Intensive Monitoring of Adherence to Treatment Helps to Identify “True” Resistant Hypertension


  • Walnéia Aparecida De Souza PhD,

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author
  • 1 Maricene Sabha PharmD, PhD,

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author
  • 2 Fabrício De Faveri Favero PharmD, MSc,

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author
  • 2 Gun Bergsten-Mendes MD, PhD,

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author
  • 2 Juan Carlos Yugar-Toledo MD, PhD,

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author
  • and 2 Heitor Moreno Jr, MD, PhD 2

    1. From the Department of Pharmacy, Federal University of Alfenas, Alfenas, MG, Brazil ; 1 and the Section of Cardiovascular Pharmacology and Hypertension, Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas (UNICAMP), Campinas, SP, Brazil2
    Search for more papers by this author

  • Drs De Souza and Sabha contributed equally to this work.

Walnéia Aparecida de Souza, PhD, Department of Pharmacy, Federal University of Alfenas, 91 Manoel José de Almeida Street, Colinas Park, 37130.000 Alfenas, MG, Brazil


Intensive monitoring of adherence in patients with uncontrolled hypertension was evaluated over a 6-month period. After that period, only patients well characterized as having resistant hypertension were followed for 12 months. The goal of this study was to evaluate whether adherence to a drug regimen helps to identify patients with resistant hypertension. Forty-four hypertensive patients resistant to a 3-drug regimen (average blood pressure [BP] mm Hg, mean ± standard deviation) were studied prospectively. Each patient was followed for a 12-month period. Adherence to treatment was evaluated through self-report, applying Morisky’s questionnaire and the pill count method. Ambulatory BP monitoring and office BP measures were performed. By pill count, 63.6% of the patients were adherent to treatment at the start of the survey and 94% at the end, although 59% of the patients still did not reach normal BP levels. We found that non-adherence was not associated with resistance to antihypertensive treatment. Therefore, after investigation, we concluded that patients who presented with uncontrolled arterial BP may be truly resistant hypertensive to treatment.

Hypertension is a major risk factor in the development of cardiovascular disease and poses a significant public health problem.1 In Brazil, the prevalence of arterial hypertension is not precisely known. Population-based surveys in Brazilian cities showed an arterial hypertension (≥140/90 mm Hg) prevalence of 22.3% to 43.9%; these studies were restricted to only a few areas.2 Although the exact incidence of resistant hypertension (RH) is not established, estimates derived from recent outcome studies including the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), the Valsartan Antihypertensive Long-Term Use Evaluation (VALUE), and the Controlled-Onset Verapamil Investigation of Cardiovascular End Points (CONVINCE) emphasize that this condition may be more common than previously believed.3

Uncontrolled hypertension is not synonymous with RH.4–6 Both the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7)7 and the 2007 European Society of Hypertension/European Society of Cardiology Guidelines for Management of Arterial Hypertension8 define RH as systolic or diastolic blood pressure (BP) that remains above goal levels in spite of the concurrent use of 3 antihypertensive agents of different classes, including a diuretic. True RH must be distinguished from apparent RH (difficult-to-treat hypertension), in which an important cause may be medication nonadherence.6,9

Reports on RH are based largely on observational data from specialty clinics. Evaluation to differentiate difficult-to-treat hypertension from true RH should include consideration of ambulatory BP monitoring to rule out white coat hypertension; chronic kidney disease might be ruled out. This is one of the most common associated factors in RH.6 Secondary causes (including exogenous substances),9 inadequate antihypertensive drug regimens (physician inertia), associated diseases, and use of non-steroidal anti-inflammatory drugs10,11 must also be considered. According to Burnier and colleagues4 poor adherence to antihypertensive treatment has been reported to be one of the most common causes of uncontrolled hypertension in primary care and is estimated to contribute to the lack of adequate BP control in more than two-thirds of patients with hypertension.

The goal of this study was to evaluate whether drug regimen adherence helps to identify patients with RH.


