Incidence and Predictors of Diabetic Nephropathy among Type 2 Diabetes Mellitus Patients, Southern Ethiopia

Background Diabetic nephropathy is the most common cause of end-stage renal disease, and it brings high morbidity and mortality. Globally, the predominant rise in type II diabetes prevalence significantly increases the incidence of diabetic nephropathy. Therefore, timely diagnosis and prompt management of diabetic nephropathy and early identification of predictors are essential. Thus, this study aimed to determine the incidence and predictors of diabetic nephropathy among type II diabetes mellitus patients. Methods A retrospective follow-up study was conducted among 532 type II diabetes patients who enrolled at Hawassa University Comprehensive Specialized Hospital from January 1, 2012, to December 31, 2021. A simple random sampling technique was used to select the study participants. The extracted data were entered into EpiData version 3.1 and analyzed by Stata version 14. A bivariate and multivariable Cox proportional hazard regression analysis was fitted to identify predictors of diabetic nephropathy. The Cox proportional hazards assumption was checked using the Schoenfeld residual test, and the goodness of fit of the model was checked using the Cox–Snell residual test. An adjusted hazard ratio with a 95% confidence interval and P values were used to identify statistically significant predictors. Results The overall incidence rate of diabetic nephropathy was 2.71 cases (95% CI: 2.12, 3.47) per 1,000 person-months of observation. Age (AHR = 1.027; 95% CI = 1.005, 1.049), fasting blood sugar (AHR = 1.010; 95% CI = 1.007, 1.013), and systolic blood pressure (AHR = 1.050; 95% CI = 1.031,1.069) were significant positive predictors of diabetic nephropathy, whereas the duration of diabetes longer than five years (AHR = 0.20; 95% CI = 0.09, 0.44) was a protective predictor for the development of diabetic nephropathy. Conclusion The incidence rate of diabetic nephropathy was high. Age, fasting blood sugar, systolic blood pressure, and duration of diabetes were found to be independent predictors of diabetic nephropathy. To overcome this public health problem, prompt and effective strategies should be designed based on identified predictors to prevent the development of diabetic nephropathy.


