Blood pressure response with fixed-dose combination therapy: comparing hydrochlorothiazide with amlodipine through individual-level meta-analysis

Development of a machine learning-based model for predicting individual responses to antihypertensive treatments

BACKGROUND AND AIMS

Personalized antihypertensive drug selection is essential for optimizing hypertension management. The study aimed to develop a machine learning (ML) model to predict individual blood pressure (BP) responses to different antihypertensive medications.

METHODS AND RESULTS

We used data from a pragmatic, cluster-randomized trial on hypertension management in China. Each patient's multiple visit records were included, and two consecutive visits were paired as the index and subsequent visits. The least absolute shrinkage and selection operator method was used to select index visit variables for predicting subsequent BP. The dataset was randomly divided into training and test sets in a 7:3 ratio. Model performance was evaluated using mean absolute error (MAE) and R-square in the test set. A total of 19,013 hypertension management visit records (6282 patients) were included. The mean age of the study population was 63.9 years, and 2657 (42.3%) were females. A total of 12 phenotypical features (age, sex, smoking within seven days, body mass index, waist circumference, index visit systolic BP, diastolic BP, heart rate, comorbidities of diabetes, dyslipidemia, coronary heart disease, and stroke), together with currently taking any prescribed antihypertensive medication regimens and visits time interval were selected to build the model. The Extreme Gradient Boost model performed best among all candidate algorithms, with an MAE of 8.57 mmHg and an R2 = 0.28 in the test set.

CONCLUSION

The ML techniques exhibit significant potential for predicting individual responses to antihypertensive treatments, thereby aiding clinicians in achieving optimal BP control safely and efficiently.

TRIAL REGISTRATION

ClinicalTrials.gov, NCT03636334. Registered July 3, 2018, https://clinicaltrials.gov/study/NCT03636334.

Practical application of the ATOM study: Treatment efficacy of antihypertensive drugs in monotherapy or combination (ATOM metaanalysis according to PRISMA statement); tables for the use of antihypertensive drugs in monotherapy or combination

BACKGROUND

The response to antihypertensive drugs is predictable. The absence of precise prescription recommendations to treat arterial hypertension (HT) lead to use drugs unable to reduce blood pressure (BP) to target values.We published ATOM study, in which we found significant differences in the ability to reduce BP between the different drugs.The objective of the study was to determine the expected decrease in blood pressure with the use of commercialized doses of the drugs commonly used in the treatment of HT in clinical practice, to avoid the use of drugs or combinations that even with the best response, are unable to obtain the necessary BP decrease to reach the goal.

METHODS

The analysis was based on the results of the ATOM study. To convert the mean doses of the different drugs and combinations in commercialized doses, the conclusions of the study by Law et al have been applied.

RESULTS

Based on the results, two tables were drawn, one for systolic BP and the other for diastolic BP, where the doses of the different drugs and combinations are classified according to the BP decrease that can be expected from them. In order to favor the use of the tables in clinical practice, the different drugs have been grouped in intervals of 10 millimeters of mercury (mmHg) for the decrease of the systolic BP and of 5 mmHg for the diastolic BP.

CONCLUSIONS

Recommendations for the use of antihypertensive treatments should not be limited to pharmacological families. They should also consider differences between drugs or specific combinations. From the data of the ATOM study we have implemented tables that express the effect of the drugs commonly used in clinical practice and that should allow the clinicians to choose with care the treatment to use.

Albuminuria and masked uncontrolled hypertension in chronic kidney disease

Background

Masked uncontrolled hypertension (MUCH) is associated with greater target organ damage such as left ventricular hypertrophy, increased arterial stiffness and albuminuria. Whether MUCH independently associates with greater cardiovascular end-organ damage or kidney damage is unclear. The objective of this study was to assess the strength of the relationship of MUCH (awake ambulatory blood pressure ≥135/85 mmHg and clinic blood pressure <140/90 mmHg) with target organ damage.

Methods

In a cross-sectional study at a veterans' administration medical center, clinically normotensive veterans without chronic kidney disease (CKD) (n = 29) and 287 patients with CKD and controlled hypertension (CH, n = 193), MUCH (n = 67) and uncontrolled hypertension (UCH, n = 27) had evaluation of target organ damage. Target organ damage was measured by echocardiography [left ventricular mass index (LVMI)], arterial ultrasonography [aortic pulse wave velocity (PWV)] and 24-h urine collection [albuminuria (urine albumin to creatinine ratio)] in all participants.

