Is obesity in the Asia-Pacific body mass index classification associated with the development of chronic kidney disease in non-diabetic Japanese men and women?

Yukari Okawa, Toshiharu Mitsuhashi, Toshihide Tsuda
doi: https://doi.org/10.1101/2024.01.06.23300566

Abstract

Objective Obesity is a risk factor for the development of chronic kidney disease (CKD). However, little is known about the relationship between the Asia-Pacific body mass index classification (Asia-Pacific classification) and the development of CKD in non-diabetic Asian men and women.

Methods Body mass index (BMI) was categorized in accordance with the Asia-Pacific classification: underweight (<18.5 kg/m2), normal weight (18.5–22.9 kg/m2), overweight (23.0–24.9 kg/m2), obesity class I (25.0–29.9 kg/m2), and obesity class II (≥30.0 kg/m2). Using administrative checkup data for non-diabetic Japanese adults (1998–2023), we assessed the relationship between the Asia-Pacific classification and the development of CKD by sex. The Weibull accelerated failure time model was selected because the proportional hazards assumption was violated.

Results In total, 34.2% of 3098 men and 34.8% of 4391 women had developed CKD. The mean follow-up time was 7.41 years for men and 8.25 years for women. Compared with normal weight, obesity class I was associated with a 6% (95% confidence interval [CI]: 2%–10%) shorter time to CKD development in men and a 6% (95% CI: 2%–9%) shorter time to CKD development in women. Obesity class II was associated with the shortest survival to the development of CKD by point estimate in men but had an almost null value in women (i.e., 95% CIs crossed 1 in all of the models).

Conclusions Among non-diabetic Japanese men and women, obesity class I in the Asia-Pacific classification was associated with a shorter survival time to CKD onset than normal weight. Low point estimates for obesity class II suggested a dose-response relationship in men but a less clear relationship in women. The conventional cutoff value for obesity, BMI ≥30.0 kg/m2, may be too high a risk factor for the development of CKD in non-diabetic Asians, regardless of sex.

Introduction

Worldwide, 697.5 million people have chronic kidney disease (CKD)1. According to the Kidney Disease: Improving Global Outcome (KDIGO), CKD is defined as abnormalities of kidney structure or function that persist for ≥3 months2. Generally, CKD is asymptomatic until it deteriorates irreversibly. As CKD progresses to end-stage renal failure, patients must overcome physical, psychological, and financial hardship in a context of limited treatment options, which include dialysis and kidney transplantation3. Moreover, CKD increases the risk of a variety of outcomes including death, cardiovascular events, and hospitalization2,4. Therefore, preventing the development of CKD is important.

Reported risk factors for the development of CKD include male sex, obesity, diabetes, high sodium intake, and low physical activity5–10. Diabetes in particular is a major cause of CKD because the hyperglycemia caused by diabetes damages renal vessels over time9,11. Therefore, many cohort studies that have examined the relationship between various exposures and the development of CKD have treated glycemic control status (e.g., fasting blood sugar levels, prevalent diabetes) as a confounding factor that is adjusted to reduce bias12–14. However, even after adjusting for glycemic control status, the effect of diabetes-induced bias may remain. Therefore, studies evaluating risk factors for the development of CKD in non-diabetic populations are important.

Obesity is a risk factor for developing CKD5,6 A previous meta-analysis in which the results of 13 cohort studies were combined reported that the risk of an individual with obesity developing CKD was 1.28 times greater than that of a person of normal weight; however, no clear association was observed in individuals with overweight7. These studies used multiple definitions of obesity and CKD, included participants with and without diabetes from multiple ethnic groups, and did not show sex-specific results15. Therefore, studies that include sex- stratified, mono-ethnic, and non-diabetic participants are needed to more accurately evaluate the relationship between obesity and the development of CKD7.

A previous cohort study of non-diabetic Japanese men and women examined whether a dose-response relationship exists between body mass index (BMI) and the development of CKD16. The study categorized BMI into seven groups and reported that individuals in the BMI categories ≥23.0 kg/m2 for men and ≥27 kg/m2 for women were at an elevated risk of developing CKD compared with the reference group16. This suggests that even in a non-diabetic Asian population, elevated BMI is associated with the development of CKD and that sex differences are observed. However, these results cannot simply be compared with previous findings because the studies use different BMI cutoff values and reference groups7,16. Therefore, sex-stratified studies in single-ethnic and non-diabetic populations that include a global standard BMI classification that accounts for racial differences are needed to examine the relationship between BMI and the development of CKD.

The World Health Organization Western Pacific Region proposed the Asia-Pacific BMI classification (Asia-Pacific classification)17. The classification has a lower BMI cutoff value than the conventional World Health Organization BMI classification (conventional BMI classification) that is used for all races because Asians have higher morbidity and mortality even with a lower BMI17. Therefore, the Asia-Pacific classification should be used when assessing the relationship between BMI and CKD onset in Asian populations. On the basis of the above, the study’s purpose was to investigate the relationship between BMI that was defined using the Asia- Pacific classification and the subsequent development of CKD in non-diabetic Japanese men and women.

Materials and methods

Data source

The study was an ongoing open population-based longitudinal study of citizens of Zentsuji City, Kagawa Prefecture, Japan18. We used anonymized annual health checkup data from Zentsuji City from 6 April 1998 to 19 April 2023. On 1 April 2023, the city had a population of 30431 (men: 49.7%), with 38.1% of citizens aged 60 years and older18. The city database used in our previous study was used for data extraction19. The data extraction date was 6 July 2023.

