Nomogram for predicting risk of mild renal dysfunction among general residents from rural Northeast China

Introduction

Mild renal dysfunction (MRD) is a growing concern worldwide, and there is increasing evidence that it may affect many cardiovascular diseases (CVDs) and cerebrovascular diseases. The Good Aging in Skåne (GÅS) cohort study confirmed that MRD was associated with impairment in learning and memory, language, complex attention, executive function, and global cognitive function, but not meta-memory.^[1] Park et al.^[2] reported that the probability of variations in the left ventricular (LV) geometry was higher in men than in women with MRD. The Framingham Heart study also inferred that individuals with mildly reduced estimated glomerular filtration rate (eGFR) had a higher probability of subclinical atherosclerosis, cross sectionally, and a greater risk of CVD and chronic kidney disease (CKD) progression.^[3] Similarly, a study has shown that MRD might be associated with a combined increase in ventricular systolic stiffness and arterial load in patients with known or suspected coronary artery disease (CAD), leading to an increased risk of heart failure.^[4] Therefore, it is necessary to find effective methods to predict MRD in order to reduce the risk of CVD.

There is a great variation in MRD prevalence between the developed and developing areas. Data from the Framingham Heart Study showed a high prevalence of mildly reduced eGFR (62%),^[3] which was remarkably higher than those observed in the studies conducted in developed countries like Australia (43%), the USA (52.1%), and Turkey (20%).^{[5, 6, 7]} Similarly, the prevalence of mildly reduced eGFR among urban populations in Bangladesh was 24%.^[8] A study conducted to estimate the prevalence of CKD in Cameroonians found that the subjects living in the urban areas had a significantly higher prevalence of MRD (15.1%) compared to those living in the rural areas (9.1%).^[9] In addition, Burkhalter et al.^[10] reported that the MRD prevalence in the rural regions of Haiti was 6.4%, which was lower than that in the urban areas. Previous studies also confirmed that a variety of metabolic disorders, like fasting insulin, hyperglycemia, hyperuricemia, and dyslipidemia, significantly increased the risk of mildly reduced eGFR.^{[11, 12, 13]} However, Onat et al.^[7] claimed that insulin resistance rather than metabolic disorders determined MRD, which was also supported by the National Health and Nutrition Examination Survey (NHANES).^[14] Furthermore, a community-based study conducted by Ji et al.^[15] in the urban areas of Chongqing, a large city in southwestern China, indicated that compared to the higher eGFR group, the waist circumference (WC), systolic blood pressure (SBP), fasting plasma glucose (FPG), and uric acid (UA) of the mildly reduced eGFR group were significantly increased. In addition, Hou et al.^[16] found that triglyceride (TG) levels are closely associated with a mild decline in the eGFR in the middle-aged and elderly Chinese population with normal serum lipid levels. Concerning the possible association between socioeconomic status and renal dysfunction, Shen et al.^[17] found that the prevalence of MRD was higher in subjects belonging to a farming background or with low level of education, whereas the income levels were irrelevant in the rural population of the North-Central Zhejiang province in eastern China.

Compared to the urban residents, the rural residents in China have relatively few medical resources and are also less likely to be concerned about their health.^[18] Furthermore, the New Cooperative Medical Scheme (NCMS) and the Urban Residents’ Basic Medical Insurance System (URBMI) have failed to improve the inequalities in health financing between the urban and rural residents.^[19] Our cross-sectional study showed that the awareness rates of hypertension, diabetes, and dyslipidemia were 17.1%, 3.9%, and 5.8%, respectively,^[20,21] and even less concern was expressed toward MRD. Our previous study, China Rural Hypertension Control Project, concluded that compared to intensive routine care, rural doctor-led interventions showed statistically significant improvements in blood pressure (BP) control among the rural population in China.^[22] Integrated analysis can, therefore, provide useful insights, and effective evaluation and prediction will further help the village doctors identify subjects at potential risk. However, thus far, no studies have been conducted on MRD in rural areas. The purpose of this study is to investigate the incidence and risk factors of MRD and make recommendations for its prevention based on the novel nomogram.

Materials and methods

Study population

The Northeast China Rural Cardiovascular Health Study (NCRCHS) is a community-based cohort study conducted in the rural areas of Northeast China, and the design and inclusion criteria of this study have been described previously.^[21,23] Figure 1 depicts the detailed recruitment and selection process of the participants. A total of 11,956 participants completed the study, with an 85.3% response rate. This study was approved by the China Medical University’s Ethics Committee (Shenyang, China AF-SDP-07-1, 0-01). The participants were invited to participate in the follow-up studies in 2015 and 2017. The average duration of the follow-up period was 4.66 years. All the participants provided written informed consent for the study.

