The Complex Interplay Between Dyslipidemic Subjects Living with Type 2 Diabetes and Metabolic Syndrome and Statin Treatment: A Reflection on Adherence and Persistence in the Real World

Abstract

Introduction

High levels of blood cholesterol are a major risk factor for atherosclerotic cardiovascular disease and require early and consistent utilization of statins as first-line medications to avoid disabling or even deadly consequences. Nevertheless, real-world treatment adherence is significantly lower than that reported in clinical trials. The present paper aims to improve understanding of the reasons behind this phenomenon to help physicians reduce morbidity, mortality, and healthcare costs by reversing this trend.

Methods

We conducted a real-life, observational, noninterventional, retrospective, multicenter study of statin prescription, adherence, persistence, potential side effects, and the rate of early discontinuation in 1959 individuals with type 2 diabetes mellitus (T2D), metabolic syndrome (MS), and moderate hypercholesterolemia, free of prior cardiovascular events in the observation period 2023–2025. In greater detail, we assessed three main areas: the history and features of the underlying pathology—understood as the “clinical condition”—the knowledge and beliefs, and the socioeconomic status of the patients by analyzing anonymized clinical records from the databases routinely used in our joint network units.

Results

Determinants significantly associated with insufficient or no adherence in the first area included older age, complex therapy involving multiple medications, and a history of myalgias and joint pain. In the second area, significant factors were low education, income levels, widowed/divorced, retired/unemployed conditions, and living alone. Finally, in the third area, a lack of awareness regarding cardiovascular disease (CVD) risk, previous negative experiences with statins, negative influence from social media, fear of side effects, a general rejection of medications, and a preference for “natural supplements,” along with the belief that diet is sufficient and that therapy is of short duration, were identified.

Conclusions

Our results prompted the decision to follow up with a prospective phase in which as many subjects as possible from this cohort will be contacted and interviewed to verify the consistency of the observations and to explore adherence further.

Key Summary Points

Why carry out this study?

Statins are the first-line pharmacotherapy for reducing atherosclerotic cardiovascular disease in patients with metabolic syndrome, type 2 diabetes, and hypercholesterolemia

Real-world treatment adherence to statins is significantly lower than that reported in clinical trials

The present paper aims to improve the understanding of the underlying causes of this phenomenon, thereby assisting physicians in reducing morbidity, mortality, and healthcare costs by reversing this trend

What was learned from the study?

Out of a population of 1959 patients, only 36% demonstrate regular adherence to prescribed statins, while the remaining cohort exhibits suboptimal or no adherence

Socioeconomic determinants, such as advanced age, social isolation, low income, and limited education, significantly impact adherence

Misinformation, fear of side effects, and poor risk perception further hinder adherence. These results should encourage clinicians to strengthen patient–physician relationships to overcome patient fears and misconceptions

Introduction

High levels of blood cholesterol carried by low-density lipoproteins (LDL-C) and very-low-density triglyceride-rich lipoproteins (VLDL-C) represent the most frequent dyslipidemic disorder worldwide and are a major risk factor for atherosclerotic cardiovascular disease (ASCVD) [1].

Vice versa, lipid-lowering therapy represents one of the most significant advancements in modern medicine, playing a central role in reducing morbidity and mortality associated with cardiovascular, cerebrovascular, and peripheral vascular diseases. Statins (HMG-CoA reductase inhibitors) are first-line lipid-lowering medications, as stated in the ADA 2024 guidelines [1].

The efficacy of plasma lipid reduction has been demonstrated in numerous large clinical trials involving statins conducted over the past 30 years, confirming the benefits of lowering LDL cholesterol in both primary and secondary prevention [1, 2]. Similar efficacy has been observed in randomized clinical trials (RCTs) using combination lipid-lowering therapy with statins, ezetimibe, and PCSK9 inhibitors [1]. However, it is important to consider that the efficacy observed in large RCTs may translate into “real-life” benefits only with good adherence to the prescribed therapy [3]. In the case of lipid-lowering treatment, adherence to statins—either alone or combined—is significantly lower in the real world than that reported in clinical trials. This phenomenon results in reduced effectiveness of cardiovascular prevention [4, 5]. Poor adherence to pharmacological therapy is now recognized as one of the most concerning public health issues. Indeed, it negatively impacts morbidity, mortality, and healthcare costs. Moreover, being the primary cause of the limited effectiveness of pharmacological treatments, it represents a significant social challenge [3].

Therefore, improving the understanding of the reasons behind this is essential for achieving the goal of high therapeutic adherence, given that it has an even greater impact on population health than the introduction of new therapies [6].

On the basis of these considerations, we conducted a retrospective study on individuals with type 2 diabetes (T2D), metabolic syndrome (MS), and moderate hypercholesterolemia with no history of cardiovascular events to evaluate behaviors concerning statin prescription with special attention to adherence and persistence levels, potential side effects, and rate of early treatment discontinuation. Adherence to pharmacological prescriptions represents individually unpredictable behavior and generally encompasses (i) adherence, defined as taking medications in the doses and timings indicated by the physician (correct dosage); and (ii) persistence, which refers to the continuation of pharmacological therapy over time [7].

Through our daily interactions with patients, we assessed three main areas that can influence individual behaviors: history and characteristics of the underlying pathology (i.e., “clinical condition”), patients’ knowledge and beliefs, and socioeconomic status. Indeed, we hypothesized that a significant portion of the poor or nonexistent adherence to therapy depends on these determinants, along with a lack of awareness regarding untreated hypercholesterolemia as a worsening element in individuals already exposed to a high cardiovascular risk CV-R.