In checkups, the participants underwent anthropometric measurements, a blood pressure test, a blood test, and a urine dipstick test, and answered a questionnaire about their lifestyle. The participants were not obliged to participate in a checkup, take all of the tests, or answer the questionnaire. Checkups were conducted in accordance with the protocol of the Ministry of Health, Labour and Welfare20.

Study participants

All citizens aged ≥40 years were eligible to receive a health checkup once every fiscal year (FY). In FYs 1998 and 1999 only, younger citizens aged 34–39 years were added to the eligible population on a trial basis to promote the health of younger people. In past years, approximately 30%–40% of the eligible population had received a checkup.

This study aimed to evaluate the relationship between BMI and the development of CKD in non-diabetic Japanese men and women. Therefore, the following exclusion criteria were applied: non-Japanese, those with missing exposure and/or outcome data, prevalent CKD and/or diabetes at study entry, missing hemoglobin A1c (HbA1c) values, prevalent diabetes at study entry, and only one observation. The participants who developed diabetes during the follow-up were regarded as right-censored. Diabetes was defined as HbA1c ≥6.5%21.

Measures Exposure

The exposure variable was obesity. BMI was used as the obesity definition. BMI was calculated by dividing weight (kg) by height (m) squared. Obesity was defined in accordance with the Asia- Pacific classification: underweight (<18.5 kg/m2), normal weight (18.5–22.9 kg/m2), overweight (23.0–24.9 kg/m2), obesity class I (25.0–29.9 kg/m2), and obesity class II (≥30.0 kg/m2)22.

Outcome

The outcome variable in this study was the development of CKD. The estimated glomerular filtration rate (eGFR) was used as a measure of renal function2. We used the three-variable revised Japanese equation to estimate glomerular filtration rate (GFR): eGFR(mL/min/1.73m2) = 194 x serum creatinine(mg/dL)-l.O94 x age (years)-O.287(x 0.739, iffemale)2,23–25. Prevalent CKD was defined as eGFR <60 mL/min/1.73 m22. Serum creatinine (SCr) values in mg/dL were used up to two decimal places. The enzymatic method was used to measure SCr.

Other covariates

Age, self-reported alcohol intake, self-reported smoking status, hypertension, dyslipidemia, HbA1c, and residential district were treated as covariates for adjustment7,16. Age groups were categorized as 34–59, 60–69, and 70–100 years. Self-reported alcohol intake was queried with the question, “How often do you drink?” Responses were dichotomized into “non- or seldom- drinker” and “drinker” categories. Self-reported smoking status was assessed by asking the question, “Are you currently a habitual smoker? A habitual smoker refers to those who have smoked a total of 100 or more cigarettes or have smoked for at least 6 months and have been smoking for the past month.” The answers were grouped into “non- or ex-smoker” and “smoker” categories.

Hypertension was defined as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg26. The participants were considered dyslipidemic if they met at least one of the following conditions: serum low-density lipoprotein cholesterol ≥140 mg/dL, serum high- density lipoprotein cholesterol <40 mg/dL, and/or serum triglycerides ≥150 mg/dL27. HbA1c values (%) were treated as a continuous variable. HbA1c was reported in Japan Diabetes Society units (%) from 1998 to 2012 and in National Glycohemoglobin Standardization Program units (%) from 2013 to 202328. HbA1c levels were standardized to National Glycohemoglobin Standardization Program values using the officially certified equation: HbA1cNGSP(%) = 1.02 x HbA1cJDS(%) + 0.2529.

Neighborhood indicators (e.g., food quality and availability) have been reported to be related to obesity, diabetes, and possibly CKD30. Therefore, to reduce the potential impact of where participants resided within the city, the district of residence was treated as a covariate. Zentsuji City has eight districts: East, West, Central, South, Fudeoka, Tatsukawa, Yogita, and Yoshiwara.

Statistical analysis

All analyses were conducted separately for men and women. The participants’ characteristics were summarized by BMI class. The person-years at risk were calculated from the date of the first participation to the date of onset of CKD or diabetes or the end of the last observation during the follow-up period. The participant demographics during the follow-up period were expressed as numbers by CKD onset, total person-years at risk, and incidence rate per 1000 person-years.

We created Kaplan–Meier curves by Asia-Pacific classification for men and women. We assessed the proportional hazards assumption using log–log plots and the Schoenfeld residuals31. We observed that the assumption was violated in this study. Therefore, the Weibull accelerated failure time model was selected in accordance with the Akaike and Bayesian information criteria31–33. The time ratio and its 95% confidence interval (CI) were a measure of the effect of exposure on an outcome of interest. For example, a time ratio of 0.9 signified a 0.9-fold longer survival to CKD onset than the reference group.

Four adjusted models were constructed in addition to the crude model. Model 1 adjusted for age category; Model 2 further adjusted for self-reported alcohol intake and smoking status; Model 3 further adjusted for hypertension, dyslipidemia, and HbA1c values; and Model 4 further adjusted for the residential district. To stabilize the models, a multiplicative term was added to Models 1–4 if an interaction effect between the exposure and the following variables was observed: age category, self-reported alcohol intake, self-reported smoking status, hypertension, dyslipidemia, and HbA1c values.

Given the nature of the voluntary checkup, missing values were observed for self- reported alcohol intake (missing in 31.0% of men and 28.3% of women), self-reported smoking status (missing in 24.9% of men and 23.1% of women), hypertension (not missing in men and missing in 0.02% of women), and dyslipidemia (missing in 17.2% of men and 24.4% of women). Missing measurements were complemented using multiple imputation methods with chained equations with 40 imputations34,35. Binary variables and categorical variables were imputed using logistic regression and multinominal logistic regression, respectively. Only imputed results were presented in this study because missing measurements were considered missing at random.

