Journal Article Page
European UrologyVolume 64, issue 2, pages e23-e42, August 2013
Timing of Curative Treatment for Prostate Cancer: A Systematic Review
Accepted 12 February 2013, Published online 26 February 2013, pages 204 - 215
Delaying definitive therapy unfavourably affects outcomes in many malignancies. Diagnostic, psychological, and logistical reasons but also active surveillance (AS) strategies can lead to treatment delay, an increase in the interval between the diagnosis and treatment of prostate cancer (PCa).
To review and summarise the current literature on the impact of treatment delay on PCa oncologic outcomes.
A comprehensive search of PubMed and Embase databases until 30 September 2012 was performed. Studies comparing pathologic, biochemical recurrence (BCR), and mortality outcomes between patients receiving direct and delayed curative treatment were included. Studies presenting single-arm results following AS were excluded.
Seventeen studies were included: 13 on radical prostatectomy, 3 on radiation therapy, and 1 combined both. A total of 34 517 PCa patients receiving radical local therapy between 1981 and 2009 were described. Some studies included low-risk PCa only; others included a wider spectrum of disease. Four studies found a significant effect of treatment delay on outcomes in multivariate analysis. Two included low-risk patients only, but it was unknown whether AS was applied or repeat biopsy triggered active therapy during AS. The two other studies found a negative effect on BCR rates of 2.5–9 mo delay in higher risk patients (respectively defined as any with T ≥2b, prostate-specific antigen >10, Gleason score >6, >34–50% positive cores; or D’Amico intermediate risk-group). All studies were retrospective and nonrandomised. Reasons for delay were not always clear, and time-to-event analyses may be subject to bias.
Treatment delay of several months or even years does not appear to affect outcomes of men with low-risk PCa. Limited data suggest treatment delay may have an impact on men with non–low-risk PCa. Most AS protocols suggest a confirmatory biopsy to avoid delaying treatment in those who harbour higher risk disease that was initially misclassified.
Delay of definitive therapy has an unfavourable impact on outcomes in different malignancies , , and . Because most malignant cells appear to grow exponentially with related systemic spread, we can reasonably assume that treatment delay of some tumours may risk missing the window of curability.
Prostate cancer (PCa) is generally considered a relatively slow-growing malignancy, with screening adding a considerable lead time  and . Delay between diagnosis and active therapy of PCa is often common. Unintended causes for this delay may include the need for pretreatment diagnostics or psychological and logistical reasons. Active surveillance (AS), as opposed to immediate definitive therapy, has garnered considerable support for several reasons in the treatment of low-risk disease. This strategy has introduced a new intended reason for delay in treatment  and . AS is designed to avoid unnecessary therapy in low-risk PCa, but identification of these tumours can be difficult and may miss the presence of occult higher risk disease.
We review the current medical literature to identify evidence whether treatment delay in PCa results in worse oncologic outcomes. Effects on functional outcome are not addressed.
2. Evidence acquisition
2.1. Study selection
We conducted a systematic review of the electronic databases PubMed and Embase according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis statement guidelines . Predefined search terms were used to identify articles describing the impact of a delay in treatment or extending the time interval between diagnosis and active curative therapy of PCa on pathologic, biochemical, and mortality outcomes. The literature search included papers published until 30 September 2012. Figure 1 presents the search strategy flowchart. Table 1 lists the terms used in the literature search. References of all retrieved articles were checked for additional cross-references.
|Delay [Title/Abstract]||AND||Prostate [Title/Abstract]||AND||Treatment [Title/Abstract]||AND||Outcomes [Title/Abstract]|
|Deferral [Title/Abstract]||Expectant management [Title/Abstract]||Recurrence [Title/Abstract]|
|Deferred [Title/Abstract]||Radiotherapy [Title/Abstract])||Survival [Title/Abstract]|
|Interval [Title/Abstract]||Watchful waiting [Title/Abstract]||Biochemical [Title/Abstract]|
|Radiation [Title/Abstract]||Mortality [Title/Abstract]|
|Prostatectomy [Title/Abstract]||Pathology [Title/Abstract]|
|Brachytherapy [Title/Abstract]||Results [Title/Abstract]|
|Active surveillance [Title/Abstract])|
|The same search strategy was used in Embase (“[Title/Abstract]” was replaced by “ti,ab”).|
2.2. Inclusion and exclusion criteria
We limited our search to full-text original articles published in English and available for review. Articles were independently assessed for eligibility using the following predefined criteria:
- • Study population: Patients diagnosed with PCa
- • Intervention: Surgery: open, laparoscopic, or robot-assisted radical prostatectomy; radiation therapy (RT): external-beam RT or brachytherapy
- • Study outcomes: Pathologic characteristics, biochemical recurrence rates (BCRs), or longer term outcomes such as distant metastasis (DM) rates or overall survival (OS) and cancer-specific survival (CSS).
The following studies were excluded: single-arm studies reporting outcomes without direct comparison group (eg, single-arm AS studies) or without subanalysis of the effect of delay, articles only assessing the effect of short intervals (days to weeks) between biopsies and surgery on perioperative outcomes (eg, blood loss, length of surgery, hospital stay), articles presenting overviews of wait times for curative therapy without any of the relevant outcome parameters, and articles presenting the effect of treatment delay on quality of life (QoL) only. Finally, papers outdated by more recent studies using overlapping data sets were excluded.
2.3. Data extraction
The following data were extracted from full-text articles by the first author: treatment modality, study design, selection and inclusion criteria, study details (delay definition, patient numbers, differences between groups, and time-to-event analysis details), specific reasons for treatment delay and switch of treatment, and relevant outcomes. Potential biases were analysed and recorded.