Introduction
Diabetes is a group of metabolic diseases characterized by hyperglycemia resulting from defects in insulin secretion, insulin action, or both [1].Diabetes mellitus (DM) is estimated to afect 463 million people by the year 2019 [2] and 537 million by 2021 globally, and this accounts for 10.5% of the global adult population aged 20 to 79 years.It afects 24 million people in Africa, with 1.9 million (3.3%) living in Ethiopia.According to the International Diabetes Federation, the number of individuals with DM is estimated to rise to 783 million by the year 2045 worldwide.By 2045, the number of people with diabetes will have increased by 94% in low-and middle-income countries and by 129% in Africa [3].Te increasing prevalence is mainly due to nutrition transitions, rapid urbanization, and sedentary behaviors [4].
Diabetes is a major cause of kidney failure, blindness, heart attacks, stroke, nerve damage, and lower limb amputation that signifcantly impacts the quality of life of an individual [5,6].Type 2 diabetes mellitus (T2DM) is a chronic, noncommunicable disease, and its complications have contributed tremendously to the burden of mortality and disability worldwide [7].It constitutes over 95% of people with diabetes, so the majority of people who develop diabetic nephropathy (DN) do so because of T2DM [5,8].Globally, the predominant rise in type II diabetes prevalence over the next two decades will signifcantly increase the incidence of DN [9].Diabetic nephropathy (DN), or diabetic kidney disease, is "a syndrome characterized by the presence of pathological quantities of urine albumin excretion, diabetic glomerular lesions, and loss of the glomerular fltration rate (GFR) in diabetics" [10].
Diabetic nephropathy is one of the most feared microvascular complications of DM, and it is highly prevalent throughout the globe [11,12].It is a major healthcare challenge that afects 20% to 50% of people living with diabetes and 40% of T2DM patients [8,12].Diabetes has become one of the leading causes of end-stage renal disease (ESRD) worldwide as a result of the epidemic [8,10].Approximately 20% to 40% of type II diabetes patients with microalbuminuria progress to clinical nephropathy, of which nearly 20% progress to ESRD after developing overt nephropathy [10].Te majority of the patients with DN die from ESRD [13].
All-cause mortality among diabetic patients with DN is nearly 30 times higher compared to those of diabetic patients without DN [11].Tis will have signifcant social and economic consequences, particularly in developing countries [9].Diabetes-related complications relate to the direct cost of hospitalization and indirect costs, which are the most signifcant contributors to premature mortality, disability, and absenteeism.Early detection and improved management of complications of diabetes will have benefts not only for people living with diabetes but also for the wider health economy [2].
In 2019, worldwide DN among T2DM was responsible for 2.5 million incident cases, 129.56 million patients, 405,990 deaths, and 9.87 million disability-adjusted life years (DALY), which increased by 156.49%, 94.78%, 172.39%, and 141.73%, respectively.In Africa, it was associated with 194.97 thousand incident cases, 9.9 million prevalent cases, 42,680 deaths, and 1 million DALY, which increased by 56.42 thousand, 4 million, 18,910, and 455,170 from 1990, respectively.Te increasing rate of DN is associated with age, sex, and anemia [14].Te pathogenesis for the development and progression of DN is multifactorial and complex, which is the result of metabolic disorders, hemodynamic abnormalities, and hormone synthesis [15,16].
Prior studies revealed that the predictors for the development of DN in developed countries were age, sex, albuminuria, hypertension, duration of DM, myocardial infraction, dyslipidemia, SBP, smoking, waist circumference, poor glycemic control, previous retinopathy, and previous cardiovascular disease [17,18], while in developing countries, age, sex, proteinuria, hypertension, HDL-C, body mass index, history of cardiovascular disease, poor glycemic control, type of antidiabetic medication, baseline GFR, coronary heart disease, anemia, duration of diabetes, SBP, and serum uric acid were the main predictors for the development of DN [19][20][21][22][23][24][25][26].Management of modifable risk factors may help to reduce DN incidence soon [27].Trough focusing on economic advancement, social development, and attention to the environment, all 17 sustainable development goals (SDGs) set out to address many of the structural factors that afect kidney health [28].Tis means that each SDG has the potential to improve the health of the kidney and prevent kidney disease by improving the individual's and society's general health and well-being and by protecting the environment [29,30].For reaching the SDGs, targeting DN will be an important consideration [31].
Compared to developed countries, DN is more common among people who live with diabetes in Africa due to delayed diagnosis, limited screening and diagnostic resources, poor glycemic control, and other risk factors, and inadequate treatment at an early stage [32].Terefore, timely diagnosis and prompt management of DN, and early identifcation of predictors are highly important to tackle this serious public health problem [33].As a prerequisite for appropriate policy development and fair priority setting within every country, the local burden of diabetic nephropathy and its predictors and the local capacity to early identify and manage such conditions must be determined [30].In developing countries, including Ethiopia, most epidemiological studies have been limited to the prevalence of diabetic nephropathy cross-sectionally, lacking the identifcation of predictor variables.A study on the incidence and predictors of diabetic nephropathy among T2DM may allow healthcare providers to reduce the incidence of diabetic nephropathy by targeting modifable predictors to improve the extent of the consequences, particularly in resource-constrained countries like Ethiopia.However, the incidence and predictors of diabetic nephropathy among patients with T2DM are less often studied, and to assess this issue, no follow-up study has been conducted in the study setting.Terefore, this study aimed to determine the incidence and predictors of diabetic nephropathy among type II diabetes mellitus patients at Hawassa University Comprehensive Specialized Hospital.

Sample Size Determination and Sampling Technique.
Te sample size was determined using Stata software power analysis for the Cox proportional hazards model by considering the following assumptions: probability of type I error (α) 0.05, power 80%, variability of covariates of interest 0.5, adjusted hazard ratio (AHR) 1.74 from the previous study [25], probability of event 0.1965, and proportion of withdrawals 0.1.Finally, the required sample size was 579.
Study participants who fulflled the inclusion criteria were selected by simple random sampling techniques using computer-generated random numbers from the sampling frame.

Data Collection Procedure or Tool.
Te data extraction checklist tool was adapted after reviewing diferent literature [18-20, 25, 26] that were used to extract relevant data from the patient's medical record.Te contents of the checklist were validated by expert seniors in the feld.Te checklist includes sociodemographic characteristics and clinical and treatment-related factors.Te data were collected by three BSc nurses under the supervision of one public health professional.Te data were collected by reviewing the T2DM patient's medical record.

Data Quality Assurance.
A pretest using 5% of the sample size was conducted on the same setting before the actual data collection, and adjustment was made to the data extraction checklist based on the result of the pretest.Oneday training was given on the objectives of the study and how to retrieve records as per the data extraction sheet to data collectors and supervisor before the actual data collection.Te extracted data were cross-checked with random samples for consistency and completeness.During data collection, the supervisor and the principal investigator checked the extracted data for its completeness and consistency.