We restricted the presentation to the results of the main exposure variable only to avoid the Table 2 fallacy36. A two-tailed p-value of less than 0.05 was treated as statistically significant. All statistical analyses were performed using Stata/SE 18.0 (StataCorp, College Station, TX, USA). A participant flow and a district map of the city were drawn in Python 3.11.537–40. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines41.

Sensitivity analyses

Three sensitivity analyses were conducted in this study. In the first sensitivity analysis, conventional BMI classification (underweight [<18.5 kg/m2], normal weight [18.5–24.9 kg/m2], overweight [25.0–29.9 kg/m2], obesity class I [30.0–34.9 kg/m2], obesity class II [35.0–39.9 kg/m2], and obesity class III [≥40 kg/m2]) were treated as exposure variables42. Obesity classes I–III were combined as one obesity category because of the small number of participants in this study with BMI ≥30.0 kg/m2. In the second sensitivity analysis, to minimize the possibility of reverse causation, we excluded the participants who developed CKD at the second observation. The third sensitivity analysis attempted to eliminate temporary abnormal eGFR values by including the stringent definition of CKD, in which CKD is defined by two or more consecutive observations of eGFR < 60 mL/min/1.73 m2. Because the study included data from annual checkups, the duration of renal dysfunction in the stringent CKD definition was longer than that in the KDIGO definition of 3 months or longer2.

Results

Main analysis

The initial participants were 6309 men and 9192 women, with a mean age at study entry of 70.0 years for men and 68.3 years for women. After exclusion, 3098 men (mean age at study entry: 68.4 years) and 4391 women (mean age at study entry: 65.9 years) who received checkups from 7 May 1998 to 19 April 2023 remained in the final cohort (Figure 1).

Figure 1.
Figure 1.

Participant flow chart of the study cohort.

Figure 2 shows the map of the Zentsuji City district. The Kaplan–Meier curves stratified by BMI category in men and women are provided in Figure 3. The participants’ characteristics by Asia-Pacific BMI classification during follow-up are presented in Table 1 for men and Table 2 for women. The mean follow-up time was slightly longer in women (8.25 years) than in men (7.41 years). For both men and women, obesity class II (≥30.0 kg/m2) had the shortest follow-up time, accounting for less than 3% of the total follow-up time. Men with higher BMI were more likely to have hypertension and/or dyslipidemia, and women with higher BMI were more likely to have hypertension. After 25 years of follow-up, 34.2% of men and 34.8% of women developed CKD.

Figure 2.
Figure 2.

Map of Zentsuji City districts.

Figure 3.
Figure 3.

Kaplan–Meier survival estimates by Asia-Pacific body mass index classification for men and women.

View this table:
Table 1. Descriptive statistics of all observations stratified by the Asia-Pacific BMI classification among 3098 non-diabetic Japanese male citizens of Zentsuji City (1998–2023) Asia-Pacific BMI classification
View this table:
Table 2. Descriptive statistics of all observations stratified by the Asia-Pacific BMI classification among 4391 non-diabetic Japanese female citizens of Zentsuji City (1998–2023) Asia-Pacific BMI classification

Our main analysis examined the relationship between the Asia-Pacific classification and later-onset CKD by sex (Table 3). Similar results were observed in men and women, with those in obesity class I having a 6% shorter survival to the development of CKD than those of normal weight. Men in obesity class II had the shortest survival to CKD onset with a point estimate of 9%; however, the 95% CIs were across 1 in all of the models. For women in obesity class II, the results were almost null. Overweight was associated with higher point estimates (i.e., close to 1) compared with obesity categories, especially null values in women.

View this table:
Table 3. New onset of chronic kidney disease by the Asia-Pacific BMI classification among non-diabetic Japanese citizens of Zentsuji City (1998–2023)

In this study, the Weibull accelerated failure time model was used because the PH assumption was violated. For reference, the results of the Weibull PH model as the full model were as follows: adjusted hazard ratios (aHR; 95% CI) in men with underweight, overweight, obesity class I, and obesity class II compared with men of normal weight were 0.72 (0.51–1.00), 1.10 (0.94–1.28), 1.30 (1.11–1.52), and 1.48 (0.93–2.34), respectively. The aHRs (95% CI) for the same BMI categories in women compared with women of normal weight were 0.97 (0.79– 1.19), 0.99 (0.86–1.14), 1.26 (1.11–1.44), and 1.02 (0.76–1.38), respectively.

Sensitivity analyses

In the first sensitivity analysis in which the conventional BMI classification was used, more participants (68.9% of men and 71.0% of women) were classified as having normal weight than in the main analysis in which the Asia-Pacific classification (43.3% of men and 50.6% of women) was used (Tables 1–2, S1). The results were similar for the obesity group using the conventional BMI classification and the obesity class II group using the Asia-Pacific classification (Table 3). Overweight was associated with a shorter survival than normal weight in each sex, similar to the results for obesity class I in the Asia-Pacific classification (Table 3).

The second sensitivity analysis attempted to minimize reverse causation by excluding the participants who had developed CKD at the second observation and included 2723 men and 3822 women. The mean age at study entry (61.4 years for men and 58.3 years for women) was lower than in the main analysis and in the group of excluded participants who developed CKD at the second observation (67.3 years for men and 64.4 years for women). The mean follow-up period of the second sensitivity analysis was 8.14 years for men and 9.14 years for women—slightly longer than those in the main analysis (Tables 1–2). The distribution of follow-up time by BMI category was similar to that of the main analysis. After the follow-up period, 25.1% of men and 25.2% of women developed CKD—lower than in the main analysis. The relationship between the Asia-Pacific classification and the subsequent development of CKD after exclusion is shown in Table S2. Most results were similar to those of the main analysis; however, point estimates of survival for obesity class II were longer in men and shorter in women than those in the main analysis (Table 3).

