Key words: Primary care, educational need, hypertension, Saudi Arabia, antihypertensive medications, audit

Hypertension (HTN) is a common health problem in eastern Saudi Arabia. The prevalence, among adult population, has been estimated to range from 4.75% to 25.6% [1,2] and constitutes 1.8-3.8% of total consultations to PHC.[3]

Five main drug classes are used, worldwide, to control HTN. These include diuretics, beta-blockers, calcium antagonists (CCB), angiotensin converting enzyme inhibitors (ACEI), and angiotensin II antagonists (ARB).[4] In some parts of the world, however, alpha-adrenergic blockers, reserpine and methyldopa are also used frequently.[4] Although there is no reliable or consistent evidence yet, that indicates substantive differences between drug classes in their effects on blood pressure, there are important differences in the side-effect profiles of each class.[5,6] In addition, there are important differences in the amount of evidence available from randomized controlled trials on the effect of treatment on morbidity and mortality. A large body of data demonstrated benefits of thiazide diuretics and thiazide/BB combination, while fewer data are available about calcium antagonists and ACE inhibitors.[4] Recently, doubts are rising towards the single use of BB, especially in the elderly. [7]

There is general agreement on the principles governing use of antihypertensive drugs to lower blood pressure, independent of the choice of a particular drug. These principles include the use of low doses of drugs to initiate therapy, and the use of appropriate drug combinations to maximize hypotensive efficacy while minimizing side effects. [4,8,9]

Thus, selection of appropriate drugs is of great importance because of wide variations in adverse effect, benefit, contraindication, and cost.

Little is known about the current practice of PHC physicians in eastern Saudi Arabia, in terms of choice of medications and their dosages. This study aims to describe and discuss the pattern of prescription of antihypertensive medications in eastern Saudi primary health care, which might help in the identification of an educational need of practicing physicians.

This study has been carried out, in Qatif district, on the eastern coast of Saudi Arabia, where a population of nearly 500,000 individuals are served by 26 PHC centers (PHCCs).[10]

A weighted, systematic, random sample of 13 (50%) PHCCs were chosen after stratification by the total number of hypertensive subjects registered in each center.

In each sampled center, 50% of registered male and female hypertensives were selected using systematic random sampling. Cases showing no visits in last three months were excluded. Missed medical files and inconsistent medical file numbers were treated as non-responders.

Trained nurses reviewed and collected data from medical records of 320 selected subjects. A pre-defined spreadsheet was used to collect demographic data, duration of HTN, diagnosis of co-morbidity and type and dose of antihypertensive medications used. Data was reviewed and verified by trained physicians working in the same center, and one of the authors.

Categorical data was cross-tabulated, while continuous data was re-coded into groups of interval. Data was tested for normality using kurtosis and skewness standard error. Normally distributed data was tested for significance, using Chi Square [2], Fisher's exact test and Pearson's correlation test, where applicable. Nonparametric categorical data was tested using Mann-Whitney U (MWU) test. Stepwise logistic regression was used to explain the use of high-dose medications. Confidence interval (CI) of 0.95 was calculated for different variables. A p-value of < 0.05 was considered statistically significant. Epi info statistical software version 6.0 was used for data entry, while Statistical Package for Social Sciences version 10 was used for revision and analysis.

Pilot study has been carried out in one PHCC, upon which data collection spreadsheets were modified.

Out of 320 hypertensive patients 302 (94.4%) were on anti-hypertensive medications, at the time of the study. Table 1 describes the demographic characteristics of these patients. Most (86.5%) of recently diagnosed hypertensives (duration of hypertension less than a year) were on medications, while 94.2% of older hypertensives (hypertension duration of five years or more) were on medications.

Mean ± SD number of medications used in male and female patients was 1.4 ± 0.6 and 1.3 ± 0.6, respectively. However, no significant difference was noted (p=0.11 MWU.)

Table 2 shows the number of anti-hypertensive medications used. Number of medications correlates poorly with age (Pearson correlation coefficient=.058; p=.3), in both male and female patients, while correlates positively with duration of HTN (Pearson correlation coefficient=.151; p=.007), as shown in figure 1.

Five main classes of antihypertensive medications were in use as monotherapy or in combination of two or three as illustrated in figure 2. The characteristics of hypertensive patients using each of these classes differ significantly as shown in table 3.

All diuretics used were hydrochlorothiazide (HCTh). Its total daily dose ranged from 12.5mg to 100mg. However, 89.7% of prescribed HCTh were in doses of 25mg, while 4.3% had a daily dose of more than 25 mg.

Atenolol was the main ß-blocker used by all patients using ß-blockers, except one who was using propranolol. Total daily dose of atenolol ranged from 25mg to 100mg, out of which 73.4% were prescribed a dose of 100mg.

Captopril was the main ACEI used by all patients using ACEI, except two patients who were using enalapril. Total daily dose of captopril ranged from 10mg to 100mg. Out of these, 95.9% have a daily dose of 75 mg or less.

All CCBs were nifidipine. Three patients (21.4%) were on long acting nifidipine while the remaining 78.6% were on short acting preparations. Total daily dose of nifidipine ranged from 10mg to 40mg. Out of these, 42.9% have a daily dose of more than 20 mg. Total daily dose of methyldopa ranged from 500mg to 750mg.

Use of high-dose medications was common in 153 (50.7%) patients which show significant correlation with number of medications used (Pearson correlation coefficient= 0.263; p<.001) as shown in figure 3, and use of BB and ACEI as shown in table 4. By stepwise logistic regression the use of BB (B=3.434; Wald test 48.149; p<.001) and no. of medications (B=1.101; Wald test 7.891; p=.005) were significant predicting variables for the use of high-dose medications, while age, sex, duration of hypertension, diagnosis of DM, place of follow up, use of other classes of medications, level of SBP, and level of DBP were not.

