Abstract
Orthostatic hypotension (OH) is associated with hypertension and diabetes mellitus. However, in populations with both hypertension and diabetes, its prevalence, the effect of intensive versus standard systolic blood pressure (BP) targets on incident OH, and its prognostic significance are unclear. In 4266 participants in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) BP trial, seated BP was measured 3 times, followed by readings every minute for 3 minutes after standing. Orthostatic BP change, calculated as the minimum standing minus the mean seated systolic BP and diastolic BP, was assessed at baseline,12, and 48 months. The relationship between OH and clinical outcomes (total and cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, heart failure hospitalization or death and the primary composite outcome of non-fatal myocardial infarction, non-fatal stroke and cardiovascular death) was assessed using proportional hazards analysis. Consensus OH, defined by orthostatic decline in systolic BP ≥20 mm Hg and/or diastolic BP ≥10 mm Hg, occurred at ≥1 time point in 20% of participants. Neither age nor systolic BP treatment target (intensive, <120 mm Hg versus standard, <140 mm Hg) was related to OH incidence. Over a median follow-up of 46.9 months, OH was associated with increased risk of total death (HR=1.61, 95% CI 1.11–2.36) and heart failure death/hospitalization (HR=1.85, 95% CI 1.17–2.93), but not with the primary outcome or other prespecified outcomes. In patients with type 2 diabetes and hypertension, OH was common, not associated with intensive versus standard BP treatment goals, and predicted increased mortality and heart failure events.
Keywords: type 2 diabetes mellitus, hypertension, orthostatic hypotension, randomized clinical trial, cardiovascular disease
Introduction
After assumption of upright posture from supine or seated positions, arterial blood pressure (BP) is typically maintained by increased sympathetic activity leading to vasoconstriction and increased heart rate. Orthostatic hypotension (OH) occurs when these normal compensatory mechanisms are deficient. The most widely accepted definition of OH is a decline of systolic BP (SBP) ≥20 mm Hg or a decline of diastolic BP (DBP) ≥10 mm Hg within 3 minutes of standing (1). Risk factors associated with OH include advanced age, hypertension, antihypertensive drugs, and diabetes (2,3). OH may lead to postural instability, falls, and syncope (4), and is predictive of increased risk for stroke (5), coronary events (6,7) and total mortality (7–9) in community-based populations.
Although hypertension occurs in more than 50% of adults with diabetes, there is surprisingly little information on the incidence or prognostic significance of OH in this setting. Participants in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) BP trial were at high risk for OH because of their diabetes, hypertension, and treatment with antihypertensive drugs (10). It was anticipated that those individuals randomized to the intensive SBP treatment goal of <120 mmHg might be at particularly high risk for OH because of their greater number and doses of antihypertensive medications and their lower SBP goal. The present study therefore seeks to answer two important questions: (1) What is the incidence of OH in adults with diabetes and hypertension randomized to intensive (SBP <120 mm Hg) versus standard (SBP <140 mm Hg) BP goals? (2) Does OH have independent prognostic significance for the combined primary endpoint of the trial (nonfatal myocardial infarction, nonfatal stroke, and cardiovascular death) in this population? The large size and design of ACCORD presented an excellent opportunity to answer these questions.
Methods
Study Design
ACCORD was a double 2-by-2 factorial design randomized trial conducted at 77 clinical sites organized into 7 networks in the United States and Canada (11). The trial enrolled 10251 high-risk participants with type 2 diabetes mellitus (T2DM). Participants were eligible if they had T2DM and a glycated hemoglobin level ≥7.5% and were between the ages of 40 and 79 years with cardiovascular disease (CVD) or 55–79 years with anatomical evidence of subclinical atherosclerosis, albuminuria, left ventricular hypertrophy, or ≥2 additional risk factors for CVD (dyslipidemia, hypertension, smoking, or obesity). Study exclusions included a body mass index >45 and a serum creatinine >1.5 mg/dl. All participants were randomly assigned to either intensive or standard glycemic control (the ACCORD Glycemia trial) (12). In addition, 5518 of these participants were randomly assigned to either simvastatin plus fenofibrate or simvastatin plus placebo (the ACCORD Lipid trial) (13); the remaining 4733 participants were randomly assigned to either intensive or standard BP control (the ACCORD BP trial) (10), on which the current report is based.
ACCORD BP was a non-blinded trial in which participants were randomly assigned to intensive therapy that targeted SBP of <120 mm Hg or to standard therapy that targeted SBP of <140 mm Hg. Individuals with SBP between 130 and 180 mm Hg who were taking ≤3 antihypertensive medications and who had the equivalent of a 24-hour urinary protein excretion of <1.0 g were eligible for the BP trial. The ACCORD BP trial tested a treatment strategy to achieve specific SBP goals rather than an evaluation of any specific drug regimen. Therefore, all available classes of antihypertensive medications were used to lower BP as previously described (10). The average difference in BP achieved in the in the intensive treatment group compared to the standard treatment group was 14.2 mm Hg for SBP and 6.1 mm Hg for DBP. The mean number of antihypertensive medications used was 2.1 for the standard group and 3.4 for the intensive group. The primary composite endpoint of cardiovascular death, non-fatal myocardial infarction (MI) or non-fatal stroke did not differ significantly between treatment arms although the risk for stroke, a prespecified secondary outcome, was significantly lower in the intensive treatment arm (10). Serious adverse events related to the intervention, including most commonly hypotension and syncope, were observed in 3.3% of the intensive group compared with 1.3% in the standard group (p< 0.001)(10).
