Primary hyperparathyroidism in dogs is most commonly caused by autonomous and excessive secretion of PTH from a single parathyroid gland adenoma.1,2 Less commonly, dogs with PHPT may have multiple parathyroid gland adenomas, a parathyroid gland adenocarcinoma, or parathyroid gland hyperplasia.2–9 Because excessive secretion of PTH disrupts calcium homeostasis, dogs with PHPT may develop clinical signs associated with hypercalcemia, including polydipsia, polyuria, stranguria, lethargy, anorexia, vomiting, weakness, and weight loss.1,2,10 Hypercalcemia may be detected incidentally in dogs with PHPT that do not have clinical signs associated with the disorder.5,10
Surgical resection of affected parathyroid gland tissue is a common treatment for dogs with PHPT.1,2,10,11 Despite advances in imaging techniques (eg, high-resolution ultrasonography), the small size of parathyroid gland nodules and complicated anatomy in the parathyroid gland region can make preoperative identification and determination of the laterality (ie, right, left, or bilateral) of parathyroid gland lesions challenging and dependent on the quality of ultrasonographic equipment and operator skill.4,12 Results of 1 study13 indicated agreement between results of preoperative ultrasonography and surgical findings regarding laterality of parathyroid gland lesions in 13 of the 14 PHPT-affected dogs in the study. However, because of variations in performance of ultrasonography among operators and subjectivity in interpretation of results, surgical exploration of the ventral aspect of the cervical region may be undertaken without definitive preoperative determination of the laterality of a parathyroid gland lesion when results of serum or plasma biochemical analyses are suggestive of PHPT (ie, circulating iCa concentration higher than the reference range and circulating PTH concentration within or higher than the reference range).
A clinically important complication of surgical excision of autonomously functioning parathyroid gland tissue in PHPT-affected dogs is development of severe hypocalcemia.2 Results of another study13 that included 29 PHPT-affected dogs undergoing various preoperative (with or without vitamin D treatment) and intraoperative (ethanol ablation vs surgical excision of parathyroid gland lesions) treatments indicated serum tCa concentrations of dogs before surgery were greater for dogs that became hypocalcemic after surgery, compared with dogs that had serum concentrations of tCa within the reference range after surgery. Interestingly, dogs in that study13 with evidence of renal failure after surgery had significantly higher serum calcium concentrations before surgery versus dogs that did not have evidence of renal failure after surgery.
Other authors2 have suggested that PHPT-affected dogs with a serum tCa concentration > 14 mg/dL before surgery have a high risk of developing hypocalcemia after surgery and should therefore be treated prophylactically with calcitriol during the perioperative period. To the authors' knowledge, similar findings for serum iCa concentrations in PHPT-affected dogs have not been reported. In another study9 that included 210 dogs with PHPT, only presurgical clinical and laboratory analysis findings for dogs were reported and the relationships between such findings and development of hypocalcemia in dogs after surgery were not determined.9
The objective of the study reported here was to identify preoperative variables associated with postoperative hypocalcemia in PHPT-affected dogs undergoing parathyroidectomy. Our hypothesis was that dogs with high preoperative serum tCa and iCa concentrations would develop more severe hypocalcemia after surgery, compared with dogs with low serum tCa and iCa concentrations. Another objective of the study reported here was to determine the accuracy of high-resolution ultrasonography of the cervical region for preoperative determination of the laterality of parathyroid gland lesions in dogs with PHPT. Our hypothesis for this objective was that high-resolution ultrasonography of the cervical regions of dogs with PHPT would be useful for preoperative determination of the laterality of parathyroid gland lesions in such dogs.
Materials and Methods
Case selection—Medical records of dogs that underwent surgery between January 2004 and January 2009 at the University of Wisconsin, University of Georgia, Wisconsin Veterinary Referral Center, and Oregon State University were reviewed. Inclusion criteria were a diagnosis of PHPT and a reduction in serum tCa or iCa concentrations or serum tCa or iCa concentrations within the reference range after surgical excision of parathyroid gland tissue. In addition, dogs that underwent surgery and had a histologic diagnosis of a parathyroid gland adenoma or parathyroid hyperplasia were included. Dogs with incomplete medical records, dogs that had received calcitriol before or after surgery, and dogs with a histologic diagnosis of parathyroid gland adenocarcinoma were excluded from the study.
Medical records review—Data were obtained from medical records of dogs. Preoperative period binary variables included sex; neuter status; owner-reported history of lethargy, weakness, inappetence, polyuria and polydipsia, gastrointestinal tract signs (vomiting or diarrhea), or neurologic signs (tremors or seizures); azotemia (ie, circulating BUN or creatinine concentration higher than the reference range); urine specific gravity < 1.020; and PTH concentration (within or higher than the reference range). The PTH concentrations were evaluated as binary variables rather than continuous variables because serum samples of dogs were not all submitted to the same laboratory for determination of PTH concentrations. Preoperative period continuous variables included age; weight; number of days in the hospital; serum CPP (ie, serum calcium concentration × serum phosphorus concentration) and tCa, iCa, phosphorus, BUN, creatinine, and parathyroid-related protein concentrations; blood glucose concentration; and urine specific gravity. All tCa and iCa concentrations determined for serum samples collected during the postoperative period were recorded. The reference range for tCa was 8.9 to 11.8 mg/dL, and the reference range for iCa was 1.17 to 1.38 mmol/L. Typically, serum concentrations of tCa, iCa, or both were monitored for dogs after surgery. Dogs with ≥ 2 postoperative tCa or iCa concentration values recorded in the medical record were included in analyses of postoperative serum tCa or iCa concentrations, respectively. Dogs with ≥ 2 postoperative serum tCa and ≥ 2 postoperative serum iCa concentration values recorded in the medical record were included in analyses of both tCa and iCa concentrations in serum samples obtained during the postoperative period. Postoperative period binary variables included hypocalcemia (ie, serum tCa or iCa concentration lower than the reference range) and development of clinical signs associated with hypocalcemia (facial or paw pruritus, generalized signs of pain, muscle fasciculations, stiff gait, signs of anxiety, elevated rectal temperature, tremors, ataxia, or seizures) concurrent with a serum tCa or iCa concentration lower than the reference range.
