Poster Presentation Australian Diabetes Society and the Australian Diabetes Educators Association Annual Scientific Meeting 2017

Correlates of changes in pH, chloride and ketone levels in diabetic ketoacidosis treated with weight-based insulin infusion with or without Plasma-Lyte® fluid resuscitation (#320)

Averil Y Tam 1 , Yvette C Goodgame 1 , Angela Graves 1 , Kartik R Atre 2 , Ken Yan Thong 3
  1. General Medicine, Rockingham General Hospital, Cooloongup, WA, Australia
  2. Intensive Care Unit, Rockingham General Hospital, Cooloongup, WA, Australia
  3. Department of Diabetes and Endocrinology, Rockingham General Hospital, Cooloongup, WA, Australia

Introduction

Normal saline is usually used for fluid resuscitation in diabetic ketoacidosis (DKA) but may result in hyperchloraemia or hyperchloraemic acidosis. We hypothesised that Plasma-Lyte®, a crystalloid fluid with lower amounts of chloride, would aid improvement of acidosis by avoiding hyperchloraemia in DKA.

 

Methods

We analysed all patients admitted with DKA to Rockingham General Hospital Intensive Care Unit between 1 July 2011 and 11 Jan 2015. Changes of pH, chloride and ketone levels closest to 6 hours were assessed and correlated with patients’ age, sex, baseline levels of pH, chloride, ketones, lactate and creatinine, total insulin dose used, total volume of crystalloid used and ≥50% use of Plasma-Lyte®. Significant correlates were entered into stepwise regression.

 

Results

Data on 42 patients (28 males, 14 females) with mean±SD, age 28±11 years, pH 7.19±0.11 and capillary ketones 5.9±1.0 mmol/L were analysed. 9 patients used ≥50% Plasma-Lyte® for fluid resuscitation, 30 patients did not use any Plasma-Lyte while 3 used Plasma-Lyte® comprising 3-38% of the total volume of fluid resuscitation. Mean changes from baseline for pH, chloride and ketones were 0.13±0.01, 9±5mmol/L and -3.8±1.8mmol/L, respectively. In final stepwise regression models, pH improvement correlated with lower baseline pH (p<0.001), higher baseline lactate (p=0.014) and older age (p=0.018). Chloride increase correlated inversely with ≥50% Plasma-Lyte®  use (p=0.002) and inversely with higher baseline chloride levels (p=0.006). Ketone improvement correlated with higher baseline pH (p<0.001), higher baseline ketone levels and female sex (p=0.012).

 

Conclusion

In a protocol-driven insulin treatment for DKA, the use of Plasma-Lyte® reduces the increase in chloride levels but did not aid improvement in acidosis levels compared with other crystalloid fluids.