All participants in the ACCORD BP trial were eligible for participation in the OH measurements, which began in October 2004. Standing BP was measured at baseline and at the 12-month and 48-month follow-up visits using the same automated oscillometric device (Omron HEM-907) that was used to measure seated BP. After at least 5 minutes of seated rest, BP was determined 3 times at 1-minute intervals; participants were then asked to stand. Starting when their feet touched the ground, BP was taken every minute for 3 minutes in the same arm used for the seated measurements. Blood pressure change was calculated using the minimum standing measurement minus the mean of the seated measurements (1) The person recording the standing measurements also recorded whether the participant experienced dizziness or felt lightheaded upon standing for this assessment.
The occurrence of consensus OH was defined by a decline in SBP ≥20 mm Hg or a decline in DBP ≥10 mm Hg from sitting to standing. (1) The detection of OH unaccompanied by orthostatic symptoms of dizziness, presyncope or syncope was not to influence treatment, according to the antihypertensive drug treatment algorithm. Medication adjustments were allowed for the management of symptomatic OH. For participants in the intensive therapy group, clinic visits to assess BP were scheduled once a month for 4 months and every 2 months thereafter; for participants in the standard therapy group, visits were scheduled at months 1 and 4 and every 4 months thereafter. Additional visits were scheduled as needed in both groups to monitor and ensure appropriate implementation of the study intervention strategies. Telephone follow-up was used for those unable to visit the clinic. The occurrence of a coronary event, stroke, or cardiovascular death was reported by the clinic staff and adjudicated by a Clinical Events Committee, blinded to treatment group assignment.
Statistical Analysis
Differences in baseline characteristics, including the prevalence of OH symptoms, between the group ever demonstrating consensus OH and the group that did not were compared by two sample t-test or chi-square analysis. The baseline characteristics examined included age, gender, race, body mass index, duration of diabetes, smoking status, glycated hemoglobin, hematocrit, treatment with antihypertensive drugs, and history of coronary heart disease, stroke, heart failure, or carotid/peripheral artery disease.
Prevalence of OH at a specific visit was defined based on the occurrence of consensus OH at that visit, regardless of whether OH had been identified at prior visits. Incidence of OH at a specific follow-up visit was defined as the occurrence of consensus OH at that visit among participants who had been examined previously but had not demonstrated OH. Resolution of OH at a specific follow-up visit is defined as the absence of consensus OH at that visit among participants that had consensus OH at their last assessment. Visit-specific differences in OH prevalence, incidence and resolution by assigned treatment group were evaluated using chi-square analyses. Because the cohorts of participants examined differed among visits, we also performed a similar analysis restricted to participants examined at all three visits (baseline and 12 and 48 months post-randomization).
The associations between baseline characteristics and the occurrence of consensus OH at baseline or during follow-up were examined using generalized linear mixed model assuming a binary distribution and a logit link, with presence of consensus OH as the response variable and after adjusting for subject and all variables shown in Table 4. Visit month was treated as a time dependent variable in these analyses. By treating each assessment as a separate observation while accounting for the within-in subject correlation, this model assigns more weight to participants with multiple assessments than to those assessed only a single time point. Similar analyses based on general linear models for repeated measurements with delta SBP (Standing SBP – Seated SBP) and delta DBP (Standing DBP – Seated DBP) as dependent variables are presented in the online appendix.
Table 4.
