The NHANES is a research initiative that evaluates the health and nutritional condition of individuals residing in the United States. The participants in our study were collected from 7 cycles that occurred between 2005 and 2018. We categorized all participants into two groups: those with diabetes and those without diabetes, based on the widely accepted international diagnostic criteria for diabetes. This study encompassed all individuals diagnosed with diabetes. The criteria for exclusion were as follows: (1) Individuals who are younger than 18 years old or older than or equal to 85 years old, (2) Individuals who are pregnant, (3) Participants who do not have data on plasma creatinine, urinary albumin, urinary creatinine, and statin use.

The population’s fundamental data was gathered by proficient individuals, and all experimental measurements were rigorously conducted throughout the procedure by experts in compliance with the technical requirements given on NHANES’ official website. The NHANES website provides the option to download all data and experimental procedures. The experiments were conducted at a laboratory located in Minnesota.

Data was collected on various demographic factors including age, gender, race, education, finance, marital status, smoking habits, alcohol consumption, physical activity levels. Blood pressure and body mass index(BMI), results from biochemical tests such as high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), glycosylated hemoglobin (Hba1c), alanine aminotransferase(ALT), aspartate transaminase(AST), alkaline phosphatase(ALP), albumin(ALB), globulin(GLB), total cholesterol and triglyceride were also included. Individuals who used any form of statin, either on their own or in conjunction with other medications, were classified as statin users.

Health measurements are performed in specially-designed and equipped mobile centers, which travel to locations throughout the country. DN is identified at this point in time, according to participants’ urine and blood samples which were collected in the NHANES’s mobile examination centers (MECs). Patients with an estimated glomerular filtration rate (eGFR) below 60 mL × min⁻¹ × 1.73 m⁻², as determined by the Modification of Diet in Renal Disease (MDRD) Study equation, or patients with a urine albumin-creatinine ratio(ACR) equal to or more than 30 mg/g, were diagnosed with DN. The MDRD estimate of kidney function was derived using the formula: 175 × plasma creatinine^−1.154 × age^−0.203 (× 0.742 if female; × 1.21 if black) (23).

A survey was conducted during the in-home interview to evaluate the utilization of prescribed drugs. Individuals who indicated in the questionnaire that they had previously used statins were categorized as statin users. The duration of statin usage was categorized based on the specific number of days as follows: less than one year (<365 days), one to three years (≥365 days and <1095 days), three to five years (≥1095 days and <1825 days), five to seven years (≥1825 days and <2555 days), seven to ten years (≥2555 days and <3650 days), and more than ten years (≥3650 days).

Diabetic patients were determined to have diabetes based on self-reports of diabetes, hypoglycemic drugs, or meeting the American Diabetes Association diagnostic criteria (glycosylated hemoglobin ≥ 6.5% or fasting blood glucose > 7.0mmol/L) (24, 25).

We calculated the mean blood pressure from three successive readings taken while the individual was in a relaxed state. The term “hypertension” was defined in the following manner: (1) Mean systolic blood pressure equal to or more than 140 mmHg, (2) Mean diastolic blood pressure equal to or greater than 90 mmHg, (3) Self-reported hypertension, (4) Antihypertensive drug users. The 140/90 mmHg criterion is based on recommendations made by the International Society of Hypertension (26).

Demographics encompass factors such as age, gender, race, educational level, financial situation, and marital status. Race was classified as Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other Race. The education level was categorized into three groups: low (less than a high school diploma), middle (high school graduate or equivalent), and high (college or above). The financial position was categorized into three groups: low, middle, and high, based on the poverty income ratio (PIR ≤ 1, 1<PIR<4, PIR≥4). The marital status was categorized as accompanied, separated, or never married. The body mass index is calculated by dividing an individual’s weight (in kilograms) by the square of their height (in square meters) (27).

