Reference intervals for plasma biochemical and hematologic measures in loggerhead sea turtles (Caretta caretta) from...

Biochemical and hematologic reference intervals have been reported for loggerhead sea turtles (Caretta caretta, Linnaeus 1758), but low sample numbers and simple statistical analyses have constrained their diagnostic usefulness. During June 2007-May 2008, 101 loggerhead sea turtles in Moreton Bay, Queensland, Australia, were captured by hand from boats; clinically assessed to determine health status; blood was sampled; and biochemical and hematologic variables were measured. Of these turtles, 66 were classified as clinically healthy and 23 as unhealthy. Reference intervals were calculated using data from clinically healthy turtles. Of the clinically unhealthy turtles, 82 and 45% had at least one biochemical and hematologic result, respectively, outside of at least one of the calculated intervals. However, only low proportions of unhealthy loggerhead sea turtles had abnormal results for each variable. The highest percentage of unhealthy turtles that were outside at least one estimated reference interval was 35%, for thrombocyte counts. Neither sex nor maturity category (mature versus large immature) influenced the risk of being clinically unhealthy. These are the first plasma biochemical and hematologic reference intervals reported for loggerhead sea turtles from the southwestern Pacific Ocean. We conclude that, for loggerhead sea turtles in Moreton Bay, separate reference intervals are required for mature and immature turtles for thrombocyte counts and for male and female turtles for lymphocyte, heterophil, and total white cell counts; otherwise, a single reference interval can be used regardless of age or sex. When estimating reference intervals in loggerhead sea turtles, it is desirable to use both methods for calculating reference intervals used in this study because intervals can differ substantially between methods for some variables. Joint interpretation using reference intervals from both methods allows the categorization of results as "normal," "suspect," or "abnormal."