Results

Compared to that of controls, LVMI was higher by 21.8 g/m2 (CI, 4.0-39.7 g/m2) in CH, 27.9 (CI, 8-47.8) in MUCH and 39.5 (CI, 15.7-63.2) in UCH (P < 0.01 for group differences, P < 0.01 for linear trend). Although differences persisted after adjustment for age, sex and race, they lost significance after adjustments for cardiovascular risk factors and their treatment. Compared to that of controls, PWV was different among CH, MUCH and UCH (P = 0.04 for group differences, P = 0.02 for linear trend). However, differences lost significance after adjustments for age, sex and race. Compared to that of controls, log2 UACR was higher by 2.40 mg/mg (CI, 1.28-3.52) in CH, 4.94 (CI, 3.70-6.18) in MUCH and 6.01 (CI, 4.49-7.53) in UCH (P < 0.0001 for group difference, P < 0.0001 for linear trend). Differences persisted after adjustment for age, sex and race, cardiovascular risk factors and their treatment and cardiovascular disease (P < 0.0001 for group difference, P < 0.0001 for linear trend).

Conclusions

MUCH is more strongly related to albuminuria compared with cardiovascular damage as assessed by left ventricular mass and PWV. A graded and an independent relationship of blood pressure classification status with albuminuria is consistent with the hypothesis that renal mechanisms may be more important than cardiovascular disease in mediating the pathogenesis of MUCH.

Fixed-dose combination therapy for the prevention of atherosclerotic cardiovascular diseases

BACKGROUND

Atherosclerotic cardiovascular disease (ASCVD) is the leading cause of death and disability worldwide, yet ASCVD risk factor control and secondary prevention rates remain low. A fixed-dose combination of blood pressure- and cholesterol-lowering and antiplatelet treatments into a single pill, or polypill, has been proposed as one strategy to reduce the global burden of ASCVD.

OBJECTIVES

To determine the effect of fixed-dose combination therapy on all-cause mortality, fatal and non-fatal ASCVD events, and adverse events. We also sought to determine the effect of fixed-dose combination therapy on blood pressure, lipids, adherence, discontinuation rates, health-related quality of life, and costs.

SEARCH METHODS

We updated our previous searches in September 2016 of CENTRAL, MEDLINE, Embase, ISI Web of Science, and DARE, HTA, and HEED. We also searched two clinical trials registers in September 2016. We used no language restrictions.

SELECTION CRITERIA

We included randomised controlled trials of a fixed-dose combination therapy including at least one blood pressure-lowering and one lipid-lowering component versus usual care, placebo, or an active drug comparator for any treatment duration in adults 18 years old or older, with no restrictions on presence or absence of pre-existing ASCVD.

DATA COLLECTION AND ANALYSIS

Three review authors independently selected studies for inclusion and extracted the data for this update. We evaluated risk of bias using the Cochrane 'Risk of bias' assessment tool. We calculated risk ratios (RR) for dichotomous data and mean differences (MD) for continuous data with 95% confidence intervals (CI) using fixed-effect models when heterogeneity was low (I2 < 50%) and random-effects models when heterogeneity was high (I2 ≥ 50%). We used the GRADE approach to evaluate the quality of evidence.

MAIN RESULTS

In the initial review, we identified nine randomised controlled trials with a total of 7047 participants and four additional trials (n = 2012 participants; mean age range 62 to 63 years; 30% to 37% women) were included in this update. Eight of the 13 trials evaluated the effects of fixed-dose combination (FDC) therapy in populations without prevalent ASCVD, and the median follow-up ranged from six weeks to 23 months. More recent trials were generally larger with longer follow-up and lower risk of bias. The main risk of bias was related to lack of blinding of participants and personnel, which was inherent to the intervention. Compared with the comparator groups (placebo, usual care, or active drug comparator), the effects of the fixed-dose combination treatment on mortality (FDC = 1.0% versus control = 1.0%, RR 1.10, 95% CI 0.64 to 1.89,  I2 = 0%, 5 studies, N = 5300) and fatal and non-fatal ASCVD events (FDC = 4.7% versus control = 3.7%, RR 1.26, 95% CI 0.95 to 1.66, I2 = 0%, 6 studies, N = 4517) were uncertain (low-quality evidence). The low event rates for these outcomes and indirectness of evidence for comparing fixed-dose combination to usual care versus individual drugs suggest that these results should be viewed with caution. Adverse events were common in both the intervention (32%) and comparator (27%) groups, with participants randomised to fixed-dose combination therapy being 16% (RR 1.16, 95% CI 1.09 to 1.25, 11 studies, 6906 participants, moderate-quality evidence) more likely to report an adverse event . The mean differences in systolic blood pressure between the intervention and control arms was -6.34 mmHg (95% CI -9.03 to -3.64, 13 trials, 7638 participants, moderate-quality evidence). The mean differences (95% CI) in total and LDL cholesterol between the intervention and control arms were -0.61 mmol/L (95% CI -0.88 to -0.35, 11 trials, 6565 participants, low-quality evidence) and -0.70 mmol/L (95% CI -0.98 to -0.41, 12 trials, 7153 participants, moderate-quality evidence), respectively. There was a high degree of statistical heterogeneity in comparisons of blood pressure and lipids (I2 ≥ 80% for all) that could not be explained, so these results should be viewed with caution. Fixed-dose combination therapy improved adherence to a multidrug strategy by 44% (26% to 65%) compared with usual care (4 trials, 3835 participants, moderate-quality evidence).

