Prolotherapy injections for chronic low‐back pain

Background Prolotherapy involves repeated injections of irritant solutions to strengthen lumbosacral ligaments and reduce some types of chronic low‐back pain; spinal manipulation and exercises are often used to enhance its effectiveness. Objectives To determine the efficacy of prolotherapy in adults with chronic low‐back pain. Search methods We searched CENTRAL 2006, Issue 3 and MEDLINE, EMBASE, CINAHL, and AMED from their respective beginnings to October 2006, with no restrictions on language, and consulted content experts. Literature search was updated on July 29th, 2009. No new RCTs were identified. Selection criteria We included randomised (RCT) and quasi‐randomised controlled trials (QRCT) that compared prolotherapy injections to control injections, alone or in combination with other treatments, which measured pain or disability before and after the intervention. Data collection and analysis Two review authors independently selected the trials and assessed methodological quality. Intervention protocols varied from study to study, making meta‐analysis impossible. Main results We identified five high quality studies with a total of 366 participants. All measured pain or disability levels at six months, and four measured the proportion of participants reporting a greater than 50% reduction in pain or disability scores. Three randomised controlled trials (206 participants) found that prolotherapy injections alone are no more effective than control injection for chronic low‐back pain and disability. At six months, there was no difference between groups in mean pain or disability scores (2 RCTs; 184 participants) and no difference in proportions who reported over 50% improvement in pain or disability (3 RCTs; 206 participants). These trials could not be pooled due to clinical heterogeneity. Two RCTs (160 participants) found that prolotherapy injections, given with spinal manipulation, exercise, and other therapies, are more effective than control injections for chronic low‐back pain and disability. At six months, one study reported a significant difference between groups in mean pain and disability scores, whereas the other study did not. Both studies reported a significant difference in the proportion of individuals who reported over 50% reduction in disability or pain. Co‐interventions confounded interpretation of results and clinical heterogeneity in the trials prevented pooling. Authors' conclusions There is conflicting evidence regarding the efficacy o