Introduction

Female sexual dysfunction (FSD) is a multifactorial medical problem affecting both biological and psychological areas [1]. A link between type 2 diabetes mellitus (T2DM) and sexual dysfunction has been known since the 10th century when the Persian physician and philosopher Avicenna spoke of “collapse of sexual function” as a complication of diabetic pathology. However, the search for this connection did not start until about 1950, and the studies were conducted mainly in the male gender. Studies about the FSD were neglected until 1971, when Kolodny published a first work on this issue [2]. Like males, females with T2DM are equally likely to develop complications of the disease, including sexual dysfunction. There is evidence that male erectile dysfunction is an early sign of silent coronary artery disease [3]. For this reason, we hypothesize that FSD can be an early sign of coronary artery disease for females.

The prevalence of FSD is about 20–80% in females with T2DM compared to about 40% of the general female population [4].

Maiorino et al. [5] showed that a normal response to sexual stimuli requires the integrity of the sensory and autonomic nervous system, with consequent smooth muscle release and increased blood flow in the female genitals. Pathologically, T2DM leads to FSD through hyperglycemia, vasculopathy and neuropathy. Hyperglycemia leads to a reduction in the hydration of the membranes, including the vaginal tissues, with consequent reduction of lubrication and development of dyspareunia with infections [6]. Vascular alterations, caused by atherosclerosis and microangiopathy, cause endothelial dysfunction [7, 8]. Finally, diabetic neuropathy alters the normal perception of sexual stimuli and the consequent response [9].

However, the correlation between FSD and the degree of glycemic control, the duration and complications of diabetic disease and cardiovascular risk factors are not so clear. In literature, there is still a lack of evidence, with conflicting data on which parameter correlate [10], or not [11] with FSD. On this basis, the primary aim of this study was to assess the prevalence of FSD in a sample of females with T2DM. The secondary aim was to evaluate if there is a correlation between FSD, and glycemic control, ischemic disease, endothelial dysfunction, autonomic neuropathy, and psychological conditions.

Materials and Methods

Study Design

This observational study was conducted at the outpatient clinic for the cure of Diabetes and Metabolic Diseases of the Department of Internal Medicine and Therapeutics, University of Pavia, Pavia (Italy), and Fondazione IRCCS Policlinico San Matteo, Pavia (Italy).

All procedures were conducted in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Informed consent was obtained from all participants for being included in the study. Trial registration: ClinicalTrials.gov NCT01896648.

Participants

Participants were selected among females referred to our center and identified during routine visits and or through a search of the electronic database of our clinic. We enrolled 81 females aged ≥ 18 years affected by T2DM according to the European Society of Cardiology (ESC) and European Association for the study of Diabetes (EASD) guidelines criteria. Exclusion criteria was type 1 diabetes mellitus or latent autoimmune diabetes in adults (LADA), previous uterine and/ or ovarian surgery and use of hormone replacement therapy (HRT).