Postoperative period continuous variables included serum tCa and iCa nadir concentrations. Nadir concentrations were defined as the lowest serum tCa and iCa concentration determined at any time during hospitalization after parathyroidectomy. For each dog, agreement between results of preoperative ultrasonography of the cervical region and intraoperative surgical findings were evaluated on the basis of scores (agreed, not agreed, or could not be determined from the medical records) for laterality (ie, left, right, or bilateral) of parathyroid gland lesions.
Statistical analysis—Data were evaluated via various methods. Initially, data were evaluated via simple linear regression or ANOVA to identify relationships between preoperative variables (continuous or categorical variables, respectively) and postoperative serum tCa and iCa nadir concentrations. Next, preoperative variables that had a relationship (P < 0.10) with postoperative tCa or iCa nadir concentrations were tested to identify strong correlations (r > 0.75) to determine multicollinearity among variables. Because no strong correlations were found, all tested variables were included in the multiple linear regression model. A stepwise selection method (cutoff, P < 0.10) was used to identify variables for inclusion in the model. All tests were 2-sided, and values of P < 0.05 were considered significant. Coefficients of determination (R2) were reported for complete models and for each variable in the models. Data for dogs with ≥ 2 postoperative serum tCa concentration values and data for dogs with ≥ 2 postoperative serum iCa concentration values were evaluated via separate multiple linear regression models. Analyses were performed with commercially available software.a
Continuous preoperative variables that were significantly (P < 0.05) correlated with postoperative serum tCa or iCa nadir concentrations on the basis of results of simple linear regression were further evaluated via receiver operator curve analysis to identify cutoff values of variables with the highest sensitivity and specificity for prediction of development of postoperative hypocalcemia in dogs. Values of P < 0.05 were considered significant. Data for dogs with ≥ 2 postoperative serum tCa concentration values and data for dogs with ≥ 2 postoperative serum iCa concentration values were evaluated via separate analyses.
A χ2 test was used to compare percentages of dogs with preoperative serum tCa concentrations of < 14 mg/dL, 14 to 15 mg/dL, and ≥ 15 mg/dL that developed hypocalcemia after surgery. A Student t test was used to compare preoperative urine specific gravity and serum creatinine and urea nitrogen concentrations between dogs with an owner-reported history of polyuria and polydipsia and dogs without such a history. Analyses were performed with commercially available software.b
Agreements between preoperative ultrasonography results and surgical findings regarding laterality of parathyroid gland lesions were determined and reported as percentages. Descriptive data were reported as mean ± SEM or percentage values.
Results
Signalment and descriptive data—Sixty-two dogs met the study inclusion criteria. Mean ± SEM preoperative serum tCa, iCa, and phosphorus concentrations of these dogs were 13.65 ± 0.18 mg/dL, 1.76 ± 0.03 mmol/L, and 2.86 ± 0.14 mg/dL, respectively. Of the 62 dogs included in the study, 32 (51.6%) were evaluated at the University of Wisconsin, 22 (35.5%) were evaluated at University of Georgia, 4 (6.5%) were evaluated at Wisconsin Veterinary Referral Center, and 4 (6.5%) were evaluated at Oregon State University. Mean ± SEM age and weight of dogs were 10.2 ± 0.2 years and 25.2 ± 1.2 kg (55.4 ± 2.6 lb), respectively. Thirty-one (50.0%) dogs were neutered males, 29 (46.8%) were spayed females, 1 (1.6%) was a sexually intact male, and 1 (1.6%) was a sexually intact female. Golden Retriever (n = 9 [14.5%]) was the most common breed; other breeds included mix (8 [12.9%]), Labrador Retriever (7 [11.3%]), English Springer Spaniel (7 [11.3%]), Keeshond (5 [8.1%]), Siberian Husky (4 [6.5%]), German Shepherd Dog (3 [4.8%]), Shih Tzu (3 [4.8%]), Beagle (2 [3.2%]), and Jack Russell Terrier (2 [3.2%]), and 1 dog of each of 12 other breeds (12 [19.4%]).
