The role of 5‐alpha reductase inhibitors in transurethral resection of the prostate: a meta‐analysis of randomised controlled trials

Abstract

Objective

To critically evaluate the existing evidence base surrounding the efficacy of preoperative 5‐alpha reductase inhibitor (5ARI) administration in the reduction of perioperative complication rates in transurethral resection of prostate (TURP).

Methods

In April 2025, a systematic search of on‐line databases was conducted to identify randomised controlled trials (RCTs) that compared surgical outcomes and complication rates in patients undergoing TURP for benign prostatic hyperplasia (BPH) who were treated preoperatively with 5ARI (finasteride or dutasteride) as compared to placebo/none. The efficacy of preoperative finasteride was evaluated through outcomes related to blood loss, rate of blood transfusion, and operative time. The physiological mechanism of 5ARI treatment was evaluated through microvessel density (MVD) and vascular endothelial growth factor (VEGF) expression of the resected specimen.

Results

A total of 30 RCTs met the inclusion criteria for this meta‐analysis in which a total of 2974 patients underwent TURP for BPH (1464 5ARI: 1410 Control). Intraoperative blood loss was significantly lower among 5ARI‐treated patients (Z = 6.37, mean difference [MD] = −82.58 mL, 95% confidence interval [CI] −107.98 to −57.18; P < 0.001), which was reflected in a significantly lesser haemoglobin drop on the first postoperative day (Z = 6.84, MD = −0.90 g/dL, 95% CI−1.16 to −0.64; P < 0.001). The MVD was significantly lower in resected specimens from 5ARI‐treated patients (MD = −6.18 vessels/mm³, P < 0.001), whilst expressing significantly less VEGF (MD = −3.25, P < 0.001). Patients treated with 5ARI required blood transfusion less frequently than controls (odds ratio 0.31, P < 0.001). The use of 5ARI was associated with shorter operative time (MD = −3.47 min, P = 0.02) and lower volume of irrigation agents (MD = –2.07 L, P < 0.001).

Conclusion

Preoperative administration of 5ARIs significantly reduces intraoperative blood loss and risk of requiring blood transfusion in patients undergoing TURP for BPH. Even short durations (2 weeks) of 5ARI therapy can significantly reduce prostate vascularity.

Abbreviations

5AR(I)

5‐alpha reductase (inhibitor)

df

degrees of freedom

DL

decreased libido

DUT

dutasteride

ED

erectile dysfunction

FIN

finasteride

Hb

haemoglobin

HCT

haematocrit

MD

mean difference

MVD

microvessel density

OR

odds ratio

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta‐Analyses

RCT

randomised controlled trial

TUR

transurethral resection

TXA

tranexamic acid

VEGF

vascular endothelial growth factor

Introduction

Benign prostatic hyperplasia (BPH) refers to a benign proliferation of smooth muscle and epithelial prostatic cells and represents the most common cause of urinary obstruction in males [1, 2]. Symptomatic BPH presents clinically with troublesome LUTS and has significant effect on patient quality of life [3]. Prostatic tissue is highly responsive to testosterone through genomic and non‐genomic signalling with dysregulated response to 5‐alpha reductase (5AR)‐mediated androgen stimulation forming the pathophysiological basis of BPH [4]. Pharmacological management of BPH, through the use of 5AR inhibitors (5ARIs), aims to decrease prostate size and vascularity through disruption of androgenic stimulation within the prostatic parenchyma.

The operative apparatus available for the surgical management of BPH has seen significant development over the past decades [5]. TURP remains broadly considered as the reference standard for BPH management, although alternative strategies have demonstrated favourable morbidity profiles and may permit safer and more efficient intervention in the context of larger prostate glands [5, 6, 7]. The advent of bipolar TURP and improved sophistication of endoscopic apparatus has significantly reduced perioperative bleeding and blood transfusion rates [8, 9]. However, intraprocedural blood loss during TURP remains at 2.4–4.6 mL/min, with an associated erythrocyte transfusion rate of 2–7% [10, 11]. Intraoperative bleeding during TURP can impede endoscopic visualisation, prolong operative time, and result in increased patient morbidity [12, 13]. Similarly, bleeding in the postoperative period may result in prolonged bladder irrigation, resultant delays in decatheterisation and discharge, and a return to the operating theatre. As such, limiting blood loss in the context of TURP presents potential to improve both patient outcomes and institutional productivity.

