Long-term Health-related Quality of Life in Patients on Active Surveillance for Prostate Cancer: A Systematic Review

Abstract

Context:

Active surveillance (AS) represents the preferred treatment option in patients with low-risk prostate cancer. Optimised patient selection has enabled more patients to be managed with AS for a longer time. Thus, there is growing interest in its effect on long-term quality of life compared with interventional management.

Objective:

To perform a systematic review evaluating the long-term patient-reported outcomes regarding mental health, and sexual and urinary function in patients on AS.

Evidence acquisition:

We performed a systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. We included series assessing validated patient-reported outcomes of health-related quality of life, and sexual and urinary function in AS patients followed up for at least 5 yr.

Evidence synthesis:

Our search yielded 1854 citations, including 19 papers involving 3643 patients on AS, 14 651 patients receiving surgery or radiotherapy, and 2478 controls without prostate cancer. In ten studies, major differences were observed in sexual and urinary symptoms between groups, such as better sexual function and fewer irritative urinary symptoms in patients on AS, though overall functional outcomes were comparable. In all studies, health-related quality of life for patients on AS was better than, or similar to, that for patients who had undergone surgery or radiotherapy and comparable with that for individuals without cancer.

Conclusions:

We observed differences in specific functional outcomes between patients on AS and surgery or radiotherapy, ≥5 yr after treatment. Patients on AS reported good quality of life, similar to that in individuals without prostate cancer. AS should continue to be a recommended management strategy for appropriately selected patients.

Patient summary:

Active surveillance is an accepted pathway for patients with low-risk localised prostate cancer. Previous literature has shown that it did not negatively affect short-term quality of life. This review finds that long-term quality of life for these patients is similar to that for people without prostate cancer.

1. Introduction

Active surveillance (AS) is an accepted, safe approach in patients with low-risk and select intermediate-risk prostate cancer (CaP) [1]. AS intends to avoid or defer curative treatment in patients with localised CaP, aiming to preserve urinary and sexual function without oncological compromise [2]. The major advantage of AS is avoidance of side effects associated with curative treatment by radical prostatectomy (RP) or radiation therapy (RT), such as urinary incontinence and erectile dysfunction [3]. The safety of AS has been well demonstrated regarding long-term oncological outcomes [4], including in patients with favourable intermediate-risk disease [5]. Additionally, with increasing experience in patient selection, patients are being managed safely on AS protocols for longer time [6].

Much interest has been devoted to assessing the impact of living with cancer on patients’ overall health-related quality of life (HRQoL). HRQoL and other patient-reported outcomes are particularly relevant in the setting of AS, given the associated psychosocial stressors and burden of AS protocols. AS requires intensive follow-up with regular prostate-specific antigen (PSA), digital rectal examination, and imaging, with the intention to offer active treatment upon disease progression [1]. A previous systematic review assessed HRQoL in patients on AS within 5 yr of diagnosis, and reported good HRQoL and low rates of anxiety and depression, though the authors highlighted the importance of further investigation to assess the long-term impact [7].

During shorter periods of follow-up, differences between the effects of AS and active treatment on urinary and sexual function are stark, given the known short-term effects of various treatments [8]. With longer follow-up, effects of active treatment are expected to attenuate [9], and AS patients are exposed to age-related changes in urinary and sexual function. Consequently, differences between patient-reported outcomes beyond 5 yr of follow-up are pertinent. This systematic review aims to provide an update on HRQoL in patients on AS, specifically examining patients with follow-up periods of at least 5 yr. We report on several relevant functional domains such as urinary and sexual function, as well as psychological and physical well-being.

2. Evidence acquisition

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses Protocol (PRISMA-P) [10] and assessment of multiple systematic reviews (AMSTAR) 2 guidelines [11]. This is a part of a larger project examining the effects of CaP treatment on quality of life. Prior to conducting the review, the search strategy, inclusion criteria, and risk of bias measure were registered on the PROSPERO database (CRD42021287902).

2.1. Search strategy and study selection

We searched PubMed/MEDLINE, Scopus, PsycNet, Google Scholar, Web of Science, PROSPERO, and Embase with reference to the Cochrane Handbook for Systematic Reviews [12] for papers published in English between 1974 and February 2022. References of relevant reviews and articles were searched, and experts in the field were consulted to identify additional studies. The search strategy is listed in the Supplementary material. The automated screening tool ASReview aided study selection. After screening based on study title and abstract, full text of the remaining articles was reviewed. Two reviewers (J.G.B. and N.P.) selected eligible studies after reaching a consensus on inclusion.

Studies were included if patients were aged >18 yr and surveyed for ≥5 yr following CaP diagnosis/treatment using a validated patient-reported outcome measure (PROM) instrument, that is, an instrument previously validated in at least one peer-reviewed study. Studies were excluded if those were case reports (two or fewer patients); were non-English papers; utilised an unvalidated PROM; were book chapters, theses, or preprint publications; or lacked PROM data. Conference proceedings, editorial features, news stories, and non–peer-reviewed articles were not included as these did not represent an appropriate level of evidence.

Only series including AS patients were included. Some older literature referred to any patient not receiving treatment as “watchful waiting”. These studies were included if the likely intention was deferred treatment rather than a palliative approach, for example, if the sample had localised CaP with ≥10 yr of life expectancy. Series with and without comparators were included. Comparators groups included patients with interventional treatment (eg, RP, RT, and brachytherapy) or a sample without CaP. Study design included randomised and nonrandomised design, as there were very few numbers of randomised controlled trials found by the search.

2.2. Data extraction and analysis

Data were extracted from an EndNote 20 library into an Excel spreadsheet by two authors (D.T. and J.T.) and included study year, study design, years the study was conducted, sources of funding, country/countries of the study, number and mean/median age of patients, and details of comparator groups. We further collated data on PROMs, broadly categorising these as mental health, HRQoL, physical health, urinary function, and sexual function. Data collected pertaining to each PROM included measure utilised, baseline measurements, follow-up details, and key findings. We elected not to perform a meta-analysis due to expected methodological heterogeneity and instead presented a narrative synthesis of evidence.