Operational Defnitions.
Event: Development of diabetic nephropathy within the follow-up period.
Censored: Patients who did not develop DN until the end of the study, or died, or lost to follow-up or transfer out before developing DN within the study period.
Hypertension is defned as an average SBP ≥140 mmHg or DBP ≥90 mmHg or both that were confrmed on two separate visits or are on prescription medication for hypertension [35,36].

Data Processing and Analysis.
Te collected data were checked for completeness and consistency, and then entered, coded, edited, and cleaned using EpiData version 3.1 and exported to Stata version 14 for further analysis.Te incidence rate of DN was calculated by dividing the number of people who experienced DN during the follow-up period by the person-time at risk starting from the time of T2DM diagnosis until the end of each patient follow-up period.Descriptive statistical analysis, including frequencies with percentage, mean with standard deviation (SD), and median with interquartile range (IQR) was performed.A life table was used to estimate the cumulative probability of developing DN in T2DM at diferent time intervals.Te Kaplan-Meier survival curve, together with the log-rank test, was used to compare the survival diferences between diferent categories of independent variables.A Cox regression model was used to examine the relationship between the rate of DN occurrence and the infuence of the predictor variables.Bivariate and multivariable Cox proportional hazard regression analysis were performed between dependent and independent variables to identify signifcant predictors of DN.Variables in bivariate analysis with a p value <0.25 were selected as candidate for multivariable Cox proportional hazard regression analysis to control the efect of confounders.Te model was built using the backward stepwise likelihood ratio method.Variables with a P value <0.05 in the multivariable analysis were considered statistically signifcant.Te assumption of Cox proportional hazard was checked using the Schoenfeld residuals test (Prob > chi 2 0.4085).Te goodness of model ftness was checked by using the Cox-Snell residual.Multicolinearity between variables was checked using the variance infation factor (mean VIF 1.07).An adjusted hazard ratio with a 95% confdence interval was used to show the presence and strength of the association, and P values were used to identify statistically signifcant predictors.Finally, the fndings of the study were presented in text, tables, and fgures.

Ethics Approval
Statement.An ethical clearance letter was obtained from the institutional review board of Arba Minch University with reference number IRB/1242/3022.Ten, an ofcial support letter was obtained from the Department of Public Health, and similarly, permission was secured from Hawassa University Comprehensive Specialized Hospital.Confdentiality was maintained by assigning a code rather than personal identifers.Due to the retrospective nature of the study, the requirement of informed consent was waived for reviewing the patient's medical records.

Sociodemographic Characteristics.
A total of 532 (91.9%) type II diabetes mellitus patients were included in the analysis with the exclusion of incomplete charts from the total sample size of 579 during the follow-up period.Te mean age of the study participants was 49.4 years, with a standard deviation (SD) of ±12.5 years.More than half 292 (54.89%) of the study participants were male.Among the participants, three-fourths (74.81%) were urban residents (Table 1).

Clinical and Treatment Characteristics of Study
Participants.Te study participants fasting blood sugar and systolic blood pressure median with interquartile range (IQR) was 171 (83.5-258.5)and 125 (112-138), respectively.More than one-third (44.17%) of the participants were hypertensive.Out of the total study participants, more than three out of ten, 193 (36.28%), had a family history of diabetes mellitus.Tirty-three (6.2%) of the study participants had diabetic retinopathy, and seventy-nine (14.85%) had diabetic neuropathy.Te total cholesterol level of one-third, 185 (34.77%), of the study participants' was greater than or equal to 200 mg/dl, and nearly one-fourth, 144 (27.07%), of their HDL was less than 40 mg/dl.More than half (57.33%) of the participants used oral hypoglycemic agents (OHA) (Table 2).

Comparison of Survival Probability among Categories.
Te Kaplan-Meier survival curve, together with the log-rank test, was used to compare the survival diferences among various categories of predictor variables.Type II DM patients who had hypertension had a statistically signifcant lower survival probability compared to those who had no hypertension (p value 0.0001).Tose who had a positive baseline proteinuria had a statistically signifcant lower survival probability compared to their counterparts (p value 0.0008).Type II DM patients with diabetic neuropathy had a statistically signifcant lower survival probability compared to those who didn't have diabetic neuropathy (p value 0.0397).Type II DM patients with triglyceride levels greater than or equal to 150 mg/dl had a statistically signifcant lower survival probability compared to those whose triglyceride level was less than 150 mg/dl (p value 0.0013).Tose whose HDL level was greater than or equal to 40 mg/dl had a statistically signifcant higher survival probability compared to those whose HDL level was less than 40 mg/dl (p value 0.0289).In addition, T2DM patients whose LDL level was greater than or equal to 100 mg/dl had a statistically signifcant lower survival probability compared to those whose LDL level was less than 100 mg/dl (p value 0.0124) (Figure 3).