Table S3 shows the results of the third sensitivity analysis, which was based on the stringent definition of CKD and in which two consecutive observations of eGFR <60 mL/min/1.73m2 were regarded as CKD. Using the stringent definition of CKD, more participants (3380 men and 4963 women) remained in the final cohort, the mean follow-up time was longer (8.13 years for men and 9.29 years for women), and fewer participants developed CKD (20.7% of men and 21.1% of women) compared with the results of the main analysis (Tables 1–2). The mean age at study entry (62.4 years for men and 58.9 years for women) was lower than in the main analysis and in the group of participants who were excluded because of prevalent CKD at study entry (70.8 years for men and 69.1 years for women). Regardless of sex, point estimates for overweight, obesity class I, and obesity class II were similar to those in the main analysis, including the category with the shortest follow-up period, BMI ≥30.0 kg/m2 (Tables 3, S3).

Discussion

Main analysis

In the current analysis based on annual health checkups from 1998 to 2023 in non-diabetic Japanese adults (3098 men and 4391 women), obesity class I as defined by the Asia-Pacific classification was associated with the earlier development of CKD. The 95% CIs for overweight and obesity class II were across 1 in all models; however, the point estimates implied a dose- response relationship, especially in men (Table 3). In all of the analyses, the results for the crude and adjusted models were similar, suggesting a lack of solid evidence that covariates affect the association between exposure and outcome.

These results partly agreed with those of the Ibaraki Prefecture Health Study (IPHS), a previous cohort study in diabetes-free Japanese participants16,43. The IPHS showed the dose- response relationship between BMI categories and the onset of CKD: aHRs were higher in BMI categories 23.0–24.9, 25.0–26.9, 27.0–29.9, and ≥30.0 kg/m2 in men and 27.0–29.9 and ≥30.0 kg/m2 in women compared with the BMI category 21.0–22.9 kg/m2. However, no clear evidence of a dose-response relationship was observed in our study, and only obesity class I shortened survival to CKD.

Three factors may have influenced the difference between the results of these two studies16. First, the number of participants in our study (n = 7489) was smaller than that of the IPHS (n = 89760). Both studies included non-diabetic Japanese adults and reflected the Asian characteristic of low obesity prevalence (2%–3% of the total population)16,44,45. Therefore, the study’s power may have been insufficient to detect a true effect in a small number of participants with BMI ≥30.0 kg/m2. Second, our participants were older than those in the IPHS. The age of our study participants ranged from 34 to 100 years compared with a range of 40 to 79 years in the IPHS. Consequently, efforts were made to mitigate bias by adjusting for age, thus reducing the likelihood of significant over- and underestimation from older adults’ tendency toward a lower BMI5,46. Notably, however, the generalizability of this study’s findings to populations of different ages is limited. Third, the observational period for our study was longer (1998–2023) than that of the IPHS (1993–2006). As the follow-up period increases, participants’ susceptibility to outcomes decreases47,48. Specifically, our study had more built-in selection bias than the IPHS, which may have led to an underestimation of our results16.

In this study, men had shorter survival by point estimate at a BMI ≥23.0 kg/m2, whereas women had shorter survival at a BMI ≥25.0 kg/m2. Therefore, men had a shorter survival time to CKD onset than women with the same BMI. This trend was consistent with that in the IPHS and previous findings and is partly explained by sex differences16,49. First, men’s higher muscle mass compared with women can lead to systematic errors in eGFR in men when the higher muscle is misclassified as low eGFR50,51. Consequently, the results are likely to be overestimated in men. However, the overestimation may have been attenuated by the large number in our study of older participants (i.e., those aged ≥70 years), whose low BMI and muscle mass may have contributed to the underestimation (Tables 1–2)46. Second, men had shorter follow-up times than women, potentially resulting in larger random errors and wider 95% CIs than women, especially in the category (≥30.0 kg/m2) with the shortest follow-up time.

According to the checkup data for the Japanese general population in FY 2020, the prevalence of CKD (eGFR <60 mL/min/173m2) for men and women was 3.97% and 2.57%, respectively, in their 40s; 7.48% and 6.82%, respectively, in their 50s; 15.8% and 14.8%, respectively, in their 60s; 27.7% and 31.8%, respectively, in their 70s; and 44.6% and 36.1%, respectively, in their 80s and older52. The prevalence of CKD in this study (i.e., the prevalence of those who had CKD at study entry and were excluded from the analysis) was higher than in the Japanese general population (Figure 1). Therefore, unmeasured factors may have influenced the prevalence of CKD among the participants in this study.

Sensitivity analyses

In the first sensitivity analysis, the conventional BMI classification was used as exposure (Table S1). Obesity had the lowest point estimates for men and an almost null value for women, with all 95% CIs crossing 1. Compared with the findings of the main analysis, overweight, which corresponds to obesity class I in the Asia-Pacific classification, was associated with shorter survival to CKD (5% shorter in men and 6% shorter in women). Thus, the cutoff value of BMI ≥30.0 kg/m2 for obesity may be too high a risk factor for CKD in a non-diabetic Asian population.