Thirty-four (54.8%) dogs had an owner-reported history of polyuria and polydipsia. Dogs with a history of polyuria and polydipsia had a preoperative mean ± SEM urine specific gravity of 1.013 ± 0.002, serum creatinine concentration of 1.15 ± 0.08 mg/dL, and serum BUN concentration of 18.8 ± 1.9 mg/dL. The 28 (45.2%) dogs without a history of polyuria and polydipsia had a preoperative mean ± SEM urine specific gravity of 1.016 ± 0.002, serum creatinine concentration of 1.08 ± 0.09 mg/dL, and serum urea nitrogen concentration of 19.0 ± 2.4. No significant differences were detected between dogs with a history of polyuria and polydipsia and dogs without such a history regarding preoperative urine specific gravity (P = 0.27) or serum creatinine (P = 0.59) or urea nitrogen (P = 0.95) concentration.
Results of histologic analysis of surgically excised parathyroid gland tissue indicated 49 (79.0%) dogs had parathyroid gland adenomas, 10 (16.1%) dogs had parathyroid gland hyperplasia, 2 (3.2%) dogs had lesions attributable to either parathyroid gland hyperplasia or adenoma, and 1 (1.6%) dog had histologically normal parathyroid gland tissue. The dog with histologically normal parathyroid gland tissue was included in the study because the preoperative serum PTH, tCa, and iCa concentrations for that dog were higher than the reference range, and concentrations of those analytes were within the reference range after surgical removal of the parathyroid gland.
Forty-two (67.7%) dogs underwent excision of 1 parathyroid gland; 16 (25.8%) underwent excision of 2 parathyroid glands, and 4 (6.5%) underwent excision of 3 parathyroid glands. Of the dogs that underwent excision of 2 parathyroid glands, 10 had adenomatous glands, 4 had hyperplastic glands, 1 had parathyroid glands for which a diagnosis of adenomatous or hyperplastic parathyroid gland tissue could not be determined on the basis of the surgical histopathology report, and 1 had both an adenomatous parathyroid gland and a hyperplastic parathyroid gland on the basis of the histopathology report. Of the dogs that underwent excision of 3 parathyroid glands, 2 dogs had only adenomatous parathyroid glands, 1 had adenomatous and hyperplastic parathyroid glands, and 1 had adenomatous and histologically normal parathyroid glands that were removed.
Of the 62 dogs in the study, 38 (61.3%) had a preoperative serum tCa concentration > 14 mg/dL; mean ± SEM serum tCa concentration for these dogs was 12.78 ± 0.13 mg/dL. Of these 38 dogs, 12 (31.6%) had at least 1 postoperative serum sample with an iCa or tCa concentration less than the reference range. Of the 62 dogs in the study, 12 (19.4%) had a preoperative serum tCa concentration between 14 and 15 mg/dL; mean ± SEM serum tCa concentration for these dogs was 14.5 ± 0.07 mg/dL. Of these 12 dogs, 4 (33.3%) had at least 1 postoperative serum sample with an iCa or tCa concentration less than the reference range. Of the 62 dogs in the study, 11 (17.7%) had a preoperative serum tCa concentration ≥ 15 mg/dL; mean ± SEM serum tCa concentration for these dogs was 15.75 ± 0.24 mg/dL; of these 11 dogs, 7 (63.6%) had at least 1 postoperative serum sample with an iCa or tCa concentration less than the reference range. A preoperative serum tCa concentration was not recorded in the medical record for 1 dog. No significant (P = 0.15) differences were detected among each of those preoperative tCa concentration groups (< 14 mg/dL, 14 to 15 mg/dL, and ≥ 15 mg/dL) regarding percentages of dogs that developed hypocalcemia after surgery.
Postoperative serum tCa concentration values were available for 32 dogs. Before surgery, these dogs had a mean ± SEM serum tCa concentration of 13.76 ± 0.27 mg/dL, serum iCa concentration of 1.80 ± 0.05 mmol/L, and serum phosphorus concentration of 2.68 ± 0.13 mg/dL. When calcium status was determined on the basis of serum tCa concentrations, dogs typically became normocalcemic 1 to 8 hours after surgery and the lowest serum tCa concentrations were detected 3 to 4 days after surgery (Figure 1). Seven of these 32 (21.9%) dogs had at least 1 postoperative serum sample with a tCa concentration less than the reference range. Of these 7 dogs, 2 (28.6%) developed clinical signs of hypocalcemia.
Mean ± SEM tCa concentrations in serum samples obtained from dogs with PHPT at various times after parathyroidectomy. Dashed lines represent upper and lower limits of the serum tCa concentration reference range. Data at 0 hours are preoperative tCa concentrations; data at 1 hour are serum tCa concentrations determined immediately after surgery. Data at all other times were determined for serum samples collected from dogs at the indicated time after surgery. Numbers in parentheses indicate number of dogs with data at each time.
Citation: Journal of the American Veterinary Medical Association 242, 4; 10.2460/javma.242.4.507
Postoperative serum iCa concentrations were available for 50 dogs. Before surgery, these dogs had a mean ± SEM serum tCa concentration of 13.60 ± 0.2 mg/dL, serum iCa concentration of 1.75 ± 0.04 mmol/L, and serum phosphorus concentration of 2.86 ± 0.17 mg/dL. Typically, serum iCa concentrations were within the reference range between 16 and 24 hours after surgery and the lowest serum iCa concentrations were detected approximately 2 to 3 days after surgery (Figure 2). Eighteen of these 50 (36.0%) dogs had at least 1 postoperative serum sample with an iCa concentration less than the reference range. Of these 18 dogs, 5 (27.8%) developed clinical signs of hypocalcemia.
