Frequency & factors associated with apparent resistant hypertension among Ghanaians in a multicenter study

Abstract

Apparent resistant hypertension (ARH) is rife among people living with hypertension and is associated with significant morbidity and mortality. There is however paucity of data from sub‐Saharan Africa on the burden of ARH. We sought to report on the frequency and factors associated with ARH among a cohort of Ghanaians with hypertension. A cross‐sectional study involving 2912 participants with hypertension enrolled at five health facilities in Ghana. ARH was defined as either office BP ≥ 140/90 mm Hg on 3 or more antihypertensive medications or on 4 or more antihypertensive medications regardless of BP. Factors associated with ARH were evaluated in a multivariate logistic regression model. We found 550 out of 2,912 (18.9%) of study participants had ARH. Out of these 550 subjects, 511 (92.9%) were on 3 or more antihypertensive medications with BP ≥ 140/90 mm Hg and 39 (7.1%) were on 4 or more antihypertensive medications with BP ≥ 140/90 mm Hg. The prevalence of ARH was 15.5% among elderly aged 75 + years (n = 341), 20.7% among 65‐74 years (n = 588), and 18.9% among those ≤ 64 years (n = 1983). The adjusted odds ratio (95% CI) of factors independently associated with ARH was duration of hypertension, 1.05 (1.03‐1.06) for each year rise; eGFR < 60 mL/min, 1.73 (1.33‐2.25); and diabetes mellitus, 0.59 (0.46‐0.76). Attaining secondary level education and residence in a peri‐urban setting were significantly associated with ARH though not in a dose‐dependent manner. ARH is rife among Ghanaians and may negatively impact on cardiovascular outcomes in the long term.

1. INTRODUCTION

Although effective treatments are available for the management of hypertension (HPT), blood pressure (BP) control is sub‐optimal worldwide and the greatest burden of uncontrolled BP is reported in low‐ and middle‐income countries (LMICs) where cardiovascular disease (CVD) rates are rapidly rising. ¹ , ² The need for improvement in HPT control is particularly urgent in LMICs with two‐thirds of global burden of hypertension. ³ In particular, the highest estimated effect size associating hypertension with stroke causation worldwide has recently been reported in sub‐Saharan Africa (SSA). ⁴ , ⁵ Hypertension is often unrecognized, undertreated, and uncontrolled in a significant proportion of the adult population in SSA with consistently < 30% of hypertensive patients achieving target BP control on treatment across the continent. ⁶ , ⁷ Efforts have been made to educate and screen the population for hypertension, encourage linkage to care centers, initiate therapy for patients, and retain them on treatment in order to achieve optimal blood pressure control. ⁸ , ⁹ However, a major challenge to this goal is the emergence of resistant hypertension.

Resistant hypertension is defined as a condition in which blood pressure control is not achieved despite adherence to adequate doses of three or more first‐line antihypertensive medications, one of which is a diuretic. ¹⁰ , ¹¹ Approximately 10%‐20% of patients with hypertension are estimated to have resistant hypertension. ¹² Since variants of it include pseudo‐hypertension, controlled resistant hypertension, and white coat hypertension, these should be excluded by, for example, measuring a mean 24‐hour ambulatory blood pressure. ¹⁰ , ¹³ It may also be defined as uncontrolled BP with concurrent use of three or more antihypertensive medication classes or use of four or more antihypertensive medication classes regardless of BP level. ¹⁴

The term apparent resistant hypertension is therefore used to define office BP > 140/90 mm Hg in hypertensive patients who are taking at least three medications. ¹³ , ¹⁵ True resistant hypertension is defined as a properly measured office BP > 140/90 mm Hg with a mean 24‐hours ambulatory BP > 130/80 mm Hg in a patient confirmed to be taking ≥ 3 antihypertensive medications. ¹³ The phenomenon of apparent resistant hypertension is quite intriguing, and there has been a growing interest in the subject due to its association with a high risk of cardiovascular complications such as stroke, myocardial infarction, and heart failure. ⁸ Thus it is imperative to achieve effective control of blood pressure to target to avert adverse cardiovascular outcomes. ¹⁶ Due to an aging population, increased prevalence of sleep disorders, obesity, diabetes mellitus, and chronic kidney disease, there is a projected surge in true resistant and apparent resistant hypertension. ¹⁷ For instance, in 1988‐1994, the estimated prevalence of apparent resistant hypertension was 15.9% and this increased to 28.0% in 2005‐2008. ¹⁷ Most published prevalence data have been among European and American populations with very little done among an African cohort of patients. ⁹ , ¹⁸ This study therefore sought to narrow the gap in knowledge on ARH in Africa by reporting on the frequency and factors associated with apparent resistant hypertension among Ghanaians.

