2025 Volume 7 Issue 3 Pages 71-79
BACKGROUND
Diabetic nephropathy is a common complication of diabetes. We investigated the risk factors for diabetic nephropathy in individuals newly diagnosed with type 2 diabetes.
METHODS
Data from the Japanese Diagnosis Procedure Combination in-patient database (April 2008 to December 2018) were analyzed. The endpoint was subsequent diabetic nephropathy diagnosis or as the time when estimated glomerular filtration rate become <60 ml/min/1.73 m2. Candidate risk factors included age, Hemoglobin A1c, log-transformed triglyceride, total cholesterol, and high-density lipoprotein cholesterol levels, body mass index, and estimated glomerular filtration rate. Eligible individuals with type 2 diabetes without complications who had pre- and post-diagnosis Hemoglobin A1c and serum creatinine measurements, and a history of hypertension or cardiovascular disease pre-diagnosis. Those with pre-existing kidney diseases, nephropathy onset pre-diagnosis, estimated glomerular filtration rate <60 ml/min/1.73 m2 on or before diabetes diagnosis, or age <20 years at diabetes diagnosis were excluded. A multivariate Cox proportional hazards model (p = 0.2 backward selection) was employed.
RESULTS
Of 2,664 eligible individuals (1,775 men, 889 women), 325 men and 175 women developed diabetic nephropathy during follow-up. Cumulative incidence within 5 years was 29.0% in men and 32.5% in women. Age and estimated glomerular filtration rate in both sexes, and total cholesterol in men were significant.
CONCLUSIONS
Age, estimated glomerular filtration rate, and lipid pose potential risks for diabetic nephropathy onset within 5 years of diabetes diagnosis in individuals with hypertension. Collectively, our findings highlight the importance of early monitoring and intervention in this high-risk.
The prevalence of type 2 diabetes and other lifestyle-related diseases in Japan has been increasing annually1), resulting in the onset of various associated complications. Diabetic nephropathy, a typical complication of diabetes, resulted in 127,745 individuals undergoing dialysis in 2018, accounting for 39.0% of all dialysis cases, with an increasing trend2). The increase in diabetes and diabetic nephropathy cases poses a significant societal challenge owing to its impact on both quality of life and healthcare expenditures. Yokoyama et al.3) reported that approximately 1% and 2% of individuals developed diabetic nephropathy within 2 and 5 years post-diabetes onset, respectively based on a survey of 958 individuals with type 2 diabetes between 1965 and 1990. Furthermore, the incidence of diabetes was higher in individuals with hypertension than in those without hypertension4). Combined with the heightened risk of chronic hypertension-induced atherosclerosis in kidney blood vessels4), individuals with hypertension are at high risk of developing diabetic nephropathy. Upon the onset of diabetic nephropathy, kidney function declines, and recovery becomes challenging. Therefore, early diagnosis and intervention are imperative to prevent kidney function decline4).
The United Kingdom Prospective Diabetes Study investigated risk factors for diabetic nephropathy through a randomized controlled trial involving individuals with normoalbuminuria who were newly diagnosed with type 2 diabetes and received hypoglycemic drugs5). Consequently, baseline systolic blood pressure, urinary albumin concentration, serum creatinine levels, and ethnicity (Indian or Asian) were identified as risk factors for microalbuminuria. The Nagano Study (2003) by Katakura et al.6) helped identify blood glucose, blood pressure, and smoking as risk factors for diabetic nephropathy among older Japanese adults. The Japan Diabetes Complications Study by Sone et al.7), which monitored approximately 2,000 individuals with diabetes (mean age, 59 years) for 8 years, helped identify systolic blood pressure, Hemoglobin A1c (HbA1c), age, and triglyceride levels as risk factors. Although previous studies have identified crucial risk factors for diabetic nephropathy, their focus has primarily been on individuals who had already developed diabetes (“prevalent cases”). In prevalent cases, the observation periods for the development of diabetic nephropathy tend to be shorter than in incident cases, potentially leading to an overestimation of the cumulative incidence of diabetic nephropathy among adults and introducing selection bias8), commonly referred to as incidence–prevalence bias9–11). This issue is also recognized as left truncation12–14) in survival analysis excluding the follow-up period could selectively exclude the individuals with higher hazard of incidence12). While the ideal approach is to conduct individual follow-ups starting from the date of diabetes onset, many studies have studied on prevalent cases.