There has been increased interest surrounding the use of pharmacological adjuncts aimed at reducing perioperative morbidity in TURP, such as intraoperative tranexamic acid (TXA) use to limit blood loss [13, 14]. Preoperative 5ARI therapy has demonstrated significant utility in the reduction of perioperative blood loss also. However, no studies have demonstrated statistically significant reduction [15, 16]. Androgenic stimulation of prostatic parenchyma results in increased prostatic microvessel density (MVD) through androgen dependent vascular endothelial growth factor (VEGF) stimulation. A growing number of studies have demonstrated 5ARIs significantly reduce both prostatic VEGF expression and MVD in patients undergoing TURP, adding credence to the proposed role of 5ARI in the reduction of perioperative blood loss [17, 18, 19].

The 5ARIs are commonly employed in the medical management of BPH and are the first treatment utilised in 15–25% of cases [20]. However, symptomatic improvement is often protracted when BPH is managed with 5ARIs, taking from 1 to 3 months for patients to report improvement in LUTS [21]. The utility of such agents as preoperative adjuncts would likely be diminished if effects on MVD and blood loss were to take a similarly protracted course. However, there is a growing body of evidence indicating that even short durations of 5ARI therapy may significantly decrease MVD and perioperative blood loss [15, 22, 23]. Several authors have reported meta‐analyses relevant to the role of preoperative 5ARIs in reducing TURP associated morbidity [24, 25, 26]. In the interim, several high‐quality RCTs have been published allowing for enhanced certainty of evidence. This systematic review and meta‐analysis aimed to critically appraise, synthesise, and update the current evidence base surrounding the safety and efficacy of 5ARIs pre‐TURP and provide insights regarding the physiological mode of action, optimal dose, timing and duration of 5ARIs in the pre‐TURP setting.

Methods

This systematic review and meta‐analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines [27]. All authors contributed to the formation of the study protocol, which was prospectively registered with the Prospective Register of Systematic Reviews (PROSPERO – CRD420251036239).

Population, Intervention, Comparison, and Outcome (PICO) framework

Population: adult, male patients undergoing TURP for BPH.

Intervention: preoperative administration of 5ARI (finasteride [FIN] or dutasteride [DUT]).

Comparison (Control): placebo or no 5ARI administered preoperatively.

Outcome: primary outcomes included perioperative blood loss, perioperative blood transfusion requirement, and resection time. Secondary outcome measures were aimed at assessing the mode of action of 5ARIs including MVD and VEGF expression of the resected tissue. We also sought to evaluate the duration of 5ARI therapy employed in the preoperative period and the incidence of 5ARI‐associated psychosexual dysfunction.

Search Strategy and Outcome

A systematic search of on‐line electronic databases (PubMed, Scopus, and Embase) was performed by two independent reviewers (C.M.H. and G.G.C.). The search terms: alpha reductase inhibitor, transurethral resection and prostate were combined using the Boolean operator ‘AND’. The search was not limited by year of publication. Duplicate studies were manually removed (C.M.H.) before screening titles, abstracts, and full texts. Relevant papers for inclusion were identified using a predetermined search strategy devised by senior author (N.F.D.). Retrieved studies were reviewed to ensure they met the inclusion criteria (minimum of one outcome of interest). The reference lists of the included studies were manually screened to identify additional relevant papers for inclusion (C.M.H. and G.G.C.).

Inclusion Criteria

1. RCTs comparing patients undergoing TURP treated preoperatively with 5ARI to patients treated with placebo/none.

2. At least one outcome of interest is reported in both intervention and control groups.

3. Full text available for review in English language.

Exclusion Criteria

1. Non‐RCT study design.

2. Inadequate statistical information provided for inclusion in meta‐analysis.

3. Full text not available.

4. Published abstracts from conference proceedings, review articles, case reports, and editorial articles.

Data Extraction

The following data were extracted and collated from studies meeting the inclusion criteria: (i) first author name, (ii) year of publication, (iii) journal of publication, (iv) study design, (v) country of origin, (vi) number of patients – intervention/control/total, (vii) outcomes of interest expressed as mean (SD) for continuous variables, (viii) outcomes of interest expressed as incidence for dichotomous variables. In studies where outcomes were reported as median and range, mean and SD were estimated utilising the Hozo et al. [28] formula. In studies reporting mean and standard error of the mean (SEM), the SD was calculated using the formula $\mathrm{SD}=\mathrm{SEM}\times \sqrt{n}$ [29]. Where the differences between two paired sets of mean values was to be determined, the variance addition rule was employed to calculate the change in mean and associated SD. Where patient level data were reported, data were manually transcribed with means and SDs calculated using Microsoft Excel (Microsoft Corp., Redmond, WA, USA).

