Updating guidelines on stable angina

Doctors vary greatly in the way they assess and manage angina.¹ The problem is common and serious so valid, clinically useful guidelines are welcome. The North of England Guidelines Group has an impressive record of producing rigorous guidelines for primary care. Today, the updated version of the group's 1996 guideline for managing stable angina is published in full on bmj.com.² The new guideline extends the literature search from 1994 to 1997. It fulfils the 1996 commitment to an update after three years. How do the two versions compare?

Recommendations for assessment and referral have changed little from 1996. They are closely aligned with those in the British government's national service framework for coronary heart disease.³ The continuing challenge is to implement them. In particular, many areas of the United Kingdom will still find it difficult to meet the recommendation for prognostic exercise testing for all patients with stable angina.

There are few changes to the recommendations for controlling symptoms of angina. The guideline advises β blockers as the first choice for patients needing regular treatment. This recommendation, previously based on indirect evidence from patients with myocardial infarction, now draws on trial evidence showing that β blockers reduce morbidity in patients with stable angina.⁴ Verapamil remains the preferred alternative for those who cannot tolerate β blockers. The guideline continues to find little evidence to differentiate between the other types of calcium channel blocker or nitrates. It contains limited new evidence about combination therapy. The firmest recommendations are to offer a patient uncontrolled on a β blocker a dihyopyridine or diltiazem, or, if intolerant of the calcium channel blockers, isosorbide mononitrate. However, because some of the new trials have conflicting results, the guideline rates the underlying evidence for these recommendations as less strong than in 1996.

The guideline continues to highlight the importance of aspirin for secondary prevention. There are minor changes in recommendations for modifying weight, physical activity, and smoking. There is a new emphasis on a Mediterranean diet and on fatty fish intake, though the guideline points out that the trials in which these diets reduced mortality were conducted in survivors of myocardial infarction. There is a helpful discussion of the efficacy and safety of nicotine replacement in helping patients with angina to stop smoking. The 1997 cut off for the literature search unfortunately excludes consideration of bupropion.

The guideline avoids specific recommendations on cholesterol lowering drugs. Contentiously, it advises that treatment of risk factors should be based on a patient's absolute risk of further cardiovascular events and death, taking into account multifactorial risk. This is accepted primary prevention but is controversial for patients with known coronary artery disease. Most guidelines treat existing disease as evidence of high risk. For example, the national service framework sets targets for cholesterol reduction in all patients with angina. In the absence of a reliable tool for risk stratification, this seems a more workable approach to secondary prevention.

A practice following the 1996 guideline would make relatively small changes to its protocols after this update. Was a comprehensive revision needed after three years? Decisions about when to revise are usually driven (as this one was) by arbitrary date cut offs. Shekelle et al recently published the first empirical evidence on how quickly guidelines become outdated.⁵ In their study of 17 clinical practice guidelines published by the US Agency for Healthcare Research and Quality, they found about half the guidelines were out of date in 5.8 years. There was, however, much variation. Guidelines in slowly evolving areas remained valid for longer, while in others they were rapidly out of date. Resources used to update guidelines are wasted if they do not change clinical practice.

Determining the optimal trade off between frequency of updating and cost is important. Though there is no simple solution, there are some possible strategies. Updates may be scheduled in response to changes in the evidence, either predicted by registers of research in progress, or identified retrospectively by rapid appraisals based on expert opinion and “sentinel literature.”⁵ Another approach is to evaluate the strength of evidence for recommendations in the baseline guidelines, prioritising updates to areas where new evidence is more likely to change practice.⁶ Brownman suggests that the solution to updating may be to do it more, rather than less frequently, taking advantage of electronic publishing to build continuously on the initial investment in developing a guideline.⁶ Identifying and appraising evidence is the most costly part of guideline development. Better collaboration between groups maintaining systematic reviews and those producing guidelines could also help to increase efficiency.⁷ Further research to evaluate these and other strategies is important to maximise the clinical payback on investment in evidence based guidelines.