Abstract
Background
Assessing adherence to antihypertensive medication remains a major challenge for healthcare professionals. The objective of this study was to compare the evaluation of adherence to antihypertensive medication between direct and indirect methods, and to identify variables associated with nonadherence.
Methods
A cross-sectional study was conducted with 253 persons with hypertension followed at a hypertension unit of a tertiary teaching hospital. Participants aged over 18 years and under antihypertensive drug treatment were included. Persons with secondary hypertension, end-stage chronic kidney disease, or pregnancy were excluded. Biopsychosocial, anthropometric, lifestyle, comorbidity, and medication data were collected. Indirect assessment of treatment adherence was performed using the 4-item Morisky–Green Medication Adherence Scale, while direct assessment was based on urine analysis using high-performance liquid chromatography coupled with mass spectrometry. Variables with P < 0.20 in the bivariate analysis were included in the multiple logistic regression model. The significance level was set at 5%. Cohen’s Kappa coefficient (CKC) was used to assess agreement between the methods.
Results
The sample consisted predominantly of women (61.7%), white individuals (63.2%), and married participants (52.8%), with a mean age of 65 ± 13.3 years and median education of 12.0 (7–12) years. Participants were using an average of 3.2 ± 1.3 antihypertensive drugs, and 69.2% had controlled blood pressure (BP). The prevalence of adherence according to the direct method was 32.4%, while 90.1% were adherent according to the indirect method. Agreement between the direct and indirect methods was low (CKC = 0.014). Variables independently associated with nonadherence according to the direct method were regular physical activity (odds ratio [OR], 0.370; 95% confidence interval [CI], 0.158–0.853), proteinuria (OR, 2.406; 95% CI, 1.066–5.751), number of antihypertensive medications (OR, 1.570; 95% CI, 1.127–2.225), and systolic BP (OR, 1.051; 95% CI, 1.010–1.097). Indirect adherence assessment was associated only with the presence of common mental disorders (OR, 1.163; 95% CI, 1.035–1.307).
Conclusions
The prevalence of adherence to antihypertensive drug treatment assessed by the indirect method was substantially higher than that observed with the direct method. The results were inconsistent between the 2 approaches. Moreover, the direct method identified a broader range of variables associated with nonadherence.
Keywords: Hypertension, Adherence, Control, High-performance liquid chromatography, Nursing
BACKGROUND
Arterial hypertension is one of the main modifiable causes of premature death worldwide [1,2,3,4]. The positive impact of controlled blood pressure (BP) in reducing cardiovascular events and target-organ damage is well established [5]. However, poor adherence to antihypertensive treatment remains a major barrier to treatment success. Assessing adherence and understanding the factors that limit its effectiveness continue to be significant challenges for healthcare professionals.
Adherence to antihypertensive medication is inadequate in most countries [6], with nonadherence rates ranging from 8.2% to 74.9%, depending on the assessment method used [7]. Evaluating treatment adherence in hypertension is not a simple task, and this heterogeneity in prevalence estimates reflects the complexity of the process. There are various methods of assessment and different cutoff points, and to date, no gold standard method exists that encompasses the multiple dimensions involved in the adherence process.
Assessment of treatment adherence can be performed using direct or indirect methods. Direct assessment includes biological analysis of blood or urine and supervised medication intake, whereas indirect assessment may involve appointment attendance, pill counts, patient self-report, physician perception, or the use of validated self-report instruments [8]. Indirect methods, particularly those using self-report instruments, are the most commonly employed in clinical practice. However, biological analysis is increasingly recommended as a more objective method, although it also presents important limitations [9].
The World Health Organization (WHO) recommends a multimethod approach as the state-of-the-art for assessing adherence behaviour [10]. However, studies that employ both direct and indirect methods remain scarce. Therefore, the main objective of this study was to compare the prevalence of adherence to antihypertensive medication in a cohort of persons with hypertension followed in an outpatient service using both direct and indirect assessment methods, and to identify factors associated with nonadherence.
METHODS
Study design and sampling
A prospective, cross-sectional study was carried out with 253 persons with hypertension in the hypertension unit of a tertiary teaching hospital in São Paulo, Brazil. The sample size calculation was based on an adherence rate of 50% [10], a maximum margin of error of 5%, and a 95% confidence level. Persons with hypertension aged over 18 years, who had been followed in the service for at least six months and were receiving antihypertensive medication, were included. Patients with secondary hypertension, end-stage chronic kidney disease (CKD) (glomerular filtration rate < 15 mL/min/1.73 m2), or pregnancy were excluded (Fig. 1). The study protocol was approved by the Research Ethics Committee of the University of São Paulo, Nursing School and University of São Paulo, Medical School (protocol No. 2,831,454 and No. 3,003,912), and all participants provided written informed consent. The original dataset is part of a database previously collected for a master’s thesis at the University of São Paulo, School of Nursing (Mayra Cristina da Luz Pádua Guimarães, 2019).
Fig. 1. Flowchart of the sampling process, exclusions, and inclusions related to the study.
Measurements
Data collection was conducted between April and July 2019. Eligible patients were recruited during their scheduled medical appointments. After providing written informed consent, participants underwent an interview using a structured instrument that included information on the study variables, supplemented by a review of medical records for laboratory results and prescribed medications. Urine samples for the analysis of antihypertensive drug metabolites were collected on the same day as the interview.
Outcomes
Direct assessment of adherence was performed using biological material, employing an analytical method for qualitative analysis of antihypertensive compounds identified in urine via high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS/MS). Mass spectrometry is a quantitative and qualitative analytical technique that provides information on the molecular weight and structure of molecules through the analysis of compounds in biological samples. HPLC-MS/MS is highly sensitive, rapid, and selective, and can be used to detect commonly prescribed antihypertensive drugs [11]. Urine samples for analysis were collected on the day of the interview in the designated data collection office, between 07:00 and 10:00 a.m. The analytical method involved: adjusting the mass spectrometer detector parameters to optimize the response or signal of the compound in “Multiple Reaction Monitoring” (MRM) mode; developing a chromatographic method to separate the analyzed compounds and minimize matrix effects; and detecting the compounds through extraction efficiency tests in urine and sensitivity assessment.
Results were expressed qualitatively as “positive” (compounds detected above the Method Limit of Detection [MLOD]) or “negative” (below the MLOD). Nonadherence to antihypertensive treatment was defined as the presence of less than 80% of the prescribed antihypertensive medications in urine. Participants whose urine analysis confirmed the presence of more than 80% of the prescribed medications were classified as adherent [12,13].
Indirect assessment of adherence was performed using the 4-item Morisky–Green–Levine Medication Adherence Scale [14]. This scale consists of four questions designed to determine adherence based on participants’ yes/no responses. The questions are: “Do you ever forget to take your medicine?”; “Are you sometimes careless about taking your medicine?”; “When you feel better, do you sometimes stop taking your medicine?”; and “Sometimes, if you feel worse when you take your medicine, do you stop taking it?” Participants were considered adherent to treatment when all responses were negative (total score = 0) and nonadherent if the score was ≥ 1.