Factors associated with orthostatic hypotension after adjustment for other variables shown
| Baseline Characteristic | Odds Ratio | 95%Confidence Interval | P-value |
|---|---|---|---|
| BP Intervention – Intensive vs. Standard | 0.925 | 0.798, 1.074 | 0.31 |
| Glycemia Intervention – Intensive vs. Standard | 0.947 | 0.816, 1.098 | 0.47 |
| 12 Month Visit vs. Baseline | 0.554 | 0.459, 0.668 | <0.0001 |
| 48 Month Visit vs. Baseline | 0.702 | 0.591, 0.835 | <0.0001 |
| Age (years) | 1.003 | 0.991, 1.016 | 0.61 |
| Female vs. Male | 1.251 | 1.054, 1.485 | 0.01 |
| Hispanic vs. Non-Hispanic White | 0.863 | 0.634, 1.174 | 0.34 |
| African American vs. Non-Hispanic White | 0.667 | 0.544, 0.819 | 0.0001 |
| Cardiovascular Disease History | 0.902 | 0.719, 1.131 | 0.37 |
| Prior Myocardial Infarction | 1.261 | 0.968 ,1.644 | 0.09 |
| Prior Stroke | 1.120 | 0.804, 1.560 | 0.50 |
| Prior Heart Failure | 1.026 | 0.713, 1.478 | 0.89 |
| Prior Coronary Revascularization | 1.499 | 0.159, 14.122 | 0.72 |
| Diabetes Duration (years) | 1.006 | 0.996, 1.017 | 0.24 |
| Peripheral Neuropathy | 1.076 | 0.920, 1.257 | 0.36 |
| Left Ventricular Hypertrophy | 0.989 | 0.703, 1.393 | 0.95 |
| Hemoglobin A1c (%) | 1.088 | 1.014, 1.167 | 0.02 |
| Body Mass Index (kg/m2) | 0.997 | 0.982, 1.012 | 0.68 |
| Systolic BP (mmHg) | 1.013 | 1.007, 1.018 | <0.0001 |
| Diastolic BP (mmHg) | 1.007 | 0.998, 1.016 | 0.13 |
| LDL-Cholesterol (mg/dL) | 1.000 | 0.998, 1.002 | 0.88 |
| HDL-Cholesterol (mg/dL) | 1.003 | 0.996, 1.010 | 0.39 |
| Triglycerides (mg/dL) | 1.000 | 0.999, 1.000 | 0.58 |
| Current vs. Never Smokers | 1.382 | 1.095, 1.743 | 0.006 |
| Former vs. Never Smokers | 1.164 | 0.984, 1.377 | 0.08 |
| Renin Angiotensin System Inhibitors | 0.956 | 0.811, 1.126 | 0.59 |
| Diuretics | 0.993 | 0.841, 1172 | 0.93 |
| Beta Blockers | 1.240 | 1.033, 1.288 | 0.02 |
| Calcium Channel Blockers | 1.035 | 0.850, 1.260 | 0.73 |
| Alpha Blockers | 1.701 | 1.018, 2.843 | 0.04 |
| Statins | 1.033 | 0.872, 1.223 | 0.71 |
| Fibrates | 1.146 | 0.862, 1.523 | 0.35 |
| Insulin | 1.242 | 1.030, 1.497 | 0.02 |
| Metformin | 0.938 | 0.800, 1.101 | 0.43 |
| Sulfonylureas | 0.994 | 0.840, 1.176 | 0.95 |
| Thiazolidinediones | 1.098 | 0.915, 1.318 | 0.32 |
Abbreviations: BP= blood pressure; HDL= high density lipoprotein; LDL= low density lipoprotein
The associations between consensus OH and subsequent cardiovascular events were evaluated using proportional hazards models accounting for left truncation, with follow-up for each participant beginning at the time of the first standing BP assessment and OH status treated as a time dependent covariate based on the occurrence of consensus OH at the most recent assessment for each participant. All models included covariates controlling for BP treatment group assignment, glycemia treatment group assignment and history of CVD at baseline. We also tested for interactions between BP treatment group assignment and consensus OH. In sensitivity analyses, we also examined systolic OH alone, time dependent definitions of OH based on any occurrence (ever OH) or based on an updated value defined as the number of visits with OH divided by the number of assessments completed at or prior to the most recent visit. In all cases, results were qualitatively similar to the model with OH based solely on the last visit; the results for the systolic OH analyses are included in the online appendix, as are analyses of delta SBP and delta DBP as predictors in place of OH. For all analyses, a 2 tailed p<0.05 was required to reject the null hypothesis.
Results
In total, 4266 ACCORD BP participants had orthostatic BP measurements on at least one visit, including 1321 at baseline, 2625 at 12 months, and 3702 at 48 months, with 926 having measurements at all 3 time points (Figure 1). The smaller number of orthostatic BP measurements at baseline and12 months was because the OH measurements did not commence until 44 months after ACCORD BP began. Consensus OH occurred on at least one time point in 852 individuals (20.0%). Individuals with OH were more likely to be women, non-Hispanic white, have had prior MI, and use beta blockers and insulin. They also had higher seated SBP and DPB at baseline. Those with OH were less likely to use sulfonylureas (Table 1).
Figure 1.
Consort diagram showing the number of participants undergoing orthostatic blood pressure measurements at baseline and at 12 months and 48 months after randomization
Table 1.
