Psychometric Evaluation of the Cushing’s Quality-of-Life Questionnaire

Abstract

Cushing's disease (CD) is a rare disorder of chronic hypercortisolism due to an adrenocorticotropic hormone (ACTH)-secreting pituitary corticotroph adenoma. Because hypercortisolism symptoms are wide ranging, it is important to assess a variety of outcomes including both clinical factors, such as cortisol levels, and health-related quality of life (HR-QOL), to better understand the severity and impact of CD on patients and the potential efficacy of CD treatment. Pasireotide, a somatostatin analog that targets somatostatin receptors on the pituitary adenoma, is under development as a treatment for CD. A phase III clinical trial was conducted to investigate its safety and efficacy in patients with CD. In this trial, HR-QOL was assessed with the Cushing's Quality-of-Life (CushingQOL) questionnaire, specifically developed and validated in patients with Cushing's syndrome. Reliability, validity, the ability to detect change, and a minimal important difference (MID) were evaluated for the CushingQOL questionnaire using data from patients diagnosed with CD who participated in the phase III clinical trial designed to assess the safety and efficacy of different doses of pasireotide. Adult patients (n=162) with CD participated in a randomized, double-blind, multinational, phase III clinical trial. Patients received subcutaneous pasireotide (600μg or 900μg) twice daily for 3months (double blind). After 3months, some patients were unblinded based on their mean urinary free cortisol (mUFC) levels and were given the chance to increase their dosage, while the other patients remained blinded. At month 6, an open-label 6-month period began. The CushingQOL questionnaire was self-administered four times (baseline [n=160], and at months 3 [n=134], 6 [n=113], and 12 [n=76]). A confirmatory factor analysis (CFA) was conducted. Reliability estimates were calculated for internal consistency (coefficient alpha) and test retest (intraclass correlation coefficients [ICCs]) for patients with stable hypercortisolism at month 3 and month 6. Construct validity hypotheses (correlations), mean differences in known groups (ANOVAs), and responsiveness effect sizes (Guyatt's) were estimated based on measures of cortisol, body mass index (BMI), waist circumference, weight, facial rubor (redness), striae (stretch marks), bruising, supraclavicular fat pad, dorsal fat pad, and results of the Beck Depression Inventory II (BDI-II). The half-standard deviation distribution method was used to estimate MID. CFA loadings supported a one-factor solution for the CushingQOL questionnaire items. Internal consistency reliability (0.87-0.88) and ICCs (0.87) were high. Construct validity hypotheses were in the anticipated direction. Changes in CushingQOL scores were moderately correlated with changes in mUFC levels, in BMI, and in weight. Mean scores for minimally depressed patients were significantly higher (indicating better HR-QOL) than for severely depressed patients. Moderate Guyatt's responsiveness effect sizes were observed for patients who achieved reductions in weight, BMI, and waist circumference. Using the half-standard deviation method, an estimate of the MID was computed as 10.1. This study provided evidence within the context of a longitudinal design that the CushingQOL questionnaire is a reliable, valid, and responsive instrument for the assessment of HR-QOL in adults with CD in accordance with recommendations set forth by regulatory agencies in the USA and Europe.