Introduction

Active surveillance (AS) is a recommended management option for patients with low- or very low-risk localised prostate cancer (LRPC), aiming to avoid or delay radical treatments and their complications [1, 2]. Evidence on long-term clinical outcomes from large cohorts of LRPC in patients undergoing AS [3,4,5] and from randomised controlled trials (RCTs), comparing observation [6] or active monitoring [7] with radical treatments at diagnosis, is reassuring.

However, fewer studies have examined the psychological burden associated with living with untreated prostate cancer, as well as the various dimensions of quality of life (QoL) when comparing AS with radical treatments. Systematic reviews [8,9,10], including only one randomised trial [11], suggest AS offers the most favourable profile for cancer-specific QoL, avoiding the negative impacts of surgery and radiation treatments on urinary and sexual function and of radiotherapy (RT) on bowel function, although some uncertainties remain regarding psychological aspects.

Common limitations in the available literature include variations in inclusion criteria, follow-up protocols, evolving treatment techniques, the measures used for health-related quality of life (HRQoL) and the representativeness of included patients in relation to the general population.

Despite being recommended by guidelines (including a regional guideline published in 2009) [12], the uptake of AS in the Piemonte and Valle d’Aosta oncology network (two regions in NW Italy with a population of about 4.5 million) was still extremely limited in 2015. Consequently, the START project, a prospective, population-based comparative cohort study, was launched in 2015 with the aim of promoting the widespread adoption of AS in clinical practice and comparing clinical and QoL outcomes according to initial treatment choice. A previous report detailed the general adoption of AS in clinical practice, factors associated with different patient choices and the main clinical outcomes [13].

This analysis focuses on longitudinal changes in general HRQoL dimensions, symptoms of anxiety and depression and specific urinary, bowel and sexual functions in LRPC patients according to their initial treatment choice.

Materials and methods

Study design

START was a prospective, population-based cohort study of LRPC patients meeting detailed inclusion criteria (eTable 1). At diagnosis, patients received both verbal and written information about their prognosis and the advantages and risks of AS, radical prostatectomy (RP) and RT. Patients undergoing RP or RT were followed up every 6 months. AS follow-up included PSA tests every 3 months, clinical evaluations every 6 months and re-biopsies at 12 and 48 months (eTable 2). Patients opting for AS could switch to a different treatment at any time, either for clinical reasons or on a voluntary basis (eTable 3).

The study protocol complied with the Declaration of Helsinki and received approval from all the regional Ethics Committees. All participants provided signed informed consent.

Participants

Patients residing in the Piemonte and Valle d’Aosta regions with a new diagnosis of prostate cancer between June 2015 and December 2021 were included if they met specific age (≤75 years), clinical cancer stage (T1c or T2a), PSA value (<10 ng/ml), biopsy and Gleason score (ISUP 2005 3 + 3, or 3 + 4 if the patient was >70 years) criteria. Exclusion criteria included prior prostate cancer treatment, contraindications to RP or RT or inability to provide informed consent (eTable 1).

Patients were asked to complete paper-based questionnaires at baseline and at 6, 12, 18 and 24 months and then annually until 60 months, regardless of any changes in treatment during follow-up. Patients who completed all the baseline questionnaires were included in the analyses.

Patient-reported outcomes measures (PROMs) are described in detail in eTable 4.

PROMs related to four domains were assessed:

  • Domain A: general HRQoL, using the EORTC QoL Questionnaire (QLQ-C30).

  • Domain B: anxiety and depression, using the Hospital Anxiety and Depression Scale (HADS) [14, 15].

  • Domain C: sexual function, including sexual activity and sexual functioning, using the QLQ-PR25 questionnaire and erectile function, using the International Index of Erectile Function (IIEF-5).

  • Domain D: urinary and bowel functions, using the relevant scales of the EORTC QLQ-PR25 questionnaire and urinary obstruction and voiding difficulties, using the International Prostatic Symptoms Score (IPSS).

