Predictors of chronic kidney disease among Indonesian adult population: Results from the 2018 Indonesia Basic Health Research [version 1; peer review: awaiting peer review]

Background: Three out of ten cases of chronic kidney disease (CKD) are cause by type 2 diabetes mellitus (T2DM). In addition, nutritional status, consumption of unhealthy foods, high blood pressure, and smoking habits were also previously identified as risk factors to CKD but there is a scarcity of data from Indonesia. Therefore, we sought to determine the risk factors using a predictive model for the incidence of CKD in Indonesia based on the Indonesia Basic Health Research 2018. Methods: Secondary data from the 2018 Indonesia Basic Health Survey with 300,000 respondents was used. Respondents with complete socio-demographic, food consumption data, anthropometric data and chronic disease status (i.e. T2DM and hypertension) type 2 were analyzed (n=96,098). Binary logistic regression model was performed to assess the association between socio-demographic, food group consumption, health and nutritional status with CKD. Results: Based on the final model of logistic regression analysis, ranging from the most and least


Introduction
Deaths due to non-communicable diseases have increased in the last decade and are predicted to increase to more than 380 million cases by 2025, especially in Southeast Asia, the Mediterranean and the Middle East and Africa. Chronic Kidney Diseases (CKD) is one of emerging public health problem that manifests by loss of kidney function, heightened cardiovascular diseases (CVD), and premature death. 1,2 KDIGO in 2020 define CKD as an abnormalities of kidney function or anatomy structure with specific symptoms presents for more than three months and has implication for further health condition. While the burden of CKD is reasonably well defined in developed countries, 1 increasing evidence indicates that the CKD burden may be even greater in developing countries. 3 The 2014-2018 Indonesia National Health Insurance Statistics results show a substantial increase in the number of patients, the number of visits, the average number of visits per person/year and the cost of claims for dialysis procedures, 4 inherent with the increasing financial burden for CKD treatment, with an average claim value per patient in a year as much as Rp. 48,714,515 or equivalent to USD 3,365. 5 In Indonesia, the prevalence of CKD in 2013 was reported as 2%. 6 In 2018, it was reported from the number of CKD patients, 19% received hemodialysis. 7 Three out of 10 cases of CKD are caused by type 2 diabetes mellitus. 8 In addition to diabetes, other conditions that are risk factors for CKD are nutritional status, consumption of unhealthy diets, high blood pressure, and smoking habits. In general, risk factors for CKD are divided into four main groups, namely susceptibility factors which are defined as a person's susceptibility factor to develop CKD, second is initiation factors including risk factors that can directly initiate kidney damage, then the progression factors include several factors that cause progressive kidney damage and accelerate the decline in kidney function; the last is end-stage factors which are factors that can increase morbidity and mortality due to CKD. 9 A prospective study on 23,543 respondents in Maryland, USA indicated that most of the risk of CKD in this population was associated with stage 1 hypertension (23%) and smoking (31%). 10 The prevalence of CKD in diabetic patients is also increasing worldwide. CKD affects 50% of individuals with type 2 diabetes mellitus (T2DM). 11 There is a strong correlation between the pathophysiology of kidney and heart disease in type 2 diabetes, this aspect is expressed by cardiovascular risk factors: T2DM, obesity, smoking, dyslipidaemia, hypertension, genetic factors, etc. 12 Obesity is also a strong risk factor for kidney disease. Obesity increases the risk of major risk factors for CKD such as hypertension and diabetes. In obesity, the kidneys have to work harder to filter more blood to meet metabolic needs due to weight gain. This increase in function can damage the kidneys and increase the risk of developing CKD in the long term. 13 Research on obese students at the Faculty of Medicine, Sam Ratulangi University, Indonesia revealed a strong relationship between body mass index (BMI) and a decrease in glomerular filtration rate (p < 0.001). 14 Moreover, the prevalence of obesity in Indonesia has increased significantly from 2013 to 2018 from 26.3% to 35.4%. This can lead to an increase in the risk of CKD in Indonesian society and increase the financing burden.
Different conditions in urban and rural areas can also affect differences in risk factors for CKD. 15 This study aimed to determine the risk factors for CKD and derived a predictive equation for the incidence of CKD in Indonesia based on the results of basic health research 2018 that represents each region in Indonesia. This analysis is expected to provide significant results and have an impact on appropriate CKD prevention efforts in Indonesia. CKD prevention efforts as early as possible can have a positive impact, not only maintaining productivity, maintaining quality of life, reducing morbidity and mortality but also reducing the burden of disease due to CKD.

