Judith Mekler

Predictors of All-Cause Mortality in Clinical Ambulatory Monitoring: Unique Aspects of Blood Pressure During Sleep

The prognostic value of sleep blood pressure reported by recent studies is variable. Our aim was to examine the relationship of sleep blood pressure, measured by 24-hour ambulatory blood pressure monitoring, with all-cause mortality. We studied a cohort of 3957 patients aged 55±16 (58% treated) referred for ambulatory monitoring (1991–2005). Sleep, including daytime sleep, was recorded by diary. Linkage with the national population register identified 303 deaths during 27 750 person-years of follow-up. Hazard ratios (HRs) for mortality in Cox proportional hazards models that included age, sex, hypertension, and diabetes treatment were 1.32 (95% CI: 0.99 to 1.76) for awake hypertension (≥135/85 mm Hg), and 1.67 (95% CI: 1.25 to 2.23) for sleep hypertension (≥120/70 mm Hg). By quintile analysis, the upper fifths of systolic and diastolic dipping during sleep were associated with adjusted HRs of 0.58 (95% CI: 0.41 to 0.82) and 0.68 (95% CI: 0.48 to 0.96), respectively. In a model controlling for awake systolic blood pressure, hazards associated with reduced systolic dipping increased from dippers (>10%; HR: 1.0), through nondippers (0% to 9.9%; HR: 1.30; 95% CI: 1.00 to 1.69) to risers (<0%; HR: 1.96; 95% CI: 1.43 to 2.96). Thus, in practice, ambulatory blood pressure predicts mortality significantly better than clinic blood pressure. The availability of blood pressure measures during sleep and, in particular, the pattern of dipping add clinically predictive information and provide further justification for the use of ambulatory monitoring in patient management.

Dipping Status May Be Determined by Nocturnal Urination

Nondipping, ie, failure to reduce blood pressure by ≥10% during the night, is considered an important prognostic variable of 24-hour ambulatory blood pressure monitoring. However, some people wake up at night to urinate. Usually, 24-hour ambulatory blood pressure monitoring–derived blood pressure includes these rises in the nighttime blood pressure mean. We identified 97 subjects undergoing 24-hour ambulatory blood pressure monitoring who reported waking up at night to urinate. We assessed the 24-hour ambulatory blood pressure monitoring first using total daytime and total nighttime means and then using actual daytime awake and nighttime asleep (as reported by the patient) means. Nocturnal decline in blood pressure was 14.4±8.5/11.8±6.1 mm Hg with the first method and 17.1±8.3/13.8±5.9 mm Hg with the second one (P <0.00001). Although the absolute difference between the nocturnal blood pressure declines calculated by the 2 methods was small, the effect on nocturnal dip was profound. Average systolic blood pressure dipping was 10.1% by the total day–total night method and 12.0% by the actual day awake–night asleep method (P ≤0.00001), and that of diastolic blood pressure was 14.2% and 16.7%, respectively (P ≤0.00001). The prevalence of systolic blood pressure nondipping decreased from 42.2% by the first method to 31.9% by the second method (P ≤0.0056), and that of diastolic blood pressure nondipping decreased from 22.6% to 11.3% (P ≤0.00001). Inclusion of awake blood pressure measurements during the night obscured the normal dipping pattern in people who woke up to urinate. Thus, taking into account people’s actual behavior increases the accuracy of the results.

The association of a simple blood pressure-independent parameter derived from ambulatory blood pressure variability with short-term mortality

We explored the predictive ability of the blood pressure variability ratio (BPVR), defined as the ratio of 24-h ambulatory systolic blood pressure variability to diastolic variability, and evaluated its predictable relation with blood pressure and the Ambulatory Arterial Stiffness Index (AASI). A total of 3433 consecutive patients were followed up to 16 years for all-cause mortality. Blood pressure variability was expressed by the standard deviation. BPVR, which is the systolic-on-diastolic slope estimated by a known type of symmetric regression (‘reduced major axis’), was compared with other expressions of this slope and with AASI using other regression procedures. Time-dependent Cox proportional hazard models, adjusted for demographics, 24-h mean blood pressure, 24-h pulse pressure and dipping were used to assess the association of BPVR and slope-related parameters with all-cause mortality. We found that Pearsons correlation between BPVR and the symmetric slope was 0.957, and between 1–1/BPVR (an AASI-equivalent expression) and the symmetric version of AASI was 0.973. BPVR was entirely independent of mean arterial pressure (r=0.013). Systolic and diastolic ambulatory blood pressure variability was not significantly associated with mortality. Over 16 years, BPVR predicted all-cause mortality [hazard ratio=1.21 (95% CI 1.05–1.40) per 1 s.d. increase]. In time-dependent models, increased BPVR was strongly associated with an 18-month mortality, weakly related to 7 years mortality, showing no effect thereafter. Thus, the ratio between 24-h systolic and diastolic blood pressure variability, readily available from ambulatory monitoring reports, is an easy-to-calculate systolic-on-diastolic slope. It is a blood pressure-independent measure believed to express an arterial property, with prognostic power similar to that of AASI.

