Physical education (PE)-based interventions are a popular method to target children's physical activity (PA) and fitness; however, little is known about their effectiveness or what factors lead to successful interventions. This paper: (1) systematically reviews studies examining PE interventions designed to impact PA, fitness, and/or body composition; and (2) makes recommendations for new research directions based upon these findings. Our systematic review was limited to experimental and quasi-experimental studies conducted in elementary schools. We conducted literature searches using predetermined keywords in 3 databases, identified a total of 4964 potentially relevant studies, and screened their abstracts and full texts for eligibility. This resulted in 12 relevant studies. We used criteria established by Downs and Black (1998) to assess each study's methodological quality. PE interventions consistently showed increases in moderate-to-vigorous PA or vigorous PA during PE class but were less consistent in impacting leisure-time PA. PE interventions affected body composition differentially, depending on the assessment used (i.e., body mass index or skinfold thickness). Half of the studies assessing fitness did not show a significant impact; however, those that did were designed to influence fitness outcomes. Few studies assessed psychosocial determinants regarding PA, and no study demonstrated significant impacts on constructs other than knowledge. Interventions often contained multiple components (e.g., diet, family) implemented alongside PE interventions. Identifying effective intervention components was difficult due to lack of process evaluation. We identify the need for future research to use more objective and accurate PA measurements and adiposity, incorporate measurement of psychological constructs, expand interventions' theoretical basis, and include strong process evaluation.