Methods

Ours was an observational, noninterventional, retrospective, multicenter study conducted in a real-life setting. The patients analyzed were recruited from the records of specialized diabetes centers, all operating on the same digital database, following the same operational protocol, and affiliated with the Nefrocenter Research support network.

The above mentioned database (Nefro-Bank) included detailed clinical and social information concerning patients and, in greater detail, allowed physicians to select one or more items from a dropdown menu concerning the following socioeconomic determinants, functional limitations, and adherence-hindering factors:

(a) Socioeconomic determinants: education (poorly educated subjects, diploma, degree), employment status (unemployed, employed, retired), income bracket (low, medium, high), marital status (single, married/cohabiting, divorced/widowed), and household composition (one member, multiple members).

(b) Functional limitations: musculoskeletal issues (primarily myalgia, osteoarthritis, fractures, muscle weakness, sarcopenia), neurological disorders (neuropathies, Parkinson’s disease, stroke, multiple sclerosis), and sensory deficits (auditory, visual).

(c) Adherence-hindering factors: too many pills to take, cardiovascular risk unawareness, negative experiences, internet consultation, distrust of medications, preference for natural products, fear of side effects, forgetfulness, “diet is enough,” lipids normalization, one package is enough, myalgia, joint pain, gastrointestinal disorders, headache, and asthenia.

The study was conducted in accordance with the original Declaration of Helsinki and further amendments, and it was approved by the Scientific Review Committee of Nefrocenter Research and the Ethics Committee of “Campania South” Local Health Authority (protocol code no. 66118, 10 March 2026), as the ethics committee of reference for all the territorial structures participating in the study. All collaborating canters refer to the Scientific Review Committee of Nefrocenter Research and the Ethics Committee of “Campania South” Local Health Authority. The use of Nefro-Bank data has been previously authorized by the Board of Directors of Nefrocenter Research.

All data used in this study were deidentified to ensure patient confidentiality by linking a randomly generated alphanumeric code to personal details with access restricted to a single privacy guarantor. All participants were white and had previously signed consent for the processing of personal data in compliance with privacy regulations for scientific research purposes.

The inclusion criteria required a diagnosis of T2D associated with MS and moderate dyslipidemia characterized by total cholesterol levels between 200 and 240 mg/dL, LDL cholesterol between 100 and 150 mg/dL, and triglycerides exceeding 150 mg/dL. All participants were over the age of 18, independent of gender, with BMI >30 kg/m².

Exclusion criteria were pregnancy or breastfeeding conditions, previous cardiovascular events or revascularization procedures, severe active organ/systemic diseases, and mental disorders. Diagnostic criteria followed the ICD-11 codes [8]. Additionally, all subjects were required to have visited the reference facilities at least four times at 4-month to 5-month intervals and had received a prescription for oral lipid-lowering medications to prevent diabetes-related cardiovascular complications. We limited patients to those prescribed 20 mg of simvastatin, 10 mg of atorvastatin, or 10 mg of rosuvastatin, which are the most frequently used dosages in daily practice as a first-line therapy for moderate hypercholesterolemia. Statin choice depended entirely on individual specialists.

In the monitoring protocol adopted by the Diabetes Units, at each step of the follow up, all subjects underwent a thorough clinical examination, including measurement of waist circumference and assessment of visceral fat with Body Metrix BX2000, Intela-Metrix, Inc. (Brentwood CA, USA), a validated instrument based on ultrasound technology, which allows direct and immediate measurement of fat thickness [9].

We collected the following data at enrollment and after 12 months: body mass index (kg/m², BMI), waist circumference (WC) visceral fat (VF), HbA1c (%), triglycerides (mg/dL, TGL), total cholesterol (mg/dL, T-CH), low-density lipoprotein (LDL), HDL cholesterol (mg/dL, HDL), aspartate transaminase (U/L, AST), alanine transaminase (U/L, ALT), gamma-glutamyl transferase (U/L, γ-GT), alkaline phosphatase (U/L, ALP), creatinine (mg/dL, CRE), creatine phosphokinase (U/L, CK), type of lipid-lowering medication prescribed, and other medications taken. We analyzed data on adherence, persistence, patient-reported reasons for nonadherence/persistence to lipid-lowering treatment, side effects, and any rechallenge of lipid-lowering treatment at 6 and 12 months.

Statistical Procedures

We considered ‘‘adherent and persistent” as those taking at least 80% of the prescribed daily dosage for at least 80% of the observation time (classified as adherent, Group A, or Ad) [10, 11]. Those taking less than 80% of the expected daily dosage in terms of amount and timing were considered to have low adherence/persistence (classified as low adherent, Group B, or LA). Finally, those who took the drug less than 0–50% in terms of time and duration were considered nonadherent and nonpersistent (classified as nonadherent, group C, or NA). Percent adherence was assessed directly by physicians during visits using a structured questionnaire available in the routinely used electronic database. Continuous descriptive data were examined for normality; normally distributed data were reported as mean ± SD. Categorical data were presented as n (%). Univariate logistic regression was performed to screen for potential covariates associated with adherence. Covariates with p < 0.01 were entered into the multivariate model using the forward condition method. A two-sided p-value <0.05 was considered as the lowest level of statistical significance. All analyses were conducted using the SAS System version 9.4 (SAS Institute, Cary, NC, USA).

Results

Examination of databases from the participating study centers identified 5560 patients with cholesterol/triglyceride levels within the desired range. Among these, 1870 had a history of previous cardiovascular events and/or revascularization and were excluded. Of the remaining 3690 patients, only 3215 had all the required information per the study protocol, including adherence/persistence data, side effects attributable to or not attributable to lipid-lowering therapy, and reasons for discontinuation/withdrawal. A further 1156 patients taking lipid-lowering medications other than statins or statins combined with other molecules were also excluded. The remaining 1959 patients who met all inclusion criteria were included in the study in the observation period 2023–2025 (Fig. 1).

