Criteria for Selection of Cases

Computer database searches at the University of Minnesota Small Animal Veterinary Medical Center were used to identify retrospectively dogs with a cytologic or histologic diagnosis of blastomycosis between February 1997 to November 2006. Medical records were evaluated for laboratory data and information about patient demographics, treatment, and concurrent diagnoses. Dogs were included in the study if they had measurements of tCa, total protein, and iCa concentrations determined at the University of Minnesota Veterinary Clinical Pathology Laboratory at the first hospital admission or at a subsequent hospital admission for disease progression. It was necessary to perform these measurements on the same day. Dogs were excluded if calcium concentrations were only available after the patient had been treated with an antifungal drug unless disease progression was documented in the medical record as deteriorating clinical signs or radiographic evidence of progressive disease. Records were examined for evidence of previous treatment with corticosteroids. Some dogs in the study did have a history of previous empiric treatment with antimicrobials or antitussives and fluid therapy prior to hospital admission. Examination of the medical records did not reveal any concurrent diagnoses that would have been expected to alter calcium status, such as neoplasia, hyper- or hypoadrenocorticism, parathyroid disorders, renal failure, hyper- or hypovitaminosis D, malabsorptive enteropathies, or pancreatitis.¹

Procedures

Analysis of calcium concentration—Blood iCa concentration was measured by use of ion-sensitive biosensors (coefficients of variation < 3%).^a-c Blood samples were collected and handled anaerobically into syringes containing dry heparin.^d The protocol for use of the syringes requires filling to just below the 1-mL mark; overfilling must be avoided to prevent clotting. Serum tCa, total protein, and albumin concentrations were measured by use of spectrophotometric methods (coefficients of variation ≤ 2%).^e,f

Adjusted serum tCa concentrations (mg/dL) were calculated on the basis of serum total protein concentration (mg/dL) or serum albumin concentration (mg/ dL) by use of the following formulas¹⁰:

View Expanded

Calcium status—If multiple measurements were performed during hospitalization, only the first value was used to minimize the impact of treatment on data. Dogs were classified as hypercalcemic, normocalcemic, or hypocalcemic on the basis of established reference intervals. For blood iCa concentration, the established reference interval was 4.9 to 6.0 mg/dL, which is similar to published ranges.¹¹ Several changes in reagent systems occurred during the study, including replacement of 1 chemistry analyzer with another model. Reference intervals for serum tCa, total protein, and albumin concentrations were updated appropriately with each modification of method, and patient values were categorized as high, low, or normal, compared with the appropriate reference interval for the method used. The AdjCa-Alb and AdjCa-TP were calculated, and dogs were reclassified with the corrected value to determine how many would have a change in calcium status. Classifications based on serum tCa concentration, AdjCa-Alb, AdjCa-TP, and blood iCa concentrations were compared to determine whether serum tCa concentration, AdjCa-Alb, or AdjCa-TP most accurately predicted blood iCa status.

Statistical analysis—Linear regression analysis and correlations were performed by use of statistical analysis software^g to determine how closely serum tCa concentration, AdjCa-Alb, and AdjCa-TP correlated with blood iCa concentration. Linear regression analysis and correlations were also performed to evaluate the relationship between the classifications of high, low, and normal for concentrations of tCa, iCa, total protein, and albumin. Correlations were performed between tCa, iCa, total protein, and albumin concentrations and survival to discharge. Significance for all analyses was defined as the probability that the null hypothesis (eg, no relationship or no difference) was rejected, although true; values of P < 0.05 were considered significant.

Results

Study population—Breed representation included 15 Labrador Retrievers; 3 mixed-breed dogs; 2 Golden Retrievers; 2 Malamutes; 2 Rottweilers; and 1 of each of the following: Australian Shepherd, Bloodhound, Boxer, Brittany Spaniel, Cocker Spaniel, Doberman Pincher, German Shepherd Dog, German Shorthair Pointer, Great Dane, Husky, Irish Wolfhound, Sharpei, Springer Spaniel, and Weimeraner. Twenty dogs were neutered males, 6 were sexually intact males, 10 were spayed females, and 2 were sexually intact females. Findings from 22 of these dogs are presented elsewhere as a subset of data on the radiographic manifestations of pulmonary blastomycosis and on their response to medical treatment.^12,13 Mean age of dogs at the time of first hospital admission was 4.4 years (median, 3.5 years; range, 7 months to 9.8 years). Mean weight of dogs at hospital admission was 32.22 kg (71 lb; [median, 31.6 kg {69.5 lb}; range, 14 to 64 kg {30.8 to 140.8 lb}]).