The second sensitivity analysis was conducted to minimize the possibility of reverse causation by excluding the participants who had developed CKD at the second observation (Table S2). The results for all but the highest category showed a similar trend to that in the main analysis (Table 3), suggesting no clear evidence of reverse causation between the BMI categories of less than 30.0 kg/m2 and CKD onset. However, these findings indicate a possible weak inverse association between the BMI ≥30.0 kg/m2 category and CKD onset that may be partly due to the shorter follow-up period in the BMI ≥30.0 kg/m2 category (449.9 person-years for men [2.03% of the total] and 1027.6 person-years for women [2.94% of the total]) compared with those in the other categories (Table S2). As a result, the relationship between BMI ≥30.0 kg/m2 and the development of CKD could not be fully assessed in all of the models because of reduced power—especially in men, who had the shortest survival to CKD onset by point estimate in the main analysis. Therefore, a type II error may exist in the BMI ≥30.0 kg/m2 category, and a dose- response relationship may exist between BMI and the development of CKD in this analysis.

The third sensitivity analysis, which was based on the stringent CKD definition and in which two consecutive observations of eGFR < 60 mL/min/1.73m2 were considered CKD, showed a similar trend to that of the main analysis (Tables 1–2, S3), suggesting a robust relationship between BMI and the subsequent development of CKD.

Strengths and limitations

A strength of this study was that it used the Asia-Pacific BMI classification to evaluate the impact of elevated BMI on the development of CKD in non-diabetic monoracial Asian men and women using longitudinal data with a long follow-up period of approximately 25 years.

However, several limitations must be specified. First, because East Asians tend to have a lower BMI than Westerners, the number of participants with the highest category of BMI (≥30.0 kg/m2) was low in this study, and the relationship between BMI and subsequent development of CKD could not be fully assessed44. Second, high BMI is not equivalent to obesity. BMI is simply a measure that is calculated based on weight and height and not body composition such as body fat percentage. When obesity is defined as body fat percentages of ≥25% for men and ≥35% for women in an Asian population, detecting obesity using a BMI ≥30.0 kg/m2 has high specificity (97.3% for men and 95.0% for women) but low sensitivity (6.7% for men and 13.4% for women)53. Therefore, a BMI cutoff of ≥30.0 kg/m2 misses most participants with high body fat and results in a misclassification that underestimates the highest BMI category.

Third, our results are not generalizable because this was a longitudinal study that was conducted in a single city and included only 7489 participants. Fourth, the exact date of CKD onset could not be known—a major and unresolved shortcoming of cohort studies that are based on annual checkup data and in which participants can decide when to be screened. These factors lead to interval censoring, which requires extremely complex statistical models to address54.

Fifth, measurement errors in renal function may have occurred. Inulin clearance values, the gold standard for measuring GFR, could not be obtained in our study because the test is too time-consuming and expensive for inclusion in regular checkups by local governments49. Instead, we used SCr to estimate GFR using the three-variable revised Japanese equation, which is more accurate than the Modification of Diet in Renal Disease equations23–25. Unlike inulin clearance, SCr is affected by factors such as moderate/intense physical activity, muscle mass, and high protein intake; participants with such characteristics may be classified as having CKD despite not having the disease51,55,56. However, the mean age of the participants at study entry was late 60s. Participation in sports tends to decline with age, suggesting that the effect of these characteristics, which are more common in young people, was small57.

Sixth, our study participants were healthier than the general population. A previous study reported that those who voluntarily attended checkups were more likely to be health-conscious and healthier than the general population58. This selection bias may be responsible for the underestimated results. Seventh, the presence of competing risks that may have modified the occurrence of an event such as death or hospitalization could not be assessed because the necessary data were unavailable.

Finally, indicators of central and visceral fat, such as waist circumference or waist-to-hip ratio, were not available59. In the Hallym Aging Study, a cohort study of 454 adult Korean participants, elevations in these factors were more strongly associated with CKD than BMI regardless of prevalent diabetes59. In the future, these variables should be used together with BMI to create an obesity index that more clearly indicates the risk of developing CKD by race for men and women without diabetes.

Conclusion

Our study of 7489 non-diabetic Japanese adults in which the Asia-Pacific BMI classification was used showed that men and women with obesity class I had 6% shorter survival to subsequent CKD than those of normal weight. Obesity class II had the shortest survival to CKD onset by point estimates in men and an almost null value in women, with all 95% CIs across 1. The results support previous findings of a dose-response relationship between BMI and the development of CKD, especially in men. Our results add new evidence to support that the BMI cutoff values for obesity that are based on the conventional BMI classification are too high when considering the association between obesity and future CKD development in non-diabetic Asian men and women.

Ethics

All data were anonymized before receipt. The Ethics Committee of Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences and Okayama University Hospital approved this study (No. K1708-040). The Ethics Committee waived the need for informed consent. Study procedures were performed in accordance with the Declaration of Helsinki and Japanese Ethical Guidelines for Medical and Biological Research Involving Human Subjects.

Data statement

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Data Availability

All data produced in the present work are contained in the manuscript and its supplementary information files.

Author contributions

T.M.: Methodology; Supervision; and Writing - review & editing. T.T.: Methodology; Supervision; and Writing - review & editing. Y.O.: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Visualization; Roles/Writing - original draft; and Writing - review & editing.

Competing interests

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Y.O. declares a relationship with Zentsuji City that includes employment. T.M. and T.T. declare no potential competing interests.

Acknowledgments

We are grateful to all participants of this study, Ayaka Nakatsu, Masako Matsumoto, Mayumi Kitadani, and all local government officers of Zentsuji City for their support and contribution. We thank Anahid Pinchis, BSc, MBA, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Footnotes

  • The Methods and Discussion sections were revised in response to reviewers' comments.