Statistical Analysis and Quality Assessment

Data were expressed as dichotomous or binary outcomes, reported as odds ratios (ORs) or mean differences (MDs) expressed with 95% CIs following estimation using the Mantel–Haenszel method. Inter‐study heterogeneity was assessed through the τ² statistic, the Cochrane Q Statistic (chi‐squared) statistic and through visual inspection of resultant Forest plots in order to assess statistical dispersion. In accordance with BJUI meta‐analytical recommendations, the I ² value was not utilised to determine the statistical effect model employed [30]. A random‐effects model was determined to be most appropriate given the presence of clinically relevant inter‐study variability with respect to 5ARI agent, dose, and duration of 5ARI therapy and mean prostatic volume at baseline. All tests of significance were two‐tailed with P < 0.05 indicating statistical significance. Meta‐analysis was performed using Review Manager (RevMan), version 5.4 (Nordic Cochrane Centre, Copenhagen, Denmark) [31]. Sensitivity analysis was performed using the leave‐one‐out method in which the meta‐analysis was iteratively repeated following exclusion of each study in sequence to assess whether individual studies had disproportionate effects on the pooled estimate. Results were considered statistically robust if the magnitude and direction of the pooled effect remained consistent across iterations. Meta‐regression was performed to evaluate the effect of covariates on key outcome measures and to identify contributing factors to inter‐study heterogeneity. Risk of bias and methodology quality assessment was performed in accordance with the Cochrane guidelines through utilisation of the risk‐of‐bias for randomised trials 2 (RoB 2) tool, with representative figures generated using ‘robvis’ visualisation tool [32].

Results

The search of on‐line databases identified a combined total of 1074 publications. Following removal of duplicate records, the titles and abstracts of 705 articles were screened for inclusion. Following screening of titles and abstracts, 311 full texts were critically appraised for inclusion (C.M.H. and G.G.C.). Of these texts, 279 failed to meet the inclusion criteria. A further two papers were excluded due to concerns regarding the quality of evidence presented and/or potentially severe risk of bias. A total of 30 RCTs were included in this meta‐analysis [15, 16, 17, 18, 19, 22, 23, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54]. The identification, screening and inclusion of relevant papers is summarised in a PRISMA flow diagram (Fig. S1).

Study Characteristics

The data of the 30 RCTs were included in this meta‐analysis, published between 2001 and 2025. In all, 22 of the included studies were placebo controlled [15, 16, 17, 18, 19, 23, 34, 35, 36, 37, 38, 39, 40, 41, 42, 49, 50, 52, 53, 54], with patients from the remaining eight studies receiving no 5ARI therapy serving as the control arm [22, 43, 44, 45, 46, 47, 48, 51] (Table 1 [15, 16, 17, 18, 19, 22, 23, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54]).In all, 12 of the included studies utilised DUT in the preoperative setting, 17 studies utilised FIN, while Allahyani et al. [48], utilised DUT/FIN interchangeably within the intervention group (Table 1). The included studies demonstrated broad geographical spread with 12 studies published from European centres and the remaining 18 published from countries within Asia (Table 1). Risk‐of‐bias assessment of the included studies is outlined in Figs S2 and S3.

Table 1.

Outline of characteristics of included studies.

