Association of blood pressure and long‐term change with chronic kidney disease risk among Chinese adults with different glucose metabolism according to the 2017 ACC/AHA guidelines

Abstract Whether the definition of hypertension according to 2017 AHA/ACC guidelines and blood pressure (BP) changes was related to the increased risk of chronic kidney disease (CKD) remained debated. This prospective cohort study aimed to investigate the association of BP and long‐term BP change with CKD risk with different glucose metabolism according to the new hypertension guidelines. This study examined 12 951 participants and 11 183 participants derived from the older people cohort study, respectively. Participants were divided into three groups based on blood glucose and the risks were assessmented by the logistic regression model. During a 10 years of follow‐up period, 2727 individuals developed CKD (21.1%). Compared with those with BP < 130/80 mmHg, individuals with increased BP levels had significantly increased risk of incident CKD. Participants with BP of 130–139/80–89 or ≥140/90 mmHg had 1.51‐ and 1.89‐fold incident risk of CKD in patients with diabetes mellitus (DM). Compared with individuals with stable BP (−5 to 5 mmHg), the risk of CKD was reduced when BP decreased by 5 mmHg or more and increased when BP increased ≥5 mmHg among normoglycemia and prediabetes participants. Similar results were observed for rapid estimated glomerular filtration rate (eGFR) decline. In conclusion, the BP of 130–139/80–89 mmHg combined with prediabetes or DM had an increased risk of incident CKD and rapid eGFR decline in older people. Long‐term changes of BP by more than 5 mmHg among normoglycemia or prediabetes were associated with the risk of incident CKD and rapid eGFR decline.


INTRODUCTION
Chronic kidney disease (CKD) is an important global public health problem. It affects 10-15% of the population worldwide and is now recognized as the most rapidly increasing contributor to global burden of disease. 1,2 Diabetes mellitus (DM) and hypertension are two of the most powerful risk factors of CKD. 3 From a public health perspective, effective and safe control of blood pressure (BP) and blood glucose is essential for the prevention and treatment of CKD.
In 2017, the American College of Cardiology and the American Heart Association (ACC/AHA) Guideline on BP in adults was released 4 ; however, there has been controversy surrounding the applicability of the guidelines' definition of hypertension (BP ≥ 130/80 mmHg). [5][6][7][8][9] Whether the new definition of hypertension was related to the increased risk of CKD remained debated, [10][11][12][13] and data are lacking on what extent these BP stratums affect CKD risk in Chinese adults. More importantly, there has been disagreement with BP reduction for CKD risk with different glycemic status. 14,15 It was showed that intensive lowering of BP (target BP less than 120/80 mmHg) did not provide significant benefit for CKD development among patients with both CKD and hypertension but not with diabetes. 16 However, a recent study reported intensive lowering of BP increased incident CKD risk in individuals with and without type 2 diabetes. 17 Meanwhile, the impact of prediabetes on the effects of BP control on kidney events also remained uncertain. 18 Therefore, it is necessary to investigate how BP levels of 130-139/80-89 mmHg affect CKD risks with different glycemic status.
In addition, BP levels at a single time point may not be sufficient to evaluate the disease risk. Prospective studies have pointed out that a worse change in BP control categories over 1 year was associated with increased occurrence of renal outcomes. 19 However, different initial BP and observation time will affect the effect of BP changes on CKD. 20 Meanwhile, there is limited evidence on the relationships between long-term BP changes on CKD in individuals with different glycemic status, especially when BP is elevated to the hypertension category defined by the 2017 ACC/AHA.
In the current study, we aimed to explore the association of newly defined hypertension criteria by the 2017 ACC/AHA and long-term BP changes with 10-year incident CKD risk in middle-aged and elderly Chinese with different glycemic status.