The white coat phenomenon is benign in referred treated patients: a 14-year ambulatory blood pressure mortality study.

Objective Previous reports on the prognosis of white coat hypertension are ambiguous. We aimed to determine the prognostic implications of the white coat phenomenon in treated patients. Methods Our 14-year hospital-based ambulatory blood pressure (BP) monitoring prospective database was analyzed for all-cause mortality. The relationships of the white coat and masking effects with mortality were assessed both categorically (controlled awake versus clinic BP) and in a continuous mode (clinic–awake BP difference). Results During the follow-up period, 2285 treated patients (aged 61 ± 13 years, 57% women) were monitored (17 621 patient-years, 286 deaths). Mean BMI was 27.8 ± 4.5 kg/m2 and 13% were treated for diabetes. Controlled hypertension (normal clinic and awake BP) was found in 15.8%, high clinic BP (with controlled awake BP; namely, white coat uncontrolled hypertension) in 12.1%, awake hypertension (with controlled clinic BP; namely, masked uncontrolled hypertension) in 11.8%, and sustained hypertension (both clinic and awake) in 60.3%. Compared with white coat uncontrolled hypertension, age-adjusted Cox-proportional all-cause mortality hazard ratios were 1.42 (0.81–2.51) for controlled hypertension, 1.88 (1.08–3.27) for masked uncontrolled hypertension, and 2.02 (1.30–3.13) for sustained hypertension. Hazards ratios per 1% increase in the clinic–awake BP difference were 0.992 (0.983–1.002) for systolic BP and 0.981 (0.971–0.991) for diastolic BP, adjusted for age, sex, diabetes, and either systolic or diastolic awake BP, respectively. Conclusion In treated hypertensive patients referred for ambulatory BP monitoring, the white coat effect is benign compared with the reverse (masking) phenomenon, which has a poorer prognosis.

Disparate effects of exercise training on glucose tolerance and insulin levels and on ambulatory blood pressure in hypertensive patients

Objective: To assess the relationship of insulin levels and glucose tolerance to blood pressure in hypertension. Design: An open, prospective trial of exercise training with ambulatory blood pressure monitoring and intravenous glucose tolerance testing before and after a 14-week training programme. Patients: Twenty sedentary, untreated, non-obese, normoglycaemic individuals of both sexes with uncomplicated essential hypertension, of whom 16 completed the study. Intervention: Fourteen weeks of supervised, low-intensity, group exercise of three 1-h sessions per week. Main outcome measures: Ambulatory and clinic blood pressure, and glucose and insulin responses to an intravenous glucose tolerance test. Results: Maximal work capacity on a bicycle ergometer increased by 20% (P < 0.001); 24-h ambulatory blood pressure was 143 ± 12/87 ± 5 mmHg before and 142 ± 13/87 ± 7 mmHg after training. Clinic blood pressure decreased from 166 ± 14/103 ± 5 mmHg to 157 ± 12/99 ± 6 mmHg (P < 0.03). Two-way analysis of variance indicated significant decreases in both glucose (P < 0.04) and insulin (P < 0.03), fasting and throughout the intravenous glucose tolerance test. Conclusions: Although mild exercise reduced clinic blood pressure significantly, it did not affect ambulatory blood pressure despite a marked reduction in glucose and insulin levels. This finding argues against a determinant role of insulin in the 24-h maintenance of blood pressure in hypertensive patients under the conditions of the study.