References	Year	Country	Journal	Study design	Placebo Control/none	5ARI	Intervention, n	Control, n	Total, N
Kravchick et al. [33]	2009	Israel	Urology	RCT	None	DUT	22	21	43
Hahn et al. [15]	2007	Sweden	BJU International	RCT	Placebo	DUT	71	70	141
Tuncel et al. [34]	2009	Turkey	Scandinavian Journal of Urology and Nephrology	RCT	Placebo	DUT	27	27	54
Pastore et al. [35]	2013	Italy	Journal of Endourology	RCT	Placebo	DUT	71	71	142
Busetto et al. [19]	2015	Italy	BJU International	RCT	Placebo	DUT	120	120	240
Nugroho et al. [36]	2015	Indonesia	Maj Obat Tradis	RCT	Placebo	DUT	20	20	40
Bansal et al. [37]	2017	India	Journal of Endourology	RCT	Placebo	DUT	147	150	297
Koul et al. [38]	2018	India	International Journal of Research in Medical Sciences	RCT	Placebo	DUT	30	30	60
Rahman et al. [39]	2019	Bangladesh	Journal of Biosciences and Medicines	RCT	Placebo	DUT	35	35	70
Khoso et al. [40]	2019	Pakistan	Pakistan Armed Forces Medical Journal	RCT	Placebo	DUT	26	26	52
Rammah et al. [41]	2023	Egypt	Current Urology	RCT	Placebo	DUT	31	34	65
Ali et al. [42]	2025	Pakistan	Cureus	RCT	Placebo	DUT	66	66	132
Özdal et al. [43]	2005	Turkey	Prostate Cancer and Prostatic Diseases	RCT	None	FIN	20	20	40
Aminsharifi et al. [44]	2016	Iran	Urology Journal	RCT	None	FIN	20	25	45
Khwaja et al. [22]	2016	Pakistan	J Coll Physicians Surg Pak	RCT	None	FIN	40	40	80
Ali et al. [45]	2018	Egypt	African Journal of Urology	RCT	None	FIN	42	56	98
Sumartoyo et al. [46]	2019	Indonesia	Indian Journal of Public Health Research and Development	RCT	None	FIN	13	13	26
Abdelbaki et al. [47]	2021	Egypt	African Journal of Urology	RCT	None	FIN	20	20	40
Allahyani et al. [48]	2024	United Arab Emirates	Int. Journal of Life Sciences, Biotechnology and Pharma Research	RCT	None	FIN/DUT	100	100	200
Sandfeldt et al. [49]	2001	Sweden	Urology	RCT	Placebo	FIN	42	56	98
Donohue et al. [50]	2002	UK	The Journal of Urology	RCT	Placebo	FIN	32	36	68
Crea et al. [51]	2005	Italy	Urologia Internationalis	RCT	Placebo	FIN	30	30	60
Donohue et al. [18]	2005	UK	BJU International	RCT	Placebo	FIN	31	33	64
De Berardinis et al. [17]	2008	Italy	Current Prostate Reports	RCT	Placebo	FIN	100	100	200
Lund et al. [16]	2009	Denmark	Scandinavian Journal of Urology and Nephrology	RCT	Placebo	FIN	16	17	33
Häggström et al. [52]	2009	Sweden	Scandinavian Journal of Urology and Nephrology	RCT	Placebo	FIN	15	13	28
Bansal et al. [37]	2017	India	Journal of Endourology	RCT	Placebo	FIN	146	50	296
Liu et al. [53]	2017	China	Tropical Journal of Pharmaceutical Research	RCT	Placebo	FIN	47	47	94
Penumatcha et al. [54]	2020	India	International Journal of Surgery Science	RCT	Placebo	FIN	50	50	100
Dutt et al. [23]	2021	India	Urology Annals	RCT	Placebo	FIN	34	34	68
Total, N							1464	1410	2974

Patient Characteristics

Baseline patient characteristics were comparable between the intervention and control groups in all studies, with no statistically significant difference in patient age, prostatic volume, preoperative haemoglobin (Hb), American Society of Anesthesiologists (ASA) grade, anticoagulant/ antiplatelet nor TXA use (P > 0.05) between intervention and control groups [15, 16, 17, 18, 19, 22, 23, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54]. There were significant differences noted in the mean prostatic size between studies (P > 0.05), ranging from 34 to 90.2 g. The dose, duration and mode of TURP (bipolar vs unipolar) was heterogenous between included studies (Table 2 [15, 16, 17, 18, 19, 22, 23, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54]).

Table 2.

Outline of the 5ARI agent, associated daily dose, and duration of therapy employed in the included trials.

References	Year	5ARI	Dose, mg/day	Duration, weeks
Kravchick et al. [33]	2009	DUT	0.5	4
Hahn et al. [15]	2007	DUT	0.5	2
Tuncel et al. [34]	2009	DUT	5	5
Pastore et al. [35]	2013	DUT	0.5	6
Busetto et al. [19]	2015	DUT	0.5	8
Nugroho et al. [36]	2015	DUT	0.5	2
Bansal et al. [37]	2017	DUT	0.5	4
Koul et al. [38]	2018	DUT	0.5	4
Rahman et al. [39]	2019	DUT	0.5	4
Khoso et al. [40]	2019	DUT	0.5	2
Rammah et al. [41]	2023	DUT	0.5	4
Ali et al. [42]	2025	DUT	0.5	2
		Mode	0.5	4
		Median	0.5	4
		Range	0.5–5	2–8
Özdal et al. [43]	2005	FIN	5	4
Aminsharifi et al. [44]	2016	FIN	5	2
Khwaja et al. [22]	2016	FIN	5	2
Ali et al. [45]	2018	FIN	5	6
Sumartoyo et al. [46]	2019	FIN	5	4
Abdelbaki et al. [47]	2021	FIN	5	2
Sandfeldt et al. [49]	2001	FIN	5	12
Donohue et al. [50]	2002	FIN	5	2
Crea et al. [51]	2004	FIN	5	8
Donohue et al. [18]	2005	FIN	5	2
De Berardinis et al. [17]	2008	FIN	10	8
Lund et al. [16]	2009	FIN	5	12
Häggström et al. [52]	2009	FIN	5	12
Bansal et al. [37]	2017	FIN	5	4
Liu et al. [53]	2017	FIN	5	1
Penumatcha et al. [54]	2020	FIN	5	2
Dutt et al. [23]	2021	FIN	5	2
Allahyani et al. [48]	2024	DUT/FIN	5	12
		Mode	5	2
		Median	5	4
		Range	5–10	1–12

