Urinary Albumin-Creatinine Ratio (UACR), Even within Normal Range, and Risk of Hypertension (HTN), Type 2 Diabetes Mellitus (T2DM), HTN with T2DM, Dyslipidemia and Cardiovascular Diseases in the Chinese Population: A Report from REACTION Study

Background: Albuminuria has been widely considered as a risk factor for cardiovascular diseases (CVDs)(cid:0) which is associated with hypertension (HTN), type 2 diabetes mellitus (T2DM), HTN with T2DM and dyslipidemia. However, it is unclear the association between albuminuria and HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs. Thus, this study is aimed to thoroughly explore the association of albuminuria, even within the normal range, with the abovementioned diseases in the Chinese population. Methods: This study included 40188 participants aged over 40 years from seven centers across China. Urinary albumin-creatinine ratio (UACR) was rstly divided into the ≥ 30 mg/g group, indicating kidney damage, and < 30 mg/g group. Furtherly, UACR was divided into ve groups: the < 20% group, the 20–39% group, the 40–59% group, the 60–79% group and the ≥ 80% group, according to quintile division of participants within the normal range. Propensity score matching was used to reduce bias, and multiple logistic regression models were conducted to examine the association between UACR and HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs. Results: Multivariable regression analysis revealed that UACR, even within the normal range, is signicantly associated with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs, and the association between UACR and HTN with T2DM was most signicant in model 3 even after adjusting for confounding factors (HTN: OR: 1.76, 95%CI: 1.65-1.88, P<0.0001; T2DM: OR: 1.98, 95%CI: 1.84-2.12, P<0.0001; HTN with T2DM: OR: 2.37, 95%CI: 2.19-2.57, P<0.0001; Dyslipidemia: OR: 1.08, 95%CI: 1.01-1.14, P=0.0154; CVDs: OR: 1.14, 95%CI: 1.02-1.27, P=0.0244). In the stratied analysis, high normal UACR was signicantly associated with HTN, T2DM, HTN with T2DM, dyslipidemia in subgroups. Conclusions: We conclude the higher prevalence of HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs in abnormal UACR and reveal a signicant association of UACR, even within the normal range, with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs. T2DM, HTN with T2DM, dyslipidemia and CVDs.


Introductions
There is a growing population of elderly adults with hypertension (HTN), diabetes, dyslipidemia and cardiovascular diseases(CVDs) of increasingly complexity and a corresponding rise in health care burdens. Indicators that stratify risk for the general population across the metabolic abnormalities would possess great clinical value.
Albuminuria has been widelyrecommended as anindicator of renal damage. Emerging data has shown that albuminuria is not only an initial manifestation of renal function loss, but also a nonnegligible risk factor for CVDsespecially in population with diabetes, HTN or dyslipidemia [1][2][3]

. Among diabetic adults in the United
States, the prevalence of diabetic kidney disease (DKD) is about 34.5%, and 16.8% present with albuminuria 4 .Further, evidence is increasing that the presence of albuminuria indicates a 2.5-fold increased risk of stroke, which is consistentwith the ndings delivered by Norfolk'sresearch 5,6 .Moreover, albuminuriaas measured by urine albumin to creatinine ratio (UACR), even within the normal range, is aneffective predictor of HTN 7 . Similarly, a 5-year-follow-up study conducted in Korean men, demonstrated that high normal albuminuria (UACR<30mg/g)could predict the increased risk of diabetes 8 . A robust body of literature has reported a strong, positive association between albuminuria and dyslipidemia in prediabetic and general population 9,10 . Hence, albuminuria is valuable in identifying the general population at risk for CVDs, diabetes, HTN and dyslipidemia in clinical practice.
Previous studies have described an association between albuminuria and related metabolic diseases. However, little literature placed focus on the prevalence of diabetes, HTN, diabetes with HTN and dyslipidemia in the Chinese population with different UACR level. The present study is the rst populationbased study in the Chinese population and may uncover the incidence of diabetes, HTN, diabetes with HTN and dyslipidemia in individuals with different UACR level. Therefore, this current study is aimed to investigate the prevalence of diabetes, HTN, diabetes with HTN and dyslipidemia in different albuminuria range and explore the internal association between albuminuria and metabolic abnormalities.

