Serum gastrin concentrations in dogs with primary hyperparathyroidism

2.1 Animals and case definitions

Ours was a 2-part prospective cohort study. The study protocol was approved by the Institutional Animal Care and Use Committee at Michigan State University (MSU). Informed consent was obtained. In phase 1, serum gastrin concentrations were measured in residual serum from dogs with PHPT. Cases were identified by reviewing “malignancy panel” sample submissions to the MSU Veterinary Diagnostic Laboratory. A diagnosis of PHPT was based on ionized hypercalcemia (>1.55 mmol/L), with a mid to upper-normal range or high parathyroid (PTH) hormone concentration, a PTH-related protein (PTH_rp) concentration of 0 mmol/L, and a serum phosphorous concentration within or below the reference interval (RI).^{5, 15} Because cases originated from many veterinary practices and utilized several different laboratories or analyzers, the stated RI of the laboratory that measured the serum phosphorus concentration was utilized when evaluating cases. If a surplus of >0.5 mL of serum was available, the primary veterinarian who submitted the sample was contacted to obtain permission for medical record review and potential case enrollment. Dogs then were excluded if they were azotemic, had concurrent disease, or had received acid suppressant medications (eg, famotidine, omeprazole) within 7 days of sample collection.¹⁶ Dogs also were excluded if medical records were unavailable for review, informed consent was not granted, or if a 12-hour fast before sample acquisition had not been performed. Dogs with PHPT were categorized as having clinical signs related to GI tract disease if ≥1 of vomiting, diarrhea, change in appetite, weight loss, constipation, or abdominal pain were noted in the medical record within 30 days of study enrollment.

Dogs that underwent corrective management of their PHPT (eg, ethanol ablation or surgical parathyroidectomy), were eligible for phase 2 of the study. Treatment decisions were made by the attending clinician and were not influenced or dictated by the study investigators. In phase 2, dogs had repeat quantification of serum gastrin, PTH, and iCa concentrations, 28 (±2) days after surgical parathyroidectomy or ultrasound-guided ethanol ablation. These samples were collected by the primary care veterinarian. Serum samples then were packaged with ice packs in insulated containers and shipped overnight to MSU. Once received, samples were immediately placed in a −20°C freezer until batch analysis of serum gastrin, PTH, and iCa concentrations.

2.2 Laboratory methods

2.3 Statistical analysis

A priori sample size calculations indicated that 139 dogs were needed to determine a 95% confidence interval (CI) with a precision of 5%, assuming a prevalence of 10% and an infinite population. Similarly, it was determined that 24 dogs were needed for phase 2 based on a 2-tailed paired t test assuming an alpha of 0.05, a power of 0.9, and an effect size of 20 (mean difference) and SD of 30. The prevalence, effect size, and SD used in the sample size calculation were conservative estimates based on review of pre-existing data in humans.^20-22 Data sets were assessed for normality using Shapiro–Wilk testing and inspection of normal probability plots. Normally distributed data were reported as means ± SD, whereas non-normally distributed data were reported as medians and interquartile ranges (IQRs). Exceptions included patient age and weight, in which range was used as the measure of spread. The effect of treatment on serum gastrin, iCa, and PTH concentrations was assessed using Wilcoxon signed rank sum tests. Spearman's rank correlation coefficients were calculated to investigate potential relationships between serum gastrin and PTH and iCa concentrations. For Spearman testing, a significant correlation score of ±0.3 to 0.5 was considered a weak correlation, ±0.5 to 0.7 was considered a moderate correlation, and ±0.7 to 1.0 was considered a strong correlation.²³ Fisher exact testing was used to compare the proportion of dogs with hypergastrinemia in dogs with and without signs related to the GI tract. Statistical analyses were performed using commercially available software (GraphPad Prism Version 9.0, GraphPad Software Inc, San Diego, California). Significance for all statistical comparisons was set at P ≤ .05.

