The aim of this review was to identify studies on patterns of associative multimorbidity, defined as the nonrandom association between diseases, focusing on the main methodological features of the studies and the similarities among the detected patterns.

Studies were identified through MEDLINE and EMBASE electronic database searches from their inception to June 2012 and bibliographies.

The final 14 articles exhibited methodological heterogeneity in terms of the sample size, age and recruitment of study participants, the data source, the number of baseline diseases considered, and the statistical procedure used. A total of 97 patterns composed of two or more diseases were identified. Among these, 63 patterns were composed of three or more diseases. Despite the methodological variability among studies, this review demonstrated relevant similarities for three groups of patterns. The first one comprised a combination of cardiovascular and metabolic diseases, the second one was related with mental health problems, and the third one with musculoskeletal disorders.

The existence of associations beyond chance among the different diseases that comprise these patterns should be considered with the aim of directing future lines of research that measure their intensity, clarify their nature, and highlight the possible causal underlying mechanisms.