Fig. 1.

Flowchart of the study

General Data

Table 1 presents general data for the 1959 patients included in the study. Their ages ranged from 25 to 85 years, with a mean age of 60 ± 27 years. The standard age distribution shows that there were 190 young adults (YA, 25–44 years), 686 middle-aged individuals (MA, 45–59 years), 826 elderly individuals (E, 60–74 years), and 257 seniors (S, >75 years), with a slight prevalence among the two older age groups, totaling 1084 patients (55.3%). Female patients outnumbered male patients by a ratio of 1.27 across all age groups. The duration of diabetes averaged 11 ± 7.3 years; the BMI was 35.5 ± 6.5 kg/m²; waist circumference and visceral fat was outside the normal range in both males and female patients; 415 were smokers (21.2%); and 65 (3.3%) had high alcohol consumption (>20 g daily for men and >10 g daily for women). Total cholesterol (T-C) was 233.4 ± 11.4 mg/dL, HDL cholesterol (HDL-C) was 34.3 ± 6.5 mg/dL, triglycerides (TRI) were 188.7 ± 18.9 mg/dL, and LDL cholesterol (LDL-C) was 136.4 ± 19.6 mg/dL. All other laboratory parameters, such as aspartate transaminase (AST), alanine transaminase (ALT), γ-glutamyl transferase (γ-GT), alkaline phosphatase (ALP), and creatine phosphokinase (CK), were within the normal range. As expected from the narrow ranges of inclusion criteria, no significant differences in baseline blood concentrations were detected among groups. Statins were prescribed without a predetermined criterion, but solely at the discretion of the physician: simvastatin (S) at a dose of 20 mg/day was given to 856 subjects (43.7%), atorvastatin 10 mg/day (At) was given to 742 (38.9%), and rosuvastatin 10 mg/day (R) was given to 382 (18.4%). The total number of drugs taken for comorbidities was 6.6 ± 3.5, ranging from 3 to 14, indicative of comorbidities. Among these, hypertension was present in 56% of cases, respiratory diseases in 27%, sleep apnea in 29%, skin or mucosal fungal infections in 9%, and insulin-related lipodystrophies in 37% of patients treated with insulin.

Table 1.

General characteristic of investigated subjects at baseline (data are numbers or mean ± SD)

Subject number	1959
Male/female (n)	870/1098
Age (years) mean ± SD Range Ratio	60 ± 27 25–85 1.27
Diabetes duration (years)	11.4 ± 7.3
Body mass index (Kg/m2, BMI)	35.5 ± 5.1
Waist circumference (cm) Male Female	119.5 ± 4.6 94.4 ± 4.8
Visceral fat (%)	25.5 ± 5.7
Systolic blood pressure (mm Hg)	132.6 ± 21.7
Diastolic blood pressure (mm Hg)	86.5 ± 9.6
Total cholesterol (mg/dl)	233.4 ± 11.4
HDL cholesterol (mg/dl)	34.3 ± 6.5
Triglycerides (mg/dl)	188.7 ± 18.6
LDL cholesterol (mf/dl)	136.4 ± 19.6
Aspartate transferase (U/L, AST)	25.5 ± 6.5
Alanine transferase (U/L, ALT)	30.7 ± 7.3
γ-glutamyl transferase (γ-GT),	27.7 ± 8.2
Alkaline phosphatase (U/L, ALP),	41.6 ± 7.3
Creatine phosphokinase (U/L, CK)	71.6 ± 12.7
Creatinine (mg/dl, CRE)	0.84 ± 0.3
Number of daily tablets (n) Range	9.1 ± 3.5 3–12
Smoking habit (%)	415 21.2%
High alcohol consumption** (%)	65 3.3%
Statin [n (%)] Simvastatin Atorvastatin Rosuvastatin	856 (43.7) 742 (38.9) 361 (18.4)
Education Primary Diploma Degree	912 (46.5) 832 (42.5) 215 (11.0)
Employment status [n (%)] Unemployed Employed Retired	119 (6.1) 898 (45.8) 942 (48.1)
Income bracket [n (%)] Low Medium High	725 (37.1) 1121 (57.2) 86 (4.4)
Marital status [n. (%)] Single Married/cohabiting Divorced/widowed	219 (11.2) 881 (45.0) 859 (43.8)
Household composition [n (%)] One member Multiple members	493 (25.2) 1466 (74.8)
Functional limitations [n (%)] * Primarily myalgia, muscle weakness Neurological disorders and sensory defects None	807 (41.2) 119 (6.1) 212 (10.8)

(*) 821 subjects, mostly less than 65 years old, were free of any functional limitations

(**) >20 g daily for men and 10 g for women (binge drinkers)

Socioeconomic Determinants

Regarding the socioeconomic determinants, 912 (46.5%) were poorly educated subjects, 832 (42.5%) were high school graduates, and 215 (11.0%) held a university degree; 119 (6.1%) were unemployed, 898 (45.8%) were regularly employed, and 942 (48.1%) were retired. In terms of income, 187 (9.6%) fell within the lower poverty bracket (no tax area), 565 (28.8%) had a low income, totaling 749 (38.2%), while 1121 (57.2%) had a medium income, and 86 (4.4%) had a high income.

Of the study participants, 219 (11.2%) were single, 881 (45.9%) were married or cohabiting, and 859 (43.8%) were divorced or widowed; 493 (25.2%) lived alone, while 1466 (74.8%) lived with others. Furthermore, 807 (41.2%) reported a history of musculoskeletal issues (osteoarthritis, previous fractures, muscle weakness, and sarcopenia), 119 (6.1%) had neurological disorders (neuropathies, stroke, Parkinson’s disease, multiple sclerosis), and 212 (10.8%) experienced balance and sensory issues (dizziness, hearing impairment, visual deficits).

