OBJECTIVE Microleakage testing of dental materials is a commonly accepted evaluation technique of margin integrity. Thermocycling has been utilized by many researchers to study the influence of temperature extremes on the marginal gap of cemented restorations. The aim of this investigation was to analyze microleakage data on cemented crowns, published in the dental literature until Dec 2009, to identify methodological factors that might potentially affect the results of in vitro microleakage tests and to compare the results. METHODS The following databases were included: Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations and Ovid MEDLINE(R) 1950 to Present, Ovid-MEDLINE(R) 1950 to Present with Daily Update, EMBASE, EBM Reviews - Cochrane Database of Systematic Reviews and Pub Med. The search was limited to articles in English, French, Italian and German published until the end of 2009. Classical reviews, comments, animal studies, in vivo articles and studies investigating restorative materials or milk teeth were excluded. 33 different studies were finally selected. The studies were entered in a database and compared using selected literature criteria: sample, restoration procedures, thermocycling and mechanical cycling, evaluation method. For statistical analysis only 16 studies could be applied. RESULTS It was not possible to make a quantitative synthesis of most of the data, due to the heterogeneity of the studies concerning methods, treatment and outcome variables. Comparing the main groups of materials (ceramics, gold alloys and base metal alloys), no significant difference in the proportion of teeth without microleakage was found. Furthermore no significant difference in the proportion of teeth showing microleakage less than two third of the wall or teeth showing microleakage including the entire wall was found. Using the mean values in the meta-analysis instead of the proportions does not change the results. Confidence intervals could only be calculated for two materials (gold alloy, metal alloy). No difference between materials was found. CONCLUSIONS Comparison of the results from different studies is critical, since there are no generally accepted standards for experimental parameters, such as type and concentration of the storage solution, time of storage, temperature during storage, type and duration of thermal cycling and/or mechanical cycling, and the scoring criteria. There is lack of standardization of experimental conditions, which would ensure confidence in the studies and would further allow better comparability of various results.