Study population and design
The present study wasdrawn from the REACTION (Risk Evaluation of Cancers in Chinese Diabetic Individuals) study, which was conducted to investigate the association of diabetes and prediabetes with the risk of cancer in the Chinese population.Detailed information of the REACTION study has been described previously 11 . The REACTION study was set up as a multi-center prospective observational study, and our study used baseline data from seven centers across China. A total of 47808 participants aged over 40 years were recruited from May and December 2012. (Liaoning 10140, Gansu 10026, Guangzhou 9743, Sichuan 8105, Shanghai 6821, Henan 1978, Hubei 995). Participants diagnosed with kidney diseases, cancer, fatty liver, viral hepatitis, cirrhosis, those using ACEI/ARB medicinesand those with missing data were excluded.Then, 41757participants were enrolled.Given differences in the baseline characteristics between the two different UACR groups, the propensity score matching was performed to reduce potential bias. Ultimately,40188 eligible participants were included in this nal analysis. (Figure 1).
The staff received extensive training related to the study questionnaire and outcome measures before the investigation. The study was conducted in accordance with Declaration of Helsinki, and the protocol was approved by the Clinical Research Ethics Committee of Rui-Jin Hospital a liated with the School of Medicine, Shanghai Jiao Tong University. Written informed consents were obtained from all participants before the study.

Data collection and measurements
Data collection was performed by the well-trained staff, which included a standardized questionnaire, anthropometric measurements, blood collection, urine collection and a 75 g OGTT or 100 g steamed-bread meal test. The self-administered questionnaire consisted of demographic information, the history of diabetes, HTN, dyslipidemia, kidney diseases, hepatic diseases, CVDs (coronary heart disease (CHD), myocardial infarction (MI), stroke), the current use of drugs, lifestyle including alcohol consumption and smoking consumption. Alcohol consumption was de ned as follows: never; occasional drinkers who drank less than once a week; regular drinkers who drank at least once a week for over six months. Smoking consumption were de ned as follows: never; occasional smokers who smoked less than one cigarette per day or less than 7 cigarettes per week; regular smokers who smoked at least one cigarette per day.
Anthropometric measurements, including the measurements of height, weight, waist circumference (WC) and blood pressure, were performed by the same well-trained staff. All participants were required to be in light clothing and take off shoes when weight and height were measured to the nearest 0.1 cm and 0.1 kg. WC was measured to the nearest 0.1 cm at the umbilical levelwhen participants were in a standing position 12 . Body mass index (BMI) was calculated using the following formula: BMI = body weight/ height 2 (kg/m 2 ). Blood pressure and heart rate (HR) were recorded three times consecutively by the same welltrained staff at 5-min intervals after participants were in a seated position for at least 5 min at rest. The three measurements of blood pressure and HR were averagedfor the nal analysis.
Blood samples were collected by the experienced nursesin the morningafter at least 12h of overnight fasting. Participants with or without a history of diabetes underwent a 100 g steamed-bread meal test or 75 g OGTT, respectively. After 2h, the second venous blood samples were obtained by the same welltrainednurses. Fasting plasma glucose (FPG), 2 h post-load blood glucose (PBG), Haemoglobin A1c (HbA1c), serum triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), lowdensity lipoprotein cholesterol (LDL-C), aspartate aminotransferase (AST), alanine aminotransferase (ALT) and creatinine (Cr) were measured at every center.
The estimated glomerular ltration rate (eGFR) was calculated on the basis of the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 13 .