General Changes at End Follow Up

Table 2 summarizes the changes in the main parameters observed at the 12-month follow up, in relation to adherence levels, categorized into three groups: adherent (Ad) at 35.9%, significantly (p < 0.01) higher than the nonadherent (NA) group at 13.7%, but significantly lower than the low adherence (LA) group at 50.4% (p < 0.01). Statin utilization was superimposable in the three groups. In greater detail, Simvastatin was the 35%, 32%, and 33% choice in Groups A, B, and C, respectively; Atorvastatin was the 32%, 35%, and 34% choice in Groups A, B, and C, respectively; Rosuvastatin was the 33%, 34%, and 34% choice in Groups A, B, and C, respectively.

Table 2.

Mean values ± SD, median (Q1–Q3), and range [min–max] of clinical parameters by adherence and persistence groups at the end follow up, and significance of the differences, where **p < 0.01 versus group

Parameter	Adherent (N = 705)	Low adherent (N = 989)	Nonadherent (N = 265)
Male/female	1.22	1.25	1.28
Age (years)	55.5 ± 22.3 Med: 55.5 (40.4–70.6) Range: [45.7–74.4]	62.7 ± 7.8* Med: 62.7 (57.4–68.0) Range: [47.1–78.3]	66.5 ± 8.7** Med: 66.5 (60.6–72.4) Range: [49.1–83.9]
Diabetes duration (years)	10.4 ± 8.3 Med: 10.4 (4.8–16.0) Range: [3.0–27.0]	11.4 ± 5.7 Med: 11.4 (7.6–15.2) Range: [3.7–22.8]	13.4 ± 5.8 Med: 13.4 (9.5–17.3) Range: [3.8–25.0]
Body mass index (BMI)	35.4 ± 7.1 Med: 35.4 (30.6–40.2) Range: [21.2–49.6]	36.5 ± 6.8 Med: 36.5 (31.9–41.1) Range: [22.9–50.1]	34.6 ± 6.3 Med: 34.6 (30.3–38.9) Range: [22.0–47.2]
Waist Circ. (cm)—male	118.2 ± 4.5 Med: 118.2 (115.2–121.2) Range: [109.2–127.2]	120.3 ± 6.2 Med: 120.3 (116.1–124.5) Range: [107.9–132.7]	119.3 ± 5.6 Med: 119.3 (115.5–123.1) Range: [108.1–130.5]
Waist Circ. (cm)—female	96.3 ± 4.3 Med: 96.3 (93.4–99.2) Range: [87.7–104.9]	95.5 ± 3.9 Med: 95.5 (92.9–98.1) Range: [87.7–103.3]	94.8 ± 6.2 Med: 94.8 (90.6–99.0) Range: [82.4–107.2]
Visceral fat (%)	25.8 ± 3.7 Med: 25.8 (23.3–28.3) Range: [18.4–33.2]	25.2 ± 4.3 Med: 25.2 (22.3–28.1) Range: [16.6–33.8]	24.9 ± 4.4 Med: 24.9 (21.9–27.9) Range: [16.1–33.7]
Total cholesterol (mg/dl)	178.5 ± 10.3 Med: 178.5 (171.5–185.5) Range: [157.9–199.1]	228.5 ± 8.8** Med: 228.5 (222.6–234.4) Range: [210.9–246.1]	247.8 ± 12.5** Med: 247.8 (239.4–256.2) Range: [222.8–272.8]
HDL cholesterol (mg/dl)	44.5 ± 7.5 Med: 44.5 (39.4–49.6) Range: [29.5–59.5]	35.8 ± 4.4* Med: 35.8 (32.8–38.8) Range: [27.0–44.6]	33.3 ± 4.7** Med: 33.3 (30.1–36.5) Range: [23.9–42.7]
Triglycerides (mg/dl)	128.2 ± 14.6 Med: 128.2 (118.3–138.1) Range: [99.0–157.4]	188.7 ± 10.5 ** Med: 188.7 (181.6–195.8) Range: [167.7–209.7]	190.3 ± 12.6 ** Med: 190.3 (181.8–198.8) Range: [165.1–215.5]
LDL cholesterol (mg/dl)	87.6 ± 12.6 Med: 87.6 (79.1–96.1) Range: [62.4–112.8]	139.6 ± 10.7 ** Med: 139.6 (132.4–146.8) Range: [118.2–161.0]	141.5 ± 12.7 ** Med: 141.5 (132.9–150.1) Range: [116.1–166.9]
AST (U/L)	23.8 ± 7.5 Med: 23.8 (18.7–28.9) Range: [8.8–38.8]	26.5 ± 7.0 Med: 26.5 (21.8–31.2) Range: [12.5–40.5]	28.7 ± 5.3 Med: 28.7 (25.1–32.3) Range: [18.1–39.3]
ALT (U/L)	41.97 ± 9.3 Med: 42.0 (35.7–48.3) Range: [23.4–60.6]	40.4 ± 5.8 Med: 40.4 (36.5–44.3) Range: [28.8–52.0]	36.5 ± 7.3 Med: 36.5 (31.6–41.4) Range: [21.9–51.1]
$\gamma$-GT (U/L)	33.8 ± 8.2 Med: 33.8 (28.3–39.3) Range: [17.4–50.2]	27.6 ± 6.6 * Med: 27.6 (23.1–32.1) Range: [14.4–40.8]	28.9 ± 6.7 * Med: 28.9 (24.4–33.4) Range: [15.5–42.3]
ALP (U/L)	41.6 ± 7.3 Med: 41.6 (36.7–46.5) Range: [27.0–56.2]	40.8 ± 5.4 Med: 40.8 (37.2–44.4) Range: [30.0–51.6]	41.3 ± 7.1 Med: 41.3 (36.5–46.1) Range: [27.1–55.5]
CK (U/L)	119.8 ± 12.7 Med: 119.8 (111.2–128.4) Range: [94.4–145.2]	67.7 ± 8.3 ** Med: 67.7 (62.1–73.3) Range: [51.1–84.3]	70.9 ± 9.6 ** Med: 70.9 (64.4–77.4) Range: [51.7–90.1]
CRE (mg/dl)	0.84 ± 0.3 Med: 0.84 (0.64–1.04) Range: [0.24–1.44]	0.83 ± 0.6 Med: 0.83 (0.43–1.23) Range: [0.00–2.03]	0.90 ± 0.2 Med: 0.90 (0.76–1.04) Range: [0.50–1.30]
Daily tablets (n)	4.5 ± 3 Med: 4.5 (2.1–6.9) Range: [0.0–11.5]	7.7 ± 2.4 * Med: 7.7 (6.1–9.3) Range: [2.9–12.5]	9.4 ± 4.3 ** Med: 9.4 (6.5–12.3) Range: [0.8–18.0