2.3. Risk of bias assessment

Bias was assessed independently by two authors (D.T. and J.G.B.), and disagreements were resolved by discussion, utilising the ROBINS-I tool [13]. This tool assesses for bias from six domains: confounding, participant selection, classification of interventions, missing data, outcome measurement, and selection of the reported result.

3. Evidence synthesis

3.1. Included studies

We identified 1854 citations relating to long-term quality of life in patients diagnosed with CaP. Of these, 1503 were excluded as these did not include patients with follow-up ≥5 yr. A further 166 were excluded as these included a time point at ≥5 yr, but did not include a PROM, used an unvalidated PROM, or did not report data. Of the remaining 185 studies analysed, 19 contained data relating to deferred treatment. The PRISMA flow diagram is shown in Figure 1, and a summary of the 19 included studies is in Table 1. The studies involved 3643 patients on AS, 14 651 patients received surgery or RT, and 2478 were CaP-free controls.

Fig. 1 –

Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram.

Table 1 –

Overview of quality of life studies of men on active surveillance for prostate cancer

Author	QoL study design	Assessment period	AS protocol; country; total sample/response rate	Sources of funding	Conversion to active treatment	AS men, no. (mean/median age, yr)	Men in comparator group, no. (mean/median age, yr)
Arredondo (2008) [14]	Observational, cohort	1990–2001	CaPSURE; USA; 310/85%	CaPSURE was funded by TAP Pharmaceuticals until 2007, funded by UCSF after 2007	None. Patients who received active therapy were removed from this study	310 (74.7a)	None
Bill-Axelson (2013) [30]	Randomised, cohort	1989–2007	SPCG-4; Sweden, Finland, Iceland; 347/78%	Swedish Cancer Society and the Swedish Prostate Cancer Foundation	Not reported	136 (64.5a)	136 RP (64.4a)
Carlsson (2016) [29]	Population, cross sectional	2011	National Prostate Cancer Register; Sweden; 7003/63%	Swedish Research Council, Swedish Cancer Foundation, and Lion’s Cancer Research Foundation	Not reported	670 (65a, 78b)	2640 RP (61a, 74b), 857 post-RP (64a, 76b), 1139 ADT ± RP/RT, 1000 control (age matched ± 1 yr)
Donovan (2016) [20]	Randomised, cohort	1999–2014	ProtecT; UK; 1643/85%	U.K. National Institute for Health Research Health Technology Assessment Program	At 6 yr, 22.2% received RP,13% RT	545 (62a)	553 RP (62a), 545 RT (62a)
Egger (2018) [21]	Population, cohort	2000–2012	Prostate Cancer Care and Outcome Study; Australia; 341/59%	Australian Commonwealth Department of Veterans Affairs and the National Health and Medical Research Council of Australia	At 10 yr, 36 men were still on AS, 27 had converted to active treatment	63 (73a)	221 RP (68a). 25 RT and/or HDR (73a), 32 LDR (71a)
Fransson (2009) [23]	Randomised, cohort	1986–2006	UMEA1; Sweden; 72/75%	Swedish Cancer Society	Not reported	27 (78b)	27 RT (77b)
Hoffman (2020) [9]	Population, cohort	2011–2017	CaPSURE and others; USA; 1386/77%	Agency for Healthcare Research and Quality, Patient-Centered Outcomes Research Institute, and the National Cancer Institute	At 5 yr, 89/363 had converted (44 EBRT, 37 RP, 5 ADT, 3 ablative therapy)	363 (67a)	Favourable risk: 675 RP (60a), 261 EBRT (68a), 87 LDR (65a) Unfavourable risk: 402 RP (64a), 217 EBRT with ADT (71a)
Hoffman (2017) [15]	Population, cohort	1994–2010	Prostate Cancer Outcomes Study; USA; 934/69%	National Cancer Institute and Holden Comprehensive Cancer Center	Not reported	92 (66.4a)	696 RP (60.6a), 146 RT (65.8a)
Johansson (2018) [32]	Randomised	1989–2016	SPCG-4; Sweden; 681/83%	National Institutes of Health, Swedish Cancer Society, Johanna Hagstrands Foundation	At 12.2 yr: 15% had RP, 28% had ADT, 23% were receiving other antiandrogens	167 (65.3a, 77.2b)	182 RP (64.1a, 76.2b), 214 control (age [4-yr interval] and region matched [72.3b])
Johansson (2011) [31]	Randomised	1989–2008	SPCG-4; Sweden; 681/83%	National Institutes of Health, Swedish Cancer Society, Johanna Hagstrands Foundation	As above	167 (65.3a, 77.2b)	182 RP (64.1a, 76.2b), 214 control (age [4-yr interval] and region matched [72.3b])
Kasperzyk (2011) [27]	Cohort, cross sectional	2005	Physicians’ Health Study; USA; 1230/41%	National Cancer Institute and National Heart, Lung, and Blood Institute	9.6% RP, 15.2% EBRT alone, 8.8% brachytherapy only, 9.6% ADT + RT, 16% ADT alone	125 (72.3a)	1105 active treatment (68.1a)
Litwin (1995) [19]	Population, cross sectional	1994	NR; USA; 214/79%	Robert Wood Johnson Foundation	Not reported	60 (75.2b)	98 RP (69.7b), 56 RT (76.2b), 273 prostate cancer–free control (age and region matched; 72.5b)
Lokman (2021) [16]	Prospective cohort	2006–2019	PRIAS; Finland; 348/80%	Finnish Cancer Foundation, and the Jane and Aatos Erkko Foundation	Not reported	348 (68b)	Population data
Mazariego (2020) [28]	Population, cohort	2000–2017	Prostate Cancer Care and Outcome Study; Australia; 2428/30%	Cancer Council New South Wales, Australian Department of Veterans Affairs, National Health and Medical Research Council of Australia	Not reported	200 (NR)	192 NSRP (NR), 141 non-NSRP (NR), 43 EBRT/HDR (NR), 25 LDR (NR), 45 ADT (NR), 786 control (age matched)
Merrick (2020) [25]	Hospital series, cohort	2005–2016	NR; USA; 226/100%	Not reported	Not reported	226 (65.7b)	None
Mols (2006) [17]	Population, cross sectional	2004	Eindhoven Cancer Registry; The Netherlands; 964/81%	Rotary Tilburg Triborch and the Comprehensive Cancer Center South, The Netherlands	Not reported	56 (NR)	193 RP (NR), 263 RT (NR), 60 ADT (NR) Population data
Punnen (2015) [18]	Observational, cohort	1995–2010	CaPSURE; USA; 3294/81%	CaPSURE was funded by TAP Pharmaceuticals until 2007, funded by UCSF after 2007	Not reported	64 (72.5a)	1139 NSRP (60.0a), 860 non-NSRP (62.9a), 684 BT (68.3a), 386 EBRT (71.3a), 161 ADT (73.6a)
Thong (2010) [24]	Registry, cohort	1994–2008	Eindhoven Cancer Registry; The Netherlands; 128/55%	Not reported	1 RP, 5 RT, 2 ADT, 1 ADT+ RT	71 (75.8b)	71 RT (75.9b) Population data
Venderbos (2017) [22]	Cross sectional	2015	PRIAS; The Netherlands; 879/70%	Prostate Cancer Research Foundation, Rotterdam, The Netherlands	2 patients converted to active therapy	120 (65.3a, 71.9b)	69 RP (70.0a, 76.0b), 219 RT (65.9a, 74.5b), 205 control (74.5b)