Supplementary eTable 4 provides more details of the questionnaires and their scoring ranges.

Statistical analysis

Details of the sample size calculation (at least 400 men in the AS group were required to estimate the 24-months treatment-free-survival with sufficient precision) and other aspects of study design and statistical methods have been reported in a previous article [13].

The present analyses were conducted according to the initial treatment choice. Due to the small number of patients in the RT group, the main comparison was between AS and RP. Secondary analyses compared AS with both local treatments combined (RP or RT).

Descriptive longitudinal changes in PROMs over time according to treatment were presented graphically with means or percentages and 95% confidence intervals (95% CIs).

Two-level random effects models were used to account for centre and within subject correlations. Generalised linear models with variance components were used for continuous variables (linear models) or dichotomous variables (logistic models), assuming normal distributions of random effects. Adjusted effects were expressed as beta regression coefficients (β) for continuous scores (interpretable as absolute score differences), or odds ratios (ORs) for dichotomous variables.

All models included age, education, presence of a partner, PSA level at baseline, Gleason score, tumour stage, Charlson comorbidity index and baseline PROM values to adjust comparisons between groups. The impact of missing data was assessed with a sensitivity analysis after excluding centres with a mean questionnaire completion rate below the regional median.

Data analysis was performed using SAS 9.4 (SAS Institute, Cary, NC, USA).

Results

Of 852 eligible patients, 651 (76.4%) with completed baseline PROM questionnaires were included (eFig. 1). Patient characteristics according to initial treatment choice are described in Table 1. Patients undergoing RT were older (mean = 71.6 years) than those undergoing RP (mean = 67.4) or AS (mean = 68.6). The AS group had the lowest proportion of Gleason score 3 + 4 (14%) and stage T2a (12%). No clinically significant differences in PROMs across treatment groups were observed at baseline.

Table 1 Baseline patients’ characteristics and QoL measures by treatment group (No, %).
Full size table

During a median follow-up of 37 months, a total of 2095 PROM questionnaires were analysed (55% of expected). The response rate to PROM questionnaires was strongly associated with centres and period (with a clear decline in 2020–2021 due to the COVID-19 pandemic) and less so with patient characteristics. Longitudinal changes in PROMs according to initial treatment are presented in Figs. 1–3 for the four domains. The adjusted effects of AS on PROMs are reported in Table 2, compared with RP (Table 2A) and with RP or RT (Table 2B).

Fig. 1: General Health-related quality of life and mental health.
figure 1

Longitudinal changes in Domain A (General Health-related quality of life-EORTC QLQ-C30) and in Domain B (Anxiety and depression-HADS) by initial treatment: A Global health status (EORTC QLQ-C30). B Anxiety (HADS). C Depression (HADS).

Full size image
Fig. 2: Sexual function.
figure 2

Longitudinal changes in Domain C (Sexual function-PR25, IIEF) by initial treatment: A Sexual activity (PR25). B Sexual functioning (PR25). C Erectile function (IIEF).

Full size image
Fig. 3: Urinary, prostatic and bowel symptoms.
figure 3

Longitudinal changes in Domain D (Urinary, prostatic and bowel symptoms-PR25, IPSS) by initial treatment: A Use of incontinence aids (PR25). B Presence of moderate-severe prostatic symptoms (IPSS). C Presence of bowel symptoms (PR25).

Full size image
Table 2 Comparison of active surveillance versus radical prostatectomy (A) or versus any local treatment (prostatectomy or radiotherapy) in relation to different domains of HRQoL measures.
Full size table

Domain A: general HRQoL

Neither the EORTC QLQ-C30 global health status (Fig. 1A) nor the other five functional scales of HRQoL dimensions (eFig. 2A–E) showed any meaningful difference between AS and RP during follow-up. These results were confirmed by multivariable modelling (Table 2A), with all adjusted scores ranging from −0.80 to 1.61, on scales ranging from 0 to 100. The absence of differences in general HRQoL measures was also confirmed in comparisons between AS and RP or RT (Table 2B).