Typically, diabetic nephropathy develops 10 years post-diabetes onset; however, some individuals experience early declines in kidney function and face an elevated risk of early dialysis (early decliners), constituting approximately 14% of cases in Japan15). Given the presence of clinically significant early decliners and the rapidly increasing number of individuals on dialysis, it is important to understand the incidence of diabetic nephropathy and explore risk factors at the early stage of diabetes3). A large database is required to conduct studies with sufficient accuracy on the incidence of diabetic nephropathy in individuals with early stage of diabetes, to obtain conclusive findings with sufficient statistical power, as the incidence of diabetic nephropathy in the early stage of diabetes is reported to be relatively low3).
Accordingly, in this study, we aimed to identify individuals with new-onset diabetes during hypertension treatment, estimate the incidence of diabetic nephropathy, and screen for risk factors using a comprehensive payment inclusion database including laboratory, medical, and endpoint data.
We utilized the nationwide Japanese Diagnosis Procedure Combination database provided by Medical Data Vision, including data from approximately 2.8 million people from 318 hospitals. The period covered was approximately 10 years, from April 2008 to December 2018. This comprehensive database includes information on payments, blood tests, suspected cases, and cases involving blood tests performed during visits to other departments. Using this information, we identified the time of diabetes diagnosis based on changes in blood test values before and after diabetes diagnosis. This analysis was limited to individuals who were being treated for hypertension. The database lacks information on visits to other hospitals, blood pressure, and urinalysis results.
RISK FACTORSThe risk factors for diabetic nephropathy listed in the United Kingdom. Prospective Diabetes Study5), the Nagano Study (2003)6), and the Japan Diabetes Complications Study7), items that exist in this database were considered as candidates. Therefore, candidate risk factors included age, sex, HbA1c (National Glycohemoglobin Standardization Program) levels on or before diabetes diagnosis, blood glucose, total cholesterol, high-density lipoprotein cholesterol, triglyceride levels, estimated glomerular filtration rate (eGFR), and body mass index. HbA1c levels, which represent the diabetes status in recent months, were selected as a measure of blood glucose. For lipids, we included the log-transformed value of triglycerides, total cholesterol, and high-density lipoprotein cholesterol levels. Low-density lipoprotein cholesterol levels, despite being measured in 51.4% of individuals, were omitted owing to their high correlation with total cholesterol. eGFR was calculated using the formula recommended by the Japanese Society of Nephrology16) as follows:
eGFR (ml/min/1.73 m2) = 194 × creatinine−1.094 × age0.287 (×0.739 for women).
STUDY PARTICIPANTS SELECTIONInclusion criteria comprised the 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) code “E119”, which defined as type 2 diabetes without complications. Then we limited to individuals who had blood tests including pre- and post-diabetes HbA1c and serum creatinine measurements, which would be used as reference for blood test results in the diagnosis of diabetes and diabetic nephropathy. Target individuals were also limited to those who had a history of hypertension or cardiovascular disease (ICD-10 code “I”) preceding diabetes diagnosis. That is individuals treated for hypertension are at high risk of developing diabetic nephropathy4). Only 274 individuals were not treated for hypertension in this database. Therefore, they were considered a different cohort. We excluded 10,998 individuals based on the following exclusion criteria: kidney-related injuries or diseases (except acute pyelonephritis and simple kidney cysts before diabetes diagnosis), nephropathy onset before diabetes diagnosis, eGFR <60 mL/min/1.73 m2 on or before diabetes diagnosis, and age <20 years at diabetes diagnosis. Since there were no urinalysis results in this database, we assumed that eGFR alone could confirm diabetic nephropathy in individuals with eGFR <60 mL/min/1.73 m2 17,18).
ENDPOINTSThe endpoint of this study was the duration from the initial diabetes diagnosis to the onset of diabetic nephropathy. This onset was defined by assigning ICD-10 codes E112, E132, and E142 related to “diabetic nephropathy” to confirmed injuries or diseases, or as the time when eGFR become <60 mL/min/1.73 m2. Supportive analysis was conducted based on changing the eGFR cut-off value to <30 mL/min/1.73 m2. For individuals who did not develop diabetic nephropathy during the study period, the endpoint was censored at the discontinuation date, defined as the first day of the last visit month (day 1).