2. METHOD

2.1. Study design

The Ghana Access and Affordability Program (GAAP) study is a prospective cohort study involving adults who were 18 years or older with either hypertension (HPT) or hypertension and diabetes mellitus (HPT + DM). The study was conducted at five sites in total including two tertiary medical centers Komfo Anokye Teaching Hospital, (KATH) and Tamale Teaching Hospital, (TTH); two secondary level health facilities namely the Agogo Presbyterian Hospital, (APH) and Atua Government Hospital, (AGH), and one primary level hospital‐ Kings Medical center, (KMC). Ethical approval was obtained from all study sites, and the study protocol has been published elsewhere. ¹⁹

2.2. Evaluation of study participants

We obtained written informed consent from all study participants prior to enrollment into the study. A Standard Operation Procedure was employed across sites to collect demographic information including age, gender, educational attainment, employment status; and lifestyle behaviors such as alcohol use, cigarette smoking, frequency and daily quantities of fruits and vegetable consumption as well as table added salt were assessed through interviews and responses collected on a questionnaire. A detailed medical history including duration of hypertension or diabetes diagnosis and current medications lists taken was obtained. Anthropometric assessments including measurement of weight, height, and waist circumference were performed by Study nurses. Body mass index (BMI) of each participant was then derived by dividing the weight in kilograms by the square of the height in meters. ¹⁹

2.3. Laboratory measurements

An International Organization for Standardization (ISO)‐certified and quality‐assured laboratory was contracted to run all biochemical panels for study participants. Samples were transported to the laboratory by trained phlebotomists on the same day of collection often within 4 hours or where not feasible (KMC and AGH sites), and samples were stored in a freezer before transported to the laboratory the next day.

2.4. Diagnosis of hypertension

All patients had been previously diagnosed with hypertension using cutoff value of ≥140/90 mm Hg and were on antihypertensive therapy at enrollment into the study. Each site was provided with an automated blood pressure device (Omron HEM‐907XL) for blood pressure measurement at enrollment and during follow‐up. Each study participant rested for at least 5 minutes prior to blood pressure measurements while sitting in a chair with both feet flat on the floor. Both arms were supported at the level of the heart on a table. Two consecutive blood pressure readings from the same arm taken 2 minutes apart were recorded, and averaged baseline BP readings were used for the present analysis.

Apparent resistant hypertension was defined as either office BP ≥140/90 mm Hg on 3 or more antihypertensive medications or on 4 or more antihypertensive medications regardless of blood pressure.

Compliance with hypertension treatment was assessed using the 14‐item version of Hill‐Bone compliance to high blood pressure therapy scale with good validity and internal reliability. ²⁰ The scale has 3 subscales under behavioral domains of hypertension treatment including medication adherence, reduced salt intake, and appointment keeping with each question/item answered with a four‐point Likert scale ranging from 1 to 4 (1 = none of time, 2 = some of the time, 3 = most of time, and 4 = all the time) [ref]. Total score ranged from 14 (perfect adherence) to 56 (non‐adherence) with higher scores denoting overall poorer adherence.

2.5. Renal impairment

Renal impairment was defined using estimated glomerular filtration rate (eGFR) calculated from baseline serum creatinine measurement using the Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) formula. ²¹

2.6. Alcohol use

Alcohol use was categorized into current users (users of any form of alcoholic drinks) or never/former drinker while alcohol intake was categorized as low drinkers (1‐2 drinks per day for female and 1‐3 drinks per day for male) and high drinker (>2 drinks per day for female and > 3 drinks per day for male. One (1) drink or 1 unit of alcohol = 8 g of alcohol). ²²

2.7. Smoking status

Smoking status was defined as current smoker (individuals who smoked any tobacco in the past 12 months) or never or former smoker. ²² Vegetable and fruit intake were assessed based on number of daily servings per week.

