Postprandial Plasma Glucose Measured from Blood Taken between 4 and 7.9 H Is Positively Associated with Mortality from Hypertension and Cardiovascular Disease

1. Introduction

Cardiovascular disease (CVD) is the leading cause of death globally, responsible for 17.9 million deaths each year [1]. The global expenditure on CVD ranges between 7.6% and 21.0% of national health expenditures [2]. In the US, CVD costs approximately $320 billion per year [3]. Therefore, there is an urgent medical need to identify new risk factors and effective prevention strategies for CVD mortality.

Diabetes affects 8.5% of adults according to the World Health Organization [4]. It is well known that patients with diabetes have an increased risk of CVD mortality [5,6]. However, the underlying mechanism is not well understood. Postprandial plasma glucose (PPG) is believed to play an important role in diabetes-associated complications [7,8,9]. Therefore, it is of value to investigate the association of PPG with CVD mortality.

To the best of my knowledge, only one study investigated PPG and CVD mortality [10]. That study found that PPG measured from blood taken between 3 and 7.9 h (PPG_3-7.9h) was positively associated with CVD mortality [10]. However, the PPG at 3-3.9 h did not return to the baseline level and PPG_3-3.9h was higher than PPG_4-7.9h [10]. A recent study showed that PPG returned to baseline four hours after a meal regardless of meal type (normal or high carbohydrate) and mealtime (breakfast, lunch and dinner) [11]. Therefore, the use of PPG_3-7.9h is inferior to PPG_4-7.9h and the association between PPG_4-7.9h and CVD mortality needs to be investigated.

In addition, it has been shown that patients with diabetes have an increased risk of hypertension incidence [12]. However, whether PPG_4-7.9h is associated with hypertension mortality is unknown, and so is whether PPG_4-7.9h is associated with diabetes mortality.

This study aimed to investigate these unaddressed questions, i.e., whether PPG_4-7.9h is associated with mortality from hypertension, diabetes, and CVD, using a representative cohort of US adults who attended the third National Health and Nutrition Examination Survey (NHANES III) from 1988 to 1994.

2. Materials and Methods

2.1. Participants

A total of 4,926 adults aged ≥ 20 years who attended the NHANES III had recorded postprandial plasma glucose data measured from blood taken between 4 and 7.9 h. Those who did not have a follow-up time (N=3) or hemoglobin A_1c (HbA_1c, N=27) were excluded. Therefore, the remaining 4,896 participants were included in this cohort study, including 343 participants with a prior diagnosis of myocardial infarction or stroke (Figure 1).

2.2. Measurement of Plasma Glucose

Plasma glucose was measured using the hexokinase-mediated reaction method as previously described [13].

2.3. Mortality

Data on mortality from CVD (I00-I09, I11, I13, I20-I51, I60-I69), diabetes (E10-E14), and hypertension were directly retrieved from NHANES-linked mortality files [14]. To evaluate mortality status and the cause of death, the National Center for Health Statistics conducted probabilistic matching [15] to link the NHANES data with death certificate records from the National Death Index (NDI) records. The NHANES-linked mortality files used the Underlying Cause of Death 113 (UCOD_113) code to recode all deaths according to the International Classification of Diseases, 9th Revision (ICD-9) or the International Classification of Diseases, 10th Revision (ICD–10) for the underlying cause of death [14]. Follow-up time was the duration from the time when the participant was examined at the Mobile Examination Center until death, or until the end of follow-up (December 31, 2019), whichever occurred first.

2.4. Covariates

Confounding factors included age (continuous), sex, ethnicity, body mass index (continuous), education, poverty-income ratio, survey periods, physical activity, alcohol consumption, smoking status, systolic blood pressure (continuous), total cholesterol (continuous), high-density lipoprotein (HDL) cholesterol (continuous), HbA_1c (continuous), family history of diabetes, and fasting time (continuous), as described previously [14,16].

2.5. Statistical Analyses

Data were presented as mean and standard deviation for normally distributed continuous variables, median and interquartile range for not normally distributed continuous variables, or number and percentage for categorical variables to describe the baseline characteristics of the cohort [17]. Differences in continuous variables between two groups were analyzed using Student’s t-test (normally distributed), or Mann Whitney U test (not normally distributed). Differences among categorical variables were analyzed using Pearson's chi-square test [18]. The difference in hourly PPG_4-7.9h was analyzed using Kruskal Wallis one-way ANOVA.