3. Evidence synthesis
3.1. Search results
Our literature search identified 17 original articles that were included in the qualitative analysis , , , , , , , , , , , , , , , , and . Differences in study design, inclusion criteria, and outcome parameters between studies and the lack of study population details prevented combining data for quantitative analysis. Table 2 presents an overview of the 17 included studies. Four relevant studies were outdated by more recent publications using the same patient study group , , , and .
|Study||Therapy||Study design||Period||Selection/inclusion criteria||Study details; delay definition, patient numbers, differences between groups, time-to-event analysis details||Specific reasons for treatment delay and switched treatment||Outcomes|
|Abern et al. ||RP||Retrospective cohort. SEARCH database||1988–2011||D’Amico low- (52%) and intermediate-risk (48%) groups||Delay definition: <3 mo vs 3–6 mo vs 6–9 mo vs >9 mo
Patient numbers: Total 1561. Low risk: 510, 225, 45, 33. Intermediate risk: 427, 250, 46, 25
Differences groups: Low- and intermediate-risk groups: Longer delay associated with surgery in later year and shorter follow-up. Intermediate group: Longer delay associated with lower rate in white race
Time-to-event analysis details: Outcome: Time between RP and BCR
|Not described||Low risk: Delay not related to BCR, ECE, PSM, pathologic upgrade
Intermediate risk: Delay >9 mo significantly related to BCR (p = 0.01) and PSM (p < 0.01)
Delay >9 mo associated with BCR also in subsets of men with biopsy Gleason ≤3 + 4 (HR: 2.51; p < 0.01), PSA ≤6.0 (HR: 2.82; p = 0.06), and low tumour volume (HR: 2.59; p = 0.06)
Crude BCR-free survival at 5 yr in intermediate-risk groups and subgroups was worse for >9 mo delay (37% vs 70%)
|Andrews et al. ||RT||Retrospective cohort||1981–2001||T stage 1–4 (12 patients stage 3 or 4)
PSA <10 to >20 (26 patients PSA >20)
Gleason 2–10 (7 patients Gleason 8–10)
Low risk 555; intermediate risk 515; high risk 252
|Delay definition: <3 mo vs 3–6 mo vs 6–9 mo vs >9 mo less vs more median delay of 3.1 mo
Patient numbers: Total 1322
Differences groups: Not described
Time-to-event analysis details: Outcome 5-yr rates of OS, CSS, DM, freedom from BCR from diagnosis and from median delay of 3.1 mo. Kaplan-Meier analyses also included
|Not described||No statistical differences in OS, CSS, DM; freedom from BCR
No difference freedom from BCR or DM between less or more median delay in low-, intermediate-, or high-risk groups. Time to treatment did not predict worse OS, CSS, DM, freedom from BCR
|Dall’era et al. ||RP||Retrospective cohort||1996–2008||T1–2
Positive biopsy cores ≤33%
Single cores ≤50% involved
Some other patients also on AS
|Delay definition: <6 mo vs after AS
Patient numbers: Total 1408. Low risk: 311; high risk: 1097. Direct group (median delay: 3 mo): 278 low risk, 1067 high risk. Delayed group (median delay: 18 mo, on AS): 33 low risk, 30 high risk
Differences groups: Delay group: Higher free PSA, more biopsy cores taken, higher presurgical Gleason score due to repeat biopsies during AS
Time-to-event analysis details: Outcome: Time between RP and BCR. Analysis in subgroup only
|AS protocol: DRE, PSA, TRUS, repeat biopsy
Not all patients underwent repeat biopsy (intervention based on biopsy in 23 of 63 patients)
|Rates Gleason ≥7 (compared with most recent biopsy), T3, PSM did not differ significantly from the immediate RP group
Multivariate: Treatment group not associated with Gleason upgrading, non–organ-confined disease or PSM
AS >18 mo vs <18 mo no significantly different outcomes
Low-risk direct comparable to low-risk delay; similar results for high risk
|Graefen et al. ||RP||Retrospective cohort||1992–2000||T1–T3a (4.3% T3)
PSA: 9.8% 0–4.0; 47.6% 4.1–10.0; 26.7% 10.1–20.0; 14.8% >20
Gleason 5–10 (3.7% 8–10)
|Delay definition: Time between diagnosis and treatment analysed as continuous variable. Mean delay: 62 d; median 54; range 5–518. Kaplan-Meier analysis between <31 d and >70 d
Patient numbers: Total 795
Differences groups: Not described
Time-to-event analysis details: Outcome: time between RP and BCR; not specified
|Not described||Time between biopsy and treatment not significantly related to BCR rate
Subgroup analysis <1 mo vs >4 mo even showed protective effect of delay
Even in high-grade disease subgroup no difference
|Holmström et al. ||RP||Retrospective cohort. NPCR Sweden||1997–2002||T1–2
≤70 yr of age
|Delay definition: Initial treatment labelled “expectant management” in database
Patient numbers: Total 2566. 2344 direct (median delay: 3.5 mo), 222 delayed (median delay: 19.2 mo)
Differences groups: Delay group: Older, lower PSA, lower clinical stage
Time-to-event analysis details: Only OS and CSS analysed; from moment of diagnosis
|Expectant management; no fixed protocol, no fixed triggers for switch treatment. Switch due to PSA in 50%; other signs progression in 9%; other causes in 39%||Upgrading Gleason lower after primary vs deferred RP (25% vs 38%; p < 0.001).
