Abstract
BACKGROUND: Ionized calcium (iCa) is the biologically active fraction of total calcium (tCa) with clinical relevance to evaluate calcium homeostasis, but not all primary veterinarians have access to serum iCa. Formulas that adjust tCa to correct for variability in serum protein concentrations were not designed to predict iCa and are considered unreliable surrogates for iCa. OBJECTIVES: To determine whether adjusted calcium concentration (aCa) can predict ionized hypocalcemia in hypoalbuminemic dogs without hyperphosphatemia. ANIMALS: A total of 262 hypoalbuminemic dogs without hyperphosphatemia. METHODS: Retrospective review of paired tCa and iCa. Patients were included if serum albumin concentration was ≤2.5 g/L and serum phosphorus concentration was ≤5 mg/dL. The aCa was calculated using tCa (mg/dL) - serum albumin concentration (g/dL) + 3.5 (g/dL). Sensitivity, specificity, positive (PPVs) and negative (NPVs) predictive values, and accuracy were determined for tCa and aCa at predicting any (<1.13 mmol/L) and moderate (<1.02 mmol/L) ionized hypocalcemia. Patients also were stratified into mild-to-moderate (2.0-2.5 g/dL) and severe hypoalbuminemia (<2.0 g/dL). RESULTS: A total of 4296 dogs had paired results of which 262 met the inclusion criteria. Of these, 35 (13.4%) dogs had iCa < 1.13 mmol/L and 13 dogs (5.0%) had concentrations <1.02 mmol/L. The sensitivity, specificity, NPVs and PPVs of a decreased tCa and aCa for detecting moderate ionized hypocalcemia were 100% and 92.3%, 57.8% and 94.8%, 100% and 99.6%, and 11.0% and 48.2%, respectively, and accuracy was 60.0% and 94.7%, respectively. CONCLUSIONS: A low aCa was useful to detect ionized hypocalcemia in hypoalbuminemic nonhyperphosphatemic dogs. A normal aCa indicated that moderate ionized hypocalcemia was unlikely.