Association of Blood Pressure with Fasting Blood Glucose Levels in Northeast China: A Cross-Sectional Study

Hypertension and diabetes mellitus (DM) have become major public health issues, and previous studies have shown that there is an association between hypertension and DM. However, there is a lack of detailed information about this association. This study aimed to explore how different blood pressure (BP) levels were associated with fasting blood glucose (FBG) levels. A cross-sectional survey with adults aged 18 to 79 years was conducted in Jilin Province, China in 2012. Lambda-mu-sigma (LMS) was used to preliminarily explore the associations of BP with FBG. Quantile regression (QR) was performed to identify the specific associations by adjusting for confounding factors. The distributions of systolic blood pressure (SBP) (χ2 = 710.76, P < 0.001) and diastolic blood pressure (DBP) (χ2 = 460.20, P < 0.001) were different according to gender. LMS showed that the associations of BP with FBG became stronger when the FBG levels were close to 5.6 mmol/L. QR showed that FBG was positively associated with SBP (P30 to P90) and DBP (P20 to P90) in males. In females, FBG was positively associated with SBP from only P85 to P90. In summary, FBG was positively associated with BP in a gender-dependent manner.

Results of LMS analysis. The smoothed percentile curves of SBP and DBP in males and females are shown separately in Figs 2 and 3. The shapes of the curves were different according to gender, and all the percentiles of BP in males were substantially and consistently higher compared with those in females. Moreover, the associations of BP with FBG became stronger when the FBG levels were close to 5.6 mmol/L.

Variables
Male (n = 5571) Female (n = 6307) Z/χ 2 P-value  Results of the QR model in males. The coefficients of FBG in different quantiles of SBP and DBP are shown separately in Tables 3 and 4. Table 3 shows that there were positive associations of SBP with FBG, and the coefficients increased from P 10 to P 90 (model 1). In addition, FBG was positively associated with SBP from P 30 to P 90 after adjusting for age, BMI, WC, smoking status, alcohol consumption and dyslipidaemia (model 2). Table 4 shows that FBG was positively associated with DBP from P 20 to P 90 (model 1). In addition, FBG was still positively associated with DBP from P 20 to P 90 after adjusting for age, BMI, WC, smoking status, alcohol consumption and dyslipidaemia (model 2).
Results of the QR model in females. The coefficients of FBG in different quantiles of SBP and DBP are shown separately in Tables 5 and 6. As shown in Table 5, FBG was positively associated with SBP from P 10 to P 90 (model 1), whereas FBG had positive associations with SBP from only P 85 to P 90 after adjusting for age, BMI, WC, smoking status, alcohol consumption and dyslipidaemia (model 2). As shown in Table 6, there were positive associations of DBP with FBG from P 10 to P 30 and from P 75 to P 90 (model 1). Yet the associations of DBP with FBG were not statistically significant after adjusting for age, BMI, WC, smoking status, alcohol consumption and dyslipidaemia (model 2).

Discussion
The findings of this study could be summarized as follows. First, there were positive associations of BP with FBG, and the associations were different according to gender. FBG was associated with BP in almost all quantiles in males, while in females, only SBP in high quantiles had positive associations with FBG. Second, the associations of BP with FBG became stronger when the FBG levels were close to 5.6 mmol/L. In our study, we found that FBG was positively associated with BP, which was consistent with the results of similar previous studies 13, 14 . The possible mechanism is that as the FBG level increases, hyperglycaemia with  insulin resistance, overweight and metabolic disorders could alter the rennin-angiotensin system (RAS), then leading to an effect on BP 15,16 .
In addition, we found that the associations of BP with FBG were different in males and females; FBG was positively associated with both SBP and DBP in almost all quantiles in males, whereas FBG had positive associations with only SBP in the high quantiles in females. The finding implied that the positive associations of BP with FBG were more notable in males than in females. There are several reasons for these associations. First, the risk of hypertension in females is lower than that in males 17 . Second, evidence has shown that the lower endogenous oestrogen levels found in males might be related to higher insulin resistance when compared with females 18 , which means that the same FBG levels in males and females may lead to different physiological effects due to different levels of insulin resistance. Thus, the finding might suggest that control of one's FBG level should be more emphasized and strengthened among males than females, except females with high SBP.
The cut-off value for the definition of impaired fasting glycaemia (IFG) is still controversial at present; the American Diabetes Association (ADA) Expert Committee reduced the cut-off value of IFG to 5.6 mmol/L in 2003, while the World Health Organization (WHO) retained the previous cut-off value of IFG (6.1 mmol/L) 19 . Further, IFG has been shown as an independent risk factor for hypertension 20,21 ; however, whether the FBG level begin to become a potential risk factor for hypertension when it is close to 5.6 mmol/L is uncertain. Fortunately, our study showed that the positive associations of BP with FBG became stronger when FBG levels were close to 5.6 mmol/L, which implied that FBG level might begin to become a potential risk factor for hypertension when it was close to 5.6 mmol/L. A possible explanation is that the body may develop hyperinsulinaemia before FBG level reaches 6.1 mmol/L, and the presence of hyperinsulinaemia could directly contribute to an elevation in BP by increasing renal sodium retention 22 . Therefore, having awareness of controlling FBG levels when they    approach 5.6 mmol/L might have potential benefits, such as early control of BP. Furthermore, having awareness of controlling FBG levels might also assist in preventing the development of co-occurrence of hypertension and DM to some extent. Some limitations should be noted in this study. First, participants in the study were from Jilin Province in Northeast China, and the results may not be generalizable to other areas. Second, some of the subjects were excluded because they did not suit the purpose of the study and/or they were with incomplete data on BP, thus, selection biases might be present. Third, other confounding factors (such as genes) may have impacted the results. Finally, the data were obtained from a cross-sectional survey; therefore, further studies are needed to explore the associations in a longitudinal setting.