ADT = androgen deprivation therapy; AS = active surveillance; BT = brachytherapy; CaPSURE = Cancer of the Prostate Strategic Urologic Research Endeavor; EBRT = external beam radiation therapy; HDR = high dose rate brachytherapy; LDR = low dose rate brachytherapy; NR = not reported; NSRP = nerve-sparing radical prostatectomy; PRIAS = Prostate Cancer International Active Surveillance; ProtecT = Prostate Testing for Cancer and Treatment; QoL = quality of life; RP = radical prostatectomy; RT = radiotherapy; SPCG-4 = Scandinavian Prostate Cancer Group Study Number 4; UCSF = University of California San Francisco.

^aAt study enrolment or cancer diagnosis/treatment.

^bAt the last follow-up or instrument completion.

3.2. Instruments

The instruments to measure general HRQoL were the Short Form Survey 36 (SF-36) used in seven studies [9,14–19], Short Form Survey 12 (SF-12) [20–22], and EuroQal Visual Analogue Scale [22]. Cancer-related HRQoL was measured using European Organisation for Research and Treatment of Cancer Core Quality of Life Questionnaire [20,23], Quality of Life—Cancer Survivors (QoL-CS) [17,24], Cancer Rehabilitation Evaluation System—SF (CARES-SF) [19], and Functional Assessment of Cancer Therapy—General (FACT-G) [19]. Three additional tools were used to gauge mental health separately from general HRQoL, including Center for Epidemiological Disease—Depression [25], State-Trait Anxiety Inventory 6 [22], and the Hospital and Anxiety Depression Score [20]. Decision regret using a validated scale [26] was implemented in one study [15]. Multiple tools were used to evaluate functional outcomes in CaP patients, including Expanded Prostate Cancer Index Composite 26 [9,20–22,24,27,28], International Consultation on Continence Questionnaire [20], the University of California Los Angeles Prostate Cancer Index (UCLA-PCI) [14,15,18,21,27,28], the Prostate Cancer Symptom Scale [23,29], and custom questionnaires in the Scandinavian Prostate Cancer Group Study 4 [30–32] and Dutch Sexual Activities module [24]. Measures used to evaluate lower urinary tract symptoms (LUTSs) included the International Continence Society “Male” Short Form Questionnaire [20] and the International Prostate Symptom Score [25,29]. Owing to the nature of questionnaire studies, risk of measurement bias exists as the studies involve patients’ self-reporting outcomes, as evidenced by the ROBINS-I tool result of a “serious” risk of bias (Supplementary Fig. 1).

3.3. Outcomes

3.3.1. General HRQoL

All AS cohorts reported good overall HRQoL at 5–15 yr after diagnosis (Table 2). Two studies without comparator groups also found that there was no appreciable deterioration in HRQoL over time [25,27]. In the 12 studies comparing AS with interventional treatment groups, none observed any overall difference in HRQoL between treatment groups [9,15,17–21,23,24,30–32]. Two papers reported that a change in treatment strategy to active treatment did not adversely affect patients’ HRQoL [16,27]. For example, Kasperzyk et al [27] reported that patients initially on AS who subsequently underwent treatment had comparable HRQoL outcomes than patients who had been treated immediately.

Table 2 –

Measurement, outcomes, and key findings on overall PROMs in patients on active surveillance for prostate cancer