De nition of Variables
Fasting midstream urine samples were collected in the morningto measure the concentration of urine albumin and urine creatinineby using chemiluminescence immunoassay in every center.UACR was calculated by dividing urine albumin in milligrams by urine creatinine in grams. The same range and units of UACR measurementwere used in all seven centers. According to the KDIGO CKD guidelines, increased albuminuria was de ned as UACR ≥ 30 mg/g, indicating kidney damage 14 .
According to the WHO guidelines, type 2 diabetes mellitus (T2DM) was de ned as FBG ≥ 7.0 mmol/L, or PBG ≥ 11.1 mmol/L, or diagnosed as T2DMby clinicians and meanwhile undergoing hypoglycemicmedication therapy.HTNwas de ned as systolic blood pressure (SBP) ≥ 140 mmHg or diastolic blood pressure (DBP) ≥ 90 mmHg, or diagnosed as HTN by clinicians and meanwhile undergoing antihypertensive-medication therapy. Dyslipidemia was de ned as increased TC (≥6.20 mmol/L), LDL-C(>4.13 mmol/L), TG (>2.25 mmol/L),decreasedHDL-C (<1.03 mmol/L),or a combination of the above lipid abnormalities. HTNwith T2DM was de ned as a combination of HTNand T2DM. Stroke was de ned as a self-reported history of language or physical dysfunction lasting over 24 h and ischemic or hemorrhagic stroke by imageologicaldiagnosis. CHD events were de ned as a self-report history of myocardial infarction or angina, or coronary revascularization by clinicians. CVDs were de ned as a self-reported history of CHD, stroke, or myocardial infarction events.
Participants were divided into three groups according to their smokingfrequency: no: never or have already quit smoking; occasional: smoking less than once a week or less than 7 cigarettes weekly; frequently: smoking one or more cigarettes daily for at least a half year. Similarly, participants were divided into three groups according to their alcohol intake frequency: no: never or have already quit drinking; occasional: drinking less than once a week; frequently: drinking more than once a week for at least a half year.

Statistical analysis
Empower(R) (www.empowerstats.com, X&Y Solutions Inc., Boston, MA) and R (http://www.Rproject.org) were used to perform the statistical analyses. Given differences in the baseline characteristics between eligible participants between the two groups of UACRs, propensity score matching was employed to control for potential bias. Matching was performed using a 1:7 matching protocol to match all covariates, with a calliper width equal to 0.05 of the SD of the logit of the propensity score.
Continuous variables with a non-normal distribution were presented as median (Q1-Q3), and those with a normal distribution were presented as means ± the standard deviations (SD). Categorical variables were expressed as n%. Differences incontinuous variables were compared using the Kruskal-Wallis test, and when variableswere categorical, the χ2 test was used. Multivariate logistic regression analysis was performed to control potential confounding factors for identifying the associations of UACR with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDsin three models. Model 1was unadjusted. Model 2 was adjusted for age and BMI. Model 3 was further adjusted for sex; SBP; DBP; HR; ALT; AST; eGFR; smoking habits, drinking habits, FBG, PBG, TC, TG, HDL-C, LDL-C and history of medication.The odds ratio (OR) and corresponding 95% con dence intervals (95% CI) were calculated.
To thoroughlyexplore the associations between UACR and HTN, T2DM,HTN with T2DM, dyslipidemia and CVDs, multivariate logistic regression analysis was also conducted in participants with normal albuminuria (UACR<30 mg/g). Furtherly, in order to examine the internal link of the relationship between UACR and HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs, subgroups were strati ed by the different history of HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs in the strati ed analyses. All statistical tests were two-sided, and P values < 0.05 were considered statistically signi cant.
Similarresults were found inHTN with T2DM. Similarly, participants with dyslipidemia had an inferior control of blood pressure, glucose, lipid, a higher prevalence of HTN, T2DM, CVDs, and were more frequent smokers, drinkers (Table S1-S4).
Associations of UACR with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDsin the total population Multiple logistic regression models that consider the association of albuminuria with the individual prevalence of HTN, T2DM, HTN with T2DM, dyslipidemia andCVDs were constructed. Table 2 shows OR and 95% CI of HTN, T2DM, HTN with T2DM, dyslipidemia,CVDswith continuous UACR and categories of UACR in the total populationof three different models. As seen in Table 2 Associations of UACR with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDsin participants with normalUACR (<30mg/g) Tothoroughly explore the association of albuminuria with such diseases, multiple logistic regression models were also constructed in participants with normal range of albuminuria (UACR<30mg/g)as shown in Table   3. These results indicate that compared with participants with lowerUACR levels, participants with higher normal UACR levels (HTN:the second quintile to the fth quintile; T2DM:the second quintile to the fth quintile; HTN with T2DM: the second quintile to the fth quintile; Dyslipidemia: the third quintile to fth quintile; CVDs:the fth quintile) were also signi cantly associated with HTN, T2DM, HTN with T2DM, dyslipidemia andCVDs, even after adjustments for confounding factors in model 3.In model 3, the ORs for HTN were increased signi cantly from the second quintile, with ORs Strati ed analysis of associations between UACR andHTN, T2DM, HTN with T2DM, dyslipidemia and CVDsin participants with normal UACR (<30mg/g) Strati ed analyses were conducted in the different subgroupsofHTN, T2DM, dyslipidemia and CVDsto validate the abovementioned results,shown in Table 4. The present study found that compared with lower UACR, higher normal UACR (thethird, fourth and fth quintiles)was closely associated with HTNinboth subgroups of HTN, T2DM, dyslipidemia and CVDs.To be noted, these associations were most signi cant in participants that were both in the subgroup of the fth quintile of UACR and the subgroups ofnormal blood