Subpressor Dose of L-NAME Unmasks Hypertensive Effect of Chronic Hyperinsulinemia

We previously found that chronic exogenous hyperinsulinemia without sugar supplementation does not elevate blood pressure. This may be partially explained by the ability of insulin to release nitric oxide and cause vasodilatation. To test this hypothesis, we studied 4 groups of rats: 9 rats (body weight, 213±14 g) treated with a gradual increase of a sustained-release subcutaneous insulin pellet; 9 rats (body weight, 213±9 g) treated with NG-nitro-l-arginine methyl ester (L-NAME) in drinking water 50 mg/L; 19 rats (body weight, 217±11 g) treated with the combination of L-NAME and insulin; and 9 control rats (body weight, 218±11 g). Blood pressure was followed weekly for 6 weeks, and then rats were studied in metabolic cages. Weight gain was not different during the 6 weeks. Renal function did not differ between the 4 groups, but 24-hour urinary nitrite/nitrate excretion was lower (P <0.02) in L-NAME–treated and higher in insulin-treated rats. Plasma insulin doubled (P <0.002) in the insulin-treated rats, but there was no hypoglycemia and, by week 6, fructosamine levels were 2.1±0.2, 2.1±0.2, 2.3±0.1, and 2.3±0.2 mmol/L in control rats and rats treated with L-NAME, insulin, and L-NAME plus insulin, respectively. Systolic blood pressure, which did not differ at baseline, at week 3 was 122±17, 118±17, and 118±24 mm Hg in the control, L-NAME, and insulin groups and 136±14 mm Hg (P <0.03) in the combination group. At week 6, systolic blood pressure was 128±14, 127±15, and 118±13 mm Hg in the control, L-NAME, and insulin groups, respectively, and 150±14 mm Hg (P <0.0005) in the combination group. In a subsequent experiment, l-arginine 2 g/L abrogated the effects of L-NAME and insulin combination. In conclusion, chronic exogenous hyperinsulinemia does not affect blood pressure but may cause hypertension when endothelial function is compromised.

How should patients treated with α-blockers be followed? Insights from an ambulatory blood pressure monitoring database

Objective Adrenergic alpha-antagonists have been suggested to confer lesser protection, compared to diuretics, when used as first agents for hypertension. While differences in clinic blood pressure may be partly responsible, this inferiority is unexpected in light of the metabolic advantages of α-blockade. The aim of this study was to evaluate the relationship between use of α-blockers and blood pressure dipping. Methods A database of a 24-h ambulatory monitoring service was cross-sectionally evaluated for associations between antihypertensives and dipping. There were 681 treated subjects during a 3-year period (age 63 ± 14, 57% female). Results Overall, 78 of 681 treated hypertensive subjects used α-blockers (11%). Nine per cent of dippers and 16% of nondippers were treated with α-blockade, odds ratio 2.0. Whereas clinic, 24-h, and awake blood pressures were similar in α-blocker users and nonusers, sleep blood pressure was significantly higher in the former group. Furthermore, significantly fewer subjects given α-blockers had a controlled sleep blood pressure. Among α-blocker nonusers sleep blood pressure was the best controlled category, whereas in α-blocker users manual blood pressure had the highest rate of control. Generally, accounting for covariates of α-blockade (age, gender, diabetes, total number of medications) did not influence the above-mentioned trends. Finally, a limited negative dose–response relationship between α-blockade and dipping magnitude was also noticed. Conclusions We found a significant negative association between adrenergic α-blockade and the magnitude of sleep-related blood pressure decline. Awaiting results from interventional studies, this may suggest a need to perform ambulatory monitoring in patients given alpha-blocking agents (or at least supine and standing measurements), and may partially clarify the inferiority of doxazosin in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).

Effect of Acute N-Nitro-l-Arginine Methyl Ester (L-NAME) Hypertension on Glucose Tolerance, Insulin Levels, and [3H]-Deoxyglucose Muscle Uptake

This study was conducted to test the hypothesis that acute, widespread N-nitro-L-arginine methyl ester (L-NAME) induced vasoconstriction and hypertension may affect glucose tolerance and insulin sensitivity in normal rats. Comparisons were made of blood pressure, intravenous glucose tolerance, and insulin response and [3H]-deoxyglucose tissue uptake between L-NAME and control treated rats. Chronically instrumented, awake rats were administered L-NAME (30 mg/kg) (n = 8) or saline (0.3 mL) (n = 8) intravenously. After blood pressure stabilized, a bolus injection containing glucose (300 mg/kg) and trace [3H]-deoxyglucose was administered. Arterial blood was sampled for evaluation of glucose tolerance, insulin response, and [3H]-deoxyglucose muscle uptake. L-NAME treated rats had a persistent 54 +/- 4 mm Hg blood pressure rise while fasting, and postload plasma glucose and insulin responses did not differ, nor did heart and striated muscle [3H]-deoxyglucose uptake differ. In conclusion, acute L-NAME induced hypertension does not result in glucose intolerance, hyperinsulinemia, or decreased [3H]-deoxyglucose muscle uptake.