2.8. Statistical analysis

Baseline characteristics of patients with apparent resistant hypertension and those without were compared by using the Student's t test. Proportions were compared using the Chi‐squared tests or Fisher's exact test for proportions with subgroupings < 5. Factors associated with apparent resistant hypertension were assessed using a multivariable logistic regression model. Variable selection in this model was based on empirical data from bivariate analyses and knowledge of risk factors of apparent resistant hypertension from literature. In all analyses, two‐tailed P‐values < .05 were considered statistically significant. Statistical analysis was performed using SPSS and GraphPad Prism version 7.

3. RESULTS

3.1. Prevalence of apparent resistant hypertension

We found 550 out of 2912 (18.9%) of study participants had apparent resistant hypertension. Of these 550 subjects, 511 (92.9%) were on 3 or more antihypertensive medications with BP ≥ 140/90 mm Hg and 39 (7.1%) were on 4 or more antihypertensive medications with BP < 140/90 mm Hg. The prevalence of ARH across the five participating hospitals was 18.8% at Agogo Presbyterian Hospital (n = 869), 26.2% at Atua Government Hospital (n = 282), 22.5% at the Komfo Anokye Teaching Hospital (n = 1219), 0.6% at Kings Medical Centre (n = 166), and 10.1% at the Tamale Teaching Hospital (n = 378). By age, ARH had a prevalence of 15.5% among elderly aged 75 + years (n = 341), 20.7% among 65‐74 years (n = 588), and 18.9% among those ≤64 years (n = 1983). Furthermore, 111 out of 383 (29.0%) subjects with renal impairment (ie, eGFR < 60 mL/min) had ARH compared with 338 out of 2024 (16.7%) with adequate renal function (ie, eGFR ≥ 60 mL/min). This was found to be statistically significant (P < .0001).

3.2. Comparison of demographic and clinical characteristics of subjects with and without ARH

The mean age ± SD of patients with ARH of 58.8 ± 11.4 years was not significantly different from those without ARH of 58.6 ± 12.5 years (P = .66). Those with ARH were less likely to reside in rural setting 29.5% vs 36.7%, (P = .004) more likely to have attained secondary level education 40.2% vs 33.1% (P = .02) compared with those without ARH. (Table 1) There were non‐significant differences between the two groups with respect to sex, employment status, and monthly income. Regarding lifestyle, there were non‐significant differences in cigarette smoking and alcohol use; however, table added salt was less frequently reported among those with ARH at 9.5% vs 19.7% (P < .0001) and had higher daily fruit servings than among those without ARH. Furthermore, anthropometric markers of obesity, namely body mass index and waist circumference, were significantly higher among those with ARH than those without. Those with ARH were more likely to have a longer duration of hypertension, a lower mean estimated glomerular filtration and less likely to have diabetes co‐morbidity 28.4% vs 39.25% (P < .0001). Mean adherence score to antihypertensive therapy was comparable between the two groups (Table 1).

Table 1.

Characteristics of subjects with and without apparent resistant hypertension in Ghana

Characteristic	Apparent resistant hypertension (n = 550)	No apparent resistant hypertension (n = 2362)	P‐value
Age, mean ± SD	58.8 ± 11.4	58.6 ± 12.5	.66
Male gender, n (%)	128 (23.3)	550 (23.3)	1.00
Location of residence
Rural	162 (29.5)	868 (36.7)	.004
Peri‐urban	137 (24.9)	495 (21.0)
Urban	251 (45.6)	999 (42.3)
Educational status
Tertiary	56 (10.2)	277 (11.7)	.02
Secondary	221 (40.2)	783 (33.1)
Primary	85 (15.5)	397 (16.8)
None	188 (34.1)	905 (38.4)
Unemployed, n (%)	202 (36.7)	771 (32.6)	.07
National Health Insurance coverage for medications, n (%)	280 (50.9)	1086 (46.0)	.04
Monthly income
>1000 GHS	37 (6.7)	184 (7.8)	.20
210‐1000 GHS	144 (26.2)	618 (26.2)
< 210	229 (41.6)	880 (37.3)
Unknown	140 (25.5)	680 (28.8)
Cigarette use
Current use	2 (0.4)	12 (0.5)	.90
Former use	36 (6.5)	157 (6.6)
Never use	512 (93.1)	2193 (92.8)
Current alcohol use	53 (9.6)	173 (7.3)	.07
Table added salt	52 (9.5)	465 (19.7)	< .0001
Fruit intake, daily serving mean ± SD	1.78 ± 1.58	1.63 ± 1.54	.04
Vegetable intake, daily serving, mean ± SD	2.29 ± 1.60	2.19 ± 1.53	.22
Body mass index, mean ± SD	27.5 ± 5.6	26.6 ± 5.6	.0005
Waist circumference, mean ± SD	85.1 ± 26.5	82.1 ± 26.5	.01
eGFR, mean ± SD	71.7 ± 18.5	77.0 ± 15.8	< .0001
Diabetes co‐morbidity, n (%)	156 (28.4)	934 (39.5)	< .0001
Duration of hypertension in years, mean ± SD	9.9 ± 7.7	7.4 ± 7.0	< .0001
Adherence score Hill‐Bone, mean ± SD	18.13 ± 3.69	18.15 ± 4.04	.91