Exposure and covariates
To characterize the study sample, the following data were collected: biosocial and psychoemotional variables, risk factors, lifestyle habits, anthropometric measurements, and laboratory tests (lipid profile, fasting glucose, glycated hemoglobin [HbA1c], urea, creatinine, estimated glomerular filtration rate [eGFR] calculated using the Modification of Diet in Renal Disease [MDRD] equation, and proteinuria). The presence of comorbidities and prescribed medications was obtained directly from electronic medical records. Socioeconomic classification was determined using the “Brazilian Economic Classification Criteria” [15], which estimates the purchasing power of individuals and urban households. Lifestyle habits were assessed as follows: smoking status (current smoker, former smoker, or never smoker); alcohol consumption using the Alcohol Use Disorders Identification Test (AUDIT) [16], which classifies intake according to score as low-risk, at-risk, harmful use, or probable dependence; and physical activity using the short version of the International Physical Activity Questionnaire (IPAQ) [17], classifying participants as very active, active, irregularly active, or sedentary. Psychoemotional characteristics were assessed using the Self-Reporting Questionnaire-20 (SRQ-20) [18], which screens for common nonpsychotic mental disorders through 20 dichotomous items, with a cutoff point of 6 for men and 8 for women. All instruments used are validated for Brazilian Portuguese.
Casual BP was measured in the interview room using a validated semi-automatic device [19]. Arm circumference was determined to select the appropriate cuff for each patient, and measurements were performed according to national guideline recommendations [20,21]. Three readings were taken with a minimum interval of 1 minute between them, and the average of the last 2 measurements was used for analysis.
Data analysis
Analyses were performed using R software version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria), and a P-value < 0.05 was considered statistically significant.
Descriptive data are presented as absolute (n) and relative (%) frequencies for categorical variables, while continuous variables are presented as means ± standard deviation or medians with interquartile ranges (IQR). Bivariate analyses were conducted to assess the relationship between the dependent variable—adherence to antihypertensive treatment, evaluated by direct and indirect methods—and independent variables. The χ2 or Fisher’s exact tests were used for categorical variables, and Wilcoxon–Mann–Whitney, Brunner–Munzel, or t-tests were applied for continuous variables. To compare the performance of the indirect method with the direct method in identifying adherence to antihypertensive treatment, diagnostic performance measures were calculated considering the direct method as a reference for comparative purposes. Cohen’s kappa coefficient, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were therefore estimated to assess agreement between methods.
Additionally, for comparative purposes, the direct method was treated as a reference approach; however, it was not considered a gold standard, given the absence of a definitive method for adherence assessment.
For multivariate analysis, a logistic regression model was used for both final models. Independent variables with P < 0.20 in the bivariate analysis were included. Candidate predictors of nonadherence according to the direct assessment method were: age, education, monthly income, history of stroke, HbA1c, physical activity, alcohol consumption, body fat, body weight, glucose, urea, creatinine, eGFR (MDRD), proteinuria, diabetes mellitus, heart failure, arrhythmia, number of prescribed medications, number of antihypertensive drugs, systolic BP, and diastolic BP. Candidate predictors of nonadherence according to the indirect assessment method were: age, socioeconomic status, presence of common mental disorders, waist-to-hip ratio, CKD, history of stroke, HbA1c, and proteinuria.
Multicollinearity among independent variables was assessed using the variance inflation factor, with values below 5 considered indicative of absence of significant multicollinearity.
RESULTS
Adherence to antihypertensive treatment according to direct and indirect methods
The prevalence of adherence to antihypertensive treatment was 32.4% according to the direct method and 90.1% according to the indirect method, with 29.6% of participants adherent by both methods. In addition, 2.8% of participants were classified as adherent only by the direct method, 60.5% only by the indirect method, and 7.1% were nonadherent according to both methods (Fig. 2).
Fig. 2. Venn diagram representing the overlap between direct adherence and indirect adherence to antihypertensive treatment among people with hypertension.
Sociodemographic characteristics of the study population
The majority of the sample were women (61.7%), of white ethnicity (63.2%), married (52.8%), with a mean age in the 6th decade (65 ± 13.3 years), and just under half (47.0%) belonged to the lower socioeconomic stratum C2. The median monthly income was US$ 341 (range: US$179–US$538), and the median years of education was 12.0 (IQR: 7–12) (Table 1).
Table 1. Biosocial characteristics of persons with hypertension according to adherence to antihypertensive treatment assessed by direct and indirect methods.
| Variables | Adherence assessment methods | |||||||
|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Total (n = 253) | ||||||
| Yes (n = 82) | No (n = 171) | P-value | Yes (n = 228) | No (n = 25) | P-value | |||
| Sex | 0.667a | 0.493a | ||||||
| Female | 49 (31.4) | 107 (68.6) | 139 (89.1) | 17 (10.9) | 156 (61.7) | |||
| Male | 33 (34.0) | 64 (66.0) | 89 (91.7) | 8 (8.2) | 97 (38.3) | |||
| Age (yr) | 63.3 ± 14.3 | 65.8 ± 12.8 | 0.135b | 65.3 ± 13.3 | 61.6 ± 13.0 | 0.159b | 65.0 ± 13.3 | |
| Ethnicity | 0.968a | 0.934a | ||||||
| White | 52 (32.5) | 108 (67.5) | 144 (90.0) | 16 (10.0) | 160 (63.2) | |||
| Non-White | 30 (32.3) | 63 (67.7) | 84 (90.3) | 9 (9.7) | 93 (36.8) | |||
| Marital Status | 0.813a | 0.848a | ||||||
| Single | 15 (36.6) | 26 (63.4) | 37 (90.2) | 4 (9.8) | 41 (16.7) | |||
| Married | 41 (30.8) | 92 (69.2) | 120 (90.2) | 13 (9.7) | 133 (52.8) | |||
| Divorced | 8 (28.6) | 20 (71.4) | 24 (85.7) | 4 (14.3) | 28 (11.1) | |||
| Widowed | 18 (36.0) | 32 (64.0) | 46 (92.0) | 4 (8.0) | 51 (19.9) | |||
| Education (yr) | 12.0 [9–12] | 12.0 [5–12] | 0.146d | 12.0 [8–12] | 12.0 [5–12] | 0.675d | 12.0 [7–12] | |
| Individual monthly income (US$) | 372 [223–372] | 333 [185–558] | 0.144d | 344 [185–103] | 372 [223–483] | 0.570d | 341 [179–538] | |
| Socioeconomic status | 0.860c | 0.142c | ||||||
| A1 | 1 (100.0) | 0 (0.0) | 1 (100.0) | 0 (0.0) | 1 (0.4) | |||
| B1 | 1 (100.0) | 0 (0.0) | 0 (0.0) | 1 (100.0) | 1 (0.4) | |||
| B2 | 20 (64.5) | 11 (35.5) | 26 (83.9) | 5 (16.1) | 31 (12.2) | |||
| C1 | 40 (62.5) | 24 (37.5) | 58 (90.6) | 6 (9.4) | 64 (25.3) | |||
| C2 | 82 (68.9) | 37 (31.1) | 112 (94.1) | 7 (5.6) | 119 (47.0) | |||
| D-E | 27 (72.9) | 10 (27.0) | 30 (81.1) | 7 (18.9) | 37 (14.6) | |||
Values are presented as mean ± standard deviation, median [interquartile range], or number (%).
aPearson’s χ2 test; bStudent’s t-test; cFisher’s exact test; dWilcoxon-Mann-Whitney test.
Lifestyle characteristics and association with adherence
Regarding lifestyle habits, most participants (56.1%) reported being non-smokers, and just under half reported low-risk alcohol consumption. More than one-third (39.9%) engaged in physical activity irregularly. The body mass index was elevated, with 38.3% classified as overweight and 43.1% as obese. In the univariate analysis, adherence to antihypertensive treatment assessed by the direct method was associated with sedentary behavior and alcohol consumption. Sedentary behavior was more prevalent among nonadherent participants compared to adherent participants (76.3% vs. 23.7%, P < 0.05), and nonadherent individuals reported a higher proportion of alcohol consumption (57.8% vs. 42.2%) compared with adherent participants (Table 2).
