Repeated measurements of blood lactate concentration as a prognostic marker in horses with acute colitis evaluated with classification and regression trees (CART) and random forest analysis.
Authors: Petersen M B, Tolver A, Husted L, Tølbøll T H, Pihl T H
Journal: Veterinary journal (London, England : 1997)
Summary
# Editorial Summary Acute colitis carries a grave prognosis in equine practice, yet prognostic tools remain limited. Researchers evaluated 66 adult horses (including ponies and Icelandic horses) admitted with acute colitis over a nine-year period, measuring blood lactate, ionised calcium, packed cell volume and plasma total protein at admission and again at 6 hours post-admission, then analysing these variables using logistic regression, classification and regression trees, and random forest modelling. Blood lactate concentration emerged as the only individual parameter significantly associated with survival (P<0.001), with admission values ≥4.3 mmol/L combined with persistently elevated lactate (>2 mmol/L) at 6 hours predicting non-survival with 72% sensitivity and 80% specificity. For practitioners, these findings suggest that a single lactate measurement lacks sufficient predictive power alone—the key to more reliable prognostication lies in the trajectory of lactate levels in the first 6 hours post-admission, allowing clinicians to better counsel owners on realistic survival expectations in this high-mortality condition. This sequential approach could substantially improve decision-making and communication in acute colitis cases.
Read the full abstract on PubMed
Practical Takeaways
- •Measure blood lactate at admission and again at 6 hours post-admission in colitis cases—horses with initial lactate ≥4.3 mmol/L that remain >2 mmol/L at 6 hours have substantially worse prognosis and should be managed accordingly
- •Lactate is a more reliable individual prognostic indicator than ionised calcium, PCV, or total protein in acute colitis—use it as your primary biochemical prognostic tool
- •Serial lactate trends matter more than single values; persistent elevation despite treatment signals poor survival likelihood and should prompt honest owner communication about prognosis
Key Findings
- •Blood lactate concentration at admission was the only individual parameter significantly associated with survival probability (P<0.001)
- •A lactate cut-off value of 7 mmol/L at admission had sensitivity 0.66 and specificity 0.92 for predicting survival in training data
- •Combined assessment using lactate at admission ≥4.3 mmol/L and lactate >2 mmol/L at 6 hours post-admission achieved sensitivity 0.72 and specificity 0.8 for identifying non-survivors
- •Repeated lactate measurements at admission and 6 hours improved prognostic evaluation compared to single measurements in this high-mortality population (38% survival rate)