THE PRESENCE OF HYPERLIPIDEMIA PREDICTS THE FUTURE RISK OF HYPERTENSION
Hypertension is commonly associated with other cardiovascular risk factors such as obesity, diabetes, and dyslipidemia. Observational studies that incorporate treating several of these risk factors simultaneously have found improvement in cardiovascular outcomes in individuals with established hypertension. The presence of these risk factors commonly results in endothelial dysfunction. Whether their presence results in the future development of hypertension has been unclear. To evaluate whether the presence of hyperlipidemia contributes to the future development of hypertension, investigators from the Physicians' Health Study prospectively examined healthy male physicians to determine whether total cholesterol (TC), high‐density lipoprotein cholesterol (HDL‐C), and non—high‐density lipoprotein cholesterol (non‐HDL‐C) increase the risk of developing future hypertension.
The National Cancer Institute and the National Heart, Lung, and Blood Institute study enrolled 22,071 male physicians between the ages of 40 and 84 years (average age, 48 years) in 1982. Of these, 3110 subjects had measurements of TC and HDL‐C and were initially free from hypertension (systolic blood pressure [BP] <140 mm Hg or diastolic BP <90 mm Hg); a history of angina, coronary revascularization, or transient ischemic attack; and did not have a history of myocardial infarction or stroke or any current or past treatment for hyperlipidemia. Subjects completed questionnaires at 6 months and annually thereafter. The primary outcome was time to development of self‐reported hypertension, defined as treatment for hypertension or a systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg. Follow‐up data were available on 99% of subjects over the duration of the trial.
For the purposes of analysis, each lipid parameter (TC, HDL‐C, non‐HDL‐C, and TC/HDL‐C ratio) was categorized in two ways: 1) by quintiles; and 2) by clinical cutpoints as suggested by the National Cholesterol Education Project (NCEP). A multivariate model adjusted for baseline risk factors included age, body mass index, diabetes, vigorous exercise, smoking status, alcohol use, parental history of myocardial infarction, and baseline systolic BP (<120 vs. 120–139 mm Hg) and diastolic BP (<80 vs. 80–89 mm Hg) according to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) definition of prehypertension.
During a follow‐up of 18.6 years (mean, 14.1 years), 1019 men (38%) developed hypertension. In models adjusted for lifestyle and clinical risk factors, men in the highest quintile for TC (>243 mg/dL) were 23% more likely to develop hypertension (p=0.0067) compared with those men in the lowest quintile (≤180 mg/dL). Men in the highest quintile for non—HDL‐C (>201 mg/dL) were 39% more likely to develop hypertension (p=0.0001) than those in the lowest quintile (≤135 mg/dL); men with TC/HDL‐C in the highest quintile (>6.97) were 54% more likely to develop hypertension (p<0.0001) compared with those in the lowest quintile (≤3.76), and men in the highest quintile for HDL‐C (>53 mg/dL) had a 32% decreased chance of developing hypertension (p<0.0001) compared with those in the lowest quintile (≤31 mg/dL). When lipid parameters were divided by NCEP clinical cutpoints, the relative risks were comparable to those seen during the analysis of the quintiles.
This prospective study demonstrates that after the exclusion of multiple potential confounders as well as the presence of diabetes and obesity, higher levels of plasma TC, non‐HDL‐C, and TC/HDL‐C are independently associated with an increased risk and higher levels of HDL‐C are associated with a decreased risk of incident hypertension in apparently healthy men. Plasma lipids may be useful in identifying men at risk for developing hypertension—Halperin RO, Sesso HD, Ma J, et al. Dyslipidemia and the risk of incident hypertension in men. Hypertension. 2006;47:45–50.