Surgical Outcomes and Complication Rates

Intraoperative Blood Loss

Intraoperative blood loss was reported by 11 of the included studies (four:seven:one, DUT:FIN:DUT/FIN), which reported on 1200 patients (595:605, 5ARI:Control). Intraoperative blood loss was objectively quantified in all studies through analysis of Hb concentration in the irrigation fluid. Blood loss was significantly less in 5ARI‐treated patients (MD = −82.58 mL, 95% CI −107.98 to −57.18 mL; P < 0.001). There were no significant differences between DUT‐ and FIN‐treated patients (Fig. 1). Sensitivity analysis indicated findings were highly robust with the MD ranging from −72.83 to −92.16 mL, with P < 0.001 in all instances.

Fig. 1.

Forest plots demonstrating the effect of preoperative 5ARIs vs placebo/none on outcomes (A) intraoperative blood loss (mL), (B) relative intraoperative blood loss (mL)/g of resected prostate specimen, (C) perioperative serum Hb reduction (g/dL), (D) perioperative transfusion rate.

Normalised Blood Loss/g of Prostate Resected

Intraoperative blood loss was normalised relative to the mass of prostatic tissue resected in eight studies (four:four, DUT:FIN). The decrease in intraoperative blood loss observed in 5ARI‐treated patients remained statistically significant when normalised/g of prostatic tissue resected (MD = −2.60 mL/g, 95% CI −3.79 to −1.41 mL/g; P < 0.001]. There were no statistically significant differences in normalised blood loss between the DUT‐ and FIN‐treated groups (Fig. 1B). Sensitivity analyses revealed high robustness of findings, with the MD ranging from −3.03 to −1.99 mL/g, P < 0.001.

Perioperative Blood Loss

Perioperative blood loss was calculated by determining the difference between preoperative and postoperative Day 1 Hb and haematocrit (HCT) levels. Delta Hb was reported by 12 studies (seven:five, DUT:FIN), which included 1198 patients (592:606, 5ARI:Control). Patients treated with 5ARIs demonstrated a significantly lower drop in Hb (MD = −0.90 g/dL, 95% CI −1.16 to −0.64 g/dL; P < 0.001) (Fig. 1C). Delta Hb demonstrated highly robust evidence utilising the leave‐one‐out approach, with the MD difference ranging from −0.96 to −0.83 g/dL, P < 0.001.

Delta HCT was reported by four studies (two:two, DUT:FIN), which reported data representative of 315 patients (156:159, 5ARI:Control). Patients treated with 5ARIs demonstrated significantly lesser decrease in their serum HCT in the perioperative period (MD = −1.76, 95% CI −3.18 to −0.34; P = 0.004) (Fig. S4) Sensitivity analysis suggested that the overall effect was sensitive to the inclusion of Dutt et al. [23]. Exclusion of Dutt et al. [23] altered the overall effect (MD = −1.97, 95% CI −4.10 to 0.16), rendering the finding non‐significant (P = 0.07). Subgroup analysis revealed no significant difference in delta Hb or delta HCT between DUT‐ and FIN‐treated patients.

Blood Transfusion Rates

The incidence of blood transfusion among patients undergoing TURP was reported by 14 studies, which reported data on 1181 patients (575:606, 5ARI:Control). Patients treated preoperatively with 5ARIs required blood transfusion significantly less frequently (OR 0.33, 95% CI 0.17–0.66; P = 0.002) (Fig. 1D). Sensitivity analysis demonstrated findings to be highly robust, with odds of requiring blood transfusion for 5ARI treated patients ranging from 0.27 to 0.37, P < 0.01. Subgroup analysis demonstrated no significant difference between patients treated with DUT and FIN.

Resection Time

The mean resection time was reported by 16 studies (nine:five:one, DUT:FIN:DUT/FIN), which reported data representative of 1692 patients who underwent TURP (833:859, 5ARI:Control). Resection time was significantly shorter in cases where patients had received preoperative 5ARI therapy (MD = −3.47 min, 95% CI −6.29 to −0.65 min; P = 0.02) (Fig. 2). The reduction in resection time demonstrated among patients treated with 5ARIs did not reach statistical significance when groups were divided to DUT and FIN, with P = 0.08 and P = 0.11, respectively. Sensitivity analysis confirmed robustness of findings with the MD ranging from −3.83 to −3.14 min, P < 0.05. Removal of Dutt et al. [23] utilising a leave‐one‐out approach, resulted in the reduction in resection time associated with FIN to reach statistical significance (MD = −9.25 min, 95% CI −17.49 to −1.07 min; P = 0.03), indicating this study may have disproportionately affected meta‐analysis [23]. Subgroup analysis demonstrated no significant difference between patients treated with DUT and FIN.

Fig. 2.

