One hundred and seventy studies were included in the review. The methodological quality of the included studies varied considerably.
Acute chest pain - clinical symptoms and signs (21 studies): No clinical feature in isolation was useful for diagnosing or excluding acute coronary syndrome. The most informative signs were pleuritic pain (positive LR 0.19, 95% CI 0.14 to 0.25) and pain on palpation (positive LR 0.23, 95% CI 0.08 to 0.30).
Acute chest pain - resting ECG (53 studies): The presence of ST elevation was highly specific for the diagnosis of myocardial infarction (positive LR 13.1, 95% CI 8.28 to 20.6). Completely normal ECG was useful for ruling out myocardial infarction (positive LR 0.14, 95% CI 0.11 to 0.20). Clinical interpretation of ECG (15 'black box studies' of real time decision making in evaluation of acute coronary syndrome) showed high positive likelihood ratios but low sensitivity.
Chronic chest pain - resting ECG (13 studies): Resting ECG features had limited usefulness. The positive likelihood ratio for the presence of Q-waves was 2.56 (95% CI 0.89 to 7.30). One study showed an likelihood ratio of 9.96 (95% CI 2.58 to 38.5) for QRS notching.
Chronic chest pain - exercise ECG (111 studies): With a 1mm cut-off the presence of ST depression had a positive likelihood ratio of 2.79 (95% CI 2.53 to 3.07) and a negative likelihood ratio of 0.44 (95% CI 0.40 to 0.47). The test performed better in men (LR 2.92, 95% CI 2.17 to 3.93) than in women (LR 1.92, 95% CI 1.72 to 2.24). With a 2mm cut-off, the positive likelihood ratio was 3.85 (95% CI 2.49 to 5.98) and the negative LR was 0.72 (95% CI 0.65 to 0.81).
Rapid assessment chest pain clinic (nine studies): There were no true evaluation studies. There was weak evidence that suggested that rapid assessment clinics were associated with reduced hospital admissions in patients with non-cardiac pain, better recognition of acute coronary syndrome, earlier assessment by specialists of exertional angina and earlier diagnosis of non-cardiac chest pain.