Erectile Dysfunction in Individuals with Type 1 Diabetes is Associated with Long-term Metabolic Control and Diabetic Complications: A Cross-Sectional Study

Abstract

Background  Erectile dysfunction (ED) affects approximately 38% of individuals with type 1 diabetes (T1DM). Skin autofluorescence (AF) reflects skin advanced glycation end product (AGE) deposits and is a marker of long-term glycemia control.

Objective  The study investigates the relationship between ED and diabetes control in patients with T1DM.

Methods  Adult patients with T1DM visiting the Diabetology Department were cross-sectionally investigated. Medical history, anthropometric features, and laboratory findings were collected. All individuals filled the International Index of Erectile Function (IIEF-5). IIEF-5 total score < 22 represented the presence of ED. AF was measured on the volar aspect of the forearm using AGE Reader. Insulin resistance (IR) was assessed by the estimated glucose disposal rate. Descriptive statistics and multivariate logistic regression analyses were performed. The adjusted covariates were general risk factors of ED.

Results  Of a total of n  = 70 patients, n  = 30 (42.9%) suffered from ED. The presence of ED was associated with higher glycated hemoglobin level (OR, 95% CI; 1.62, 1.02–2.60; p  = 0.043), presence of at least one diabetic complication (3.49, 1.10–11.03; p  = 0.03), and skin AF (9.20, 1.60–52.94; p  = 0.01), but not with IR (0.78, 0.57–2.60; p  = 0.12). Skin AF values ≥ 2.2 indicates presence of ED with a sensitivity of 70.0% and a specificity of 77.5%. Area under the curve was equal to 0.72 (95% CI: 0.60–0.85).

Conclusions  The presence of ED in individuals with T1DM is associated with HbA1c, the presence of at least one diabetic complication, and skin AF.

Introduction

Type 1 diabetes mellitus (T1DM) is a chronic disease of an autoimmune origin. The autoinflammatory process destroys the β-cells of the pancreas and leads to an absolute insulin deficiency. Standard treatment for individuals with T1DM is an intensive insulin therapy that allows imitating the natural secretion of insulin. ¹ Nowadays, the main challenge in the treatment of T1DM is the maintenance of optimal long-term control of glycemia to prevent the development of insulin resistance (IR), diabetic complications, and preterm mortality. ²

Erectile dysfunction (ED) is a durable inability to achieve and/or maintain an adequate erection to perform sexual intercourse. ED in people with diabetes is multifactorial and involves vascular, neurologic, hormonal and psychological disturbances. ³ The risk of ED is higher in older subjects with higher body mass index (BMI), or with sleep disturbances, those spending more time sitting as well as in smokers and could be related to extensive alcohol consumption. ^{4 5 6 7 8 9 10} ED is present in 37.5% of adult males with T1DM and is more prevalent in this group than in healthy subjects. ¹¹ ED in people with diabetes is a challenge because the efficacy of the treatment is less efficient than in nondiabetic individuals. ^{12 13} Chronic hyperglycemia results in the formation of advanced glycation end products (AGEs), the accumulation of which is related to the progress of diabetic complications. ¹⁴ AGEs are involved in inflammation pathways and the oxidative stress, and may accelerate the atherosclerotic process (diabetic vasculopathy) as well as the neuropathic course. ¹⁵ AGE accumulation levels provide information on glycemic control in a more extended period than one measurement of glycated hemoglobin (HbA1c). ¹⁶ To date, only Kouidrat et al studied the relationship between AGE accumulation and ED in a limited sample of individuals with T1DM. ¹⁷ They found that skin autofluorescence (AF), an indirect marker for AGE, is associated with the presence of severe ED. ¹⁷

The relationship between IR, metabolic syndrome, and ED was previously observed, but it was never investigated in the T1DM population. ¹⁸ Chen et al revealed that IR (assessed by using quantitative insulin sensitivity check index—a surrogate of gold standard hyperinsulinemic-euglycemic clamp) was positively associated with the severity of ED. ¹⁹ Homeostasis model assessment IR index was proven to be an independent predictor of ED. However, indicators mentioned above cannot be used in the assessment of IR among individuals with T1DM. ED concerned people without other clinical symptoms and could be one of the first warning signs indicating IR. ²⁰ IR is associated with impaired gene regulation and protein production—many harmed genes are involved in erectile function. ²¹

Here, we hypothesize that the presence of ED may be associated with markers of long-term diabetes control: HbA1c, AGE skin accumulation assessed by skin AF, and consequences of diabetes such as IR and presence of diabetic complications.

The study investigates the relationship between ED and diabetes control in patients with T1DM.

