the validation study. Demographic data, the clinical classification of patients according to their skin status and the etiology of the pruritus are shown in Table 2.

Test-retest reliability
Test-retest reliability was excellent for the LMS measuring both worst and average pruritus intensity according to the Cronbach α, while ICCs revealed an almost perfect agreement for both scales. Similar reliability scores were observed for the NRS ( Table 3).

Convergent validity
Strong to very strong correlations were observed between the LMS and NRS assessing the average pruritus intensity in the previous 24 hours (r = 0.808, P = 0 < 0.001, n = 169) and between LMS and NRS assessing the worst pruritus intensity in the previous 24 hours (r = 0.783, P = 0 < 0.001, n = 169). Moderate correlations were recorded for the correlations between the LMS and the VAS ( Table 4).

Sensitivity to change
Comparing baseline and follow-up scores of the LMS and NRS a significant reduction of the intensity in both measuring instruments were found (Table 5). However, in patients with severe pruritus at baseline the LMS scores had a stronger reduction than the NRS scores ( Table 6).

Discussion
We validated for the first time the LMS for the use in CP. In previous studies, the LMS was used in acute experimentally induced pruritus and proved to be more reliable than the VAS [11] . In our investigations, the LMS assessing the worst and average chronic pruritus intensity of the previous 24 hours showed a very good convergent validity with strong to very strong correlations with the NRS, while test-retest reliability was excellent for the LMS assessing both average and worst pruritus intensity in the last 24 hours. Moreover, this scale showed sensitivity to change, as improvement of itch after treatment with an emollient containing menthoxypropanediol resulted in significant lower LMS    scores. Thus, the LMS is a suitable instrument to assess pruritus intensity in CP conditions, augmenting the spectrum of indications of the LMS. The innovative aspect of the LMS is its ability to reduce the ceiling effect observed in linear scales (NRS, VAS), due to its quasi-logarithmic distribution of the verbal anchors [6,10] . Indeed, in patients with severe or very severe pruritus (ie, VRS ≥ 3/4) at baseline, we recorded a 30.2% improvement of average pruritus intensity using the LMS compared with an improvement of 25.0% using the NRS, and a 27.8% improvement of worst pruritus intensity using the LMS compared with an improvement of 11.1% using the NRS. However, this difference was not observed for the whole study population who suffered from mild to very severe pruritus intensity according to the VRS. As such, the LMS is particularly helpful to assess pruritus intensity over time in patients with severe pruritus, as it can detect variations of pruritus intensity at the higher end of the scale, which are not depicted by the NRS or VAS. These findings are in agreement with previous studies showing higher sensitivity of the LMS for the assessment of high intensities of pain [10] and taste perception [12] . Moreover, owing to the construct of the LMS, the median scores recorded in the LMS were lower than in the NRS (average pruritus: 3. The LMS assessing average pruritus showed a slightly better test-retest reliability compared with the LMS assessing worst pruritus. This may be explained by the longer intervals of the verbal anchors at the higher end of the scale. The average pruritus intensity is in most instances lower than the worst pruritus intensity and the higher concentration of verbal anchors at the lower end of the scale provides more orientation to patients, leading to better reliability scores.
The excellent reliability of the LMS constitutes an advantage when comparing to the VAS, which performs less well in this regard [11,[13][14][15] . The lack of visual anchors in the VAS, which difficult the choosing of a precise score, may explain the worse reliability of this scale compared with the LMS [16] . In our study, the reliability of the NRS was superior to that of the LMS. A better training of CP patients in the NRS may explain this finding. CP patients often associate their pruritus intensity with a number from 0 to 10, since this scale is regularly used in clinical practice either in routine questionnaires or verbally during the consultations. Furthermore, the numerical anchors of the NRS may suffice in guiding the patient to choose a precise score.
As for the convergent validity, we recorded stronger correlations between the LMS and the NRS (average pruritus: r = 0.808, worst pruritus: r = 0.783) compared with the correlations between the LMS and VAS (average pruritus: r = 0.570, worst pruritus: r = 0.535). Previous studies have shown strong correlations between the NRS and VAS [4,14] . In our study, different recall periods used for the NRS (24 h as in the LMS) and the VAS (4 wk) may explain the observed discrepancies.
Patients with CP considered the LMS an appropriate tool and easy to use, with only a minority of patients manifesting difficulties in deciding on a score. A feasibility analysis from the perspective of attending physicians would also be desirable and should be considered in future studies.
The use of verbal anchors in the LMS constitutes one possible limitation. As the LMS aims to depict the continuum of a sensory sensation, it is not desirable that patients limit their response to marking one of the visual anchors, a phenomenon termed clustering [17] . In our study, the majority of patients (75%) marked their scores in-between verbal anchors. This suggests that the visual anchors were generally used as an orientation, and that clustering did not occur in most cases.
In conclusion, the LMS is a suitable tool to assess CP with high acceptance from patients. Physicians should consider resorting to the LMS to assess possible variations in pruritus intensity in patients with severe CP.  To assess the sensitivity to change, the labeled magnitude scale and numerical rating scale scores before and after treatment with an emollient containing menthoxypropanediol were compared via Wilcoxon Test.  The labeled magnitude scale and numerical rating scale scores from patients with severe or very severe pruritus according to the verbal rating scale at baseline were compared via Wilcoxon Test. IQR indicates interquartile range.

Sources of funding
This study was supported by Beiersdorf AG.