Estimation of perinatal mortality in dairy calves: A review.

• Perinatal mortality is defined as death of a full term calf either during parturition or shortly after (up to 24 h). • The unweighted mean of mortality estimates was 6.2%. • The review indicated that previous retrospective studies are likely to have underestimated perinatal mortality. Dairy calf mortality is of interest worldwide as animal welfare continues to increase in importance to the consumer. This review summarises overall estimations of perinatal mortality from published literature, then further explores the limitations of each publication as well as its applicability, and challenges faced in use of the generated data for comparisons. The first limitation is the definition of perinatal mortality. There is no agreed definition of the period over which perinatal mortality can occur and, depending on the source, estimates may include late-gestation abortions and deaths up to 48 h. In 26 studies identified by the literature search, 11 explicitly mentioned that they excluded abortion, by excluding calvings with too short a gestation length, or in a situation where calves were born in a tight cluster, excluded calvings that were more than 3 weeks prior to the planned start of calving. One study separated out the stillbirth percentage and abortion percentage. One study stated that abortion was included in their perinatal mortality data, and the remaining 13 studies included stillbirth but never mentioned any exclusion criteria for abortion. Seventeen studies measured mortality up to 24 h or less, six up to 48 h and three had unknown time periods. The second issue is study design. Of the 26 studies, seven were prospective studies and 19 were retrospective analyses. There are clear advantages to prospective studies as protocols can be put in place to maintain data quality. In contrast, ensuring data quality on retrospective studies is more difficult. One clear issue is the inclusion of data from herds with no perinatal mortality information. In a prospective study, herds that are not reporting any mortality can be easily identified and specific checks put in place to ensure recording is accurate; this cannot happen in a retrospective analysis. Thus retrospective studies which include data from herds with no perinatal mortality (especially when herd size is >100) are likely to be underestimating perinatal mortality. Across the studies, the unweighted mean of mortality estimates was 6.2%. However, if the objective is to compare countries, farming management, years or more, results must always be interpreted with caution. The variability in methods of data collection, validation, definition and even just applicability of the studies is large and if comparisons are to be made close examination of the methods of the studies being used as a comparison or benchmark for farms, regions or countries is crucial. [ABSTRACT FROM AUTHOR]