The association between fasting plasma glucose and glycated hemoglobin in the prediabetes range and future development of hypertension

Hypertension (HTN) and diabetes mellitus (DM) are very common pathological entities, which very frequently co-exist [1‐3] Both DM and HTN are associated with increased cardiovascular (CV) morbidity and mortality [4]. HTN and DM demonstrate substantial similarities in their clinical manifestations as well as the underlying pathophysiological mechanisms, such as inflammation and oxidative stress, obesity, and insulin resistance [1, 5, 6]. As much as one-third of hypertensive subjects are also insulin resistant and fulfill the diagnostic criteria for the metabolic syndrome [7]. The American Diabetes Association (ADA) defines prediabetes as either impaired fasting glucose (IFG) (i.e., fasting glucose levels of 100–125 mg/dl), impaired glucose tolerance (IGT) (i.e., glucose levels post 2-h oral glucose loading of 140–199 mg/dl), or an elevated hemoglobin A1c (HbA1c) level of 5.7% to 6.4% [8]. Prediabetes is a well-established risk factor for progression to overt DM [8], which is in turn associated with development of HTN [1, 9] and vice versa [10]. However, the role of prediabetes as an independent risk factor for the development of HTN necessitates further elucidation [11‐19]. Furthermore, data are scarce and somewhat conflicting regarding the specific prognosticator role of HbA1c in the prediabetic range and the development of future HTN [17‐20]. Therefore, in this current study, we aimed to evaluate the association between both fasting glucose and HbA1c levels in the prediabetes range and development of HTN among a large cohort of normotensive subjects.

The study population comprised 5016 individuals of whom 69.4% were men (Table 1). Mean follow up was 3.7 ± 2.9 years and mean age was 53.5 ± 9.9 years (Table 1). Among study participants 2619 (52.2%) were normoglycemic, and 2397 (47.8%) had prediabetes (Table 1). Notably, normoglycemic participants were significantly younger, were more likely to be women, had lower BMI, as well as lower SBP and DBP, lower TG levels and higher HDL levels compared with participants with prediabetes (Table 1). Those with prediabetes also had higher rates of statin use and lower rates of physical activity. There were no clinically significant differences between study groups concerning LDL levels and smoking status (Table 1).

In this large retrospective cohort study, we demonstrated an independent significant association between FPG levels in the prediabetes range and future development of HTN. These findings further emphasize the role of insulin resistance in the pathogenesis of HTN. Several plausible pathophysiological mechanisms may underlie the association of impaired glycemic control and development of HTN; Elevated FPG has been previously associated with arterial stiffness [21‐23], most likely via oxidative stress and accumulation of glycation end products, and alterations in activities of vasoactive substances [24‐26]. Furthermore, in patients with DM, the integrity of the vascular wall is more susceptible to damage, specifically in the presence of CV risk factor and these macrovascular changes are also evident in the prediabetic phase [27, 28]. In addition to the well-established effects of insulin resistance and hyperglycemia on the macrocirculatory dysfunction, one must not ignore the role of glycemic dysregulation on the microcirculation. In particular, its role in the development of endothelial dysfunction and alterations in vascular extracellular matrix most likely secondary to inflammatory and pro-fibrotic alterations [28]. IFG and glycated hemoglobin represent different pathophysiological aspects of glycemic control; whereas IFG expresses increased hepatic glucose production in the fasting state, HbA1c encompasses glycemic control in the fasting state as well as post-prandial glucose levels [29]. Therefore, we also assessed the differential risk for HTN development among patients with prediabetes diagnosed by either isolated HbA1C criterion, isolated IFG criterion or both criteria. We found that elevated IFG was independently associated with increased risk of HTN, whereas impaired HbA1c as a sole criterion for prediabetes (i.e., with normal FPG) was not independently associated with increased risk for HTN. While HbA1c is a very useful tool in the diabetes management armamentarium, it has few notable caveats. First, it is a fairly crude measurement of glycemic control, and as such a single HbA1c percentage may represent a wide spectrum of glucose levels [8]. Second, because HbA1c represents the average glycated hemoglobin level, it often fails to properly account for extreme glucose values and for the glucose variability as well [8]. Furthermore, HbA1C levels are susceptible to false elevations and false reductions secondary to various medical comorbidities and substance abuse [8]. Several studies have previously described the association between IFG and HTN development [11‐19]. However, these studies have mainly been conducted among Asian participants [12, 14‐17, 19]. While Lee et al. found no association between IFG, IGT and HTN [12], other studies in Asian subjects found IFG to be an independent risk factor for development of HTN [14‐17, 19]. There are also studies conducted among Caucasians participants; however, these studies yielded conflicting results [11, 13, 30]. Interestingly, a predictive association of elevated fasting serum insulin with incident HTN in European Americans has also been previously reported [31]. As previously mentioned, FPG and HbA1c levels identify different pathological abnormalities in glucose metabolism [32‐34]. We currently found that individuals who had isolated impaired HbA1c had a distinctly different profile. They were significantly older and more likely to be women, had lower BMI, higher HDL cholesterol levels but lower TG levels and lower SBP compared with those who had IFG with normal HbA1c (Table 3). Similar findings were also previously reported in a large study of patients without DM [35]. The prognosticator role of HbA1c in the incidence of cardiovascular morbidity and mortality events is unclear; its’ associated risk is mostly attributed to confounding risk factors [36‐38]. In a recent Japanese population study, Kuwabara et al. did not find HbA1c to be an independent risk factor for developing HTN, whereas each 10 mg/dl increase in FPG was associated with a 42.2% increased risk for developing HTN over 5 years [17]. A similar pattern of a differential association of IFG rather than elevated HbA1c and incident HTN was also previously reported by Heianza et al. [19]. In addition, Britton et al. have shown that in women without diabetes HbA1c levels were not significantly associated with increased risk of HTN after adjustment for BMI [20]. These findings concur with the results of our current study showing a 4.6% increased risk for HTN for every 1 mg/dl increment of FPG and a twofold increased risk for HTN independently associated with IFG. Notably, in our study, the risk for HTN increased significantly by 19% when both measures of glycemic control, i.e., FPG and HbA1C were both impaired. Bower et al. have demonstrated that higher HbA1c in the prediabetic range was independently associated with incident self-reported HTN [18].