Materials and Methods

Data Collection

The data comes from the project Poznań Atherosclerosis in Adult Patients with long-term Type 1 Diabetes Mellitus Study (PARADISE T1DM Study), which was approved by the local Ethical Committee (consent # 67/19). The design of the study follows the principles of the 1964 Declaration of Helsinki. All individuals gave their written, informed consent for participation in the study.

One hundred consecutive male patients with T1DM, from the Department of Internal Medicine and Diabetology, were recruited from February 2019 to November 2019. The study was cross-sectional. Inclusion criteria were as follows: at least 5 years' history of the disease confirmed in the past with positive T1DM-antibodies status, aged between 18 and 65 years. Patients with unstable hyper/hypothyroidism (thyroid-stimulating hormone [TSH] beyond normal range) and/or other endocrinological disorders, contagious diseases, unstable rheumatological disorders, renal or liver diseases, alcoholism, or diagnosed malignancy, and those using medication for ED, with a history of urological procedures and without a complete dataset were excluded. The minimal number of entities was set at 70, because it is widely accepted as the minimal number of records for the multivariate logistic regression model with seven predictors. ²²

All enrolled subjects completed a questionnaire including details of age, medical history, duration of diabetes, daily insulin dose, treatment, history of diabetic complications, history of smoking, alcohol intake, and lifestyle including the type of occupation (shift work vs. nonshift work), sleeping time (hours per day), physical work time (hours per working day), sitting time (hours per day) and sports activity (hours per week).

Participants completed the International Index of Erectile Function (IIEF-5). ²³ The questionnaire consists of 5 questions, measured on the Likert scale ranging from 1 to 5 points. The questions are as follows:

Over the past 6 months:

• 1) “How do you rate your confidence that you could get and keep an erection?”

• 2) “When you had erections with sexual stimulation, how often were your erections hard enough for penetration?”

• 3) “During sexual intercourse, how often were you able to maintain your erection after you had penetrated your partner?”

• 4) “During sexual intercourse, how difficult was it to maintain your erection to completion of intercourse?”

• 5) “When you attempted sexual intercourse, how often was it satisfactory for you?”

Higher values (maximum 25) represent good sexual function, while lower (minimum 5) ones indicate ED. An IIEF-5 value below 22 indicates ED. The Polish version of the questionnaire utilized in previous studies was used. ^{24 25 26}

The patients underwent a complete physical examination with anthropometric (weight, height, waist and hip circumferences) and blood pressure (BP) measurements. BP was measured three times by the Korotkoff method in the sitting position, after 10 minutes rest, using a manual sphygmomanometer. BMI and waist-to-hip ratio (WHR) were calculated from the following equations: BMI = weight (kg) / squared height (m ² ); WHR = waist circumference (cm) / hip circumference (cm).

Laboratory measurements included lipid profile, TSH, creatinine, transaminases, and C-reactive protein (CRP). Low-density lipoprotein cholesterol (LDL-C) level was calculated by the Friedewald formula. ²⁷ HbA1c was assessed with a turbidimetric inhibition immunoassay (Cobas 6000, Roche Diagnostics). Additionally, the urine albumin to urine creatinine ratio (ACR) was analyzed.

Participants were screened for diabetic retinopathy using direct ophthalmoscopy through dilated pupils, followed, if necessary, by fluorescent angiography. Renal function was assessed by creatinine level, estimated glomerular filtration rate (GFR), and presence of albuminuria. Albuminuria was defined as a urinary albumin excretion rate between 30 and 300 mg/24 hour in two of the three samples collected over 3 months. Diabetic nephropathy was defined as the presence of albuminuria in connection with diabetes of over 10 years' duration or with diagnosed earlier diabetic nephropathy. Neuropathy was assessed using pressure sensation with 10 g monofilament perception, vibration perception with a 128 Hz tuning fork, temperature sensation, pain sensation with a pin, the patellar reflex, and Achilles' reflex. Diabetic neuropathy was diagnosed in individuals with two or more of the following four components: the presence of symptoms of neuropathy, the absence of Achilles' reflex, abnormal scores for pressure, and/or of vibration perception. Macroangiopathy was defined as the presence of peripheral artery disease, coronary artery disease, experienced myocardial infarction, or ischemic stroke in connection with diabetes. Diabetic foot syndrome was defined as the presence of a foot ulcer, neuropathic osteoarthropathy, or history of lower limb amputation due to diabetes complications. ¹

