Twenty-two studies (11,297 participants) were included, of which two were randomised trials. It was not stated how many of the remainder were prospective studies and how many were retrospective.
It should be noted that prognostic groups, and high and low doses, were defined differently in different studies.
Of the 11 studies addressing the outcome of local control, all showed improvement and 9 showed statistically-significant improvements associated with an increasing dose. When comparing the high- and low- dose results (expressed as percentages) for different prognostic groups with different follow-ups, the differences found in local control rates varied from +1 to +65%.
Of the 12 studies addressing the outcome of biochemical control, all showed statistically-significant improvements associated with an increasing dose. When comparing the high- and low-dose results (expressed as percentages) for different prognostic groups with different follow-ups, the differences found in biochemical control rates varied from +15 to +83%.
Of the 4 studies addressing the outcome of pathologic control, 2 showed statistically-significant improvements associated with an increasing dose. The effect sizes were not reported.
Of the 6 studies addressing the outcome of cause-specific survival, 3 showed statistically-significant improvements associated with an increasing dose, and all 6 showed some improvement. When comparing the high- and low-dose results (expressed as percentages) for different prognostic groups with different follow-ups, the differences found in cause-specific survival rates varied from +3 to +36%. Five out of the 9 studies that measured overall survival showed a statistically-significant improvement. The 2 RCTs (n=300 and n=200) showed a small and statistically non significant improvement in overall survival.
The authors stated that despite inconclusive results, patients with poor risk features were most likely to benefit from an increasing dose with respect to each outcome. An optimal RT dose, and the magnitude of the benefit of dose escalation, could not be identified.