STATISTICAL ANALYSISThe cumulative incidence of diabetic nephropathy was estimated using the Kaplan–Meier method for men and women, separately. Cox proportional hazard models, with candidate risk factors as covariates, were used for identifying key prognostic factors. Hazard ratios and 95% confidence intervals were calculated per 10 units for age, total cholesterol, high-density lipoprotein cholesterol levels, and eGFR, and per 5 units for body mass index. We considered full model including those six factors and backward variable selection model. The significance level for removing a covariate form the model was set at p = 0.2 based on Wald test. Note that age was not considered for backward variable selection to be always included in the model. Those model constructions and estimations were done for men and women, separately. All analyses were conducted by SAS version 9.4 (Cary, NC, USA).
Among the 10,998 individuals first diagnosed with type 2 diabetes who met the inclusion criteria (Fig. 1), 1,124 individuals with kidney-related disease, 167 individuals with nephropathy onset before diagnosis, 2,749 individuals with eGFR <60 before diagnosis and 20 individuals aged under 20 years at the diagnosis, were excluded. In addition, individuals without age, triglyceride, total cholesterol, high-density lipoprotein cholesterol, or body mass index were excluded. Thus 2,664 individuals were included in the analysis population. Table 1 presents the baseline characteristics of the individuals stratified according to sex, comprising 1,775 men and 889 women. The mean age was 65.7 years for men and 69.6 years for women. The mean follow-up period was 2.14 years for men and 2.18 years for women (Table 2). Throughout the follow-up period, 325 men and 175 women developed diabetic nephropathy.
Abbreviations: ICD-10, the 10th revision of the International Statistical Classification of Diseases and Related Health Problems; HbA1c, Hemoglobin A1c; HDL, high-density lipoprotein; eGFR, estimated glomerular filtration rate
| Variables | Units | Mean | Standard deviation | Minimum | Median | Maximum | |
|---|---|---|---|---|---|---|---|
| Men (n = 1,775) |
Age | years | 65.7 | 12.1 | 20 | 67 | 92 |
| HbA1c | % | 7.3 | 1.9 | 4.1 | 6.8 | 17.0 | |
| Triglyceride | mg/dL | 151.0 | 116.0 | 11 | 119 | 1477 | |
| log (Triglyceride) | 4.8 | 0.6 | 2.4 | 4.8 | 7.3 | ||
| Total cholesterol | mg/dL | 180.8 | 43.2 | 36 | 178 | 431 | |
| HDL cholesterol | mg/dL | 46.8 | 15.2 | 1.9 | 44 | 131 | |
| eGFR | mL/min/1.73 m2 | 81.9 | 19.3 | 60.0 | 77.1 | 239.8 | |
| Body mass index | kg/m2 | 24.7 | 4.5 | 13.5 | 24.2 | 57.5 | |
| Women (n = 889) |
Age | years | 69.6 | 12.7 | 20 | 71 | 98 |
| HbA1c | % | 7.2 | 1.8 | 4.4 | 6.6 | 16.1 | |
| Triglyceride | mg/dL | 130.3 | 80.6 | 21 | 108 | 654 | |
| log (Triglyceride) | 4.7 | 0.5 | 3.0 | 4.7 | 6.5 | ||
| Total cholesterol | mg/dL | 195.3 | 49 | 72 | 190 | 515 | |
| HDL cholesterol | mg/dL | 53.6 | 16.3 | 6 | 53 | 118 | |
| eGFR | mL/min/1.73 m2 | 83.8 | 21.9 | 60.0 | 78.1 | 238.8 | |
| Body mass index | kg/m2 | 24.4 | 5.1 | 12.8 | 23.8 | 47.1 |
Abbreviations: HbA1c, Hemoglobin A1c; HDL, high-density lipoprotein; eGFR, estimated glomerular filtration rate
| N | Number of events | Follow-up periods | |||
|---|---|---|---|---|---|
| Mean | Minimum | Maximum | |||
| Men | 1,775 | 325 | 2.14 | 0.00 | 7.44 |
| Women | 889 | 125 | 2.18 | 0.00 | 8.10 |
The cumulative incidence of diabetic nephropathy was 21.0% in men and 20.6% in women at 3 years, with a subsequent increase to 29.0% in men and 32.5% in women within 5 years (Fig. 2). Cox regression analyses were conducted separately for men and women, based on full model and backward variable selection model (Table 3). The variable selection model selected age and eGFR that were statistically significant factor both for men and women (age, hazard ratio: 1.30, 95% confidence interval: 1.17–1.46, p < 0.001 in men; hazard ratio: 1.31, 95% confidence interval: 1.13–1.52, p < 0.001 in women. eGFR, hazard ratio: 0.66, 95% confidence interval: 0.60–0.73, p < 0.001 in men; hazard ratio: 0.75, 95% confidence interval: 0.67–0.83, p < 0.001 in women). Additionally, total cholesterol was statistically significant for men (hazard ratio: 0.96, 95% confidence interval: 0.93–0.99, p = 0.005). Also, log-triglyceride was selected for both sexes (hazard ratio: 1.23, 95% confidence interval: 1.00–1.52, p = 0.052 in men; hazard ratio: 1.24, 95% confidence interval: 0.91–1.69, p = 0.166 in women), and total cholesterol was selected for women (hazard ratio: 0.97, 95% confidence interval: 0.93–1.00, p = 0.082) (Table 3, left).