Participant characteristics by orthostatic hypotension status
| Baseline Characteristic | Ever Consensus OH | Never Consensus OH | P-value | ||
|---|---|---|---|---|---|
|
| |||||
| N | Value* | N | Value* | ||
| Age (years) | 852 | 62.1 (6.8) | 3414 | 62.1 (6.8) | 0.94 |
| Gender - % Male | 852 | 50.2 (428) | 3414 | 54.2 (1849) | 0.04 |
| Ethnicity - % Non-Hispanic White | 852 | 68.3 (582) | 3414 | 60.5 (2065) | <.0001 |
| Ethnicity - % Hispanic | 852 | 6.5 (55) | 3414 | 6.6 (225) | 0.94 |
| Ethnicity - % African American | 852 | 19.1 (163) | 3414 | 23.5 (803) | 0.006 |
| Prior Cardiovascular Event (%) | 852 | 35.1 (299) | 3414 | 32.7 (1115) | 0.18 |
| Prior Myocardial Infarction (%) | 852 | 16.7 (142) | 3414 | 12.9 (439) | 0.004 |
| Prior Stroke (%) | 852 | 6.8 (58) | 3414 | 5.9 (201) | 0.34 |
| Prior Heart Failure (%) | 843 | 4.9 (41) | 3381 | 3.9 (130) | 0.20 |
| Prior Coronary Revascularization (%) | 852 | 0.12 (1) | 3414 | 0.15 (5) | 1 |
| Diabetes Duration (median, years) | 843 | 10 (11) | 3380 | 10 (10) | 0.076 |
| Peripheral Neuropathy (%) | 851 | 42.1 (358) | 3410 | 38.9 (1326) | 0.092 |
| Left Ventricular Hypertrophy (%) | 851 | 5.3 (45) | 3395 | 5.2 (175) | 0.86 |
| Body Mass Index (kg/m2) | 852 | 32.3 (5.3) | 3414 | 32.1 (5.6) | 0.44 |
| Systolic Blood Pressure (mmHg) | 852 | 141.6 (17.5) | 3414 | 138.5 (15.2) | <.0001 |
| Diastolic Blood Pressure (mmHg) | 852 | 76.9 (11.09) | 3414 | 75.7 (10.1) | 0.003 |
| LDL -Cholesterol (mg/dL) | 850 | 109.6 (37.0) | 3393 | 109.5 (36.5) | 0.93 |
| HDL -Cholesterol (mg/dL) | 850 | 46.7 (13.3) | 3393 | 46.0 (13.6) | 0.21 |
| Triglycerides (median, mg/dL) | 850 | 151 (129) | 3393 | 147 (130) | 0.39 |
| Former Smokers (%) | 850 | 44.0 (374) | 3412 | 42.4 (1448) | 0.20 |
| Current Smokers (%) | 850 | 14.0 (119) | 3412 | 12.4 (424) | 0.20 |
| Antihypertensives (%) | 852 | 87.4 (745) | 3414 | 87.3 (2981) | 0.95 |
| RAS Inhibitors (%) | 852 | 67.1 (572) | 3414 | 68.5 (2339) | 0.46 |
| Diuretics (%) | 852 | 34.7 (296) | 3414 | 34.5 (1177) | 0.90 |
| Beta Blockers (%) | 852 | 30.1 (256) | 3414 | 24.7 (843) | 0.0016 |
| Calcium Channel Blockers (%) | 852 | 18.4 (157) | 3414 | 18.3 (624) | 0.92 |
| Alpha Blockers (%) | 852 | 2.2 (19) | 3414 | 1.7 (58) | 0.31 |
| Statins (%) | 852 | 67.0 (571) | 3414 | 65.3 (2229) | 0.35 |
| Fibrates (%) | 852 | 8.2 (70) | 3414 | 6.7 (227) | 0.11 |
| Insulin (%) | 852 | 41.9 (357) | 3414 | 35.5 (1211) | 0.0005 |
| Metformin (%) | 852 | 55.1 (469) | 3414 | 58.0 (1981) | 0.12 |
| Sulfonylureas (%) | 852 | 44.7 (381) | 3414 | 48.6 (1658) | 0.046 |
| Thiazolidinediones (%) | 852 | 22.1 (188) | 3414 | 21.2 (722) | 0.57 |
Values represent mean (SD) for continuous variables and % (n) for discrete variables.
Abbreviations: HDL= high density lipoprotein, LLDL= low density lipoprotein, RAS=renin angiotensin system
Table 2 compares the prevalence, incidence and resolution of OH in the intensive versus standard BP treatment groups. The most noteworthy finding is the lack of significant difference in OH prevalence, incidence or resolution between the treatment groups at all time points. A similar lack of treatment group difference in OH occurrence was observed in the subset with standing BP measurements at all 3 points. Although OH prevalence at baseline averaged 17.8% in the total sample with baseline measurements, prevalence at 12 and 48 months declined to 10.4% and 12.8%, respectively, with resolution rates between 70 and 82%. Similar time trends were seen in the subset assessed at all 3 points.
Table 2.