Data source and study population
All participants of this study were derived from the Dongfeng-Tongji (DFTJ) cohort and the detailed baseline profiles have been described previously. 21 This study followed the Strengthening the Reporting of  Figure S2).
The present study has been approved by the Ethics and Human Sub-

Covariates
The epidemiological questionnaires were used to collect information Restricted cubic spline regression (RCS) linked to logistic regression models were conducted to accommodate nonlinearity in the associations of continuous BP values and CKD incidence.
To evaluate the association of long-term BP change with CKD at different blood glucose categories, we constructed three categories of SBP or DBP changes as <−5, −5 to 5, and ≥5 mmHg. Multivariate logistic regression models were performed to calculate RRs (95% CIs) after adjusted for covariates (BP change of −5 to 5 mmHg as reference).
In the sensitivity analyses, we repeated the multivariable logis-

RESULTS
Distribution of the characteristics by BP categories at baseline is presented in Table 1 (Table S1). After exclusion of participants who were taking antihypertensive or antidiabetic medications at baseline, the significant associations still remained in both 10-year and 5-year follow-up periods (Tables S2-S3). We further divided BP levels into five groups, incident risk of CKD increased with BP levels increased (all p trend < .001) among the whole participants, participants with normoglycemia, prediabetes, and diabetes groups (Tables S4-S5).
In Figure 1, RCS analysis indicated significant nonlinear associations for SBP and incident CKD (p nonlinearity = .040), but the result of DBP Compared with individuals who were with stable SBP (−5 to 5 mmHg), the risk of incident CKD reduced when SBP decreased by more than 5 mmHg, and increased when SBP increased ≥5 mmHg among the whole participants, normoglycemia, and prediabetes participants, no significant association was observed in DM cases (Figure 2, Tables S6-S7). Similar findings were obtained for DBP. After exclusion of participants with antihypertensive or antidiabetic treatment, similar results were obtained except that the association between SBP levels change ≥5 mmHg and incident CKD risk (Tables S8-S9). There was a similar result when we constructed five categories of SBP or DBP changes as <−15, −15 to −5, −5 to 5, 5-15, and ≥15 mmHg (Tables S10-S11).
Subgroup analysis for incident CKD risk was performed according to baseline characteristics including age, sex, BMI, smoking status, drinking status, and baseline eGFR. No significant interaction was found between blood glucose and BP for CKD across groups for various baseline characteristics (all p Interaction > .05) (Tables S14-S16).

DISCUSSION
In the present prospective study, we found that a BP of  did not provide additional benefit for renal outcomes in patients with CKD but without DM. 29 Another study, used averaged BP, found that the relationship between BP ≥ 130/80 mmHg and increased CKD risk still existed in participant with DM. 30 It can be seen that the association of stage 1 hypertension and CKD risk at different glycemic status was still unclear and short of comprehensive and systematic research.
Therefore, we combined with BP categories and glycemic status and found that stage 1 hypertension was significantly associated with a higher CKD risk in total, prediabetes, or DM participants. Our study  34 Secondly, the new-onset CKD was based on the new occurrence of eGFR < 60 ml/min/1.73 m 2 in the follow-ups so participants with acute kidney injury were also included in our study. However, it might have little effect on the results because there was only a small part of participants with acute kidney injury among middle-aged and elderly people, 33 in particular, kidney disease was excluded at baseline. In addition, in the present study, we did not collect urinary albumin data, which would lead to false negative for CKD with an eGFR above 60 ml/min/1.73 m 2 . Thirdly, we measured BP levels only once at baseline and follow-ups, which might misclassify the BP group. More studies were warranted to validate our findings. Fourthly, our analyses were restricted to the middle-aged and elderly Chinese participants and thus may not be generalized to other age groups or ethnicities. Fifthly, we could not explore the impact of time effect on the results because the time record of CKD incidence could not be very accurate by follow-up time of every 5 years. Finally, although a range of potential confounding factors were adjusted in multivariate analysis, we still could not eliminate the unmeasured and residual confounding.

CONCLUSIONS
The present study found that the BP of 130-139/80-89 mmHg combined with prediabetes or DM had an increased risk of incident CKD and rapid eGFR decline in Chinese adults. The long-term changes of BP by more than 5 mmHg among normoglycemia or prediabetes were associated with the risk of incident CKD and rapid eGFR decline.

ACKNOWLEDGMENTS
We thank all study subjects for participating in the present DFTJcohort study and all volunteers for assisting in collecting samples