The Healthy Lifestyle encompassed data regarding cigarette smoking, alcohol intake, and physical activity. Participants were categorized into two groups based on their self-reported smoking history: Individuals who self-reported having smoked a minimum of 100 cigarettes during their lives, as well as those who smoked fewer than 100 cigarettes. Annual alcohol consumption was classified as “12 cups or more” or “less than 12 cups.” “Average daily physical activity,” “vigorous recreational activity,” and “moderate recreational activity” were utilized to classify physical activity in the questionnaire as either active or inactive. Individuals exhibiting HDL (high-density lipoprotein) levels below 40 mg/dL, total cholesterol levels equal to or beyond 240 mg/dL, triglyceride levels surpassing 200 mg/dL, or LDL levels at or above 160 mg/dL were categorized as having dyslipidemia (28).

The biochemical indicators comprise BUN, Hba1c, HDL, ALB, ALT, AST, ALP, BUN, SCR, Total cholesterol, Triglyceride, LDL, and GLB. Subsequently, all quantities were measured and expressed about internationally recognized standard units.

Data processing and analysis were conducted using R version 4.3.0 (2023-04-21), in conjunction with the Storm Statistical Platform (www.medsta.cn/software). As shown in Prevention (2023) (29), the NHANES reporting requirements adjust and recognize the NHANES sample weights.

Statistical significance was determined for P values less than 0.05. Continuous variables are depicted through comprehensive sample descriptions accompanied by a 95% confidence interval. Discrete counts and weighted proportions represent categorical variables. The atypical distribution is depicted by the median and the interquartile range (Q1-Q3). The participants were categorized into two groups: the DN group and the non-DN group, based on whether their ACR exceeded 30 or their eGFR was below 60. We employed the “multiple imputation” technique to address the issue of missing covariates, mitigating the potential selection bias that could arise from eliminating participants with incomplete data. We aimed to ascertain whether the presence of DN was influenced by the utilization of statins, in comparison to not using them.

Following the construction of the weighted single-factor and weighted multifactor logistic regression models, four distinct models were generated in accordance with the variable types. Model 1, which is the single-factor logical regression model, does not involve any variable adjustments. Model 2 was modified to account for variables such as gender, race, BMI, and finance. Model 3 incorporated the variables from model 2, along with adjustments for smoking, drinking, physical activity, and dyslipidemia. Model 4 included additional variables compared to model 3, namely Hba1c, HDL, ALB, AST, ALT, ALP, LDL, Triglyceride, Total cholesterol, and GLB.

We conduct a more detailed analysis in the group of patients with diabetes to investigate the impact of different types of statins and the length of treatment on DN. In order to evaluate the strength of the results, weighted stratified logistic regression and subgroup analyses were conducted on each of the subgroups. The duration of statin usage was subsequently converted to a categorical variable. Furthermore, to examine for interaction terms between subgroups during effect correction tests, likelihood ratio tests were applied. The nomogram model was constructed using R. Using receiver operating characteristic curve (ROC), the discriminatory authority of the nomogram model in DN risk detection was assessed.

After undergoing screening by the aforementioned stringent criteria ( Figure 1 ), a total of 6483 diabetic patients (mean age: 61.14 ± 13.67 years) were selected from the NHANES 2005–2018 population (n = 70,190). The aforementioned groups comprised over 200 million adults residing in the United States. The baseline characteristics are specified in Supplementary Table 1 . Individuals who presented with DN were found to be older (67.00 (58.00 - 76.00) years), male (53.83%), non-Hispanic white (37.83%), accompanied (87.69%), alcoholics (46.83%), hypertensive (89.83%), possess a high level of education (38.19%), have a moderate financial situation (58.07%), engage in inactive physical activity (74.31%), and utilize statins (53.94%). The glycosylated hemoglobin (7.00 (6.30, 8.20) %, GLB (3.10 (2.80, 3.40) g/dL), ALP (74.00 (60.00, 94.00) g/L), BUN (18.00 (13.00, 25.00) mg/dL), SCR (1.10 (0.83, 1.40) mg/dL), and triglyceride (132.00 (92.00 - 197.00) mg/dL) of patients with DN were all significantly higher than those of patients without DN. The levels of BMI and AST did not differ significantly between the two groups. We also found that in this study, 3,116 participants used statin medication, while 3,367 did not ( Supplementary Table 1 ). There were significant differences between the two groups in terms of age, gender, race, marital status, economic status, hypertension, physical activity, diabetic nephropathy, and several biochemical indicators (P<0.05). Specifically, statin users had a higher incidence of diabetic nephropathy (45.03% vs 35.58%, P<0.001).