Main ndings
Asfar as we all know, this is the rst study conducted in a Chinses general population to observe the prevalence of HTN, T2DM, HTN with T2DM, dyslipidemia and CVDswith different UACRleveland explore the associations between albuminuria and the above diseases, even when albuminuria within thenormal range.
The following are the main ndings of this current study:(1) compared with participants with normal albuminuria (UACR<30mg/g), participants with abnormal albuminuria (UACR ≥30mg/g) had a higher prevalence of HTN (

UACR and HTN
It is widely accepted that HTN is an important risk factor contributing to mortality worldwide and albuminuria plays a crucial role in the initiation and progression of HTNin previous studies 15,16 . This research has described that albuminuria excretion more than 6mg per day can effectively predict the progression of HTN.Notably, a growing number of studies performed in western population have pointed out that a signi cant association between albuminuria and HTNwas not restricted to abnormal albuminuria (UACR≥30mg/g); albuminuria below the normal threshold (UACR<30mg/g) was also found associated with HTN. TheFramingham Heart Studyhas reported that men with UACR >6.66 mg/gand women with>15.24 mg/ghad an approximately 2-fold risk of HTN, indicating UACR being a useful biomarker for identifying individuals at high risk for HTN 17 . Moreover, a positive association between UACR within the normal range and HTN was revealed in postmenopausal women without diabetes, suggesting the revaluation of normal albuminuria excretion 18 .
Similarly, we found that not only abnormal UACR, but also an increase in UACReven within the normal rangeis closely associated with HTN, especially in people with T2DM and CVDs.Although the prevalence of HTN was higher in people with abnormal UACR than those with normal UACR (61.11%VS 40.26%), people with high normal UACR (the fth quintile) have a 1.55-fold risk of HTN than those with low UACR (the rst quintile) within the normal range in our study (Table 4).Our results showed clearly that high normal UACR is closely associated with HTN and the association is independent of eGFR levels, which was consistent with previous studies. Systemic and glomerular vascular abnormalities was thought to be a physiologiclink between albuminuria and HTN 19,20 . The presence of albuminuria could be caused by physiologic abnormalities of glomerular endothelial cells, the glomerular basement membrane, or podocytes, leading to increased ltration of albumin. It is likely that increased albuminuria re ects generalized microvascular endothelial cell damage 21 , which possibly predispose to an increased atherogenic lipoproteins accumulation within subendothelial cell space 22 .A cohort study based on Japanese population, which followed 412 normotensive individuals without diabetes for a median 6.7 years,observed that a slight increase in UACR wasclosely associated with the incidence of HTN and isa predictor of increased blood pressure and incident HTN 23 , suggesting that increased UACR in this current study is partly due to increased blood pressure below the level of diagnosis of HTN.

