The comparative analyses of Grimm et al.  defined reasonable inclusion criteria for studies to be included in their analysis but the results for high-risk patients appear to be biased, because consecutive reports of certain institutions were included multiple times, some surgical and external-beam radiotherapy (EBRT) series were not included and other errors were made as outlined below.
Six series referred to as ‘EBRT + seeds + androgen-deprivation therapy (ADT)’ in Figure 3 of the article, merely reflect the experience of only two centres. Reference #58 seems to contain both prior reported #18 and #59. Reference #20 comes from the same institution, describes however treatment without ADT and only 127 patients were high-risk. For the remaining references #23 and #60 the included sample size might be reconsidered.
Series referred to as ‘EBRT + seeds’ in fact did not use EBRT at all (Reference #26 and #62), or only in a subset of patients (#22). One study actually used high-dose-rate (HDR) brachytherapy in conjunction with EBRT (#44) and at least in a subset of patients ADT was used (References #11, #12, #41, #42, #44, #58, #61, #62). Inclusion of multiple reports on the same patient cohorts necessitates recalculation also in this treatment group.
For the treatment group ‘HDR’, Reference #65 should refer to Deger et al. European Urology 2005, Reference #45 might be replaced by Reference #44 (which was referred to as EBRT + seeds) and in Reference #66 the 3.3-year (not 8-year) biochemical control was 88.5% (compare Figure 3).
For series referred to as ‘EBRT’, Reference #56 used RT doses <72 Gy and should not be eligible for the present analysis. The sample size of Reference #35 should be reconsidered, as some of the included patients underwent RT with <72 Gy. Reference #33 might be excluded as it remains unclear how many of the conventional high-risk patients (n = 272) received a total dose of ≥72 Gy. The stated sample size based on the MD Anderson
dose escalation trial should be reconsidered
(n = 53 not n = 1256, Figure 3, Reference #67). Importantly, results from other dose-escalation trials, meeting the inclusion criteria for the present analysis, were not included: The Medical Research Council RT01 trial reported a 5-year biochemical progression-free survival rate of 57% for 184 patients treated within the dose-escalated arm after neoadjuvant ADT  and the Dutch trial described a 6-year freedom-from-failure rate of 49%, based on 179 patients from the experimental arm, most of which underwent either short- or long-term ADT .
Importantly, series referred to as ‘Surgery’ should also be reconsidered with respect to multiple inclusion of the same patient cohorts (References #50 and #54 as well as #5 and #57). Reference #57 described a 15-year biochemical recurrence-free survival of 38%, which seems not accounted for in Figure 3 of the present article. It should be noted that inclusion of surgical series using prostatectomy specimen Gleason score for risk stratification, might introduce a bias when compared with RT studies (Reference #50, #52, #53, #54). For unclear reasons, surgical reports, meeting the inclusion criteria and reporting favourably comparing results were not included. Freedland et al.
 described a 15-year biochemical progression-free survival rate of 49% for 56 patients with clinical stage cT3a disease and Spahn et al.  described a 10-year biochemical progression-free survival rate of 52% for 712 patients with a PSA level of >20 ng/mL treated within a multi-institutional study.
We also wonder why the numbers of high-risk studies mentioned in Table 3 does not match the study count listed in Figure 3 for some of the treatment groups.
The authors should be acknowledged for attempting a comprehensive meta-analysis. However, consideration of the concerns raised above may lead to different results and conclusions regarding treatment of patients with high-risk prostate cancer. We fear that the article in its present form fails to provide ‘evidence-based prostate cancer treatment comparisons’, as stated by the authors.