Forest plots demonstrating perioperative outcomes of (A) resection time (min), (B) volume of intraoperative irrigation agent utilised (L).

Volume of Intraoperative Irrigation Agent and Transurethral Resection (TUR) Syndrome

The volume of intraoperative irrigation was recorded in 12 studies (eight:four, DUT:FIN), which reported data representative of 1109 patients who underwent TURP for BPH (544:565, 5ARI:Control). A significantly lower volume of irrigation agent was required during TURP in 5ARI‐treated patients (MD = −2.07 L, 95% CI −3.06 to −1.08 L; P < 0.001) (Fig. 2B). Subgroup analysis demonstrated no significant difference between groups treated with FIN and DUT. TUR syndrome rates were reported in five of the included studies. The incidence of TUR syndrome was 0% (0/261) and 1% (3/277) in the 5ARI and Control groups, respectively (Fig. S8). Meta‐analysis was not possible due to the low incidence rate of TUR syndrome, rendering most studies non‐estimable.

Secondary Outcomes

The MVD of the Resected Specimen

In all, 10 studies reported the MVD of the resected specimen (four:six, DUT:FIN), which reported data representative of 1227 patients who underwent TURP for BPH (611:616, 5ARI:Control). MVD was significantly reduced in the resected specimens of 5ARI‐treated patients (MD = −6.27 vessels/mm³, 95% CI −8.28 to −4.26 vessels/mm³; P < 0.001) (Fig. 3A). Sensitivity analysis demonstrated the findings to be robust when 5ARIs were evaluated collectively, with the MD ranging from −4.54 to −8.27, P < 0.001. However, when sub‐grouped based on the specific 5ARI subtype, DUT results demonstrated statistical fragility on removal of Busetto et al. [19] and Hahn et al. [15]. Upon performing leave‐one‐out analysis, the overall effect was (MD = −8.10 vessels/mm³, 95% CI −19.16 to 2.97 vessels/mm³; P = 0.15) and (MD = −4.54 vessels/mm³, 95% CI −9.71 to −0.65 vessels/mm³; P = 0.09) on removal of Busetto et al. [19] and Hahn et al. [15] data, respectively. Subgroup analysis revealed no significant differences in the reduction of MVD between the DUT‐ and FIN‐treated cohorts.

Fig. 3.

Forest plots demonstrating histopathological specimen characteristics associated with 5ARI treatment vs placebo/none in TURP. (A) MVD of resected prostatic specimen (vessels/mm³), (B) VEGF expression of resected prostatic specimen, (C) MVD characteristics sub‐grouped relative to duration of 5ARI therapy ≤2 weeks vs >2 weeks.

Vascular Endothelial Growth Factor Expression of the Resected Specimen

Four studies evaluated the expression of VEGF within the resected specimen (one:three, DUT:FIN), which captured data representative of 530 patients (264:266, 5ARI:Control) who underwent TURP for BPH. Preoperative treatment with 5ARI was associated with a significant decrease in VEGF expression within the resected specimen (standardised MD = −4.24, 95% CI −7.14 to −1.35; P < 0.001) (Fig. 3B). There were no statistically significant differences between DUT and FIN on subgroup analysis. Sensitivity analysis revealed the meta‐analysis results to be statistically robust, with overall effect ranging from −5.59 to −3.56, P < 0.01.

Duration and Dosage of 5ARI Therapy

The duration of 5ARI therapy was heterogenous between studies, ranging from 2 to 8 weeks (mode = 4 weeks, median = 4 weeks) and 1–12 weeks (mode = 2 weeks, median = 4 weeks) in DUT and FIN, respectively (Table 2). Subgroup analysis was performed to identify whether there were any statistically significant differences with respect to intraoperative blood loss and MVD of the resected specimen in patients treated with 5ARI for ≤2 weeks and those treated for >2 weeks.

Intraoperative blood loss was significantly reduced among patients treated with ≤2 weeks of 5ARI therapy (MD = −57.26 mL, 95% CI −81.20 to −33.31 mL; P < 0.001). Although the magnitude of change was greater among patients who received extended (>2 weeks) preoperative therapy of 5ARI (MD = −90.69 mL, 95% CI −136.83 to −44.56 mL; P < 0.001), there was no significant difference between groups (P = 0.21) (Fig. S5).

The MVD of the resected specimen was also significantly reduced among patients both treated with ≤2 weeks (MD = −17.15 vessels/mm³, 95% CI −26.32 to −7.99 vessels/mm³; P < 0.001) and >2 weeks of 5ARI therapy (MD = −4.06 vessels/mm³, 95% CI −6.13 to −1.99 vessels/mm³; P < 0.001) (Fig. 3C). Short‐term 5ARI therapy was associated with significantly greater reduction in MVD than therapy for >2 weeks on subgroup analysis (chi‐squared = 7.46, P < 0.01).