Predictors of Diabetic Nephropathy. In bivariate Cox
proportional hazard regression analysis variables with p value <0.25, age, sex, residence, FBS, SBP, DBP, hypertension, proteinuria, diabetic neuropathy, cardiovascular disease, duration of diabetes, total cholesterol, triglyceride, HDL, and LDL were signifcantly associated with DN and were included in multivariable Cox proportional hazard regression analysis.However, age, fasting blood sugar, systolic blood pressure, and duration of diabetes were found to be independent predictors of diabetic nephropathy among T2DM patients.
For each one-year increase in age, the hazard of developing DN increased by 2.7% at any time during the follow-up period (AHR � 1.027, 95% CI � (1.005, 1.049)).As the fasting blood sugar level increases by 1 mg/dl, the hazard of developing DN increases by 1% at any time during the follow-up period (AHR � 1.010, 95% CI � (1.007, 1.013)).As systolic blood pressure increases by 1 mmHg, the hazard of developing DN increases by 5% at any time during the follow-up period (AHR � 1.050, 95% CI � (1.031, 1.069)).Among T2DM patients with a duration of diabetes longer than or equal to fve years, the hazard of developing DN decreased by 80% when compared to their counterpart at any time during the follow-up period (AHR � 0.20, 95% CI � (0.09, 0.44)) (Table 4).

