DOES DIABETES HAVE A NEGATIVE IMPACT ON ERECTILE FUNCTION RECOVERY AFTER RADICAL PROSTATECTOMY?

Abstract

Objectives:

To quantify the risk of long-term post-radical prostatectomy (RP) erectile dysfunction (ED) in men with diabetes mellitus (DM).

Methods:

We included men who underwent RP and were followed for ≥ 24 months at our institution; men were excluded if they received androgen deprivation therapy or radiation therapy. Erectile function recovery (EFR) was assessed using the International Index of Erectile Function (IIEF) Erectile Function Domain (EFD) score pre-RP and serially during follow-up. We performed logistic regression analysis to investigate a potential association between erectile function 24 months post- RP.

Results:

Of 2,261 men included. 8% were diabetic. Men in the diabetic group tended to present with more vascular comorbidities. For men with DM, the median time from diagnosis was 4 years pre-RP, and the median hemoglobin A1c pre-RP was 6.7%. After 24 months post-RP, EFR was significantly lower among the diabetic group. The median EFD was 7. Men with DM had a lower proportion of functional EFR (17%) and a greater proportion of severe ED (57%). In the univariable logistic regression model to analyze DM diagnosis was a significant predictor of functional EFR (OR 0.43, p<0.001) and severe ED (OR 1.85, p<0.001) 24 months post-RP. Furthermore, this was not observed for a multivariable analysis.

Conclusions:

24 months after RP, EFR is compromised in individuals with DM

INTRODUCTION

Despite the low mortality associated with radical prostatectomy (RP), it is a potentially morbid procedure which carries risks of iatrogenic urinary incontinence and erectile dysfunction (ED).¹ Several studies have reported rates of long-term, post-RP ED ranging from 14–90%.² While appropriate counseling regarding the potential for post-RP ED can help mitigate procedure regret,² in current practice, personalized and comprehensive risk assessments are usually not performed.^3–6

Prior studies have indicated that baseline erectile function (EF), patient age, nerve-sparing during surgery, and vascular comorbidities are associated with EF recovery (EFR) after RP.^7–9 Coronary artery disease, dyslipidemia, sleep apnea, diabetes mellitus (DM), and hypertension, as well as cigarette smoking status are known factors associated with ED in general.^{2, 8, 10, 11} Studies have reported that a higher number of comorbidities are associated with poorer EFR after RP; however, there is a lack of evidence describing how each comorbidity in general and DM in particular, affects EFR post-RP.

DM is highly prevalent in the US with estimates up to 14% of the general population.^12–14 When accounting for variability in severity and duration of DM, the prevalence of ED among diabetics ranges from 20–85%.^{9, 15, 16} Mechanisms of ED development in men with DM include: autonomic neuropathy, endothelial dysfunction, cavernosal hypercontractility due to heightened sympathetic activation, and venocclusive dysfunction related to corpus cavernosal smooth muscle collagenization. Despite this, it remains poorly understood how DM impacts EFR after RP. We aimed to evaluate the association between the diagnosis of DM and functional EFR 24 months after RP.

MATERIALS AND METHODS

Study Population:

After institutional review board approval was obtained (IRB 16–459), we reviewed our institutional database for all patients undergoing RP between 2008 and 2020 who completed the International Index of Erectile Function (IIEF) at baseline and 24 months post-surgery. As part of the institutional penile rehabilitation protocol to maximize EFR after surgery, all patients are instructed to start with daily low-dose PDE5i before and after surgery; starting in the third week after surgery the patient will continue with a low-dose PDE5 inhibitor (PDE5i) plus a challenge of maximum dose PDE5i. Those men who achieve functional erections with PDE5i are encouraged to continue with oral medications and full dose PDE5i up to 3 times a week; men who are not achieving functional erections commence intracavernosal injection therapy immediately and use these up to 3 times a week plus low dose PDE5i (on non-ICI days) for a duration of 1 year after surgery. At the beginning of the second year after RP, patients resume trial of maximum dose PDE5i and if and when they are capable of penetration hardness erection with PDE5i they cease ICI if they wish.