Table 2. Lifestyle and habit characteristics of persons with hypertension according to adherence to antihypertensive treatment assessed by direct and indirect methods.
| Variables | Adherence assessment methods | |||||||
|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Total (n = 253) | ||||||
| Yes (n = 82) | No (n = 171) | P-value | Yes (n = 228) | No (n = 25) | P-value | |||
| Smoking | 0.823a | 0.893a | ||||||
| Yes | 5 (26.3) | 14 (73.7) | 17 (89.5) | 2 (10.5) | 19 (7.5) | |||
| No | 46 (32.4) | 96 (67.6) | 127 (89.4) | 15 (10.6) | 142 (56.1) | |||
| Ex-smoker | 31 (33.7) | 61 (66.3) | 84 (91.3) | 8 (8.7) | 92 (36.4) | |||
| Physical activity | 0.022 a | 0.333a | ||||||
| Active | 25 (45.4) | 30 (54.6) | 51 (92.7) | 4 (7.3) | 55 (21.7) | |||
| Irregularly active | 34 (33.7) | 67 (66.3) | 93 (92.1) | 8 (7.9) | 101 (39.9) | |||
| Sedentary | 23 (23.7) | 74 (76.3) | 84 (86.6) | 13 (13.4) | 97 (38.3) | |||
| Do you currently drink alcohol? | 0.007 a | 0.211a | ||||||
| Yes | 43 (42.2) | 59 (57.8) | 89 (87.3) | 13 (12.7) | 102 (40.3) | |||
| No | 39 (25.8) | 112 (74.2) | 139 (92.0) | 12 (8.0) | 151 (59.7) | |||
| AUDIT final score | ||||||||
| Low-risk consumption | 80 (32.0) | 170 (68.0) | 226 (90.4) | 24 (9.6) | 250 (98.8) | |||
| Risky consumption | 2 (66.7) | 1 (33.3) | 2 (66.7) | 1 (33.3) | 3 (1.2) | |||
| Body mass index (kg/m2) | 28.8 [25.8–32.2] | 28.8 [25.9–33.1] | 0.572b | 28.6 [25.9–32.7] | 30.7 [24.9–33.1] | 0.561b | 28.8 [25.8–32.8] | |
| Nutritional status | ||||||||
| Underweight | 1 (50.0) | 1 (50.0) | 2 (100.0) | 0 (0.0) | 2 (0.8) | |||
| Eutrophic | 15 (33.3) | 30 (66.7) | 38 (84.4) | 7 (15.6) | 45 (17.8) | |||
| Overweight | 33 (34.0) | 64 (67.0) | 93 (95.9) | 4 (4.1) | 97 (38.3) | |||
| Obesity | 33 (30.3) | 76 (69.7) | 95 (87.2) | 14 (12.8) | 109 (43.1) | |||
Values are presented as median [interquartile range] or number (%). Bold styled P-values indicate statistically significant.
AUDIT, Alcohol Use Disorders Identification Test.
aPearson’s χ2 test; bWilcoxon-Mann-Whitney test.
Renal function and association with treatment adherence, comorbidities and clinical characteristics
Renal function assessment showed altered eGFR in 31.6% of hypertensive participants, elevated urea levels in 17.8%, elevated creatinine in 38.3%, and the presence of proteinuria in 28.1%. In univariate analysis, these parameters were associated (P < 0.05) only with adherence evaluated by the direct method. Nonadherent hypertensive participants, compared to adherent participants, had higher urea levels (37.0 ± 17.0 vs. 32.5 ± 16.0), lower eGFR (67.5 ± 22.8 vs. 74.3 ± 22.7 mL/min/1.73 m2), and a higher proportion of individuals with proteinuria (80.3% vs. 19.7%) (Table 3).
Table 3. Laboratory markers of persons with hypertension according to adherence to antihypertensive treatment assessed by direct and indirect methods.
| Variables | Adherence assessment methods | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Total (n = 253) | |||||||
| Yes (n = 82) | No (n = 171) | P-value | Yes (n = 228) | No (n = 25) | P-value | ||||
| Fasting blood glucose (mg/dL) | 99 [92.2–114.7] | 105 [94–122] | 0.162a | 104 [93–118.2] | 96 [93–115] | 0.226a | 104 [93–118] | ||
| Normal | 43 (39.5) | 66 (60.5) | 92 (84.4) | 17 (15.6) | 109 (43.1) | ||||
| Increased risk for diabetes | 26 (26.8) | 71 (73.2) | 92 (94.8) | 5 (5.2) | 97 (38.3) | ||||
| Diabetes mellitus | 13 (27.7) | 34 (72.3) | 44 (93.6) | 3 (6.4) | 47 (15.6) | ||||
| HbA1c | 5.9 [5.5–6.5] | 5.9 [5.5–6.8] | 0.371a | 5.9 [5.5–6.7] | 5.7 [5.3–6.1] | 0.082a | 5.9 [5.5–6.7] | ||
| Normal | 32 (36.0) | 57 (64.0) | 77 (86.5) | 12 (13.5) | 89 (35.2) | ||||
| Pré–diabetes | 31 (35.6) | 56 (64.4) | 80 (91.9) | 7 (8.1) | 87 (34.4) | ||||
| Diabetes mellitus | 19 (24.7) | 58 (75.3) | 71 (92.2) | 6 (7.8) | 77 (30.4) | ||||
| Triglycerides (mg/dL) | 120.0 [86.7–178.7] | 126 [87.5–168.5] | 0.829a | 125 [86–177.2] | 120 [89–162] | 0.786a | 124 [86–171] | ||
| Desirable level | 53 (33.0) | 108 (67.0) | 148 (91.9) | 13 (8.1) | 161 (63.6) | ||||
| Total cholesterol (mg/dL) | 175.5 [152.2–203.5] | 174 [150–191.5] | 0.531a | 174.5 [149–194] | 174 [161–188] | 0.694a | 174 [151–194] | ||
| Desirable level | 54 (30.7) | 122 (69.3) | 157 (89.2) | 19 (10.8) | 176 (69.6) | ||||
| LDL-C (mg/dL) | 98.5 [74–117] | 93 [74–112.5] | 0.390a | 93 [72–114] | 98 [83–113] | 0.287a | 94 [74–113] | ||
| Great level | 41 (28.7) | 102 (71.3) | 132 (92.3) | 11 (7.7) | 143 (56.5) | ||||
| HDL-C (mg/dL) | 53.0 [45–64] | 55.0 [42–65] | 0.955a | 53.5 [43–66] | 55 [43–61] | 0.733a | 54 [43–65] | ||
| Desirable level | 71 (34.5) | 135 (65.5) | 185 (89.8) | 21 (10.2) | 206 (81.4) | ||||
| Serum urea (mg/dL) | 32.5 [27–43] | 37.0 [29–46] | 0.049 a | 36 [27–44.2] | 36 [31–54] | 0.274a | 36 [28–45] | ||
| Normal | 70 (33.6) | 138 (66.4) | 191 (91.8) | 17 (8.2) | 208 (82.2) | ||||
| Altered | 12 (26.7) | 33 (73.4) | 37 (82.2) | 8 (17.8) | 45 (17.8) | ||||
| Creatinine (mg/dL) | 0.88 [0.7–1.1] | 0.93 [0.7–1.2] | 0.112a | 0.9 [0.7–1.1] | 1.0 [0.8–1.2] | 0.289a | 0.91 [0.7–1.1] | ||
| Normal | 55 (35.3) | 101 (64.7) | 141 (90.4) | 15 (9.6) | 156 (61.7) | ||||
| Altered | 27 (27.8) | 70 (72.2) | 87 (89.7) | 10 (10.3) | 97 (38.3) | ||||
| eGFR MDRD (mL/min/1.73 m2) | 74.3 ± 22.7 | 67.5 ± 22.8 | 0.028 c | 69.8 ± 22.1 | 69.5 ± 25.7 | 0.952c | 69.8 ± 22.1 | ||
| MDRD (mL/min/1.73 m2) | |||||||||
| < 60 | 21 (26.2) | 59 (73.8) | 71 (88.7) | 9 (11.3) | 80 (31.6) | ||||
| 45 to 59 | 13 (27.7) | 34 (72.3) | 43 (91.5) | 4 (8.5) | 47 (18.6) | ||||
| 30 to 44 | 5 (20.8) | 19 (79.2) | 20 (83.3) | 4 (16.7) | 24 (9.5) | ||||
| 15 to 29 | 3 (33.3) | 6 (66.7) | 8 (88.9) | 1 (11.1) | 9 (3.5) | ||||
| Proteinuria | 0.007 b | 0.163b | |||||||
| Absent | 68 (37.4) | 114 (62.6) | 167 (91.8) | 15 (8.2) | 182 (71.9) | ||||
| Present | 14 (19.7) | 57 (80.3) | 61 (85.9) | 10 (14.1) | 71 (28.1) | ||||
Values are presented as mean ± standard deviation, median [interquartile range], or number (%). Bold styled P-values indicate statistically significant.