Two of the included studies utilised a comparatively high‐dose of 5ARI therapy prior to TURP, with Tuncel et al. [34] and De Berardinis et al. [17] prescribing 5 mg DUT once daily and 10 mg FIN once daily, respectively [17, 34]. All other included studies utilised standardised dosage regimens of 0.5 mg/day DUT or 5 mg/day FIN. Subgroup analyses were employed to evaluate the effect of higher dosages on MVD and intraoperative blood loss. The decrease in MVD was greater in patients treated with standard dose 5ARI therapy (MD = −9.96 vessels/mm³, 95% CI −13.94 to −5.99 vessels/mm³; P < 0.001) vs high‐dose 5ARI (MD = −0.42 vessels/mm³, 95% CI −6.17 to 5.33 vessels/mm³; P = 0.89). The subgroup differences were statistically significant (chi‐squared = 7.17, P < 0.01). Intraoperative blood loss was only reported in one of the RCTs that employed a high‐dose 5ARI regimen [34]. As such, subgroup analysis was not performed.

Psychosexual Dysfunction

Psychosexual dysfunction associated with 5ARI use was evaluated by four studies, which reported on data representative of 391 patients who underwent TURP for BPH (194:197, 5ARI:Control). Erectile dysfunction (ED) was more commonly reported among 5ARI‐treated patients (2.6% [5/194]) compared to Control (0% [0/197]). Meta‐analysis demonstrated a non‐significant increased OR of ED among 5ARI‐treated patients (OR 4.58, 95% CI 0.76–27.60; P = 0.10) (Fig. S6). Decreased libido (DL) was also reported with increased frequency among patients treated preoperatively with 5ARIs (4.1% [8/194]) compared to controls (0% [0/197]). Patients treated with 5ARIs were significantly more likely to experience DL (OR 6.99, 95% CI 1.22–40.04; P = 0.03) (Fig. S7). In all studies reporting follow‐up on outcomes relevant to psychosexual function, patients who reported ED or DL reported resolution of symptoms following cessation of 5ARI at the 3‐month follow‐up [34, 37, 43].

Meta‐Regression

Meta‐regression analyses were conducted to evaluate the effect of covariates 5ARI agent (DUT/FIN), duration of 5ARI therapy, prostate volume, TURP subtype (monopolar/bipolar), patient age, and geographical location on the following outcome measures: intraoperative blood loss, perioperative blood loss, and operative time.

Intraoperative Blood Loss

Univariate meta‐regression demonstrated no single covariate to be a significant contributor of heterogeneity in blood loss effect (P > 0.05). Multivariate modelling utilising prostate volume, mean age and 5ARI dose as covariates accounted for 74% (R ² = 0.74) of inter‐study heterogeneity (P < 0.001, k = 9, degrees of freedom [df] = 5).

Perioperative Blood Loss

Univariate meta‐regression demonstrated higher MVD to be significantly associated with greater delta Hb values (P = 0.04), no other covariate was significant at univariate analysis (P > 0.3). Multivariate regression model with prostate volume, mean age and 5ARI duration accounted for a significant proportion of inter‐study heterogeneity (R ² = 0.52, P = 0.02, k = 9, df = 5).

Operative Time

No covariable was significant following univariate meta‐regression. Combination of 5ARI dose, duration, prostate volume and patient age accounted for a moderate proportion of heterogeneity (R ² = 0.14). The model was non‐significant (P = 0.72, k = 15, df = 9).

Discussion

This meta‐analysis demonstrates preoperative 5ARI therapy is effective in significantly reducing intraoperative blood loss, perioperative blood loss, and blood transfusion requirement in patients undergoing TURP for BPH. The mechanism of the 5ARI induced reduction in blood loss is likely owed to decreased prostatic vascularity, as demonstrated by the significant reduction in MVD and VEGF expression in the resected prostatic specimens of 5ARI‐treated patients. Although erythrocyte transfusion rates following TURP are reportedly low (2–7%), the significant morbidity associated with blood product transfusion should not be overlooked [11, 55]. As such, preoperative 5ARI therapy presents an attractive risk mitigation strategy in patients undergoing TURP.

However, perioperative morbidity in TURP is not limited to blood loss alone. Although rates of TUR syndrome following TURP have declined in the past decades, the risk of TUR syndrome remains directly proportional to both the length of operating time and the volume of irrigation agent utilised [11, 56]. Patients treated with preoperative 5ARI therapy in this study demonstrated significant reduction in both operative duration and volume of irrigation agent required.