IR was assessed using the estimated glucose disposal rate (eGDR) formula ²⁸ :

eGDR (mg/kg/min) = 24.31–(12.22 × WHR) – (3.29 × arterial hypertension) – (0.57 × HbA1c)

where arterial hypertension if present, count as 1, and if not present, count as 0, and HbA1c—glycated hemoglobin (%).

eGDR formula is easy to utilize in clinical practice. The calculation was developed based on clinical risk factors in adults with T1DM in the epidemiology of diabetes complications (EDC) study and was validated using the euglycemic-hyperinsulinemic clamp. ²⁸

Skin AGE levels were assessed using a noninvasive device measuring skin AF (AGE-Reader, DiagnOptics BV, Groningen, The Netherlands). This method was validated by Meerwaldt et al. ²⁹ The device was calibrated using the manufacturer's instruction. The standard protocol, which was used in a previous study, was followed. ¹⁷ The investigator had to detect healthy skin on the volar aspect of the forearm. These measurements should not be performed on skin with any lesions. The device covers 4 cm ² of skin to protect against an external source of light. The covered area is illuminated by the AGE Reader. The emitted light and the excitation light reflected by the skin are measured with a spectrometer in the 300 to 600 nm range. The trained researcher performed three measurements and calculated the mean value. The final outcome of AGE Reader is expressed in arbitrary units. Higher values represent higher AGE levels. The method has limited value for people with dark skin. However, all of the participants were Caucasian; thus, our study was not limited by this limitation of the AGE Reader.

Data Analysis

We used the R-programming language (version 3.6.1.; Vienna, R Project) for statistical analysis. A descriptive analysis and comparison using the Chi-square test and the Mann–Whitney U test were performed. The categorical data is presented as number (percentage), while numerical data is preset as median (lower quartile to upper quartile). We compared the group with versus without presence of ED as well as based on median of skin AF value.

The total IIEF-5 score as well as the scores (1–5) of all five questions with the numerical characteristics of the study group were correlated. For this purpose, corrplot package of R was used to visualize the R Spearman rank correlation test. ³⁰

Both univariate and multivariate logistic regression analyses were performed. The dependent variable was the presence of ED (IIEF-5 score < 22) versus no ED (IIEF-5 score: 22–25). The presence of ED was coded as one, while no ED as null. The independent variables were as follows: eGDR, HbA1c, the presence of at least one diabetic complications, and skin AF value. HbA1c, eGDR, the presence of at least one diabetic complications, and skin AF value were separately adjusted to universal ED risk factors such as age, BMI, sleeping time, sitting time, current smoking status and alcohol intake. ^{4 5 6 7 8 9 10} In total, four models were created (skin AF/HbA1c/eGDR/the presence of at least one diabetic complication + age, BMI, sleeping time, sitting time, current smoking, alcohol intake). All models were checked for multicollinearity using the variance inflation factor (VIF). The cutoff of VIF was set at 5.0.

The receiver operating characteristic (ROC) curve was analyzed to determine the optimal skin AF value to discriminate individuals with ED versus those without ED. The optimal cutoff point was established using Youden's index. The analysis was performed using the ROCit package of R. ³¹

We performed a sensitivity analysis. We analyzed the association between the presence of ED and skin AF value, stratified by age. We performed an additional ROC curve analysis of the model adjusted for age.

Results

Initially, n  = 100 males signed informed consent. Of these, n  = 17 (17.0%) did not complete the IIEF-5 questionnaire, and n  = 13 (13.0%) were excluded due to the diagnosis of the condition, which was exclusion criteria or missing data. Finally, there were a total of n  = 70 (100.0%) males aged 39 (29–49) years who fulfilled the inclusion criteria. Of them, n  = 40 (57.1%) had no ED (22–25 points), n  = 21 (30.0%) mild ED (17–21 points), n  = 6 (8.6%) mild-to-moderate ED (12–16), n  = 2 (2.9%) moderate ED (8–11), and n  = 1 (1.4%) severe ED (1–7). The characteristics of the study group and the comparison of individuals with ED versus those without ED are presented in Table 1 . None of the patients took other lipid-lowering drugs other than statins. Patients with ED were significantly older, with a longer history of T1DM, and more commonly suffering from at least one diabetic complication such as diabetic retinopathy, diabetic neuropathy, or having higher aspartate aminotransferase (AST) serum level, AF and lower eGDR than people without ED ( Table 1 ). The median value of skin AF value was equal to 2.1. The comparison of groups with higher (≥ 2.1) and lower (< 2.1) skin AF values are presented in Table 2 . The individuals with higher skin AF values were older, with longer T1DM history, more commonly had diabetic complications, had higher HbA1c levels, and lower IIEF-5 total scores and eGDR values.