Study (publication date)	HRQoL outcomes	Standardised measures	Baseline assessment, follow-up	Key findings	QoL on AS
Arredondo (2008) [14]	HRQoL, PCa-specific functional outcomes	UCLA-PCI, SF-36	Baseline–5 yr	Mental health in men on AS was stable over 5 yr, other HRQoL domains declined over time	Overall HRQoL declined over time, more so than would be expected with normal ageing
Bill-Axelson (2013) [30]	HRQoL, PCa-specific functional outcomes, anxiety, depression	Study-specific questionnaire	Baseline–8 yr	All PCa groups showed reduction in HRQoL over time. RP had worse functional outcomes, whereas AS group had more anxiety	HRQoL was impacted negatively by cancer in both AS and treatment groups, though in different ways. Identifying symptoms at baseline is important for tailored decision-making
Carlsson (2016) [29]	PCa-specific functional outcomes	IPSS, IIEF-5, PCSS	No baseline; 12 yr	AS group had similar functional outcomes to the control group of men without PCa. Erectile function was worse in all treatment groups than control group and AS patients. Urinary incontinence was associated with a treatment regime including RP	AS patients had good functional HRQoL, similar to control group and better than treated groups
Donovan (2016) [20]	HRQoL, PCa-specific functional outcomes, anxiety, depression	ICIQ, EPIC, ICSmaleSF, SF-12, HAD-S, EORTC QLQ-C30	Baseline–6 yr	Urinary and sexual function declined over time in men on AS, but less than in men in treatment groups (RP or RT). HRQoL domains recovered in the first 12 mo and thereafter remained stable over time in all groups	AS patients had good overall HRQoL that did not diminish over time. Anxiety and depression rates were low in all groups
Egger (2018) [21]	HRQoL, PCa-specific functional outcomes, anxiety, depression	EPIC-26, UCLA-PCI, SF-12	Mean of 3 mo–10 yr	AS group had similar functional outcomes to treatment groups and better continence than RP group. AS group reported more distress and disease-related hyperarousal than RT groups	AS patients had good overall HRQoL and functional outcomes, scoring similarly to treatment groups in most domains
Fransson (2009) [23]	HRQoL, PCa-specific functional outcomes	PCSS, QLQ-C30	6–16 yr	AS and RT had comparable urinary bother; AS group reported less sexual bother. AS patients had worse physical but similar emotional health	AS patients had similar overall HRQoL to the group treated with RT, though there were some small differences in certain functional outcomes
Hoffman (2020) [9]	HRQoL, PCa-specific functional outcomes	EPIC-26, SF-36	Baseline–5 yr	AS group had better sexual function and continence than RP group, and less urinary and bowel bother than BT group. No group reported a decline in physical or emotional well-being over time	HRQoL scores were high across all treatment groups, indicating good quality of life. This did not change over time. AS group had better functional outcomes than treatment groups in certain domains
Hoffman (2017) [15]	HRQoL, PCa-specific functional outcomes, treatment regret	UCLA-PCI, SF-36, EPIC-26	6 mo–15 yr	AS group had similar functional outcomes to treatment groups. AS group was less likely to report “excellent” or “very good” health than RP or RT group, though also less likely to report decision regret	Overall HRQoL was similar between AS, RP, and RT groups. Low rates of treatment regret were seen, with 8.2% in AS group, 15% in RP group, and 16.6% in RT group
Johansson (2018) [32]	HRQoL, PCa-specific functional outcomes, anxiety, depression	Study-specific questionnaire	7–19 yr	AS patients had similar anxiety and depression levels to a control group (moderate-high anxiety in 35% AS, 33% control, moderate-high depression in 40% AS, 41% control). Functional outcomes were good in AS patients	HRQoL in AS group was similar to that in a control group of men without PCa (“high” QoL reported in 44% AS, 36% RP, 45% control). Good patient counselling and informed decision-making are important
Johansson (2011) [31]	HRQoL, PCa-specific functional outcomes, anxiety, depression	Study-specific questionnaire	Median 3.7–13.4 yr	AS and RP groups had worse erectile function than control group. AS group had better continence but worse stream and nocturia than RP group. Both groups had more anxiety than control group (moderate-high anxiety in 43% of AS, 43% of RP, 33% of control groups)	Overall HRQoL was generally high in AS and RP groups and not significantly different from control group (“high” QoL in 34% AS, 35% RP, 45% control, “high” well-being in 44% AS, 41% RP, 52% control)
Kasperzyk (2011) [27]	HRQoL, PCa-specific functional outcomes	EPIC-26, UCLA-PCI	Range of 1–23 yr	AS patients had better continence and erectile function than treatment groups (3.5% vs 10% incontinence, 68% vs 78% impotence), more obstructive voiding symptoms (22% vs 13%)	AS patients had similar HRQoL outcomes when compared to treatment groups and also scored better in certain functional domains
Litwin (1995) [19]	HRQoL, PCa-specific functional outcomes	SF-36, CARES-SF, FACT-G	No baseline; range of 3–33 yr	AS patients had better urinary and sexual function than RP group, similar urinary and sexual function to RT group, worse sexual function and bother than control group. AS group more likely to report role limitations due to emotional issues than control or RP/RT groups	AS patients had similar overall HRQoL outcomes to control and treatment groups. AS patients scored higher in sexual and urinary function than RP patients, but were also more likely to report role limitations due to emotional issues
Lokman (2021) [16]	HRQoL	SF-36	Baseline–11 yr	AS patients scored highly on RAND-36 and scores remained stable over time. When comparing men on AS and men previously on AS who later underwent RP, QoL scores were similar	HRQoL was good in AS group and comparable with or better than age-stratified reference values. AS patients who later required treatment also had good QoL despite delay in treatment
Mazariego (2020) [28]	HRQoL, PCa-specific functional outcomes	UCLA-PCI, EPIC-26	Baseline–15 yr	AS patients had better erectile function and continence than RP patients but worse than control group. AS group was the only PCa group to report that mental health was unaffected by PCa	HRQoL was worse in AS patients when than a control group of men without PCa; however, AS patients scored better than men who had undergone RP
Merrick (2020) [25]	HRQoL, PCa-specific functional outcomes, fatigue, depression	IPSS, FAS, IIEF-6, CES-D	Baseline–8 yr	Erectile function in men on AS worsened over time (in patients fully potent at baseline, 73% remained so at 8 yr). No deterioration in other functional domains. Depression rates were low	HRQoL remained good for men on AS over the 8-yr follow-up, with the only decline seen in erectile function
Mols (2006) [17]	HRQoL	SF-36, QoL-CS	Range of 5–10 yr	Men on AS reported better mental health scores than those of treatment groups, but these were not felt to be clinically significant. AS patients over age 75 yr reported worse physical health than men in treatment groups	Overall HRQoL in men on AS was similar to or better than that seen in treatment groups or in the general population. In men >75 yr old, those treated with RP had better physical and social functioning than other groups
Punnen (2015) [18]	HRQoL, PCa-specific functional outcomes	SF-36, UCLA-PCI	Baseline–10 yr	AS group had better functional outcomes than RP or RT group. No differences in mental or physical health between any of the groups	Overall HRQoL remained stable over time in men on AS and was comparable with that of both treatment groups
Thong (2010) [24]	HRQoL, PCa-specific functional outcomes	QoL-CS, EPIC, SAc	10–14 yr	AS group had similar urinary function to RT group, both worse than general population. AS group had better physical and sexual function than RT group, similar to general population	Overall HRQoL was similar between men on AS, men who had undergone RT, and age-matched men from the general population
Venderbos (2017) [22]	HRQoL, PCa-specific functional outcomes, anxiety	EPIC-26, SF-12, STAI-6, EQ-VAS	Mean of 6.6 yr	Functional outcomes of AS group were similar to or better than those of treatment groups and comparable with those of control group. Anxiety scores were low in all groups	Overall HRQoL in AS group was similar to that in control group. AS patients scored higher than treated patients in some functional domains, such as urinary continence and erectile function

