Timo Soeterik

54 CHAPTER 3 example. Hospital 1 had the lowest percentage of patients complying with 1-yr repeat biopsy (48%) but the highest percentage of patients with concordant PSA monitoring (83%). The rate of metastasis was relatively low in this hospital (2/248, 1%) in comparison to the other clinics. These findings suggest that deviating from the repeat biopsy protocol may be safe as long as PSA kinetics are monitored closely. The strengths of our study include the large sample size, a study population representing the real-world clinical situation, and evaluation of AS management strategies including a wide range of follow-up tests. Besides the strengths of the study, some limitations should be acknowledged. First, the retrospective nature of the study is a limitation and carries a risk of bias due to confounding by indication, especially concerning allocation of patients to a low- or high-intensity monitoring strategy. However, given the fact that it is possible that patients with more aggressive tumours received closer monitoring than patients with less aggressive tumours, we still found significant differences regarding rates of metastasis. Thus, this form of bias has not altered our conclusions. Second, we only evaluated whether AS follow-up was concordant with the PRIAS follow-up guidelines. We did not evaluate deviations from the PRIAS protocol regarding recommendations for discontinuation of AS (ie, Gleason sum score ≥7 on repeat biopsy or >2 cores positive). Therefore, we cannot assess the potential impact of this on our observed outcomes, as it was unclear which patients remained on AS despite Gleason upgrading or a substantial increase in tumour volume. However, we did collect data for an individual’s PSA course, which also provides important information on tumour aggressiveness. CONCLUSIONS Compliance with the PRIAS protocol in a real-world cohort is low: 43% of patients on AS in daily clinical practice receive PRIAS-concordant follow-up. Noncompliance with the PRIAS follow-up protocol was associated with a higher rate of metastasis among PRIAS- ineligible patients. Higher rates of metastasis were found in patients with discordant PSA monitoring, but not among patients not complying with the biopsy schedule. This suggests that low biopsy compliance seems to be “compensated” by high PSA compliance. The fact that discordant follow-up was not associated with a higher rate of metastasis among PRIAS-eligible patients suggests that less strict monitoring may be safe in this subgroup.