We excluded patients who were treated with androgen deprivation therapy or radiation therapy before or after RP. We gathered demographic, comorbidity, and surgical data. We report nerve sparing status (NSS) during surgery, each nerve being independently graded and scored between 1–4: 1 = completed preserved, 2 = partially preserved, 3 = partially damaged, and 4 = completed resected. Nerve sparing scored as 1/2 was deemed as nerve preservation, while scores of 3/4 were deemed non-preserved. For predictive modeling, we defined nerve sparing as bilateral (1+1, 1+2, 2+2), unilateral (3 or 4 on one side) and non-nerve-sparing (3 or 4 bilaterally).

Diabetes:

DM was characterized including time since diagnosis, number of medications used, insulin utilization, hemoglobin A1c before RP, and the report of end-organ damage.

Erectile Function Assessment:

Erectile function was assessed using the IIEF erectile function domain (EFD) score, pre-RP and at 24 months post-RP. Question 1–5 and 15 of the IIEF represent the EFD with scores ranging from 6–30. By definition, a score of ≤10 indicates severe ED. For this study, we utilized a score ≥24 to define functional erections as this has been shown to correlate with high patient satisfaction with erection.¹⁷ Patients also reported PDE5i exposure after RP. Patients completed a PDE5i use inventory scored as: 0 = never, 1 = sometimes (less than once a week), 2= regularly (once or twice a week), 3 = routinely (every day or nearly every day).

Statistical Analyses:

Medians with interquartile range (IQR) for continuous variables and as percentages for categorical variables were used to report descriptive assessment. To assess the association between DM and EFD scores at 24 months post-RP, we used Fischer’s exact tests for categorical variables and rank sum test for continuous variables. Similarly, to assess the association between degree of DM (uncontrolled vs controlled) and EFD scores at 24 months post-RP we used Fischer’s exact tests for categorical variables and rank sum test for continuous variables. To evaluate PDE5i usage in the 24-month post-operative period, we defined PDE5i use as any survey within the first 24 months indicating that they had taken a PDE5i. To assess duration of PDE5i in the first 24 months, we assumed when a patient reported taking PDE5i that they had used them for the full time period since they last answered a survey (or since radical prostatectomy for a patient’s first survey). To assess predictors of poor/severe ED (EFD ≤10) and functional EFR (EFD ≥24) after RP, we utilized univariable and multivariable logistic regression assessments. Predictors included in the model were: DM diagnosis, number of comorbidities, baseline EFD score, time on PDE5i post-RP, age at RP and NSS. All statistical analyses were conducted using STATA 16.1 (StataCorp, College Station, TX).

RESULTS

Patient Population:

The total number of eligible patients was 2,261, 8% (181 patients) being diabetic. Patient characteristics are summarized in Table 1. The median age was similar in the diabetic and non-diabetic control groups, respectively. Men in the diabetic group tended to present with more vascular comorbidities, with significantly higher rates of hypertension, hyperlipidemia, and obstructive sleep apnea, compared to the control group. Men with DM had lower median testosterone levels, 304 vs 367 ng/dL, which while statistically significant are likely not clinically meaningful. The median duration of DM diagnosis was 4 (1, 15) years pre-RP; median medication number was 1 (1, 2); 17% reported the use of insulin; median hemoglobin A1c was 6.7% (6.2%, 7.4%); 43% had a hemoglobin A1c ≥7% pre-RP, and 21% reported end-organ damage. Before RP, most of the men in both groups reported never or sometimes using PDE5i (94% for the non-diabetic group and 97% for the diabetic group).

Table 1:

Baseline Patient Characteristics by Diabetes Diagnosis.