The use of 5ARIs has recently been increasingly correlated with adverse psychological effects, with 5ARI‐treated patients demonstrating significantly elevated risk of depression, anxiety, and suicidal ideation [57, 58]. However, the risks of 5ARI‐associated self‐harm and suicidality are not significantly increased among patients without a prior medical history of mood disorder, with many authors calling for pre‐treatment assessment of psychological risk to be protocolised [58, 59]. Psychosexual adverse effects, such as ED and DL, are also commonly reported among patients treated with 5ARIs [60]. A subset of patients will reportedly experience persistence of these symptoms following cessation of 5ARI therapy [61]. However, it is important to distinguish that proposed application of 5ARIs in the included trials are as preoperative adjunctive therapies, with a median therapeutic duration of 4 weeks. All patients who reported de novo psychosexual dysfunction associated with 5ARI therapy reported resolution at the 3‐month follow‐up.

The use of 5ARIs in the medical management of BPH, is limited primarily by the duration of therapy required for patients to report symptomatic benefit. Peak‐reduction in LUTS following initiation of 5ARI therapy is reported to occur at 12–18 months, with most patients requiring at least 1–3 months of therapy to achieve any noticeable symptomatic benefit [21, 62]. As such, many urologists may be hesitant to initiate preoperative 5ARI therapy due to the perception that prolonged therapeutic duration may be required. To address said perception, this meta‐analysis has conclusively demonstrated that even short duration (≤2 weeks) of 5ARI therapy demonstrates a significant reduction in MVD and intraoperative blood loss in TURP. The short therapeutic duration required to generate significant reduction in intraoperative blood loss and MVD represents the most clinically applicable finding of this meta‐analysis. An ideal adjunctive pharmacological therapy should augment but not delay surgical intervention, rendering preoperative 5ARI therapy an ideal candidate agent pre‐TURP. The results of this systematic review demonstrate DUT and FIN to be equivalent when employed pre‐TURP, but these drugs are not equivalent from a pharmacological perspective. DUT binds and inhibits both type 1 and type 2 isoenzymes of 5AR, whereas FIN is only capable of inhibiting the type 1 isoenzyme [63]. DUT has demonstrated superior capacity in the reduction of prostatic volume, the reduction of PSA, and the improvement in patient‐reported severity of LUTS [64, 65]. However, there remains insufficient evidence to support the choice of an optimal pre‐TURP 5ARI agent. Further research should focus on directly comparing FIN and DUT, in a non‐inferiority/superiority trial setting.

In this systematic review, we have identified multiple statistically significant findings with a high degree of clinical applicability for practicing urologists. These findings have demonstrated high robustness on sensitivity analyses. Further research on the role of preoperative 5ARI therapy in TURP should focus on delineating the optimal dose and duration of therapy.

This study is subject to limitations. The most notable limitation is that of heterogeneity within the dataset, with a mean I ² of the included analyses of 70.2%. There are multiple likely contributors to the high I ² values. Increased prostate size is strongly correlated to perioperative blood loss, and increased transfusion requirements [66]. The included studies with respect to geographical location as well as dose and duration of pre‐TURP 5ARI therapy, further confounding analysis. Although meta‐regression provided insight into the sources of inter‐study heterogeneity, the statistical powering was low with high degrees of freedom. The results of meta‐regression analyses should be interpreted as exploratory rather than explanatory. The sample size available for evaluation of 5ARI‐associated psychosexual dysfunction among TURP patients was small and warrants cautious interpretation.

Conclusion

Preoperative administration of 5ARIs significantly reduces intraoperative blood loss and risk of requiring blood transfusion in patients undergoing TURP for BPH. The significant decrease in perioperative blood loss is likely owed to the role of 5ARIs in reducing MVD of prostatic tissue. Even short durations (i.e., 2 weeks) of 5ARI therapy can significantly reduce prostate vascularity.

Disclosure of Interests

All authors have declared no conflicts of interest.

Funding

No funds have been received for the purpose of this research output.

Supporting information

Fig. S1. The PRISMA flow diagram detailing the process of study identification, screening and inclusion/exclusion.

Fig. S2. Risk of bias traffic light plot.

Fig. S3. Risk of bias summary plot.

Fig. S4. Forest plots demonstrating the effect of preoperative 5ARIs vs placebo/none on perioperative serum HCT reduction.

Fig. S5. Forest plots demonstrating the effect of preoperative 5ARIs vs placebo/none on outcome intraoperative blood loss (mL) sub‐grouped by duration of 5ARI therapy ≤2 weeks vs >2 weeks.

Fig. S6. Patient reported incidence rates of ED among 5ARI‐treated and control cohorts.

Fig. S7. Patient reported incidence rates of DL among 5ARI‐treated and control cohorts.

Fig. S8. Incidence of TUR syndrome in 5ARI‐treated patients vs control undergoing TURP for BPH.