A (A); *p < 0.05 versus Group A (A); °°p < 0.01 versus Group B (NA)

Mean treatment duration 2.3 ± 1.2 months (range 1–4)

All significant differences are reported in the table using asterisks, and the bold characters are there only to be more visible within the other figures in the table

The men-to-women ratio was similar across the various classes, reflecting the initial condition at enrollment. In contrast, patients in group Ad were significantly younger (55.5 ± 22.3 years, p < 0.01) than those in the other groups (LA = 62.7 ± 7.8 years, NA = 66.5 ± 8.7 years), with a progressively increasing trend. The NA group had diabetes for more years compared with the other groups, also showing a progressive increase with age (Group Ad versus Group NA, RR = 1.87, 95%CI 1.02–2.98).

The BMI was substantially similar across the three groups and did not differ from enrollment; smoking was highly prevalent in the NA group (239 patients, 90.1%) compared with the LA (n = 126, 12.7%; p < 0.001) and Ad groups (n = 50, 7.1%; p < 0.001). High alcohol consumption (>20 g daily for men and >10 g daily for women) was significantly higher (p < 0.01) in the NA group (n = 32, 12.1%) compared with the LA group (n = 20, 2.0%) and the Ad group (n = 12, 1.8%).

Laboratory Parameters at End Follow Up

Laboratory parameters, including AST, ALT, γ-GT, ALP, and creatinine, were similar across the three groups and comparable to baseline values. However, lipid parameters varied with adherence. In fact, in the Ad group (adherent), a significant change was observed compared with enrollment, with a reduction of 23.5% in total cholesterol (p < 0.01), 37.1% in triglycerides (p < 0.01), 35.7% in LDL cholesterol (p < 0.01), and an increase of 29.7% in HDL cholesterol (p < 0.01). These values were all significant compared with the other groups, in which the changes from baseline were insignificant, remaining comparable to baseline levels, with no differences observed between the low adherence group (LA) and the nonadherent group (NA), documenting that in subjects with low to no adherence, there was no effect of statins. Finally, the type of statin taken did not correlate with adherence.

The CK value of Group Ad was significantly higher (p < 0.001) than baseline and both B and C groups at the end of follow up but did not exceed the maximum normal value (171 U/L in men and 145 U/L in women, respectively).

Finally, the NA patients took significantly more medications compared with the LA (p < 0.01) and Ad groups (p < 0.01), with a relative risk (RR) of 3.72 (95%CI 2.23–5.86).

Regarding age, the patients were distributed across various age groups (classified as young adult 25–44 years = 190 subjects, middle age 45–59 years = 685 patients, elderly 60–74 years = 836 subjects, and senior 75+ years = 385 patients, respectively) in relation to adherence levels. The adherence rate was uniformly low across all age groups, ranging from 20% in middle age to 27% in seniors. In comparison, the level of nonadherence was high and similar across age groups (4% YA, 41% MA, 38% E, 44% S, respectively). Low adherence (LA) rates were equal in the extreme age groups (29% in both YA and seniors) and significantly lower (p < 0.05) than in the middle age (39%) and elderly (37%) groups. The association between advanced age and adherence had an RR of 1.55 (95%CI 0.67–1.98).

Table 3 shows the distribution of socioeconomic determinants across the three adherence classes. Regarding “education,” the Ad group had the highest frequency of patients with high school diplomas, while the LA group had the lowest at the basic level (25%). In contrast, the NA group had the highest frequency of poorly educated patients (45%) and the lowest among graduates (5%), indicating that lower education levels correlate with increased nonadherence (RR = 1.78, 95%CI 1.10–2.78).

Table 3.

Socioeconomic status of the three adherence and persistence groups expressed as %

Subject N (%)	Group adherent 705 (35.9)	Group Low adherent 989 (50.0)	Group Nonadherent 265 (13.7)
Education Primary Diploma Degree	25 63 12	45 49 6	53 43 4
Employment status [%] Unemployed Employed Retired	17 52 31	20 47 33	26 36 38
Income bracket [%] Low Medium High	40 55 15	54 34 12	78 20 2
Marital status [%] Single Married/cohabiting Divorced/widowed	19 40 32	27 33 37	30 35 56
Household composition [%] One member Multiple members	41 59	50 50	68 32
Functional limitations [%] Primarily myalgia, muscle weakness Neurological disorders and sensory defects None	21 12 52	40 12 40	47 13 33

In terms of “employment,” unemployed individuals and retirees combined were more frequent in the NA (56%) group than in the LA (53%) and Ad (38%) groups. In comparison, employed individuals accounted for 52% in the Ad group, 47% in the LA group, and 36% in the NA group, indicating a significant association between unemployment or retirement and nonadherence (RR 2.21, 95%CI 1.63–3.12).

