OBJECTIVES: When patients take medications as prescribed, they are more likely to achieve clinical goals and reduce their use of otherwise avoidable healthcare services.  A growing literature suggests that good medication adherence can also reduce healthcare costs for these patients.  Our objective was to review the literature to determine for which diseases the evidence for cost offsets was strongest, and identify why some studies show no cost offsets. METHODS: We conducted a literature review on medical cost offsets associated with medication adherence.  Key words searched were ("medication adherence" OR "medication compliance") AND ("medical care costs" OR "cost savings" OR "cost offset" OR "healthcare costs").   We limited the search to English language articles published between 1/1/2005-3/31/2010.   Studies were included if they reported the findings of original research. RESULTS: We identified 573 article titles, of which 100 met the inclusion criteria.  Overall, 44 studies directly measured both medication adherence and costs, and of them 81% found statistically significant reductions in costs associated with better adherence.  The range of cost offsets varied by disease, with the greatest offsets noted for congestive heart failure (costs were $4000-$8000 less per year for patients with a medication possession ratio [MPR] >80%), diabetes (range was $3700-$5700 less for those with an MPR>80%), and hypertension (range was $549-$3900 less for those with an MPR>80%).    Characteristics of studies where findings were cost neutral among adherent patients included:  conditions with high medication costs such as multiple sclerosis, and diseases where spending was shifted from medical care to prescription drugs without increasing total costs such as depression and asthma. CONCLUSIONS: Many studies confirm the clinical benefits of medications and their ability to reduce utilization.  A strong body of new studies now link those benefits to cost-offsets.  Even when cost-offsets are not found, medications tend to be cost neutral.