3.3. Antihypertensive drug classes

Calcium channel blockers were the most frequently prescribed class of antihypertensives, followed by ACE inhibitors (ACE‐I) and hydralazine was the least. Approximately 76.2% of all the study participants were prescribed calcium channel blockers; 46.2% were on ACE‐I and those on hydralazine represented 1.5%. Among patients with ARH, in decreasing order, the frequency of class of medications was calcium channel blockers (88.7%), diuretics (59.1%), ACE‐I (56.4%), methyldopa (49.1%), angiotensin receptor blockers (ARBs) 45.3%, beta‐blockers 28.2%, and hydralazine (7.1%) shown in Table 2. Only 46.2% were on the recommended triple combination of CCB plus diuretic plus either ACE‐I or ARB.

Table 2.

Comparison of antihypertensive drug classes among subjects with and without Apparent Resistant Hypertension in Ghana

Classes of antihypertensive, n (%)	Apparent resistant hypertension (n = 550)	No apparent resistant hypertension (n = 2362)	P‐value
Calcium channel blockers
Yes	488 (88.7)	1731 (73.3)	< .0001
No	62 (11.3)	631 (26.7)
Diuretics
Yes	325 (59.1)	571 (24.2)	< .0001
No	225 (40.9)	1791 (75.8)
ACE‐I
Yes	310 (56.4)	1036 (43.9)	< .0001
No	240 (43.6)	1326 (56.1)
Methyldopa
Yes	270 (49.1)	180 (7.6)	< .0001
No	280 (50.9)	2182 (92.4)
ARB
Yes	249 (45.3)	599 (25.4)	< .0001
No	301 (54.7)	1763 (74.6)
Beta‐blockers
Yes	155 (28.2)	125 (5.3)	< .0001
No	395 (71.8)	2237 (94.7)
Hydralazine
Yes	39 (7.1)	6 (0.3)	< .0001
No	511 (92.9)	2356 (99.7)
ACE‐I + CCB + Diuretics
Yes	147 (26.7)	60 (2.5)	< .0001
No	403 (73.3)	2302 (97.5)
ARB + CCB + Diuretics
Yes	107 (19.5)	36 (1.5)	< .0001
No	443 (80.5)	2326 (98.5)

3.4. Factors associated with ARH

Table 3 shows the factors associated with apparent resistant hypertension among hypertensive patients in Ghana. Bivariate analyses identified location of residence, educational attainment, availability of antihypertensive medications on the national health insurance scheme, fruit intake, body mass index, duration of hypertension diagnosis, estimated glomerular filtration rate, and diabetes mellitus co‐morbidity as potential factors associated with ARH. After multivariable adjustment, the adjusted odds of apparent resistant hypertension was independently predicted by longer duration of hypertension diagnosis (aOR 1.05; 95% CI 1.03‐1.06) for each year increase and decreased eGFR (aOR 1.73; 95% CI 1.33‐2.25) increase the likelihood of a patient having apparent resistant hypertension. On the other hand, having diabetes mellitus decreased the likelihood (aOR 0.59; 95% CI 0.46‐0.76) of having apparent resistant hypertension. There was also non‐dose‐related association between educational attainment and location of residence with ARH where secondary education was significantly associated with ARH (aOR 1.58; 95% CI 1.08‐2.32) and also living in a peri‐urban area (aOR 1.46; 95% CI 1.07‐2.01).

Table 3.