AS = active surveillance; BT = brachytherapy; CARES-SF = Cancer Rehabilitation Evaluation System; CES-D = Centre for Epidemiological Studies Depression; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Core Quality of Life Questionnaire; EPIC = Expanded Prostate Cancer Index Composite; EQ-VAS = EuroQal Visual Analogue Scale; FACT-G = Functional Assessment of Cancer Therapy—General; FAS = Fatigue Assessment Scale; HAD-S = Hospital Anxiety and Depression Scale; HRQoL = health-related quality of life; ICIQ = International Consultation on Incontinence Questionnaire; ICSmaleSF = International Continence Society “Male” Short Form; IIEF-5/6 = International Index of Erectile Function 5/6; IPSS = International Prostate Symptom Score; PCa = prostate cancer; PCSS = Prostate Cancer Symptom Scale; PROM = patient-reported outcome measure; QoL = quality of life; QoL-CS = quality of life cancer survivors; RP = radical prostatectomy; RT = radiotherapy; SAc = sexual activities module; SF-12/36 = Short Form Survey 12/36; STAI-6 = State-Trait Anxiety Inventory 6; UCLA-PCI = University of California Los Angeles Prostate Cancer Index.

Nine studies compared HRQoL in patients on AS with CaP-free patients. Seven observed no substantial differences between groups [16,17,19,22,24,29,32]. Two described worse HRQoL in patients on AS compared with the general population, but this result was also seen in active treatment groups [14,31].

Several studies assessed baseline predictors of long-term HRQoL within the AS group of patients. Bill-Axelson et al [30] found that patients with good function and mental health at baseline were more likely to report favourable HRQoL after 8 yr of AS. Hoffman et al [15] identified multiple factors predicting higher HRQoL, including older age at diagnosis, feeling of well-informed decision-making, less sexual or bowel bother at baseline, and less PSA anxiety. To maintain good long-term general HRQoL, a common recommendation was that ensuring that patients are well-informed active participants in the decision-making process to commence AS.

3.3.2. Mental health

Fourteen studies examined mental health, ten demonstrated no statistically significant long-term difference between patients on AS compared with those undergoing active treatment [9,15,18,20–24,31,32]. Egger et al [21] found initially similar distress and hyperarousal scores in AS patients than in the RT group (adjusted mean difference = 5.9 [95% confidence interval {CI} 0.5–11.3] vs 5.4 [95% CI 0.2–10.5]), and there were no statistically significant differences in other mental health domains. Two studies observed slightly superior mental health in AS patients to those who underwent RP, with Mols et al [17] reporting scores of 7.3 versus 7.0 in the QoL-CS psychological domain for patients in the AS versus RP group. This difference was well within half the standard deviation of this measure and therefore unlikely to be clinically significant. Using SF-12, Mazariego et al [28] found a small disparity in mental well-being scores particularly within 3 yr of diagnosis, with mean differences of −1.3 and −0.8 in, respectively, the nerve-sparing and non–nerve-sparing RP groups when compared with AS, though this attenuated by the end of the 15-yr follow-up period. Hoffman et al [15] reported that patients on AS had less decision regret: 8.2% of AS patients versus 15% of RP patients.

In contrast, Bill-Axelson et al [30] reported higher odds of anxiety and distress in AS patients than in RP patients (odds ratio = 0.60, 95% CI 0.38–0.96), which was attributed to the close monitoring protocol of AS. There were no consequential differences in other psychological parameters, such as depression, guilt, and self-confidence [30]. Additionally, Litwin et al [19] concluded that AS patients were more likely than treated patients or controls to report limitations in day-to-day life because of emotional distress, a result found to be clinically meaningful (RAND-36 score 57.4 vs 70.2 for RP and 75.6 for RT), which authors attributed to anxiety and uncertainty about the future. However, the same study did not observe a difference in other mental health parameters or emotional domains assessed by SF-36, CARES-SF, and FACT-G [19].

Of the nine studies with controls or population data as a comparator, six did not find a difference in mental health outcomes [14,16,22,24,28,32]. Conversely, Johansson et al [31] observed similar mental health outcomes when comparing AS and RP, but found that both groups experienced higher anxiety than the control group (risk ratio = 1.42, 95% CI 1.07–1.88), with no differences in other psychological outcomes such as depression, sense of meaningfulness, and well-being.

3.3.3. Physical health

Six of 13 papers did not detect a difference in physical health measures between the AS and treatment groups [9,19–21,30,32]. Venderbos et al [22] found similar outcomes between AS and RP, though AS patients had better physical health than RT patients (SF-12 Physical Component Score = 50.7 vs 47.8, p = 0.011), a finding corroborated by Thong et al [24]. Another study found differences between AS and non–nerve-sparing RP favouring AS (odds ratio of clinically significant physical decline = 2.7, p < 0.01), but not between the AS group and the nerve-sparing RP or RT group [18].

