Data source
This cross-sectional study analyzed the 2007–2018 NHANES data. NHANES is a nationally representative biennial survey in the US implemented by National Center for Health Statistics (NCHS) under the Centers for Disease Control and Prevention. The NHANES utilized a multi-staged cluster sampling design. The details of NHANES including survey design, methodologies, sampling process, and variables included in the datasets were published previously [12, 13]. The datasets were downloaded and then merged using the unique identification numbers.
Outcome variable
The eligible population for this study was men and women aged ≥ 20 years with ≥ 3 blood pressure measurements. Participants who don’t have at least 3 measurements (i.e., 1 or 2 measurements), we obtained BP levels from the available measurements. Blood pressure status was the outcome of interest in this study. Blood pressure was measured by trained research physicians using factory-calibrated Baumanometer® (W.A.Baum Co., Copiague, NY, USA) mercury true gravity wall model sphygmomanometers after the participants remained in seating position for at least 5 min. The appropriate cuff size was used [14].
Prevalence and control of hypertension was defined as per 2017 American College of Cardiology/American Heart Association (ACC/AHA) guideline. An individual was considered as hypertensive if any of the following three conditions were met: SBP ≥ 130 mm Hg, DBP ≥ 80 mm Hg or self-report of taking antihypertensive drugs [15]. Among individuals who reported that they were taking antihypertensive drugs, uncontrolled hypertension was defined as SBP ≥ 130 mm Hg and/or DBP ≥ 80 mm Hg [15].
Study variables
Height/stature was the main explanatory variable of interest. It was measured in meters (m) standing position with a wall-mounted digital stadiometer [16]. During analyses, it was converted into centimeters (cm) and was further divided into quartiles: Q1 (135.3–159.2 cm), Q2 (159.3–166.2 cm), Q3 (166.3–173.6 cm), Q4 (173.7–204.5 cm) [6].
Based on literature review, the following covariates were considered: age, gender, race/ethnicity, family income, education level, cholesterol level, High-density lipoprotein (HDL) level, CKD status, diabetes status, smoker, leisure-time physical activity (LTPA), and survey period.
Age (in years), gender, race/ethnicity, family income, and education level were reported by the participants. Age was categorized into 20–39 years, 40–59 years, and ≥ 60 years. Gender was dichotomized into male and female. Race/ethnicity was categorized into non-Hispanic whites, non-Hispanic blacks, Mexican-Americans, and others. Family income was divided into three equal-sized strata: low, middle, and high. Education level was categorized into below high school, high school, and college graduate or above.
Cholesterol level was categorized into ‘no high cholesterol’ [< 200 mg/ deciliter (mg/dl)], ‘borderline elevated’ (200–239 mg/dl), ‘high cholesterol’ (≥240 mg/dl). Self-reported intake of cholesterol-lowering drugs were also included in the ‘high cholesterol’ group. HDL levels were categorized as low (< 40 mg/dl for men and < 50 mg/dl for women) and normal. CKD was defined as having an albumin-creatinine ratio ≥ 30 mg/gram (mg/g) or a glomerular filtration rate (GFR) < 60 ml/minute (ml/min) per 1.73 m2 (m2) [17]. Diabetes mellitus was defined as fasting plasma glucose ≥126 mg/dl or self-report of taking antidiabetic drugs.
Smoking status was dichotomized into yes and no. Participants were asked about the usual amount of time spent to perform moderate and vigorous aerobic recreational physical activity (PA) in a week. LTPA was calculated by summing the minutes spent to perform vigorous PA multiplied by two with the minutes spent to perform moderate PA. Aerobic LTPA was categorized into no (0 min/week), some (> 0 to < 150 min/week), and high (≥150 min/week) [5]. To increase sample size, the survey years were merged and categorized into 2007–2010, 2011–2014, and 2015–2018. Supplemental Table 1 describes study variables.
Statistical analysis
Descriptive analyses were carried out and the findings were presented in unweighted frequencies and weighted percentages. Then, bivariate analyses were carried out to observe the distribution of the covariates according to hypertension status and according to height quartiles. Also, the uncontrolled hypertension status among those who were taking blood pressure lowering drugs were presented, according to the distribution of the selected covariates. Additionally, the distribution of covariates across the height quartiles among those who were taking blood pressure lowering drugs was presented. In order to find out the association of height with hypertension prevalence and control, unadjusted and multivariable logistic regression analyses were conducted. The variables which yielded a p < 0.05 (which was considered enough to control residual confounding in the multivariable model) in the crude analysis were included in the multivariable logistic regression analyses [18]. Both crude odds ratios (CORs) and adjusted odds ratios (AORs) were reported with 95% confidence intervals (CIs). With separate logistic regression models, we also examined the interaction between age and height. Lastly, we ran three separate multiple linear regression model considering SBP, DBP, and PP as outcomes and height as the main predictor among the individuals with who reported that they were not taking any antihypertensive drugs. As the proportion of missing data was low (< 10%), we used complete case analysis approach to handle missing data. During the analyses, the mobile examination center (MEC) weights of NHANES was used. All the analyses were done using Stata 14.0 (College Station, Texas, USA).
Ethical approval
The National Center for Health Statistics approved the protocol for NHANES. Informed consent was taken from the participants before data collection. This study utilized NHANES data which are publicly available and de-identified, hence was deemed exempted from review by the institutions’ Institutional Review Board (IRB).