References

  1. 1.↵
    Bikbov B, Purcell CA, Levey AS, et al. Global, regional, and national burden of chronic kidney disease, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. The Lancet. 2020;395(10225):709-733. doi:10.1016/S0140-6736(20)30045-3
    OpenUrlCrossRefPubMed
  2. 2.↵
    Stevens PE, Ahmed SB, Carrero JJ, et al. KDIGO 2024 Clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney International. 2024;105(4):S117-S314. doi:10.1016/j.kint.2023.10.018
    OpenUrlCrossRef
  3. 3.↵
    Liyanage T, Ninomiya T, Jha V, et al. Worldwide access to treatment for end-stage kidney disease: a systematic review. The Lancet. 2015;385(9981):1975-1982. doi:10.1016/S0140-6736(14)61601-9
    OpenUrlCrossRefPubMed
  4. 4.↵
    Go AS, McCulloch CE. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. The New England Journal of Medicine. Published online 2004:10.
  5. 5.↵
    Yamagata K, Ishida K, Sairenchi T, et al. Risk factors for chronic kidney disease in a community-based population: a 10-year follow-up study. Kidney International. 2007;71(2):159–166. doi:10.1038/sj.ki.5002017
    OpenUrlCrossRefPubMedWeb of Science
  6. 6.↵
    Tsai WC, Wu HY, Peng YS, et al. Risk factors for development and progression of chronic kidney disease: a systematic review and exploratory meta-analysis. Medicine. 2016;95(11):e3013. doi:10.1097/MD.0000000000003013
    OpenUrlCrossRefPubMed
  7. 7.↵
    Garofalo C, Borrelli S, Minutolo R, Chiodini P, De Nicola L, Conte G. A systematic review and meta-analysis suggests obesity predicts onset of chronic kidney disease in the general population. Kidney International. 2017;91(5):1224–1235. doi:10.1016/j.kint.2016.12.013
    OpenUrlCrossRefPubMed
  8. 8.
    Weldegiorgis M, Woodward M. The impact of hypertension on chronic kidney disease and end-stage renal disease is greater in men than women: a systematic review and meta-analysis. BMC Nephrol. 2020;21(1):506. doi:10.1186/s12882-020-02151-7
    OpenUrlCrossRef
  9. 9.↵
    Shen Y, Cai R, Sun J, et al. Diabetes mellitus as a risk factor for incident chronic kidney disease and end-stage renal disease in women compared with men: A systematic review and meta-analysis. Endocrine. 2017;55(1):66–76. doi:10.1007/s12020-016-1014-6
    OpenUrlCrossRefPubMed
  10. 10.↵
    Kelly JT, Su G, Zhang L, et al. Modifiable lifestyle factors for primary prevention of CKD: a systematic review and meta-analysis. JASN. 2021;32(1):239–253. doi:10.1681/ASN.2020030384
    OpenUrlAbstract/FREE Full Text
  11. 11.↵
    Thomas MC, Brownlee M, Susztak K, et al. Diabetic kidney disease. Nat Rev Dis Primers. 2015;1(1):15018. doi:10.1038/nrdp.2015.18
    OpenUrlCrossRefPubMed
  12. 12.↵
    Wijarnpreecha K, Thongprayoon C, Thamcharoen N, Panjawatanan P, Cheungpasitporn W. Association of coffee consumption and chronic kidney disease: A meta-analysis. Int J Clin Pract. 2017;71(1):e12919. doi:10.1111/ijcp.12919
    OpenUrlCrossRef
  13. 13.
    Thomas G, Sehgal AR, Kashyap SR, Srinivas TR, Kirwan JP, Navaneethan SD. Metabolic syndrome and kidney disease: a systematic review and meta-analysis. CJASN. 2011;6(10):2364–2373. doi:10.2215/CJN.02180311
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    Fabrizi F, Verdesca S, Messa P, Martin P. Hepatitis C virus infection increases the risk of developing chronic kidney disease: a systematic review and meta-analysis. Dig Dis Sci. 2015;60(12):3801–3813. doi:10.1007/s10620-015-3801-y
    OpenUrlCrossRefPubMed
  15. 15.↵
    Flegal KM, Shepherd JA, Looker AC, et al. Comparisons of percentage body fat, body mass index, waist circumference, and waist-stature ratio in adults. The American Journal of Clinical Nutrition. 2009;89(2):500–508. doi:10.3945/ajcn.2008.26847
    OpenUrlAbstract/FREE Full Text
  16. 16.↵
    Tsujimoto T, Sairenchi T, Iso H, et al. The dose-response relationship between body mass index and the risk of incident stage ≥3 chronic kidney disease in a general Japanese population: The Ibaraki Prefectural Health Study (IPHS). Journal of Epidemiology. 2014;24(6):444–451. doi:10.2188/jea.JE20140028
    OpenUrlCrossRefPubMed
  17. 17.↵
    World Health Organization Regional Office for the Western Pacific Region. The Asia-Pacific Perspective_J: Redefining Obesity and Its Treatment. SydneyL: Health Communications Australia; 2000. Accessed April 18, 2024. https://iris.who.int/handle/10665/206936
  18. 18.↵
    Zentsuji City. Zentsuji City official website. Accessed April 18, 2024. https://www.city.zentsuji.kagawa.jp
  19. 19.↵
    Okawa Y, Suzuki E, Mitsuhashi T, Tsuda T, Yorifuji T. A population-based longitudinal study on glycated hemoglobin levels and new-onset chronic kidney disease among non-diabetic Japanese adults. Sci Rep. 2023;13(1):13770. doi:10.1038/s41598-023-40300-8