Multivariate logistic regression analysis to identify risk factors for apparent resistant hypertension among Ghanaians

Predictors	Unadjusted OR (95% CI)	P‐value	Adjusted OR (95% CI)	P‐value
Increasing age	1.00 (0.99‐1.01)	.69	–	–
Male gender	1.00 (0.80‐1.25)	.98	–	–
Residence location
Rural (referent)	1.00		1.00
Peri‐urban	1.48 (1.15‐1.91)	.002	1.46 (1.07‐2.01)	.02
Urban	1.35 (1.08‐1.67)	.007	1.11 (0.85‐1.45)	.45
Educational status
Tertiary (referent)	1.00		1.00
Secondary	1.40 (1.01‐1.93)	.04	1.58 (1.08‐2.32)	.02
Primary	1.06 (0.73‐1.53)	.76	1.22 (0.78‐1.89)	.38
None	1.03 (0.74‐1.43)	.87	1.19 (0.80‐1.76)	.39
Antihypertensive medications not available on NHIS	1.25 (1.04‐1.51)	.02	1.23 (0.99‐1.52)	.06
Current alcohol use	1.30 (0.94‐1.79)	.11	–	–
Cigarette use	0.96 (0.67‐1.39)	.84	–	–
Fruit intake	1.06 (1.00‐1.12)	.04	1.01 (0.95‐1.08)	.67
Vegetable intake	1.04 (0.98‐1.10)	.22	–	–
Body mass index, each 5‐year increase	1.15 (1.06‐1.25)	.0005	1.03 (0.94‐1.13)	.53
Duration of hypertension	1.04 (1.03‐1.06)	< .0001	1.05 (1.03‐1.06)	< .001
eGFR < 60 mL/min	2.04 (1.59‐2.61)	< .0001	1.73 (1.33‐2.25)	< .0001
Diabetes mellitus	0.61 (0.49‐0.74)	< .0001	0.59 (0.46‐0.76)	< .0001

4. DISCUSSIONS

4.1. Prevalence of ARH

The goal of hypertension management is to achieve optimal blood pressure (BP) control < 140/90 mm Hg and to prevent cardiovascular complications associated with persistently raised BP. ¹⁸ , ²³ The global burden of poor BP control varies from one region to the other. Published data show that for patients with hypertension, only 8.1% are controlled (BP < 140/90) in China and 10.7% in Vietnam. ²⁴ In Europe, BP control is reported as approximately 10%, 29% in the United States, 17% in Canada, and 7% in sub‐Saharan Africa. ²⁵ , ²⁶ In Ghana, only 2.8% of semi‐urban and rural dwellers and 4% of urban dwellers with hypertension have BP controlled on medications. ²⁴ It is posited that a sizable proportion of individuals living with hypertension on medications may have uncontrolled hypertension due to resistant hypertension.

We found in our study that 18.9% of Ghanaians with hypertension from five hospitals had apparent resistant hypertension (ARH). This is marginally higher than the 4.6% to 17.5% of ARH prevalence estimated in Africa. ²⁷ It has been reported to be 49.7% among individuals aged 45 years or older with different geographic and racial backgrounds living in the United States ²⁸ and 28%‐30% of all patients with uncontrolled BPs. ²⁹ A meta‐analysis of data from 3.2 million patients reported prevalence of ARH of 14.7% in the general population, approximately 14% in America, Asia, and Europe, 20.3% in Eastern Mediterranean, and 19.2% in the Western Pacific regions. ²⁷ The varying prevalence rates have been attributed to the observation that most of the studies (70%) on true, apparent, and pseudo‐resistant hypertension used only office BP measurement. ²⁷ This can influence the BP recordings, result in masked hypertension or inconsistent BP from “white coat” effect. ³⁰ Other reasons for the varying prevalence rates may be different definitions used for ARH, different study designs (observational, cross‐sectional, retrospective, or case‐control studies), and the sociodemographic diversity of study populations (age groups, gender, ethnicity, or race). ¹⁸ , ²⁷