Our study is observational; hence we can only postulate regarding the underlying mechanisms responsible for the differential association of IFG and HbA1c and incident HTN. First, as HbA1c levels are influenced by the glycation of proteins in the body due to hyperglycemia, it has been suggested that glucose concentrations may more directly reflect the pathogenesis of the HTN development compared to HbA1c which is an indirect measure of dysglycemia [19]. Furthermore, a discordance between glucose level in the prediabetic or diabetic range and HbA1c values has been previously described by others [35, 39, 40], and thus may provide a possible explanation for our findings.

Our study has several strengths and limitations. First, our database relies on annual clinical and laboratory evaluation, which enables us to perform a time-based survival analysis for HTN development, which is expressed as Hazard Ratio, rather than a cumulative 5-years incidence rate as reported by previous studies [11‐17, 19]. We hypothesize that our annual data gathering and analysis permits superior delineation of HTN development over time compared to a cumulative 5-years incidence rate. Unlike previous studies [11‐17, 19], our definition of HTN was not based on a single BP measurement and in addition, we excluded the presence of white coat HTN and other etiologies that may cause falsely elevated BP measurement using a strict definition for new-onset HTN. To further emphasize and verify our results we performed a subgroup analysis with a stricter HTN free exclusion. Our findings support a higher hazard risk for HTN in our subgroup analysis as compared to the all-cohort analysis (Additional file 1: Tables S1, S2). There are several possible explanations for this finding. There could be a dominance of patients with white coat HTN, a paucity (n = 25) of extremely high BP levels, i.e., above 160/100 mmHg, or the relatively small number of participants and event rates in this group. Therefore, the subgroup analysis highlights our finding of prediabetes and IFG in particular as an independent and major risk factor for HTN. Our study has several limitations. First, this is a retrospective observational study. Second, our cohort comprises of middle-aged men and women with a higher-than-average socioeconomic status, higher rates of physically activity and lower BMI levels, all of which may limit the generalization capacity of our findings and attributes to the relatively low events rates of newly HTN in this cohort. In addition, our cohort had a relatively lower rate of women participants, which may preclude a significant sex effect on the development of HTN. Moreover, while our study is notable for using both PFG and HbA1C as indices of glycemic control, we did not have specific data regarding IGT.

In conclusion, our study demonstrates that FPG in the prediabetes range, albeit not glycated hemoglobin, is independently and significantly associated with future development of HTN. Therefore, our findings further highlight the pivotal predictive role of IFG for HTN development as opposed to the limited independent role of abnormal HbA1c levels.