Setting, Patients, and Study Design

Forty-four patients with RH (11 men and 33 women) were included in this study. The patients were referred to the Cardiovascular Pharmacology and Hypertension Clinic at the University Hospital, State University of Campinas (UNICAMP), São Paulo, Brazil, for evaluation and treatment of RH. All patients had been treated for at least 6 months at the hospital with the same antihypertensive medication during 4 weeks that preceded the study. To be included in the study patients had to be hypertensive (systolic BP [SBP] >140 mm Hg and/or diastolic BP [DBP] >90 mm Hg), according to the mean of 2 consecutive visits separated by at least 35 to 40 days despite their prescribed triple antihypertensive drug regimen. Secondary hypertension, sleep apnea, insulin resistance, myocardial infarction, and cerebral vascular accident were exclusion factors.

Data Collection

Data collection was performed in 2 phases. The visit was performed every 35 to 40 days at the Clinical Hospital, where arterial pressure measurements were performed and patients were oriented to pharmacologic and nonpharmacologic treatments. All participants gave written informed consent and the local ethics committee approved the study protocol.

The first phase consisted of follow-up visits during a 6-month period to evaluate and optimize drug usage. Clinical interview, demographic, and anthropometric characteristics (sex, age, ethnicity, weight, and height), cardiovascular risk factors (smoking, obesity, physical inactivity, dyslipidemia, diabetes, family history of premature coronary heart disease) from all patients were recorded. Laboratory evaluation including fasting glycemia, insulin resistance, serum creatinine, electrolytes, triglycerides, total cholesterol, and high-density lipoprotein cholesterol was performed every 6 months or whenever necessary.

In the second phase of the study, adherence was evaluated by a standard questionnaire.12 Patients were invited to come back to the hospital each 35 to 40 days, when the antihypertensive drug and/or its doses were adjusted in an effort to reduce the BP to <140 mm Hg (systolic) and 90 mm Hg (diastolic). The antihypertensive medication and/or doses were adjusted individually by our group of specialists in hypertension.

BP Measurement

BP was measured at regular clinic appointments by a trained physician with patients in the sitting position using a calibrated mercury sphygmomanometer. Three BP measurements were performed at each visit and the average of the last 2 readings was used for the analysis. Mean BP measurements at the start of the first phase (baseline measurements) and at the start and at the end of the second one were compared. Mean systolic and diastolic office BP values <140/90 mm Hg7 indicated controlled hypertension.

BP measurement was also obtained through ambulatory BP monitoring (ABPM). All patients underwent ABPM (Spacelabs 90207; Spacelabs Healthcare, Issaquah, WA) for 24 hours. ABPM was performed at the start (baseline measurement) and at the beginning and the end of the adherence study. A reading was taken every 15 minutes throughout the day and every 30 minutes at night. Mean BP measurements between the baseline measurement and the beginning and the end of the adherence study were compared. Mean systolic and diastolic BPs were considered (24-hour). Values of <130/80 mm Hg were indicative of controlled hypertension. Nocturnal and awake arterial pressures were also analyzed.7


Adherence was measured by pill count each time patients returned for their scheduled appointments. Patients with pill counts ≥80% for each prescribed medicine were classified as adherent.13,14 In order to follow the protocol, each patient received the medications and was asked to bring the empty bottles for a pill count. After each bottle return, adhesion to treatment was calculated.

Adherence was also assessed using Morisky’s questionnaire,12 a 4 “yes or no” question test that evaluates attitudes regarding medication intake. Only patients who scored 4 points were considered highly adherent to treatment. Patients answered the test just before starting the second phase and at the end of the study.