Prevalence and incidence of consensus orthostatic hypotension by visit and treatment group
| Participants Analyzed | Intensive Group % (n/N) | Standard Group % (n/N) | P-Value | |
|---|---|---|---|---|
| All Participants | Prevalence | |||
| Baseline | 19.3 (127/658) | 16.1 (107/663) | 0.15 | |
| 12 Months | 9.5 (126/1324) | 11.4 (148/1301) | 0.13 | |
| 48 Months | 12.2 (223/1830) | 13.5 (252/1872) | 0.26 | |
| Incidence | ||||
| 12 Months* | 8.0 (36/453) | 9.9 (47/473) | 0.30 | |
| 48 Months† | 9.9 (95/958) | 11.0 (104/942) | 0.45 | |
| Resolution | ||||
| 12 Months‡ | 81.7 (89/109) | 78.2 (68/ 87) | 0.54 | |
| 48 Months§ | 70.5 (74/105) | 71.2 (89/125) | 0.90 | |
| Participants with data at all 3 visits | Prevalence | |||
| Baseline | 18.7 (86/461) | 16.8 (78/465) | 0.49 | |
| 12 Months | 10.2 (47/461) | 11.6 (54/465) | 0.53 | |
| 48 Months | 12.2 (56/461) | 12.9 (60/465) | 0.77 | |
| Incidence | ||||
| 12 Months* | 7.7 (29/375) | 9.8 (38/387) | 0.31 | |
| 48 Months† | 9.8 (34/346) | 8.6 (30/349) | 0.58 | |
| Resolution | ||||
| 12 Months‡ | 79.1 (68/ 86) | 79.5 (62/ 78) | 0.95 | |
| 48 Months§ | 70.2 (33/ 47) | 75.9 (41/ 54) | 0.52 | |
Restricted to participants assessed and free from OH at baseline
Restricted to participants assessed at baseline or 12 months and free from OH
Restricted to participants with OH at baseline
Restricted to participants with OH at prior visit
At the baseline examination, 31 of 676 (4.6%) intensive group participants and 35 of 679 (5.2%) standard group participants reported feeling dizzy upon standing for the BP measurement (p=0.71). During follow-up, these numbers were 71 of 1337 (5.3%) of intensive group participants and 65 of 1308 (5.0%) of standard group participants (p=0.72) for the 12 month exam, and 106 of1860 (5.7%) intensive group participants and 77 of 1896 (4.1%) standard group participants (p=0.02) at the 48 month exam. Thus, symptomatic dizziness on standing was modestly but significantly higher in the intensive compared with the standard group only during the at 48 month exam. Although <10% of participants with consensus OH at any time point reported dizziness on standing during the exam, at the 12 and 48 month visits participants with consensus OH were about twice as likely to report symptoms of dizziness upon standing during the BP exam compared to those without OH (Table 3). Of participants reporting dizziness when standing for the BP exam at baseline, only 11 of 66 (16.7%) had consensus OH. These numbers were 25 of 136 (18.4%) at 12 month follow-up and 39 of 183 (21.3%) at 48 month follow-up. Figure 2 displays histograms of the orthostatic change in SBP and DBP at the baseline visit, showing a normal distribution for both variables centered around 0 to -5 mmHg , with a similar proportion of individuals demonstrating reductions in SBP ≥20 mm Hg and/or reductions in DBP ≥10 mm Hg.
Table 3.
Dizziness on standing as a function of measured orthostatic hypotension and treatment arm.
| Time Point | Consensus OH | No Consensus OH | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| Overall | Intensive Therapy | Standard Therapy | Overall | Intensive Therapy | Standard Therapy | |
| Baseline Exam | 11/242 (4.6) | 6/132 (4.6) | 5/110 (4.6) | 55/1113 (4.9) | 25/544 (4.6) | 30/569 (5.3) |
| 12 Month Exam | 25/276 (9.1)* | 12/128 (9.4) | 13/148 (8.8) | 111/2369 (4.7) * | 59/1209 (4.9) | 52/1160 (4.5) |
| 48 Month Exam | 39/487 (8.0)* | 22/230 (9.6) † | 17/257 (6.6) † | 144/3269 (4.4) * | 84/1630 (5.2) † | 60/1639 (3.7) † |
Entries are number with event / number assessed (% with event). Abbreviation: OH=orthostatic hypotension
p<0.05 comparing overall percent with dizziness among participants with Consensus OH to the overall percent with dizziness among participants without Consensus OH using fisher’s exact test.
p<0.05 comparing percent with dizziness among intensive therapy participants to the percent with dizziness among standard therapy participants after stratifying on Consensus OH status using Cochran-Mantel-Hanzel statistics.
Figure 2.
Histograms showing distribution of orthostatic change in systolic (Panel A) and diastolic (Panel B) blood pressure at baseline
Table 4 shows the baseline characteristics independently associated with consensus OH on fully adjusted regression analysis. Female sex, higher SBP and hemoglobin A1c, current smoking, and use of beta blockers, alpha blockers, and insulin were associated with greater likelihood of OH, whereas African American race was associated with a lesser likelihood of OH. Of note, the assigned BP treatment group was not associated with OH. In addition, individuals were less likely to show OH at the 12 and 48 month visits than at baseline. In Supplement Tables S3 and S4, the baseline characteristics associated with delta SBP and delta DBP as continuous variables are generally similar to those for consensus OH.
Table 5 examines the relationship between consensus OH at the most recent visit and the primary combined outcome and its components and with other secondary outcomes after controlling for treatment assignment for both the blood pressure and glycemia interventions, CVD history at baseline and clinical network (primary outcome only). No significant relationship was observed between OH and the combined primary outcome of nonfatal MI, nonfatal stroke, and cardiovascular death or its components. However, compared with individuals without OH, those with OH had 85% higher risk for heart failure deaths or hospitalizations (p=0.01) and 62% higher risk for total mortality (p=0.02). Similar findings were observed for the subset of individuals with systolic OH (Supplement Tables S1 and S2) except that only the higher hazard for mortality was significant (HR =1. 80 , p<0.01). There was no evidence of an interaction between the effect of treatment group and the effect of consensus OH for any of the reported outcomes (p>0.13 in all cases).
Table 5.