Data source and respondents
This was a cross-sectional study which used secondary data from the 2018 cohort Indonesia Basic Health Survey (RISKESDAS) upon request and approval from the National Institute of Health Research and Development, Ministry of Health, Republic of Indonesia with a total number of respondents of 300,000. RISKEDAS is a national representative household health survey conducted every three years by the Indonesia National Institute of Health Research and Development (NIHRD), Ministry of Health Indonesia. Data for Riskesdas were collected by qualified enumerator that collected from representative areas of Indonesia, both rural and urban. Respondents who provided sociodemographic, food consumption, type 2 diabetes status (self-reported), hypertension status (self-reported), and body mass index data were further analyzed. We excluded respondents who were aged <18 years to minimize the sampling bias. Further, we included only respondents who had no missing data. After the inclusion criteria were applied, 96,098 respondents were included in this study.

Measurement of CKD
Diagnosis of CKD was noted by the self-reported declaration on whether any physicians ever diagnosed and informed them that they had CKD with symptoms persists for a minimum of three months based on the result of glomerulus filtration rate (GFR). 7

Measurement of covariates
Sociodemographic characteristics including marital status, level of education, geographical areas of living, sex, and working status were reported using categorical data with the exception for age, accessed using ratio data. The respondents' marital status was categorized into never been married, married, and divorced. The respondents' level of education was categorized into low (<6 years of school attainment), middle (6-12 years of school attainment) and high (>12 years of school attainment). Living area was categorized into urban and rural; while working status was classified as working and not working.
We also obtained information on T2DM and hypertension as the risk factors of CKD. Respondents answered the selfreported questionnaire on whether they have been diagnosed with T2DM and hypertension. 7 The body mass index (in kg/m 2 ) was classified into four groups (underweight <18.5, normal 18.5-25.0, overweight 25.1-27.0, and obese >27.0). 16 In addition, food group consumption was measured using a Food Frequency Questionnaire (FFQ). Respondents were asked the frequency of consumption of non-healthy foods in a day, week or month including consumption of sugary foods and beverages, salty foods, fatty/fried foods, grilled foods, processed meats, seasonings, soft-drinks or carbonated drinks, energy drinks, and instant noodles.

Statistical analysis
The respondents' sociodemographic data were presented as means (standard deviation) for the continuous data and numbers (percentages) for the categorical data. For the bivariate analysis, chi-square test was used to analyse the relation between independent and dependent variables in categorical data; while t-test analysis was used for continuous data. Moreover, we used a binary logistic regression model to assess the association between socio-demographic, food group consumption, health and nutritional status with CKD. All the analyses were conducted using SPSS statistical software (V 26; SPSS Inc., IBM Corporation). Values were significantly different from adult with and without CKD: * p < 0.05, ** p < 0.01, *** p < 0.001. P was obtained with χ 2 tests. + Analyze using T-test.

Results
Among 96,098 adults were involved in this study, 0.5% or 501 adults had CKD (Table 1). There was a significant (p < 0.001; 2.000-2.003% CI) difference in age between adults with CKD and non-CKD. The older adults group had a 2.002 times higher risk of getting CKD than the younger adults group. So that age factor was one of the greatest risk among other factors.
A total of 54,844 respondents lived in rural and had significantly less risk (p < 0.05; 0.665-0.956% CI) of 0.798 times lower than respondents who lived in urban. In other words, adults who lived in an urban area had a 1.25 times higher risk to get CKD than adults who lived in rural areas. Meanwhile, sex was not identified as a risk factor of CKD (p > 0.005), and yet, the majority of respondents in this study were male (94.6%). The majority of respondents were married. Interestingly, marital status was identified as one of the risk factors of CKD (p < 0.001; 0.498-0765% CI), in which unmarried adults had 1.62 times greater risk of developing CKD than their married counterparts. Most of the respondents in the study had a low education status in which they never attended a formal school or only completed elementary school. Lower educational background posts a risk towards CKD (p < 0.05; OR 1.229; 1.061-1.425% CI). There was significant difference between adults with CKD and non-CKD with regards to their employment status (p < 0.01; OR:1.438; 1.099-1.881% CI) in which a higher risk was observed for working adults (1.438 times) than the non-working adults.
In total, five dietary patterns from total of 10 food groups were identified to be significantly different between adults with and without CKD (p < 0.05) ( The present of comorbidities particularly T2DM and hypertension were associated with CKD (p < 0.000) (  Values were significantly different from adult with and without CKD: * p < 0.05, ** p < 0.01, *** p < 0.001. P was obtained with χ 2 tests. 0: not present); education (1: not school/not graduate elementary/elementary; 0: middle to high education); living area (1: rural; 0: urban); and sugary drinks consumption (1: 1Â/day or more; 0: weekly/monthly/never).