Figure 1

Flow chart of participants included and excluded in this study. eGFR: estimated glomerular filtration rate.

Results

Development of the nomogram for MRD

A nomogram incorporating the five significant risk factors was created (Figure 2), and the total scores were used to assess the risk for MRD. For instance, a rural resident, consuming tea one or two times/day, suffering from abdominal obesity and hypertension, earning 5000–20,000 CYN/year, and without hyperglycemia contributed 281.25 points (= 95 + 17.5 + 86.25 + 30 + 52.5) and had a 36.8% probability of developing MRD. The overall predictive accuracy of the nomogram, as measured by the bootstrap-corrected receiver operating characteristic (ROC) curve, was 0.705, indicating good discrimination (Figure 3). The calibration plot revealed that the nomogram is well calibrated and there is no significant difference between prediction probabilities and observation probability (Hosmer-Lemeshow test, P = 0.232) (Figure 4).

Figure 2

Nomogram for predicting risk of mildly reduced eGFR among rural Northeast Chinese. eGFR: estimated glomerular filtration rate.

Figure 3

ROC analysis of the nomogram for mildly reduced eGFR among rural Northeast Chinese. AUC: area under curve; ROC: receiver operating characteristic; eGFR: estimated glomerular filtration rate.

Figure 4

Calibration plot of the nomogram for mildly reduced eGFR among rural Northeast Chinese. Hosmer-Lemeshow test, P-value 0.232. eGFR: estimated glomerular filtration rate.

Discussion

The primary finding of our study is that the proportion of MRD is relatively high in the Chinese rural general population. Therefore, effective prediction is needed for early diagnosis and prevention of renal damage. MRD was associated with annual income, abdominal

obesity, high BP, hyperglycemia, and tea consumption frequency. A new nomogram model was established to better evaluate the risk of MRD in the rural Northeast Chinese population.

Many previous studies have shown that decreased renal function is a predictor of hospitalization,^[27] cognitive impairment,^[28] and poor quality of life, often accompanied by an increased risk of cardiovascular events like arterial stiffness, coronary artery calcium, and myocardial hypertrophy.^[29] In addition, it has been an important issue in the management of CVD and cerebrovascular diseases. According to some previous studies, the prevalence of MRD varies by 2.8%–15.1%;^[30,31] however, in our study, a relatively higher cumulative incidence was observed, with 889 (18.0%) out of 4944 subjects experiencing MRD after 4.6 years of follow-up. The relative rates of hypertension (51.1%), diabetes (6.5% previously diagnosed; 8.7% undiagnosed), and dyslipidemia (36.9%) among the rural residents were higher than those recorded in the previous studies and could be responsible for the increased incidence of MRD during the follow-up.^{[21, 32, 33]} Hypertension, glucose levels, and lipid metabolism disorders have been estimated to increase the risk of MRD.^{[34, 35, 36, 37]} Furthermore, studies have linked an increase in eGFR to a steady decline in cardiovascular morbidity and mortality in subjects with MRD, suggesting an inverse relationship between eGFR and cardiovascular risk.^[38,39] Therefore, it is particularly important to assess the risk factors associated with MRD and make recommendations for its prevention.

Multifactorial etiologies were associated with MRD. In the current study, univariate and multivariate analyses identified six independent risk factors for MRD. Previous studies reported that the prevalence of dyslipidemia and diastolic BP variability was higher in subjects with MRD than in subjects with normal renal function.^[15,40] As previously claimed, elevated TG (odds ratio [OR]: 1.25), decreased HDL-C (OR: 1.23), and obesity (OR: 1.22) were associated with increased risk of mildly reduced eGFR, while elevated FPG (OR: 0.64) was associated with lower risk of mildly reduced eGFR.^[41] Our results support that abdominal obesity, hypertension, and hyperglycemia are strongly associated with MRD, and that abdominal obesity is an independent risk factor for MRD, while hyperglycemia is negatively associated with MRD. Many pathophysiological pathways (e.g., chronic inflammation, elevated oxidative stress, and hyperinsulinemia) or a set of risk factors (e.g., insulin resistance, hypertension, and dyslipidemia) may explain the possible association between obesity and renal damage.^[42] Our results are consistent with those of Hu et al.^[41] who claimed that hyperglycemia was associated with a relatively lower risk of MRD, which is inconsistent with several previous studies.^[43,44] This is possibly attributed to the hyperfiltration due to elevated FPG and glycated hemoglobin (HbA1c) levels resulting in a temporary increase in eGFR.^[41] Another possibility is the relatively short follow-up duration, since the results can vary with time. Therefore, large prospective studies with relatively longer follow-up durations should be conducted to explore the possible association between hyperglycemia and mildly reduced eGFR.