Relationship between consensus orthostatic hypotension at most recent visit and cardiovascular outcomes
| Endpoint | OH | No OH | OH (Yes / No) | ||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| # of Events | % / yr | # of Events | % / yr | Hazard Ratio | 95%Confidence Interval | P-value | |
| Primary Outcome | 35 | 2.19 | 219 | 1.87 | 1.13 | (0.81,1.59) | 0.47 |
| Nonfatal MI | 25 | 1.56 | 140 | 1.19 | 1.31 | (0.87,1.98) | 0.21 |
| Nonfatal Stroke | 4 | 0.24 | 46 | 0.38 | 0.71 | (0.28,1.79) | 0.45 |
| CVD Mortality | 8 | 0.48 | 50 | 0.41 | 1.19 | (0.62,2.27) | 0.60 |
| Total Mortality | 28 | 1.67 | 125 | 1.03 | 1.62 | (1.10,2.36) | 0.02 |
| Total Stroke | 4 | 0.24 | 51 | 0.43 | 0.64 | (0.26,1.61) | 0.32 |
| HF Death or Hospitalization | 21 | 1.30 | 77 | 0.65 | 1.85 | (1.17,2.93) | 0.01 |
Abbreviations: CVD = cardiovascular disease; HF = heart failure; MI = myocardial infarction
Discussion
In the ACCORD BP Trial, consensus OH, defined by a decline in systolic BP ≥20 mm Hg or a decline in diastolic BP ≥10 mm Hg, occurred at one or more of the 3 time points in 20% of participants. Independent factors associated with OH were female sex, white race, current smoking, higher baseline systolic BP and hemoglobin A1c, and use of alpha blockers, beta blockers, and insulin. Of note, neither age nor assignment to intensive versus standard BP treatment goals was associated with OH, although dizziness on standing during the exam was slightly more common in the intensive than the standard group at the 48 month exam. Individuals were less likely to experience OH at 12 or 48 months of follow-up than at baseline. OH based on standing BP measurements was poorly correlated with orthostatic symptoms – only 17–20% of those reporting lightheadedness or dizziness on standing at the time of the exam had consensus OH, and <10% of those with consensus OH reported dizziness on standing during the exam. Although no significant relationship was found between OH and the combined primary endpoint of non-fatal MI, nonfatal stroke or cardiovascular death or its components, OH was associated with higher total mortality and higher rates of heart failure hospitalization and death.
The current findings in a large cohort with T2DM and hypertension share both similarities and noteworthy differences from prior studies examining the epidemiology of OH. Higher rates of OH have been associated with higher systolic BP (5,7,14,15), which may reflect both exaggerated baroreflex changes and regression to the mean of repeated BP measurements. Alpha adrenergic blocking drugs reduce BP by arterial vasodilation, which is exacerbated by venodilation upon standing; their ability to cause OH is widely recognized (16). Beta blockers have also been associated with OH, likely related to blunting of the compensatory cardioacceleration that occurs on standing (16). Association of OH with higher hemoglobin A1c and insulin use may reflect more advanced diabetes, with higher prevalence of later stages of cardiovascular autonomic neuropathy when OH is more likely to occur (17). In addition, insulin-induced vasodilatation, possibly mediated via endothelium-dependent mechanisms, may also contribute (18, 19). This explanation is further supported by trends for longer duration of diabetes and higher prevalence of peripheral neuropathy in persons with OH on univariate analysis, although both were non-significant on multivariable analysis. Higher rates of OH occurrence in women than men and in whites versus African Americans have been previously observed (20, 21).
A noteworthy difference between the current findings and prior studies was the lack of increase in OH with age in ACCORD. In epidemiologic studies OH prevalence is generally low in younger persons and increases substantially with age (5,8,9,14), probably secondary to age-associated autonomic dysfunction. The likely higher prevalence of autonomic dysfunction in the ACCORD cohort with longstanding T2DM plus the upper age cutoff of 80 years for trial eligibility may partially explain the absence of a relationship of OH with age in this study.
The observation that OH incidence was unrelated to assigned BP treatment group intensity and the lower rates of OH on follow-up than at baseline are important and reassuring findings. Although assignment to intensive BP control did not significantly affect the composite primary endpoint in ACCORD BP, there was a substantial reduction in the annual stroke rate in this group, 0.32% versus 0.52% per year, p=0.01 (10). The current data are therefore reassuring that a decision to treat individuals with diabetes who are at high stroke risk to a lower BP target would not increase the risk of OH. Furthermore, a similar lack of increased OH incidence in nondiabetic individuals at high cardiovascular risk treated to a systolic BP target of 120 mmHg was shown in the SPRINT trial, in which mortality and major cardiovascular events were reduced by 25% in persons assigned to this BP target (22).
Our findings that OH was independently associated with increased risk of total mortality and heart failure deaths or hospitalizations have also been observed in prior studies (7–9, 17,18). Increased total mortality in persons with OH was seen in men 71–93 years old in the Honolulu Heart Program (8), in a general population in Malmo, Sweden (7), and in the Hypertension Detection and Follow-up Program (9). In the latter study, the association of OH with accelerated mortality was particularly strong in the diabetic subset. A higher rate of heart failure was observed in persons with OH in the predominantly middle-aged Atherosclerosis Risk In Communities (ARIC) cohort (23) and the Cardiovascular Health Study participants of mean age 74 years (24). In contrast with ARIC (5), we did not observe an increased stroke risk in persons with OH despite the high risk profile of ACCORD participants. Although OH was associated with a higher rate of coronary events in prior epidemiologic studies (7–9), the rate of MI was not significantly increased with OH in ACCORD.