Discussion
Older age was reported to be associated with increased risk of reduced renal function and they have high prevalence of more severe stages of CKD in elderly individuals. 17,18 Increasing age affects the anatomy, physiology and cytology of the kidneys. Decreasing the estimated glomerular filtration rate (eGFR) is a "normal aging" process. The kidneys cannot regenerate new nephrons, so that if the kidneys are damaged or in the aging process, there will be a decrease in the number of nephrons. After the age of 40, the number of functioning nephrons decreases by about 10% every 10 years and after the age of 80 the kidneys have only 40% of the functioning. [19][20][21] The cohort of the Framingham Offspring Study in 2,585 subjects without CKD followed for 12 years, showed an age-related decrease in GFR (OR = 2.36 per 10-year increase in age; 95% CI 2.00-2.78) 22 concurred with the findings from our study that age is the biggest predictor of CKD.
The lifestyle of urban adults has experienced adaptation and sedentary. This is caused by the type and work environment, population density in settlements and the influence of the media, especially television. 23 In India, urbanisation is associated with risk factor of non-communicable diseases. 24 Compared with their rural counterparts, men working in factories have approximately 1.5-fold increase in odds of a sedentary lifestyle, three-fold increase in odds of obesity, and two-fold increase in odds of T2DM. 25 The results of this study is in line with Singh et al. 26 that showed that patients with CKD were older, more likely to be male, more likely to have a high school diploma, more likely to reside in urban areas, less likely to have a low income, more likely to be overweight or obese, to present with comorbidities including diabetes, hypertension and cardiovascular disease than patients without CKD.
One of the important factors in determining health status is educational background. The higher the level of education, the higher the awareness of the importance of health. 27 Different education level will affect level of knowledge and it  causes differences in responses to a health problem. In addition, there will be different levels of understanding of information that conveyed about the disease suffered. People with low educational background tend to have awareness to do early detection of CKD. This can increase the awareness of patients with CKD because in the early stages the signs and symptoms are not felt specifically. Most of the patients came with severe complaints and at the time of further examination were already at the end stage (stage 5). 28 Being single or divorced/widowed increased the OR 3 fold compared to those who are married. Additionally in sex stratified analysis among women, being single or divorced/widowed relative to being married were positive predictors for incident of CKD stages 3 to 5. 29 This is in line with the study by Novak et al. 30 that showed that people with depression had a significantly higher risk of developing CKD and those who depressed were younger, more likely to be divorced and had higher eGFR. Thus, it is postulated that someone who is single/divorced has potential to experience stress, inherit to increase risk of other chronic diseases such as hypertension, T2DM that can lead to chronic kidney failure. 31 Most people with CKD have a lower quality of life than non-CKD people. 32 Most patients cannot return to their previous activities or work. Decreased body function and limitations in carrying out activities are the cause, so they prefer to rest and not work. 33 Dietary pattern is reflected the habitual dietary intake which can be classified into healthy and unhealthy pattern. Metaanalysis by Bach et al. show intake of low red and processed meats, sodium, also sugar sweetened foods associated with lower odds ratio of CKD incident. 34 Meanwhile high consumption of refined grains, high fat dairy foods, meat, beans highly associated with albumin to creatinine ratio which give higher impact in kidney injury. 35 Consumption of several unhealthy foods such as added fats, sugars, salt, and refined grains are associated with adverse metabolic outcome, including CKD. 36,37 Excess calorie intake is an underlying factor of the development of obesity which may give consequences for long term kidney health. The higher calorie intake, the higher body mass index, also the higher prevalence of kidney injury and increase risk of CKD incident. [38][39][40][41][42][43][44] Consumption of drinks containing high amounts of sugar especially fructose (sugar cane based sugar, the most used kind in Indonesia) is associated with the increasement risk of kidney disease. 45 High consumption of refined sugar may increase triglycerides level and increase body mass index. The elevation of blood glucose also give consequences in kidney failure by metabolic impairment like induce insulin resistance and produce more uric acid. 46 Besides, polyol pathway from glucose to fructose conversion can cause kidney damage. 47 This study has the same result as a study in Iran which showed high fat and sugar food and drink choices significantly associated with the increased of incident risk of CKD by 46%. 48 Consumption of artificial sweetened soda showed association with incidence of albuminuria and rapid GFR decline. 