Statistical Analysis

The data were collected and stored into the Epi-Info version 6.04d (Centers for Disease Control and Prevention, Atlanta, GA) database. Descriptive data were expressed as mean ± standard deviation (SD) and percentage. To compare arterial pressure variables (ABPM and office BP) between different times (beginning and end of the study) we used analysis of variance (ANOVA) for repeated measures, with rank transformation due to the nonexistence of data normality. When the difference was significant, multiple comparison tests were performed to identify the differences. McNemar test was used to compare antihypertensive medications and adherence to treatment by Morisky’s questionnaire. Social-demographic profile, lifestyle modifications, clinical characteristics of controlled and uncontrolled hypertensive patients, and comorbidities were compared using chi-square or Fisher exact test. Glycated hemoglobin, basal insulin, and fasting glycemia were analyzed using t test. Adherence to treatment measured by pill count was compared using regression linear test. All statistical analysis was done using SPSS for Windows, version 10.0.7 (SPSS Inc, Chicago, IL) and SAS for Windows, version 8.02 (SAS Institute, Cary, NC). A P<.05 indicated statistical significance.


Demographic and Clinical Characteristics

Table I shows the characteristics of all patients included in this study. Participants were mainly women (aged 20–64 years). The major presence of women is explained by the fact that Brazilian women are more committed to appointments than men. The mean ± SD age of the 44 patients was 48.6±9.0 years; 93% were younger than 60 years. The remaining characteristics were similar among controlled and uncontrolled groups, except for age. Among obese patients, none presented with insulin resistance. Several comorbidities were found and the most common in the controlled group were diabetes (22.2%) and psychiatric (depression) problems, accounting for 27.8% in the controlled group and 19.2% in the uncontrolled group.

Table I.   Distribution of Hypertensive Patients, Controlled and Uncontrolled, According to Social, Structural, and Clinical Characteristics and Lifestyle
Controlled (n=18)Uncontrolled (n=26)Total (N=44)
No.%No.%No.%P Value
  1. Abbreviations: CHD, coronary heart disease; SD, standard deviation. aP<.05. bAbout US $200 (monthly).

Age, ya
 18–50 527.82076.92556.8.002
 >501372.2 623.11943.2
 Male 316.7 934.61227.3.303
Marital status
 Not married 738.9 726.91431.8
 Unemployed 316.7 415.4715.9.707
 Employed 211.1 519.2715.9
 Retired 844.4 726.91534.1
 Employment disability 527.81038.51534.1
Educational level
 ≤4 y  527.8 311.5818.2.240
 >4 y 1372.22388.53681.8
Family income (minimum wage)b
 >3 527.8 830.81329.5
Hypertension duration
 1–5 y 527.8 415.4920.5.092
 6–10 y 211.1 0024.5
 >10 y1161.12284.63375.0
Length of treatment
 1–5 y 527.8 415.4920.5.066
 6–10 y 316.7 27.7511.4
 >10 y1055.52076.93068.2
Alcohol consumption 15.6 27.736.81.0
Sedentary life1372.21765.43068.2.632
Cigarette smoker 211.1 415.4613.6.252
Family history of premature CHD181002284.64090.9.133
Body mass index
 Normal 316.7 27.7511.4.367
 Overweight 844.4 830.81636.4
 Obese 738.91661.52352.3
Laboratory values, mean (±SD)
 Basal insulin13.7 (7.41)12.9 (5.9).372
 Glycated hemoglobin6.0 (1.1)5.6 (1.0).282
 Fasting glucose101.3 (18.4)94.9 (13.8).194

Long-Term Follow-Up

The mean BP of the 44 patients was consistent with uncontrolled arterial hypertension. The mean ± SD office systolic and diastolic BP and the systolic and diastolic BP measured with 24-hour ABPM are shown in Figure 1. Adjustments in therapy resulted in further decreases of mean office BP and ABPM. Office BPs decreased from baseline to the end of study from 163/102.7±18.8/12.6 mm Hg to 144.5/93.6±13.2/10.2 mm Hg (ANOVA: P=.001 for systolic and diastolic BP). ABPM decreased between baseline and the end of the study from 141.9/90.5±18.4/15.2 mm Hg to 131.7/84.1± 18.4/12.6 mm Hg (ANOVA: P=.001 for systolic and P=.004 for diastolic BP), respectively.

