Between August 2015 and October 2017, questionnaires were sent to the 799 patients and 849 controls from the EPICAP cohort for whom a postal address was available. Among these 1648 subjects, 6 had died and 106 were non-eligible for the EPICAP-QALY study. Responses to questionnaires were received from 376/704 eligible patients (53.4%) and 364/832 eligible controls (43.8%) (Figure 1). Patient profiles did not significantly differ between participants and non-participants to the study according to age, Gleason score or BMI (Supplementary Table). The controls were age-matched in a 2:1 ratio with the patients (n=188). The average patient age was 67.9 years old. Baseline characteristics at inclusion were similar between groups, except for PSA level, which was much lower in the patient group as anticipated due to treatment (Table 1). The most common treatments for patients were radical prostatectomy (RP) (68.6%) and EBRT (23.9%).
Primary outcome: QLQ-C30 scores were high and did not differ between groups; median summary scores were respectively 94.87 [87.44; 98.72] and 94.15 [89.66; 98.21] for patients and controls, p=1 (Table 2). No significant difference in the QLQC30 was highlighted in univariate analysis.
Estimation of confounding factors in the multivariate regression model could only be performed for QLQ-C30 summary score, global health status/QOL, emotional functioning, cognitive functioning, social functioning, fatigue, pain and insomnia, for which enough data were collected. Multivariate analysis of QLQ-C30 summary score was conducted on 540 participants using the following variables: age, group (patient vs. control), type of lodging, and presence of a chronic illness. No significant difference of QLQ-C30 summary score was observed between groups (p=1). For global health status/QOL, emotional functioning, cognitive functioning, fatigue and insomnia, multivariate analysis confirmed the absence of difference shown in univariate analysis with a reduced level of statistical significance using Holm’s correction, with respectively p=1, p=1, p=0.91, p=1, p=1, p=0.29, p=0.23 for global health status/QOL, emotional functioning, cognitive functioning, fatigue and insomnia, social functioning and pain.
Secondary outcomes: The univariate analysis showed no difference between medical care for QLQ-C30 scores between treatments.
Responses to the IPSS questionnaire showed no significant difference between patients and controls for urinary symptoms: the median score was 4 [2-8] for patients vs 3 [1-7] for controls, with the majority of subjects in each group classified as suffering from mild urinary symptoms (72.4% patients vs. 78.8% controls; p=1) (Table 2). The results remain non-significant when adjusted for age (p=1).
The ICIQ-MLUTS questionnaire showed no difference in voiding score (VS) between patients and controls (median of 2 [0; 5] and 2 [1; 5] respectively; p=1), but a significantly higher incontinence score (IS) in patients (median score 3 [1; 6] vs. 2 [0; 3] respectively, p=0.0025). Age adjustment confirmed these results. Frequency of diurnal and nocturnal urination did not differ between groups (p=1 for both). The majority of subjects in both groups reported frequency of urination between 1-6 times per day (58.4% for patients vs. 66.3% controls), with only 12.3% patients and 10.7% controls reporting a frequency ≥9 times per day. Similarly, for night frequency, patients and controls mostly reported 1 voiding per night (51.9% vs. 50.5% respectively). Only 4 questions about bladder weakness (questions 8, 9, 10 and 12) showed differences between patients and controls in terms of level of bother experienced.
In contrast, the IIEF-6 questionnaire revealed a significant difference (p=0.0025) in erectile dysfunction between groups with 65.5% of patients having severe erectile dysfunction compared to 32% of controls. Only 11.3% patients were free of dysfunction versus 34.3% of controls. Adjustment for age did not affect the significance of the results.
The HADS questionnaire showed no difference in the number of participants with probable anxiety and/or depression between groups. Median HADS total score was 8 [5; 13] for patients versus 9 [6; 13] for controls, thus 80.8% patients and 85.6% controls showed absence of anxio-depressive symptoms (p=1).
QLQC30, IPSS, IIEF-6, HADS and ICIQ-MLUTS scores were analyzed in 370 out of 376 patients (medical care of 5 patients was missing, and 1 patient treated with hormonotherapy alone was not eligible) according to the different treatments: 10.5% (n=39) in AS, 50.5% (n=187) with a RP alone, 14.6% (n=54) with EBRT, brachytherapy, HIFU or ADT and 24.3% (n=90) with CC.
No significant difference was highlighted for QLQC30 scores (Table 3). The IPSS score, frequency of day or night urination and HADS scores did not differ between the different medical cares. For the ICIQ-MLUTS questionnaire, VS score and IS score were both significantly different (p=0.00253 and p=0.0025, respectively) between treatments and no potential confounder was highlighted. The Bonferonni-post-hoc-analysis (threshold p<0.0083) showed significant differences between RP and EBRT or brachytherapy or HIFU or ADT for VS score (p=0.0009) and IS score (p≤0.0001), whereby VS and IS scores were lowest for patients treated with RP and RT, respectively. IS score was significantly different between AS and RP (p=0.0013).
Erectile dysfunction differed between groups, with severe erectile dysfunction for 33.3%, 80.8%, 64.8% and 72.2% respectively for AS, RP, EBRT or brachytherapy or HIFU or ADT and CC. The analysis with the 5 classes of erectile dysfunction could not be tested, but when grouping into 3 classes (No dysfunction; Mild, Mild to moderate or Moderate dysfunction; Severe dysfunction), the Bonferroni-Holm corrected p-value confirmed a difference (p=0.0025). Multivariate analysis was not performed due to insufficient patients with no dysfunction.