Context: Feeding behaviour is an important tool for the early detection of diseases in dairy cows. Aims: The aim of this study was to evaluate whether the variation in time spent feeding in the prepartum and postpartum periods may be used to detect the occurrence of metritis and subclinical ketosis (SCK) before the onset of the clinical symptoms at the postpartum. Methods: The research was conducted in four electronic databases, including Scopus, Science Direct, Pubmed, and Web of Science. The inclusion criteria for citations were original research, evaluation of daily time spent feeding in dairy cows, and use of this indicator for early identification of metritis and/or SCK in dairy cows in the prepartum and postpartum periods. A random-effect meta-analysis (MA) was conducted for metritis with the time spent feeding means of control (healthy) and treated (sick) groups measured in the prepartum and postpartum periods. The analysis was conducted with the values of daily time spent feeding before and after calving in both groups. Key results: In total, 26 trials from six papers, involving 1037 dairy cows, were included in the statistical analysis. No data were obtained for SCK to conduct a MA, while for metritis, 16 trials from six papers (prepartum) and 10 trials from three papers (postpartum) were considered. The heterogeneity between studies on metritis was moderate (I2 = 57.5%) in the prepartum period and low (I2 = 10.0%) in the postpartum period. The mean difference feeding time for healthy and unhealthy animals was greater during postpartum (21.14 min/day, P