Figure 1.

 (A) Blood pressure (BP) of the patients: time course ambulatory BP monitoring (ABPM) in patients with treatment of resistant hypertension. Drug regimens were changed during the study in order to achieve the therapeutic goal. Results are mean ± standard deviation (SD); N=44; bP<.05 vs baseline. (B) BP of the patients: time course of office BPs in patients with treatment of resistant hypertension. Drug regimens were changed during the study in order to achieve the therapeutic goal. Results are mean ± SD; N=44. bP<.05 vs baseline. cP<.05 vs start of adherence study.

Figure 2 shows the number of patients achieving the therapeutic goal or improving office SBP and DBP during the 12 months of follow-up. Therapeutic optimization and adherence monitoring resulted in normalization of office DBP in 31.8% and of SBP in 43.2%. In addition, significant improvements in BP control (ie, a decrease in office SBP >20 mm Hg or a reduction in DBP of >10 mm Hg) were found in 36.4% and 34.1% of the patients, respectively. ABPM showed normalization of DBP in 40.9% of patients and of SBP in 52.3% and a decrease in SBP >20 mm Hg or a reduction of DBP of >10 mm Hg in 27.3% and 36.4% of the patients, respectively. However, when the percentage of patients achieving the therapeutic goal was analyzed based on office BP and ABPM, 18 (40.9%) patients were controlled, but 26 (59.1%) patients remained resistant.

Figure 2.

 Monitoring of blood pressure (BP) results and achievement of therapeutic goal. The upper panel shows ambulatory BP monitoring (ABPM) results for systolic BP (SBP) and diastolic BP (DBP) and the lower panel shows office BP.

Effect of Adherence Monitoring on BP Control

During the adherence study, in the last 6 months, a small significant decrease in BP measured by ABPM was observed, from 133/85±18.7/13.6 mm Hg at the start of the study to 131.7/84.1±18.4/ 12.6 mm Hg at the end of the study. This change was not statistically significant. Office BP showed a significant reduction, falling from 158/100±24.4/13.2 mm Hg to 144.5/93.6 ± 13.2/10.2 mm Hg (ANOVA: P=.001 for SBP and P=.004 for DBP). The results are shown in Figure 1. There was a significant decrease of arterial pressure in both the controlled and uncontrolled group 6 months after the therapeutic optimization. The BP decrease in the controlled patients was smaller than in the uncontrolled patients. Results are shown in Table II.

Table II.   Ambulatory Blood Pressure at Baseline, Start of Adherence, and at the End of the Study
ABPM, mm HgGroupBaselineStart of AdherenceEnd of StudyP Value
  1. Abbreviation: ABPM, ambulatory blood pressure monitoring. Values are expressed as mean ± standard deviation. P<.05 for controlled and uncontrolled patients.

24-Hour systolicControlled132.6±14127.1±11124.8±6<.001
Day systolicControlled133.3±15130.9±10120.8±7<.001
Night systolicControlled128.7±13123.2±14113.2±7<.001
24-Hour diastolicControlled82.9±1077.1±976.3±6<.004
Day diastolicControlled83.5±1077.0±1077.5±7<.001
Night diastolicControlled76.1±1068.4±10 69.2±7<.001

Antihypertensive Treatment

Patients took a mean ± SD of 5.4±1.08 medications daily. During optimization of the therapeutic regimen there were alterations in controlled and uncontrolled patients’ medications. The most frequently prescribed medications were diuretics (100%), calcium channel blockers (72.7%), and angiotensin II receptor antagonists (65.8%), as shown in Table III. Sixty-two percent of the participants reported that they had been taking some antihypertensive medications for more than 10 years. Some increase in the number of prescriptions for hypertension therapy was noted at the end of the study (3.34±0.68 at the start vs 3.54±0.79 at the end); however, there were no significant differences (P=.09).