Mean ± SEM iCa concentrations in serum samples obtained from dogs with PHPT at various times after parathyroidectomy. Dashed lines represent upper and lower limits of the serum iCa concentration reference range. Data at 0 hours are preoperative iCa concentrations; data at 1 hour are serum iCa concentrations determined immediately after surgery. Data at all other times were determined for serum samples collected from dogs at the indicated time after surgery. Numbers in parentheses indicate number of dogs with data at each time.
Citation: Journal of the American Veterinary Medical Association 242, 4; 10.2460/javma.242.4.507
Of the 62 dogs in the study, 24 (38.7%) had at least 1 postoperative serum sample with a tCa or iCa concentration less than the reference range. Of these 24 dogs, 6 (25%) developed clinical signs of hypocalcemia. The mean ± SEM preoperative serum tCa, iCa, and phosphorus concentrations for these 6 dogs were 14.9 ± 0.7 mg/dL, 1.89 ± 0.15 mmol/L, and 4.08 ± 0.86 mg/dL, respectively. For 1 of these dogs, only serum tCa concentrations were determined after surgery (serum tCa nadir concentration, 6 mg/dL). For 1 of these dogs, serum tCa and iCa concentrations were determined after surgery (serum tCa and iCa nadir concentrations, 7.1 mg/dL and 0.94 mmol/L, respectively). For 4 of these dogs, only serum iCa concentrations were determined after surgery (serum iCa nadir concentrations of 0.75, 0.83, 0.83, and 0.92 mmol/L).
Relationship between preoperative variables and postoperative serum tCa and iCa nadir concentrations—Preoperative variables that had a relationship (P < 0.10) with postoperative serum tCa and iCa nadir concentrations for dogs were summarized (Tables 1 and 2). No strong correlations (r > 0.75) were detected among the preoperative variables that had relationships with postoperative serum tCa and iCa nadir concentrations. Preoperative urine specific gravity values were excluded from multiple linear regression analysis because data were available for only 33 of the 62 dogs in the study. Variables with significant results in the multiple linear regression models were summarized (Table 3).
Preoperative patient variables identified via ANOVA (binary variables) or simple linear regression (continuous variables) significantly related to serum tCa nadir concentrations (ie, lowest detected postoperative serum tCa concentration) for 32 PHPT-affected dogs after parathyroidectomy for which postoperative serum tCa concentrations were determined.
| P value | Slope of regression line | Dogs without variable | Dogs with variable | |||
|---|---|---|---|---|---|---|
| No. (%) of dogs | Mean tCa nadir concentration (mg/dL) | No. (%) of dogs | Mean tCa nadir concentration (mg/dL) | |||
| Binary variables | ||||||
| Weakness | < 0.05 | — | 25 (78.1) | 9.80 | 7 (21.9) | 8.55 |
| Gastrointestinal tract signs | < 0.05 | — | 27 (84.4) | 9.32 | 5 (15.6) | 10.58 |
| PTH concentration greater than reference range | < 0.01 | — | 11 (34.3) | 10.3 | 21 (65.6) | 9.10 |
| Continuous variables | ||||||
| Age | < 0.05 | –0.73 | — | — | — | — |
| Serum CPP | < 0.001 | 13.50 | — | — | — | — |
| Serum iCa concentration | < 0.1 | –0.08 | — | — | — | — |
— = Not applicable.
Preoperative patient variables identified via ANOVA (binary variables) or simple linear regression (continuous variables) significantly related to serum iCa nadir concentrations (ie, lowest detected postoperative serum iCa concentration) for 50 PHPT-affected dogs after parathyroidectomy for which postoperative serum iCa concentrations were determined.
| P value | Slope of regression line | Dogs without variable | Dogs with variable | |||
|---|---|---|---|---|---|---|
| No. (%) of dogs | Mean iCa nadir concentration (mmol/L) | No. (%) of dogs | Mean iCa nadir concentration (mmol/L) | |||
| Binary variables | ||||||
| Sexually intact | < 0.05 | — | 48 (96.0) | 1.47 | 2 (4.0) | 1.21 |
| Polyuria and polydipsia | < 0.05 | — | 21 (42.0) | 1.14 | 29 (58.0) | 1.28 |
| Renal azotemia* | < 0.1 | — | 6 (14.3) | 1.15 | 36 (85.7) | 1.21 |
| Continuous variables | ||||||
| Body weight | < 0.01 | 14.24 | — | — | — | — |
| Serum urea nitrogen concentration | < 0.05 | –9.49 | — | — | — | — |
| Urine specific gravity | < 0.1 | –0.02 | — | — | — | — |
Renal azotemia was defined as a serum creatinine concentration > 2.0 mg/dL or a serum urea nitrogen concentration > 30 mg/dL and a urine specific gravity < 1.020. Results regarding renal azotemia are reported for 42 dogs because urinalysis was not performed concurrent with preoperative CBC and serum biochemical analysis assays for 8 dogs.
See Table 1 for remainder of key.
