Skip to main content
. 2020 Mar 14;8(2):214–222. doi: 10.1016/j.esxm.2020.01.010

Table 3.

Univariable logistic regression: potential predictors of MCID in IIEF-EF at 6 months

Total patients
MCID in IIEF-EF at 6 months
Or (95% CI) P
N No
Yes
N % N %
All 80 38 47.5 42 52.5
Group
 Group A 40 17 42.5 23 57.5 1.50 (0.62, 3.61) .371
 Group B 40 21 52.5 19 47.5 Reference
Age, in years
 <60 58 25 43.1 33 56.9 1.91 (0.70, 5.16) .204
 ≥60 22 13 59.1 9 40.9 Reference
 1-year increase - - - - - 0.97 (0.93, 1.01) .104
BMI (m/kg2)
 18.5–24.9 13 5 38.5 8 61.5 1.49 (0.39, 5.67) .556
 25–29.9 38 19 50.0 19 50.0 0.93 (0.35, 2.45) .889
 ≥30 29 14 48.3 15 51.7 Reference
 18.5–24.9 13 5 38.5 8 61.5 1.55 (0.46, 5.24) .478
 ≥25 67 33 49.3 34 50.7 Reference
HTN
 No 55 25 45.5 30 54.5 1.30 (0.50, 3.35) .587
 Yes 25 13 52.0 12 48.0 Reference
Dyslipidemia
 No 64 28 43.8 36 56.3 2.14 (0.69, 6.61) .185
 Yes 16 10 62.5 6 37.5 Reference
DM
 No 73 33 45.2 40 54.8 3.03 (0.55, 16.64) .202
 Yes 7 5 71.4 2 28.6 Reference
Hemoglobin A1C at baseline
 Normal 51 20 39.2 31 60.8 2.54 (0.99, 6.48) .052
 Pre-diabetes/diabetes 29 18 62.1 11 37.9 Reference
No. of comorbidities
 0 34 13 38.2 21 61.8 2.49 (0.70, 8.81) .158
 1 34 17 50.0 17 50.0 1.20 (0.29, 4.93) .800
 2+ 12 8 66.7 4 33.3 Reference
Comorbidities Y or N
 0 34 13 38.2 21 61.8 2.22 (0.90, 5.49) .083
 1+ 46 25 54.3 21 45.7 Reference
ED
 Mild 56 26 46.4 30 53.6 1.15 (0.44, 3.01) .769
 Moderate 24 12 50.0 12 50.0 Reference

BMI = body mass index; HTN = hypertension; DM = diabetes mellitus; ED = erectile dysfunction; IIEF-EF = International Index of Erectile Function-Erectile Function domain; MCID = minimal clinically important difference; OR = odds ratio.

Pre-diabetes: HbA1C: 5.7–6.4%.

MCID was defined as increase in IIEF-EF total score of ≥2 for patients with baseline mild ED (baseline IIEF-EF score of 17–25), and ≥5 for patients with baseline moderate ED (score of 11–16).

95% CI: 95% CI for OR. P: P-value from Wald's test.

OR estimates from univariable logistic regression modeling the probability of MCID = “Yes” at 6 months.

A multivariable analysis was conducted testing all the variables listed in this table, using stepwise model selection with entry criteria of 50% and retention criteria of 10%. Given the small sample size and data sparseness, only A1C at baseline was retained in the model.