The prescription pattern of first-line antihypertensives among family medicine trainees in Hong Kong - Part 3: in the presence of hyperlipidemia

Martin C S Wong 黃至生, Roger Y Chung 鍾一諾

HK Pract 2005;27:123-132

Summary

Objective: There has been no authoritative, universal pharmacological recommendations on the best first-line antihypertensive in the presence of hyperlipidemia. This cross-sectional survey studies how family medicine trainees prescribe first-line antihypertensives in patients with known hyperlipidemia.

Design: Telephone-administered interviews conducted from December 2003 to January 2004.

Subjects: All basic and higher trainees of the Hong Kong College of Family Physicians (HKCFP) as of December 2003.

Main outcome measures: (1) Class of antihypertensives prescribed among basic and higher trainees in different patient age and gender groups in the presence of hyperlipidemia; (2) Comparison of prescription proportion of thiazide diuretics and -blockers between basic and higher trainees; (3) The effect of practice experience on the choice of thiazide diuretics and b-blockers as first-line agents.

Results: 101 complete interviews were conducted. (1) In young patients diuretics were the least commonly prescribed (17.1%) among basic trainees, while calcium channel blockers (CCBs) were least popular (12%) among higher trainees. Both basic and higher trainees most commonly chose b-blockers as first-line agents. (2) For elderly patients, the prescription proportions of the 4 major antihypertensive classes were similar among basic trainees, but diuretics became the most commonly prescribed (40%) among higher trainees. Patient gender had zero effect on the prescription proportions in both trainee groups irrespective of patient age. There was no difference in the prescription proportions of diuretics and -blockers between basic and higher trainees, while trainees prescribing diuretics as first line agents had more practice experience (0.02 < p < 0.05 and p < 0.001 for young and elderly patients respectively) as compared to those choosing alternative options.

Conclusion: This study described the baseline prescription pattern of first-line antihypertensives among trainees of HKCFP for patients with hyperlipidemia, revealing heterogeneity of choices. We suggested future prospective, longitudinal studies to evaluate the characteristics of antihypertensives on Asian patients and their clinical outcomes so as to assist the drawing up of cost-effective, evidence-based guidelines in the Asian region.

Keywords: First-line, Antihypertensives, Prescription, Hyperlipidemia

摘要

目的： 家庭醫生對高血脂的高血壓患者未有常規性的降壓藥處方。本文旨在研究正接受家庭醫學訓練的醫生對此類病人的第一線治療方式。

設計： 2003年12月至2004年1月期間經電話進行訪問。

研究對象： 所有於2003年12月正接受香港家庭醫學院基礎或高級程度訓練的醫生。

主要測量內容： (1)有關醫生為不同年齡、性別之高血脂患者所選擇第一線降壓藥之類別。 (2)比較在基礎及高級程度訓練中醫生在處方利尿藥及受體阻斷劑習慣上的差異。 (3)行醫經驗對處方利尿藥及受體阻斷劑的影響。

結果： 研究共進行了101個電話訪問。(1)於年輕病者中，利尿藥為基礎訓練中醫生最少處方的(17.1%)， 而鈣拮抗劑藥為高級程度訓練中醫生最少處方的藥物(12%)。在基礎及高級程度訓練中醫生最常選擇b受體阻斷劑為第一線降壓藥。 (2)在高齡病者，基礎訓練中醫生在處方四類主要降壓藥時的比率相若，但高級程度訓練中醫生多選擇利尿藥(40%)。 不論病者年齡，病者的性別沒有影響醫生所選擇藥物的比率。在比較基礎及高級程度訓練中醫生分別在處方利尿藥及受體阻斷劑的比率， 並沒發現任何差異，但處方利尿藥之訓練生比選擇他類藥物者有更多行醫經驗(年輕病者p值介乎0.02到0.05；老年病者p值少於0.001)。