Results of multiple regression analysis indicating preoperative variables significantly associated with postoperative serum tCa (n = 32 dogs) or iCa (50) nadir concentrations in 62 PHPT-affected dogs that underwent parathyroidectomy.
| Variable | tCa | iCa |
|---|---|---|
| Complete model | 0.51 | 0.31 |
| Serum PTH concentration | 0.02 | NS |
| Age | 0.07 | NS |
| Body weight | NS | 0.23 |
| Serum CPP | 0.42 | NS |
| Serum urea nitrogen concentration | NS | 0.08 |
Values of P < 0.05 were considered significant. Data are R2 values. Complete model R2 values are the percentage of variability in serum tCa or iCa concentrations accounted for by all variables in the model. The R2 values for each variable in the model are also reported.
NS = Not significant.
Receiver operator characteristic curve analysis—On the basis of results of simple linear regression analysis, the following continuous preoperative variables were selected for receiver operator characteristic curve analysis: age and serum CPP (for postoperative serum tCa nadir concentrations) and body weight and serum urea nitrogen concentration (for postoperative serum iCa nadir concentrations). Results for serum CPP and urea nitrogen concentrations were not significant. Age was a significant factor for postoperative serum tCa nadir concentrations, and body weight was a significant factor for postoperative serum iCa nadir concentrations. The variable with the highest combined sensitivity and specificity for prediction of a postoperative serum tCa concentration less than the reference range was an age of > 10.5 years (Table 4). The variable with the highest combined sensitivity and specificity for prediction of a postoperative serum iCa concentration less than the reference range was a body weight of < 23.85 kg (< 52.47 lb; Table 5), although sensitivity and specificity were low.
Receiver operator curve analysis results for various age cutoff values as predictors of detection of at least 1 serum tCa concentration less than the reference range during the postoperative period for 32 PHPT-affected dogs that underwent parathyroidectomy and for which postoperative serum tCa concentrations were determined.
| Age (y) | Sensitivity (%) | 95% CI (%) | Specificity (%) | 95% CI (%) | Likelihood ratio |
|---|---|---|---|---|---|
| > 9.5 | 100.0 | 59.0–100.0 | 44.0 | 24.4–65.1 | 1.79 |
| > 10.5* | 71.4 | 29.0–96.3 | 60.0 | 38.7–78.9 | 1.79 |
| > 11.5 | 42.9 | 9.9–81.6 | 68.0 | 46.5–85.1 | 1.34 |
| > 12.5 | 42.9 | 9.9–81.6 | 96.0 | 79.7–99.9 | 10.71 |
| > 13.5 | 42.9 | 9.9–81.6 | 100.0 | 86.3–100.0 | — |
Results of the analysis were significant (P = 0.038); the area under the curve was 0.760 (SE, 0.098; 95% CI, 0.568 to 0.952). Serum tCa concentrations less than the reference range were not detected for 25 dogs, and a tCa concentration less than the reference range was detected in at least 1 serum sample for 7 dogs after surgery.
The highest combined sensitivity and specificity were associated with a cutoff value of 10.5 years.
CI = Confidence interval.
See Table 1 for remainder of key.
Receiver operator curve analysis results for various body weight cutoff values as predictors of detection of at least 1 serum iCa concentration less than the reference range during the postoperative period for 50 PHPT-affected dogs that underwent parathyroidectomy and for which postoperative serum iCa concentrations were determined.
| Body weight (kg [lb]) | Sensitivity (%) | 95% CI (%) | Specificity | 95% CI (%) | Likelihood ratio |
|---|---|---|---|---|---|
| < 5.95 (13.09) | 0.0 | 0.0–18.53 | 96.77 | 83.3–99.92 | 0.00 |
| < 17.10 (37.62) | 27.78 | 9.7–53.48 | 83.87 | 66.27–94.55 | 1.72 |
| < 23.85 (52.47)* | 66.67 | 40.99–86.66 | 67.74 | 48.63–83.32 | 2.07 |
| < 30.10 (66.22) | 77.78 | 52.36–93.59 | 45.16 | 27.32–63.97 | 1.42 |
| < 32.75 (72.05) | 100.0 | 81.47–100.0 | 32.26 | 16.68–51.37 | 1.48 |
Results of the analysis were significant (P = 0.029); the area under the curve was 0.688 (SE, 0.077; 95% CI, 0.538 to 0.839). Serum iCa concentrations less than the reference range were not detected for 32 dogs, and an iCa concentration less than the reference range was detected in at least 1 serum sample for 18 dogs after surgery.
The highest combined sensitivity and specificity were associated with a cutoff value of 23.85 kg (52.47 lb).
See Table 4 for remainder of key.
Agreement between ultrasonography results and surgical findings—Fifty-eight of the 62 (93.5%) dogs in this study underwent ultrasonography of the ventral aspect of the cervical region prior to surgery. For those 58 dogs, results of 44 (75.9%) ultrasonographic reports agreed with surgical findings regarding laterality of abnormal parathyroid gland tissue. For 11 (19.0%) dogs, ultrasonography results did not agree with surgical findings regarding laterality of parathyroid gland lesions. Agreement between ultrasonographic reports and surgical findings could not be determined for 3 (5.2%) dogs because laterality of parathyroid gland lesions could not be determined on the basis of ultrasonographic reports for these dogs.
