The purpose of this study was to determine if patients who do not receive manipulation for their low back pain (LBP) are at an increased risk for worsening disability compared to patients receiving an exercise intervention without manipulation. One hundred and thirty-one consecutive patients with LBP were randomly assigned to receive manipulation and an exercise intervention (n = 70) or an exercise intervention without manipulation (n = 61). Patients were classified as to whether they had experienced a worsening in disability upon follow-up. Relative risk and number needed to treat (NNT) statistics and associated 95% confidence intervals (CI) were calculated. Patients who completed the exercise intervention without manipulation were eight (95% CI: 1.1, 63.5) times more likely to experience a worsening in disability than patients who received manipulation. The NNT with manipulation to prevent one additional patient from experiencing a worsening in disability was 9.9 (95% CI: 4.9, 65.3) and 4 weeks with manipulation was 11.6 (95% CI: 5.2, 219.2). The results of this study offer an additional perspective for considering the risks and benefits of spinal manipulation and help to inform the integration of current evidence for spinal manipulation into healthcare policy.