Four studies observed worse self-reported physical health in AS patients compared with treated patients. Hoffman et al [15] showed that 29.2% of AS patients reported SF-36 health rating of “excellent” or “very good” 15 yr after diagnosis, compared with 48.3% and 42.2% of RP and RT patients, respectively. Fransson et al [23] reported similar findings comparing physical functioning in the AS and RT groups, with only the AS group showing a decline over time (deterioration in QLQ-C30b score of −11.5, p = 0.041). Mols et al [17] demonstrated that physical health outcomes in patients aged <75 yr were comparable between the AS and RP groups, though older patients on AS reported worse physical health, which the authors attributed to the association between age and choice of treatment. Considering all patients in the Mols et al’s [17] study, physical functioning appeared superior in the RP group to that in the other treatment modalities. Lastly, while Mazariego et al [28] found similar outcomes up to 15 yr of follow-up, AS patients reported a sharper decline in physical health than treated patients, which was both statistically and clinically significant (mean difference = −3.7, 95% CI −6.7 to −0.8).

In studies comparing the AS cohort with a CaP-free cohort, five observed no appreciable difference in self-reported physical health [16,19,22,24,32]. Mols et al [17] reported on subsections of the SF-36 noting statistically significantly better “bodily pain” scores in AS patients than in an age-matched sample, though this difference was also noted for active treatment groups.

Bowel function was assessed in some studies with an RT cohort. Two reported that RT patients had initially worse bowel bother than AS patients at 1 yr of follow-up [9,20], while an additional four noted that this difference persisted in the long term [15,18,21,24].

3.3.4. Urinary function

Ten papers compared urinary function in AS and RP cohorts (Table 3). Only one study found no difference between groups using the UCLA-PCI, which provides an overall urinary bother score rather than separate scores for continence and LUTSs [15]. Five studies demonstrated clinically better overall urinary function, particularly with respect to continence [18–22]. Four papers found that continence was better in the AS group, but these were more likely to report irritative or obstructive symptoms [9,28,30,31]. Kasperzyk et al [27] also observed better continence than all treatment groups (moderate to severe incontinence in 3.5% AS vs 10.0% treatment, p = 0.024).

Table 3 –

Measurement, outcomes, and key findings on urinary function in patients on active surveillance for prostate cancer

Study (publication date)	Urinary function outcomes	Standardised measures	Baseline assessment, follow-up	Key findings
Arredondo (2008) [14]	Urinary function, urinary bother	UCLA-PCI	Baseline–5 yr	Decline in urinary function over time but similar to expected age-related change
Bill-Axelson (2013) [30]	Urinary leakage, obstruction, urinary tract pain, urgency, catheter problems, haematuria	Study-specific questionnaire	Baseline–8 yr	More incontinence in RP group than in AS group (OR = 8.92 in years 0–3, 4.02 in years 4–9); RP group had less obstructive voiding symptoms (OR = 0.34)
Carlsson (2016) [29]	Urinary urge, urinary incontinence	IPSS, PCSS	No baseline; 12 yr	No difference between AS and control
Donovan (2016) [20]	Urinary incontinence, voiding symptoms, frequency, nocturia	ICIQ, EPIC, ICSmaleSF	Baseline–6 yr	AS had no effect on continence or urinary frequency. All groups saw similar worsening in nocturia over time
Egger (2018) [21]	Urinary incontinence, urinary bother	EPIC, UCLA-PCI	Mean of 3 mo–10 yr	More incontinence in RP group than in AS group (AMD 9.1). No difference in urinary bother scores
Fransson (2009) [23]	Urinary bother, incontinence, frequency, urgency, nocturia, starting problems, dysuria	PCSS	6–16 yr	AS group reported better stream and less nocturia than RT group, though no difference in overall urinary bother
Hoffman (2020) [9]	Urinary incontinence, irritative symptoms	EPIC	Baseline–5 yr	AS group had better continence, worse irritative symptoms than RP group (AMD 10.9 and −5.7), better continence and irritative symptoms than BT group (7 and 7). Comparable urinary symptoms in AS and EBRT
Hoffman (2017) [15]	Urinary bother	UCLA-PCI	Baseline–15 yr	No difference between AS and any treatment group (RP or RT)
Johansson (2011) [31]	Urinary incontinence, stream, nocturia	Study-specific questionnaire	Median 3.7–13.4 yr	AS group had worse stream and nocturia than RP or control groups, but better continence than RP group
Kasperzyk (2011) [27]	Urinary incontinence, nocturia, frequency, stream, haematuria	EPIC, UCLA-PCI	No baseline; range of 1–23 yr	AS group less likely to report incontinence than treatment groups (OR = 0.33). No difference seen in other parameters
Litwin (1995) [19]	Urinary function, continence, bother	CARES-SF	No baseline; range of 3–33 yr	AS group had better urinary function, continence, and bother than RP group. No difference between AS and RT or AS and control
Mazariego (2020) [28]	Incontinence, nocturia, stream, frequency, haematuria, bother	EPIC, UCLA-PCI	Baseline–15 yr	No difference between AS and control groups. AS less likely than RP group to have incontinence (AMD 16.4), more likely to have urinary bother
Merrick (2020) [25]	Overall IPSS score	IPSS	Baseline–8 yr	No change in urinary function over 8 yr
Punnen (2015) [18]	Urinary function, urinary bother, continence	UCLA-PCI	Baseline–10 yr	AS group had better continence than RP group. Comparable urinary function observed in AS and treatment groups (RP, RT, ADT)
Thong (2010) [24]	Urinary function, urinary bother	EPIC	10–14 yr	No difference in urinary function between AS and RT groups, though both reported worse function than population data
Venderbos (2017) [22]	Urinary incontinence, urinary bother	EPIC	Mean of 6.6 yr	AS group had better urinary function than RP group (mean score 93 vs 80). No difference between AS, RT, and control groups

ADT = androgen deprivation therapy; AMD = adjusted mean difference; AS = active surveillance; BT = brachytherapy; CARES-SF = Cancer Rehabilitation Evaluation System; EBRT = external beam radiation therapy; EPIC = Expanded Prostate Cancer Index Composite; ICIQ = International Consultation on Incontinence Questionnaire; ICSmaleSF = International Continence Society “Male” Short Form; IPSS = International Prostate Symptom Score; OR = odds ratio; PCSS = Prostate Cancer Symptom Scale; RP = radical prostatectomy; RT = radiotherapy; UCLA-PCI = University of California Los Angeles Prostate Cancer Index.