    OpenUrlCrossRef
  20. 20.↵
    Ministry of Health, Labour and Welfare. Specific health examination and specific health guidance. Accessed April 18, 2024. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000161103.html
  21. 21.↵
    ElSayed NA, Aleppo G, Aroda VR, et al. 2. Classification and diagnosis of diabetes: standards of care in diabetes—2023. Diabetes Care. 2023;46(Supplement_1):S19-S40. doi:10.2337/dc23-S002
    OpenUrlCrossRefPubMed
  22. 22.↵
    WHO expert consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. The Lancet. 2004;363(9403):157-163. doi:10.1016/S0140-6736(03)15268-3
    OpenUrlCrossRefPubMedWeb of Science
  23. 23.↵
    Matsuo S, Imai E, Horio M, et al. Revised equations for estimated GFR from serum creatinine in Japan. American Journal of Kidney Diseases. 2009;53(6):982–992. doi:10.1053/j.ajkd.2008.12.034
    OpenUrlCrossRefPubMedWeb of Science
  24. 24.
    Matsuo S, Yasuda Y, Imai E, Horio M. Current status of estimated glomerular filtration rate (eGFR) equations for Asians and an approach to create a common eGFR equation: approach to eGFR equation for Asians. Nephrology. 2010;15:45–48. doi:10.1111/j.1440-1797.2010.01313.x
    OpenUrlCrossRefPubMed
  25. 25.↵
    Fujii R, Pattaro C, Tsuboi Y, et al. Comparison of glomerular filtration rate estimating formulas among Japanese adults without kidney disease. Clinical Biochemistry. 2023;111:54–59. doi:10.1016/j.clinbiochem.2022.10.011
    OpenUrlCrossRef
  26. 26.↵
    Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: Executive summary: a report of the American college of cardiology/American heart association task force on clinical practice guidelines. Hypertension. 2018;71(6):1269-1324. doi:10.1161/HYP.0000000000000066
    OpenUrlCrossRef
  27. 27.↵
    Kinoshita M, Yokote K, Arai H, et al. Japan Atherosclerosis Society (JAS) guidelines for prevention of atherosclerotic cardiovascular diseases 2017. J Atheroscler Thromb. 2018;25(9):846–984. doi:10.5551/jat.GL2017
    OpenUrlCrossRefPubMed
  28. 28.↵
    NGSP. NGSP Home. Accessed April 18, 2024. https://ngsp.org
  29. 29.↵
    Kashiwagi A, Kasuga M, Araki E, et al. International clinical harmonization of glycated hemoglobin in Japan: from Japan Diabetes Society to National Glycohemoglobin Standardization Program values. Journal of Diabetes Investigation. 2012;3(1):39–40. doi:10.1111/j.2040-1124.2012.00207.x
    OpenUrlCrossRefPubMedWeb of Science
  30. 30.↵
    Lapedis CJ, Mariani LH, Jang BJ, Hodgin J, Hicken MT. Understanding the link between neighborhoods and kidney disease. Kidney360. 2020;1(8):845-854. doi:10.34067/KID.0001202019
    OpenUrlAbstract/FREE Full Text
  31. 31.↵
    Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69(1):239–241. doi:10.1093/biomet/69.1.239
    OpenUrlCrossRefWeb of Science
  32. 32.
    Akaike H. A new look at the statistical model identification. IEEE Trans Automat Contr. 1974;19(6):716–723. doi:10.1109/TAC.1974.1100705
    OpenUrlCrossRef
  33. 33.↵
    Schwarz G. Estimating the dimension of a model. Ann Statist. 1978;6(2). doi:10.1214/aos/1176344136
    OpenUrlCrossRefWeb of Science
  34. 34.↵
    von Hippel PT. How many imputations do you need? A two-stage calculation using a quadratic rule. Sociological Methods & Research. 2020;49(3):699–718. doi:10.1177/0049124117747303
    OpenUrlCrossRef
  35. 35.↵
    Young R, Johnson DR. Handling missing values in longitudinal panel data with multiple imputation. Fam Relat. 2015;77(1):277–294. doi:10.1111/jomf.12144
    OpenUrlCrossRef
  36. 36.↵
    Westreich D, Greenland S. The Table 2 fallacy: presenting and interpreting confounder and modifier coefficients. American Journal of Epidemiology. 2013;177(4):292–298. doi:10.1093/aje/kws412
    OpenUrlCrossRefPubMedWeb of Science
  37. 37.↵
    Python Software Foundation. Python. Python.org. Published April 5, 2023. Accessed April 18, 2024. https://www.python.org/
  38. 38.
    Delker CJ. Schemdraw 0.17 documentation. Accessed April 18, 2024. https://schemdraw.readthedocs.io/en/latest/
  39. 39.
    Bossche JV den, Fleischmann M, McBride J, Ward B, Wolf L, Richards M. GeoPandas 0.14.1 documentation. Accessed April 18, 2024. https://geopandas.org/en/stable/index.html
  40. 40.↵
    National Statistics Center. e-Stat: portal site of official statistics of Japan. Accessed April 18, 2024. https://www.e-stat.go.jp/gis
  41. 41.↵
    Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. Epidemiology. 2007;18(6):805–835. doi:10.1097/EDE.0b013e3181577511
    OpenUrlCrossRefPubMedWeb of Science
