Predictors of switching from beta-blockers to other anti-hypertensive drugs: a review of records of 19,177 Chinese patients seen in public primary care clinics in the New Territory East, Hong Kong

Hypertension is a global health problem and represents one of the most important, modifiable risk factors for renal and cardiovascular diseases [1] It affects more than 30% of the general US population [2], and estimates of the cost of the disorder are around $66.4 billion worldwide in 2007 [3]. Its prevalence is high over the globe, including Asia Pacific countries [4]. Despite the effectiveness of antihypertensive pharmacotherapy to mitigate its complications [5, 6], the control of hypertension remained suboptimal [7]. In family practice, its total yearly costs were approximately £76.5 million, of which £26.9 million was attributed to patients who discontinue or switch their pharmacotherapy [8].

Beta-blockers are commonly used in family practice to manage a wide variety of chronic diseases, including hypertension. Together with calcium channel blockers, beta-blockers are one of the most commonly prescribed antihypertensive agents in primary care settings [9]. The clinical guidelines used in the clinics of the public sector in Hong Kong were mainly from the 7^th Joint National Committee Seventh Report. Therefore, the major antihypertensive drug classes were regarded as equally acceptable as first-line prescriptions for management of arterial hypertension since 2003 [10]. In our recent studies on switching of antihypertensive drug therapies in Hong Kong, we found that among more than 93,000 subjects with hypertension, 5.7% switched their medications to another antihypertensive drug class within 180 days of their prescriptions [11]. Subjects at higher risks of medication switching were evaluated, and the risk factors included young age, female gender, attendance in family medicine clinics, new visitors and those prescribed thiazide diuretics as well [11]. However when we conducted stratified analysis of each antihypertensive drug class in the same group of patients, the associated factors of medication switching were different which signified a drug-class specific heterogeneity [12, 13]. Beta-blocker is a distinct group of antihypertensive agent which has different efficacy depending on demographic groups and other comorbid conditions [14].

Previous studies among Caucasian populations have identified the risk factors associated with antihypertensive drug switching [15‐19]. Nevertheless, ethnicity might be a factor which limits the generalizability of these findings, since patients in the Asia Pacific region might have a different profile of medication switching due to different pharmacological actions of beta-blockers [20, 21]. In particular, studies on patients of Chinese ethnicity are particularly scarce despite the fact that they constitute one-fifth of the world's population. Chinese patients were shown to be more sensitive to beta-blockade effect when compared to white patients [22].

The primary objective of this study is to identify the independent factors associated with switching of beta-blockers in family practice. We tested the association between age, gender, socioeconomic status, service setting (general out-patient clinic [GOPC] or family medicine specialist clinic [FMSC] or staff clinic), district of residence, visit type (initial or follow-up visits) and the number of co-morbidities, which were also the variables tested in our previous studies [11‐13].

A total of 19,177 eligible patients were included in the present study. 41.5% were aged ≥ 60 years, and the majority were female (62.8%) (Table 2). Most were fee-payers (73.5%), attended GOPCs for the consultations (90.2%), lived in more urbanized regions (Shatin and Taipo: 70.8%), and were first-ever visitors (61.6%). Among them, 83.2% had no co-morbidities, and most prescriptions were issued in the calendar years 2004 (51.2%) and 2005 (22.7%).

Table 3 shows the characteristics of patients who switched compared with those who continued their beta-blocker prescriptions. A total of 4.0% switched their medication within 180 days. When compared with drug continuers, those who switched their prescriptions were older, more were female (64.0% vs. 62.8%), fee payers (73.9% vs. 73.4%), and visitors of GOPCs (91.9% vs. 90.1%) (all p < 0.001). In addition, more patients who switched lived in more rural districts (the North region; 24.4% vs. 22.0%), were new visitors (74.7% vs. 61.1%) and had more co-morbidities (17.3% vs. 16.7%) (all p < 0.001).