We evaluated “income levels” by grouping them into three categories for simplification purposes: the “no tax area” plus “low income,” the second group as “medium income,” and the third group as “high income.” The proportion of the first income category was highest in the NA group (56%), lowest in the Ad group (38%), and intermediate in the LA group (53%). Conversely, the highest income category was most frequently observed in the Ad group (52%), the lowest in the NA group (22%), and intermediate in the LA group, with a significant relationship between low income and nonadherence (RR = 2.93, 95%CI 1.81–3.96).

In terms of “marital status,” married or cohabiting individuals were fairly uniformly distributed across the three adherence classes (41% in Group Ad, 33% in Group LA, and 35% in Group NA), while the divorced/widowed status progressively increased from Group A (23%) to Group LA (37%) to Group NA (56%), with an RR of 2.68 (95%CI 1.70–3.95). The single rate was 19% in Group Ad, 27% in Group LA, and 30% in Group NA, while the divorced/widowed rate was 32%, 37%, and 59%, respectively. Cohabiting/married individuals were 40%, 33%, and 35%, respectively, indicating that nonadherent and low adherent subjects were significantly more numerous among those living alone (RR 1.84, 95%CI 0.92–2.69). In contrast, those living with family were distributed relatively evenly across the three adherence classes.

“Functional limitations” (primarily myalgia) were significantly more frequent in the NA group than in the LA and Ad groups (p < 0.01), with a high RR for functional limitations and nonadherence of 3.25 (95%CI 2.27–5.47). Neuropathies and sensory deficits, however, were comparable across groups.

An annotation of “behavioral motivations” for adherence defects is illustrated in Table 4. A quick examination reveals a significantly (p < 0.01) lower frequency of any motivations and side effects in Group Ad compared with Groups LA and NA. Conversely, in Groups LA and NA, the frequency of many responses is comparable, except for “forgetting,” “discontinuation upon reaching the target,” and “stopping therapy at the end of the initially prescribed package;” factors that were not reported in Group NA. The discontinuation of therapy at the end of the initial packages is present only in Group LA in 35% of cases, indicating a lack of understanding regarding the duration of therapy. Patients in both Groups LA and NA report the simultaneous occurrence of various side effects approximately three times as often as in Group Ad.

Table 4.

Determinants of adherence in the three adherence and persistence groups expressed as %

	Group Adherent	Group Low adherent	Group Nonadherent
Subject N (%)	705 (35.9)	989 (50.4)	265 (13.6)
Too many pills to take	16**	40	42
Cardiovascular risk unawareness	27**	66	78
Negative experience	2**	39	47
Internet consultation	4**	53	69
Distrust of medications	5**	39	46
Preference for natural products	16**	30	35
Fear of side effects	2**	57	65
Forgetfulness	10**	38	–
“Diet is enough”	1**	22	38
Lipids normalization	1**	46	–
One package is enough	1**	31	–
Myalgia	6**	38	31
Joint pain	9**	42	49
Gastrointestinal disorders	7**	27	31
Headache	3**	29	16
Asthenia	12**	34	52
Enzyme abnormalities (CK, ALT)	67 **	5	1

More than one response was provided for each subject

^**p < 0.001 versus Group B and C

It is noteworthy that “musculoskeletal disorders” (primarily myalgias) reported before starting statin therapy are significantly lower in Group Ad (6%, p < 0.01) than in Group LA (38%) with limited adherence and Group NA (31%), with no significant difference between the latter two groups. A similar pattern is observed for: (i) gastrointestinal disorders, 7% in Group Ad (p < 0.01 versus 29% in Group LA, and 31% in Group NA, p < 0.01 Ad versus LA and NA); (ii) headaches, with 3%, 29%, and 26%, respectively (Ad versus LA and NA, p < 0.01); (iii) asthenia, 12% in Ad, 34% in LA, and 42% in NA, respectively (Ad versus LA and NA, p < 0.01).

“Changes in CK” (creatine kinase) levels from baseline were more frequent in Group LA (38%, p < 0.01) compared with both Group Ad (6%) and Group NA (11%). However, there was no corresponding increase in γ-GT levels. Indeed, γ-GT was elevated in only 5% of subjects in this group. Figures 2 and 3 display the main responses of the three groups with differing adherence levels.

Fig. 2.

Percent reasons (%) for nonadherence/nonpersistence at the end of follow up. More than one response was provided for each subject. **p < 0.01 versus B and C; °p < 0.05 versus C. (Group A = adherent), Group B = low adherent, Group C = nonadherent). Items: 1 = too many pills to take, 2 = cardiovascular risk unawareness, 3 = negative experience, 4 = internet consultation, 5 = distrust of medications, 6 = preference for natural products, 7 = fear of side effects

Fig. 3.

Percent reasons (%) for nonadherence/nonpersistence at the end of follow up. More than one response was provided for each subject. **p < 0.01 versus B and C. Group A = adherent, Group B = low adherent, Group C = nonadherent. Items: 8 = forgetfulness, 9 = “diet is enough,” 10 = lipids normalization

In general, all determinants of clinical condition, beliefs and knowledge, and socioeconomic status show comparable patterns in subjects with low and nonadherence, prompting us to consider them together in a single class compared with those with good adherence. Table 5 presents the independent variables significantly correlated with low or no adherence.

Table 5.