  42. 42.↵
    World Health Organization. A healthy lifestyle - WHO recommendations. Accessed April 18, 2024. https://www.who.int/europe/news-room/fact-sheets/item/a-healthy-lifestyle---who-recommendations
  43. 43.↵
    Ibaraki Prefectural Health Plaza. Ibaraki Prefectural Health Study: IPHS. Accessed April 18, 2024. https://www.hsc-i.jp/05_chousa/iphs.htm
  44. 44.↵
    Abarca-Gómez L, Abdeen ZA, Hamid ZA, et al. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128·9 million children, adolescents, and adults. The Lancet. 2017;390(10113):2627–2642. doi:10.1016/S0140-6736(17)32129-3
    OpenUrlCrossRefPubMed
  45. 45.↵
    The GBD 2015 Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med. 2017;377(1):13-27. doi:10.1056/NEJMoa1614362
    OpenUrlCrossRefPubMed
  46. 46.↵
    Tarui I, Okada E, Okada C, Saito A, Takimoto H. Trends in BMI among elderly Japanese population: findings from 1973 to 2016 Japan National Health and Nutrition Survey. Public Health Nutr. 2020;23(11):1907–1915. doi:10.1017/S1368980019004828
    OpenUrlCrossRef
  47. 47.↵
    Hernán MA, Hernández-Díaz S, Robins JM. A structural approach to selection bias. Epidemiology. 2004;15(5):615–625. doi:10.1097/01.ede.0000135174.63482.43
    OpenUrlCrossRefPubMedWeb of Science
  48. 48.↵
    Hernán MA. The hazards of hazard ratios. Epidemiology. 2010;21(1):13–15. doi:10.1097/EDE.0b013e3181c1ea43
    OpenUrlCrossRefPubMedWeb of Science
  49. 49.↵
    Kidney Disease: Improving Global Outcomes (KDIGO). Chapter 1: definition and classification of CKD. Kidney International Supplements. 2013;3(1):19-62. doi:10.1038/kisup.2012.64
    OpenUrlCrossRefPubMed
  50. 50.↵
    Janssen I, Heymsfield SB, Wang Z, Ross R. Skeletal muscle mass and distribution in 468 men and women aged 18–88 yr. Journal of Applied Physiology. 2000;89(1):81–88. doi:10.1152/jappl.2000.89.1.81
    OpenUrlCrossRefPubMedWeb of Science
  51. 51.↵
    Nankivell BJ, Nankivell LFJ, Elder GJ, Gruenewald SM. How unmeasured muscle mass affects estimated GFR and diagnostic inaccuracy. EClinicalMedicine. 2020;29–30:100662. doi:10.1016/j.eclinm.2020.100662
    OpenUrlCrossRef
  52. 52.↵
    Ministry of Health, Labour and Welfare. NDB Open Data. NDB Open Data. Accessed April 18, 2024. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000177182.html
  53. 53.↵
    Goh VHH, Tain CF, Tong TYY, Mok HPP, Wong MT. Are BMI and other anthropometric measures appropriate as indices for obesity? A study in an Asian population. Journal of Lipid Research. 2004;45(10):1892–1898. doi:10.1194/jlr.M400159-JLR200
    OpenUrlAbstract/FREE Full Text
  54. 54.↵
    Chen L, Qiu B. Analysis of lengthLbiased and partly intervalLcensored survival data with mismeasured covariates. Biometrics. Published online July 17, 2023:biom.13898. doi:10.1111/biom.13898
    OpenUrlCrossRef
  55. 55.↵
    Baxmann ACAA, Ahmed MS, Marques NAAC, et al. Influence of muscle mass and physical activity on serum and urinary creatinine and serum cystatin C. Clinical Journal of the American Society of Nephrology. 2008;3(2):348–354. doi:10.2215/CJN.02870707
    OpenUrlAbstract/FREE Full Text
  56. 56.↵
    Ko GJ, Obi Y, Tortorici AR, Kalantar-Zadeh K. Dietary protein intake and chronic kidney disease. Current Opinion in Clinical Nutrition & Metabolic Care. 2017;20(1):77–85. doi:10.1097/MCO.0000000000000342
    OpenUrlCrossRefPubMed
  57. 57.↵
    McPhee JS, French DP, Jackson D, Nazroo J, Pendleton N, Degens H. Physical activity in older age: perspectives for healthy ageing and frailty. Biogerontology. 2016;17(3):567–580. doi:10.1007/s10522-016-9641-0
    OpenUrlCrossRefPubMed
  58. 58.↵
    Bjerregaard AL, Maindal HT, Bruun NH, Sandbæk A. Patterns of attendance to health checks in a municipality setting: the Danish ‘Check Your Health Preventive Program.’ Preventive Medicine Reports. 2017;5:175–182. doi:10.1016/j.pmedr.2016.12.011
    OpenUrlCrossRefPubMed
  59. 59.↵
    Oh H, Quan SA, Jeong JY, Jang SN, Lee JE, Kim DH. Waist circumference, not body mass index, is associated with renal function decline in Korean population: Hallym Aging Study. Vliegenthart R, ed. PLoS ONE. 2013;8(3):e59071. doi:10.1371/journal.pone.0059071
    OpenUrlCrossRef
Back to top
Posted May 08, 2024.
Download PDF

Supplementary Material

Data/Code
Email
Is obesity in the Asia-Pacific body mass index classification associated with the development of chronic kidney disease in non-diabetic Japanese men and women?
Yukari Okawa, Toshiharu Mitsuhashi, Toshihide Tsuda
medRxiv 2024.01.06.23300566; doi: https://doi.org/10.1101/2024.01.06.23300566
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area