Table 1. Comparison of groups with and without self-declared ED, according to the IIEF-5.

Feature	All participants n  = 70 (100.0%)	ED group n  = 30 (42.9%)	non-ED group n = 40 (57.1%)	p -Value
General
• Age (years)	39.0 (29.0–49.0)	46.5 (38.0–53.5)	35.5 (28.0–42.2)	0.01
• Diabetes duration (years)	18.0 (11.0–25.0)	22.0 (17.2–29.5)	13.0 (9.0–24.2)	0.009
• BMI (kg/m 2 )	26.5 (23.4–28.9)	26.2 (22.6–30.4)	26.5 (24.1–28.3)	0.74
• Waist circumference (m)	0.94 (0.86–1.05)	0.96 (0.86–1.08)	0.94 (0.87–1.02)	0.351
• WHR	0.9 (0.9–1.0)	0.9 (0.9–1.0)	0.9 (0.8–0.9)	0.05
• SBP (mm Hg)	130.0 (120.0–136.0)	130.0 (126.2–138.0)	128.2 (120.0–133.0)	0.07
• DBP (mm Hg)	80.0 (77.8–86.3)	81.2 (79.2–86.5)	80.0 (76.8–85.4)	0.43
Comorbidities
• At least one diabetic complication ( n ) (%)	37 (52.9)	22 (73.3)	15 (37.5)	0.006
• Diabetic retinopathy ( n ) (%)	25 (35.7)	16 (53.3)	9 (22.5)	0.02
• Diabetic nephropathy ( n ) (%)	6 (8.6)	4 (13.3)	2 (5.0)	0.42
• Diabetic neuropathy ( n ) (%)	24 (34.3)	15 (50.0)	9 (22.5)	0.03
• Diabetic foot syndrome ( n ) (%)	3 (4.3)	3 (10.0)	0 (0.0)	0.15
• Diabetic macroangiopathy ( n ) (%)	3 (4.3)	2 (6.7)	1 (2.5)	0.80
• Hypertension ( n ) (%)	30 (42.9)	16 (53.3)	14 (35.0)	0.20
Drugs
• Multiple daily injection ( n ) (%)	65 (92.9)	30 (100.0)	35 (87.5)	0.12
• Daily insulin intake (insulin units/day/kg)	0.5 (0.5–0.6)	0.5 (0.4–0.6)	0.6 (0.5–0.7)	0.09
• Metformin ( n ) (%)	14 (20)	7 (23.3)	7 (17.5)	0.76
• Statin ( n ) (%)	16 (22.9)	8 (26.7)	8 (20.0)	0.71
Atorvastatin 10 mg	2 (2.9)	1 (3.3)	1 (2.5)
Atorvastatin 20 mg	4 (5.7)	2 (6.6)	2 (5.0)
Atorvastatin 40 mg	4 (5.7)	2 (6.6)	2 (5.0)
Rosuvastatin 5 mg	2 (2.9)	0 (0.0)	2 (5.0)
Rosuvastatin 10 mg	2 (2.9)	1 (3.3)	1 (2.5)
Rosuvastatin 40 mg	1 (1.4)	1 (3.3)	0 (0.0)
Simvastatin 10 mg	1 (1.4)	1 (3.3)	0 (0.0)
• ACEI/ARB ( n ) (%)	26 (37.1)	14 (46.7)	12 (30.0)	0.24
• Beta-blocker ( n ) (%)	11 (15.7)	5 (16.7)	6 (15.0)	1.00
• Diuretic ( n ) (%)	7 (10)	5 (16.7)	2 (5.0)	0.23
Lifestyle
• Current smoker ( n ) (%)	16 (22.9)	4 (13.3)	12 (30.0)	0.18
• Packyears	0.0 (0.0–7.1)	0.0 (0.0–6.4)	0.0 (0.0–7.0)	0.99
• Alcohol intake (units/week)	1.0 (0.0–2.5)	1.0 (0.0–5.0)	1.0 (0.0–2.0)	0.56
• Shift work ( n ) (%)	22 (32.4)	10 (35.7)	12 (30.0)	0.82
• Sleeping (hours/day)	7.0 (6.5–8.0)	7.0 (6.5–8.0)	7.0 (6.5–7.6)	0.59
• Physical work (hours/day)	4.0 (1.0–8.0)	6.0 (2.5–8.8)	3.0 (0.4–8.0)	0.09
• Sitting (hours/day)	5.0 (4.0–6.9)	5.0 (3.4–6.1)	5.0 (4.1–7.0)	0.70
• Sport activity (hours/week)	3.0 (1.0–5.8)	2.0 (0.1–4.9)	3.8 (1.5–6.0)	0.16
Laboratory findings and others
• HbA1c (%)	7.9 (7.0–9.0)	8.3 (7.2–9.7)	7.8 (6.8–8.6)	0.14
• TC (mmol/L)	4.6 (4.1–5.2)	4.6 (4.2–4.9)	4.6 (4.0–5.3)	0.87
• LDL-C (mmol/L)	2.5 (1.9–2.9)	2.4 (2.1–2.7)	2.5 (1.9–3.0)	0.67
• HDL-C (mmol/L)	1.6 (1.3–1.8)	1.6 (1.3–1.8)	1.5 (1.4–1.8)	0.82
• TG (mmol/L)	91.0 (74.0–125.0)	92.0 (79.5–109.8)	88.0 (73.0–129.5)	0.63
• ALT (UI/L)	20.0 (16.0–25.8)	21.5 (18.0–32.8)	19.0 (14.0–25.0)	0.11
• AST (UI/L)	19.0 (16.0–25.0)	21.0 (17.0–25.8)	17.0 (15.0–23.2)	0.04
• Creatinine (µmol/L)	82.7 (76.0–88.2)	81.3 (71.6–87.3)	83.1 (76.0–88.4)	0.48
• ACR (mg/mmol)	0.4 (0.3–0.7)	0.4 (0.3–0.9)	0.3 (0.3–0.6)	0.33
• C-reactive protein (nmol/L)	7.6 (5.5–14.9)	8.0 (6.3–13.8)	7.3 (5.4–16.7)	0.76
• Skin AF	2.1 (1.8–2.5)	2.3 (2.0–2.6)	1.9 (1.8–2.2)	0.006
• eGDR (mg/kg/min)	7.5 (4.3–8.6)	5.8 (3.7–7.9)	7.9 (5.5–8.8)	0.04