結論： 此研究描述正接受香港家庭醫學院訓練的醫生在處方第一線降壓藥予高血脂患者時的基礎模式， 顯示對藥物選擇的差異。我們建議將來要為亞洲族裔患者評估其對使用降壓藥的特性及臨床效果作前瞻性縱向研究， 以制定適合亞洲而合乎成本效益之實證指引。

主要詞彙： 第一線，抗高血壓藥物，處方，高血脂。

Introduction

Many randomized controlled trials (RCTs),^1,2 meta-analyses,^3,4 clinical guidelines^5,6 and expert commentaries^7-9 have pointed towards diuretics as the best first-line antihypertensive agent, and in a few instances b-blockers. Professor Michael Alderman has put it well that according to his meta-analysis published in Lancet, "Low-dose diuretics, which have proven efficacy and low cost, should continue to be the standard therapy for hypertension¹⁰..."

Arguably, most of hypertensive patients would require an "antihypertensive treatment cocktail" which includes a thiazide diuretic,⁷ and trials like Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial¹ (ALLHAT) have affirmed the notion that thiazide diuretic is in an older population with hypertension at least as effective as more expensive options.^7,11 In addition, from a cost minimization analysis, major cost savings could be achieved¹² if we switch the more expensive drugs to thiazides, which potentially "...may be over $1 billion per year in the United States¹³".

In summary, there is no room for cost-quality trade-off _ "the most effective treatment (thiazide diuretic) was also the least expensive".⁸ The clinical and public health implications are so enormous once diuretics are adopted as first-line drugs,¹⁴ argued by Professor Lawrence Appel of the Johns Hopkins University, Baltimore.

However, there are relatively few literatures addressing the best first-line antihypertensives in patients presented with hyperlipidemia in primary care settings. No explicit pharmacological recommendation could be observed when hyperlipidemia is an associated concomitant risk factor in major guidelines, including the Seventh Report of the Joint National Committee⁶ (JNC 7th). Indeed, although dyslipidemia is not a compelling contraindication for prescribing thiazides (and b-blockers), many authoritative textbooks^15,16 have highlighted their possible adverse effects on lipid profile, which has been recognized more than a decade ago.^17-19 It has also been raised that hyperlipidemia presenting before commencing antihypertensive treatments may bear prognostic impact in middle-aged, treated hypertensive men,²⁰ despite no major impact was yet found in coronary heart diseases.

Analysis from the Copenhagen Male Study²¹ has revealed that blood pressure (BP) level did not predict the risk of ischaemic heart disease (IHD) in those with high Triglyceride (TG) or low High Density Lipoprotein _ Cholesterol (HDL-C) level.²² Normalizing TG or HDL-C in metabolic syndrome X seems, therefore, more important than lowering BP to minimize the risk of IHD.

Hence it was not clear whether in patients with known hyperlipidemia, starting thiazide diuretics as first-line antihypertensives would still be the most favourable option to our patients given the latest research findings. Doctors might be cautious on the lipid modifying effect of thiazides and b-blockers, which could escalate lipid level in a patient with marginal control to a degree requiring lipid-lowering agents unnecessarily and causing potential complications like stroke and IHD.

As a first step to address the above issue, the present survey studies the class of antihypertensive prescribed to hyperlipidemia patients so as to obtain a baseline prescription pattern among family medicine trainees in Hong Kong.

Methods

We conducted 101 complete telephone interviews from December 2003 to January 2004.

The questionnaire enquired the preferred first-line antihypertensive prescribed by the participants in a patient newly diagnosed as hypertension. They presented with different ages (<65 y.o. and >65 y.o.) and genders with co-existing hyperlipidemia (i.e. questions 5 and 6; see appendix which was reproduced partly from the full survey²³). The questions posed a clinical scenario where patients had only hyperlipidemia with no other complications.

Sampling involved all the basic and higher family medicine trainees registered in the Hong Kong College of Family Physicians (HKCFP) in the year 2002-2003.

