The use of pneumococcal polysaccharide vaccine has remained controversial since licensure, especially in the elderly. Observational studies form much of the evidence base. We conducted a systematic review of observational studies and compared results with those obtained from an earlier review of randomised controlled trials (RCTs). Estimates of protection against invasive disease from observational studies were consistent, homogenous and compatible with sparse information obtained from RCTs. Studies were of moderate quality. From 13 observational studies the estimate of vaccine efficacy against invasive disease was 53% (46-59%) compared with 38% (-4 to 63%) from nine RCTs. Estimates of protection against all-cause pneumonia were based on fewer, heterogeneous studies that were not consistent with the findings from RCTs for this outcome. From five studies combined efficacy was 32% (7-50%) compared with 3% (-16 to 19%) from 13 RCTs.