HbA1c, glycated hemoglobin; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; MDRD, Modification of Diet in Renal Disease.
aWilcoxon-Mann-Whitney test; bPearson’s χ2 test; cStudent’s t-test.
The most prevalent comorbidity was dyslipidemia (71.5%), followed by diabetes (40.7%), with lower frequencies of CKD (19.0%), neoplasia (15.4%), and heart failure (13.4%). Adherence assessed by the direct method showed that nonadherent participants had a higher prevalence of heart failure (82.3% vs. 17.7%, P < 0.05) and arrhythmia (87.0% vs. 13.0%, P < 0.05). The median duration of hypertension diagnosis was 2 decades (20.0 ± 20.0 years), and 23.7% of participants were classified as having common mental disorders. Adherence assessment using the indirect method showed a significantly higher presence of common mental disorders (P = 0.003) among nonadherent participants compared with adherent participants (5.44 ± 3.4 vs. 3.44 ± 3.5) (Table 4).
Table 4. Personal health history of persons with hypertension according to adherence to antihypertensive treatment assessed by direct and indirect methods.
| Variables | Adherence assessment methods | |||||||
|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Total (n = 253) | ||||||
| Yes (n = 82) | No (n = 171) | P-value | Yes (n = 228) | No (n = 25) | P-value | |||
| Dyslipidemia | 121 (66.9) | 60 (33.1) | 0.691a | 162 (89.5) | 19 (10.5) | 0.603a | 181 (71.5) | |
| Diabetes mellitus | 27 (26.2) | 76 (73.8) | 0.082a | 94 (91.3) | 9 (8.7) | 0.614a | 103 (40.7) | |
| CKD | 14 (29.1) | 34 (70.8) | 0.594a | 40 (83.3) | 8 (16.7) | 0.081a | 48 (19.0) | |
| Neoplasia | 10 (25.6) | 29 (74.4) | 0.327a | 35 (89.7) | 4 (10.3) | 0.932a | 39 (15.4) | |
| Heart failure | 6 (17.7) | 28 (82.3) | 0.048 a | 30 (88.2) | 4 (11.8) | 0.693a | 34 (13.4) | |
| Arrhythmia | 3 (13.0) | 20 (87.0) | 0.038 a | 21 (91.3) | 2 (8.7) | 0.842a | 23 (9.1) | |
| Stroke | 6 (27.3) | 16 (72.7) | 0.591a | 18 (81.8) | 4 (18.2) | 0.173a | 22 (8.7) | |
| Acute myocardial infarction | 5 (23.8) | 16 (76.2) | 0.380a | 19 (90.5) | 2 (9.5) | 0.954a | 21 (8.3) | |
| Depression | 4 (25.0) | 12 (75.0) | 0.514a | 14 (87.5) | 2 (12.5) | 0.717a | 16 (6.3) | |
| Peripheral vascular disease | 4 (33.3) | 8 (66.7) | 0.944a | 11 (91.7) | 1 (8.3) | 0.854a | 12 (4.7) | |
| Chronic airway disease | 3 (33.3) | 6 (66.7) | 0.952b | 6 (66.7) | 3 (33.3) | 0.558b | 9 (3.6) | |
| Digestive diseases | 3 (37.5) | 5 (62.5) | 0.755b | 8 (100.0) | 0 (0.0) | 1.000b | 8 (3.2) | |
| Time since diagnosis of arterial hypertension (yr) | 20 [10.2–30] | 20.0 [10–30] | 0.579c | 20 [10–30] | 20 [10–30] | 0.796c | 20 [10–30] | |
| SRQ–20 total score | 3.3 ± 3.4 | 3.7 ± 3.6 | 0.417d | 3.4 ± 3.5 | 3.4 ± 3.5 | 0.003 d | 3.6 ± 3.6 | |
| Common mental disorders | ||||||||
| Yes | 20 (33.3) | 40 (66.7) | 50 (83.3) | 10 (16.7) | 60 (23.7) | |||
| No | 62 (32.1) | 131 (67.8) | 178 (92.2) | 15 (7.8) | 193 (76.3) | |||
Values are presented as mean ± standard deviation, median [interquartile range], or number (%). Bold styled P-values indicate statistically significant.
CKD, chronic kidney disease; SRQ-20, Self-Reporting Questionnaire-20.
aPearson’s χ2 test; bFisher’s exact test; cWilcoxon-Mann-Whitney test; dStudent’s t-test.
Antihypertensive treatment and medication use, BP control and association with adherence
The mean number of prescribed antihypertensive medications per patient was 3.2 ± 1.3 daily, and the median total number of medications was 7.0 (IQR: 5.0–9.0). The most commonly prescribed antihypertensive drugs were diuretics (78.3%), calcium channel blockers (68.8%), angiotensin II receptor blockers (61.7%), and beta-adrenergic blockers (50.2%). Adherence assessment using the direct method showed that nonadherent participants had higher numbers of antihypertensive medications and total prescribed medications compared with adherent participants (3.5 ± 1.4 vs. 2.6 ± 1.0, 7.0 [5–9] vs. 5.5 [4.2–7.7]; P < 0.05) (Table 5).