Assessments

Participants underwent an initial evaluation that included personal anamnesis and identification of the presence of factors associated with FSD including age, T2DM duration and its complications, glycemic control, hypertension and anti-hypertensive drugs, smoking status and alcohol use. Participants also underwent a physical examination to assess the presence of distal neuropathy and autonomic disorders. Then, these participants underwent the modified Neuropathy Disability Score (NDS), Neuropathy System Score (NSS) and several test such as lying to standing, Valsalva test, deep breathing, and orthostatic blood pressure measurement. We self-submitted participants the Female Sexual Function Index (FSFI) questionnaire to assess the presence of FSD and the evaluation of determinants of the sexual sphere (desire, excitement, lubrication, satisfaction, pain, and orgasm). We also administered the Self Rating Anxiety Scale (SAS) questionnaire and the Self Rating Depression Scale (SDS) questionnaire. All of these were validated in Italian language. We also assessed body weight and body mass index (BMI), waist, abdominal, and hip circumferences, systolic blood pressure (SPB), diastolic blood pressure (DBP), ankle brachial index (ABI), heart rate, glycated hemoglobin (HbA1c) and its mean in the previous years, fasting plasma glucose (FPG), capillary pre- and post-prandial glycemia, fasting plasma insulin (FPI), homeostasis model assessment of insulin resistance index (HOMA-IR), total cholesterol (TC), low density lipoprotein-cholesterol (LDL-C), high density lipoprotein-cholesterol (HDL-C), triglycerides (Tg), homocysteine, lipoprotein(a) [Lp(a)], plasminogen activator inhibitor-1 (PAI-1), soluble intercellular adhesion molecule-1 (sICAM-1), soluble vascular cell adhesion molecule-1 (sVCAM-1), E-selectin, metalloproteinase 2 and 9, high sensitivity-reactive C protein (Hs-CRP). All blood parameters were determined after a 12-h overnight fast except for capillary post-prandial glycemia, determined 2 h after a standardized meal. Venous blood samples were taken for all participants between 08.00 and 09.00 and were drawn from an antecubital vein with a 19-gauge needle without venous stasis. We used plasma obtained by addition of Na2-EDTA, 1 mg/ml, and centrifuged at 3000 g for 15 min at 4 8 C. Immediately after centrifugation, the plasma samples were frozen and stored at −80 °C for no more than 3 months. All measurements were performed in a central laboratory. Body mass index was calculated as weight in kilograms divided by the square of height in meters. Blood pressure measurements were obtained from each participant (using the right arm) in the seated position, using a standard mercury sphygmomanometer (GEIMA) (Korotkoff I and V) with a cuff of appropriate size. Blood pressure was measured by the same investigator, in the morning, after the participant had rested for ≥ 10 min in a quiet room. Three successive blood pressure readings were obtained at 1-min intervals, and the mean of the 3 readings was calculated. Capillary pre- and post-prandial glycemia were measured using the FreeStyle Freedom Lite1 Blood Glucose Monitoring System (Abbott Laboratories, Abbott Park, Illinois, U.S.A.). For a description of how various laboratory parameters were assessed, see our previous work [12].

Questionnaires

The FSFI is a validated questionnaire of 19 items grouped in 6 domains: desire, excitation, lubrication, satisfaction, orgasm and pain to assess the presence of FSD. In the FSFI the score for each item ranges from 0, or 1, to 5 and a total score ≤ 26.55 is suggestive for FSD. The SAS questionnaire quantifies level of anxiety. SAS is a validated 20 item self-report assessment device which include measures of state and trait anxiety. Each question is scored on a scale of 1–4; a total score > 44 is suggestive for anxiety. The SDS questionnaire quantifies level of depression. There are 20 items that rate the four common characteristics of depression: the pervasive effect, the physiological equivalents, other disturbances, and psychomotor activities. Each question is scored on a scale of 1 through 4 and a total score > 49 is suggestive for depression.

Statistical Analysis

Quantitative data were expressed, if normally distributed, as mean and standard deviation (Shapiro test), median and interquartile range if not normally distributed. The qualitative variables were described as counts and percentage. The comparison of quantitative variables between two groups was performed with the Student t test; the chi-square test was used for comparisons among qualitative variables. The connection among quantitative variables were analyzed whit Pearson correlation coefficient [r]. All tests were two-tails, and the limit of significance was 5% (p < 0.05). Analysis was made using the software STATA (version 14.0) (Stata Corporation, 2016, 4905 Lakeway Drive, College Station, Texas 77,485, USA).

Results

Study Sample and FSD Prevalence

Of the 81 enrolled participants, 4 did not complete all the study scheduled procedures leaving 77 participants in the final analysis, 67 (87%) with FSD and 10 (13%) without FSD. The average age of the overall sample was 58.72 ± 9.73 years, 54.8 ± 9.18 years for the participants without FSD, and 59.31 ± 9.73 years for the participants with FSD.

Concomitant Diseases

Comparing the group with and without FSD, there was a significant difference regarding anxiety, with higher prevalence in the group affected by FSD (p = 0.043). There was no difference between two groups regarding depression or other diseases. However, when we stratified the sample according to the presence or absence of a certain disease and analyzed the general FSFI score obtained at the questionnaire, we noticed that participants affected by anxiety, depression and ischemic heart disease obtained a lower FSFI score compared to participants not affected by the diseases (Table 1).

