The ProCaSP Study

Quality of Life Outcomes of Prostate Cancer Patients After Radiotherapy or Radical Prostatectomy in a Cohort Study

Nora Eisemann; Sandra Nolte; Maike Schnoor; Alexander Katalinic; Volker Rohde; Annika Waldmann


BMC Urol. 2015;15(28) 

In This Article


Patients and Tumour Characteristics

516 of the initial 529 patients had complete or partially complete HRQOL data and were included. Approximately one of five observations for each HRQOL outcome (23.9%) and one of the ten baseline covariate values (13.3%) were missing for every patient. 256 patients received nnsRP, 127 nsRP, and 133 patients received RT, of which 44 were treated by brachyRT, 52 by externRT, and 37 by combRT. More than half of those patients treated with RT, where information on pre-baseline androgen deprivation therapy was available, additionally received androgen deprivation therapy (ADT) (52%).

Compared to patients receiving other treatment options, patients who underwent nsRP surgery were younger, were more often employed, more often living in a rural area, and had a better baseline sexual functioning (Table 1). Tumours that could not be treated nerve-sparingly during surgery had more often an advanced stage. RT patients were on average older than patients in the other treatment groups, more often not employed, living in urban areas, had a higher PSA level, more often a small tumour stage, and a lower sexual functioning. Comparison of D'Amico risk stratification revealed a much lower risk for the RT treatment groups. More than 80% of the nnsRP and nsRP patients were considered at high risk compared to less than 40% of the RT patients (data after multiple imputation, not shown). Pre-treatment PSA levels were highest in the combRT group and lowest in the nsRP group, while Gleason levels were highest in the nnsRP group and lowest in the brachyRT group.

Sexual Functioning - IIEF

Figure 1 shows that the average sexual functioning of the RT patients remains on a similar low level over the whole observation period. Patients receiving nnsRP start off with a higher sexual functioning, but end up with the lowest scores. Patients receiving nnsRP have the highest baseline scores, but drop by more than 20 points, ending after a small recovery at a 24-month score similar to the baseline score in RT patients. Additional file 1 presents the trends for the three RT subgroups, which end with very similar scores after 24 months of follow-up.

Figure 1.

Sexual functioning (IIEF). Mean scores of nnsRP, nsRP, and RT cancer patients at baseline and during the 24-month follow-up period after multiple imputation of missing values (solid line: nnsRP, dashed line: nsRP, dotted line: RT).

Generic HRQOL – QLQ-C30

Several of the baseline functioning or symptom scales differed to a clinically relevant extent across the treatment groups (physical, role, emotional, and social functioning, fatigue, pain, dyspnoea, insomnia, and financial difficulties). In all cases, the RT group – especially the combRT group – had the least favourable baseline values (Figure 2 and Additional file 2 The most favourable value was generally found in the nsRP group, except for emotional functioning, social functioning, and insomnia, which was best in the externRT group. However, both the nsRP and the nnsRP group most often showed higher functioning and fewer symptoms than the total RT group.

Figure 2.

Health-related QoL (QLQ-C30). Mean scores of nnsRP, nsRP, and RT cancer patients at baseline and during the 24-month follow-up period after multiple imputation of missing values (solid line: nnsRP, dashed line: nsRP, dotted line: RT).

Three months after baseline, typically a decrease in functioning scales and an increase in symptom scales were seen, followed by a recovery (Figure 2 and Additional file 2 Due to the smaller number of patients, the time trends of the individual RT groups show a larger variability and the typical time trend can be seen less clearly than in the total RT group. When comparing nnsRP, nsRP, and the total RT group, the following deviations from this pattern were observed: In contrast to the more specific functioning and symptom scales, the global health status was hardly affected by treatment: It remained nearly unchanged for the total RT group but further increased over time for the two RP groups. The increase in emotional functioning was clinically relevant in the nsRP group at six, twelve, and twenty-four months and in the nnsRP group twelve months after baseline. Physical functioning, role functioning, and cognitive functioning worsened over time for the RT group, a group of older patients compared to those in the surgery groups. Among the RT groups, the largest differences were observed for dyspnoea, insomnia, and financial difficulties.

In the multiple regression analysis, only diarrhoea was statistically significantly associated with treatment option after adjusting for baseline HRQOL, age, and other demographical and clinical data (Table 2). RT was estimated to increase the score on the diarrhoea symptom scale by 7 points, while the estimate for nsRP was −0.9. The RT effect was mostly driven by the brachyRT group (effect of 7.8 compared to 4.1 and 4.2 for combRT and externRT). The difference was not clinically relevant.

Disease-specific Quality of Life – PORPUS-P

At baseline, the PORPUS-P score was highest in nsRP and lowest in RT patients, in particular in patients treated with combRT (Figure 3 and Additional file 3 The difference was clinically relevant. The PORPUS-P score decreased three months after baseline, followed by a partial recovery.

Figure 3.

PORPUS-P. Mean scores of nnsRP, nsRP, and RT cancer patients at baseline and during the 24-month follow-up period after multiple imputation of missing values (solid line: nnsRP, dashed line: nsRP, dotted line: RT).

In the covariate-adjusted multiple regression analysis a statistically significant but not clinically relevant association of main treatment option and HRQOL during the follow-up period was observed (Table 2), with nsRP patients having a significantly higher HRQOL and RT patients also having a (not statistically significant) higher HRQOL than nnsRP patients.

Comparison of Generic HRQOL to German Norm Values

Overall, scores of the treatment groups at baseline and twenty-four months after treatment were mostly comparable to German norm values for the EORTC QLQ-C30[29] (Table 3). However, patients treated with RP had clinically relevant lower scores of fatigue and pain at baseline, and the low pain score persisted for the nsRP patients for twenty-four months. The nsPR patients had a clinically relevant higher role functioning than expected at baseline and at the end of the follow-up. No clinically relevant deviations from the norm values were observed for the total group of patients treated with RT. However, the brachyRT group had a clinically relevant higher role functioning at baseline than expected. The externRT group had a higher role functioning at baseline as well, and lower symptoms of pain, which slightly decreased to a clinically non-relevant level until twenty-four months later. The combRT group had clinically relevant lower emotional and social functioning and higher insomnia scores at baseline, and less pain but more insomnia than the reference population twenty-four months after baseline.

Sensitivity Analysis – Complete Case Analysis

The complete case analysis was based on 248 to 254 patients, depending on the HRQOL measure. Results were mostly similar to the results after multiple imputation presented above. Global health status, emotional functioning, and appetite loss, but not diarrhoea, were additionally found to be significantly (but not clinically relevantly) related to main treatment option.