Table 5. Therapeutic classes of antihypertensive medications of persons with hypertension according to adherence to antihypertensive treatment assessed by direct and indirect methods.
| Variables | Adherence assessment methods | |||||||
|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Total (n = 253) | ||||||
| Yes (n = 82) | No (n = 171) | P-value | Yes (n = 228) | No (n = 25) | P-value | |||
| Pharmacological classes – antihypertensives | ||||||||
| Diuretic | 55 (27.8) | 143 (72.2) | 0.003 a | 181 (91.4) | 17 (8.6) | 0.191a | 198 (78.3) | |
| CCBs | 49 (28.1) | 125 (71.8) | 0.032 a | 155 (89.1) | 19 (10.1) | 0.412a | 174 (68.8) | |
| ARBs | 45 (28.9) | 111 (71.1) | 0.125a | 144 (92.3) | 12 (7.7) | 0.140a | 156 (61.7) | |
| BB adrenergic inhibitors | 35 (27.6) | 92 (72.4) | 0.099a | 115 (90.5) | 12 (9.5) | 0.817a | 127 (50.2) | |
| ACE inhibitor | 27 (35.0) | 50 (65.0) | 0.552a | 70 (90.9) | 7 (9.1) | 0.781a | 77 (30.4) | |
| Centrally acting agents | 4 (10.0) | 36 (90.0) | 0.001 a | 36 (90.0) | 4 (10.0) | 0.978a | 40 (15.8) | |
| Vasodilators | 2 (8.0) | 23 (92.0) | 0.006 a | 21 (84.0) | 4 (16.0) | 0.281a | 25 (10.0) | |
| Alpha-blockers | 1 (5.6) | 17 (94.4) | 0.012 a | 16 (88.9) | 2 (11.1) | 0.856a | 18 (7.1) | |
| Blood pressure control | 0.035 a | 0.894a | ||||||
| Yes | 64 (36.6) | 111 (63.4) | 158 (90.3) | 17 (9.7) | 175 (69.2) | |||
| No | 18 (23.0) | 60 (76.9) | 70 (89.7) | 8 (10.3) | 78 (30.8) | |||
| Number of antihypertensives prescribed | 2.6 ± 1.0 | 3.5 ± 1.4 | < 0.001 b | 3.2 ± 1.3 | 3.0 ± 1.2 | 0.555b | 3.2 ± 1.3 | |
| One | 10 (41.7) | 14 (58.3) | 23 (95.8) | 1 (4.2) | 24 (9.5) | |||
| Two to three | 53 (41.7) | 74 (58.3) | 112 (88.2) | 15 (11.8) | 108 (42.7) | |||
| Four or more | 19 (18.6) | 83 (81.4) | 93 (91.2) | 9 (8.8) | 121 (47.8) | |||
| Number of medications prescribed | 5.5 [4.2–7.7] | 7.0 [5–9] | 0.001 c | 7 [5–9] | 8 [4–9] | 0.526c | 7 [5–9] | |
| BP (mmHg) | ||||||||
| Systolic BP | 136.5 [130.5–139.5] | 139.0 [136–142.7] | 0.001 c | 138.5 [134–140.5] | 138.5 [131–145] | 0.916c | 138.5 [134.0–140.5] | |
| Diastolic BP | 80.5 [75.5–84.8] | 81.5 [77.5–86.5] | 0.116c | 81 [76.3–86] | 82.5 [89–87.5] | 0.210c | 81 [77.0–86.0] | |
Values are presented as mean ± standard deviation, median [interquartile range], or number (%). Bold styled P-values indicate statistically significant.
CCB, calcium channel blocker; ARB, angiotensin II receptor blocker; BB, beta-blocker; ACE, angiotensin-converting enzyme; BP, blood pressure.
aPearson χ2 test; bBrunner-Munzel test; cWilcoxon-Mann-Whitney test.
The majority of participants (69.2%) had controlled BP, and control was significantly associated (P < 0.05) with adherence assessed by the direct method. Nonadherent participants, compared with adherent participants, had a higher proportion of uncontrolled BP (76.9% vs. 23.0%) and a higher median systolic BP (139 [136–142.7] vs. 136.5 [130.5–139.5] mmHg) (Table 5).
Multivariate analysis of factors associated with nonadherence
Multivariate regression analysis showed that, according to the direct assessment method, the variables that remained significantly associated (P < 0.05) with nonadherence to antihypertensive medication were physical activity, proteinuria, number of prescribed antihypertensive drugs, and systolic BP. Physically active participants had a 63.0% lower likelihood of nonadherence compared with sedentary participants, while those with proteinuria had 2.4 times higher odds of nonadherence. Each additional antihypertensive medication increased the odds of nonadherence by 57.0%, and each mmHg increase in systolic BP increased the odds by 5.1%. For the indirect assessment method, only the presence of common mental disorders remained significantly associated (P < 0.05) with nonadherence to antihypertensive treatment. Each one-point increase in the total score of the SRQ-20 increased the odds of nonadherence by 16.3% (Table 6).
Table 6. Logistic regression model: variables independently associated with nonadherence to treatment according to direct and indirect assessment methods.
| Variables | OR | 95% CI | P-value | |||
|---|---|---|---|---|---|---|
| Adherence assessed by the direct method | ||||||
| Physical activity | ||||||
| Sedentary | 1 | |||||
| Irregularly active | 0.697 | 0.338–1.422 | 0.323 | |||
| Active | 0.370 | 0.158–0.853 | 0.020 | |||
| Proteinuria | ||||||
| Absent | 1 | |||||
| Present | 2.406 | 1.066–5.751 | 0.040 | |||
| Number of antihypertensive medications | 1.570 | 1.127–2.225 | 0.009 | |||
| Systolic blood pressure | ||||||
| Sedentary | 1.051 | 1.010–1.097 | 0.019 | |||
| Adherence assessed by the indirect method | ||||||
| Common mental disorders (SRQ-20) | 1.163 | 1.035–1.307 | 0.011 | |||
Bold styled P-values indicate statistically significant.
OR, odds ratio; CI, confidence interval; SRQ-20, Self-Reporting Questionnaire-20.
No significant multicollinearity was observed among the independent variables included in the models (all VIF values < 5).
Agreement and diagnostic performance of the adherence assessment methods
The direct and indirect methods showed minimal agreement, with a Cohen’s kappa coefficient of 0.014, and only 29.6% of participants were classified as adherent by both methods. Considering the direct method as the reference standard, the indirect method showed a sensitivity of 91.5% and a specificity of 10.5% for identifying adherence to antihypertensive medication. The PPV was low: only 32.9% of participants classified as adherent by the indirect method were truly adherent according to the direct method. Conversely, the NPV indicated that 72.0% of participants classified as nonadherent by the indirect method were also classified as nonadherent by the direct method (Table 7).
Table 7. Level of agreement between direct and indirect methods for assessing adherence to antihypertensive medication.
| Adherence indirect method, No. (%) | CKC | Sensitivity | Specificity | PPV | NPV | |||
|---|---|---|---|---|---|---|---|---|
| Yes | No | |||||||
| Direct adherence method | 0.014 | 91.5% | 10.5% | 32.9% | 72.0% | |||
| Yes | 75 (29.6) | 7 (2.8) | ||||||
| No | 153 (60.5) | 18 (7.1) | ||||||
CKC, Cohen’s kappa coefficient; PPV, positive predictive value; NPV, negative predictive value.
DISCUSSION
This study evaluated adherence to antihypertensive treatment using both direct and indirect methods, demonstrating low agreement between the approaches and substantial differences in adherence estimates. While most participants were classified as adherent by the indirect method, biochemical detection indicated a lower proportion of adherence. Physical activity, proteinuria, number of antihypertensive medications, and systolic BP were significantly associated with nonadherence when adherence was assessed using the direct method. In contrast, when using the indirect method, only common mental disorders remained in the final model.