3.1 Animals

Medical records were sought from 273 client-owned dogs. One hundred and fifty-one (55%) of those dogs then were entered into phase 1 of the study. Reasons for non-enrollment are documented in Figure 1. The median age of enrolled dogs was 12 years (range' 4-17 years). Of the enrolled dogs, 89 were male (87 neutered, 2 intact), 62 were female (61 spayed, 1 intact). The median weight of enrolled dogs was 16.8 kg (range, 2.7-48.0 kg). Breeds enrolled consisted of mixed breed (n = 55; 38%), Shih Tzu (n = 14; 9.2%), Dachshund (n = 13; 8.6%), and Siberian Husky (n = 10; 6.6%). Fewer than 5% of each of the following breeds were represented: Labrador Retriever (n = 5), Pug (n = 6), Beagle (n = 5), Chihuahua (n = 4), Pitbull (n = 4), Boxer (n = 3), Lhasa Apso (n = 2), Weimaraner (n = 1), Maltese (n = 2), Miniature Poodle (n = 1), Miniature Schnauzer (n = 1), Great Pyrenees (n = 1), American Bulldog (n = 1), Yorkshire Terrier (n = 1), Jack Russell Terrier (n = 1), West Highland Terrier (n = 1), Cairn Terrier (n = 1), Bearded Collie (n = 1), Bernese Mountain Dog (n = 1), Brittany Spaniel (n = 1), German Wire-haired Pointer (n = 1), Whippet (n = 1), Bichon (n = 1), English Bulldog (n = 1), French Bulldog (n = 1), Australian Shepherd (n = 1), Miniature Doberman Pinscher (n = 1), Pomeranian (n = 1), Rat Terrier (n = 1), Water Spaniel (n = 1), Cockapoo (n = 1), Wire-haired Fox Terrier (n = 1), and Border Collie (n = 1).

Thirty-three (33) dogs were screened for phase 2 of the study and 24 dogs were enrolled (20 dogs that underwent surgical parathyroidectomy, and 4 dogs that underwent ethanol ablation). Reasons for non-enrollment for the 9 dogs are presented in Figure 2.

3.2 Correlation between gastrin, iCa, and PTH concentrations

Serum gastrin concentrations were not significantly correlated with iCa (P = .92; Figure 3) or PTH concentrations (P = .60; Figure 4).

3.3 Pre- and post-treatment PTH and iCa concentrations

The median pre-treatment PTH concentration was 14.3 pmol/L (IQR' 14.4 pmol/L; RI, 1.1-10.6 pmol/L). The median pre-treatment iCa concentration was 1.68 mmol/L (IQR, 0.14 mmol/L; RI, 1.25-1.45 mmol/L). The median post-treatment PTH concentration was 6.45 pmol/L (IQR, 6.40 pmol/L). The mean post-treatment iCa concentration was 1.36 ± 0.10 mmol/L. Treatment of PHPT significantly decreased serum PTH (P < .001) and iCa (P < .001) concentrations (Figure 5). All but 4 dogs had PTH concentrations return to within the RI after treatment. All but 1 dog had resolution of ionized hypercalcemia after treatment (pretreatment: PTH, 26.9 pmol/L; iCa, 1.67 mmol/L; gastrin, 64 ng/L; post-treatment: PTH, 16.3 pmol/L; iCa, 1.65 mmol/L; gastrin, 68 ng/L). The dog that did not have resolution of ionized hypercalcemia had undergone a parathyroidectomy.

3.4 Pre- and post-treatment gastrin concentrations

Twenty-seven dogs (27/151, 17.9%) had increased pre-treatment serum gastrin concentrations (RI ≤ 78.9 ng/L). The median pre-treatment serum gastrin concentration was 45.0 ng/L (IQR, 20.0 ng/L). Three dogs (3/24, 12.5%) had an increased post-treatment gastrin concentration. The mean post-treatment gastrin concentration was 54.0 ± 27.5 ng/L. In the 24 dogs with paired pre-treatment and post-treatment gastrin concentrations no difference was found in serum gastrin concentrations before and after treatment (P = .15; Figure 5).

3.5 Gastrointestinal signs

Thirty-nine dogs had clinical signs associated with GI tract disease within 30 days of diagnosis of PHPT. The proportion of dogs with hypergastrinemia with and without GI clinical signs was not different (P = 1.00). The specific clinical signs included changes in appetite (16 dogs), diarrhea (13 dogs), weight loss (7 dogs), vomiting (6 dogs), abdominal pain (3 dogs), and constipation (2 dogs). Dogs may have had >1 clinical sign of GI tract disease concurrently. The proportion of dogs with hypergastrinemia with and without specific clinical signs was not different (P = .46 to 1.00).