4.2. Risk factors for ARH

Studies have shown ARH to be lower in women than men. ¹⁴ , ²⁹ This, notwithstanding, with advancing age (usually after 65 years), there is a higher female prevalence of hypertension ³¹ and history of oral contraceptive pill usage and pregnancy increase the risk of hypertension and stroke. ³¹ , ³² In our hospital‐based study, however, there was non‐significant gender effect in the prevalence of ARH. We found that participants who lived in peri‐urban communities were significantly more likely to have ARH relative to those in rural dwellings. Urbanization promotes sedentary lifestyle and consumption of high calorie and processed food. In our study, however, the significant association between urban dwelling and ARH observed in bivariate analysis was attenuated into non‐significance upon adjusting to confounders. It is also intriguing that secondary level education was independently associated with ARH relative those the referent group of tertiary level education. Other lower strata of educational attainment were not significantly associated with ARH. This is contrary to findings from the REasons for Geographic And Racial Differences in Stroke (REGARDS) study which reported a high prevalence of ARH among patients who have had less than high school education. ²⁸ This may be attributed to the fact that standard and quality of education at various levels differ from country to country. This can impact on patients’ understanding of disease conditions, drug adherence, and their contribution to long‐term outcome. Thus, our study highlights the roles of educational status and location of residence as potential contributors to ARH in Ghana.

Apart from sociodemographic characteristics, participants’ co‐morbidities such as chronic kidney disease (CKD), obesity, diabetes, and duration of hypertension were found to be associated with ARH. Our study showed that estimated glomerular filtration rate (eGFR) has an association with ARH. This is in consonance with previous studies that found high prevalence of ARH in patients with impaired kidney function. ¹⁴ , ²⁹ Among patients with CKD (eGFR < 60 mL/min per 1.73 m²), prevalence rates of ARH varied from 24.9% for eGFR 45‐59 mL/min and 33.4% for eGFR 45 mL/min per 1.73 m². ¹⁴ It was 15.8% in those with eGFR > 60 mL/min per 1.73 m² and 28% for eGFR < 60 mL/min per 1.73 m². ¹⁴ In a bigger study, ARH was as high as 56% in patients with CKD. ²⁷ We found a 73% higher adjusted risk of ARH in patients with eGFR < 60 mL/min per 1.73 m². It is therefore imperative that patients diagnosed of CKD should be given adequate tolerable doses of antihypertensives to achieve optimal control of their BPs. This would prevent stroke, coronary artery disease, and myocardial infarction, which are complications of both ARH and CKD. Within this cohort, we have reported the prevalence of CKD to be 25.7% ³³ and CKD was independently associated with risk of incident stroke. ³⁴

Obesity negatively affects effective BP control; for every 1.7 kg/m² increase in BMI and a 4.5 cm increase in waist circumference for men, there is a 1 mm Hg rise in systolic BP. ³⁵ In women, an increase in BMI of 1.3 kg/m² and 2.5 cm in waist circumference result in a similar outcome. ³⁵ Obesity contributes to high aldosterone levels and increased adiposity which may lead to resistant hypertension, as well as mortality. ³⁶ , ³⁷ It has been found to be one of the independent risk factors of ARH. ³⁸ Other factors attributed to obesity and difficult‐to‐treat hypertension are increased level of retention of sodium, raised cardiac output, occurrence of obstructive sleep apnea, and overactivity of the sympathetic nervous system. ³⁹ , ⁴⁰ This may explain why in our study, BMI and waist circumference were found to be significantly associated with ARH in bivariate analyses but not in multivariate analysis.

Lifestyles have been documented to contribute to inability to achieve optimal BP control through the conduit of high sodium intake, excess alcohol consumption and inadequate intake of fruit and vegetables. ⁴⁰ In this study, fruit intake was found to have an association with ARH in bivariate analyses. Table added salt increased the risk of ARH. Apart from increasing the intravascular volume and blood pressure, high sodium intake blunts the effectiveness of the various antihypertensive medications. ⁴⁰ A previous study could not establish a significant association between unhealthy lifestyle and ARH. ²⁸