The diagnostic performance analysis provided additional insight into the discrepancies observed between adherence assessment methods. For comparative purposes, and considering the direct method as the reference standard, the indirect method showed high sensitivity but extremely low specificity. The low specificity and reduced PPV indicate a high rate of misclassification of individuals as adherent when indirect methods are used. This pattern is consistent with previous evidence showing that self-reported measures are subject to social desirability and recall biases, leading to overestimation of treatment adherence [22]. Importantly, these findings should be interpreted with caution, as no universally accepted gold standard exists for adherence assessment. Rather than reflecting true diagnostic accuracy, these results highlight the low agreement between methods and the fact that each approach captures different dimensions of adherence behavior. Taken together, these findings reinforce that direct and indirect methods should not be considered interchangeable, but rather complementary in the assessment of antihypertensive treatment adherence.
Previous studies have demonstrated substantial variability in the estimated prevalence of adherence to antihypertensive treatment, depending on the population studied and, particularly, on the method used for assessment [23]. Studies based only on indirect measurements highlight this heterogeneity: a meta-analysis conducted in low- and middle-income countries identified prevalences between 6.7% and 85.5% [24]. Globally, another review found rates ranging from 8.2% to 74.9% [8], while a national meta-analysis estimated prevalence ranging from 4.5% to 97.7% with high heterogeneity (I2 = 97.9%; P < 0.001) [25]. Such findings may result from the multiplicity of instruments used to assess adherence and highlight the complexity inherent in assessing adherence.
When considering only the direct method of assessing adherence, no systematic reviews were identified that exclusively included studies based on biochemical measurements. In a meta-analysis [23] that sought to estimate the global epidemiology of adherence to antihypertensive treatment, 161 studies were analyzed, of which only 5 assessed adherence directly and the average prevalence found was 73.0%, higher than that observed in the present study. An observational study with uncontrolled hypertensive patients also identified higher adherence rates (51.0%) based on biochemical urine analysis [26].
Assessing treatment adherence has historically been challenging. One of the main challenges in behavior measurement, such as evaluating adherence to treatment, is the possibility that the act of measurement itself influences people’s attitudes. This phenomenon, known as the Hawthorne effect, refers to the active modification of behavior when participants are aware that they are being observed or monitored [26]. Consequently, patients may overestimate their adherence, a phenomenon known as “white coat adherence.” When assessing adherence using the indirect method, mainly using self-report instruments, people are more exposed to this risk.
Although biochemical analysis provides objective evidence of recent medication intake, this approach also has important limitations, such as the point-in-time assessment of adherence and dependence on the pharmacokinetic profile of drugs. Furthermore, as it reflects a specific moment, it may not capture adherence patterns over time, which can also facilitate the Hawthorne effect, as the person can only use the medication on the day of the appointment. Additionally, the detection of antihypertensive drugs in urine may be influenced by pharmacokinetic variability across different drug classes. Factors such as half-life, metabolic pathways, and renal excretion profiles may affect the presence and detectability of compounds at the time of sample collection. Consequently, drugs with shorter half-lives or rapid clearance may be underdetected, potentially leading to misclassification of adherence. These aspects should be considered when interpreting results obtained through biochemical methods. On the other hand, indirect methods are widely accessible and allow longitudinal assessment, despite being subject to memory biases and social desirability [10,22,26].
Despite the challenges, identifying adherence to antihypertensive medications and, above all, non-adherence, is essential to reduce the burden of the disease and associated costs. Significant efforts have been reported in different countries. A multicenter study in different countries estimated that, if at least 70% of patients used more than 80% of prescribed medications, approximately 6,553 complications related to hypertension would be avoided and there would be savings of around 36 million euros over a decade [27].
Adherence to treatment can be influenced by several factors, including those related to the hypertensive person, highlighting unfavorable socioeconomic conditions, gender, age, living in rural areas, disaster or pandemic situations, denial of the disease, absence of symptoms, inadequate perception of the benefits of treatment, insufficient knowledge about the disease and its management, forgetting to take medications, fear of dependence and the adverse effects of antihypertensive drugs, in addition to lack of family support. With regard to drug treatment, difficulties in acquiring medications, adverse effects, complexity of therapeutic regimens, and the need for continuous treatment stand out. Factors related to the team and the health system include ineffective communication, individualized treatment, failure to identify non-adherence and therapeutic inertia [28,29].
Our study showed that physically active people had a lower chance of non-adherence compared to sedentary people. The physical and functional benefits resulting from regular physical activity may be related to these findings. A similar result was observed in a study, which identified a significant association between adherence to antihypertensive medication treatment and physical activity; active people had a lower chance of non-adherence (P = 0.003) [30].
Participants who had protein in their urine were 2.4 times more likely to not adhere to antihypertensive medication treatment. Proteinuria is an important marker of kidney damage and may be frequently associated with complications of hypertension. Effective BP control in people with CKD associated with hypertension has been associated with a lower risk of adverse cardiovascular events and mortality [31,32]. Recent studies corroborate our findings, indicating that proteinuria not only reflects kidney damage, but is also related to lower levels of adherence to antihypertensive treatment [33,34].
And with each antihypertensive added to treatment, the chance of non-adherence increased by 57%. This result suggests that a greater number of medications may be associated with greater difficulty in adhering to treatment, as has been described in the literature [35,36,37], and points to a relevant public health issue in the country, since the use of combined medications is not a common practice among free antihypertensive treatment schemes. However, only around 30% of hypertensive people respond to treatment with monotherapy [38], justifying the need to add other classes of antihypertensives, which contributes to the increase in polypharmacy, especially in more complex [39] and elderly hypertensive patients [40]. In clinical practice, integrated care strategies, focused on managing prescription activities, in addition to individualized therapeutic approaches, have shown effects on the success of hypertension treatment [41,42].
Elevation in systolic BP was associated with a greater chance of non-adherence, which has been evidenced in the literature, such as in an evaluation study using self-report questionnaires in which adequate pressure control increased the chance of adherence to treatment [24]. This relationship can be bidirectional.
While the direct method identified multiple clinical and behavioral determinants associated with non-adherence, the indirect method was restricted to a single factor related to self-reported emotional state. The presence of common mental disorders was associated with a greater chance of non-adherence. In this context, mental health is an essential element in the multidimensionality that characterizes adherence and a recent study also showed an increase in this correlation according to the degree of depression [43].
Our study has several strengths. The main one is the multimethod approach, considered state of the art for assessing adherence by the WHO [10], and still very little used in clinical studies. Some limitations must also be considered. First, the cross-sectional design is the main one, due to the impossibility of establishing cause and effect relationships. Another limitation may be related to the methods of assessing medication adherence, considering the weaknesses of both the indirect and direct methods. However, this limitation was mitigated by opting for the instrument most used to assess adherence indirectly and the assessment technique currently most recommended (HPLC-MS/MS) to assess adherence directly. Finally, although the sample size was statistically determined and adequate for the study objectives, the fact that the study was conducted in a single tertiary care center may limit the generalizability of the findings to other healthcare settings.
CONCLUSIONS
The findings of this study indicate that the estimate of adherence to antihypertensive treatment varies depending on the assessment method. The results were inconsistent between the two approaches. The direct method demonstrated greater sensitivity for identifying variables associated with non-adherence. However, both methods have limitations, reinforcing the importance of a multimodal approach in assessing adherence. Future studies are needed to better understand these relationships and support the development of effective strategies to effectively control hypertension and thereby reduce the public health burden of hypertension.
Acknowledgements
Not applicable.