Domain B: anxiety and depression

Longitudinal scores of the HADS questionnaire are shown in Fig. 1B (anxiety) and 1C (depression). The AS group showed a stable average score for depression of around 4 points (on a scale ranging from 0 to 21), with an adjusted difference of 0.16 (95% CI: −0.01 to 0.33) (Table 2A). The corresponding adjusted score for AS versus RP or RT was 0.36 (95% CI: 0.13–0.60) (Table 2B).

Domain C: sexual function

After excluding men without sexual activity in the last 6 months before diagnosis, those included in the analyses showed similar baseline values. Men in the AS group experienced stable or slowly declining scores. In contrast, those who underwent RP reported significant declines in sexual activity (Fig. 2A), sexual functioning (Fig. 2B) and erectile function (Fig. 2C) during the first 6–12 months. These differences remained relatively constant thereafter. Adjusted regression coefficients confirmed that men who chose AS had better scores for sexual functioning (β = 8.27, 95% CI: 5.57–10.96) and sexual activity (β = 6.70, 95% CI: 4.19–9.20) (Table 2A). Slightly smaller differences were observed when comparing AS with RP or RT (Table 2B).

Domain D: urinary and bowel symptoms

The use of incontinence aids, which was similar at baseline, remained stable in the AS group but increased significantly in patients who underwent RP (OR = 0.14; 95% CI: 0.09–0.23) (Fig. 3A and Table 2A). Prostatic symptoms (mainly obstructive) remained stable during follow-up in the AS group, while a reduction was evident in the RP group (OR = 2.77; 95% CI: 1.52–5.06) (Fig. 3B and Table 2A). Bowel symptoms showed a worse baseline score in the AS group but no significant differences during follow-up compared with other treatments (Fig. 3C and Table 2A, B).

Results of sensitivity analyses comparing AS versus RP, limited to the 14 centres with a completion rate of the questionnaires higher than the median value (60%), are reported in eTable 5. The re-estimated coefficients based on these 14 centres, with an average completion rate of 89.5%, were very close to the main analyses. The few showing some differences suggest that in centres with a lower proportion of missing data the results were even more clear.

Discussion

During follow-up, LRPC patients who opted for AS rather than RP had a significantly lower risk of urinary incontinence, although they exhibited a stable higher prevalence of moderate or severe obstructive prostate symptoms. AS was associated with significantly better scores in sexual function, with no notable differences in other dimensions of HRQoL. Differences in depression and anxiety scores were minimal. These results were confirmed when comparing AS with RP or RT, despite the small number of RT patients and when the analyses were limited to half of the centres with a higher response rate.

Apart from two older RCTs comparing clinical outcomes of early versus watchful waiting or delayed treatments in prostate cancer patients [6, 16], the only available RCT with long-term PROM comparisons between active monitoring, RP and RT of localised PC is the PROTECT trial [11, 17]. At 12 years since diagnosis, patients randomised to active monitoring continued to report significantly less urinary leakage and sexual dysfunction than those assigned to RP and fewer bowel symptoms than those assigned to RT.

Systematic reviews of comparative studies on QoL [8,9,10] found only one randomised trial [11] comparing HRQoL between AS and active treatments. The most recent systematic review [10] included 21 prospective studies comparing AS with active treatments and concluded that no differences were reported in overall HRQoL outcomes and mental health PROMs, but AS showed better urinary continence and sexual function.

Some studies, including patients treated with more recent surgical or radiation techniques, confirm the same pattern of differences in HRQoL outcomes between AS and radical treatments [18], without meaningful differences according to the surgical approach [19].

Considering the general HRQoL dimension measured with the EORTC QLQ-C30, our results show a good and stable QoL during follow-up, regardless of the initial treatment choice. This finding is reassuring about the favourable prognosis of these low-risk patients. Interestingly, the differences reported by patients in other specific HRQoL domains (especially urinary and sexual functioning) are not reflected in these more general assessments. These findings are consistent with all comparative studies of AS included in recent systematic reviews [9, 10] and updated results of comparative studies with long follow-up times [17, 20].