Independent variables related to nonadherence: R and 95%CI

	R	95%CI
Advanced age	1.55	0.67–1.98
Several daily tablets	3.72	2.23–5.86
Low education	2.69	1.93–4.51
Unemployed/divorced	1.87	0.92–2.76
Low income	3.21	1.90–5.24
Single/widowed	2.68	1.70–3.95
Single-person household	1.84	0.92–2.69
Functional limitations	3.25	2.17–5.47
History of myalgia	3.18	2.27–6.31
History of arthralgia	2.12	1.78–4.39

Discussion

Despite the universally acknowledged fundamental role of statins in primary and secondary cardiovascular prevention, several patients fail to achieve therapeutic goals due to low adherence [4–6, 12, 13]. This issue is longstanding and remains unresolved, heavily impacting morbidity, mortality, and healthcare costs [6], particularly in high-risk categories such as patients with T2D, MS, and dyslipidemia [14]. Moreover, most of the latter display only moderate abnormalities which, as such, often go unnoticed. From an epidemiological perspective, such an element has significant cardiovascular risk implications at the population level in cases of nonadherence-related protection gaps.

Numerous studies already published on adherence, whether derived from RCTs or dedicated research, have patient attitudes as a central issue. Adherence to pharmacological prescriptions represents an individually unpredictable behavior and generally encompasses: (i) adherence, defined as taking medications in the doses and timings indicated by the physician (correct dosage); and (ii) persistence, which refers to the continuation of pharmacological therapy over time [15].

Both adherence and persistence determine appropriate long-term medication intake. Motivations for low or absent adherence can range from general distrust of all medications to a lack of awareness of cardiovascular risk and fear of side effects [6, 7, 12, 13, 16–18].

This situation requires reflection on the role of physicians in patient care, including providing precise, detailed explanations of therapies and implementing rechallenge (modification or resumption of therapy) after adequate education and training [19]. However, even this measure has often proven ineffective [19–29].

It is also essential to consider the physician’s perception of the extent of the adherence-persistence phenomenon in lipid-lowering therapy when devising operational improvement measures. Furthermore, without periodic inquiries into their patients’ databases, they cannot accurately assess the magnitude of the phenomenon nor be aware of their patients’ actual condition [30].

Our investigation pertains to individuals from a region in Southern Italy characterized by dietary patterns inspired by the Mediterranean diet but often influenced by various external factors, including the globalization of consumption [31]. Other researchers, before us, have addressed the issue of adherence to statins in subjects with and without metabolic diseases; yet they did so in populations significantly different from ours, noting substantial influences on the risk of CVD.

In a study of over 40,000 adults without cardiovascular disease or diabetes, the authors concluded that statin therapy in healthy adults over 70 is associated with reduced cardiovascular events and mortality without an increased risk of developing diabetes [32]. These findings could support the broader use of this class of drugs in primary prevention for individuals over 70 [1]. These data are further corroborated by other studies involving subjects with metabolic diseases [33–35].

In ten observational studies (nine cohorts and one case-control study; n = 815,667) sharing our own inclusion criteria, in older people (aged >65 years) without CVD, statin therapy was associated with a 14% lower risk of CVD death, a 20% lower risk of all-cause mortality, and a 15% lower risk of stroke. In individuals with DM of both genders, the risk of all-cause mortality remained significantly lower even at higher ages (>75 years). Such observations call for RCTs on the effects of statins in patients aged 75 years or older.

In relation to the significant reduction in cholesterol levels among subjects with good adherence and the null effect among those with low adherence or complete nonadherence, our data confirm that low or no adherence does not reduce the risk of cardiovascular disease (CVD). Specifically, the assessment of individual risk over 10 years according to the Cuore Project [36] indicates that, with a mean risk of 22.74% at enrollment for males and 10.15% for females, the group with optimal adherence demonstrates a risk reduction to 11.05% and 5.05%, respectively. In the group with insufficient adherence, there is no risk reduction; instead, an increase is noted (due to the higher average age of these subjects), amounting to 25.75% and 10.45%, respectively, and even more pronounced in the nonadherent group at 35.11% for males and 15.85% for females.

The reasons for nonadherence have also been widely evaluated in the literature, yet the populations studied differed in lifestyle habits and socioeconomic status from ours. Moreover, the Italian National Health System, established in 1978 as a universal organization based on principles of universality, equality, and equity, is free of charge and thus excludes patient out-of-pocket costs as a possible reason for noncompliance.

Determinants significantly associated with insufficient or no adherence in the clinical area included older age, complex therapy involving multiple medications, and a history of myalgias and joint pain. In the knowledge and attitudes area, significant factors were low education and income levels, being widowed/divorced, being retired/unemployed, and living alone. Finally, in the socioeconomic area, we identified a lack of awareness regarding CVD risk, previous negative experiences with statins, negative influence from social media, fear of side effects, a general rejection of medications, and a preference for “natural supplements,” along with the belief that diet is sufficient and that therapy is of short duration. Additionally, smoking and excessive alcohol consumption were predominantly found among subjects with insufficient or no adherence, completing the phenotype of this category of individuals. Outside socioeconomic factors we could not identify a priori, i.e., before follow-up results, possible factors predisposing to a specific adherence behavior, probably because, per inclusion criteria, baseline lipid parameters were within a narrow, homogeneous range.

As our data suggest, the most common and feared effect of statins is myalgia, which has been the focus of various studies, not excluding other possible side effects. The study by Pergolizzi et al. [37] states that, while there may be a causal link between statin use and muscle pain in several individuals, elevated CK levels supporting the statin–pain relationship are inconsistent, ambiguous, and not particularly strong [37].

According to other researchers, the risk of serious muscle injury caused by statins, including rhabdomyolysis, is <0.1%, and the risk of serious hepatotoxicity is 0.001%. Furthermore, the risk of newly diagnosed diabetes mellitus in relation to the population risk of DM is 0.2% per year of treatment and is influenced by the underlying risk of developing DM [38].