The proportion of HTN patients on non-pharmacological regimens (5.6%) is, relatively, low. This is in comparison to other studies, which documented a higher range of 9.49% to 38.4%. [11,12,13,14] However, the same finding has been reported in a hospital-based study in Hong Kong.[15] Studied PHCC used to refer most newly diagnosed cases of hypertension to hospital for evaluation and start of treatment. This might explain this low proportion of hypertensive patients on no medication. However, this has to be ascertained, in view of the regionally reported high prevalence of metabolic risk factors in hypertensive patients.[16,17]

The mean number of anti-hypertensive medications is 1.3 and increases with the duration of HTN, which might reflect increasing age, co-morbidities and worsening of control. However, this relationship has been shown widely.[15,18] Patients on monotherapy constituted 64%, which is comparable to data presented from northern Saudi Arabia, Bahrain and few European countries.[19,20,21] However, it is far less than those figures (80.4-82.1%) presented in other parts of eastern Saudi Arabia ?22,?23 and is far more than those (29.1%-50.8%) presented in Lebanon, Finland, Italy and United States. [12,14,18,24] This relative low proportion of patients on combination therapy might be reflected as lower control of BP, [8,9,25] which might be worsened, further, by the wide variation noticed in practice of different centers.

On the other hand, irrational combinations were found in 5% of the patients, who were using BB and ACEI. [4]

ß-blockers were the main drug of choice (62.2%) for both male and female patients, whether as single therapy or in combination with other medications. These patients tend to be younger in age, which might be explained by increasing hemodynamic adverse effect at older age.[26] High use of ß-blockers is demonstrated, as well, in Bahrain (65.5%) Khobar (55.4%), Finland (51%) and Hong Kong (51%),[18,20,22,23] while lower rates were reported in Lebanon, United States and many European studies (12.9-27%) which show higher use of ACEI, thiazide diuretics or CCB. [11,12,14,18,19] The extensively high proportion of ß-blockers used in this study might be attributed to affordability of the medication and relative younger age of our population. However, it is worrying in the context of its diabetogenic effect and the increasing doubts on the use of non-combined atenolol. [7]

Thiazide diuretics were second in popularity (36.3%) in this study, with higher tendency in older age patients and non-diabetics. Apart from higher proportions (42%) reported in Trinidad, remarkably lower proportions were reported in different studies (4%-27.4%). [11,12,14,19] Though a high proportion of patients used thiazide in comparison to other studies, this remains far from the international recommendation of having HCTh part of all HTN patients, unless contraindicated or other classes are compellingly indicated. [4,8] The tendency for older age might be attributed to its favorable effects in older age. [26] On the other hand, physicians might avoid prescribing thiazide diuretic for their diabetic hypertensives due to its dose-related adverse metabolic effect on glucose and lipids. [6]

One fifth of non-diabetics were using captopril ACEI, while only one third of diabetics were using it. This is far from what is expected and recommended for the use of ACEI in DM-HTN for its protective effect on propagation of diabetic nephropathy.[27] However a similar finding is shown in Bahrain [28] and the United States (39.3%).[14] This low use of ACEI might be attributed to unavailability of single- dose formula, unawareness of the practicing physician to its protective effects or intolerability to cough which is a well known adverse effect of the drug.[29] In such cases, the lack of an alternative angiotensin receptor blockade (ARB) in PHCCs makes the choices even more limited.[4]

CCBs were used in small proportion (4.4%) in comparison to other studies (8%-35.9%).[12,14,19] This low proportion might be related to the concerns of the association of SA-CCB with increased risk of myocardial infarction in elderly and ischemic heart disease patients. [30] Despite the fact that SA-CCB are not approved for management of HTN patients, but LA-CCB,[30] short acting nifidipine is the major CCB used in this study. In-affordability of LA-CCB and lack of updated information might be reasons for such practice.[31]

Methyldopa was used by only 1.9% of patients, which is similar to figures noted in the United States (1.8%).[14] Its limited use is well understood in view of its frequent CNS adverse effects.[32] This limited use is, however, much lower than figures reported in other studies from Saudi Arabia (4.5%),[22] Bahrain (8.5%),[20] Lebanon (13.6%),[12] Hong Kong (3.5%),[25] and Trinidad (33%). [11]

Dose of anti-HTN medications was high in half of this study population. However, it is BBs which were found to be the main predictor for this practice. This might be influenced by the available form of BB in studied PHCCs, which was a 100 mg tablet of atenolol. This conclusion is further supported by the low use of high dose HCTh (4.3%) and captopril (4.1%) which are available as medium-dose 25 mg HCTh tablets and low-dose 25 mg captopril tablets. This must be considered in view of the association of high doses of anti-HTN medications with higher incidence of metabolic and hemodynamic AE. [33] Variation in the practice of PHCCs is wide, as noticed in the average number of medications used and proportion of patients on diuretics and CCB. This variation might reflect variable background, variable training of practicing physicians and the absence of common guidelines that addresses these issues. However, these concerns are worth further exploration.

This study shows many drawbacks in management of hypertension, which constitute a major educational need. This includes low use of combination anti-hypertensive medications, low use of ACEI in diabetic hypertensives, use of SA-CCB in place of LA-CCB, high use of maximum or high-dose medications and wide variation between studied practices.

Such pitfalls need to be addressed by decision makers, practicing physicians, supervisors and educational bodies. Putting updated common guidelines into practice with purposeful training, replacing SA-CCB with LA-CCB and affording low-dose formulation of thiazide and atenolol BB are suggested recommendations.

Acknowledgements:
We would like to thank doctors and chronic disease nurses working in the sampled centers for their valuable contribution in collection of data.