UACR and T2DM
There is strong evidence that albuminuria could be well indicative of microvascular dysfunction 21 .
Compared with individuals without T2DM, the microvascular function of individuals with T2DM is markedly impaired 24,25 .Louiset al. pointed out that the levels of albuminuria were independently associated with the severity of cardiac macrovascularfunction in individuals with T2DM 26 .Although among diabetic individuals with normal ventricular diastolic function,the prevalence ofcardiac macrovascular dysfunction was higher, especially in those with abnormal albuminuria, which was in line with our results. UACR is signi cantly higher in individuals with T2DM than those without T2DM in the present research (Table 1), indicating a close link between UACR and T2DM. It was well proved that UACR is not only aknown indicator of kidney damage but alsoan effective predictor of atherogenic state. Accordingly, the results of population-based studies supported that UACR is valuablein predictingcardiovascular outcomes in clinical practice 27,28 .
Interestingly, we noted that albuminuria, even within the normal range,is closely associated with T2DMin our study. Participants with abnormal albuminuria(UACR≥30mg/g) had a 1.98-fold risk of T2DM than those with the normal range (UACR<30mg/g) (P<0.0001).When UACR within the normal range, participants with high normal albuminuria (the fth quintile)were still more likely to have the incidence of T2DM (OR 2.17, 95%CI 1.90-2.48, P<0.0001), and the association between high normal albuminuriaand T2DM was more signi cant in participants without HTN and dyslipidemia. This difference may be explained, in part, by smaller sample size of HTN and dyslipidemia group than non-HTN and non-dyslipidemiagroupin our study.Further large sample and prospective studies are necessary to clarify the association between UACR and the incidence of T2DM in different levels of blood pressure and lipids.