| Variables | backward selection | full model | |||||||
|---|---|---|---|---|---|---|---|---|---|
| P value | Hazard ratios | 95% confidence intervals | P value | Hazard ratios | 95% confidence intervals | ||||
| Men (n = 1,775) (event = 325) |
Age (by 10years) | <.001 | 1.30 | 1.17 | 1.46 | <.001 | 1.28 | 1.14 | 1.44 |
| HbA1c (%) | 0.623 | 1.02 | 0.95 | 1.09 | |||||
| log (Triglyceride) | 0.052 | 1.23 | 1.00 | 1.52 | 0.083 | 1.22 | 0.97 | 1.53 | |
| Total cholesterol (by 10 mg/dL) | 0.005 | 0.96 | 0.93 | 0.99 | 0.026 | 0.96 | 0.93 | 1.00 | |
| HDL cholesterol (by 10 mg/dL) | 0.549 | 0.97 | 0.89 | 1.06 | |||||
| eGFR (by 10 mL/min/1.73 m2) | <.001 | 0.66 | 0.60 | 0.73 | <.001 | 0.66 | 0.60 | 0.73 | |
| Body mass index (by 5 kg/m2) | 0.213 | 0.90 | 0.77 | 1.06 | |||||
| Women (n = 889) (event = 175) |
Age (by 10years) | <.001 | 1.31 | 1.13 | 1.52 | <.001 | 1.32 | 1.13 | 1.55 |
| HbA1c (%) | 0.923 | 1.00 | 0.90 | 1.10 | |||||
| log (Triglyceride) | 0.166 | 1.24 | 0.91 | 1.69 | 0.294 | 1.19 | 0.86 | 1.65 | |
| Total cholesterol (by 10 mg/dL) | 0.082 | 0.97 | 0.93 | 1.00 | 0.189 | 0.97 | 0.94 | 1.01 | |
| HDL cholesterol (by 10 mg/dL) | 0.541 | 0.97 | 0.87 | 1.08 | |||||
| eGFR (by 10 mL/min/1.73 m2) | <.001 | 0.75 | 0.67 | 0.83 | <.001 | 0.75 | 0.67 | 0.84 | |
| Body mass index (by 5 kg/m2) | 0.510 | 1.06 | 0.90 | 1.25 | |||||
Abbreviations: HbA1c, Hemoglobin A1c; HDL, high-density lipoprotein; eGFR, estimated glomerular filtration rate
Age and eGFR remained as a statistically significant factor in the full model both for men and women (age, hazard ratio: 1.28, 95% confidence interval: 1.14–1.44, p < 0.001 in men; hazard ratio: 1.32, 95% confidence interval: 1.13–1.55, p < 0.001 in women. eGFR, hazard ratio: 0.66, 95% confidence interval: 0.60–0.73, p < 0.001 in men; hazard ratio: 0.75, 95% confidence interval: 0.67–0.84, p < 0.001 in women). Total cholesterol also remained as a statistically significant factor for men (hazard ratio: 0.96, 95% confidence interval: 0.93–1.00, p = 0.026)(Table 3, right).
