Efficacy and safety of PDE5 inhibitors in the treatment of diabetes mellitus erectile dysfunction: a systematic review and meta-analysis

ABSTRACT Background: To systematically evaluate the efficacy and safety of the PDE5 inhibitors in diabetes Methods:Random were searched. Two researchers independently screened the literature, extracted the data and checked the results, and used the risk assessment tool to conduct a methodological quality assessment, and finally conducted a meta-analysis. The primary outcome, the erectile function scores of the International Index of Erectile Dysfunction (IIEF-EF), was recorded, while secondary outcomes (IIEF-Q3, IIEF-Q4, SEP 2 and 3, GAQ) and safety outcomes also needed attention. Results: Twelve studies included 3,124 patients and six PDE5 inhibitors were included. Compared with placebo, PDE5 inhibitors significantly improved male erectile function in IIEF-EF (WMD = 5.73; 95% CI: 3.62 to 7.84), IIEF-Q3 (WMD = 1.14; 95% CI: 0.87 to 1.40), IIEF-Q4 (WMD=1.28; 95% CI: 1.04 to 1.51), SEP2 (WMD=21.24; 95% CI: 15.50 to 26.98), SEP3 (WMD=25.77; 95% CI: 18.78 to 32.77), GAQ (RR) =2.98; 95% CI: 2.06 to 4.32), and with the safety (WMD=5.73; 95% CI: 3.62 to 7.84). Conclusions: PDE5 inhibitors can significantly improve the patient's erectile function, but cannot ignore their side effects.

inhibitors can significantly improve the patient's erectile function, but cannot ignore their side effects.

Background
Diabetes mellitus erectile dysfunction (DMED) is one of the common chronic complications of diabetes mellitus (DM), [1] which is also an important type of erectile dysfunction (ED), with the main manifestation of not being able to achieve and/or maintain adequate erection to achieve a satisfactory sexual life. [2,3] Compared with healthy men, the probability of ED in diabetic patients is 1.9-4 times that of healthy men. [4] As long-term hyperglycemia can damage the patient's vascular endothelium, nerves and endocrine, finally the erectile function of the corpus cavernosum is also seriously affected. [5,6] According to research, approximately 75% of diabetic men suffered from ED. [7] However, due to the complex effects of blood vessels and neuropathy, DMED is more difficult to treat than normal ED. [8] Currently, Phosphodiesterase-5 (PDE5) inhibitors are the first line of treatment for ED. Nowadays, more and more PDE5 inhibitors, such as avanafil, lodenafil, mirodenafil, sildenafil, tadalafil, udenafil and vardenafil, meets the sexual demanding of DMED patients and their partners. [9] And every coin has two sides, PDE5 inhibitors also have some side effects like dyspepsia and headache. [10] Studies have confirmed that PDE5 inhibitors can improve the clinical symptoms of DMED patients.
Previous literature search showed that there had been two meta-analyses to explore the efficacy and safety of PDE5 inhibitors in the treatment of DMED, but the above meta-analysis had some limitations such as outcomes and relatively old literature. [11,12] This article intends to use the current latest outcome indicators to conduct meta-analysis of relevant randomized controlled trials, systematically evaluate their effectiveness and safety, and provide more accurate evidence-based medical evidence for the clinical application of PDE5 inhibitors in the treatment of DMED.

Data Sources and Search Strategy
This systematic review and meta-analysis had registered on PROSPERO (No. CRD42018095185), and the whole protocol were published. [13] The protocol obeyed from Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement guidelines.

Study selection and data extraction
All the studies were managed by EndNote X7. Two dependent reviewers (FC and YFL) would read the the topics and abstracts of the studies, and decide whether they met the inclusion criteria. Any disagreement were discussed by another rich-experienced reviewer (HSL). We would extract necessary information from the included studies: trial characteristics (first author, publication year, sample size, duration of therapy, type and dose of drug), participants' baseline (age, country, type of diabetes) and outcomes. If the data in the literature was inadequate, we would try to contact the corresponding authors by email.

Risk of bias assessment
Risk of bias (ROB) assessment tool would be used to evaluate the quality of each literature. There were six biases including sequence generation, allocation concealment, blinding of outcome assessment, incomplete outcome data, selective reporting and others bias according to the Cochrane handbook. Two trained review authors (JSW and HHD) gave the article a low bias, unclear, or high bias evaluation. If there was a disagreement among reviewers, it would be discussed with another reviewer (BW) to make a decision. ROB assessment were performed with Review Manager statistical software package (Version 5.3.5).

Data synthesis and statistical analysis
Data synthesis were performed by the STATA statistical software package (Version 12.0). Mean difference (MD) and 95% confidence interval (CI) would be recorded for continuous variable outcome.
For dichotomous outcomes, we recorded the relative risk (RR) and 95% CI. Statistical heterogeneity was represented by a standard χ2 test and I 2 test with a significance level of p<.1. When I 2 < 50% and P ≥.1, a fixed-effect model will be used and a random-effects model will be use when I 2 ≥50% or P<.1. If the heterogeneity is high, we will analyze the cause of the heterogeneity and subgroup analysis will be performed.

Subgroup analysis and Sensitivity analysis
We would make subgroup analysis if there was significant heterogeneity in the final outcomes. And sensitivity analysis would preformed to test the robustness of meta-analysis results.