Table III.   Antihypertensive Medications Prescribed at Baseline and at the End of the Study
Antihypertensive DrugsBaseline %End %P Value
  1. Abbreviation: ACE, angiotensin-converting enzyme. aP<.05 considered statistically significant.

 Thiazide + nonthiazide4.54.5
β-Adrenergic antagonists61.356.8.46
Calcium channel blockers63.872.7.41
ACE inhibitorsa43.125.002
Angiotensin II receptor antagonistsa2565.8.001
 Half-life (12 h)15.954.5
 Half-life (>12 h)9.111.3
Centrally acting antiadrenergic agentsa20.511.3.001

Patients taking antihypertensive medications in fixed-dose combinations increased from 7 (15.9%) at the beginning to 22 (50.0%) at the end of the study. In both study periods, the 2 most frequently prescribed fixed-dose combinations were losartan plus hydrochlorothiazide and pindolol plus clopamide. In the second step, 2 fixed-dose combinations of angiotensin-converting enzyme inhibitors plus calcium channel blockers were prescribed: amlodipine plus enalapril (2 patients) and ramipril plus amlodipine (2 patients). Fifty-nine percent of the patients who remained with uncontrolled levels of BP at the end of the study were prescribed 3.69±0.93 antihypertensive medications, compared with 3.22±0.73 medications prescribed to 18 (40.9%) patients with controlled BP; the difference was not significant (P=.19).


Adherence evaluation via pill count showed that at the first return, 28 patients (63.6%) had used ≥80% of every prescribed medication and were classified as adherent. At the second appointment, the number of adherent patients increased to 35 (79.5%), at the third visit the number increased to 40 (91%), and at the end of the study 42 (95.5%) patients were adherent by this criterion. Controlled patients had an adherence variation from 83.3% at the beginning to 88.9% at the end of the study. Uncontrolled patients presented a variation of 50% at the beginning and 100% at the end of the study. Linear correlation was moderate for controlled patients (r=.635) and high for uncontrolled patients (r=.989). Results are shown in Figure 3.

Figure 3.

 Adhesion results according to each hypertensive patients’ return, after their controlled and uncontrolled identification, to the hospital.

Morisky’s questionnaire12 indicated that only 16 (36.4%) patients scored 4 points at the start of the adherence study and were classified as adherent. At the end of the study, this figure rose to 30 patients (68.2%; P=.001). There was a trend toward better adherence among controlled patients than among uncontrolled ones (P=.07).


In this study, during a year of pharmacotherapeutic intervention, there was a significant BP reduction in office BP and ABPM (P<.05) (Figure 1). This may have been due to changes in therapeutic regimen during the study. However, BP remained uncontrolled in 59% of the patients. Among the controlled patients, 4 were classified as having white coat hypertension based on ABPM. The diagnosis of true/white coat RH is important to avoid over-treatment. All RH patients should undergo ABPM to rule out cases with office RH but controlled BP at home.15,16

Since the definition of RH is dependant on the fact that patients adhere to their multidrug antihypertensive regimens, medication nonadherence is not a cause of true RH, but rather a form of difficult-to-treat RH, oftentimes with multiple causes.6 Although physicians may claim to recognize nonadherence in their patients, studies have shown that physician recognition of nonadherence is quite poor and no better than simple random guessing.4

There are several ways to monitor adherence. In our study, an indirect method and subjective measurements were utilized respectively. These methods are cheap and simple to use. Neither indirect nor subjective measures, however, offer definitive proof that a medication has been administered.17 Electronic monitors have shown excellent results, being considered the gold standard,4 but this method is difficult to carry out and is expensive.

Through adherence analysis, according to Morisky’s questionnaire,12 we found that there was a significant improvement in adherence at the end of the study in patients with controlled BP (P=.07). According to the pill count method, it was responsible for an increase in adherence from 63.6% to 95.4%. When controlled and uncontrolled patients were compared, however, it was observed that treatment adherence in controlled patients at the beginning of the study was 83.3% and at the end was 88.9%, and for uncontrolled patients it was 50% at the beginning of the study and 100% at the end. Controlled patients already presented a high adherence rate at the start of the study (r=.635). On the other hand, while uncontrolled patients presented a low adherence rate at the start of the study, on follow-up there was a significant improvement with a high and positive correlation (r=.958). Adherence to therapy apparently had not made a great difference in BP control in these patients.