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; ACR, albumin to creatinine ratio; AF, autofluorescence; ALT, alanine transaminase; ARB, angiotensin receptor blocker; ASA, acetylsalicylic acid; AST, aspartate transaminase; BMI, body mass index; DBP, diastolic blood pressure; eGDR, estimated glucose disposal rate; HbA1c, glycated hemoglobin; IIEF, International Index of Erectile Function Questionnaire; HDL-C, high-density lipoprotein; IQR, interquartile range; LDL-C, low-density lipoprotein; SBP, systolic blood pressure; TG, triglycerides; TC, total cholesterol; WHR, waist-to-hip ratio.

Data presented as median (IQR) / n (%).

Table 2. Comparison of groups with higher (≥ 2.1) or lower (< 2.1) skin AF values.

Feature	AF above/equal median (≥ 2.1) n  = 40 (57.1%)	AF below median (< 2.1) n  = 30 (42.9%)	p -Value
General
• Age (years)	43.5 (37.5–52.0)	33.0 (25.0–39.8)	0.001
• Diabetes duration (years)	22.0 (13.0–27.0)	13.0 (9.0–21.8)	0.045
• BMI (kg/m 2 )	25.9 (23.2–31.0)	26.5 (24.1–28.0)	0.90
• Waist circumference (m)	0.9 (0.87–1.09)	0.94 (0.86–1.02)	0.45
• WHR	0.9 (0.9–1.0)	0.9 (0.8–0.9)	0.051
• SBP (mm Hg)	130.0 (122.5–136.1)	128.0 (120.0–133.0)	0.30
• DBP (mm Hg)	81.2 (78.2–86.0)	80.0 (76.6–86.3)	0.52
Comorbidities
• At least one diabetic complication ( n ) (%)	28 (70.0)	9 (30.0)	0.002
• Diabetic retinopathy ( n ) (%)	21 (52.5)	4 (13.3)	0.002
• Diabetic nephropathy ( n ) (%)	4 (10.0)	2 (6.7)	0.95
• Diabetic neuropathy ( n ) (%)	19 (47.5)	5 (16.7)	0.015
• Diabetic foot syndrome ( n ) (%)	3 (7.5)	0 (0.0)	0.35
• Diabetic macroangiopathy ( n ) (%)	2 (5.0)	1 (3.3)	1.00
• Hypertension ( n ) (%)	28 (70.0)	9 (30.0)	0.10
Drugs
• Multiple daily injection ( n ) (%)	37 (92.5)	28 (93.3)	1.00
• Daily insulin intake (insulin units/day/kg)	0.6 (0.5–0.7)	0.5 (0.5–0.6)	0.42
• Metformin ( n ) (%)	12 (30.0)	2 (6.7)	0.035
• Statin ( n ) (%)	16 (40.0)	0 (0.0)	< 0.001
• ACEI/ARB ( n ) (%)	19 (47.5)	7 (23.3)	0.07
• Beta-blocker ( n ) (%)	10 (25.0)	1 (3.3)	0.03
• Diuretic ( n ) (%)	6 (15.0)	1 (3.3)	0.23
Lifestyle
• Current smoker ( n ) (%)	10 (25.0)	6 (20.0)	0.84
• Pack years	0.3 (0.0–12.0)	0.0 (0.0–1.9)	0.12
• Alcohol intake (units/week)	1.0 (0.0–2.6)	1.0 (0.0–2.4)	0.33
• Shift work ( n ) (%)	16 (42.1)	6 (20.0)	0.09
• Sleeping (hours/day)	7.0 (6.5–8.0)	7.0 (6.5–8.0)	0.77
• Physical work (hours/day)	3.2 (0.4–8.2)	4.5 (1.6–8.0)	0.78
• Sitting (hours/day)	5.4 (4.5–7.2)	4.6 (3.4–6.3)	0.04
• Sport activity (hours/week)	3.0 (1.0–6.2)	2.4 (1.0–5.0)	0.68