The current findings should be interpreted in light of the study’s strengths and limitations. To our knowledge, the ACCORD cohort represents the largest sample of individuals with diabetes in whom the prevalence and incidence of OH have been reported. The standardized protocol using multiple measurements of seated and standing BP and the detailed participant follow-up and adjudication of cardiovascular events and cause of death are additional study attributes. However, the use of seated BP may have resulted in lower rates of OH than had supine BP been used prior to standing. Nevertheless, seated BP is more representative of clinical practice. A second limitation was the implementation of standing BP measurements well after the trial began, resulting in a smaller sample with orthostatic BP determination at all 3 time points. Finally, the absence of individuals aged 80 years and older was based on the parent ACCORD Glycemia Trial exclusion criteria, thus eliminating the age group with the highest expected rate of OH.
Supplementary Material
Perspectives.
The prevalence and incidence of consensus OH in this ACCORD cohort with both hypertension and T2DM were relatively high. However, incident OH rate decreased over time and was not related to aggressive versus standard BP target assignment. Factors associated with OH in this setting were white race, female sex, current smoking, higher seated systolic BP and hemoglobin A1c, and use of alpha blockers, beta blockers, and insulin. Occurrence of OH was an independent marker for total mortality and heart failure death or hospitalization but not nonfatal myocardial infarction, stroke, cardiovascular death or their composite. The current findings are reassuring that a decision to treat high risk individuals with diabetes to a lower BP target should not increase the risk of OH.
Novelty and Significance.
What is New
In patients 40–79 years old with type 2 diabetes and hypertension, Incident OH is not related to age or to intensive versus standard SBP targets. OH independently predicts total mortality and heart failure death/hospitalization.
What is Relevant
Treatment of elevated SBP to an intensive target of 120mmHg does not increase OH. However, OH remains a marker for adverse prognosis in this population.
Summary
In patients with type 2 diabetes and hypertension, OH was common, not associated with intensive versus standard BP treatment goals, and predicted increased mortality and heart failure events.
Acknowledgments
Source of Funding: The ACCORD Blood Pressure trial was funded by a contract from the National Heart, Lung, and Blood Institute.
Footnotes
Disclaimer: The content of this manuscript is solely the responsibility of the authors and does not necessarily reflect the views of the National Institutes of Health or the Department of Health and Human Services or the Department of Veterans Affairs.
Disclosures: Gregory Evans is supported by an institutional grant from AstraZeneca; Dr. Cushman receives institutional grant support from Merck and Eli Lilly and is an unpaid consultant to Takeda. All other authors have no disclosures to report.
References
- 1.Consensus Statement. The definition of orthostatic hypotension, pure autonomic failure and multiple system atrophy. Clin Autonomic Res. 1996;6:125–126. doi: 10.1007/BF02291236. [DOI] [PubMed] [Google Scholar]
- 2.Matthias CJ. Orthostatic hypotension: causes, mechanisms and influencing factors. Neurology. 1995;45(suppl 5):S6–S11. [PubMed] [Google Scholar]
- 3.Wu JS, Lu FH, Yang YC, Chang CJ. Postural hypotension and postural dizziness in patients with non-insulin dependent diabetes. Arch Int Med. 1999;159:1350–1356. doi: 10.1001/archinte.159.12.1350. [DOI] [PubMed] [Google Scholar]
- 4.Mader SL. Orthostatic hypotension. Med Clin North Am. 1989;73:1337–1349. doi: 10.1016/s0025-7125(16)30602-2. [DOI] [PubMed] [Google Scholar]
- 5.Eigenbrodt ML, Rose KM, Cooper DJ, Arnett DK, Smith R, Jones D. Orthostatic hypotension as a risk factor for stroke. The Atherosclerosis Risk in Communities (ARIC ) Study 1987–1996. Stroke. 2000;3:2307–2313. doi: 10.1161/01.str.31.10.2307. [DOI] [PubMed] [Google Scholar]
- 6.Rose KM, Tyroler HA, Mardo CJ, Arnett DK, Light KC, Rosamond W, Sharett AR, Szklo M. Orthostatic hypotension and the incidence of coronary heart disease: The Atherosclerosis Risk in Communities Study. Am J Hypertension. 2000;13:571–578. doi: 10.1016/s0895-7061(99)00257-5. [DOI] [PubMed] [Google Scholar]
- 7.Federowski A, Stavenow, Hedblad B, Berglund G, Nilsson PM, Melander O. Orthostatic hypotension predicts all-cause mortality and coronary events in middle-aged individuals (The Malmo Preventive Project) Eur Heart J. 2010;31:85–91. doi: 10.1093/eurheartj/ehp329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Masaki KH, Schatz IJ, Burchfield CM, Sharp DS, Chico D, Foley D, Curb JD. Orthostatic hypotension predicts mortality in elderly men. The Honolulu Heart Program. Circulation. 1998;98:2290–2295. doi: 10.1161/01.cir.98.21.2290. [DOI] [PubMed] [Google Scholar]