Out of 4,896 participants, a total of 115 (2.3%) had missing data including body mass index (N=14), systolic blood pressure (N=11), total cholesterol (N=53), or HDL cholesterol (N=93). The missing data were imputed via multiple imputation by chained equations, with 20 imputed data sets being created [19]. Little’s test showed that the missing data were not missing completely at random (P<0.001). In all the regression analyses, body mass index, systolic blood pressure, total cholesterol, HDL cholesterol, and HbA_1c were natural log transformed to improve data distribution.

Cox proportional hazards models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) of PPG_4-7.9h for mortality from hypertension, diabetes, and CVD [20]. PPG_4-7.9h was treated as a continuous variable (natural log-transformed) or categorical variable. Further analyses were conducted in the sub-cohort of participants without a prior diagnosis of myocardial infarction or stroke.

Sensitivity analyses were conducted when the imputed data were not used, i.e., by excluding those 115 (2.3%) participants with missing data from the analysis, or when those with a follow-up time of <1 year (N=45) or those who were prescribed with insulin or other anti-diabetic medications (N=250) were excluded.

The null hypothesis was rejected for two-sided values of P<0.05. All analyses were performed using SPSS version 27.0 (IBM SPSS Statistics for Windows, Armonk, NY, IBM Corporation).

3. Results

3.1. General Characteristics

This cohort included 4,896 adult participants with a mean (standard deviation, SD) age of 49 (19) years. Those who had higher PPG_4-7.9h were older and had higher body mass index, systolic blood pressure, and total cholesterol (Table 1). In addition, they were less physically active, had lower HDL cholesterol, and received less education and income (Table 1). Hourly PPG_4-7.9h was similar (Figure 2).

3.2. Association of PPG_4-7.9h with Mortality

This cohort was followed up for 106,300 person-years with a mean follow-up of 21.7 years. During the follow-up, 337 hypertension deaths, 70 diabetes deaths, and 835 CVD deaths were recorded.

A 1-natural-log-unit increase in PPG_4-7.9h was associated with a higher multivariate-adjusted risk of mortality from hypertension (HR, 3.50; 95% CI, 2.34-5.24), diabetes (HR, 11.7; 95% CI, 6.85-20.0), and CVD (HR, 2.76; 95% CI, 2.08-3.68), after adjustment for all the tested confounders except HbA_1c (Model 1; Figure 3). After further adjustment for HbA_1c (Model 2, Figure 3), PPG_4-7.9h remained positively associated with mortality from hypertension (HR, 2.15; 95% CI, 1.13-4.08) and CVD (HR, 1.62; 95% CI, 1.05-2.51). Similar results were obtained when PPG_4-7.9h was treated as a dichotomous variable using the top decile as the cutoff (Figure 4). The use of the top decile as the cutoff is based on the estimate from the World Health Organization that 8.5% of adults have diabetes [4]. Subgroup analyses showed that similar results were obtained in those participants without a prior diagnosis of myocardial infarction or stroke (Figure 5).

Sensitivity analyses showed that PPG_4-7.9h remained positively associated with mortality from hypertension and CVD when imputed data were not used, i.e., by excluding those 115 participants with missing data (Figure 6), or when those with a follow-up time of <1 year were excluded (Figure 7), or when those who were prescribed with anti-diabetic medications were excluded (Figure 8).

4. Discussion

Using a general cohort of US adults, this study, for the first time, demonstrated that PPG_4-7.9h was positively associated with mortality from hypertension and CVD, independent of HbA_1c. In addition, these positive associations remained in the sub-cohort of participants who did not have a prior diagnosis of myocardial infarction or stroke.

This study found that PPG_4-7.9h was positively associated with hypertension mortality. However, the underlying mechanism is unknown. It is well known that diabetes and hypertension often co-exist in many individuals [21], and these two conditions share some risk factors such as obesity [22,23] and physical inactivity [24,25]. It has been shown that baseline fasting plasma glucose [26], fasting plasma glucose change trajectory [27], and diabetes [12] are positively associated with risks of hypertension incidence [12], suggesting that high blood glucose may disturb blood pressure homeostasis. Consistently, the current study showed that PPG_4-7.9h was positively associated with hypertension mortality, independent of well-known confounders including body mass index, physical activity, total cholesterol, and HDL cholesterol, supporting a causal role of high plasma glucose in worsening hypertension outcomes. It has been reported that high plasma glucose may lead to oxidative stress and endothelial dysfunction [28,29]. Whether increased oxidative stress and endothelial dysfunction play a role in mediating the positive association between PPG_4-7.9h and hypertension mortality needs to the investigated in the future. So does whether lowering PPG_4-7.9h is effective in improving blood pressure control and hypertension mortality.