Conclusions
There were positive associations of BP with FBG, and these associations were different according to gender. FBG was associated with BP in almost all quantiles in males, yet in females, FBG had positive associations with only SBP in the high quantiles. Moreover, the associations of BP with FBG became stronger when the FBG levels were close to 5.6 mmol/L.

Methods
Study population. Data was derived from a cross-sectional study of chronic disease conducted by the Jilin University School of Public Health and the Jilin Department of Health in Jilin Province of China in 2012. In this study, a total of 23,050 subjects who had lived in Jilin Province for more than 6 months and were 18-79 years old were selected through multistage stratified random cluster sampling 23 . First, 1,615 subjects whose questionnaires were invalid were excluded. Second, 9,421 subjects were excluded because they did not suit the purpose of the study (4,993 subjects were not tested for FBG and 4,428 subjects controlled their BP and/or FBG level with medicine). Third, 136 subjects were excluded due to incomplete data on systolic blood pressure (SBP) (75 subjects) and diastolic blood pressure (DBP) (61 subjects). Finally, 11,878 subjects were included in this study (Fig. 4). All participants provided written informed consent, and the study was approved by the Institutional Review Board of the Jilin University School of Public Health. In addition, all analyses were performed in accordance with the relevant guidelines and regulations.
Data Collection and Measurement. The data in this study were collected by investigators who had been uniformly trained. The information included demographics (gender, age), health-related behaviours (smoking status, alcohol consumption), anthropometric measurements (SBP, DBP and waist circumference (WC), etc.) and biochemical tests (FBG, total cholesterol (TC) and triglycerides (TG), etc.). BP was measured using a mercury sphygmomanometer by trained professionals, and the subjects were required to have rested for at least 5 min before BP was measured. Two readings of SBP and DBP were recorded, and the average values were used for the data analysis. If the first two measurements differed by more than 5 mmHg, additional readings were taken. FBG levels were measured by the Bai Ankang fingertip blood glucose monitor (Bayer, Leverkusen, Germany).
Serum lipid levels (TG, TC, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C)) were measured by the MODULE P800 biochemical analysis machine (Roche Co., Ltd., Shanghai, China) in the morning after participants had fasted for 10 or more hours overnight. The participants' height and weight were measured according to a standardized protocol with the participants wearing clothing but no shoes, and WC was measured at the midpoint between the lowest rib margin and the iliac crest with the participants standing. Body mass index (BMI) was calculated using the following formula: BMI = Weight (kg)/ Height (m) 2 . Assessment Criteria. Hypertension was defined as resting SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg; high-normal BP was defined as SBP 120-139 mmHg or DBP 80-89 mmHg; and normal BP was defined as SBP < 120 mmHg and DBP < 80 mmHg, with no antihypertensive treatment 24 . Smoking status was categorized into never smoker (had never smoked cigarettes or had smoked fewer than 100 cigarettes in one's lifetime), former smoker (had smoked at least 100 cigarettes in one's lifetime but was not currently smoking), and current smoker (had smoked at least 100 cigarettes in one's lifetime and was still smoking) 25 . Alcohol consumption was defined as consuming any type of purchased or homemade alcohol-containing beverages on average more than once per week 26 . Regarding serum lipid level, any of the following criteria was defined as dyslipidaemia: high TC: TC ≥ 6.22 mmol/L; high TG: TG ≥ 2.26 mmol/L; low HDL-C < 1.04 mmol/L; and high LDL-C ≥ 4.14 mmol/L 27 .
Statistical Analysis. Continuous variables were presented as medians (inter-quartile range) because they did not have normal distributions. Wilcoxon rank sum tests were used to make comparisons according to gender. Categorical variables were presented as counts or percentages and compared with the Rao-Scott Chi-square test. Sex-specific percentile curves of BP with FBG were constructed by LMS method developed by Cole and Green 28 . The 3rd, 5th, 10th, 25th, 50th, 75th, 90th, 95th and 97th 29 percentile BP curves were constructed to explore the associations of BP with FBG and to determine the general trend of those associations. QR, which is very flexible, especially for data with a heterogeneous conditional distribution, was performed to establish 2 models to explore how different BP levels were associated with FBG levels according to gender (model 2 was adjusted for age, BMI, WC, smoking status, alcohol consumption and dyslipidaemia). Meanwhile, QR was applied to build a series of regression equations in all quantiles of BP, and thus, the extreme data, such as the high BP quantiles, could be analysed 30 . R version 3.3.3 (University of Auckland, Oakland, New Zealand) was used to perform the statistical analyses. Statistical significance was set at P-value < 0.05. Data Availability. The survey was implemented by School of Public Health, Jilin University and Jilin Center for Disease Control and Prevention in Jilin Province in 2012. According to the relevant regulations, we were sorry that the data can't be shared.