Other studies suggest that the overall incidence of muscle symptoms without significantly elevated CK in statin-treated subjects compared with placebo-treated subjects is <1% in RCTs [37], and that CK elevation, when present, may correlate with antiHMG-CoA antibodies [39, 40]. However, according to largely anecdotal reports, <10% of patients partaking in RCTs stop treatment due to muscle cramps without CK alterations [38, 39].

Moreover, reliable data indicate that CK levels do not exceed 10 times normal values in patients taking high doses of atorvastatin, and this is a rare event (<2%; 40). Intolerance-related discontinuation and nonadherence are the most relevant factors that negatively affect primary and secondary prevention [38, 41]. The risk of statin-related myalgia is 10–30% in real life, with atorvastatin being the most commonly implicated, and it also varies with the duration of treatment, significantly higher than that reported in RCTs [38, 41].

In particular, our data show that a light elevation of CK levels—yet keeping within normal limits—is evident in subjects with good adherence. However, 6% reported a prior history of myalgias and 9% reported joint pain. In contrast, in subjects with low or no adherence, CK values were significantly lower than those of adherent subjects, although those reporting myalgia constituted 38% and 31%, respectively, with little to no correlation with statin use.

Limitations and Future Perspectives

The data collected are drawn from historical databases of the diabetes clinics participating in this research and represent a snapshot of elements gathered in real-life settings without the direct presence of patients. Consequently, they are subject to various limitations, including record accuracy. This aspect explains the significant reduction in the number of eligible cases relative to the theoretically much larger recruitable cohort, following careful analysis of data completeness and consistency. Therefore, our results will need confirmation in a further, specifically designed study.

Percent adherence was assessed directly by physicians during visits using a structured questionnaire available in the routinely used electronic database. Such a method may be interpreted as a limitation, although the behavior of different lipid components provides an indirect demonstration of the reliability of adherence indicators, with individual medical interpretations.

However, in the analysis of the repeatedly addressed adherence topic, it is perhaps the first time that socioeconomic determinants and patients’ perceptions/beliefs have been taken into account in a population with specific general characteristics alongside circulating lipid levels, whose variations over time are indirect markers of adherence.

Another interesting element is that, in a geographically favorable country for climate, sunlight, and pollution, preexisting myalgias at the beginning of therapy are a strong predictor of adherence, even in the absence of changes in CK levels.

Interestingly enough, rating our initial approach to the issue of adherence as particularly valuable, the researchers participating in the study decided to follow it up with a prospective phase in which as many subjects as possible from this cohort will be contacted and interviewed to verify the consistency of the observations made and to explore the issue of adherence further.

Conclusions

Our investigation reveals the discouraging finding that only 36% of subjects with metabolic diseases derive tangible benefits from proper adherence to statin therapy. In comparison, the remaining 64% are not protected from CVR. Insufficient information, prejudices, and unjustified fears or distrust of these medications, often amplified by false information disseminated by the media, coupled with unfavorable socioeconomic status, significantly hinder their use.

Further consideration must be given to the negative influence of fake news conveyed through social media, which lacks critical evaluation and verification by qualified professionals. In our study, this was evident in 53% of the low adherence group and 69% of the nonadherent group. The impact of these negative factors across different age groups is uncertain. We can only hypothesize that younger individuals may be more affected by fake news. At the same time, older adults may be more susceptible to misinformation, although this remains speculative in the absence of evidence.

The issues discussed so far are well known and remain unresolved, requiring greater commitment from healthcare providers. Increased surveillance of adherence is necessary, along with implementing actions to enhance physicians’ awareness of adherence issues by improving doctor–patient relationships and adopting a more vigorous approach to sensitizing healthcare personnel to adherence challenges. Lastly, a suitable regulatory framework is needed to effectively control the media’s dissemination of health information, which is often shared without prior critical scrutiny.

Author Contributions

Sandro Gentile: conceptualisation, writing—original draft preparation; Felice Strollo: methodology, writing—review and editing; Giuseppina Guarino, Ersilia Satta, Giuseppe Caccavale, Maria Chiarello, and Rosina De Luca: investigation, data curation, writing—review editing and analysis; Roberta De-Martino and Carmine Romano: software and data curation; Raffaele Marfella: methodology, data curation, and analysis; Maria Carla Sibilio: methodology, data analysis, and validation; Roberta Di Martino: formal analysis, writing—review and editing; and Carmine Romano: supervision. All authors have critically read and agreed to the published version of the manuscript.

Funding

No funding or sponsorship was received for this study or publication of this article.

Data Availability

The datasets analysed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethical Approval

The study was conducted in accordance with the original Declaration of Helsinki and further amendments, and it was approved by the Scientific Review Committee of Nefrocenter Research and the Ethics Committee of “Campania South” Local Health Authority (protocol code no. code no. 66118, 10 March 2026), as the ethics committee of reference for all the territorial structures participating in the study. All collaborating canters refers to the Scientific Review Committee of Nefrocenter Research and the Ethics Committee of “Campania South” Local Health Authority. The use of Nefro-Bank data has been previously authorized by the Board of Directors of Nefrocenter Research. Previous consent was obtained from all subjects registered in the clinic records to use data from research.

Conflicts of Interest

Felice Strollo and Sandro Gentile are editorial board members of Diabetes Therapy. Felice Strollo and Sandro Gentile were not involved in the selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions. The following authors—Giuseppina Guarino, Ersilia Satta, Giuseppe Caccavale, Maria Chiarello, Rosina De Luca, Roberta Di Martino, Carmine Romano, Maria Carla Sibilio, and Raffaele Marfella—have nothing to disclose.