The association of diabetes with CVD incidence and mortality are well documented. Diabetes is an independent risk factor for CVD [30]. In addition, sodium–glucose cotransporter 2 (SGLT2) inhibitors, a class of anti-diabetic medication, decrease CVD events and mortality [31,32,33]. The mechanism underlying the association of diabetes with CVD events and mortality is not well understood.

A few studies have investigated the association of PPG with cardiovascular events. PPG at 1 or 2 h after breakfast [34,35] or 2 h after lunch [36,37] were reported to be positively associated with CVD events. However, those studies did not investigate CVD mortality. In addition, measuring glucose at 1 or 2 h after a meal may not be ideal, as variation in diet could change PPG by more than 20 mg/dL [11] and variation in blood collection time (± 0.5 h in practice [38]) could introduce bias as PPG is time-sensitive around 1 to 2 h [11].

Only one study investigated PPG and CVD mortality, which found that PPG measured from blood taken between 3 and 7.9 h was positively associated with CVD mortality [10]. However, the use of PPG_3-7.9h is inferior to PPG_4-7.9h, as PPG at 3-3.9 h did not return to the baseline level and PPG_3-3.9h was higher than PPG_4-7.9h [10]. In addition, PPG returned to baseline four hours after a meal regardless of meal type and mealtime [11]. Moreover, the current study confirmed that hourly PPG_4-7.9h was similar across the duration from 4 to 7.9 h. Therefore, it is necessary to investigate the association between PPG_4-7.9h and CVD mortality.

Some studies have investigated the association between fasting plasma glucose and CVD mortality and the results are inconsistent: some show a positive association [39,40] whereas others show no association [41,42]. The reason for this inconsistency is unknown. This may be due to poor reproducibility of fasting plasma glucose [43]. For instance, only 75% of adults were classified into the same diabetes category (normal, prediabetes, or diabetes) based on two consecutive measures of fasting plasma glucose which were conducted 6 weeks apart [43].

The current study showed that PPG_4-7.9h was positively and independently associated with CVD mortality, and such a positive association remained in those without a prior diagnosis of myocardial infarction or stroke. Given its stability and reproducibility_, PPG_4-7.9h may be a better predictor of CVD mortality than fasting plasma glucose and PPG_3-7.9h. Whether lowering PPG_4-7.9h is a primary prevention strategy to decrease CVD mortality needs to be investigated in the future.

This study found that PPG_4-7.9h was positively associated with diabetes mortality and such an association disappeared after future adjustment for HbA_1c, suggesting that HbA_1c could explain the association between PPG_4-7.9h and diabetes mortality.

Some guidelines have started to recommend non-fasting lipids (triglycerides and various forms of cholesterol) as the standard for cardiovascular risk assessment [44,45]. Consistently, the current study suggests that non-fasting plasma glucose (PPG_4-7.9h) may be used for cardiovascular risk assessment. The non-fasting plasma glucose test is more convenient than a fasting glucose test. More research is needed to establish whether non-fasting plasma glucose could be eventually used in the clinic for CVD risk assessment. For example, studies replicating the results of the current study using different populations from different countries are needed.

Strengths and limitations One strength of this study is its analysis of PPG after meals of free choice in a large representative cohort of US adults. Another strength is its prospective study design with a long follow-up (mean, 21.7 years). A third strength is its adjustment for a large number of confounders. This study also has several limitations. First, mortality outcomes were ascertained by linkage to the National Death Index (NDI) records with a probabilistic match, which could result in misclassification [46]. However, this matching method has been shown to be highly accurate (accuracy, 98.5%) [47]. Second, PG was only measured at one timepoint for each participant, which may lead to bias. Nevertheless, in epidemiological analysis, this bias tends to result in an underestimate rather than an overestimate of risk due to regression dilution [48].

5. Conclusions

This study found that PPG_4-7.9h is positively associated with mortality from hypertension and CVD and such positive associations remain in those without a prior diagnosis of myocardial infarction or stroke. Therefore, lowering PPG_4-7.9h may be a primary prevention strategy to decrease CVD mortality.