No difference in PSM or ECE
No differences in any one or more adverse characteristics
|Khan et al. ||RP||Retrospective cohort||1989–1994||T1c–T3a (35 patients T3a)
PSA 0.4–41.8 (55 patients >20)
Gleason 2–10 (43 patients 8–10)
|Delay definition: <60 d vs > 60 d. Also comparison between 61 and 90, 91 and 120, 121 and 150, >151 d
Patient numbers: Total 926; <60 d: 162, 61–90: 268, 91–120: 247, 121–150: 130, >151: 119
Differences groups: Longer delay: Lower stage (<60 d: 35% T1, 121–150 d: 48%, >151 d: 57%), lower Gleason (>151 d: 86% 2–6, <151 d: 65% 2–6), lower PSA
Time-to-event analysis details: Outcome: Time between RP and BCR
|Not described||Delay group: Higher BCR-free survival (>151 d group), lower rates positive LN, SVI (both 61–90 d and 91–120 d group)
No differences when stratified according to preoperative characteristics
|Korets et al. ||RP||Retrospective cohort||1990–2009||T1c–T3 (0.8% T3)
Gleason ≤7 to >7 (10.8% >7)
D’Amico low, intermediate, high risk (12.1% high risk)
|Delay definition: ≤60 d vs 61–90 d vs >90 d
Patient numbers: Total 1568; 1098 ≤60 d, 303 61–90 d, 167 >90 d
Differences groups: Longer delay: Lower Gleason, lower D’Amico risk group
Time-to-event analysis details: Outcome: Time between RP and BCR; not specified
Men with >180-d delay were individually reviewed to ensure that none had been placed on AS with delayed curative intervention protocols
|Not described, but AS patients were excluded||Delay >60 d not associated with adverse pathologic findings at RP, not with BCR
5-yr survival rates similar
Higher age, risk group, PSA, stage, Gleason and African American did give higher risk BCR
|Kwan et al. ||RT||Retrospective cohort||1993–2001||T1c–T3a
Low PSA: >20
261 patients low risk; 481 intermediate risk; 274 high risk
|Delay definition: Less than vs more than median delay of 3.7 mo
Patient numbers: Total 1024
Differences groups: Longer delay: Lower PSAs, lower Gleason scores, higher radiation dose (no p values)
Time-to-event analysis details: Outcome: Time between diagnosis and BCR
|Not described, but AS is mentioned as frequently applied strategy||No evidence that a longer time interval between diagnosis and radiation therapy was associated with poorer PSA control
Intermediate- and high-risk patients with longer intervals better biochemical control, but not in multivariate analysis
|Lee et al. ||RP||Retrospective cohort||2001–2004||PSA mean 7.2
Mean Gleason 6.3
Mean age 59 yr
|Delay definition: less or more than median delay of 56 d
Patient numbers: Total 169
Differences groups: Not described
Time-to-event analysis details: Outcome: Time between RP and BCR; not specified
|Not described||No significant differences in operative duration, intraoperative blood loss, nerve-sparing rate, transfusion, hospital time, PSM, postoperative complications, continence
Biopsy/treatment interval did not predict outcomes
|Nam et al. ||RP||Retrospective cohort||1987–1997||T1–2
PSA <4 to >20.0 (10.3% >20)
Gleason 2–10 (10.2% 8–10)
|Delay definition: <3 mo vs ≥3 mo. Median delay: 68 d
Patient numbers: Total 645; 456 <3 mo; 189 ≥3 mo
Differences groups: Delay patients: Higher PSA
Time-to-event analysis details: “Landmark analysis”; analysis starts at 1 yr after diagnosis for both direct and delayed group
Only patients without evidence of residual tumour following surgery were included. BCRs 0–1 yr after diagnosis were not included
|Not described.||Delay group had lower 10-yr BCR-free survival (61.3% vs 74.6%; p = 0.05)
Crude HR for developing BCR 1.58 (p = 0.04), adjusted (for grade, stage, PSA) HR 1.46 (p = 0.09) of delay compared with direct treatment; so after adjusting no longer significant
|Nguyen et al. ||RT||Retrospective cohort||1992–2001||T1c–T2
PSA <4 to >20 (9.6% >20)
Gleason <6–10 (9.6% 8–10)
|Delay definition: Less than vs more than median delay of 2.5 mo
Patient numbers: Total 460
Differences groups: In high-risk groups (defined as any of the following criteria: T ≥2b, PSA >10, Gleason >6, >34–50% positive biopsy cores) no differences between direct or delay. No analysis of other groups
Time-to-event analysis details: Outcome: Time between diagnosis and BCR
|Not described||Treatment delay independently predicted time to PSA failure following diagnosis for high-risk (p = 0.029) but not for low-risk (p = 0.31) patients
High risk: Delay <2.5 mo 55% 5-yr PSA-free survival, delay >2.5 mo 39% (p = 0.014).
Adjusted HR 1.08 per month delay
|O’Brien et al. ||RP||Retrospective cohort||1989–2009||D’Amico low risk PCa (T1c/T2a, PSA <10.0, Gleason ≤6)||Delay definition: < 6 mo vs ≥6 mo
Patient numbers: Total 1111; 1052 <6 mo; 59 ≥6 mo
Differences groups: Longer delay: Older, more African Americans, less T2a vs T1c
Time-to-event analysis details: Outcome: Time between RP and BCR; not specified
|Not described||Delay group: More high-grade (Gleason 7–10) disease (27% vs 47% upgrade), lower BCR-free survival (5% vs 12%)
PSA, clinical stage, and delay >6 mo significant predictors of BCR
No difference organ-confined disease, ECE, PSM, SVI, LN metastasis
|Phillips et al. ||RP/RT||Retrospective cohort. RP: 62%. RT: 38%||1991–2004||Stage T1 to ≥T2b (9% ≥T2b)
PSA <10 to >20 (3% >20)
Gleason 2–10 (3% 8–10)
59% low risk; 6% high risk
|Delay definition: ≤3 mo vs >3 mo
Patient numbers: Total 245. RP: 210 ≤3 mo; 35 >3 mo. RT: 100 ≤3 mo; 48 >3 mo
Differences groups: Delay group: Lower Gleason, more RT
Time-to-event analysis details: Outcome: Time between diagnosis and BCR
|Not described||BCR-free survival not associated with time to diagnosis/treatment
Outcomes for RP/RT not separately presented
|Sun et al. ||RP||Retrospective cohort. SEER database||1995–2005||T1–2
≥66 yr of age
|Delay definition: ≤3 mo vs >3 mo (>75th percentile delay)
Patient numbers: Total 17 153; 14 577 direct (median delay: 1.6 mo); 2576 delayed (median delay: 11.5 mo)
Differences groups: Delay group: Lower age, fewer whites, higher socioeconomic status, fewer married, more metropolitan area, more ADT. Multivariate analyses: Higher age and higher comorbidity less likely to have treatment delay. Patients of black race and unmarried more likely to undergo delayed RP. Patients with baseline erectile dysfunction less likely to have delayed RP. Delayed RP more frequently applied in more recent years
Time-to-event analysis details: Only OS and CSS analysed; t = 0 not described
|Not described||Delay not associated with pathologic upstaging or survival
Delay was associated with more postoperative urinary incontinence and erectile dysfunction and procedures