AUTHORS' CONCLUSIONS

The effects of fixed-dose combination therapy on all-cause mortality or ASCVD events are uncertain. A limited number of trials reported these outcomes, and the included trials were primarily designed to observe changes in ASCVD risk factor levels rather than clinical events, which may partially explain the observed differences in risk factors that were not translated into differences in clinical outcomes among the included trials. Fixed-dose combination therapy is associated with modest increases in adverse events compared with placebo, active comparator, or usual care but may be associated with improved adherence to a multidrug regimen. Ongoing, longer-term trials of fixed-dose combination therapy will help demonstrate whether short-term changes in risk factors might be maintained and lead to expected differences in clinical events based on these changes.

Treatment efficacy of anti-hypertensive drugs in monotherapy or combination: ATOM systematic review and meta-analysis of randomized clinical trials according to PRISMA statement

BACKGROUND

The relative efficacy of antihypertensive drugs/combinations is not well known. Identifying the most effective ones and the patients' characteristics associated with best performance of the drugs will improve management of hypertensive patients.

OBJECTIVE

To assess the blood pressure (BP) reduction attributed to antihypertensive drugs and identify characteristics associated with BP decrease.

DATA SOURCES

MEDLINE, Cochrane Central Register of Controlled Trials from inception through July 2012 and selected papers.

STUDY ELIGIBILITY CRITERIA

Double-blind, randomized clinical trials whose main result was the reduction in BP by antihypertensive treatment, with study population ≥50 or ≥25 if the study was a crossover, follow-up of at least 8 weeks, and available required data.

STUDY APPRAISAL AND SYNTHESIS METHODS

Study data were independently extracted by multiple observers and introduced in an electronic database. Inconsistencies were resolved by discussion and referral back to the original articles. Meta-analysis was performed according to PRISMA statement and using a Bayesian framework.

MAIN OUTCOME(S) AND MEASURE(S)

Mean decrease in systolic (SBP) and diastolic blood pressure (DBP) achieved by each drug or combination.

RESULTS

Two hundred eight trials including 94,305 patients were identified. In monotherapy, most drugs achieved 10 to 15 mm Hg SBP and 8 to 10 mm Hg DBP decreases.Olmesartan/amlodipine, olmesartan/hydrochlorothiazide, felodipine/metoprolol, and valsartan/hydrochlorothiazide were the combinations leading to the greatest mean SBP reductions (>20 mm Hg). Female sex and body mass index >25 kg/m were associated with more pronounced SBP and DBP reductions, whereas Afro-American ethnicity was associated with BP reductions smaller than the median. Results were adjusted by study duration, cardiovascular disease, and diabetes mellitus. Still, the estimation was performed using the mean administered doses, which do not exactly match those of the available drug formats.

LIMITATIONS

Data corresponded to those obtained in each of the included trials; the analysis of the combinations was limited to the most recent ones; estimations were performed using the mean administered doses.

CONCLUSIONS AND IMPLICATIONS

Certain drug combinations achieve BP reductions ranging from 20 to 25/10 to 15 mm Hg. Sex, ethnicity, and obesity are associated with antihypertensive response. This information can contribute to better selection of the antihypertensive drug, depending on the magnitude of pretreatment BP elevation. Guidelines should be revised.

Comparison of olmesartan combined with a calcium channel blocker or a diuretic in elderly hypertensive patients (COLM Study): safety and tolerability

The cardiovascular effects of combined therapy with the angiotensin receptor blocker (olmesartan) and a dihydropyridine calcium channel blocker (CCB) or a diuretic were compared in high-risk elderly Japanese hypertensive patients by performing a randomized, open label, blinded-endpoint study of morbidity and mortality (the COLM study). Here we report the results obtained with respect to safety and tolerability. High-risk hypertensive patients aged 65–84 years were enrolled and were randomized to receive olmesartan combined with either a CCB (amlodipine or azelnidipine) or a low-dose diuretic for at least 3 years. The primary endpoint was a composite of fatal and non fatal cardiovascular events, whereas adverse events (AEs) and the percentage of patients who discontinued the allocated treatment were evaluated as secondary endpoints. A total of 5141 patients were randomized. Both combination regimens achieved a similar reduction of cardiovascular morbidity and mortality. The incidences of AEs, serious AEs, drug-related serious AEs and discontinuation due to serious AEs were lower in the olmesartan plus CCB group than in the olmesartan plus diuretic group. Serum levels of uric acid and creatinine were significantly higher in the olmesartan plus diuretic group than in the olmesartan plus CCB group. Olmesartan combined with a CCB was significantly superior to olmesartan plus a diuretic with regard to the frequency of AEs and discontinuation of treatment.