	Non-Diabetics (N=2080; 92%)	Diabetics (N=181; 8%)	p-value
Age at RP (Years) (Median IQR)	62 (57, 67)	64 (59, 68)	0.001
Number of Comorbidities (Median IQR)	1 (0, 2)	3 (2, 3)	<0.001
Hypertension	43%	72%	<0.001
Hyperlipidemia	49%	73%	<0.001
Obstructive Sleep Apnea	18%	25%	0.023
Coronary Artery Disease	5%	9.4%	0.024
Smoking Status
Never	53%	56%	0.8
Former	38%	37%
Current	8.6%	7.2%
Pre-RP PSA (ng/mL) (Median IQR)	5.5 (4.1, 7.8)	5.3 (4.2, 7.4)	0.9
Baseline Testosterone (ng/dL) (Median IQR)	367 (282, 487)	304 (226, 385)	0.002
Baseline Glucose (mg/dL) (Median IQR)	95 (86, 107)	122 (100, 159)	<0.001
Pre-RP Glucose ≥180	0.9%	19%	<0.001
BMI (Median IQR)	28 (25, 31)	30 (27, 33)	<0.001
Grade Group at RP
1	13%	7.2%	0.019
2	65%	62%
3	17%	22%
4	2.4%	5%
5	3%	4%
Pathologic Stage at RP
pT2	65%	53%	0.004
pT3	35%	46%
pT4	0.2%	0.6%
Nerve Sparing Status
Non-sparing	6%	12%	<0.001
Bilateral	78%	63%
Unilateral	16%	25%
Nerve Sparing Score (Median IQR)	3 (2, 4)	4 (2, 5)	<0.001

RP = Radical Prostatectomy, PSA = Prostate Specific Antigen, IQR = Interquartile range,

Erectile Function Recovery:

In our cohort, the median baseline EF was significantly higher (4-point difference) in non-diabetic men than in men with DM (Table 2). No between group differences were seen for pre-RP PDE5i use, 84% for non-diabetic group and 82% for diabetic group reporting never using PDE5i prior to their operation. At the 24-month follow-up time-point, no difference was observed between DM and non-DM patients in PDE5i exposure, with a median PDE5i utilization score of 1 (0, 2), designating “sometimes”. After 24 months post-RP, EFR was significantly higher among the non-diabetic group than among men with DM (7-point difference). Additionally, based on the EFD score category, it was observed that men with DM had a lower rate of functional EFR (17% vs. 32%, p< 0.001) and a greater proportion of severe ED (57% vs. 42%, p<0.001) compared to the non-DM group 24 months post-RP.

Table 2:

Erectile Function Recovery at 24 Months after Radical Prostatectomy by Diabetes Status

	Non-Diabetics (N=2080; 92%)	Diabetics (N=181; 8.0%)	p-value
Baseline EFD score (Median IQR)	27 (18, 30)	23 (12, 28)	<0.001
EFD score at 24 months post RP (Median IQR)	14 (5, 26)	7 (3, 18)	<0.001
EFD score at 24 months post RP
≤10	42%	57%	<0.001
11–23	26%	26%
≥24	32%	17%

RP = Radical Prostatectomy, IQR = Interquartile range, EFD = erectile function domain

Predictors of Erectile Function Recovery:

On univariable assessment (Table 3), at 24 months post-RP, diabetics were less likely to have recovery of functional erections (OR 0.43, p<0.001) and were more likely to have severe ED (OR 1.85, p<0.001). The same findings were observed for: comorbidity number (excluding DM), DM plus any comorbidity, baseline EFD score, patient age, and nerve sparing status. However, on multivariable analysis (Table 4), only baseline EFD score, patient age and nerve sparing status were predictive of the recovery of functional erections and severe ED were predictive.

Table 3:

Univariable Analysis of Predictors of Erectile Function Recovery 24 Months after Radical Prostatectomy

	Functional Erectile Function Recovery (IIEF score ≥24)			Poor Erectile Function Recovery (IIEF score ≤10)
Predictor	OR	95% CI	p-value	OR	95% CI	p-value
Diabetes (yes)	0.43	0.29, 0.64	<0.001	1.85	1.36, 2.51	<0.001
Number of Comorbidities excluding diabetes (per one increase)	0.75	0.69, 0.83	<0.001	1.25	1.15, 1.36	<0.001
Diabetes + any comorbidity (yes)	0.42	0.28, 0.64	<0.001	1.87	1.36, 2.57	<0.001
Baseline EFD score (per point increase)	1.16	1.14, 1.19	<0.001	0.91	0.90, 0.92	<0.001
PDE5i Exposure score (per point increase)	1.04	0.99, 1.09	0.14	0.90	0.86, 0.94	<0.0001
Age at RP (per year increase)	0.92	0.91, 0.93	<0.001	1.10	1.09, 1.12	<0.001
Nerve Sparing Status
Not Preserved	Ref.	.	.	Ref.	.	.
Bilateral	6.02	3.31, 11.0	<0.001	0.15	0.10, 0.23	<0.001
Unilateral	2.59	1.36, 4.93	0.004	0.35	0.22, 0.54	<0.001

