Despite concerns that post-RP PS may be irreversible [5, 6], our results demonstrate that 59.4% and 60.2% of patients returned to their baseline stretched and flaccid PL, respectively, suggesting that PS is not a permanent consequence. Additionally, our study demonstrated that younger age, high preoperative erectile function, and consistent PDE5i use were independent predictors of complete PL recovery. It appears that these therapeutic effects of PDE5i were only appreciated for long-term users, and in individuals with medium preoperative SHIM scores. Neither prostate anatomy nor metabolic risk factors were significant predictors of complete PL recovery.
Older patients were less likely to reach complete PL recovery when controlled for other variables. This finding agrees with the results of a recent report that analyzed postoperative EF and continence, confirming the correlation between senility and delayed recovery from surgical trauma [17]. Among other variables associated with complete recovery was a high baseline preoperative EF, which has also been previously identified as a protective factor in PL recovery [8, 9]. Similarly, our multivariate logistic regression model revealed that high preoperative SHIM, a marker for intact erectile function, was an independent predictor of complete PL recovery.
NS techniques have frequently been reported as an independent predictor of reduced PL loss [5]. One study demonstrated that patients undergoing NS RP had no changes in penile measurements at 6-months postoperative visit when compared to preoperative baseline [9]. Our univariate logistic regression showed that both types of NS RP approach were marginally predictive of complete PL recovery (OR 2.511; 95%CI 0.969–6.502; p = 0.058). The lack of association at the multivariate level could be related to the distribution of our study cohort where 96.6% of men were operated on using a NS technique - either IF or AIR. No difference was appreciated in PL recovery between these two types of NS techniques (p = 0.24) although AIR was found to be a superior NS technique when compared to IF in terms of preserving sexual function in our earlier report [14].
Prostate anatomic factors have been implicated in the pathophysiology of PL shortening because a part of prostatic urethra will be resected during surgery [13]. However, recent studies demonstrated that prostate size and weight were not correlated with PS, suggesting that the length of prostatic urethra is fixed at the urogenital diaphragm [2, 5]. However, no study has investigated the effects of post-surgical anatomic alteration using more than two parameters. The incorporation of comprehensive anatomic parameters in our study has shown that prostate length (4.4 cm vs. 4.3 cm; p = 0.93), volume (63.7 ml vs. 64.1 ml; p = 0.46), and weight (49.1 g vs. 50.7 g; p = 0.40) were not significantly different between the CR and the IR groups, demonstrating that prostate size itself does not affect PL recovery.
We also investigated the effects of metabolic derangement and CV risk factors on PL recovery. In our findings, the CR group had a smaller proportion of men affected by hypertension and CAD when compared to the IR group (40.3% vs. 50.8%; p = 0.047 and 6.9% vs. 13.7%; p = 0.026, respectively), but the associations with CV factors did not remain significant in logistic regression analyses, suggesting that they might be an age-related phenomenon. Similarly, the proportion of diabetic patients who reached a complete recovery were similar in both the CR and the IR groups (13.0% vs. 12.1%; p = 0.877).
The literature suggests that patients can improve PL recovery by consistently taking PDE5i. Recently, Brock et al. reported on 423 patients who were randomized to receive 1) 5-mg tadalafil once daily (OaD), 2) 20-mg tadalafil on-demand (“pro renata”, PRN), or 3) placebo [12]. The authors found that at the end of 9 months, PS was significantly less for patients treated with tadalafil OaD than those treated with placebo, with a least-squares mean difference in stretched PL change from preoperative PL of 4.1 mm (95% CI, 0.4–7.8; P = 0.032). No significant difference in PL change was observed between tadalafil PRN and placebo.
Our results are in line with previous studies and support the routine use of PDE5i after RP. Our multivariate logistic analysis demonstrated that consistent use of PDE5i was predictive of PL recovery (OR 1.998; 95%CI 1.166–3.425; p = 0.012). Moreover, when the study population was analyzed according to PDE5i use, the difference in penile recovery was only significant at 12-month postoperative visit (Fig. 2).
Our study is unique in that it not only attempted to describe longitudinal patterns of PL recovery between the CR and the IR groups, but also analyzed comprehensive factors associated with complete recovery within a one-year follow-up. Our study has confirmed many previous findings, and also has tested claims that were largely speculative. Importantly, our study further supports the long-term efficacy of consistent PDE5i use for PL recovery. This benefit of PDE5i in this patient population was first proved in the recent clinical trial that demonstrated that tadalafil has therapeutic potential beyond its conventional indication. To add to this, our study has found that the PL protective effect is not just unique to patients using tadalafil, but also to patients using sildenafil and vardenafil, thereby better representing the current clinical practice where different types of PDE5i are utilized in a heterogeneous patient population. For example, our study population is not limited to patients with Gleason 6 disease and PSA < 10 ng/mL, two of the inclusion criteria listed for the clinical trial. To our knowledge, our study is also the first to describe the temporal patterns of PDE5i efficacy and identify SHIM scores that would benefit from PDE5i use for PL recovery, defining a target population for intervention. The results of our study are also important to patients, as PL shortening has been consistently associated with a decreased quality of life, including reduced self-esteem, interference with close relationships, and ultimately treatment regret [18, 19]. The potential of PDE5is to potentially improve PL recovery, and consequently quality of life, is a concept that many patients may wish to embrace.
Nonetheless, our study is limited by several weaknesses. First, limited by its retrospective design with inherent selection bias, our study necessitates a randomized clinical trial for a higher level of evidence and confirmation of our results. Second, it is possible that the frequency of sexual intercourse and patient/partner satisfaction are correlated with the rate of PL recovery, but this information was unattainable in our study. Third, a longer follow-up beyond 1 year would be more useful to understand patterns of PL recovery in patients who demonstrated slower PL recovery. Fourth, measurement errors were inevitable. Although inter-observer bias was eliminated as a single surgeon evaluated patients throughout the study period, intra-observer variability may not have been well-controlled for as repeat measurements were not made for each visit. Fifth, the design of our penile habitation protocol also makes a direct comparison of the three PDE5I agents not possible because our PDE5i assignment regimen entails that a patient would take all three types in an assigned, sequential order. But it also provided an opportunity for patients to explore because it is reported that up to 40% patients are kept from best drug of choice if they only try one type of PDE5Is [20]. Sixth, patients who received adjuvant or salvage radiation therapies were not excluded from the study, but the relationship between penile length recovery and radiation therapy was not statistically significant in our study cohort (Additional file 2: Table S2). The cost of PDE5is is a potential financial limitation and may be prohibitive to some patients, though a generic version of sildenafil is now currently available. Lastly, our penile measurement did not include penile girth, and the standardized approximation of erect PL from SPL could be a source of systemic bias. However, it is not feasible to perform direct measurement of erect PL, and SPL is considered an accepted alternative [13].