"Basic trainees" are defined as doctors undergoing hospital- or community-based basic training offered by HKCFP as of December 2003.

Eligibility as "higher trainees" includes: (a) medical doctors who had completed four years of HKCFP basic training; (b) current higher trainees of HKCFP; or (c) had just completed the higher training in 2002 who were certified and acknowledged by the Board of Vocational Training and Standards in the college's Annual Report 2002-2003.²⁴ This definition is different from "higher trainees of HKCFP" as the latter involves only category (b).

One of the authors contacted corresponding trainee representatives (or equivalent coordinators) for basic trainees practicing in various clusters, and study group leaders for higher trainees. We requested to invite trainees' participation by telephone interviews. We were committed to obtain trainees' telephone numbers only where explicit and informed verbal consent of each trainee was available from the representatives.

Eligible participants were then contacted using the list of telephone numbers. When the interview started, we further assured our interviewees' anonymity, confidentiality and the sole purpose of data collection for research.

We studied the prescription pattern of the various classes of antihypertensive among basic and higher trainees. Also, we compared the prescription proportion of (a) diuretics and (b) b-blockers between basic and higher trainees using c² tests of homogeneity. These two agents were chosen for analysis as they may aggravate dyslipidemia and we hypothesized that their prescription proportions differed between the 2 groups of trainees. We also studied the effect of practice experience on the prescription of diuretics and b-blockers respectively using student's t-test, assuming normal distribution. Since these were all pre-planned hypotheses, corrections for multiple statistical comparisons were not attempted. An a value of <0.05 was considered significant, and all tests of continuous variables were two-sided.

Results

We obtained 142 telephone numbers from our contact persons, and a single investigator conducted 101 complete telephone surveys (Table 1). During the study period, the other 41 potential participants could not be contacted. We did not come across any refusal or withdrawal during the telephone interviews after clear explanation of the nature of our study and the ethical issues mentioned above. Further clarifications of survey questions were requested from a minority of trainees, which were responded using our criteria stated at the end of the appendix.

For the young patients, diuretics were the least popular choice (17.1%) among the basic trainees irrespective of gender (Table 2). The prescription proportions of the other three drug classes were approximately equal (26.3% - 28.9%). CCBs were the least commonly prescribed (12.0%) among the higher trainees. For both basic and higher trainees, b-blockers (28.9% and 32.0% respectively) enjoyed the greatest popularity.

In elderly patients, the prescription proportions of all 4 drug classes were similar (22.4% - 26.3%) among the basic trainees, but diuretics significantly surpassed the other drug groups as the preferred first-line agent (40.0%) among the higher trainees.

It is noteworthy that exactly the same figures were shown (the lower 2 panels for Q6 compared with the upper 2 panels for Q5 in Table 2) when patient gender was now assumed to be female. Indeed, we further tested the degree of agreement on the choice of drug class for every individual participant (n = 101, both basic and higher trainees) using patient gender as a dependent variable, resulting in a kappa (k) value ( S.D.) of 1.0 0.0. This means there is perfect intra-rate agreement in drug choice for both patient genders. In other words, all of the trainees individually kept their same choice of antihypertensive class irrespective of patient gender, physicians' gender and training location, as well as their practice experience.

Comparing the prescription pattern of thiazide diuretics between basic and higher trainees (Table 3), we observed greater proportion of higher trainees prescribing diuretics than that of basic trainees, although not statistically significant. This finding included all 4 age and gender groups. In the elderly patients, both male and female, the difference reached marginal significance (p~0.08).

For the b-blockers, no significant differences in prescription proportion were found between basic and higher trainees in all patient groups. There seemed to have more basic than higher trainees choosing b-blockers (25.0% vs. 16.0%) in the elderly patients, but the findings failed to reveal statistical difference.