|van den Bergh et al. ||RP||Retrospective cohort. ERSPC section Sweden||1995–2009||T ≤2
PSA density <0.2
Gleason ≤3 + 3 = 6
1–2 positive biopsies
(PRIAS study inclusion criteria)
|Delay definition: Registered in database as WW or AS as initial treatment
Patient numbers: Total 227; 158 direct (median delay: 0.5 yr), 69 delayed (median delay: 2.6 yr)
Differences groups: Delay RP: Higher prostate volume, lower PSA density, lower positive biopsies, lower mm PCa tissue, higher percentage T1c, higher age, higher PSA at time of surgery
Time-to-event analysis details: Kaplan-Meier curves with both diagnosis as t = 0 and treatment as t = 0
|AS but without standardised follow-up; no fixed thresholds; regular PSA, repeat biopsies, mainly in men with rising PSA||No statistically significant differences between direct or delayed group
ORs for delayed vs direct RP: Gleason score >6 1.54 (p = 0.221), capsular penetration 2.45 (p = 0.091), tumour volume 0.099 ml (p = 0.155), BCR (0.185, p = 0.689)
|Vickers et al. ||RP||Retrospective cohort||1987–2002||T1c–T3 (3% T3)
PSA median 6.3
Gleason 6–10 (5% 8–10)
Age median 61 yr
|Delay definition: Outcomes presented per month extra delay after diagnosis
Patient numbers: Total 3149; 29% >3 mo delay, 4% >6 mo, 1% >9 mo
Differences groups: Not analysed
Time-to-event analysis details: BCR as binary end point at 3, 5, 8, and 10 yr after diagnosis
Patients treated >12 mo after diagnosis were excluded from study analysis (n = 36)
|Not described||No clear evidence of significant effect of delay to diagnosis on BCR at 3-, 5-, 8-, 10-yr follow-up|
|Warlick et al. ||RP||Retrospective matched case-control study||1995–2005||PSA 2.0–12.0
PSA density 0.03–0.30
1–2 positive cores
1–50% of core involved
|Delay definition: Delayed group were patients included in AS program
Patient numbers: Total 188; 150 direct (median delay: 3.0 mo), 38 delayed (median delay: 26.5 mo)
Differences groups: Delay group: Higher PSA density, lower maximal core involved with PCa
Time-to-event analysis details: No time-to-event analysis; chance of BCR-free survival as outcome
|AS program, switch if upgrading annual surveillance biopsy, 23% switched due to other reasons (PSA no trigger)||After adjusting for age and PSA density, the risk for incurable cancer did not differ significantly between direct and delayed treatment|
ADT = androgen-deprivation therapy; AS = active surveillance; BCR = biochemical recurrence; CSS = cancer/cause-specific survival; DM = distance metastasis; DRE = digital rectal examination; ECE = extracapsular extension; ERSPC = European Randomised Study of Screening for Prostate Cancer; HR = hazard ratio; LN = lymph node; NPCR = National Prostate Cancer Registry; OR = odds ratio; OS = overall survival; PCa = prostate cancer; PSA = prostate-specific antigen; PSM = positive surgical margin; RP = radical prostatectomy; RT = radiation therapy; SEER = Surveillance, Epidemiology and End Results; SVI = seminal vesical invasion; TRUS = transrectal ultrasound; WW = watchful waiting.
3.2. Study analysis
Thirteen studies presented outcomes after RP, 3 after RT, and 1 study combined RP and RT; none were found on brachytherapy. The reports include 34 517 patients with PCa diagnosed from 1981 to 2009. All studies excluded patients with adjuvant therapy. All were retrospectively designed. Summary data occasionally hindered detailed data extraction.
3.3. Patient population
Some of these studies focused on patients with low-risk PCa only; others had wider criteria for eligibility (clinical stage T3–4 disease, prostate-specific antigen [PSA] >20, Gleason scores 8–10). Treatment delay was often associated with older patient age or more favourable tumour characteristics (lower D’Amico risk groups, lower PSA, Gleason scores, clinical stage, less positive biopsies). In one study the delayed treatment group actually showed higher Gleason scores but explicitly stated it was due to the effect of repeat biopsies during AS . In general, the number of men receiving immediate treatment was much higher than the number of men undergoing delayed interventions. If described, the staging work-up such as the total number of biopsies differed between studies but not between the direct and delayed treatment groups within one study. Comorbidity status was presented in two studies  and , with significant differences between groups found in one .
3.4. Definition of interval and delay
Treatment delay was defined according to the following definitions: time interval shorter and longer than the median delay time (n = 4; range: 56 d to 3.7 mo), less than or more than a delay of 3 mo (n = 3), analysis per extra month delay (n = 3), <6 mo or >6 mo (n = 2), <60 d or >60 d (n = 2), or comparison of different 3-mo intervals (n = 2). Three studies considered men in the treatment delay category if they were labelled so in the study database; this mainly involved AS-like strategies , , and . One study used treatment delay as a continuous variable for the main analysis (other studies used this variable in subanalyses).
It was unclear whether delay was intended or unintended in 11 studies. Two studies reported that the delay group was followed with a fixed AS protocol  and , two reported on expectant management without fixed criteria  and , and one stated that AS was frequently applied . Korets et al. excluded men on AS .
3.5. Definition of outcome/end point
The selected studies used one or more of the following end points as the main outcome variables: PSA BCR rates after radical prostatectomy (RP) or RT (n = 15), pathologic characteristics after RP (n = 8), OS and/or CSS (n = 3), and lymph node or DM rates (n = 2). One study used a previously constructed nomogram predicting the chance of incurable cancer (<75% chance of remaining BCR free 10 yr postsurgery) as the outcome . This was also the only study that did not include any form of a time-to-event analysis. Fourteen studies presented a time-to-event analysis with BCR as the outcome. Of these, seven studies used the time between RP and BCR as the outcome time parameter, four studies used the time between diagnosis and BCR, and one included both. The potential effect on outcomes of this important difference in time-to-event analyses is discussed later. The remaining two studies used other analyses: a landmark analysis in which analysis started at 1 yr after diagnosis for both direct and delayed treatment or binary end point BCR yes or no at a fixed period after treatment for direct or delayed treatment.
