Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Are my patients like this?


One of the questions we are supposed to ask when interpreting the results of randomised trials into clinical practice is whether the patients in the trial are like our patients? That may be a relatively simple judgement, or it might be rather more difficult.

Especially in the elderly the presence of age-related changes, plus the likely presence of other conditions makes it all a bit complicated. For instance, is a 65 year old patient with heart failure likely to be the same as the 85 year old? A large study from the US [1] suggests we may need our thinking caps on more often.


The study was part of the US National Heart Failure Project. Hospital charts of patients discharged over the year to March 1999 with a principle diagnosis of heart failure were examined. The goal was to examine 800 discharge charts for each State. When fewer than 800 were available, all were examined. When more than 800 were available, 800 were chosen randomly. The discharge codes used had been shown to have high specificity for heart failure, and this was confirmed in a review of 100 charts by a cardiologist.

From the charts, 192 items of information were collected, including administrative data and information on history, physical examination and laboratory results.


There were 34,587 charts available for patients with a discharge diagnosis of heart failure and who were 65 years or older. The average age was 79 years, and 58% were women. The mean left ejection fraction was 69%, and 31% had a left ejection fraction above 40%.

Figure 1: Age of 35,000 heart failure patients

Other disorders were frequently present. There was a history of hypertension in 61%, coronary artery disease in 56%, diabetes in 38%, COPD in 33%, atrial fibrillation in 30%, history of stroke in 18% and dementia in 9%. About 10% would have a creatinine above 225 μmol/L.

There were important age-related changes across these patients. Firstly, within this restricted age range there were more older than younger patients, with a preponderance of over-85 year olds (27%). The presence of other conditions also varied with age. Though a history of hypertension was static at about 60%, coronary artery disease, diabetes and COPD were less common in over-85s than younger patients, while atrial fibrillation was more common (Figure 2). Dementia rose steadily from 3% in the youngest age group to 16% in over-85s.

Figure 2: Co-morbid conditions in heart failure patients, by age


Now this is a particular sample of patients who have been in hospital, primarily with problems relating to their heart failure. It may not be exactly similar to patients with heart failure who have not been admitted to hospital. But it serves to remind us that elderly patients with heart failure are likely to have co-morbid conditions and poor renal function. Even within the restriction of the over-64s there are age-related differences.

Interpreting the results of clinical trials in the light of this large survey will be interesting. It also impacts on development of guidance for treating older patients with heart failure.


  1. EP Havranek et al. Spectrum of heart failure in older patients: results of the National Heart Failure Project. American Heart Journal 2002 143: 412-417.
previous or next story in this issue