Previous studies have shown that the rates of elevated diastolic blood pressure were significantly higher in subjects with mildly reduced eGFR than in subjects with normal eGFR.^[15,40] In addition, hypertension has been proven to be a well-established risk factor for the progression of CKDs.^[14,45] As the economy grows, the prevalence of metabolic disorders, such as obesity, hypertension, and hyperglycemia, is increasing rapidly in both urban and rural areas.^[20,46,47] However, the rates of awareness, treatment, and control are significantly lower among the rural Chinese population compared to the urban population.^[48,49] Additionally, the management of diabetes and hypertension in the Chinese population is far from satisfactory. According to a Chinese national survey, only 32.2% of patients with diabetes mellitus (DM) received treatment and only 49.2% of the treated patients had < 7.0% HbA1c level.^[50] In addition, the rural Chinese population had significantly lower rates of awareness (4.6%) and treatment (3.7%). According to the China Hypertension Survey (2012–2015), 23.2% of Chinese adults suffer from hypertension, 46.9% are aware of their hypertension diagnosis, 40.7% take prescribed antihypertensive medications, and 15.3% have controlled hypertension.^[51] Similarly, our previous study on the rural Chinese population revealed that the awareness, treatment, and control rates for new cases of hypertension were 29.9%, 19.5%, and 1.5%, respectively.^[52] Low rates of awareness, treatment, and control in rural areas can lead to serious complications and cardiovascular mortality.

Interestingly, frequent tea consumption also increased the risk of MRD in the rural population. Tea consumption has been prevalent in Asia for centuries, especially in Japan and China, where it is the most consumed beverage after water. Tea is a rich source of pharmacologically active molecules and has many health benefits.^[53] Cumulative evidence from cell culture and animal studies confirms the beneficial effects of tea polyphenols in a wide range of pathological diseases, including cancer, diabetes, and CVDs.^[54,55] In addition, tea consumption prevents increased creatinine clearance and has antiproteinuric effects.^[56,57] However, most of the previous studies on tea consumption were conducted in cell cultures or experimental animal models. One recent cross-sectional study of the general Japanese population found that consumption of green tea may increase the risk of mildly reduced eGFR in Japanese men with NADH dehydrogenase subunit-2 237 leucine/methionine (ND2-237 Leu/Met) polymorphism.^[58] In contrast, a large population-based study on the older Chinese population showed no clear evidence between renal function and tea consumption.^[59] Therefore, it might be one possible explanation for the inconsistent results on the relationship between tea consumption and renal function, and further studies are required to confirm this association.

Nomograms are graphical tools or models that are used to estimate the probability of an event and accurately predict the outcomes for individual subjects. We developed a nomogram with six predictors that displayed good accuracy and discrimination in predicting the risk of MRD, with a corrected area under the ROC curve (AUC) of 0.705. Using this model, rural clinicians can accurately estimate the risk of MRD in rural subjects and tailor treatment to those who require close follow-up. To our knowledge, our study is the first to assess the risk of MRD in rural subjects. However, this study has a few limitations: (1) we did not assess the gene polymorphism among rural subjects, which could have explained the association between tea consumption and MRD; (2) the eGFR calculation was based on a single blood test, which can introduce bias; (3) although our analyses were adjusted for many possible confounding factors, residual confounding attributed to the measurement error in the assessment of confounding factors, unmeasured factors, such as family history of DM, hypertension, CVD, and other chronic diseases and antihyperglycemic, antihypertensive, and antihyperlipidemic therapeutic agents, cannot be excluded; (4) we only assessed the renal function by the eGFR equation; therefore, further studies are required to identify the risk factors for MRD by considering the other important indicators such as proteinuria, microalbuminuria, or urine microalbumin– creatinine ratio; and (5) even though we excluded the subjects with MRD at the baseline study, we did not adjust some factors that could possibly affect eGFR, such as the medicine consumed.

Conclusion

The results of our study indicate that MRD is prevalent among rural residents in Northeast China. Additionally, MRD is more prevalent in rural subjects with relatively higher annual income, abdominal obesity, hypertension, and higher tea consumption frequency. We have constructed a novel nomogram that is a good predictor of MRD and can serve as a guide for epidemiological screening.