The results of supportive analysis based on the endpoint eGFR cut-off value of 30 mL/min/1.73 m2 showed the similar trend in point estimates of hazard ratios (Table 3-A).
| Variables | backward selection | full model | |||||||
|---|---|---|---|---|---|---|---|---|---|
| P value | Hazard ratios | 95% confidence intervals | P value | Hazard ratios | 95% confidence intervals | ||||
| Men (n = 1,775) (event = 105) |
Age (by 10years) | 0.189 | 1.12 | 0.95 | 1.33 | 0.429 | 1.08 | 0.89 | 1.32 |
| HbA1c (%) | 0.795 | 1.01 | 0.91 | 1.13 | |||||
| log (Triglyceride) | 0.260 | 1.25 | 0.85 | 1.84 | |||||
| Total cholesterol (by 10 mg/dL) | 0.117 | 0.96 | 0.91 | 1.01 | 0.060 | 0.95 | 0.89 | 1.00 | |
| HDL cholesterol (by 10 mg/dL) | 0.043 | 0.86 | 0.74 | 1.00 | 0.100 | 0.87 | 0.74 | 1.03 | |
| eGFR (by 10 mL/min/1.73 m2) | 0.272 | 0.94 | 0.84 | 1.05 | |||||
| Body mass index (by 5 kg/m2) | 0.303 | 0.87 | 0.66 | 1.14 | |||||
| Women (n = 889) (event = 63) |
Age (by 10years) | 0.006 | 1.41 | 1.11 | 1.81 | 0.006 | 1.43 | 1.11 | 1.86 |
| HbA1c (%) | 0.383 | 1.06 | 0.93 | 1.22 | |||||
| log (Triglyceride) | 0.152 | 1.46 | 0.87 | 2.44 | 0.162 | 1.45 | 0.86 | 2.45 | |
| Total cholesterol (by 10 mg/dL) | 0.103 | 0.95 | 0.89 | 1.01 | 0.085 | 0.94 | 0.88 | 1.01 | |
| HDL cholesterol (by 10 mg/dL) | 0.011 | 0.79 | 0.66 | 0.95 | 0.019 | 0.80 | 0.67 | 0.96 | |
| eGFR (by 10 mL/min/1.73 m2) | 0.775 | 0.98 | 0.87 | 1.11 | |||||
| Body mass index (by 5 kg/m2) | 0.525 | 1.09 | 0.84 | 1.42 | |||||
Abbreviations: HbA1c, Hemoglobin A1c; HDL, high-density lipoprotein; eGFR, estimated glomerular filtration rate
In this study, we estimated the incidence of diabetic nephropathy subsequent to diabetes diagnosis and examined the risk factors associated with its onset in individuals undergoing hypertension treatment. Leveraging the extensive database from Medical Data Vision, we successfully extracted information for 500 cases of early diabetic nephropathy, a task challenging to achieve with conventional datasets. Since the onset of diabetes is a continuous pathological change, estimating its exact commencement date is challenging. To address this, we utilized ICD-10 codes, definite injury/illness names, eGFR, HbA1c levels, and other variables to reasonably estimate the onset of diabetes, ensuring an adequate number of events as possible to consider diabetic nephropathy as an endpoint.
The cumulative 5-year incidence of diabetic nephropathy was 29.0% in men and 32.5% in women, that was notably higher than that reported by Yokoyama et al.3) (~2%). Unlike Yokoyama et al.3) who define diabetic nephropathy by “persistent” proteinuria, the present study in which albuminuria is not available in the data has some limitations in defining outcomes, as we will discuss later. Additionally, this discrepancy may be influenced by the inclusion of individuals who visited the university hospital and those who were referred from other hospitals in the study by Yokoyama et al. These individuals may have had diabetes for a longer period than the diagnosis indicated.
Age and eGFR level in both sexes, and total cholesterol in men were risk factors for the onset of diabetic nephropathy with hypertension who were newly diagnosed with type 2 diabetes, while it was suggested that triglycerides in both sexes, and total cholesterol in women may also be relevant . Triglyceride levels have been linked to an increased risk of developing chronic kidney disease19,20). Concordantly, Sone et al.7) reported triglyceride levels as a risk factor for both the onset and progression of diabetic nephropathy. Jinyang et al.21) highlighted the therapeutic effect of statins on early diabetic nephropathy because they could eliminate the underlying causes, emphasizing the role of lipids in the onset and progression of diabetic nephropathy.
Unlike conventional studies that include individuals who have already developed diabetes, this study included only incident cases of diabetes. This allowed us to avoid the incidence-prevalence bias. It has been noted that studies enrolling already exposed individuals are prone to selection bias12,13), and that omission of the period at the beginning of diabetes due to left-sided censoring may miss the detection of risk factors in the early period. Brookhart13) pointed out the superiority of follow-up from the beginning is similar to the intention-to-treat analysis in a parallel randomized controlled trial.