The results of univariate logistic regression analysis ( Table 1 ) indicated that the following factors were associated with DN: educational situation, finance, smoking, drinking, physical activity, dyslipidemia, GGT, statin use, Hba1c, BUN, SCR, ALP, GLB, and ALB. Individuals with a high level of education or financial resources are at a reduced risk of developing DN compared to those low or middle. Non-smokers have a reduced risk of developing DN in comparison to smokers. Conversely, individuals who consume alcohol, are physically inactive or have dyslipidemia are at an elevated risk of developing DN. A negative correlation was observed between age, Hba1c, BUN, SCR, ALP, and GLB with the incidence of DN. Conversely, there existed a positive correlation between ALB and the incidence of DN. Furthermore, pharmacological indicators (specifically, statin use) demonstrated a significant association with the exacerbation of DN symptoms (OR=1.32, 95%CI (1.147,1.52), p<0.001) in comparison to those who did not use statins. This association carries a 32% increased risk of developing DN.

Utilizing weighted multivariate logistic regression, the relationship between statin use and DN was investigated. Table 2 details the relationship between statin use and DN, and four logistic regression models were developed, with the effect value expressed as odds ratio and 95% confidence interval. Depending on the magnitude of the effect size, patients treated with statins may have a relatively increased risk of developing DN. The unadjusted analysis (model 1) revealed a statistically significant association between statin use and DN (OR 1.32; p<0.001). This association corresponds to a 32% escalation in the risk of DN among patients who were prescribed statins. The effect value in model 2, which was marginally adjusted, was (OR=1.349, 95%CI (1.168,1.558), p<0.001). This value indicated that patients taking statins had a 34.9% increased risk of developing DN. The effect value in further adjusted model 3, denoted as (OR=1.351, 95%CI (1.165,1.566), p<0.001), suggested that patients undergoing statin treatment faced a 35.1% increased risk of developing DN. The effect value in fully adjusted model 4 was (OR=1.405, 95%CI (1.199,1.647), p<0.001), indicating that the use of statins significantly increased the risk of DN by 40.5% in comparison to those who did not use statins.

To ensure the results’ stability, subgroup and sensitivity analyses were performed. We examined whether DN and statin use differed with regard to age, gender, race and finance. The findings indicated that the association between statin use and DN remained consistent across all subgroups ( Figure 2 ), and no interaction was observed (P > 0.05).

Upon adjusting for model 4, individuals who had used statins for 5 to 7 years had a significantly higher and more pronounced risk (OR=1.472, 95%CI (1.057,2.048), p=0.022) of developing DN compared to those who had used them for < 1 year (OR=1.266, 95%CI (1.054,1.522), p = 0.012), 1 to 3 years (OR=1.334, 95%CI(1.058,1.682), p=0.015) or 3 to 5 years of statin use (OR=1). However, as the duration of use increases (7 to 10 years, p = 0.852; or > 10 years, p = 0.077), this association gradually weakens and becomes statistically insignificant after more than 7 years of use. According to NHANES, the five most frequently prescribed statins in the entire sample from 2005 to 2018 ( Supplementary Table 1 ) were simvastatin, atorvastatin, pravastatin, rosuvastatin, and lovastatin. A significant risk of developing DN was associated with atorvastatin (OR=1.443, 95%CI (1.168,1.784), p = 0.001). Furthermore, simvastatin, the most frequently prescribed statin, exhibited noteworthy and even the most pronounced ascending effects on the development of DN (OR=1.448, 95%CI (1.177, 1.78), P<0.001) ( Table 2 ). We selected variables (age, statin use, alcohol drinking, cigarette smoking, Hba1c, ALB, GLB, ALP) that were more clinically and statistically significant to construct a nomogram model ( Figure 3 ). The discriminatory capacity of the nomogram model developed in this study was confirmed by the area under the curve (AUC) of 68.4% (95% CI: 65.3%–71.5%) as shown in the ROC curve (testing) results ( Figure 4 ).