RP = Radical Prostatectomy, PDE5i = Phosphodiesterase 5 inhibitor, OR = Odds Ratio, CI = Confidence Interval

Table 4:

Multivariable Analysis of Predictors of Erectile Function Recovery 24 Months after Radical Prostatectomy

	Functional Erectile Function Recovery (IIEF score ≥24)			Poor Erectile Function Recovery (IIEF score ≤10)
Predictor	OR	95% CI	p-value	OR	95% CI	p-value
Diabetes diagnosis (yes)	0.50	0.19, 1.31	0.2	1.69	0.78, 3.67	0.2
Number of Comorbidities (excluding diabetes, per ne increase)	0.91	0.75, 1.11	0.4	1.07	0.88, 1.30	0.5
Baseline EFD score (per point increase)	1.13	1.09, 1.18	<0.0001	0.92	0.90, 0.94	<0.0001
Time on PDE5i post-RP (per month increase)	0.99	0.94, 1.05	0.8	0.94	0.89, 0.99	0.013
Age at RP (per year increase)	0.94	0.92, 0.97	0.0002	1.06	1.02, 1.09	0.001
Nerve Sparing Status
Not Preserved	Ref.			Ref.
Bilateral	4.59	0.92, 22.84	0.063	0.20	0.06, 0.71	0.013
Unilateral	2.42	0.45, 13.04	0.3	0.35	0.09, 1.32	0.12

RP = Radical Prostatectomy, PDE5i = Phosphodiesterase 5 inhibitor, OR = Odds Ratio, CI = Confidence Interval

DISCUSSION

Endothelial dysfunction is a significant cause of ED in the general population and serves as a risk factor for failure of EFR after RP.^{5, 7, 9} The presence of vascular comorbidities (hypertension, smoking, dyslipidemia, DM mellitus, coronary artery disease) negatively impacts EFR after RP.^{7, 9} DM and hypertension are highly prevalent in men with ED. Selvin et al. reported the prevalence and risk factors for ED in the US.⁹ This study described that after adjustment for age, men aged 40 years and older with dyslipidemia or with DM had a higher prevalence of ED compared to other cardiovascular risks; 49% and 31% of these men had ED, respectively.⁹ It is known that all these cardiovascular risk factors are potential threats to penile endothelium and penile smooth muscle leading to ED due to functional and structural changes.⁸ However, there is a lack of evidence published on each single cardiovascular risk factor, including DM, as a risk factor for EFR post-RP. Our study showed that men with DM had more vascular comorbidities, lower testosterone levels, lower baseline erectile function scores, with a lower proportion of functional erections and greater proportion of severe ED compared to non-DM men on univariable analysis.

Moreover, based on clinical experience and data published, a higher number of these vascular risk factors present in the patient preoperatively affect the EFR post-RP. Muller et al., in a penile rehabilitation program with oral sildenafil and/or intracavernosal injection therapy following RP, demonstrated that the presence of at least two vascular risk factors resulted in a two-fold risk for failure to recover natural erections sufficient for intercourse 18 months post-RP in a cohort of men with no ED pre-RP.⁷ Teloken et al. also demonstrated that an accumulation of vascular risk factors decreases the chances of EFR.² In this study that included 984 with prostate cancer underwent RP, the preoperative information on vascular risk factors (hypertension, dyslipidemia, DM, coronary artery disease, and smoking status) were assessed to evaluate EFR 24 months post-RP. This study showed that older age (r=0.37), lower EF baseline pre-RP (r=0.41), higher NSS score (r=0.35), and the number of vascular risk factors ≥2 (r=0.15) were significantly correlated with higher probabilities of ED 24 months after RP. Mulhall et al. developed pre-op and early postoperative nomograms to predict the EFR following RP.¹⁸ Both nomograms have the number of comorbidities as the variables included to calculate the probabilities of EFR 24 months post-RP. The comorbidities profile included coronary artery disease, peripheral vascular disease, stroke, DM, hypertension, dyslipidemia, obesity, and cigarette smoking status. In both nomograms, the numbers of comorbidities had the lowest score compared to age and pre-RP EF for pre-operative nomogram and age, pre-RP EF, and NSS score for early postoperative nomogram.¹⁸ All these studies failed to report on each individual vascular risk factor’s association with lower probabilities of EFR 24 months post-RP. We highlight the research published by Sridhar et al., who reported that preoperative hypertension alone among other comorbidities (including DM), predicted significantly lower EFR 18 months post-RP.⁶