Of the 38 dogs in the study, 2 dogs had been treated with antifungal medications prior to measurement of blood iCa and serum tCa concentrations. One dog had been receiving antifungal treatment for 5 days, and the other had been receiving antifungal treatment for 7 days. The second dog had also been receiving 7 days of ophthalmic corticosteroid treatment. Of these 2 dogs that had received antifungal medications, 1 had a blood iCa concentration within the reference interval and 1 had a high blood iCa concentration. Six of the 38 dogs had received systemically administered corticosteroids: 1 for 30 days, 1 for 14 days, 2 for 5 days, 1 for 4 days, and 1 for an undetermined number of days prior to the collection of blood samples for calcium measurements. All dogs that received systemically administered corticosteroids had blood iCa concentrations values within the reference interval.

Calcium status—All but 2 dogs in the study had blood iCa concentrations that were within the reference interval, and those 2 dogs (5% of the study population) were hypercalcemic with blood iCa concentrations of 6.5 and 7.4 mg/dL, respectively (Table 1). In contrast, serum tCa concentration overestimated hypocalcemia in 57.9% (22/38) of dogs and underestimated hypercalcemia in 1 dog, compared with blood iCa concentration. The use of AdjCa-Alb correctly reclassified all dogs as normocalcemic that had been determined to be hypocalcemic by use of serum tCa concentrations; however, the use of AdjCa-Alb failed to predict hypercalcemia in the same patient that was missed by use of serum tCa concentration. The use of AdjCa-TP correctly reclassified all but 2 dogs that were classified as hypocalcemic on the basis of serum tCa concentration and failed to predict hypercalcemia in the same patient that was missed by use of serum tCa concentration and AdjCa-Alb. Thirty-one of 38 (81.6%) dogs had hypoalbuminemia, whereas the remaining 7 dogs had serum albumin concentrations within the reference interval. Five dogs were hyperproteinemic, 7 were hypoproteinemic, and 26 had serum total protein concentrations within the reference interval.

Serum total protein concentration was significantly correlated with serum albumin concentration (correlation coefficient, 0.702) and serum tCa concentration (correlation coefficient, 0.713). No significant correlation was found between blood iCa concentration and any of the following: serum tCa concentration, serum total protein concentration, AdjCa-Alb, AdjCa-TP, or serum albumin concentration. Despite the use of AdjCa-Alb and AdjCa-TP to qualitatively predict normocalcemia as defined by blood iCa concentration, quantitative linear regression analyses performed with a 95% mean prediction interval revealed poor correlation between blood iCa concentration and serum tCa concentration (R² = 0.21), AdjCa-Alb (R² = 0.36), and AdjCa-TP (R² = 0.27).

View Full Size

Classification of the calcium status of 38 dogs with blastomycosis as determined by measurement of serum tCa concentration, blood iCa concentration, AdjCa-Alb, and AdjCa-TP.

Citation: Journal of the American Veterinary Medical Association 231, 10; 10.2460/javma.231.10.1545

• Download Figure
• Download figure as PowerPoint slide

Discussion

The demographic characteristics of the population of dogs with blastomycosis in our study are similar to those in previous reports^4,5 with large-breed dogs predominating. The sex distribution is similar to that in 1 study⁴ in which sexually intact male and neutered male dogs were overrepresented; however, results of another study⁵ do not demonstrate this difference. Our results indicate that an increase in the biologically active fraction of calcium is not common in dogs with blastomycosis and may occur in approximately 5% of the population admitted to a large referral institution during a 10-year period.

The 2 dogs that had hypercalcemia in our study were 4-to 5-year-old male dogs with unremarkable medical histories: 1 neutered Bloodhound and 1 sexually intact Labrador Retriever. The Bloodhound had peripheral lymphadenopathy and an increase in lung sounds. Smears of the affected lymph nodes for cytologic examination revealed pyogranulomatous inflammation with large numbers of blastomycosis organisms. Serologic results of this dog for fungal infections were negative for histoplasma, coccidioides, and aspergillus, but positive for blastomycosis. The Labrador Retriever had coughing and enlarged lymph nodes at the time of admission; blastomycosis organisms with pyogranulomatous inflammation were observed on cytologic examination of affected lymph nodes. In both affected dogs, respiratory compromise responded to antifungal treatment and supportive care, and both dogs were discharged. Neither dog had any clinical or laboratory evidence of neoplasia, renal failure, endocrinopathy, or other potential causes of hypercalcemia other than the infection with blastomycosis.

The association between hypercalcemia and granulomatous disease appears to be well documented in the human medical literature.^14,15 It has been shown that proliferating activated macrophages in the lesions of sarcoidosis convert 25-hydroxyvitamin D to the active form 1,25-dihydroxyvitamin D in an unregulated manner.¹⁵ This mechanism has been extrapolated to explain hypercalcemia in other granulomatous diseases in people; however, people with coccidioidomycosis-asso-ciated hypercalcemia did not consistently have high serum concentrations of 1,25-dihydroxyvitamin D.¹⁵ The association between hypercalcemia and granulomatous diseases in domestic animal species is not as well documented, and mechanisms have not been thoroughly investigated. In a clinical report¹⁶ of a dog with hypercalcemia associated with granulomatous lymphadenitis of undetermined etiology, serum concentrations of 1,25-dihydroxyvitamin D were high with suppressed serum parathyroid hormone concentrations and serum concentrations of parathyroid related protein were within reference interval. Those findings support the potential for high serum 1,25-dihydroxyvitamin D concentrations to contribute to hypercalcemia in dogs with granulomatous disease, but they do not confirm this mechanism in dogs with fungal disease in general, or blastomycosis in particular.¹⁶