A more controversial aspect of HRQoL in LRPC patients is the impact of treatment choice on mental health, particularly anxiety and depression. Our results suggest that most patients, whether on AS or after active treatments, reported stable scores for anxiety (around 5 points) and depression (around 4 points) during follow-up. While most studies did not report meaningful differences in mental health between AS and actively treated patients [9, 10, 17], some cohort studies reported higher anxiety levels, especially during the initial follow-up period, which could be a reason for early abandonment of AS by many patients [21, 22]. Depressive symptoms are less frequently reported in LRPC patients and rarely show clinically meaningful differences according to the initial treatment. A cross-sectional study conducted in the UK reported higher levels of anxiety (threefold) and depression (twofold) among AS patients compared with the general population [23], but the study design limited causal inference. However, fear of living with cancer is a major determinant of anxiety in AS patients, highlighting the need for careful counselling and psychological support both at diagnosis and during follow-up [24, 25]. Noteworthy are the results of a systematic review of 14 studies on the frequency of decision regret reported by patients with LRPC [26]: 13% in those opting for AS compared to a mean of around 20% for patients overall.

The major differences in QoL measures reported in our study by LRPC patients concern sexual functioning. Compared to patients who underwent RP, who reported a rapid and clear worsening in all indicators of sexual functioning, those in AS maintained or slowly reduced their baseline levels during follow-up. These findings are largely confirmed by the available literature [9, 10], even in studies with long follow-up [17, 20] and those considering more recent surgical or radiation techniques with shorter follow-up [18, 27].

The treatment choice showed clearly different effects on urological and prostatic symptoms. While patients in AS showed a stable or slow increase in the use of pads for urinary incontinence, patients receiving RP reported an early sharp increase in incontinence. This well-known side effect of surgery for PC is reported by almost all available studies [9, 10], including those involving robot-assisted prostatectomies [18, 27]. Conversely, while obstructive prostatic symptoms remained stable in the AS cohort, these symptoms significantly reduced in those who underwent RP. This positive effect of RP, largely confirmed in the literature, is nearly specular to the increased risk of incontinence. A less clear difference between treatments was reported for bowel symptoms, which are more frequently reported in patients who underwent radiation treatments [9, 10].

Strengths and limitations

The START study has several strengths that add relevance to its findings. Firstly, it is a large, population-based cohort study involving almost all urology and radiation oncology units in the Piemonte and Valle d’Aosta oncology network. The study protocol was aligned with those of the PRIAS study and other AS cohorts regarding key aspects, including inclusion criteria, follow-up schedules and guidelines for transitioning patients on AS to radical treatments, allowing fair comparisons of results. The results, analysed according to an intention-to-treat approach, reflect the real impact of the initial treatment choice on HRQoL outcomes, including the experiences of those who discontinued AS for active treatments. Lastly, the HRQoL domains analysed included the most important aspects considered by patients and professionals in the decision-making process and in guiding supportive interventions during follow-up [28, 29].

The main limitation of our study is the imbalance between groups, with a majority opting for AS (indeed a relevant result of the study) and fewer undergoing local treatment (especially RT), which precluded separate comparisons. In addition, the relative incompleteness of the QoL questionnaires collected during follow-up, particularly during the COVID-19 pandemic, increases the risk of selective missing data, although the results of sensitivity analyses are reassuring and seem to rule out any serious bias. Another limitation was the small number of patients who completed QoL questionnaires after definitive treatment following discontinuation of AS, which precluded analysis of changes to the QoL profile in this subgroup.

Conclusions

The results of START, a pragmatic cohort study representative of a large population of LRPC patients and involving almost all urology and RT centres of the oncology network, clearly outline the benefits (preservation of urinary continence and sexual functioning) and risks (persistence of prostatic symptoms) of AS, without other relevant differences. These results, along with the existing literature, could be shared between patients and their treating physicians to support more informed decision-making when diagnosed with LRPC.