Twenty-two studies were included in the review. The number of included patients was not stated.
Eighteen studies were considered as level 1 to 3 for quality and were included in the meta-analysis.
Accuracy of the clinical examination (13 studies).
The overall clinical examination by the emergency department physician had a high positive LR (4.4, 95% CI: 1.8, 10.0) for a final diagnosis of heart failure. The odds of having heart failure when the clinical examination suggested that the patient was unlikely to have heart failure were decreased by about half (negative LR 0.45, 95% CI: 0.28, 0.73).
Medical history.
The most useful features in confirming the presence of heart failure were a medical history of congestive heart failure (positive LR 5.8, 95% CI: 4.1, 8.0), myocardial infarction (positive LR 3.1, 95% CI: 2.0, 5.9) or coronary artery disease (positive LR 1.8, 95% CI: 1.1, 2.8). Patients without a history of heart failure (negative LR 0.45, 95% CI: 0.38, 0.53), coronary artery disease (negative LR 0.68, 95% CI: 0.48, 0.96) or myocardial infarction (negative LR 0.69, 95% CI: 0.58, 0.82) were less likely to have heart failure. None of the other historical findings had statistically significant results.
Symptoms.
The most useful symptoms in confirming the presence of heart failure were paroxysmal nocturnal dyspnoea (positive LR 2.6, 95% CI: 1.5, 4.5), orthopnoea (positive LR 2.2, 95% CI: 1.2, 3.9) or dyspnoea on exertion (positive LR 1.3, 95% CI: 1.2, 1.4). Patients without symptoms of dyspnoea on exertion (negative LR 0.48, 95% CI: 0.35, 0.67), orthopnoea (negative LR 0.65, 95% CI: 0.45, 0.92) or paroxysmal nocturnal dyspnoea (negative LR 0.70, 95% CI: 0.54, 0.91) were less likely to have heart failure. None of the other symptoms had statistically significant results.
Physical examination.
The most useful physical examination features in confirming the presence of heart failure were the presence of a third heart sound (positive LR 11, 95% CI: 4.9, 25.0), jugular venous distension (positive LR 5.1, 95% CI: 3.2, 7.9), pulmonary rales (positive LR 2.8, 95% CI: 1.9, 4.1), any cardiac murmur (positive LR 2.6, 95% CI: 1.7, 4.1) or leg oedema (positive LR 2.3, 95% CI: 1.5, 3.7). Patients without pulmonary rales (negative LR 0.51, 95% CI: 0.37, 0.70), leg oedema (negative LR 0.64, 95% CI: 0.47, 0.87), jugular venous distension (negative LR 0.66, 95% CI: 0.57, 0.77) or wheezing (negative LR 0.52, 95% CI: 0.38, 0.71) were less likely to have heart failure. None of the other symptoms had statistically significant results.
Accuracy of chest radiographs (7 studies).
The most useful chest radiographic features in confirming the presence of heart failure were pulmonary venous congestion (positive LR 12, 95% CI: 6.8, 21.0), cardiomegaly (positive LR 3.3, 95% CI: 2.4, 4.7) or interstitial oedema (positive LR 12.0, 95% CI: 5.2, 27.0). Patients without cardiomegaly (negative LR 0.33, 95% CI: 0.23, 0.48) or pulmonary venous congestion (negative LR 0.48, 95% CI: 0.28, 0.83) were less likely to have heart failure.
Accuracy of ECG (7 studies).
The most useful ECG features in confirming the presence of heart failure were atrial fibrillation (positive LR 3.8, 95% CI: 1.7, 8.8), new T-wave changes (positive LR 3.0, 95% CI: 1.7, 5.3) or abnormal ECG findings (positive LR 2.2, 95% CI: 1.6, 3.1). Patients with a completely normal ECG (negative LR 0.64, 95% CI: 0.47, 0.88) were less likely to have heart failure.
Accuracy of serum BNP (11 studies).
A low serum BNP (less than 100 pg/mL) was the most useful test for excluding the presence of heart failure (negative LR 0.11, 95% CI: 0.07, 0.16). However, BNP levels should be interpreted differently for patients with renal insufficiency.
Results were also reported for a subgroup of patients in one of the included studies with a history of pulmonary disease.