Systematic review of feline diabetic remission: separating fact from opinion.

It is increasingly recognised that diabetic remission is possible in the cat. This systematic review, following Cochrane Collaboration (CC) guidelines, critically appraises the level of evidence on factors influencing remission rate and factors predicting remission. A systematic online, bibliographic search and reference list examination was conducted. A level of evidence was assigned to each identified article by five internists using the Newcastle-Ottawa Scale for follow-up, cohort, case-series and case-control studies, the CC's risk of bias tool for trials and the Cochrane Effective Practice and Organisation of Care Group risk of bias criteria for before and after trials. Twenty-two studies were included in the review, assessing influence of pharmaceutical intervention (n = 14) and diet (n = 4), as well as diagnostic tests (n = 9) and feline patient characteristics (n = 5) as predictors of remission. The current level of evidence was found to be moderate to poor. Common sources of bias included lack of randomisation and blinding among trials, and many studies were affected by small sample size. Failure to provide criteria for the diagnosis of diabetes, or diabetic remission, and poor control of confounding factors were frequent causes of poor study design. Addressing these factors would significantly strengthen future research and ultimately allow meta-analyses to provide an excellent level of evidence. No single factor predicts remission and successful remission has been documented with a variety of insulin types and protocols. Dietary carbohydrate reduction might be beneficial, but requires further study. A lack of well-designed trials prevents reliable remission rate comparison. Factors associated with remission resemble those in human medicine and support the hypothesis that reversal of glucotoxicity is a major underlying mechanism for feline diabetic remission.