BACKGROUND: Vertebral fractures are associated with increased morbidity (e.g. pain, reduced quality of life) and mortality. Therapeutic exercise is a non‐pharmacological conservative treatment that is often recommended for patients with vertebral fractures to reduce pain and restore functional movement. This is an update of a Cochrane Review first published in 2013. OBJECTIVES: To assess the effects (benefits and harms) of exercise intervention of four weeks or greater (alone or as part of a physical therapy intervention) versus non‐exercise/non‐active physical therapy intervention, no intervention or placebo among adults with a history of vertebral fractures on incident fragility fractures of the hip, vertebra or other sites. Our secondary objectives were to evaluate the effects of exercise on the following outcomes: falls, pain, physical performance, health‐related quality of life (disease‐specific and generic), and adverse events. SEARCH METHODS: We searched the following databases until November 2017: the Cochrane Library (Issue 11 of 12), MEDLINE (from 2005), Embase (from 1988), CINAHL (Cumulative Index to Nursing and Allied Health Literature, from 1982), AMED (from 1985), and PEDro (Physiotherapy Evidence Database, from 1929). Ongoing/recently completed trials were identified by searching the World Health Organization International Clinical Trials Registry Platform and ClinicalTrials.gov. Conference proceedings were searched via ISI and SCOPUS, and targeted searches of proceedings of the American Congress of Rehabilitation Medicine and American Society for Bone and Mineral Research. Search terms or MeSH headings included terms such as vertebral fracture AND exercise OR physical therapy. For this update, the search results were limited from 2011 onward. SELECTION CRITERIA: We included all randomized controlled trials and quasi‐randomized trials comparing exercise or active physical therapy interventions with placebo/non‐exercise/non‐active physical therapy interventions or no intervention implemented in individuals with a history of vertebral fracture. DATA COLLECTION AND ANALYSIS: Two review authors independently selected trials and extracted data using a pre‐tested data extraction form. Disagreements were resolved by consensus, or third‐party adjudication. We used Cochrane's tool for assessing risk of bias to evaluate each study. Studies were grouped according to duration of follow‐up (i.e. a) 4‐12 weeks; b) 16‐24 weeks; c) 52 weeks); a study could be represented in more than one group depending on the number of follow‐up assessments. For dichotomous data, we reported risk ratios (RR) and corresponding 95% confidence intervals (95% CI). For continuous data, we reported mean differences (MD) of the change from baseline and 95% CI. Data were pooled for Timed Up and Go test, self‐reported physical function measured by the QUALEFFO‐41 physical function subscale score (scale of zero to 100; lower scores indicate better self‐reported physical function), and disease‐specific quality of life measured by the QUALEFFO‐41 total score (scale of zero to 100; lower scores indicate better quality of life) at 12 weeks using a fixed‐effect model. MAIN RESULTS: Nine trials (n = 749, 68 male participants; two new trials in this review update) were included. Substantial variability across the trials prevented any meaningful pooling of data for most outcomes. Risk of bias across all studies was variable; low risk across most domains in four studies, and unclear/high risk in most domains for five studies. Performance bias and blinding of subjective outcome assessment were almost all high risk of bias. One trial reported no between‐group difference in favor of the effect of exercise on incident fragility fractures after 52 weeks (RR 0.54, 95% CI 0.17 to 1.71; very low‐quality evidence with control: 184 per 1000 and exercise: 100 per 1000, 95% CI 31 to 315; absolute difference: 8%, 95% CI 2 to 30). One trial reported no between‐group difference in favor of the effect of exercise on incident falls after 52 weeks (RR 1.06, 95% CI 0.53 to 2.10; very low‐quality evidence with control: 262 per 1000 and exercise: 277 per 1000; 95% CI 139 to 550; absolute difference: 2%, 95% CI ‐12 to 29). These findings should be interpreted with caution because of the very serious risk of bias in these studies and the small sample sizes resulting in imprecise estimates. We are uncertain that exercise could improve pain, self‐reported physical function, and disease‐specific quality of life, because certain studies showed no evidence of clinically important differences for these outcomes. Pooled analyses revealed a small between‐group difference in favor of exercise for Timed Up and Go (MD ‐1.13 seconds, 95% CI ‐1.85 to ‐0.42; studies = 2), which did not change following a sensitivity analysis (MD ‐1.09 seconds, 95% CI ‐1.78 to ‐0.40; studies = 3; moderate‐quality evidence). Exercise improved QUALEFFO‐41 physical function score (MD ‐2.84 points, 95% CI ‐5.57 to ‐0.11; studies = 2; very low‐quality evidence) and QUALEFFO‐41 total score (MD ‐3.24 points, 95% CI ‐6.05 to ‐0.43; studies = 2; very low‐quality evidence), yet it is unlikely that we observed any clinically important differences. Three trials reported four adverse events related to the exercise intervention (costal cartilage fracture, rib fracture, knee pain, irritation to tape, very low‐quality evidence). AUTHORS' CONCLUSIONS: In conclusion, we do not have sufficient evidence to determine the effects of exercise on incident fractures, falls or adverse events. Our updated review found moderate‐quality evidence that exercise probably improves physical performance, specifically Timed Up and Go test, in individuals with vertebral fracture (downgraded due to study limitations). However, a one‐second improvement in Timed Up and Go is not a clinically important improvement. Although individual trials did report benefits for some pain and disease‐specific quality of life outcomes, the findings do not represent clinically meaningful improvements and should be interpreted with caution given the very low‐quality evidence due to inconsistent findings, study limitations and imprecise estimates. The small number of trials and variability across trials limited our ability to pool outcomes or make conclusions. Evidence regarding the effects of exercise after vertebral fracture in men is scarce. A high‐quality randomized trial is needed to inform safety and effectiveness of exercise to lower incidence of fracture and falls and to improve patient‐centered outcomes (pain, function) for individuals with vertebral fractures (minimal sample size required is approximately 2500 untreated participants or 4400 participants if taking anti‐osteoporosis therapy).