Populations. The included trials involved men diagnosed with erectile dysfunction (ED)

The included trials involved men diagnosed with erectile dysfunction (ED). The total and mean (range) numbers of patients randomly assigned to trial arms were 1975 and 179 (12–569), respectively. The inclusion criteria across the trials were adult men (age ≥ 18 years) diagnosed with ED in a stable heterosexual relationship of at least 6 months of duration. The inclusion criteria in two studies^{114, 120} were restricted either to patients with arteriogenic¹²⁰ or nonarteriogenic ED.¹¹⁴ In one trial, ²⁴⁹ the inclusion criteria were restricted to patients with ED and a history of diabetes (type I or II) who had been naï ve to any ED-related treatment. The enrolled patients in one trial¹⁵⁹ had to be naï ve to ED drug therapy and had to have an IIEF-5 score ≤ 21.

The exclusion criteria in the majority of the included trials were spinal cord injury^{114, 120, 248, 253} and penile/testicular deformity.^{114, 120, 148, 159, 248, 249, 251–254} Other frequently reported exclusion criteria were uncontrolled hypertension, ^{114, 148, 251–253}diabetes, ^{148, 248, 252, 253} endocrine disease, ^{114, 248, 251–253} any major hepatic/renal disease, ^{117, 148, 251} prostatectomy, ^{148, 248, 251, 253} peptic ulcers, ¹⁵⁹ HIV-positivity/AIDS, ^{248, 252, 253} the use of nitrates, ^{114, 117, 120, 148, 159, 251} a history of cancer, ^{248, 252, 253} and serious cardiovascular diseases (e.g. angina pectoris).^{148, 159} Only one trial failed to report exclusion criteria.²⁵⁰

The mean age of the study participants across the included trials ranged from 35²⁵⁰ to 59 years.¹²⁰ The race of the study participants was reported in six trials.^{117, 159, 248, 252, 253} The proportion of Caucasians across these trials ranged from 85 percent²⁵² to 99 percent.²⁴⁸ The duration of ED was at least 3 months at study entry in all trials. The longest mean duration of ED (3.5–4.2 years) was reported in the trial by Porst et al 2007.¹⁵⁹ Most commonly reported comorbidities among the study participants were diabetes, ^{114, 120, 148, 159, 248, 252, 253} hypertension, ^{114, 120, 148, 159, 248, 249, 252, 253} ischemic heart disease, ^{114, 249} and coronary artery disease.^{120, 159, 248, 252, 253} The presence or absence of comorbidities could not be ascertained from three trials.^{117, 250, 251} The underlying cause of ED (i.e., etiology) in the patients was specified in three trials to be nonarteriogenic (all patients), ¹¹⁴ arteriogenic (all patients), ¹²⁰ and various causes (patients with organic, psychogenic or mixed origin ED).¹⁵⁹ In four trials, ^{114, 120, 249, 250} the proportion of smokers ranged from 35 percent²⁴⁹ to 95 percent.^{114, 120} In the remaining trials this proportion could not be ascertained.^{117, 148, 159, 248, 251–253}

In seven trials^{117, 159, 248, 249, 252, 253} the authors provided baseline arm-specific distributions of the patients' sociodemographic and clinical characteristics (e.g. age, race, duration/severity of ED, comorbidities) and explicitly stated that the distribution of these characteristics between the randomized arms was generally similar. The authors of the remaining five trials^{114, 120, 148, 250, 251} did not report this information.