Hyperinsulinemia Increases Placenta Endothelin-Converting Enzyme-1 Expression in Trophoblasts

BACKGROUND Exogenous hyperisulinemia causes pregnancy, induced hypertension and intrauterine growth restriction (IUGR) in pregnant rats. Hyperinsulinemia may increase production of endothelin-1 (ET-1), produced by sequential proteolysis of the big endothelin by the endothelin-converting enzyme (ECE)-1, the expression of which is examined here in the placenta, kidney, heart, and liver. METHODS Rats were rendered hyperinsulinemic by subcutaneous insulin pellet, mated and followed to the twenty-first day of pregnancy. They were then killed, and their fetuses and placentas were examined. RESULTS Hyperinsulinemic dams (HD) had higher blood pressure (BP) (130 ± 17 mm Hg in HD vs. 115 ± 16 mm Hg in normal pregnant dams (NPD), P < 0.05), lower placenta weight (0.44 ± 0.08 g in HD vs. 0.47 ± 0.08 NPD, P < 0.05) and lower fetus weight: males 4.9 ± 0.4 g in HD vs. 5.5 ± 0.4 g in NPD, P < 0.0001; females 4.7 ± 0.4 g in HD vs. 5.2 ± 0.4 g in NPD (P < 0.0001). ECE-1 expression as determined by western blot was significantly increased in the placenta and its implantation site, i.e., the mesometrial triangle (MT) of HD by 46 and 48%, respectively. In the kidney and heart of HD ECE-1, protein expression was increased by 230 and 220%, respectively, but its level in the liver was similar in both groups. Immunohistochemical staining revealed ECE-1 expression in endothelial cells and trophoblastic cells of the placenta and MT. Endothelin receptor A (ET-A), a mediator of vasoconstriction by ET-1, was also expressed in the endothelium and in trophoblasts of the placenta and MT. The expression of both ECE-1 and ET-A, as measured by automated image analysis, was generally stronger in placentas of HD. CONCLUSIONS ECE-1 and ET-A are expressed in the trophoblastic cells of the placenta and MT. This may affect local endothelin levels, BP and IUGR.

Chronic exogenous hyperinsulinaemia without sugar supplementation: acute salt-sensitive hypertension without changes in resting blood pressure.

Objective: To study the effects of chronic insulin administration without sugar supplementation on blood pressure and response to acute saline loading in normal rats. Design: Design: Comparison of blood pressure, insulin and glucose levels in 24 insulin-treated and 12 control rats on regular rat chow (not supplemented with sugar). Methods: Sustained-release insulin implants (or sham implantation for the control rats) were administered subcutaneously. The sustained-release insulin implant size was gradually increased. Tail-cuff systolic blood pressure, insulin and glucose were measured twice a week for 8 weeks, after which intra-arterial blood pressure was recorded under resting conditions and 2 h after saline loading in seven insulin-treated and seven control rats. Results: Insulin-treated rats had a 1.2- to twofold increase in insulin without hypoglycaemia, a small but significant increase in glucose levels being found at weeks 6 and 8. When the rats were killed (week 8) triglyceride and fructosamine levels were increased in the insulin-treated rats in comparison with controls. Neither tail-cuff systolic blood pressure nor resting intra-arterial blood pressure differed between the two groups. However, acute saline loading resulted in significantly higher blood pressure in the insulin-treated rats, without altering renal Na+ excretion. Conclusions: Insulin-treated rats had a 1.2- to twofold increase in insulin without hypoglycaemia, a small but significant increase in glucose levels being found at weeks 6 and 8. When the rats were killed (week 8) triglyceride and fructosamine levels were increased in the insulin-treated rats in comparison with controls. Neither tail-cuff systolic blood pressure nor resting intra-arterial blood pressure differed between the two groups. However, acute saline loading resulted in significantly higher blood pressure in the insulin-treated rats, without altering renal Na+ excretion.