To our knowledge, blood iCa concentrations of dogs with blastomycosis has not been reported previously. In a study⁴ of 47 dogs with blastomycosis, 3 of 47 (6.4%) dogs had high serum tCa concentrations, whereas 2 of 40 (2.5%) dogs in our study had high serum tCa concentrations. In another study⁵ of 115 dogs with blastomycosis, a larger proportion of dogs had high serum tCa (13.7%) concentrations after the application of quantitative correction formulas for hypoalbuminemia, whereas only 1 of 38 dogs in our study had high corrected serum tCa concentrations. It is difficult to compare our data directly with that of the other study⁵ because the formula used to correct serum tCa concentrations in that study was not reported and blood iCa concentrations were not determined. Arceneaux et al⁵ reported a similar proportion of dogs with hypoalbuminemia (77% vs 81.6% in our study), but did not directly indicate how many dogs were hypo-, normo-, or hypercalcemic on the basis of serum tCa concentrations or how many had a change in calcium status after the correction formula was applied. In that study,⁵ the same reference intervals were used during a 15-year period, whereas our reference intervals were updated several times during the 10-year period and patient data were compared with the reference interval in use at the time of hospital admission.

Our results indicate that serum tCa concentrations in dogs with blastomycosis do not correlate well with blood iCa concentrations and also overestimate the number of dogs with hypocalcemia. In our study, quantitative correction factors appeared to improve the use of serum tCa concentrations to estimate concentrations of the biologically active iCa; however, use of serum tCa concentration, AdjCa-Alb, and AdjCa-TP all appeared to slightly underestimate hypercalcemia in our patient population. Despite the improved qualitative classification of normal calcium status versus hypocalcemia by use of correction factors, results of linear regression analysis did not support a strong quantitative association between the corrected calcium concentrations and blood iCa concentrations. Therefore, measurement of blood iCa concentration should be performed if a clinical rationale exists for accurate determination of calcium status.

Current textbooks present quantitative correction factors that may be used by clinicians to adjust for the potential of low serum albumin or total protein concentrations to negatively influence the serum tCa concentration via decreases in the protein-bound fraction of calcium in the blood.^8,9 In presenting these formulas, authors qualify that the corrections have been validated only for dogs and that the adjustments may not accurately predict iCa status, particularly in patients with substantial acid-base abnormalities. Results of studies^10,11 in dogs with chronic renal failure and a variety of other disorders indicate that iCa concentration may vary widely from serum tCa concentration and that the application of quantitative correction factors may actually cause an increase in error in predicting iCa status as high, normal, or low in some patients. To our knowledge, iCa status and the diagnostic usefulness of quantitative correction factors have not been specifically evaluated in dogs with blastomycosis until now. Our findings suggest that qualitatively, hypoalbuminemia may contribute to a low serum tCa concentration despite a blood iCa concentration that is within the reference interval, but the poor correlation between the corrected value and serum tCa concentration does not support a predictable quantitative relationship. Furthermore, the use of serum tCa concentration, AdjCa-Alb, or AdjCa-TP may fail to identify a small number of patients with high blood iCa concentrations. It is likely that the impact of protein abnormalities on the use of serum tCa concentrations to predict iCa status should be evaluated for each disease process that is characterized by altered calcium metabolism and that measurement of blood iCa concentration will usually be the best way to characterize calcium concentrations in clinical patients.

Although our study provides new and clinically relevant information about the calcium status of dogs with blastomycosis, several methodologic issues should be considered when extrapolating our data to other patient populations. First, because the University of Minnesota Veterinary Medical Center is a referral hospital, the potential exists for bias towards sicker patients or patients with more severe pulmonary involvement that may have been referred for support in the intensive care unit. Likewise, much of the iCa data were collected as part of blood gas analysis, so patients with respiratory compromise requiring blood gas monitoring may have been more likely to have blood iCa and serum tCa concentrations determined on the same day. The potential exists for previous corticosteroid treatment in a small subset of patients to have lowered calcium concentrations and decreased numbers of hypercalcemic patients. Two dogs had been treated with antifungal drugs, which may also have lowered calcium concentrations⁶; however, previously treated dogs were only included if strong clinical evidence existed of disease progression despite treatment. None of these potential sources of bias should have substantially affected the assessment of correction factors or the use of serum tCa concentrations to predict iCa status in the study dogs, but they may have influenced the overall incidence of blastomycosis-associated hypercalcemia. Finally, collection of differing amounts of blood into the preheparinized syringes could have resulted in dilutional effects on measurement of blood iCa concentrations.