Of the ten studies comparing AS with RT, three observed better urinary function in the AS group [9,20,23], but this was generally only for a short term. Fransson et al [23], however, found that while urinary function in AS patients remained stable over 10 yr, RT patients reported worsening nocturia (p = 0.023). The remaining seven studies did not find any statistically significant differences in urinary function between the AS and RT groups [15,18,19,21,22,24,28].

Seven studies compared urinary function in the AS and CaP-free groups, with five showing no statistically significant differences [14,19,22,28,29]. Conversely, Johansson et al [31] reported that of AS patients 40% reported weak stream and 63% nocturia, compared with only 26% and 42% of CaP-free patients, respectively (risk ratio = 1.50, 95% CI 1.08–2.07 and risk ratio = 1.33, 95% CI 1.08–1.64, respectively). Similarly, Thong et al [24] described worse overall urinary function and more urinary bother.

3.3.5. Sexual function

All 12 papers examining this domain observed better or comparable sexual function in the AS group when compared with treatment groups (Table 4).

Table 4 –

Measurement, outcomes, and key findings on sexual function in patients on active surveillance for prostate cancer

Study (publication date)	Sexual function outcomes	Standardised measures	Baseline assessment, follow-up	Key findings
Arredondo (2008) [14]	Sexual function, sexual bother	UCLA-PCI	Baseline–5 yr	Decline of sexual function steeper in first 2 yr after starting AS. Decrease in sexual function worse than expected with normal ageing
Bill-Axelson (2013) [30]	Libido, erectile function	Study-specific questionnaire	Baseline–8 yr	AS group had a gradual decrease in erectile function and libido, but less than that in RP group
Carlsson (2016) [29]	Erectile function, sexual activity	IIEF-5	No baseline; 12 yr	No difference between men on AS and control group with regard to erectile function
Donovan (2016) [20]	Erectile function, sexual function, sexual bother	EPIC	Baseline–6 yr	Erectile/sexual function declined over time in men on AS but less than in RP + RT groups
Egger (2018) [21]	Sexual function, sexual bother, erectile function	UCLA-PCI, EPIC-26	Mean of 3 mo–10 yr	AS had similar sexual health to AT groups
Fransson (2009) [23]	Sexual bother	PCSS	6–16 yr	Men on AS had less sexual bother than RT group, but similar erectile function and libido
Hoffman (2020) [9]	Sexual function: erectile frequency/quality	EPIC 26	Baseline–5 yr	AS group had better sexual function than RT group at 3 yr, and better than BT group at 1 yr, but comparable to EBRT group at 5 yr
Hoffman (2017) [15]	Sexual bother	UCLA-PCI	Baseline–15 yr	No difference in sexual outcomes between the groups
Johansson (2011) [31]	Sexuality, erectile function, frequency of intercourse/orgasm	Study-specific questionnaire	Median of 3.7–13.4 yr	AS group had better sexual function than RP group, but both groups were worse than control
Kasperzyk (2011) [27]	Erectile function, sexual desire	EPIC-26, UCLA-PCI	No baseline; range of 1–23 yr	AS group had significantly better erectile function than AT groups, though all groups had similarly reduced sexual desire
Litwin (1995) [19]	Sexual function, sexual bother	CARES-SF	No baseline; range of 3–33 yr	AS group had better erectile function than RP groups, and similar function and bother to RP group, but worse than control population
Mazariego (2020) [28]	Sexual bother, sexual function	EPIC-26, UCLA-PCI	Baseline–15 yr	All PCa groups including AS had worse sexual function than control. AS group had better erectile function than RP group
Merrick (2020) [25]	Erectile function	IIEF-6	Baseline–8 yr	Men on AS had a decline in sexual function over time (73% potency preservation over 10 yr)
Punnen (2015) [18]	Sexual function, sexual bother	UCLA-PCI	Baseline–10 yr	AS group was less likely to report sexual dysfunction than surgery groups
Thong (2010) [24]	Erectile function	Dutch Sexual Activities module	10–14 yr	AS had similar sexual function to general population and better than that in RT group, though more problems in getting and maintaining an erection
Venderbos (2017) [22]	Sexual function, sexual bother	EPIC-26	No baseline; mean of 6.6 yr	AS group had the highest sexual function score, similar to general population. RP group was the worst

AS = active surveillance; AT = active treatment; CARES-SF = Cancer Rehabilitation Evaluation System; EBRT = external beam radiation therapy; EPIC = Expanded Prostate Cancer Index Composite; IIEF-5/6 = International Index of Erectile Function 5/6; PCSS = Prostate Cancer Symptom Scale; RP = radical prostatectomy; RT = radiotherapy; UCLA-PCI = University of California Los Angeles Prostate Cancer Index.

Of the ten papers comparing AS and RP, eight found that the AS group had better sexual function [9,18–20,22,28,30,31]. Bill-Axelson et al [30] reported that RP patients were over four times more likely to have difficulty achieving an erection (odds ratio = 4.19, 95% CI 2.63–6.68) and twice as likely to have reduced libido (odds ratio = 2.09, 95% CI 1.37–3.19) than AS counterparts. Johansson et al [31] also reported on the effect of erectile dysfunction, with 83% of RP patients feeling unable to satisfy their partners compared with 72% of AS patients (risk ratio = 1.19, 95% CI 1.03–1.36).

Nine studies compared sexual function in the AS and RT groups, of which five found no statistically significant difference [15,18,19,21,28]. Donovan et al [20] initially saw that AS patients report less sexual bother and better erections than RT patients, though this converged by the 6th year. Two studies reported better sexual function in the AS group, which appeared to persist for the duration of the follow-up period. At 10 yr, Thong et al [24] found better erectile function in the AS group than in the RT group (p < 0.01), while Fransson et al [23] showed less sexual bother (p = 0.011).