Laboratory findings and others
• HbA1c (%)	8.4 (7.4–9.2)	7.5 (6.8–8.2)	0.018
• TC (mmol/L)	4.5 (3.9–5.4)	4.6 (4.3–5.1)	0.55
• LDL-C (mmol/L)	2.5 (1.9–2.8)	2.5 (2.1–2.9)	0.70
• HDL-C (mmol/L)	1.5 (1.2–1.8)	1.7 (1.4–1.9)	0.004
• TG (mmol/L)	1.3 (1.1–2.2)	1.1 (1.0–1.4)	0.015
• ALT (UI/L)	20.0 (16.0–29.0)	20.0 (17.2–25.0)	0.92
• AST (UI/L)	19.0 (16.0–26.2)	19.0 (16.0–23.8)	0.76
• Creatinine (µmol/L)	81.3 (74.3–89.5)	83.1 (77.3–86.6)	0.84
• ACR (mg/mmol)	4.0 (2.9–6.1)	4.0 (3.0–7.0)	0.86
• C-reactive protein (nmol/L)	8.0 (6.1–14.9)	7.4 (5.0–15.0)	0.47
• IIEF total score ( n )	21.0 (17.0–23.2)	24.0 (21.2–25.0)	0.012
• eGDR (mg/kg/min)	6.0 (3.6–7.9)	8.4 (5.6–8.9)	0.009

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; ACR, albumin to creatinine ratio; AF, autofluorescence; ALT, alanine transaminase; ARB, angiotensin receptor blocker; ASA, acetylsalicylic acid; AST, aspartate transaminase; BMI, body mass index; DBP, diastolic blood pressure; eGDR, estimated glucose disposal rate; HbA1c, glycated hemoglobin; IIEF, International Index of Erectile Function Questionnaire; HDL-C, high-density lipoprotein; IQR, interquartile range; LDL-C, low-density lipoprotein; SBP, systolic blood pressure; TG, triglycerides; TC, total cholesterol; WHR, waist-to-hip ratio.

Data presented as median (IQR) / n (%).

The correlation results are presented in Fig. 1 . Total IIEF-5 score was positively related to eGDR (Rs = 0.25; p  = 0.040) while was negatively related to age (Rs = – 0.33; p  = 0.005), diabetes duration (Rs = – 0.35; p  = 0.003), AST (Rs = – 0.24; p  = 0.045) and AF (Rs = – 0.35; p  = 0.003). The correlation results between other numerical variables are presented in Supplementary Fig. S1 .

Fig. 1.

Correlation matrix. Columns names: the five questions and total score of the International Index of Erectile Function (IIEF-5). Blue tiles present positive associations, while red tiles negative associations. Asterisks within tiles corresponds to p value of associations. * p value < 0.05; ** p value < 0.01; *** p value < 0.001. Abbreviations: ACR, albumin to creatinine ratio; AF, autofluorescence, ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; DDI, daily dose of insulin; DM, diabetes mellitus; eGDR, estimated glucose disposal rate; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein; LDL-C, low-density lipoprotein; SBP, systolic blood pressure; TG, triglycerides; TC, total cholesterol; WC, waist circumference; WHR, waist-to-hip ratio.