Our study has the following limitations. First, although this study selected individuals based on a new diagnosis of diabetes, it is possible that many of the individuals were diagnosed late because of delays in visits to medical institutions, and the existence of individuals who potentially did not visit medical institutions and did not receive treatment for a long period of time after the onset of their disease cannot be ruled out. Second, it was also unclear from the data whether the participants had previously been diagnosed with diabetes and treated by another health care provider. These are limitations of this study based on real-world data, and it is necessary to explore ways to exclude such individuals in the future. Third, there was also the issue of whether the diagnosis and ICD-10 code match. Yamana et al.22) conducted a study on the correctness of diagnostic names in Japanese Diagnosis Procedure Combination data and reported a specificity of 96.7%, sensitivity of 52.2%, positive predictive value of 72.7%, and negative predictive value of 92.2% for diabetes.
There is also a major concern about the measurement of outcome. Diabetic nephropathy is considered in all diabetic individuals because it is extremely common and asymptomatic until late in life. Therefore, all diabetic individuals should be screened with regular urinalysis to detect and treat the disease as early as possible. In fact, routine HbA1c examination (≥1 per year), serum creatinine and lipid tests were reported to be conducted for most individuals, especially in the insulin injection and oral antihyperglycemic agent groups23). On the other hand, according to Mitri et al.24),which reviewed 6 studies that examined the quality status of diabetes care in Japan from 2010 to 2019, the rate of screening for urine albumin was very low. Since the serum creatinine screening was optimal, physicians could have been diagnosing nephropathy only from the serum creatinine level, without testing urine microalbumin excretion. However, it is unclear from the data whether the participant has been accurately diagnosed with diabetic nephropathy, even though they actually have diabetic nephropathy. Therefore it was possible that nephropathy was missed even in our study. In addition, it should be noted that the issue of outcome measurement may include the fact that there is no guarantee that physicians will actively assign diagnostic codes for diabetic nephropathy for insurance reimbursement. Yamana et al.22) reported that specificity was 99.7%, sensitivity was 29.4%, positive predictive value was 83.3%, and negative predictive value was 96.1% for chronic complications of diabetes. Those possibilities of under diagnosis with diabetic nephropathy could be reduced by adding the criteria of eGFR <60 mL/min/1.73 m2 to the endpoint.
We could not investigate whether diabetic retinopathy (based on ICD-10 code E113) could predict diabetic nephropathy because the individuals in this study was defined as type 2 diabetes without complications. Because the presence of diabetic retinopathy is often confirmed in daily practice to screen for microvascular complications23), it would be useful to conduct a study including diabetic retinopathy in a prediction model.
This study was limited to individuals who underwent treatment for hypertension before being diagnosed with diabetes. The impact of the factors listed as missing information in this database study should be examined again in a prospective study. Although the Medical Data Vision database did not include blood pressure measurements and could not be included in the analytical model, limiting the inclusion to individuals who were on hypertension treatment before being diagnosed with diabetes could reduce the negative effect not to include blood pressure in the model. Despite these constraints and the unavailability of all desired information, we extracted information for a substantial number of cases adhering to clinical study selection criteria and succeeded in identifying risk factors among routine medical care blood indices.
We chose to analyze the data separately by sex. It is well established that the absolute risk of cardiovascular disease is higher in men with diabetes than women with diabetes24). Sex differences in the risk of developing microvascular and macrovascular complications in diabetic individuals have also been reported27). Considering the number of diabetic nephropathy cases in the Cox regression-analyzed population (325 men and 175 women) was somewhat large, we felt that our study should also examine the onset of diabetic nephropathy and screen for risk factors separately for men and women. Our findings helped identify crucial risk factors, aligning with that of previous studies. In future investigations, we plan to replicate this study with a longer follow-up period.
The cumulative incidence of diabetic nephropathy within 5 years in new-onset type 2 diabetes individuals treated for hypertension was 29.0% in men and 32.5% in women. Age and eGFR in both sexes, and total cholesterol in men were identified as the strongest risk factor for diabetic nephropathy. Followed by an associated trend also observed with triglyceride levels in both sexes, and total cholesterol in women. Our results may contribute to reducing the number of individuals requiring dialysis by identifying individuals susceptible to diabetic nephropathy in the early stages of diabetes and providing them with intensive care in routine medical care.
The authors declare no conflicts of interest in relation to the work presented in the manuscript.
This database is available for purchase from Medical Data Vision.
We thank Dr. Hitomi Tanaka, the director of Hijiribashi Clinic and a nephrologist, for her valuable advice. This research was supported by Japan Agency for Medical Research and Development under Grant Number JP21lk0201701.