- 9.Davis BR, Langford HG, Blaufox MD, Curb JD, Polk BF, Schulman NB. The association of postural changes in systolic blood pressure and mortality in persons with hypertension: the Hypertension Detection and Follow-up Program experience. Circulation. 1987;75:340–346. doi: 10.1161/01.cir.75.2.340. [DOI] [PubMed] [Google Scholar]
- 10.The ACCORD Study Group. Effects of intensive blood-pressure control in type 2 diabetes mellitus. N Engl J Med. 2010;362:1575–1585. doi: 10.1056/NEJMoa1001286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.ACCORD Study Group. Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: design and methods. Am J Cardiol. 2007;99(Suppl 2):1i–33i. doi: 10.1016/j.amjcard.2007.03.003. [DOI] [PubMed] [Google Scholar]
- 12.The Action to Control Cardiovascular risk in Diabetes Study Group. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358:2545–2559. doi: 10.1056/NEJMoa0802743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.The ACCORD Study Group. Effects of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med. 2010;362:1563–1574. doi: 10.1056/NEJMoa1001282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Finucane C, O’Connell MDL, Fan CW, Savva GM, Soraghan CJ, Nolan H, Cronin H, Kenny RA. Age-related normative changes in orthostatic blood pressure in a large population study. Findings from the Irish Longitudinal Study on Ageing (TILDA) Circulation. 2014;130:1780–1789. doi: 10.1161/CIRCULATIONAHA.114.009831. [DOI] [PubMed] [Google Scholar]
- 15.Applegate WB, Davis BR, Black HR, Smith WM, Miller ST, Burlando AJ. Prevalence of postural hypotension at baseline in the Systolic Hypertension in the Elderly Program (SHEP) cohort. J Am Geriatr Soc. 1991;39:1057–1064. doi: 10.1111/j.1532-5415.1991.tb02869.x. [DOI] [PubMed] [Google Scholar]
- 16.Verhaeverbeke I, Mets T. Drug-induced orthostatic hypotension in the elderly. Avoiding its onset. Drug Safety. 1995;17:105–118. doi: 10.2165/00002018-199717020-00003. [DOI] [PubMed] [Google Scholar]
- 17.Spallone V, Ziegler D, Freeman R, Bernardi L, Frontoni S, Pop-Busui R, Stevens M, Kempler P, Hilsted J, Tesfaye S, Low P, Valensi P on behalf of the Toronto Consensus Panel on Diabetic Neuropathy. Cardiovascular autonomic neuropathy in diabetes: clinical impact, assessment, diagnosis, and management. Diabetes Metab Res Rev. 2011;27:639–653. doi: 10.1002/dmrr.1239. [DOI] [PubMed] [Google Scholar]
- 18.Steinberg HO, Brechtel G, Johnson A, Fineberg N, Baron AD. Insulin-mediated skeletal muscle vasodilation is nitric oxide dependent. A novel action of insulin to increase nitric oxide release. J Clin Invest. 1994 Sep;94:1172–1179. doi: 10.1172/JCI117433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cleland SJ, Petrie JR, Ueda S, Elliott HL, Connell JMC. Insulin-mediated vasodilation and glucose uptake are functionally linked in humans. Hypertension. 1999;33:554–558. doi: 10.1161/01.hyp.33.1.554. [DOI] [PubMed] [Google Scholar]
- 20.Cheng YC, Vyas A, Hymen E, Perlmuter LC. Gender differences in orthostatic hypotension. Am J Med Sci. 2011;342:221–225. doi: 10.1097/MAJ.0b013e318208752b. [DOI] [PubMed] [Google Scholar]
- 21.Strogatz DS, Keenan NL, Barnett EM, Wagner EH. Correlates of postural hypotension in a community sample of elderly blacks and whites. J Am Geriatr Soc. 1991;39:562–566. doi: 10.1111/j.1532-5415.1991.tb03593.x. [DOI] [PubMed] [Google Scholar]
- 22.SPRINT Research Group. Wright JT, Jr, Williamson JD, Whelton PK, Snyder JK, Sink KM, Rocco MV, Reboussin DM, Rahman M, Oparil S, Lewis CE, Kimmel PL, Johnson KC, Goff DC, Jr, Fine LJ, Cutler JA, Cushman WC, Cheung AK, Ambrosius WT. A randomized trial of intensive versus standard blood-pressure control. N Engl J Med. 2015;373:2103–2116. doi: 10.1056/NEJMoa1511939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Jones CD, Loehr L, Franceschini N, Rosamond WD, Chang PP, Shahar E, Cooper DJ, Rose KM. Orthostatic hypotension as a risk factor for incident heart failure. The Atherosclerosis Risk in Communities study. Hypertension. 2012;59:913–918. doi: 10.1161/HYPERTENSIONAHA.111.188151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Alagiakrishnan K, Patel K, Desai RV, Ahmed MB, Fonarow GC, Forman DE, White M, Aban IB, Love TE, Aronow WS, Allman RM, Anker SD, Ahmed A. Orthostatic hypotension and incident heart failure in commmunity-dwelling older adults. J Gerontol Biol Sci Med Sci. 2014;69:223–230. doi: 10.1093/gerona/glt086. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.