UACR and CVDs
As we all know, albuminuria is an established risk factor forCVDs morbidityand mortality both in diabetic and hypertensive individuals. Moreover, in the national andinternational guidelines, albuminuria is recommended as a routine screening parameterin individuals at high risk for CVDs [29][30][31] andhas been recognized as a signi cant indicator of the incidence generalized atherosclerosis because of the close association of albuminuria with atherosclerotic risk factors and microvascular endothelial damage 32 .
Findings from population-based studies have reporteda signi cant relationship between albuminuria and CVDs 27,28 . Studies onindividuals without T2DM and HTN also reached similar conclusions, which was in line with our ndings. 33 A prospective study, including 2484 white subjects, found that non-diabetic individuals with albuminuria have a 1.38-fold increased risk of cardiovascular mortalityafter adjustment for a wide spectrum of risk factors and a markedly high 5.68-foldincreased risk of cardiovascular mortality was observed in diabetic population 34 . This signi cant association was also assessed in the general population in this research. Additionally, a study of 40548 individuals observed that a 2-fold increase in albuminuria conferred a 1.29-fold increase in the risk of cardiovascular mortality 35 . The results of our study, which showedan association between abnormal UACR and CVDs in the general population in seven regions across China, agree with earlier ones. Participants with abnormal albuminuria(UACR≥30mg/g) had a 1.14-fold increased risk of CVDs than those with normal albuminuria in our study.
Interestingly, we noted that UACR, even within thenormal range, exhibited a signi cant association with CVDs in our study even after adjusting for confounding factors. Several studies pointed out that low-grade albuminuria can predict the incidence of CVDs events andCVDs death 36 .The Framingham Study, includingmiddle-aged non-hypertensive and nondiabetic individualswith normal UACR , found that low-grade UACR well below the abnormal threshold can effectively indicate the development of CVDs 37 . Any degree of albuminuria has been proven to be a risk factor for CVDs in diabetic and non-diabetic patients; the risk increases with albuminuria, even below the microalbuminuria cutoff 38 .Every 3.5mg/g increment in UACR conferred a 5.9% increased risk of CVDsacross a wide spectrum of UACR after adjustment for age and sex 38 . Arnlov J et al. proposed that a nearly 3-fold increased risk of CVDs in people without HTN and T2DM but with UACR ≥3.9mg/g in men, ≥7.5mg/g in women, which was equal to the sex-speci c median value 37 .
In our study, we also found a positive association between UACR within thenormal range and the risk of CVDs.CVDs events has been pronounced to be predictable by UACR variation within the normal range 39 .The discrepancies between UACR within the normal range and CVDs in different subgroups might be account for the interaction of strati cation variables with CVDs.It is documented that UACR was signi cantly associated withcomponents of metabolic syndrome, including blood glucose, pressure and lipids level 36 .ACC/AHA and ESC/EAS guidelines have recommendedLDL-C to be the most crucial risk lipid factor and therapeutic goal for CVDs 40 , and the association between UACR within the normal rangeand CVDs in participants with dyslipidemia was at the borderline signi cant level in our study.It iswell noticing that despite the achievement of optimal LDL-C level, a worrisome number of CVDs events still occurinclinical practice 41,42 .In fact, the contribution of other lipid components and subfractions to CVDs development is increasingly being recognized 43,44 .Traditionally, high HDL-C wascon rmed to be protective against the incidence and development of atherosclerosis, and low HDL-Cwas associated with increased risk of CVDs 45. However, recent clinical trials reported that low HDL-C is not a cause of atherosclerosis,as originally thought, renewed interest in elevated TGhas been generated 46 . A growing up of studies supported the theory that elevated TG has a remarkable association with increased risk of CVDs 47,48 .Moreover, reports from the CACTI Studypointed out that TG independently predicted increased odds of both related CVDs and albuminuria in patients with diabetes. Apart from this, several studies placed great importance on the average levels and ideal targets of glycemic parameters, and it was shown that individuals with CVDs can bene t from well control of blood glucose 49,50 .Although elevated glucose parameters has been treated as a modi able cardiovascular risk factor and a robust predictor of CVDs, HbA1cserves as a superior indicator ofcardiovascular events than FBG and PBG in clinical practice 51 . This might be account for the uctuation of FBG and PBGin different individuals, which could be in uenced by various factors. Lots of researches have been carried out on the relationship between HTN and CVDs.The relationship between blood pressure and the increased risk of CVDshas been reported to be graded and continuous, starting from 115/75 mmHg, well within what is thought to be the normotensive range 52 .In fact, it is of great importance to comprehensively consider the predicted risk of atherosclerotic CVDsrather than the level of blood pressure alone, as patients with high CVDs risk couldderive thebene ts fromblood pressure lowering treatment 53 .
Moreover, an association has also been reported between albuminuria, stroke and peripheral vascular diseases in several studies 6,54,55 . The presence of albuminuria may occur due to vascular damage, indicating systemic endothelial dysfunction. The abovementioned evidence may further support our ndings. Thus, early identi cationand prevention of albuminuria is of great signi cance and could contribute to reducing the risk of CVDs.

Limitations
Our study was a multi-center study based on seven-region community population, which representatively demonstrate the distribution of different regions across China.However, there are still limitations in our study. First, the variables in our study were measured at the same time. As a feature of the cross-sectional study, only associations, rather than causality, can be determined. Thus, the association of UACR with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDsshould be further explored in follow-up studies. Second, because the elderly population were from China, the association among other ethnic populations are needed to be con rmed. Third, although the participants using ACEI/ARB were excluded in our study, the possibility that other medications may partially in uence the associationcould not be eliminated. Herein, we emphasize the association between UACR, even within the normal range and increased risk of HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs, andsuch people should be vigilant about the detection, avoidance and intervention of the presence of albuminuria.

Conclusion
In summary, we observe the higher prevalence of HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs in abnormal UACR and reveal a signi cant association of UACR, even within the normal range, with HTN, T2DM, HTN with T2DM, dyslipidemia and CVDs.Thus, we propose that albuminuria might be a simple ande cient indicator of the metabolic diseases as well as CVDs and targeting the early prevention as well as intervention of albuminuria metabolism may increase the possibility of successful drug discovery in the eld of CVDs and its related diseases.

Declarations
Availability of data and materials The datasets used to support this study are not freely available due to participants' privacy protection.