Years after graduation was taken as a proxy measure for practice experience in the analysis of the prescription of diuretics and b-blockers (Table 4). It was revealed that those prescribing diuretics were more experienced physicians when compared to those choosing other drug classes (p < 0.05 for all tests). This applied to all patient groups; and of particular interest was the very significantly (p<0.001) greater number of years in practice of the "diuretic-prescribers" than "non-diuretic-prescribers" among the elderly patients. On the contrary, we observed no statistical difference in physicians' practice experience between those choosing b-blockers as compared to those choosing alternative options.

Discussion

One limitation of our present study concerns the definition and degree of "hyperlipidemia" posed in the survey. Hence a respondent may elect a different response when LDL-C rather than TG was elevated, or when the patient suffered from poorly-controlled hyperlipidemia as compared to a milder condition. However, it may not be appropriate to define disease severity in this context as there is no consensual cut-off value for both LDL-C and TG distinguishing "severe" or "mild". Indeed, no respondent requested clarification on this issue. To fully account for this ambiguity, a prescription-based rather than survey-based study is required for logistic regression analysis.

Our finding demonstrates that basic trainees were rather conservative in prescribing thiazide diuretics in a patient already presented with hyperlidemia, irrespective of patient age and gender. This is in contrast to the prescription patterns in the absence of concomitant cardiovascular risk factors of the same study, where diuretics were the most popular for both trainee groups disregarding patient age and gender.²³ This observation might therefore be contributed to the lipid modifying effects of diuretics, although why b-blockers (sharing similar modifying effects) remained popular was unknown.

Relatively few higher trainees chose CCBs for the young patients, which could be due to recent awareness of the unfavourable effects of commencing CCBs as a first-line agent. In a meta-analysis of trials involving 27743 people, intermediate- and long-acting CCBs were found to have 26% higher risk of heart attack, 25% higher risk of heart failure, and 10% higher risk of any major cardiovascular events.¹⁰ As commented by Josefen,²⁵ there is growing evidence that CCBs are inferior to other drugs, like its poorer effect in preventing strokes as compared to hydrochlorothiazides.²⁶

Of interest was the finding that most higher trainees chose diuretics in the elderly patients presented with hyperlipidemia, a result which was not observed among the basic trainees. One might refer to the guidelines of the JNC 7^th which stated that treatment recommendations for elderly should follow the same principles⁶ as in general care (i.e. thiazide diuretics as suggested by this guideline), and that elderly patients responded better to hydrocholorothiazide as argued in an article in the British Medical Journal.²⁷ Therefore higher trainees could be more exposed to recent arguments of best first-line antihypertensives favouring the use of diuretics in the elderly despite the presence of lipid disorders.

For every single participant, the response to question 5(a) was exactly the same as those to question 5(b), as well as for questions 6(a) and 6(b). This perfect, positive agreement (k = 1) revealed that patient gender was not regarded by trainees as important in starting first-line antihypertensive agents. In fact, the authors could not observe any compelling, or relative indications or contraindications of antihypertensives with respect to a particular patient sex, with possible exceptions of pregnancy and benign prostatic hyperplasia which could only occur in females and males respectively.

Whereas we found no statistical difference in prescription proportions of diuretics and b-blockers between basic and higher trainees, the prescription of diuretics was shown to be associated with practice experience. Again, hyperlipidemia has not been recognized as a contraindication to these two agents. That more experienced practitioner chose diuretics in the face of hyperlipidemia might reveal current opinions among the more senior colleagues favouring diuretics as a first-line agent. The exact reasons warranted future studies.

Our study findings depicted the baseline prescription patterns in Hong Kong. It is a challenge however to evaluate the best first-line agent in our locality since research among Asian populations is still scarce. The factors which determine the best agent are also numerous and complicated, including cardiovascular endpoints, BP lowering efficacy, tolerability profiles, costs, administrative issues and other unquantifiable parameters like patients' expectations and doctors' preference. Despite the complexity involved in clinical decision making, local data on drug characteristics and cardiovascular endpoints are still invaluable. Future studies in Asia, preferably by a prospective longitudinal design, may shed light into more favourable drug choices in the presence of hyperlipidemia and help drawing up of local evidence-based clinical guidelines.