3.6. Observations of outcome
Seven studies found no significant impact of treatment delay on oncologic outcomes , , , , , , and . Two included only men with low-risk disease  and ; the other five included some men with higher risk PCa , , , , and . The longest treatment delay interval was 2.1 yr . Four studies initially suggested an association between treatment delay and outcomes following univariate analysis but not multivariate analysis , , , and . Two studies found lower rates of upstaging in the group with delayed therapy  and .
Four studies found that treatment delay had an unfavourable impact; however, none found a relationship of delay with DM, OS, or CSS. Abern et al. noted that if RP was delayed >9 mo, BCR rates and positive surgical margins rates were higher among men with intermediate-risk disease but not among men with low-risk PCa (D’Amico risk classification) . Specific reasons for delay or triggers to switch to active therapy were not presented. Holmström et al. found that patients labelled initially as “expectant management” in the Swedish National Prostate Cancer Registry had higher Gleason scores in their surgical specimens when compared with men who had undergone surgery at the time of diagnosis (delay of 19.2 mo vs 3.5 mo) . All patients had stage T ≤2 disease, PSA <20 ng/ml, and Gleason score ≤6 at diagnosis. Treatment delay resulted from AS in these patients, but the reasons that triggered deferred radical surgery were known (eg, repeat biopsies). Nguyen et al. found that treatment delay >2.5 mo independently predicted time to PSA failure after RT in patients labelled high risk (defined as any of the following criteria: T ≥2b, PSA >10, Gleason score >6, >34–50% positive biopsy cores) . The authors did not provide reasons for delay or triggers to switch to active therapy. Finally, O’Brien et al. found that in a cohort of men with low-risk disease, a delay in performing RP >6 mo after diagnosis was associated with a higher frequency of disease upgrading (47% vs 27% upstaging to Gleason score 7–10) and lower PSA progression-free survival rates . Again, no specific reasons for treatment delay or triggers to undergo active therapy were provided.
Besides the effect of treatment delay, most studies included information on the known risk factors for worse pathologic and BCR outcomes after RP or RT.
3.7. Study evaluation
This review included studies monitoring the effect of PCa treatment delay on oncologic outcome. Most did not find any differences in outcomes between patients who received immediate or delayed RP or RT according to different definitions of delay. All are limited in varying degrees by suboptimal study design including unaccounted potential biases in the time-to-event analysis. Four studies did find a significant relation in multivariate analysis , , , and , of which two included intermediate- to high-risk PCa. One study found an unfavourable effect of delay of RP >9 mo in intermediate-risk disease groups, and one found worse outcomes in non–low-risk disease of RT >2.5 mo after diagnosis  and . These two studies suggest that in men with non–low-risk disease, oncologic outcomes may be compromised if treatment is delayed. This finding is balanced, however, by five studies that also included intermediate- to high-risk men that did not show an unfavourable impact of treatment delay on outcomes , , , , and .
The two other studies that showed a significantly unfavourable outcome associated with treatment delay included low-risk patients only  and . Selection bias in these two studies very likely accounts for this contrary conclusion. In the article by Holmström et al., patients in the delayed treatment group (19.2 mo vs 3.5 mo) are labelled as “expectant management” . The reasons why patients switched to active therapy were unclear. We hypothesise that men who initially chose expectant management may have subsequently developed unfavourable disease characteristics on repeat biopsies that resulted in their selection for RP. A similar selection bias might explain the large differences in Gleason upgrading between men undergoing direct RP and delayed RP (27% vs 47% for <6 mo vs ≥6 mo delay) in the study presented by O’Brien et al. . No reasons for delaying treatment were provided in this study that included recently treated patients with low-risk disease. Again, men may have been included who underwent repeat prostate biopsies as part of an AS-like protocol resulting in pathologic risk reclassification and deferred treatment. Dall’Era et al. specifically cite this potential bias when analysing outcomes . They used the most recent prostate biopsies when comparing outcomes between men undergoing immediate or delayed therapy. Treatment delay did not affect men with low-risk tumours in this study. This selection bias, as hypothesised earlier, probably does not explain the negative effect of treatment delay on non–low-risk disease because these patients are generally considered unsuitable for AS.
The contradictory protective effect of treatment delay as presented in two studies is also likely to be caused by selection bias  and . Tumours with more favourable characteristics are often treated later and show more favourable outcomes.
Differences in comorbidity may also introduce potential biases between patients receiving direct versus delayed treatment. Patients with more comorbidity may be more likely to be selected for initial expectant management. This may artificially favour BCR or mortality outcomes after deferred treatment because these patients will succumb to non-PCa causes earlier than healthy men. This bias was noted in both studies presenting comorbidity scores.
Most studies used pathologic characteristics and BCR as the surrogate end points rather than DM, CSS, or OS. However, these surrogate end points have been found to show a significant correlation with metastatic and mortality rates .
Unfortunately, the quality of the evidence in these studies is low. The retrospective design and lack of randomisation cause these studies to be rated level ≥3. Furthermore, only a few of these studies included patients with unfavourable (intermediate- or high-risk) disease. Most studies show evidence of a selection bias in which patients with more unfavourable risk disease and who are younger are treated earlier than older men and those men with comorbidities. Most studies failed to provide information why patients were treated later or what triggers were used to initiate delayed therapy. It is therefore not possible to generalise findings from single studies or to draw definitive conclusions on the timing of curative treatment for PCa.
Although most included studies failed to reject the null hypothesis that there is no difference in outcomes between immediate and delayed treatment, this does not necessarily imply that we should accept the hypothesis. The power of most studies to reject the null hypothesis may have been too low.
3.8. Time-to-event analyses
Fourteen studies presented a time-to-event analysis with BCR as the outcome. Different starting points (t = 0 at the moment of diagnosis versus t = 0 at the moment of treatment) in these analyses may lead to considerable bias. Only 4 of 14 studies identified this potential bias in choosing a specific time-to-event analysis , , , and .
For example, consider two patients with occult metastases at diagnosis that will grow to clinically detectable BCR, independent of the specific local treatment. The first receives RP at 3 mo after diagnosis, and the second patient at 9 mo after diagnosis.
If the moment of RP is used as the starting point of the time-to-event analysis and BCR occurs at 12 mo after diagnosis, BCR will occur at 9 mo and 3 mo after RP, respectively. This artificially increased time to recurrence may lead to a bias with more favourable results in the early treatment group.