In univariate regression analysis, eGDR was negatively associated with the presence of ED (OR, 95% CI; 0.81, 0.66–0.99; p  = 0.040), but the relationship was not significant after adjustment to covariates (0.78, 0.57–2.60; p  = 0.12) ( Fig. 2A ).

Fig. 2.

Multivariate logistic regression analysis. The dependent variable is presence of erectile dysfunction (ED) (coded as one). The outcomes presented as a dependent variable (odds ratio, 95% confidence interval). Abbreviations: AF, autofluorescence; BMI, body mass index; DM, diabetes mellitus; eGDR, estimated glucose disposal rate; HbA1c, glycated hemoglobin.

In the univariate regression analysis, HbA1c tends to associate with the presence of ED (1.37, 0.96–1.95; p  = 0.08), while the presence of at least one diabetic complications (4.58, 1.63–12.86; p  = 0.003) and AF (5.69, 1.56–20.72; p  = 0.005) were positively related to ED. After adjustment, HbA1c (1.62, 1.02–2.60; p  = 0.043), the presence of at least one diabetic complications (3.49, 1.10–11.03; p  = 0.03), and skin AF (9.20, 1.60–52.94; p  = 0.01) were positively related to the presence of ED ( Fig. 2B-D ). All predictors in all four models had VIF below 2.1.

The optimal point (Younden's index) of skin AF cutoff to discriminate the presence of ED was equal to 2.2. In the ROC curve analysis, the skin AF values equal or higher than 2.2 indicate the presence of ED with a sensitivity of 70.0% and a specificity of 77.5% ( Fig. 3 ). The area under the curve (AUC) was equal to 0.72 (95% CI: 0.60–0.85).

Fig. 3.

Receiver operating characteristic (ROC) curve analysis of skin autofluorescence (AF) as a predictor for erectile dysfunction (ED) in individuals with type 1 diabetes mellitus (T1DM). Abbreviations: AUC, area under the curve; CI, confidence interval; FPR, false positive ratio; TPR, true positive ratio.

In the multivariate model, both age (1.04, 0.99–1.09; p  = 0.08) and skin AF (3.70, 0.94–14.58; p  = 0.06) tend to associate with the presence of ED. AUC of the model was equal to 0.73 (95% CI: 0.60–0.85).

Discussion

In this cross-sectional study, associations between diabetes features in Polish individuals with T1DM and ED were analyzed. The glycemic control, diabetic complications, and skin AF value are related to the presence of ED. Moreover, skin AF measurements may predict the presence of ED.

Main Findings

The prevalence of ED in our study was 42.9%, similar to the prevalence of ED in a meta-analysis of Kouidrat et al, where it was 37.5%, as well as a prevalence of 50.0% in an assessment of Polish individuals with T1DM by Bak et al. ^{11 32} Most of the entities had mild ED, which is similar to the previous report by Maiorino et al. ³³ The patients with ED were older, had longer diabetes duration, and were more commonly suffering from some diabetic complications. Moreover, ED severity was associated with age and diabetes duration. These findings are unsurprising and have previously been described. ^{32 34 35 36} ED is regarded as a form of diabetic complication and should not be ignored. ³⁷ Only 30 to 50% of men with ED consult a physician regarding this problem. ³⁸ Therefore, doctors should delicately ask the patient about erectile function. ³⁹ In particular, those individuals with poorly controlled diabetes should be vigilantly monitored for the occurrence of ED.

There were no significant differences between patients with ED compared with those without ED in the method of insulin delivery, insulin dose, medication intake, and lifestyle characteristics. Only a minority of the entities used an insulin pump for insulin delivery or took additional drugs. Therefore, the difference between the groups was minimal. Individuals with ED tended to work physically for longer periods a day than those without ED. It may be assumed that people who work physically had a lower educational level, which might be associated with worse glycemic control. ⁴⁰

The presence and severity of ED were negatively associated with eGDR, which reflects the higher IR. However, the relationship was not significant in the adjusted model apart from low VIF for all predictors. p value in multivariate model was equal to 0.12, and odds ratio was consistent with correlation analysis. Therefore, the results may suggest that analysis of larger populations of adult males with T1DM may reveal that association may be significant. Our study does not exclude the possibility that IR may be a background for ED; instead, it seems that IR may have a minor relationship with ED in comparison to other diabetic features. IR among people with type 2 diabetes (T2DM) independently associates with ED. ^{41 42} Therefore, the association between IR and ED in individuals with T1DM requires further studies with larger populations.