Available literature suggests that patients with long‐standing hypertension (at least 11 years) have a heightened risk of ARH. ²⁸ , ⁴¹ This was confirmed in our study where each year of living with diagnosis of hypertension was associated with a 5% higher risk of ARH (95% CI of 3%‐6%). Prolonged duration of hypertension may be associated with endothelial dysfunction, impaired stimulation of the autonomic nervous system which may result in vasoconstriction and increased peripheral resistance. Furthermore, diabetes as a co‐morbidity has been found as an important contributor to ARH ²⁸ , ³⁸ and as diabetes prevalence increases with its attendant treatment challenges, so does ARH. ⁴⁰ Paradoxically, in our study, we found patients with diabetes were at a lower risk for ARH. This may be attributable to the fact that most clinicians prescribed ACE‐Is and ARBs to patients with diabetes/hypertension co‐morbidities based on recommended guidelines and literature and aimed for lower BP targets. ⁴² It can be inferred that since these medications have enhanced renoprotective effect, the benefit inured from renin‐angiotensin blockade might have improved eGFR with attendant reduction in ARH risk. ¹⁴ Furthermore, there was a higher usage of a combination of calcium channel blockers, diuretics, and either ACE‐I or ARB among those with ARH compared with those without ARH in the current study as supported by literature. ⁴³ Assessing medication adherence and relating it to ARH can be challenging partly due to the inherent weaknesses in the tools for adherence measurement. ⁴⁴ The mean score on the Hill‐Bone compliance to high blood pressure therapy scale was not significantly different between those with ARH (18.13) vs (18.15) among those without ARH. On the Hill‐Bone scale, a score of 14 is deemed perfect adherence while 56 is poor adherence. Furthermore, the Hill‐Bone scale assesses adherence in three composite domains namely medication adherence, salt intake, and appointment keeping. Thus, it may be argued that level of compliance or adherence to therapy for hypertension did not significantly differentiate those with ARH from those without ARH in the present study. However, evidence from previous studies that assessed mainly medication adherence suggests that non‐adherence to treatment may underlie ARH. ⁴⁵

4.3. Implications

Clinicians and other healthcare professionals should actively identify patients with resistant hypertension and address it to prevent morbidity and mortality. Care providers are encouraged to use guideline‐recommended combinations such as diuretics with calcium channel blocker and either ACE‐I or ARB. In the present study, < 50% of participants with resistant hypertension were on the recommended triple combination of CCB plus diuretic plus either ACE‐I or ARB. Patients should be educated to adhere to medications and encouraged to embark on self‐monitoring of home‐based BP to improve self‐efficacy. At the system‐level, there is the need for increased provision and availability of effective hypertension medications, and training of health workers in evidence‐based management of hypertension.

4.4. Limitations and strengths

Although we have reported on important findings on ARH among Ghanaian patients with hypertension, this is a cross‐sectional study hence we cannot draw causal relationships between risk factors identified and ARH. We recommend that future studies should include long‐term follow‐up of study participants to assess the outcomes of ARH. However, it is a multicenter study involving participants from teaching hospitals, secondary level hospitals, and a primary care facility which enhances generalizability of our findings. Interestingly, ARH prevalence was highest at secondary level hospitals followed by tertiary level and finally primary level hospitals. The reasons for these differences are not immediately apparent to us. This may however be due to a combination of patient‐level, provider‐level, and system‐level factors which may differ across study sites. With a big sample size (2912) and participants of different geographic and ethnic diversity, the findings of this study are a fair representation of the Ghanaian population. To our knowledge, this is the first study on the subject of ARH among Ghanaians. Hence, it is a significant contribution to the global efforts to determine the prevalence of ARH and identify factors associated with it. This will equip prescribers, health workers, and policy makers to effectively address challenges with BP control and ARH in order to enhance patients’ survival and positive disease outcomes.

5. CONCLUSION

In conclusion, the prevalence of apparent resistant hypertension is high among Ghanaian patients. The independent predictors of ARH are renal impairment, diabetes, long duration of hypertension, and sociodemographic determinants such as educational attainment and location of residence. ARH is rife among Ghanaians and may negatively impact on cardiovascular outcomes in the long term.

CONFLICT OF INTEREST

The authors do not have any competing interests.

AUTHORS CONTRIBUTIONS

FSS conceptualized this study and developed the study design. FSS undertook the fieldwork, designed the data collection tools, and collected the study data. NKAB, AM, DOA, and FSS participated in data analysis and interpretation of data. NKAB, AM, DOA, and FSS drafted the manuscript. All authors read and approved the final manuscript.

Ayisi‐Boateng NK, Mohammed A, Opoku DA, Sarfo FS. Frequency & factors associated with apparent resistant hypertension among Ghanaians in a multicenter study. J Clin Hypertens. 2020;22:1594–1602. 10.1111/jch.13974