Abbreviations
- ACE
angiotensin-converting enzyme
- AUDIT
Alcohol Use Disorders Identification Test
- BP
blood pressure
- CI
confidence interval
- CKC
Cohen’s Kappa coefficient
- CKD
chronic kidney disease
- eGFR
estimated glomerular filtration rate
- HbA1c
glycated hemoglobin
- HDL-C
high-density lipoprotein cholesterol
- HPLC-MS/MS
high-performance liquid chromatography–tandem mass spectrometry
- IPAQ
International Physical Activity Questionnaire
- IQR
interquartile range
- LDL-C
low-density lipoprotein cholesterol
- MDRD
Modification of Diet in Renal Disease
- MLOD
Method Limit of Detection
- MRM
Multiple Reaction Monitoring
- NPV
negative predictive value
- OR
odds ratio
- PPV
positive predictive value
- SRQ-20
Self-Reporting Questionnaire-20
- WHO
World Health Organization
Footnotes
Funding: This study was supported by the National Council of Technological and Scientific Development (CNPq; PhD Program), and The São Paulo Research Foundation (FAPESP; 2018/20948-2).
Competing interest: The authors declare that they have no competing interests.
Availability of data and materials: The data sets generated during and/or analyzed during the current study are available from the corresponding authors on reasonable request.
Ethics approval and consent to participate: The research project was approved by the Research Ethics Committee of the University of Sao Paulo, Nursing School and University of Sao Paulo, Medical School (protocol No. 2,831,454 and No. 3,003,912).
Consent for publication: Not applicable.
- Conceptualization: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
- Data curation: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Meira KC, Vieira da Silva G, Pierin AMG.
- Formal analysis: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Vieira da Silva G, Pierin AMG.
- Funding acquisition: da Luz Pádua Guimarães MC, Pierin AMG.
- Investigation: da Luz Pádua Guimarães MC, Pierin AMG.
- Methodology: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Meira KC, Vattimo MFF, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
- Project administration: da Luz Pádua Guimarães MC, Pierin AMG.
- Resources: da Luz Pádua Guimarães MC, Pierin AMG.
- Software: da Luz Pádua Guimarães MC, Pierin AMG.
- Supervision: da Luz Pádua Guimarães MC, Pierin AMG.
- Validation: da Luz Pádua Guimarães MC, Vattimo MFF, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
- Visualization: da Luz Pádua Guimarães MC, Coelho JC, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
- Writing - original draft: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Meira KC, Vattimo MFF, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
- Writing - review & editing: da Luz Pádua Guimarães MC, Coelho JC, dos Santos J, Meira KC, Vattimo MFF, Costa E, Vieira da Silva G, Drager LF, Pierin AMG.
References
- 1.Zhou B, Perel P, Mensah GA, Ezzati M. Global epidemiology, health burden and effective interventions for elevated blood pressure and hypertension. Nat Rev Cardiol. 2021;18:785–802. doi: 10.1038/s41569-021-00559-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.GBD 2019 Risk Factors Collaborators. Global burden of 87 risk factors in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2020;396:1223–1249. doi: 10.1016/S0140-6736(20)30752-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bundy JD, Li C, Stuchlik P, Bu X, Kelly TN, Mills KT, et al. Systolic blood pressure reduction and risk of cardiovascular disease and mortality: a systematic review and network meta-analysis. JAMA Cardiol. 2017;2:775–781. doi: 10.1001/jamacardio.2017.1421. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global burden of cardiovascular diseases and risk factors, 1990-2019: update from the GBD 2019 Study. J Am Coll Cardiol. 2020;76:2982–3021. doi: 10.1016/j.jacc.2020.11.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Writing Committee Members*. Jones DW, Ferdinand KC, Taler SJ, Johnson HM, Shimbo D, et al. 2025 AHA/ACC/AANP/AAPA/ABC/ACCP/ACPM/AGS/AMA/ASPC/NMA/PCNA/SGIM Guideline for the prevention, detection, evaluation and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Hypertension. 2025;82:e212–e316. doi: 10.1161/HYP.0000000000000249. [DOI] [PubMed] [Google Scholar]
- 6.Burnier M, Egan BM. Adherence in hypertension. Circ Res. 2019;124:1124–1140. doi: 10.1161/CIRCRESAHA.118.313220. [DOI] [PubMed] [Google Scholar]
- 7.Schneider APH, Gaedke MÂ, Garcez A, Barcellos NT, Paniz VMV. Effect of characteristics of pharmacotherapy on non-adherence in chronic cardiovascular disease: a systematic review and meta-analysis of observational studies. Int J Clin Pract. 2018;72:e13044. doi: 10.1111/ijcp.13044. [DOI] [PubMed] [Google Scholar]
- 8.Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International Society of Hypertension global hypertension practice guidelines. Hypertension. 2020;75:1334–1357. doi: 10.1161/HYPERTENSIONAHA.120.15026. [DOI] [PubMed] [Google Scholar]
- 9.Lane D, Lawson A, Burns A, Azizi M, Burnier M, Jones DJL, et al. Nonadherence in hypertension: how to develop and implement chemical adherence testing. Hypertension. 2022;79:12–23. doi: 10.1161/HYPERTENSIONAHA.121.17596. [DOI] [PubMed] [Google Scholar]
- 10.World Health Organization (WHO) Adherence to long-term therapies: evidence for action. Geneva: WHO; 2003. [Google Scholar]
- 11.Tomaszewski M, White C, Patel P, Masca N, Damani R, Hepworth J, et al. High rates of non-adherence to antihypertensive treatment revealed by high-performance liquid chromatography-tandem mass spectrometry (HP LC-MS/MS) urine analysis. Heart. 2014;100:855–861. doi: 10.1136/heartjnl-2013-305063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.van der Laan DM, Elders PJM, Boons CCLM, Beckeringh JJ, Nijpels G, Hugtenburg JG. Factors associated with antihypertensive medication non-adherence: a systematic review. J Hum Hypertens. 2017;31:687–694. doi: 10.1038/jhh.2017.48. [DOI] [PubMed] [Google Scholar]
- 13.Beaussier H, Boutouyrie P, Bobrie G, Frank M, Laurent S, Coudoré F, et al. True antihypertensive efficacy of sequential nephron blockade in patients with resistant hypertension and confirmed medication adherence. J Hypertens. 2015;33:2526–2533. doi: 10.1097/HJH.0000000000000737. [DOI] [PubMed] [Google Scholar]
- 14.Morisky DE, Green LW, Levine DM. Concurrent and predictive validity of a self-reported measure of medication adherence. Med Care. 1986;24:67–74. doi: 10.1097/00005650-198601000-00007. [DOI] [PubMed] [Google Scholar]
- 15.Brazilian Association of Research Companies (ABEP) Brazil economic classification criteria 2008. [Accessed 19 Mar 2026]. http://www.abep.org (2008)
- 16.Lima CT, Freire ACC, Silva APB, Teixeira RM, Farrell M, Prince M. Concurrent and construct validity of the audit in an urban Brazilian sample. Alcohol Alcohol. 2005;40:584–589. doi: 10.1093/alcalc/agh202. [DOI] [PubMed] [Google Scholar]