COMMENT
Hyperlipidemia and hypertension, when present together, increase the risk of developing atherosclerotic cardiovascular disease. It remains unclear, however, whether hyperlipidemia actually contributes to the development of hypertension or whether they simply coexist as a common risk factor sharing similar pathophysiologic mechanisms. Although the potential biologic mechanisms relating dyslipidemia to the development of hypertension remain poorly understood, several theories exist. Atherogenic lipid abnormalities cause endothelial dysfunction, which, through impaired NO production and alterations in endothelin production and endothelin‐receptor expression, impairs vasoregulation. The inability of the endothelium to physiologically constrict and dilate may lead to an increase in resting BP. Other mechanisms noted in people with elevations of serum lipids and BP are serum angiotensin II overexpression and sympathetic nervous system overactivity. Of course, other possible reasons for this association, including a common genetic predisposition unrelated to endothelial dysfunction or lifestyle issues that were not fully corrected for in this multivariate analysis, cannot be excluded. Since the pathophysiologic causes of both hyperlipidemia and hypertension alone are poorly understood, it is not surprising that the cause of their apparent association is difficult to elucidate.
The present study, with its long duration and high rates of follow‐up, suggests that hypertension represents an early manifestation of the atherosclerotic process. The self‐reporting of hypertension has been validated previously, and the fact that a single baseline lipid level was used (which could have been reduced over time by dietary and pharmacologic intervention) would only underestimate the true association present. The fact that all of the participants were physicians, the majority of whom were white (actual percentage not given), who had a body mass index <25 kg/m2, and rarely smoked lessens the study's generalizability to minorities and populations of lower socioeconomic status.
The authors conclude that their study identifies an independent relationship between increased lipid levels and incident hypertension, indicating that hypertension is an early manifestation of the atherosclerotic process due to hyperlipidemia. Although more studies are needed to explain this relationship, the present data suggest that perhaps treating hyperlipidemia may decrease the future risk of hypertension. Prospective studies with both lifestyle modification and pharmacologic therapy will be necessary to better understand this relationship. While other, unmeasured factors may be involved, the prevalence of hypertension in the future may decrease as we more effectively identify and treat patients with hyperlipidemia before they become hypertensive.
ROUTINE OFFICE BLOOD PRESSURE MEASUREMENTS ARE OFTEN HIGHER THAN STANDARDIZED OR AMBULATORY READINGS
Although out‐of‐office blood pressure (BP) measurement (home and 24‐hour ambulatory BP monitoring) is being used more commonly, the diagnosis and management of hypertension continues to be based on accurate BP readings determined in the office. National guidelines continue to advocate a standardized technique for measuring BP. Although a few studies have previously suggested that there is a wide variation in the accuracy of equipment and standardization of measurement technique, there are surprisingly few data on the accuracy of BP measurement in usual clinical practice as compared with that measured by trained nurses using standardized techniques (similar to clinical trials) or 24‐hour ambulatory devices. In this Canadian study designed to examine the accuracy of office BP readings, the usual determination of office BP measurement was compared with a standardized measurement technique by a trained research nurse and by readings taken with a 24‐hour ambulatory BP monitoring device.
To initiate the study, family physicians were randomly recruited from the Calgary, Alberta telephone book or selected from those who routinely referred patients to the Calgary Health Region Hypertension and Lipid Center. These physicians were asked to refer adult patients with hypertension (systolic BP >140 mm Hg and diastolic BP >90 mm Hg) for whom they would usually request ambulatory BP monitoring. Each subject underwent 24‐hour ambulatory BP monitoring using a SpaceLabs model 90207 device (SpaceLabs, Issaquah, WA). After removal of the device, BP was measured by a research nurse trained to follow the Canadian Hypertension Education Program protocol for BP measurement (including the use of a mercury sphygmomanometer, an average of two measurements, a cuff of appropriate size, and a 5‐minute rest in a quiet room before measurement). Within 3 hours, patients returned to their physicians' offices for their usual office readings recorded by either their usual office nurse or the physician. Office hypertension was defined as BP ≥140/90 mm Hg; for ambulatory monitoring hypertension was defined as an average daytime ambulatory BP ≥135/85 mm Hg or average 24‐hour BP ≥130/80 mm Hg.