Search results and study characteristics
According to the search strategy, we retrieved 341 articles from the electronic database and 8 articles from other resources. Among them, 19 articles were repeated, 296 articles were removed according to the bibliography and abstracts, and 34 papers were obtained. Finally, 12 articles [14][15][16][17][18][19][20][21][22][23][24][25] were selected according to the document exclusion criteria. The literature screening process is shown in Figure 1.
We assessed risk of bias by the Cochrane Collaboration's tool. The systematic review included studies of RCT studies and blinded approach to participants and personnel (performance bias) and outcome accessment (detection bias). Three studies reported allocation concealment. [18,24,25] In the included literature, relevant outcome measures were reported in the method, so there was no selective bias. The risk of bias assessment of the studies is depicted in Figure 2.
IIEF-Q4 represented ability to maintain erections. A total of 5 articles [14][15][16]19,22] were included in the literature comparing PDE5 inhibitors with placebo to improve IIEF-Q4 in patients with DMED. Meta-analysis showed that PDE5 inhibitors were better than placebo in improving IIEF-Q4 in patients with DMED (WMD=1.28; 95% CI: 1.04 to 1.51; Figure 4, B), with no heterogeneity (P=0.523; I 2 =0%), a fixed effect model was used. SEP2 means whether you are able to insert your penis into your partner's vagina. A total of 3 articles [20,22,23] were included in the literature comparing PDE5 inhibitors with placebo to improve SEP2 in patients with DMED. Meta-analysis showed that PDE5 inhibitors were better than placebo in improving

Discussion
With the changes in people's lifestyles and the aging process, the incidence of DM is increasing year by year, and more and more young adults are suffering from DM. [26] The various complications caused by DM seriously affect people's quality of life. ED is one of its common chronic complications and is easily overlooked when assessing DM conditions. [27] Normally, their NO will activate guanylate cyclase (GC) to convert GTP to second messenger cGMP, when males receive sexual stimulation. cGMP activates serine protein kinase, which further phosphorylates proteins and ion channels, leading to open potassium channels, hyperpolarization of muscle cell membranes, and intracellular calcium chelation of the endoplasmic reticulum. By inhibiting the calcium channel to block the calcium influx, the intracellular calcium concentration is reduced, the smooth muscle is relaxed, and the corpus cavernosum is filled with blood to induce penile erection. 28 Current studies shows that the damage of endothelial function is the possible pathogenesis of DMED. Diabetes can lead to complications such as peripheral vascular disease and autonomic neuropathy. Endothelial cellmediated smooth muscle relaxation and neurofibrillary neurogenic NOS (nNOS) activity is often the main cause of complications. In addition, some researches have demonstrated that decreased eNOS activity in endothelial cells leads to a reduction in NO level in penile blood vessels, which may result in DMED. [29] Oral type 5 phosphodiesterase (PDE5) inhibitors are the primary means of oral drug therapy for DMED. PDE5 inhibitors mainly increase the concentration of PDE5 by inhibiting the activity of PDE5, thereby enhancing the penile erectile function under sexual stimulation. [30] This study retrieved RCTs from various PDE5 inhibitors for the treatment of DMED and systematically evaluated their efficacy and safety. This systematic review updated one RCT study, revised the data entry problem, and added new observation indexes such as sexual satisfaction evaluation and sexual self-confidence evaluation. The clinical efficacy evaluation method of ED is often carried out by means of subjective questionnaires. Currently, the most commonly used internationally is the IIEF-EF score questionnaire. In addition, SEP2 and SEP3 are also commonly used criteria for evaluating the clinical efficacy of ED. The study also included several other evaluation indicators, such as IIEF-Q3, IIEF-Q4, and GAQ. The results of this systematic review showed that there was a significant difference between the treatment group and the control group (P < 0.001), in the IIEF-EF score, IIEF-Q3, IIEF-Q4, SEP2, SEP3 and GAQ scores, suggesting that PDE5 inhibitors were significantly better than the control group in improving sexual desire, sexual life satisfaction. In terms of total adverse drug reactions, the difference between the treatment group and the control group was statistically significant, suggesting that the clinical complications should be noted to prevent the possibility of the above complications, like flushing, headache or dyspepsia.
However, this meta-analysis also had certain limitations: (1) due to the different types of PDE5 inhibitors used in the study, the subjects included in the study included different ethnic groups in Asia and Europe. In addition, some studies did not clearly distinguish the type of diabetes (T1DM or T2DM).
Therefore, there is a certain heterogeneity.

Author contributions
XL and QZ contributed equally to this work and are co-first authors. XL and JW designed the systematic review. QZ and JW drafted the protocol and JW and HD revised the manuscript. FC and YFL will independently screen the potential studies, extract data and finish data synthesis. BW, JW will assess the risk of bias. BW and HL will arbitrate any disagreement and ensure that no errors occur during the review.

Ethics approval and consent to participate
Not applicable. Our manuscript is a systematic review of previous studies; therefore it does not report on or involve the use of any animal or human data or tissue during our study.

Consent for publication
Not applicable. Our manuscript is a systematic review of previous studies; therefore it does not contain any individual persons data obtained in our study.  Abbreviations: CI, confidence interval; RR, risk ratio. Figure 1 PSIMA Flow Chart The risk of bias assessment of the studies