Longitudinal studies are critical to understanding health changes following trauma. The present systematic review adopted a longitudinal perspective on intimate partner abuse (IPA) by examining characteristics and findings of studies that followed persons who recently left an abusive relationship or who were at a point where they might leave and reported on health over time. Web of Science, EBSCO, Published International Literature on Traumatic Stress, and PubMed databases were searched using combinations of terms reflecting IPA, longitudinal study design, and health outcomes. Quantitative studies that recruited adults or adolescents, had multiple time points, used a health indicator as an outcome, and where IPA was a predictor, independent variable, or inclusion criterion were included. These methods yielded 36 studies from 20 unique samples. The following domains were coded for each article: citation, demographics, risk of bias, sample setting, design, follow-up, relationship and IPA characteristics, outcomes, and analytic focus. Results showed that all samples were female, and most were help-seeking. Depression, post-traumatic stress, and physical symptoms decreased over time, while quality of life increased in most studies. Changes in anxiety symptoms over time were less consistent. Ongoing IPA and social support were the most consistent predictors of health changes over time, showing relations with indicators of poorer or better health, respectively. There was preliminary evidence that decreases in emotional and physical symptoms plateaued within 9 months of the baseline assessment. Sampling, study design, measurement, and analysis are considered, and recommendations for future research are provided.