If the moment of diagnosis is used as the starting point of the time-to-event analysis and PSA would start to rise at 6 mo in both patients, the problem is that the first patient shows BCR at 3 mo after diagnosis, whereas the second is not even at risk for BCR. Such an analysis may be biased in favour of delayed surgery.
An alternative approach is to use a landmark analysis, as applied in one of the included studies . Time to BCR is measured from a fixed starting point, such as 12 mo, after diagnosis. Patients are excluded if they are treated after the landmark time but also if they develop BCR before the landmark. This method may bias results in favour of early surgery because patients who develop disease recurrence rapidly would be excluded if they underwent early surgery but not if surgery was delayed.
BCR may also be used as a binary end point (yes vs no) at different time points after treatment . This method may introduce the same bias in which early treatment leads to more favourable results due to an artificial longer time until BCR. If the patients were treated at 3 and 9 mo after diagnosis and had a BCR at 18 mo after diagnosis but were analysed at 12 mo after treatment, the patient treated early would not show BCR, whereas the patient treated later would show BCR at this point. In analyses at ≥2 yr, however, this bias would be drastically reduced. Therefore the bias introduced in this last analysis may be the smallest.
3.9. Selection for treatment
The spectrum of clinical progression among men with PCa is very wide and depends on both disease-specific and patient-specific characteristics. Patients with high-grade PCa show an aggressive disease course with rapid spread and higher disease-specific mortality. These malignancies deserve prompt diagnosis and treatment. Vickers et al. used data from the Scandinavian Prostate Cancer Group-4 trial and found that radical treatment is most justified in high-risk patients (Gleason 8 or Gleason 7 with clinical stage II PCa) .
PCa with favourable disease features at the time of diagnosis, however, has been shown to have a relatively benign course in most cases, even when diagnosed clinically . Screening for PCa results has advanced the date of diagnosis up to 10–15 yr for men with low-risk tumours . Radical treatment for men with these malignancies may not lead to lower rates of metastases or lower PCa mortality, especially in older men or men with comorbidities. The results of the Prostate Cancer Intervention versus Observation Trial support this concept . In this trial 731 men with localised PCa were randomised to either observation or RP. No differences were found regarding OS or CSS through 12 yr of follow-up. Patients in the intervention group received RP a median of 35 d after randomisation. A subanalysis of the Prostate Cancer Intervention versus Observation Trial study showed that in men with intermediate- and high-risk disease, OS was improved with RP compared with observation .
3.10. Clinical implications
A delay of several months or even years from diagnosis to definitive therapy in men with low-risk PCa is very unlikely to have any unfavourable impact on morbidity or PCa mortality. This is not surprising considering the favourable natural history of low-risk PCa with long lead times. Most of the studies included in this review are in line with this hypothesis. On the other side of the disease spectrum, however, as also found in this review, the risk of missing the window of curability due to treatment delay is a realistic concern. In patients with high-risk or even intermediate-risk disease, a delay of 2.5–9 mo might unfavourably affect outcomes after radical treatment, although the data on this point were mixed. Due to the limitations of the studies included in this review, it is not possible to reach any conclusions on more specific risk subgroups such as men with intermediate- and high-risk disease.
Figure 2 presents schematically the effect of a delay between diagnosis and treatment in men with low-risk and higher risk disease. Treatment delay may risk missing a window of curability in men with higher risk disease but may not have an impact on results among men with lower risk tumours.
For men with intermediate- or high-risk disease, 3 mo appears to be an acceptable period in which treatment choices should be made and additional diagnostic studies and treatment performed. Waiting lists for these patients should ideally not exceed this period.
Some of the included studies specifically mention AS. Others do not present specific reasons for delay, but they can be considered to have unintended reasons for delay instead of AS. Although a similar delay would biologically have the same effect independent of the specific (intended or unintended) reason for delay, findings from this review may not be directly applicable to an AS situation.
The number of men on AS who will experience a pathologic upgrading following repeat biopsies may be as high as 28% . Men thought to have low-risk disease at diagnosis may actually harbour higher risk PC. This upgrading is most likely the result of initial undersampling and is described as risk reclassification rather than disease progression . A tumour initially assumed to be low risk but was reclassified as a higher risk tumour later may have a worse prognosis if treatment is delayed >1 yr . Based on this information, clinicians adopting AS protocols may wish to perform confirmatory studies earlier rather than later  and . This may especially hold true for patients on AS with more positive biopsy cores and higher PSA density at diagnosis . Most AS protocols recommend repeat biopsies within 1 yr of the initial diagnosis .
3.11. Additional considerations
In the United Kingdom the Prostate Testing for Cancer and Treatment study is being conducted . Patients are randomised for treatment including AS. This study may also provide answers on the effect of delay due to AS. Simulation models of differences in outcomes between direct and delayed treatment may provide an alternative to randomised studies. For example, Xia et al. modelled PCa mortality after direct RP versus RP after initial AS . It was found that AS may result in a small decline in CSS but large benefits in QoL in men with low-risk PCa.
We did not consider the potentially unfavourable effect of delayed treatment on QoL. Sun et al. noted more postoperative urinary incontinence and erectile dysfunction and associated procedures among men undergoing delayed treatment . Radomski et al. found that the treatment delay due to AS did not increase the risk of urinary incontinence , but Fujita et al. noted an unfavourable effect on erectile function possibly because of repeat biopsies . Waiting for radical treatment may also cause distress, although many patients included in an AS protocol did not show high levels of anxiety  and . Finally, other potentially negative effects of delaying therapy on physical or psychological domain outcomes are unknown. For example, delaying surgery may miss the opportunity to perform nerve-sparing surgery. Disease upgrading or upstaging after treatment following initial AS may lead to a greater need for secondary treatments, which may have a further unfavourable effect on QoL. The European and American guidelines do not specifically mention the timing of curative therapy for PCa  and .
Among men with low-risk PCa, a treatment delay of several months or even years does not appear to compromise long-term oncologic results following definitive treatment. Limited data suggest that treatment delay may compromise oncologic outcomes in men with non–low-risk PCa. The overall quality of evidence of included studies was ≥3.
