Determinants of low health-related quality of life in patients with myelodysplastic syndromes: EUMDS Registry study

Abstract

Patients with myelodysplastic syndromes (MDS) frequently experience a significant symptom burden, which reduces health-related quality of life (HRQoL). We aimed to identify determinants of low HRQoL in patients recently diagnosed with MDS, for guiding early intervention strategies. We evaluated longitudinal data in 2205 patients with MDS during their first year after diagnosis. Median values of EQ-5D 3-level (EQ-5D-3L) index (0.78) and visual analog scale (VAS) score (0.70) were used as thresholds for low HRQoL. In addition, the 5 dimensions of EQ-5D-3L were analyzed for impairments (any level vs "no problem" category). After multiple imputation of missing values, we used generalized estimating equations (GEE) to estimate odds ratios (OR) for univariable determinant screening (P < .15), and to subsequently derive multivariable models for low HRQoL with 95% confidence intervals (CI). Multivariable GEE analysis showed the following independent determinants (OR, 95% CI) for low EQ-5D index: increased age (60-75 years: 1.33, 1.01-1.75; >75: 1.84, 1.39-2.45), female sex (1.70, 1.43-2.03), high serum ferritin level (≥1000 vs ≤300 μg/L: 1.41, 1.06-1.87), comorbidity burden (per unit: 1.11, 1.02-1.20), and reduced Karnofsky performance status (KPS, per 10 units: 0.62, 0.58-0.67). For low VAS score, additional determinants were transfusion dependence (1.53, 1.03-2.29), low hemoglobin <10 g/dL (1.34, 1.12-1.61), and high body mass index (≥30 vs 23-29.9 kg/m2: 1.26, 1.02-1.57). Sex, KPS, comorbidity burden, hemoglobin count, and transfusion burden were determinants for all EQ-5D dimensions. Low HRQoL is determined by multiple factors, which should be considered in the management and shared decision making of patients with MDS. This trial was registered at www.clinicaltrials.gov as #NCT00600860.

Graphical abstract

Figure 1.

Determinants of low HRQoL – univariable analyses. Presented are the univariable analyses of the determinants of low EQ-5D index value and of low EQ-5D VAS score. ORs are colored if P value < .15 and gray if P value > .15. ∗At least 1 unit RBC transfusion for a surveillance time of 8 weeks before the HRQoL assessment; ∗∗Average number of RBC transfusions per month since the last visit or since the MDS diagnosis; ∗∗∗Progression to higher IPSS-R risk group. HCT-CI, hematopoietic cell transplantation-specific comorbidity index; IPSS-R (vs VL/L), all remaining IPSS-R groups vs very low/low risk.

Figure 2.

Determinants of impaired EQ-5D-3L dimensions - univariable analyses. Presented are the univariable analyses of the determinants of impaired dimensions of the EQ-5D questionnaire. ORs are colored if P value < .15 and gray if P value > .15. ∗At least 1 unit RBC transfusion for a surveillance time of 8 weeks before the HRQoL assessment; ∗∗Average number of RBC transfusions per month since the last visit or since the MDS diagnosis; ∗∗∗Progression to higher IPSS-R risk group. HCT-CI, hematopoietic cell transplantation-specific comorbidity index; IPSS-R (vs VL/L), all remaining IPSS-R groups vs very low/low risk.