Discussion
The objective of the present study was to identify preoperative variables associated with postoperative hypocalcemia in dogs with PHPT that underwent parathyroidectomy. Our hypothesis was that dogs with high preoperative serum concentrations of tCa and iCa would have more severe postoperative hypocalcemia versus dogs with low preoperative serum concentrations of tCa and iCa. That hypothesis was not supported, because results did not indicate a correlation between preoperative serum tCa and iCa concentrations and postoperative serum tCa and iCa nadir concentrations. No single preoperative variable fully accounted for variability in postoperative serum tCa and iCa concentrations.
Other authors2 have suggested that dogs with a circulating tCa concentration > 14 mg/dL are at greater risk of developing postoperative hypocalcemia versus dogs with lower circulating tCa concentrations. To the authors' knowledge, validation of that cutoff value via statistical analysis has not been reported in peer-reviewed literature. Results of the present study indicated 12 of 38 (31.6%) dogs with a preoperative serum tCa concentration < 14 mg/dL and 11 of 23 (47.8%) dogs with a preoperative serum tCa concentration ≥ 14 mg/dL were hypocalcemic during the postoperative period. Although the percentage of dogs that developed hypocalcemia after surgery was higher for the group of dogs with high preoperative serum tCa concentrations versus dogs with low concentrations, these data were not significantly different. This finding of the present study did not support the suggestion2 that dogs with preoperative serum tCa concentrations > 14 mg/dL are at greater risk of developing postoperative hypocalcemia versus dogs with lower concentrations; that finding may be attributable to type II statistical error because of a small number of dogs included in this study or because of a true lack of association between preoperative serum tCa concentrations > 14 mg/dL and postoperative hypocalcemia.
Variables significantly associated with low postoperative serum tCa nadir concentrations in dogs in this study included old age, history of weakness, lack of gastrointestinal tract signs, preoperative serum PTH concentration higher than the reference range, and low serum CPP value. Preoperative serum tCa concentrations were not significantly associated with postoperative serum tCa nadir concentrations, and results did not support a serum tCa concentration cutoff value of 14 mg/dL.2 Preoperative variables significantly associated with low postoperative serum iCa nadir concentrations in dogs in this study included sexually intact neuter status, low body weight, high serum urea nitrogen concentration, and absence of polyuria and polydipsia in the history. Although results indicated a significant association between an owner-provided history of polyuria and polydipsia and a low postoperative serum iCa nadir concentration, a biological reason for these findings could not be identified.
Results of this study contradict results of another study13; PHPT-affected dogs in that study13 that became hypocalcemic after parathyroidectomy had higher preoperative serum tCa concentrations versus dogs in the present study. However, 13 of 19 dogs that underwent surgery in that other study13 received vitamin D before surgery and data for those dogs were not analyzed separately; therefore, findings of that study are difficult to interpret. Additionally, results of that other study13 indicated postoperative renal failure of dogs was associated with high preoperative circulating calcium concentrations. Interestingly, results of the present study suggested that high preoperative serum urea nitrogen concentrations were associated with low postoperative serum iCa concentrations in dogs. However, interpretation of this finding is difficult because PHPT-affected dogs in the present study may have had concomitant conditions (eg, dehydration) that may have prevented concentration of urine in kidneys of those dogs; attributing azotemia in PHPT-affected dogs with low urine specific gravities to renal causes would have been inaccurate.
Results of the present study indicated serum PTH concentrations higher than the reference range were significantly associated with low postoperative serum tCa nadir concentrations in dogs. No significant correlation between preoperative plasma PTH and calcium concentrations was found in another study13 that included PHPT-affected dogs; relationships between preoperative plasma PTH concentrations and postoperative plasma calcium concentrations were not determined in that study. In the present study, we analyzed serum PTH concentrations as a binary variable (ie, concentrations within vs greater than the reference range). Results of other studies14,15 indicate the intermethod variation for measurement of circulating PTH concentrations in humans is −44.9% to 123.0%, which suggests accurate measurement of circulating PTH concentrations is challenging. Various assays for measurement of PTH differ in methodology, including the part of the PTH molecule that is detected. Additionally, PTH degrades during storage and is variably stable in plasma or serum samples.15 Because serum PTH concentrations reported in the present study were determined via various types of assays, values could not be compared. Regardless, results of this study suggested there were associations between renal function and circulating PTH and calcium concentrations in dogs. These variables should be evaluated in future prospective studies conducted to identify factors associated with risk of postoperative complications in PHPT-affected dogs.
The multiple linear regression model determined in this study accounted for 51% of the variability in postoperative serum tCa nadir concentrations; 3 variables were identified that had significant effects on postoperative serum tCa nadir concentrations. The serum CPP had the greatest effect and accounted for 42% of the variability in postoperative serum tCa nadir concentrations. The multiple linear regression model accounted for 31% of the variability in postoperative serum iCa nadir concentrations; 2 variables were identified that had significant effects on postoperative serum iCa nadir concentrations. Body weight had the greatest effect and accounted for 23% of variability in postoperative serum iCa nadir concentrations.
Decreases in circulating phosphorus concentrations in dogs with hypercalcemia secondary to PHPT are typically attributable to PTH-induced inhibition of renal tubular phosphorus resorption.16 Results of the present study indicated low serum CPP values were associated with low postoperative serum tCa nadir concentrations. Interestingly, preoperative serum phosphorus, calcium, and creatinine concentrations alone were not significantly associated with postoperative serum calcium concentrations. Regardless, results of this study suggested all systemic abnormalities of dogs resulting from preoperative hypercalcemia should be assessed for determination of the likelihood of development of postoperative hypocalcemia, rather than assessing circulating tCa or iCa concentrations alone.