Curative treatment for low-risk PCa, if indicated at all, is not an urgent matter. This creates a window for additional risk stratification including initial AS. Patients deciding whether to commit to an AS protocol should consider a confirmatory biopsy to decrease the possibility that they may harbour occult higher risk disease.
Author contributions: Roderick C.N. van den Bergh had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: van den Bergh, van der Poel.
Acquisition of data: van den Bergh.
Analysis and interpretation of data: van den Bergh, van der Poel, Vickers, Albertsen.
Drafting of the manuscript: van den Bergh.
Critical revision of the manuscript for important intellectual content: van den Bergh, Albertsen, Bangma, Freedland, Graefen, Vickers, van der Poel.
Statistical analysis: van den Bergh, Albertsen, Bangma, Freedland, Graefen, Vickers, van der Poel.
Obtaining funding: None.
Administrative, technical, or material support: None.
Supervision: van der Poel, Vickers, Albertsen.
Other (specify): None.
Financial disclosures: Roderick C.N. van den Bergh certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor: None.
-  G.S. Kulkarni, D.R. Urbach, P.C. Austin, N.E. Fleshner, A. Laupacis. Longer wait times increase overall mortality in patients with bladder cancer. J Urol. 2009;182:1318-1324
-  J. Huang, L. Barbera, M. Brouwers, G. Browman, W.J. Mackillop. Does delay in starting treatment affect the outcomes of radiotherapy? A systematic review. J Clin Oncol. 2003;21:555-563
-  N. O‘Rourke, R. Edwards. Lung cancer treatment waiting times and tumour growth. Clin Oncol (R Coll Radiol). 2000;12:141-144
-  F.H. Schröder, J. Hugosson, M.J. Roobol, et al. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med. 2012;366:981-990
-  G. Draisma, R. Boer, S.J. Otto, et al. Lead times and overdetection due to prostate-specific antigen screening: estimates from the European Randomised Study of Screening for Prostate Cancer. J Natl Cancer Inst. 2003;95:868-878
-  M.A. Dall’era, P.C. Albertsen, C. Bangma, et al. Active surveillance for prostate cancer: a systematic review of the literature. Eur Urol. 2012;62:976-983
-  R.C. van den Bergh, S. Roemeling, M.J. Roobol, et al. Outcomes of men with screen-detected prostate cancer eligible for active surveillance who were managed expectantly. Eur Urol. 2009;55:1-8
-  A. Liberati, D.G. Altman, J. Tetzlaff, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6:e1000100
-  M.R. Abern, W.J. Aronson, M.K. Terris, et al. Delayed radical prostatectomy for intermediate-risk prostate cancer is associated with biochemical recurrence: possible implications for active surveillance from the SEARCH database. Prostate. 2013;73:409-417
-  S.F. Andrews, E.M. Horwitz, S.J. Feigenberg, et al. Does a delay in external beam radiation therapy after tissue diagnosis affect outcome for men with prostate carcinoma?. Cancer. 2005;104:299-304
-  M.A. Dall’Era, J.E. Cowan, J. Simko, et al. Surgical management after active surveillance for low-risk prostate cancer: pathological outcomes compared with men undergoing immediate treatment. BJU Int. 2011;107:1232-1237
-  M. Graefen, J. Walz, K.H. Chun, T. Schlomm, A. Haese, H. Huland. Reasonable delay of surgical treatment in men with localized prostate cancer—impact on prognosis?. Eur Urol. 2005;47:756-760
-  B. Holmström, E. Holmberg, L. Egevad, et al. Outcome of primary versus deferred radical prostatectomy in the National Prostate Cancer Register of Sweden Follow-Up Study. J Urol. 2010;184:1322-1327
-  M.A. Khan, L.A. Mangold, J.I. Epstein, J.K. Boitnott, P.C. Walsh, A.W. Partin. Impact of surgical delay on long-term cancer control for clinically localized prostate cancer. J Urol. 2004;172:1835-1839
-  R. Korets, C.M. Seager, M.S. Pitman, G.W. Hruby, M.C. Benson, J.M. McKiernan. Effect of delaying surgery on radical prostatectomy outcomes: a contemporary analysis. BJU Int. 2012;110:211-216
-  W. Kwan, T. Pickles, G. Duncan, M. Liu, C. Paltiel. Relationship between delay in radiotherapy and biochemical control in prostate cancer. Int J Radiat Oncol Biol Phys. 2006;66:663-668
-  D.K. Lee, V. Allareddy, M.A. O’Donnell, R.D. Williams, B.R. Konety. Does the interval between prostate biopsy and radical prostatectomy affect the immediate postoperative outcome?. BJU Int. 2006;97:48-50
-  R.K. Nam, M.A. Jewett, M.D. Krahn, et al. Delay in surgical therapy for clinically localized prostate cancer and biochemical recurrence after radical prostatectomy. Can J Urol. 2003;10:1891-1898
-  P.L. Nguyen, R. Whittington, S. Koo, et al. The impact of a delay in initiating radiation therapy on prostate-specific antigen outcome for patients with clinically localized prostate carcinoma. Cancer. 2005;103:2053-2059
-  D. O’Brien, S. Loeb, G.F. Carvalhal, et al. Delay of surgery in men with low risk prostate cancer. J Urol. 2011;185:2143-2147
-  J.J. Phillips, M.C. Hall, W.R. Lee, P.E. Clark. Does a delay in initiating definitive therapy affect biochemical recurrence rates in men with clinically localized prostate cancer?. Urol Oncol. 2007;25:196-200
-  M. Sun, F. Abdollah, J. Hansen, et al. Is a treatment delay in radical prostatectomy safe in individuals with low-risk prostate cancer?. J Sex Med. 2012;9:2961-2969
-  R.C. van den Bergh, E.W. Steyerberg, A. Khatami, et al. Is delayed radical prostatectomy in men with low-risk screen-detected prostate cancer associated with a higher risk of unfavorable outcomes?. Cancer. 2010;116:1281-1290