- 17.Matsudo S, Araújo T, Matsudo V, Andrade D, Andrade E, Oliveira LC, et al. International Physical Activity Questionnaire (IPAQ): study of validity and reliability in Brazil. Rev Bras Ativ Fís Saúde. 2001;6:5–18. [Google Scholar]
- 18.Mari JJ, Williams P. A validity study of a psychiatric screening questionnaire (SRQ-20) in primary care in the city of Sao Paulo. Br J Psychiatry. 1986;148:23–26. doi: 10.1192/bjp.148.1.23. [DOI] [PubMed] [Google Scholar]
- 19.O’Brien E, Mee F, Atkins N, Thomas M. Evaluation of three devices for self-measurement of blood pressure according to the revised British Hypertension Society Protocol: the Omron HEM-705CP, Philips HP5332, and Nissei DS-175. Blood Press Monit. 1996;1:55–61. [PubMed] [Google Scholar]
- 20.Barroso WKS, Rodrigues CIS, Bortolotto LA, Mota-Gomes MA, Brandão AA, Feitosa ADM, et al. Brazilian guidelines of hypertension - 2020. Arq Bras Cardiol. 2021;116:516–658. doi: 10.36660/abc.20201238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Feitosa ADM, Barroso WKS, Mion Junior D, Nobre F, Mota-Gomes MA, Jardim PCBV, et al. Brazilian guidelines for in-office and out-of-office blood pressure measurement - 2023. Arq Bras Cardiol. 2024;121:e20240113. doi: 10.36660/abc.20240113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Feinstein AR. On white-coat effects and the electronic monitoring of compliance. Arch Intern Med. 1990;150:1377–1378. [PubMed] [Google Scholar]
- 23.Lee EKP, Poon P, Yip BHK, Bo Y, Zhu MT, Yu CP, et al. Global burden, regional differences, trends, and health consequences of medication nonadherence for hypertension during 2010 to 2020: a meta-analysis involving 27 million patients. J Am Heart Assoc. 2022;11:e026582. doi: 10.1161/JAHA.122.026582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nielsen JO, Shrestha AD, Neupane D, Kallestrup P. Non-adherence to anti-hypertensive medication in low- and middle-income countries: a systematic review and meta-analysis of 92443 subjects. J Hum Hypertens. 2017;31:14–21. doi: 10.1038/jhh.2016.31. [DOI] [PubMed] [Google Scholar]
- 25.Coelho JC, Guimarães MCDLP, Vaz AKMG, Meira KC, Santos JD, Lee RJW, et al. Adherence to antihypertensive drug treatment in Brazil: a systematic review and meta-analysis. Cien Saude Colet. 2024;29:e19282022. doi: 10.1590/1413-81232024298.19282022. [DOI] [PubMed] [Google Scholar]
- 26.Pucci M, Martin U. Detecting non-adherence by urine analysis in patients with uncontrolled hypertension: rates, reasons and reactions. J Hum Hypertens. 2017;31:253–257. doi: 10.1038/jhh.2016.69. [DOI] [PubMed] [Google Scholar]
- 27.Mennini FS, Marcellusi A, von der Schulenburg JMG, Gray A, Levy P, Sciattella P, et al. Cost of poor adherence to anti-hypertensive therapy in five European countries. Eur J Health Econ. 2015;16:65–72. doi: 10.1007/s10198-013-0554-4. [DOI] [PubMed] [Google Scholar]
- 28.Azharuddin M, Adil M, Sharma M, Gyawali B. A systematic review and meta-analysis of non-adherence to anti-diabetic medication: evidence from low- and middle-income countries. Int J Clin Pract. 2021;75:e14717. doi: 10.1111/ijcp.14717. [DOI] [PubMed] [Google Scholar]
- 29.Abegaz TM, Shehab A, Gebreyohannes EA, Bhagavathula AS, Elnour AA. Nonadherence to antihypertensive drugs: a systematic review and meta-analysis. Medicine (Baltimore) 2017;96:e5641. doi: 10.1097/MD.0000000000005641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Fragoulis C, Prentakis AG, Kontogianni E, Polyzos D, Leontsinis I, Mantzouranis E, et al. Dependency relationship of exercise and medication adherence in hypertensive patients. J Hypertens. 2022;40:e306 [Google Scholar]
- 31.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int. 2024;105:S117–S314. doi: 10.1016/j.kint.2023.10.018. [DOI] [PubMed] [Google Scholar]
- 32.Ku E, Lee BJ, Wei J, Weir MR. Hypertension in CKD: core curriculum 2019. Am J Kidney Dis. 2019;74:120–131. doi: 10.1053/j.ajkd.2018.12.044. [DOI] [PubMed] [Google Scholar]
- 33.Ahmed Aziz KM. Association of high levels of spot urine protein with high blood pressure, mean arterial pressure and pulse pressure with the development of diabetic chronic kidney dysfunction or failure among diabetic patients. Statistical regression modeling to predict diabetic proteinuria. Curr Diabetes Rev. 2019;15:486–496. doi: 10.2174/1573399814666180924114041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Yamazaki D, Konishi Y, Kitada K. Effects of renal denervation on the kidney: albuminuria, proteinuria, and renal function. Hypertens Res. 2024;47:2659–2664. doi: 10.1038/s41440-024-01709-4. [DOI] [PubMed] [Google Scholar]
- 35.Poulter NR, Borghi C, Parati G, Pathak A, Toli D, Williams B, et al. Medication adherence in hypertension. J Hypertens. 2020;38:579–587. doi: 10.1097/HJH.0000000000002294. [DOI] [PubMed] [Google Scholar]
- 36.Gupta P, Patel P, Štrauch B, Lai FY, Akbarov A, Marešová V, et al. Risk factors for nonadherence to antihypertensive treatment. Hypertension. 2017;69:1113–1120. doi: 10.1161/HYPERTENSIONAHA.116.08729. [DOI] [PubMed] [Google Scholar]
- 37.Pantuzza LL, Ceccato MDGB, Silveira MR, Junqueira LMR, Reis AMM. Association between medication regimen complexity and pharmacotherapy adherence: a systematic review. Eur J Clin Pharmacol. 2017;73:1475–1489. doi: 10.1007/s00228-017-2315-2. [DOI] [PubMed] [Google Scholar]
- 38.James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J, et al. 2014 Evidence-based guideline for the management of high blood pressure in adults: report from the panel members appointed to the Eighth Joint National Committee (JNC 8) JAMA. 2014;311:507–520. doi: 10.1001/jama.2013.284427. [DOI] [PubMed] [Google Scholar]
- 39.Gupta P, Voors AA, Patel P, Lane D, Anker SD, Cleland JGF, et al. Non-adherence to heart failure medications predicts clinical outcomes: assessment in a single spot urine sample by liquid chromatography-tandem mass spectrometry (results of a prospective multicentre study) Eur J Heart Fail. 2021;23:1182–1190. doi: 10.1002/ejhf.2160. [DOI] [PubMed] [Google Scholar]
- 40.Bonanno EG, Figueiredo T, Mimoso IF, Morgado MI, Carrilho J, Midão L, et al. Polypharmacy prevalence among older adults based on the survey of health, ageing and retirement in Europe: an update. J Clin Med. 2025;14:1330. doi: 10.3390/jcm14041330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Pharmaceutical Society of Australia. Guidelines for comprehensive medication management reviews. 2020. [Accessed 14 Oct 2025]. https://my.psa.org.au/s/article/guidelines-for-comprehensive-mmr .
- 42.Gupta P, Patel P, Štrauch B, Lai FY, Akbarov A, Gulsin GS, et al. Biochemical screening for nonadherence is associated with blood pressure reduction and improvement in adherence. Hypertension. 2017;70:1042–1048. doi: 10.1161/HYPERTENSIONAHA.117.09631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Liu Q, Wang H, Liu A, Jiang C, Li W, Ma H, et al. Adherence to prescribed antihypertensive medication among patients with depression in the United States. BMC Psychiatry. 2022;22:764. doi: 10.1186/s12888-022-04424-x. [DOI] [PMC free article] [PubMed] [Google Scholar]