Hypocalcemia was commonly detected during the postoperative period in dogs in this study, although development of associated clinical signs was uncommon. Approximately one-third of dogs in the study became hypocalcemic following parathyroidectomy, and approximately one-quarter of those dogs developed associated clinical signs. Only 6 of the 62 (9.7%) dogs in this study developed clinical signs of hypocalcemia after parathyroidectomy. These results were similar to results of another study17 in which 47 dogs with PHPT underwent parathyroidectomy; 18 (38.3%) dogs in that study became hypocalcemic and 5 (10.6%) had clinical signs of hypocalcemia after surgery. These findings suggested that although hypocalcemia is a common complication in dogs after surgical parathyroidectomy, adverse clinical signs typically do not develop.
The age and sex distributions of dogs in the present study were consistent with results of other studies1,2,10 including dogs with PHPT; such dogs were typically old, and distributions of dogs among sexes were similar. A significant negative correlation was detected between age of dog and postoperative serum tCa nadir concentration in the present study (old dogs had lower serum tCa nadir concentrations vs young dogs). Although age was a significant variable in the multiple linear regression analysis, the contribution of age to variability in postoperative serum tCa nadir concentrations was small. Old dogs may have been more likely than young dogs to have comorbidities (eg, chronic kidney disease) and may therefore have been unable to reestablish calcium homeostasis. Additionally, neuter status (sexually intact vs neutered; but not sex) and body weight had significant effects on postoperative serum iCa nadir concentrations. The reason for this finding could not be determined, and only 2 dogs in the study were sexually intact; therefore, these results should be interpreted cautiously.
Results of other studies1,2,10 indicate Keeshonden are predisposed to development of PHPT, and results of another study18 indicate PHPT has an autosomal dominant mode of inheritance in Keeshonden with possible age-dependent penetrance.18 Although Keeshonden were included in the present study, dogs of this breed did not seem to be overrepresented; this difference between results of the present study and those of other studies may have been attributable to variations in regional populations of dogs or other factors such as the small number of dogs in this study.
Most (79.0%) of the dogs in this study had a histologic diagnosis of parathyroid gland adenoma. Parathyroid gland hyperplasia was less commonly identified (16.1% of dogs); 2 dogs had lesions that were either parathyroid gland hyperplasia or adenoma, and 1 dog had histologically normal parathyroid gland tissue. These findings suggested accurate histologic classification of surgically excised parathyroid gland tissue was challenging. The parathyroid gland immunohistochemical labeling patterns for PTH have been determined for a small number of clinically normal dogs and dogs with PHPT by use of an avidin-biotin-peroxidase complex technique with monoclonal rat anti-human PTH antibodies.19–22 Evaluation of parathyroid glands of a large number of dogs via these techniques may be warranted. Furthermore, we chose to exclude dogs with parathyroid gland adenocarcinomas from the study because, at the time of writing, no information had been published regarding the clinical findings for dogs with such tumors. Therefore, to prevent inclusion of confounding variables in analyses, such dogs were excluded.
Most (93.5%) dogs in this study underwent high-resolution cervical ultrasonography of the ventral aspect of the cervical region before surgery to determine locations of ultrasonographically abnormal parathyroid gland tissue. Ultrasonography was performed by board-certified veterinary radiologists with various amounts of experience at 4 institutions. Therefore, only agreement regarding laterality of parathyroid gland lesions was assessed. Agreement between ultrasonographic results and surgical findings regarding laterality of parathyroid gland lesions was identified for only 75.9% of dogs; this finding supported findings of other studies4,10,12,22 that suggest interpretation of ultrasonographic findings is subjective and operator-dependent. Regardless, surgical exploration of parathyroid glands is performed via a ventral cervical approach, and each parathyroid gland is evaluated during the procedure; therefore, even if parathyroid gland abnormalities are incorrectly identified during ultrasonography before surgery, parathyroid gland lesions should be correctly identified during surgery.
The retrospective, multicenter design of the present study allowed inclusion of more dogs than would have been included in a prospective study conducted at 1 institution. However, several limitations were attributable to this study design. For example, the high variability in serum iCa assay results may have been attributable to variation among personnel at institutions regarding collection, handling, and processing of serum samples for those assays. Serum samples for measurement of iCa must be collected in anaerobic conditions and assayed within a short time after collection to prevent changes in pH caused by loss of CO2 from samples.23 Although personnel at the 4 hospitals that contributed cases to this study used standardized serum sample collection, handling, and analytic methods, inconsistencies in techniques could have resulted in variability in data. Additionally, postoperative treatments and care (eg, type of fluids administered IV or time at which food was offered after surgery) of dogs may have altered the shape of the calcium concentration-time curve. A prospective study in which consistent treatments and postoperative monitoring of dogs is used may be required to more accurately identify preoperative variables associated with development of hypocalcemia after surgery.