49,50 In addition, dietary sugar increases the risk of diabetes mellitus which being the high risk factor of CKD. 45 One of the findings in this study is that consumption of instant noodles, which are high in sodium, is significantly higher in adults with CKD. More than 75% of salt intake are coming from processed foods, one of which is instant noodle. 51 Excess salt intake is associated with adverse health outcome like hypertension, volume overload, vascular damage and ventricular hypertrophy, which exacerbate CKD. The higher sodium intake, the higher the blood pressure, external fluid volume, and albuminuria which may be implicated in CKD incidence. [52][53][54][55][56] Excess sodium intake also increases kidney plasma flow and glomerular filtration pressure which contributes to proteinuria and faster progression to kidney failure. 57,58 The association between T2DM and CKD as well as hypertension and CKD has been long known both as initiation factors or risk factors that can directly initiate kidney damage. KDOQI statement also mention that diabetes as marked by hyperglycemia which causes vascular complication and diabetic kidney disease (DKD). 59 A prospective study, NHANES, from 2009-2014 in the US revealed that CKD prevalence was progressively higher with longer duration of diabetes. Further, they explained that increase risk of CKD among people with diabetes could be mediated by hyperglycemia-induced glomerulopathy, vascular damage, treatment specific to diabetes, or other confounding such as smoking, obesity or genetic predisposition. 60 Similarly, Szczech, et al. (2014) also reported that presence of diabetes ultimately increases CKD risks by modifying urine protein excretion and decreased eGFR. 61 Moreover, mitochondrial dysfunction among TD2M individual plays an important role in the existence of diabetic complication through increasement of reactive oxygen species (ROS) and superoxide generation. 62,63 Kim et al. (2014) proposed that the risk of kidney damage in T2DM could be induced by the higher level of serum uric acid. 64 High serum uric acid increased renin-angiotensin-aldosterone (RAA) system and inflammation, also dysfunction the endothelial. Those pathways were similar to the pathway of hypertension to CKD. Hypertension and CKD have an inverse relationship, untreated hypertension increases the risk of CKD and CKD can also worsen the condition of hypertension to cause cardiovascular complications. 65,66 In all types of glomerulonephritis, hypertension is a substantial independent risk factor for progression to End-Stage Renal Diseases (ESRD). Therefore, therapeutic intervention is highly needed in hypertensive patient to prevent the kidney damage. 67 Our results did not find a significant correlation between weight status and CKD, but were significant with the presence of T2DM and hypertension. We suggest these results could imply that CKD might present to those with previously diagnosed T2DM or hypertension without the individual being overweight/obese. However, it is well-known that obesity is a major risk factor of metabolic disorders such as TD2M and hypertension that lead to increased risk of kidney failure. 68 Impaired pressure natriuresis and increase of tubular sodium reabsorption affects volume expansion through sympathetic nervous system then activate RAA-system; especially with the presence of visceral obesity, in the end, raises blood pressure. In addition, increased oxidative stress, inflammation, and lipotoxicity also increases the risk of kidney damage among obese hypertensive. 68,69 To the best of our knowledge, there were no previous studies in Indonesia which had reported predictors of CKD among Indonesian adults, which makes this report novel. Despite a cross-sectional nature, our study analyzed a huge number of samples which we believe can be generalized to the Indonesian adult population. However, we also note some limitations such as self-declaration of prior diagnosis of CKD, T2DM and hypertension that may undermine undiagnosed cases and memory bias. Despite the presence of aforementioned bias, there was no other similar study in Indonesia therefore the novelty of this study is accountable. In future research, database should be compared from year to year included more risk factors that might correlated with CKD to give better perspective on CKD trend in Indonesia.

Conclusion
This result can be used by the health professional to predict the risk of CKD based on the identified risk factors. Early prediction of CKD risk could be beneficial to prevent the development of CKD by changing dietary and lifestyle behavior; thus, prolonging the need for dialysis treatment. Due to the data limitations of the study, for further studies we suggest researchers can identify specific factors influencing the occurrence of kidney diseases in the Indonesian population.

Data availability
Data are available upon request to National Institute of Health Research and Development, Ministry of Health, Republic of Indonesia by sending request letter, data usage statement, and research proposal to Kepala Badan Litbangkes Jl. Percetakan Negara No. 29 Jakarta Pusat, Indonesia. Raw data will be sent after reviewed and approved by the Head of National Institute of Health Research and Development, Ministry of Health, Republic of Indonesia. To request the raw data, visit https://www.litbang.kemkes.go.id/layanan-permintaan-data-riset/.