From a binary logistic regression model, patients who were older (≥ 50 years; adjusted odds ratio [AOR] ranged from 1.38 to 1.82) and new clinic visitors (AOR for follow-up visitors = 0.57, 95% C.I. 0.48 to 0.68, p < 0.001) were significantly more likely to have their medication switched (Table 4). There were no statistically significant associations detected between other covariates with the outcome variable. When we used 240 days and 360 days as the yardsticks of medication switching and re-conducted the regression analyses, similar significant associations were detected for age and visit type. There existed no multi-colinearity for the regression models, suggesting their statistical robustness.

From this retrospective cohort consisting of 19,177 Chinese patients, we reported that older age and first clinic visits were significant predictors of medication switching. The other factors previously shown to be associated with antihypertensive drug switching [11], including gender, service type, district of residence, visit type and co-morbidities, were no longer statistical significant. Thus far there were no studies of this scale which evaluated the factors associated with switching from a beta-blocker to another antihypertensive prescription.

The mechanism where beta-blockers lower blood pressure could be multifactorial [26]. Some suggested mechanisms included diminution of sympathetic outflow mediated by the effect of the central nervous system, and reduced cardiac output accompanied by increased peripheral resistance [27]. Among the elderly, hypertension is characterized by compromised cardiac output and escalated peripheral vascular resistance [28, 29]. Beta-blockers used in the long term may lead to homeostatic adjustments that can improve cardiac output but will then be depressed when compared with the original state. Furthermore, there has been evidence that lipophilic beta-blockers can easily cross the blood-brain barrier, which induces more depression, sedation and sexual dysfunction in older subjects [30]. In addition, beta-blockers have strong potentials to induce conduction abnormalities, undesirable bradycardia, and development of heart failure especially among elderly with left ventricular dysfunction [30‐32].

Evidence points towards the relative ineffectiveness of beta-blockers as a treatment regimen for hypertension among the elderly populations. For instance, one meta-analysis including more than 16,000 elderly subjects 60 years or older reported that beta-blockers were incapable to prevent coronary heart diseases, fatal or non-fatal myocardial infarctions [33]. Also, when compared with other antihypertensive therapies, beta-blockers were the least effective in reducing left ventricular mass in the elderly although the effects on blood pressure lowering were similar across major antihypertensive drug classes [34‐36]. Therefore with an exception of the Joint National Committee 7^th report, beta-blocker has not been recommended as a first-line agent in the elderly without any concomitant disorders like congestive heart failure or myocardial infarction [37].

Medication switching is essentially a physician-initiated action, but the patient could have influence. The physician might perceive the initial beta-blocker regimen is better replaced by a more efficacious antihypertensive medication to control blood pressure or other cardiovascular risk factors. For instance, elderly patients are more likely to suffer from poorly controlled diabetes and impaired glucose tolerance, and switching of beta-blockers to another agent is warranted. It can also be patient-driven, like a complaint of intolerable adverse events. Our findings that older subjects were more likely to have their beta-blockers switched could be explained by either or both of the above reasons. That the new visitors were more likely to have their medication switched is not surprising, as the follow-up patients might already have refilled their medications for at least once and should be arguably free of adverse effects when compared with the new clinic attendees.

Our study included a large number of Chinese subjects. The good dispensing practice in the primary care out-patient clinics allowed an accurate analysis of prescription patterns free of recall biases. However some limitations should be mentioned. Firstly this study included a Chinese population from only one geographical region out of seven in Hong Kong, although the demographic features of our samples are similar to all residents in Hong Kong. In addition, we captured only short-term switching profiles, and longer term prescription changes could not be assessed. Furthermore, the reasons of medication switching could not be evaluated from retrospective database analyses, and there remained some confounding factors like blood pressure level, body mass index, lifestyle habits and attitudes towards taking of chronic medications which could not be controlled for.

The implication of this study is that more careful practice of prescribing beta-blockers is needed particularly among the elderly and also the new visitors, who are more likely to be drug-naive. Translated into clinical practice, family physicians should attempt closer monitoring of the medication taking behavior of their patients with these risk factors. In the broader context of primary care, a multidisciplinary team of healthcare professionals should collaborate and strengthen medication management in the community. This could involve community pharmacists who can give timely advice to patients regarding medication counseling and support.

The future direction of research should be targeted towards the reasons of medication switching, and to what degree family physicians were influenced by evidence-based guidelines regarding prescription of first-line antihypertensive agents.

not applicable