The preoperative variables associated with low postoperative serum calcium nadir concentrations in the 62 dogs with PHPT in the present study included a history of weakness, a serum PTH concentration higher than the reference range, a low serum CPP value, and a high serum BUN concentration. Preoperative serum tCa concentrations were not significantly associated with postoperative development of hypocalcemia in these dogs. Ultrasonography of the ventral aspect of the cervical region was moderately accurate for determination of laterality of parathyroid gland lesions in dogs; limitations of ultrasonography were attributed to subjective, operator-dependent performance and interpretation of this imaging technique.
ABBREVIATIONS
| CPP | Calcium-phosphorus concentration product |
| iCa | Ionized calcium |
| PHPT | Primary hyperparathyroidism |
| PTH | Parathyroid hormone |
| tCa | Total calcium |
SAS, version 9.2, SAS Institute Inc, Cary, NC.
GraphPad Prism, version 5.03 for Windows, GraphPad Software, San Diego, Calif.
References
1. Berger B, Feldman EC. Primary hyperparathyroidism in dogs: 21 cases (1976–1986). J Am Vet Med Assoc 1987; 191:350–356.
2. Feldman EC, Nelson RW. Hypercalcemia and primary hyperparathyroidism. In: Feldman EC, Nelson RW, eds. Canine and feline endocrinology and reproduction. 3rd ed. St Louis: Elsevier Science, 2004;660–715.
3. Patnaik AK, MacEwen EG, Erlandson RA, et al. Mediastinal parathyroid adenocarcinoma in a dog. Vet Pathol 1978; 15:55–63.
4. Wisner ER, Nyland TG, Feldman EC, et al. Ultrasonographic evaluation of the parathyroid glands in hypercalcemic dogs. Vet Rad Ultrasound 1993; 34:108–111.
5. DeVries SE, Feldman EC, Nelson RW, et al. Primary parathyroid gland hyperplasia in dogs: six cases (1982–1991). J Am Vet Med Assoc 1993; 202:1132–1136.
6. Krook L. Spontaneous hyperparathyroidism in the dog; a pathological-anatomical study. Acta Pathol Microbiol Scand Suppl 1957;(122):1–88.
7. Pearson PT, Dellmann HD, Berrier HH, et al. Primary hyperparathyroidism in a Beagle. J Am Vet Med Assoc 1965; 147:1201–1206.
8. Peterson ME, Randolph JF, Zaki FA, et al. Multiple endocrine neoplasia in a dog. J Am Vet Med Assoc 1982; 180:1476–1478.
9. Sawyer ES, Northrup NC, Schmiedt CW, et al. Outcome of 19 dogs with parathyroid carcinoma after surgical excision. Vet Comp Oncol 2012; 10:57–64.
10. Feldman EC, Hoar B, Pollard R, et al. Pretreatment clinical and laboratory findings in dogs with primary hyperparathyroidism: 210 cases (1987–2004). J Am Vet Med Assoc 2005; 227:756–761.
11. Flanders JA. Parathyroid gland. In: Slatter D, ed. Textbook of small animal surgery. 3rd ed. Philadelphia: WB Saunders Co, 2003;1711–1723.
12. Wisner ER, Penninck D, Biller DS, et al. High-resolution parathyroid sonography. Vet Radiol Ultrasound 1997; 38:462–466.
13. Gear RN, Neiger R, Skelly BJ, et al. Primary hyperparathyroidism in 29 dogs: diagnosis, treatment, outcome and associated renal failure. J Small Anim Pract 2005; 46:10–16.
14. Souberbielle JC, Boutten A, Carlier MC, et al. Inter-method variability in PTH measurement: implication for the care of CKD patients. Kidney Int 2006; 70:345–350.
15. Souberbielle JC, Roth H, Fouque DP. Parathyroid hormone measurement in CKD. Kidney Int 2010; 77:93–100.
16. Feldman EC. Disorders of the parathyroid glands. In: Ettinger SJ, ed. Textbook of veterinary internal medicine. 5th ed. Philadelphia: WB Saunders Co, 2000;1379–1399.
17. Rasor L, Pollard R, Feldman EC. Retrospective evaluation of three treatment methods for primary hyperparathyroidism in dogs. J Am Anim Hosp Assoc 2007; 43:70–77.
18. Goldstein RE, Atwater DZ, Cazolli DM, et al. Inheritance, mode of inheritance, and candidate genes for primary hyperparathyroidism in Keeshonden. J Vet Intern Med 2007; 21:199–203.
19. Inoue Y, Setoguti T. Immunocytochemical localization of parathormone in the mammalian parathyroid gland using the protein A-gold technique. Cell Tissue Res 1986; 243:3–7.
20. Grone A, Werkmeister JR, Steinmeyer CL, et al. Parathyroid hormone-related protein in normal and neoplastic canine tissues: immunohistochemical localization and biochemical extraction. Vet Pathol 1994; 31:308–315.
21. van Vonderen IK, Kooistra HS, Peeters ME, et al. Parathyroid hormone immunohistochemistry in dogs with primary and secondary hyperparathyroidism: the question of adenoma and primary hyperplasia. J Comp Pathol 2003; 129:61–69.
22. Feldman EC, Wisner ER, Nelson RW, et al. Comparison of results of hormonal analysis of samples obtained from selected venous sites versus cervical ultrasonography for localizing parathyroid masses in dogs. J Am Vet Med Assoc 1997; 211:54–56.
23. Schenck PA, Chew DJ Calcium: total or ionized? Vet Clin North Am Small Anim Pract 2008; 38:497–502.