Attention to adherence to therapy should be the first measure in the identification of RH.18

Our results indicate that controlled patients were mostly adherent during the whole study. Uncontrolled patients presented a high adhesion rate at the third and fourth visits (84.5% and 100%, respectively) but still had not achieved BP control. This suggests that nonadherence was not the responsible factor for resistance to treatment.

In our study some steps were taken to simplify the therapeutic regimen. Inconvenient or complex dosage regimens may contribute to low control rates19 and result in failure of continuity of medical care and patient frustration.1 A choice was made for the use of longer-lasting antihypertension drugs (Table III): more atenolol (half-life, 5–8 hours) and less propranolol (half-life, 3–5 hours), more enalapril (half-life, 11 hours) and less captopril (half-life, 2 hours), and more amlodipine (half-life, 39 hours) and less nifedipine (half-life, 1.8 hours). Antihypertensive drugs with longer half-lives remain effective for a longer time, allowing for a more comfortable drug schedule, besides maintaining BP control during occasional drug omissions.20 The option to use specific antihypertensive drugs was based on the availability of such drugs at the out-patient clinic.

The prescription of drugs that are well tolerated such as angiotensin II receptor antagonists will help overcome one cause for nonadherence. In our study there was a significant increase in the use of angiotensin II receptor antagonists (25% at baseline and 65.8% at the end of the study; P=.001). Garg and colleagues21 reported in their study of RH that angiotensin receptor blockers are better tolerated than other classes of antihypertensive agents.

In addition to the above medication adjustments, changes in diuretic therapy—including increasing or changing the diuretic—were among the most important remedial actions taken. The importance of volume and sodium regulation in the successful treatment of hypertension, and the benefits of diuretic therapy, are well known.

Another strategy in antihypertensive drug therapy, namely the prescription of fixed-dose combinations, was also employed to enhance adherence by reducing patient pill burden. At the end of the study, patients with prescriptions of fixed-dose combinations had risen from 15.9% to 50%. The inherent advantage of fixed-dose combination therapy lies in its improving adherence because fewer pills are required.22 Another goal of the fixed-dose combination therapy is to at least partially neutralize adverse effects from any single medication.

Obesity may play a role in RH.6,16,23 In our study, there were a large number of obese patients, but these patients did not have elevated basal insulin or high glycated hemoglobin levels. Weight reduction is a particularly important adjunct to drug therapy in overweight patients with RH. Progressively higher doses of antihypertensive drugs may be required to control BP as body mass index increases.15


Although elevated adherence rates were reached (95.4%), the majority of the patients who participated in this study did not achieve ideal BP levels. However, if any degree of BP reduction in hypertensive patients can decrease the number of cardiovascular events, the same may be true in resistant hypertensive patients. We conclude that monitoring of adherence helps to distinguish poor compliance but that good adherence does not necessarily improve outcome in these resistant patients. ABPM can be utilized to distinguish white coat hypertension from true RH. The history, examination, and laboratory work should focus on medication and illicit drug use and clues for identifiable causes of RH to tailor a further workup. Careful evaluation for identifiable causes should be performed in appropriately selected patients. Our study suggests that patients who present with uncontrolled arterial BP may have truly resistant hypertension.

Disclosures:  This study was supported by the Coordination of Improvement of Higher Education (CAPES), the National Counsel of Technological and Scientific Development (CNPq), the State of São Paulo Research Foundation (FAPESP), Research and Teaching Fund Support (FAEP) from the State University of Campinas (São Paulo, Brazil), and the Federal University of Alfenas (UNIFAL) (Alfenas, Brazil).