Discussion
Te main goal of this study was to determine the incidence and predictors of diabetic nephropathy among T2DM patients.According to this study, the incidence rate of diabetic nephropathy among T2DM patients was 2.71 cases per 1,000 person-month observation.Age, fasting blood sugar, systolic blood pressure, and duration of diabetes were identifed as independent predictors of DN.  observation [25], and Southwest Ethiopia, 2.29 per 1,000 person month of observation [22].Tis consistency might be due to the similarity in diagnostic methods, service provision, and follow-up period.However, the fnding was slightly higher than the fnding from the study conducted in Spain, which reported an incidence rate of 2.07 per 1,000 person-month of observation [18].Tis discrepancy might be due to the diferences in health care systems and the follow-up period of the study; in this study, 10-year data were used, whereas in a study conducted in Spain, 5-year data were used.Te fnding of this study was higher than the fndings of the studies conducted in the Netherlands, 1,213 per 100,000 person year of observation [37]; Northwest Ethiopia, 14 per 10,000 person month of observation [20]; and Amhara region, Ethiopia, 193 per 10,000 person year of observation [26].Tis diference could be related to the study setting and follow-up period [20,26,37].In addition, for the study conducted in the Netherlands, the diference might be due to the sample size and variation in diagnostic methods [37].Moreover, the black race had a higher incidence and severity of DN associated with a greater rate of GFR decline [27].Te incidence rate of DN was lower than in a study conducted in Iran, 43.84 per 1,000 person year of observation and 55.80 per 1,000 person year of observation [21].Tese diferences may be attributed to diferences in urbanization, poor adherence to treatment, lifestyle, and methods used for defning the outcome [21].Moreover, the possible explanation for the lower incidence rate of DN in our study may be due to the improvement of healthcare services nowadays compared to the previous.
In this study, age was a signifcant predictor of the development of diabetic nephropathy.Te fnding is in line with a study conducted in the United Kingdom, Spain, Ethiopia, and Iran [17][18][19]21].In older age, declining GFR is increasingly prevalent among cases with DN [27].Among patients with T2DM, low eGFR, was very high in patients older than 65 years [38].Due to the co-existence of prevalently increasing diabetes mellitus with age, renal function may be compromised in the elderly [39].Te reduction in renal function may be linked to cardiovascular hemodynamics due to endothelial cell dysfunction and changes in vasoactive mediators, resulting in increased atherosclerosis, glomerulosclerosis, and hypertension [40].For diabetes in elderly DN patients, blood glucose control would have a greater beneft [39].
According to the fndings of this study, type II diabetes patients with higher FBS levels were at a higher hazard of developing DN.Te study fnding is consistent with cohort studies conducted in Ethiopia and Iran, respectively [21,25].Journal of Nutrition and Metabolism activation [8,12].In patients with T2DM who achieved optimal glycemic control, the incidence of DN decreased [33].Tus, the risk of the onset of DN and its progression can be reduced through intensive glycemic control [8,12].
In the current study, high SBP showed a signifcant association with DN.Tis is consistent with previous studies conducted in the United Kingdom, Asia, and Spain, respectively [17,18,23].Tese might be due to the fact that higher blood pressure is a risk factor for both reduced GFR and microalbuminuria [27].Controlling blood pressure is fundamental to reducing the risk of progression of DN [8].For every 10 mmHg decrease in SBP, the development of DN is reduced by 17% [41].Terefore, to slow the progression of DN, the most important intervention in patients with type II diabetes is a sustained reduction of blood pressure [42,43].
Furthermore, the duration of diabetes mellitus is also the other predictor for the development of DN among T2DM patients.In this study, the duration of diabetes was negatively associated with the hazard of DN.Te fnding is supported by a study conducted in northwest Ethiopia [20].However, the fnding of the current study is inconsistent with those of studies done in Spain and Ethiopia, respectively [18,25].Patients with diabetes for a longer duration are more likely to develop DN than patients with a shorter duration of diabetes [44].However, this discrepancy might be due to the study participants being T2DM patients; they may have been asymptomatic for several years and come to the health facilities lately when the diagnosis was fnally made DN might have already been present [10,44].
In the interpretation of the fndings of this study, several limitations should be considered.First, due to the retrospective follow-up nature of the study, potential confounding factors like sociodemographic and behavioral characteristics (occupational status, educational status, smoking status, and physical activity) were not included due to the unavailability of the records.Second, in this study, variables like body mass index and hemoglobin A1c were incomplete.Tird, as this study was institution-based, the fndings might not be generalizable to the total population.Te strength of this study was that patients with type 2 diabetes were followed for a longer duration, the sample size of this study was sufciently powered, and the data were collected by trained health professionals.Tis result suggests that in older age type 2 diabetic patients, diabetic nephropathy development and progression increase as the previous similar studies [17][18][19]21].Tis fnding suggests that poor glycemic control accelerates the development of the incidence of DN over time, which is comparable with previous studies [21,25].Tese results build on existing evidence that uncontrolled blood pressure increases the progress towards the development of DN [17,18,23].Tese results suggest that a longer duration of T2DM reduces the risk of developing DN, similar to the previous study [20].However, this contradicts the existing evidence that a longer duration increases the hazard of DN development [18,25].Terefore, the implication to the researchers was to study the association between the duration of T2DM and the incidence of DN.
Tis fnding indicates that DN remains a public health problem that deserves full policy consideration from policymakers to address and improve type II diabetes patient outcomes.If the high incidence of DN continues, it will result in huge health-related, economic, and social consequences to people living with diabetes and the healthcare system.

Conclusion
Te study revealed that a substantial number of T2DM patients developed DN, with a high incidence rate in the study setting.Older age, high systolic blood pressure, and fasting blood sugar level were the identifed predictors of DN.Duration of diabetes longer than fve years was a negatively associated predictor with the development of DN.Tight systolic blood pressure and optimal blood sugar control could reduce the development and progression of DN among T2DM patients.Terefore, regular monitoring and counseling on blood pressure and blood sugar control should be strengthened for type II diabetes mellitus patient attendants at diabetic follow-up clinic, particularly for the older age.

Figure 2 :
Figure 2: Te Nelson-Aalen cumulative hazard estimate of type II diabetes mellitus patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2012-2022.

Table 1 :
[19]odemographic characteristics of type II diabetes patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2022 (n � 532).Te fnding of this study revealed that the incidence rate of diabetic nephropathy was 2.71 per 1,000 person-month observation.Te fnding was consistent with a study conducted in St. Paul's Hospital, Addis Ababa, Ethiopia, 21.78 per 10,000 person month of observation[19], Black Lion, Addis Ababa, Ethiopia 3.6 per 100 person year of

Table 2 :
Clinical and treatment characteristics of type II diabetes patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2022 (n � 532).Figure 1: Te overall Kaplan-Meier survival curve of type II diabetes mellitus patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2012-2022.

Table 3 :
Life table of T2DM patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2022.

Table 4 :
A bivariable and multivariable Cox proportional hazard regression analysis for predictors of time to diabetic nephropathy among T2DM patients at Hawassa University Comprehensive Specialized Hospital, Southern Ethiopia, 2022 (n � 532).