Seven studies compared AS and CaP-free men; two found no statistically significant difference [22,29], whereas the remaining five found worse sexual function in patients on AS [14,19,24,28,31]. Arredondo et al [14] concluded that while deterioration in erectile function is expected with normal ageing, this effect was more pronounced than expected in patients on AS. The other studies found that erectile function in AS groups, though better than that in treatment groups, was still worse than in CaP-free patients of similar age [19,24,28,31].

3.4. Discussion

Increasing experience using AS for the management of low-risk and favourable intermediate-risk CaP has reassured clinicians and patients of its safety if effective protocols are employed. However, such intensive strategies may be burdensome, particularly during long periods of close surveillance. The review results consistently suggest that extended duration of AS did not adversely affect HRQol and mental health PROMs when compared with patients undergoing curative treatment or controls. Similarly, urinary and sexual function PROMs remained favourable compared with definitive treatment and were comparable with general population norms.

3.4.1. HRQoL and mental health outcomes

Overall HRQoL outcomes were good in patients on AS, as was psychological well-being. All studies comparing AS with active treatment found no substantial or consistent difference in general HRQoL PROMs between groups. Safe AS depends on adherence to intensive surveillance, which may impact the psychological well-being of patients. Anxiety on surveillance has been documented, often related to the feelings of “not doing anything”, and may prompt a change to active treatment despite a lack of objective measures of disease progression [33]. These results further the existing knowledge, suggesting that long-term mental health PROMs are similar between AS and active treatment, perhaps due to increased confidence in AS by patients after the early periods of enrolment. Van Hemelrijck et al [6] assessed the reasons for discontinuing AS and evaluated patients who progressed to treatment without objective disease progression—a surrogate for anxiety-related discontinuation of AS. Their analysis noted that the proportion of these patients remained stable after 7 yr, suggesting that anxious patients were more likely to discontinue AS in the earlier years [6]. Furthermore, the impact of regular PSA testing, magnetic resonance imaging, and physical and digital rectal examinations of patients undergoing AS is substantial and may be the reason why a minority of AS patients expressed worse HRQoL than patients without cancer [14,32].

The influence of counselling and education is important, given that evidence points to better informed patients having less regret and higher HRQoL [15]. Given that negative psychological outcomes in cancer patients are related to the loss of control [33], empowering patients to make their own informed decisions using long-term data can lead to better HRQoL.

3.4.2. Urinary function

A clear advantage that AS has over active treatment, particularly RP, is preservation of continence. While this is expected, given the recognised adverse impact RP has on continence [3], it is the duration of this benefit that is particularly compelling. Results suggest that even after extended periods, continence is still considerably superior in AS patients to that in RP patients [9,18–22,28,30,31]. However, obstructive voiding symptoms were more common in patients on AS than in postoperative patients. LUTSs are common in the general population, with prevalence studies suggesting that an estimated 47–89% of male patients worldwide report at least one LUTS, representing a significant burden on patients [34]. AS patients in our review had comparable outcomes to the general population, suggesting that voiding LUTSs in these patients are not related to cancer but rather to the ageing process. Therefore, in general, prolonged AS is unlikely to result in worse urinary function when compared with the general population, but offers long-term benefits versus active treatment in the form of continence preservation. However, for younger patients on AS, for example, commencing in their 50s, even longer periods of follow-up are needed to adequately assess the effect of age-related change versus the impact of cancer.

3.4.3. Sexual function

Included studies showed sexual function in the AS group to be better than or comparable to that in the active treatment group. The observation of superior function highlights the benefit of extended AS strategies, even beyond 5 yr. Given the known impact of RP and RT on erectile function [3], particularly in the setting of neoadjuvant hormonal deprivation [35], this finding is expected. While these differences are intuitive on short-term follow-up, confirming superior sexual PROMs after longer periods of AS is important. Preservation of sexual function is a key motivation for instigating AS; however, all patients are exposed to age-related changes. Despite the advantages of AS compared with active treatment, studies comparing sexual function with that of CaP-free patients had mixed results. While two papers observed no statistically significant difference, five reported that sexual function was, at least numerically, worse in patients on AS than in CaP-free patients. This may be attributed to a loss of desire and frequency of intercourse related to psychological stressors from the cancer diagnosis. Moreover, a diagnosis of CaP may also put stress on a patient’s partner, as the majority of patients in the reviewed studies were not single.

3.4.4. Limitations

As “active surveillance” is a relatively recent term, many older studies encompassed all nontreatment patients as those under “watchful waiting”, which in contemporary practice has a different connotation. This was addressed through a review of the study methodology to ensure that “watchful waiting” referred to deferral of treatment in patients comparable with treatment groups, rather than a palliative approach in older, comorbid patients. The majority of included studies had cohort, retrospective design rather than randomised controlled trials; therefore, biases associated with study design can carry across to this review. Additionally, follow-up protocols varied considerably between studies and within patients in the same study. On occasion, studies used the same patient cohort, but these were still included as the individual study reported on different measures or subgroups of patients. Multiple different questionnaires were administered across the included papers; hence, our evidence synthesis should be considered general in nature, rather than specific to any single instrument. Furthermore, general HRQoL questionnaires such as SF-12 are unlikely to be sensitive enough to detect particular differences between AS and active treatment. Finally, narrative synthesis without a meta-analysis is prone to a reporting bias as we determined which results to present in the text; however, all included studies are listed in the tables, and we consider this method appropriate given the expected high clinical and methodological heterogeneity of included studies.

4. Conclusions

A previous systematic review found good HRQoL in patients on AS for periods <5 yr from diagnosis. After reviewing studies with longer-term follow-up, we report acceptable HRQoL outcomes in patients on AS, which are not meaningfully different from those in the general population. We further highlight the continuing preservation of urinary or sexual function when compared with intervention groups. These results indicate that in addition to being oncologically safe, AS is well tolerated with regard to long-term psychological well-being, functional outcomes, and overall quality of life. It should continue to be a recommended treatment for appropriately selected patients.

Data sharing:

Data underlying this study are available upon reasonable request to the corresponding author.