Fifteen family practitioners referred subjects to the study. A total of 121 subjects participated. Fourteen subjects were excluded (for absence of appropriate usual clinic BP readings or poor quality ambulatory BP), leaving 107 subjects for analysis (57 women and 50 men; average age, 57 years). In all, 47% percent were on antihypertensive medication. The average usual clinic BP values were a mean (95% confidence interval) 10.8 (8.0–13.6) mm Hg/4.9 (2.9–6.9) mm Hg higher than readings obtained by the research nurse. The average usual clinic Bps were also 7.7 (5.1–10.3) mm Hg/5.1 (3.0–7.1) mm Hg higher than the average daytime ambulatory BP and 12.1 (9.6–14.6) mm Hg/8.9 (6.9–10.9) mm Hg higher than the 24‐hour ambulatory BP. The correlation between usual clinic readings and those obtained by the research nurse was 0.58 for systolic BP and 0.44 for diastolic BP. BP readings obtained by the research nurse and average daytime ambulatory BP were closer, with a correlation of 0.74 for systolic BP and 0.65 for diastolic BP. Forty‐four percent of the patients assessed as hypertensive by physicians had normal Bps when measured by the trained study nurse. When clinic nurses found BP in the hypertensive range, 56% of subjects had normotensive readings when assessed by the trained study nurse.
BP is frequently overestimated by usual office readings. Many patients classified as being hypertensive by usual office readings actually are not hypertensive when standardized or ambulatory measurement techniques are employed. Significant improvement in the assessment of clinic BP is required for the diagnosis and optimal management of hypertension. Consideration needs to be given to the development of alternative methods of assessing BP in clinical practice.—Campbell N, Culleton B, McKay D. Misclassification of blood pressure by usual measurement in ambulatory physician practices. Am J Hypertens. 2005;18:1522–1527.
COMMENT
Hypertension is the leading cardiovascular risk factor for death in women and the second leading risk factor in men. The residual lifetime risk for developing hypertension is 90%. Hypertension is the most common reason an adult visits a physician. Accordingly, every decision we make regarding the diagnosis of hypertension must be based on an accurate measurement. The present study suggests a large potential for misclassification of patients' BP when assessed by usual clinic measurement compared with either a trained research nurse or 24‐hour ambulatory recordings. The overclassification and the perception of poorly controlled hypertension could lead to overtreatment of patients with antihypertensive medications. These routine measurements may fail to represent how well BP actually is controlled in the clinic setting.
Interpretation of these results, however, is subject to a number of limitations. In addition to the small number of patients, only a small number of practices (15) were included. As such, generalizability to other populations, especially those outside Canada, is open to question. Additionally, this study included only patients who were being referred for ambulatory BP monitoring and specifically excluded patients with normal BP, likely resulting in a large number of patients suspected of having white coat hypertension. As the largest differences in BP were between usual clinic office readings and office readings performed by a trained study nurse, rather than between usual clinic BP and ambulatory BP monitoring results, the inaccuracies found may not be due to a white coat effect. Furthermore, the sequence of BP readings was not randomized (casual office BP was always performed last), which could have biased the results; if anything, this would have led to lower casual office readings and not the higher BP measurements that were found. The similarity of BP readings obtained by the trained research nurse and the average daytime ambulatory BP in this study is also noteworthy. Previous research has suggested that due to the white coat effect, daytime ambulatory BP is usually lower than office readings. We typically use a lower cutoff for making the diagnosis of hypertension using ambulatory BP monitoring (≥135/85 mm Hg for daytime ambulatory BP and ≥130/80 mm Hg for 24‐hour ambulatory BP). The implications of this difference remain unclear.
The results of this study are worrisome because they suggest that we may be overdiagnosing hypertension in the usual office setting. As a 10‐mm Hg increase in systolic BP is associated with a 50% increase in risk of cardiovascular events, we may be substantially overestimating individual risk and needlessly subjecting patients to specific therapy. If the results of this small clinical trial are confirmed in larger and more diverse patient populations, the use of casual office Bps to diagnose hypertension may have to be reevaluated. The easiest solution would appear to be more careful measurement of BP in the office setting using standardized published technique. As the standardized measurement of BP by a trained research nurse, who is usually not available in most practices, usually takes approximately 8–12 minutes and the usual office visit in the United States lasts only 10–15 minutes, patients need to arrive at office visits early if accurate BP measurements are to occur. The use of out‐of‐office BP measurement with either ambulatory or home monitoring may become increasingly important, although these methods are subject to their own limitations. Practitioners must have confidence that the BP measurements taken in the office are accurate.