Even with the scarcity of local evidence for choosing the best antihypertensive in hyperlipidemia patients, primary care practitioners should continue to adopt a holistic, patient-centered approach and regard highly our patients' concerns and expectations. In this respect, we should remind ourselves that evidence-based decisions are influenced by a multitude of factors,²⁸ but not research data only.

Perhaps in this era when we still lack local evidence and are full of controversies, we could, until more evidence is available, adopt what Chan et al²⁹ has pointed out in the Jan, 2005 issue of Hong Kong Practitioner, "the crucial issue is treat-to-target rather than debating which is the best initial antihypertensive drug."

Acknowledgements

Sincere thanks to all colleagues who participated in this survey and contributed their invaluable time and advice. The principal author extends his personal gratitude to his former trainer in family medicine, Dr. Simon Kwong Ka-Wah, Senior Medical Officer of Nam Shan Elderly Centre, since he has offered many insightful commentaries and shared his experience in the area of antihypertensive prescription.

Key messages

1. There has been no universal recommendation on the best first-line antihypertensive in a patient with known hyperlipidemia.
2. In young patients diuretics and calcium channel blockers were the least popular among basic and higher trainees respectively, while b-blockers were the most popular in both groups of trainees.
3. In elderly patients, diuretics were the most popular option among higher trainees.
4. Patient gender was not regarded by trainees as a significant factor in prescribing first-line antihypertensives.
5. No difference in prescription proportion of diuretics and b-blockers exists between basic and higher trainees, but doctors who prescribed diuretics (versus those not) for all patient age and gender groups had statistically more practice experience.

Appendix

Appendix: Abridged version of Sample Questionnaire

(Reproduced from reference 23 with permission; only Questions 5-6 were analyzed in this article)

Sex (性別):______   Years after graduation (畢業後年期):_________   Training status: (醫學訓練階段)   Higher trainees (高級程度訓練)   Basic trainees, Community-based (社區為本基礎訓練)   Basic trainees, Hospital-based (醫院為本基礎訓練)         Location of practice (執業地點):________________________

(1)* In your own present practice, which drug would you mostly use as your first-line anti-hypertensive in a young (< 65 y.o.) and an elderly (> 65 y.o.) male patient newly diagnosed as having hypertension (HT)? He enjoyed good past health and had no concomitant cardiovascular risk factors. 在你現時執業的地方，當處理一位剛被診斷為高血壓的 (A) 年輕 (少於65歲)及 (B) 年長 (超過65歲)的男性病人時，你最先會選擇什麼抗血壓藥物？他以往的健康良好，亦沒有其他循環系統的高危因素。

(A) Young HT ____________________________ (Drug name) (B) Elderly HT____________________________ (Drug name)

(2) If the above patient was female, what would be your drug of choice? 若以上病患者為女性，你處方的藥物為何？

(A) Young HT ____________________________ (Drug name) (B) Elderly HT____________________________ (Drug name)

(3) If the male patient was now having diabetes mellitus, what would be your drug of choice? 若以上的男性高血壓患者同時患上糖尿病，你處方的藥物為何？

(A) Young HT ____________________________ (Drug name) (B) Elderly HT____________________________ (Drug name)

(4) If the female patient was now having diabetes mellitus, what would be your drug of choice? 若以上的女性高血壓患者同時患上糖尿病，你處方的藥物為何？

(A) Young HT ____________________________ (Drug name) (B) Elderly HT____________________________ (Drug name)

*Notes: If clarification from the respondents was met, the interviewer would highlight the following points: (1) There would be no compelling indications nor contraindications for whatever drugs subsequently chosen; (2) On subsequent physical examinations, no complications were found (3) The patient did not show any preference towards a particular medication

~ ~ ~ Thank you very much ~ ~ ~