Table 1 Concomitant diseases
Full size table

Concomitant Medications

Regarding the drugs taken chronically by the participants, we did not observe any differences when the sample was divided according to the presence or absence of FSD, in particular 60.00% of participants without FSD were taking oral anti-diabetic drugs versus 76.12 in participants with FSD (p = 0.275); 40.00% of participants without FSD were taking insulin versus 20.90% of participants with FSD (p = 0.230); 70.00% of participants without FSD were taking hypocholesterolemic agents versus 82.09% of participants with FSD (p = 0.398). 20.00% of participants without FDS were taking vasodilatators versus 26.87% of participants with FSD (p = 1.000); 40.00% of participants without FSD were in therapy with ACE-inhibitors versus 67.16% of participants without FSD (p = 0.156). 20.00% of participants without FSD were taking beta-blockers, versus 32.84% of participants with FSD (p = 0.715), 20.00% of participants without FSD were taking diuretics versus 35.82% of participants with FSD (p = 0.480). Finally, 40.00% of participants without FSD, and 55.22% of participants with FSD were taking anti-thrombotic drugs (p = 0.501). However, stratifying the sample on the assumption or not of a single therapy and evaluating the trend of the total score of the FSFI, we noticed that participants taking beta blockers had a worse performance at FSFI questionnaire with a lower score (p = 0.032) (Table 2).

Table 2 Concomitant medications
Full size table

Anthropometric Variables

There were no differences in participants with or without FSD regarding anthropometric variables. However, conducting the analysis using the correlation coefficient of Pearson, there was a positive correlation between SBP (r = 0.476) and DBP (r = 0.652) in participants without FSD (Table 3).

Table 3 Correlation between study parameters and FSFI questionnaire score
Full size table

Glyco-metabolic Control, Lipid Profile and Homocysteine

The mean values ​​of HbA1c, FPG, capillary pre- and post-prandial glycemia, and FPI were lower in the FSD group (Table 4). These differences were at the limit of significance for FPG, capillary pre- and post-prandial glycemia, and FPI, while the difference between HbA1c values ​​was significant (7.49 ± 0.71% in females without FSD versus 6.86 ± 0.88% in females with FSD, p = 0.035, respectively). No differences in lipid profile were observed between the group affected by FSD and the other one. Homocysteine value, instead, was higher in females with FSD compared to those without (17.75 ± 5.32 µmol/l versus 12.39 ± 1.98 µmol/l, p = 0.002, respectively) (Table 4).

Table 4 Glycemic parameters trend and lipid profile parameters
Full size table

Extending the evaluation to a correlation analysis with the score of the FSFI questionnaire, we found an inverse correlation (r = − 0.413) with the HbA1c values ​​in participants not affected by FSD; in the same group there were direct correlations with the values ​​of TC (r = 0.593), LDL-C (r = 0.586) and HDL-C (r = 0.541) (Table 5). When evaluating the score of the FSFI questionnaire stratified by the desire, excitement, lubrication, and satisfaction domains, there was an inverse correlation with the values of HbA1c, the duration of T2DM and HOMA-IR in participants not affected by FSD (Table 6).“

Table 5 Correlation between various parameters and the score obtained in the FSFI questionnaire
Full size table
Table 6 Correlation between glycemic parameters and the score obtained at the various area of FSFI questionnaire
Full size table

Adipocytokines

Participants with FSD had higher levels of E-selectin compared to participants without FSD. No other significant differences were recorded regarding PAI-1, sICAM-1, sVCAM-1, MMP-2, MMP-9, Hs-CPR (Table 7).

Table 7 Adipocytokines
Full size table

Neuropathy Parameters

Participants with FSD performed a lower score at deep breathing test (p = 0.008). There were no other differences in neuropathic variables (Table 8).

Table 8 Neuropathy parameters
Full size table

Discussion

Our data suggest a prevalence of FSD of about 87%, substantially in line with what reported in the literature: the published data, in fact, showed a prevalence of FSD from 20 to 80%. This wide range of prevalence is justified by the fact that the studies were conducted in different ethnic groups [13], and without univocal inclusion criteria [4]. Comparing our data with the study of Esposito et al. [11], carried out on an Italian sample and whose participants average age is similar to our sample (58.72 ± 9.73 vs. 57.9 ± 6.90 years), the prevalence is higher in our trial (87% vs. 53.4%). This difference can be justified by the fact that in the Esposito study a score < 23 was considered as a cut off for the FSF definition at the FSFI questionnaire, and not < 26.55, as performed by many Authors and to which we also conformed.

Regarding the etiology of FSD, previous studies suggested its multi-factorial genesis, with also a psychological involvement: in anxious females, as well as in males, the increased sympathetic tone, and the higher levels of catecholamines generated by this condition could interfere with the smooth muscle release mechanisms, responsible for the correct responses to sexual stimulation of the erectile genital tissue [14].