Our univariate analysis observed an association between post-RP EFR and DM. Men with DM had a significantly lower likelihood of functional EFR (OR 0.43) and higher likelihood of severe ED (OR 1.85) after RP. Nevertheless, upon adjusting for baseline EF score, duration of PDE5i utilization, age at RP, and NSS, the association failed to attain conventional statistical significance. A diagnosis of DM alone was not a predictor of severe ED nor a predictor of functional erections in a cohort of men undergoing penile rehab; based on the previous evidence reported, baseline erectile function, age, and NSS remain critical in setting patient expectations of EFR after RP. It is noteworthy, however, that clinical significance cannot be ruled out. A more extensive multivariable analysis, possibly requiring a larger sample size of DM patients, with better glycemic control characterization, may unveil statistically significant associations, and this might then shift to statistical significance. Our study also highlights that a larger proportion of DM failed to recover functional erections compared to those without DM, (57% vs 42%) at 24 months post-RP. This may be attributed to multiple factors, including the greater impact of intraoperative vascular and nerve injury in men with DM. This study should urge clinicians to discuss the impact of DM during preoperative counseling and underscores the importance of addressing baseline comorbidities to optimize EFR outcomes post-surgery. It is noteworthy that Derogar et al. have demonstrated that counseling on sexual side effects decreased the bother patients felt about ED one year after surgery.⁴

The definition of EFR post-RP in the literature that is published is highly varied.¹⁰ One definition that is widely used is erection with or without PDE5i to have sexual intercourse. Based on this definition, Bianco et al., in a study that included 785 men, reported that 70% of men had EFR 24 months post-open RP.¹⁹ In another study published by Eastham et al. using the same definition for EFR after RP, in 1,577 men with normal EF baseline pre-RP, normal EFR was seen in 56% after 24 months post-RP.²⁰ In our cohort, using the IIEF EFD score ≥ 24 to define functional recovery, we reported lower EFR prevalence compared to many studies (IIEF EFD score ≥24 = 31%), demonstrating the need for a consensus on the cut-off for defining recovery of erectile function after RP. The rate of ED in this study was found to be approximately 70% 24 months post-RP. This is even higher than the study published by Stanford et al., which demonstrated ED rates of 59.9% after RP¹, and is lower than those found by Sridhar et al., which reported 76%.⁶ The discrepancy between these values is likely attributable to variation in surgeon expertise, single vs multi-center studies, post-operative EF rehabilitation efforts, and differences in the definition of ED. Furthermore, while the range is broad, these findings are significant for patients when choosing a therapy for prostate cancer.

Our findings must be interpreted in the context of the limitations of the study, including its single-center nature, retrospective study design, no data on serial HbA1C levels, and the lack of a standard definition of well-controlled vs poorly controlled DM. This study could be improved upon by implementing prospective enrollment of RP patients, serial HbA1c and testosterone measurement for all participants, and quantification of the adherence to the institutional PDE5i penile rehabilitation protocol. This study however, benefits from study of a large cohort, use of a database that is audited and reviewed for accuracy annually by the institution, and use of an adequate and widely accepted survey to evaluate EF. Considering that sexual dysfunction is an independent predictor of a reduced quality of life after prostate cancer treatment and that maintenance of quality of life is the primary motivator for treatment choice in over 45% of patients,^{21, 22} it follows that clinicians should strive to offer personalized preoperative counseling based on readily available data such as diabetic status. Establishing realistic expectations and formulating a comprehensive post-surgery recovery strategy for men with DM undergoing ED treatment post-surgery necessitates thoroughly evaluating multifaceted factors. Including measures degree of DM control and complications.

CONCLUSIONS

At twenty-four months after RP, based on univariable analysis, EFR is compromised in individuals with DM, as evidenced by diminished odds of functional EFR and elevated odds of severe ED. These findings were not observed in multivariable analysis suggesting such analysis will require a larger DM patient number.

Figure 1:

Erectile Function Domain Score at 24 Months Post Radical Prostatectomy