AGE accumulation is related to long-term glycemic control. The noninvasive investigation of dermal levels of AGE provides new insights into diabetes progress. In our population, we recorded higher skin AF value in older patients, with longer T1DM duration, with more prevalent diabetic complications, higher HbA1c and triglycerides, and lower high-density lipoprotein (HDL) cholesterol and eGDR (higher IR). AGE accumulation in corpus cavernosum is related to ED. ⁴³ Skin AF value was independently associated with the presence and severity of ED. This relationship was previously described by Kouidrat et al. ¹⁷ This study includes more individuals with T1DM than in paper of Kouidrat et al. Furthermore, the population of this study had similar age, duration of diabetes, and BMI to the individuals with T1DM in study of Kouidrat et al, but lower HbA1c levels, smokers' prevalence, and AF skin values than those participants. Importantly, the model of Kouidrat et al involved all individuals with both T1DM and T2DM who were all older, had higher BMI, HbA1c, smoking prevalence, and skin AF than the study group analyzed in this paper. Moreover, in this study, skin AF values were used to detect the presence of ED, while Kouidrat et al found that IIEF-5 scores are strongly associated with AF values in T1DM subpopulation ( r  = – 0.52; p  = 0.004). In the study of Kouidrat et al, skin AF value ≥ 3.2 determined severe ED with a sensitivity of 60% and a specificity of 87% (AUC = 0.74; 95% CI:0.62–0.86). This study showed that skin AF might be used to detect entities with ED. Therefore, both studies showed similar results and create a background for future investigations. It is worth considering longitudinal studies that may use skin AF for the following: (1) monitoring long-term glycemic control and its association with changes in the severity of ED over time; (2) prediction of future response to the treatment of oral phosphodiesterase-5 inhibitors. A new potential line of ED treatment for individuals with T1DM could be “AGE breaker”—alagebrium/ALT-711 that improves the effects of sildenafil on ED in streptozotocin-induced T1DM in rats. ⁴⁴ For this reason, we may expect that skin AF may be helpful in the selection of subjects susceptible to the treatment of alagebrium or its derivatives; (3) There is an emerging question of whether AF value can predict ED in the future. It may be expected that a dynamic increase of skin AF values in the following measures may precede ED. However, all these hypotheses should be verified in further studies.

Strengths and Practical Implications

This study confirms that ED is prevalent among adults with T1DM. Most of the affected entities had mild ED which could potentially be more susceptible to the treatment. ⁴⁵ The clinical response is the best among young people; thus, the early detection of young individuals with diabetes and ED is crucial. ⁴⁶ Importantly, IIEF-5 was completed by a majority of the respondents. This proves the utility of the questionnaire in clinical circumstances. We highly recommend assessing patients' erectile function using IIEF-5 and monitoring further progress of ED. This study suggests that ED is associated with worse long-term glycemic management and the presence of at least one diabetic complication. It indicates the features of the patients who should be vigilantly monitored on ED. Finally, it was shown that AGE Reader might be useful to detect the presence of ED, and skin AF values are associated with ED severity. However, the study does not prove the clinical utility of the AGE Reader. The device is still used for scientific purposes. The results are encouraging, but the method requires further studies.

Limitations

The study has several limitations. First, the sample size limits the statistical power of the analysis. Up to 30% of the initial entities were excluded, mostly due to missing data such as IIEF-5. Moreover, IIEF-5 may be affected by the presence of a sexual partner, but this intimate information was not collected; thus, the analysis was not adjusted to the presence of a stable sexual relationship. ²³ Second, IR was analyzed using the eGDR formula. This is an acceptable method, but the gold standard is the hyperinsulinemic-euglycemic clamp. ⁴⁷ eGDR correlates with the clamped result and is less time-consuming than the direct method. Third, AGE skin accumulation was measured using an indirect method instead of using skin biopsy. Moreover, serum levels of AGE, a receptor for AGE, and clearance of AGE were not assessed. Fourth, the study population involved only Caucasian males with fair skin; thus, the results cannot be extrapolated for people with darker skin pigmentation. Fifth, most of the study group did not report ED or reported only mild dysfunction. Therefore, the low number of individuals with the moderate or severe form of ED limits the generalization of described relationship. We could not conclude about the relative prevalence and severity of ED in T1DM group in comparison with healthy peers. Sixth, we did not recruit healthy individuals as a control group. We aimed to assess the relationship between ED and diabetes control within T1DM population. Finally, the study has a cross-sectional character, and thus no cause-and-effect relationship can be stated.

Conclusions

ED is prevalent among people with T1DM. The presence of ED in individuals with T1DM is associated with HbA1c, the presence of at least one diabetic complication, and skin AF, but not with IR. Men with poor diabetes control should be asked about ED symptoms.