In line with literature, our data showed a significant difference between participants with and without FSD regarding the prevalence of anxiety. Also, depression plays a leading role: there are many studies that correlate depression with FSD not only in general, but also in correlation with T2DM [11, 15]. The findings of our study support these data: in our study, participants with depression performed a lower FSFI score (3.9 vs. 14.1, p = 0.005).

Vascular dysfunction proved to be one of the main causes of erectile dysfunction in males [16]. The correlation between cardiovascular disease and FSD is debated and needs further study. Our data showed a statistically significant reduction of the total score to the FSFI questionnaire (13.7 [4.20–24.55] versus 3.6 [3.6–4.0], p = 0.010) for participants with ischemic heart disease, thus recognizing a close link between the two conditions.

Considering the main cardiovascular risk factors, we did not find any correlation with arterial hypertension; this is in line with previous literature [11]. This data is even emphasized by the finding of a direct correlation between the values ​​of SBP and DBP and the total score of the FSFI questionnaire in participants not affected by FSD (respectively r = 0.476, and r = 0.652), almost suggesting a protective effect; this mechanism could be explained by the fact that higher pressure values, within certain limits and for a first period, guarantee a more adequate perfusion overcoming initial vascular damage. Based on these aspects, it is not surprising that the means of the total score on the FSFI questionnaire of the participants examined is lower in ones taking beta-blockers (4.8 [3.2–18.55] versus 14.1 [4.2–23.9], p = 0.032), a data confirmed by Miocic et al. [17], and, transversely also in humans, by Burchardt et al. [18]. Differently from what reported by other authors [19], but in agreement with others [4], we did not observe significant differences in the prevalence of dyslipidemia between the two groups, probably because most of our participants (80.51%) were already treated with lipid-lowering drugs. Regarding the most debated aspect in literature, the correlation between FSD and the glycemic control, our data suggest a possible relationship between the two aspects. At a first glance, we observed an apparent contradiction: there was a lower mean of pre- and post-prandial glycemic values, and a lower HbA1c value (6.86 ± 0.88 vs. 7.49 ± 0.71%, p = 0.035) in the group of participants with FSD, compared to the group without FSD, but we need to consider that these are the data collected at the time of enrollment. We must consider that, once developed, FSF led to a very important impact on quality of life [20, 21], this can motivate the patient and the physician to improve her lifestyle to gain a better glycemic control [22] with the improvement of HbA1c. In fact, when we considered HbA1c value trend in the years before the enrollment, we observed an inverse correlation between the total score of the FSFI questionnaire and the mean of the values ​​of HbA1c, confirming that a better glycemic control reduces the risk of developing FSD. We also observed an inverse correlation with the duration of diabetes and HOMA-IR in participants not affected by FSD, further underline that the alteration of these parameters may represent the prelude to the development of FSD.

Another important finding in our study is the higher value of homocysteine ​​in participants with FSD compared to those not affected (17.75 ± 5.32 versus 12.39 ± 1.98 µmol/l, p = 0.002). It is known that hyperhomocysteinemia is a potent risk factor for early atherosclerosis [23], and that a moderate hyperhomocysteinemia is associated with an alteration of the endothelium-mediated arterial dilatation in humans [24]. The angiopathic effect throughout which hyperhomocysteinemia causes FSD is expressed in endothelial dysfunction with consequent altered dilatator effect of nitric oxide [25]. To support the hypothesis of a role of endothelial dysfunction, especially in the presence of increased oxidative stress and inflammatory conditions [26] in the pathophysiology of FSD, there was a higher value ​​of E-selectin (41.77 ± 16.2 vs. 29.15 ± 6.22 ng/ml, p = 0.017) and, even if at the limits of significance, of MMP-9 (54.42 ± 17.41 vs. 43.85 ± 4.87 ng/ml, p = 0.061) in participants with sexual dysfunction. All these parameters showed a pro-inflammatory role in participants with T2DM [27,28,29].

Conclusion

The study showed a prevalence of FSD of 87% in a sample of females with T2DM attending our clinic, in line with literature. The factors involved in the development of FSD could be connected to psychological disorders such as anxiety and depression. There was also a correlation with glycemic control and ischemic heart disease; however, further data is needed to define whether, as for erectile dysfunction, FSD represents an early sign of coronary artery disease. Other important direct correlations were the one with autonomic neuropathy and with endothelial dysfunction, supported by the detection of high levels of E-selectin and homocysteine in participants with FSD.