-  A.J. Vickers, F.J. Bianco Jr., S. Boorjian, P.T. Scardino, J.A. Eastham. Does a delay between diagnosis and radical prostatectomy increase the risk of disease recurrence?. Cancer. 2006;106:576-580
-  C. Warlick, B.J. Trock, P. Landis, J.I. Epstein, H.B. Carter. Delayed versus immediate surgical intervention and prostate cancer outcome. J Natl Cancer Inst. 2006;98:355-357
-  S.A. Boorjian, F.J. Bianco Jr., P.T. Scardino, J.A. Eastham. Does the time from biopsy to surgery affect biochemical recurrence after radical prostatectomy?. BJU Int. 2005;96:773-776
-  S.J. Freedland, C.J. Kane, C.L. Amling, et al. Delay of radical prostatectomy and risk of biochemical progression in men with low risk prostate cancer. J Urol. 2006;175:1298-1302
-  A. Khatami, J.E. Damber, P. Lodding, C.G. Pihl, J. Hugosson. Does initial surveillance in early prostate cancer reduce the chance of cure by radical prostatectomy? A case control study. Scand J Urol Nephrol. 2003;37:213-217
-  A. Shibata, U.M. Mohanasundaram, M.K. Terris. Interval from prostate biopsy to radical prostatectomy: effect on PSA, Gleason sum, and risk of recurrence. Urology. 2005;66:808-813
-  A.V. D’Amico, J.W. Moul, P.R. Carroll, L. Sun, D. Lubeck, M.H. Chen. Intermediate end point for prostate cancer-specific mortality following salvage hormonal therapy for prostate-specific antigen failure. J Natl Cancer Inst. 2004;96:509-515
-  A. Vickers, C. Bennette, G. Steineck, et al. Individualized estimation of the benefit of radical prostatectomy from the Scandinavian Prostate Cancer Group randomized trial. Eur Urol. 2012;62:204-209
-  G.L. Lu-Yao, P.C. Albertsen, D.F. Moore, et al. Outcomes of localized prostate cancer following conservative management. JAMA. 2009;302:1202-1209
-  T.J. Wilt, M.K. Brawer, K.M. Jones, et al. Radical prostatectomy versus observation for localized prostate cancer. N Engl J Med. 2012;367:203-213
-  R.A. Godtman, E. Holmberg, A. Khatami, J. Stranne, J. Hugosson. Outcome following active surveillance of men with screen-detected prostate cancer. Results from the Göteborg randomised population-based prostate cancer screening trial. Eur Urol. 2013;63:101-107
-  J.I. Epstein, Z. Feng, B.J. Trock, P.M. Pierorazio. Upgrading and downgrading of prostate cancer from biopsy to radical prostatectomy: incidence and predictive factors using the modified Gleason grading system and factoring in tertiary grades. Eur Urol. 2012;61:1019-1024
-  V. Iremashvili, L. Pelaez, M. Manoharan, M. Jorda, D.L. Rosenberg, M.S. Soloway. Pathologic prostate cancer characteristics in patients eligible for active surveillance: a head-to-head comparison of contemporary protocols. Eur Urol. 2012;62:462-468
-  R.K. Berglund, T.A. Masterson, K.C. Vora, S.E. Eggener, J.A. Eastham, B.D. Guillonneau. Pathological upgrading and up staging with immediate repeat biopsy in patients eligible for active surveillance. J Urol. 2008;180:1964-1968
-  King AC, Livermore A, Laurila TA, Huang W, Jarrard DF. Impact of immediate TRUS rebiopsy in a patient cohort considering active surveillance for favorable risk prostate cancer. Urol Oncol. In press.
-  M. Bul, X. Zhu, R. Valdagni, et al. Active surveillance for low-risk prostate cancer worldwide: the PRIAS study. Eur Urol. 2013;63:597-603
-  J. Donovan, N. Mills, M. Smith, et al. Quality improvement report: improving design and conduct of randomised trials by embedding them in qualitative research: ProtecT (prostate testing for cancer and treatment) study. Commentary: presenting unbiased information to patients can be difficult. BMJ. 2002;325:766-770
-  J. Xia, B.J. Trock, M.R. Cooperberg, et al. Prostate cancer mortality following active surveillance versus immediate radical prostatectomy. Clin Cancer Res. 2012;18:5471-5478
-  L. Radomski, J. Gani, G. Trottier, A. Finelli. Active surveillance failure for prostate cancer: does the delay in treatment increase the risk of urinary incontinence?. Can J Urol. 2012;19:6287-6292
-  K. Fujita, P. Landis, B.K. McNeil, C.P. Pavlovich. Serial prostate biopsies are associated with an increased risk of erectile dysfunction in men with prostate cancer on active surveillance. J Urol. 2009;182:2664-2669
-  R.E. Gray, M.I. Fitch, C. Phillips, M. Labrecque, L. Klotz. Presurgery experiences of prostate cancer patients and their spouses. Cancer Pract. 1999;7:130-135
-  R.C. van den Bergh, M.L. Essink-Bot, M.J. Roobol, et al. Anxiety and distress during active surveillance for early prostate cancer. Cancer. 2009;115:3868-3878
-  A. Heidenreich, J. Bellmunt, M. Bolla, et al. EAU guidelines on prostate cancer. Part 1: screening, diagnosis, and treatment of clinically localised disease. Eur Urol. 2011;59:61-71
-  I. Thompson, J.B. Thrasher, G. Aus, et al. Guideline for the management of clinically localized prostate cancer: 2007 update. J Urol. 2007;177:2106-2131
a University Medical Centre Utrecht, Utrecht, The Netherlands
b Netherlands Cancer Institute, Amsterdam, The Netherlands
c University of Connecticut Health Centre, Farmington, CT, USA
d Erasmus University Medical Centre, Rotterdam, The Netherlands
e Durham VA Medical Centre and Duke University, Durham, NC, USA
f Martini-Clinic, Prostate Cancer Centre, University Hospital Hamburg-Eppendorf, Germany
g Memorial Sloan-Kettering Cancer Centre, New York, NY, USA
☆ Please visit www.eu-acme.org/europeanurology to read and answer questions on-line. The EU-ACME credits will then be attributed automatically.
© 2013 European Association of Urology, Published by Elsevier B.V.
Recommend this article
Currently this article has a rating of 5. Please log in to recommend it.