Gastric pH and serum gastrin concentration in age-matched healthy dogs and dogs with chronic kidney disease

2.1 Study animals

The Institutional Animal Care and Use Committee (IACUC) at Texas A&M University, North Carolina State University, and University of Tennessee approved the protocols for this study. To detect a hypothesized paired mean difference in mean pH of 0.4 between dogs with CKD and healthy dogs, assuming a SD of 0.4096 for each group, a correlation of 0.5, at an alpha of 0.05, 11 pairs of dogs were required per group to achieve 80% power. Client-owned healthy adult dogs and dogs with stable CKD were prospectively enrolled in the study from September 2019 to September 2021 from 3 veterinary university hospitals (Texas A&M University, North Carolina State University, and University of Tennessee). One affected male laboratory dog with an X-linked hereditary nephropathy was also enrolled at Texas A&M University. All owners completed a consent form before study enrollment. Inclusion criteria for dogs with CKD included a body weight >3.0 kg, Stage I to IV CKD as defined by the International Renal Interest Society (IRIS), and USG <1.015. Inclusion criteria for healthy control dogs included body weight >3.0 kg, lack of azotemia (International Renal Interest Society. IRIS Staging of CKD; 2023. http://www.iriskidney.com/pdf/2_IRIS_Staging_of_CKD_2023.pdf; ie serum creatinine <1.4), USG >1.015, a serum symmetric dimethylarginine (SDMA) within reference range (<14 μg/dL), a body condition ≥4 (out of 9; WSAVA Global Nutrition Committee. Body condition score; 2013. https://wsava.org/wp-content/uploads/2020/01/Body-Condition-Score-Dog.pdf), and no recent (<2 months) history of medication administration other than routine heartworm and flea prophylaxis. The requirement for healthy dogs having an SDMA <14 μg/dL was added following case screening and initial inclusion. Exclusion criteria in either group included the presence of other systemic diseases (eg, diabetes mellitus, hepatitis, history, or evidence of chronic primary GI disease) or the use of any other drug that might alter intragastric pH within the past 7 days before or during intragastric pH monitoring (eg, NSAIDS, steroids, antibiotics).

Given that serum gastrin concentration normalizes after a 7-day discontinuation of omeprazole and famotidine,²⁰ acid suppressants were discontinued for a minimum of 8 days before and during pH monitoring. Oral administration of aluminum hydroxide was stopped for 48 hours before and during pH monitoring due to their neutralizing effect on gastric acid. This is well beyond the likely duration of action of aluminum hydroxide, which has a duration in humans of approximately 26 minutes.²¹ As the majority of dogs with CKD were eating a renal therapeutic diet, healthy dogs were also fed a therapeutic renal diet (Royal Canin Renal Support A, Royal Canin USA Inc., St. Charles, Missouri) 3 to 5 days before the study and during pH monitoring to minimize the effect of diet as a confounding variable. All dogs had a complete blood count, serum biochemistry, urinalysis, urine culture, serum gastrin, and blood pressure measurement within 72 hours of pH monitoring. Muscle condition score was recorded for all animals (WSAVA Global Nutrition Committee. Muscle Condition Score. 2013; https://wsava.org/wp306content/uploads/2020/01/Muscle-Condition-Score-Chart-for-Dogs.pdf). Dog's IRIS stage was determined by serum creatinine. All blood tests, with the exception of serum gastrin, were performed by the respective institutional clinical pathology service. For measurement of serum gastrin concentrations, serum was collected from blood tubes after centrifugation at 250g and stored in cryovials at −80°C. Serum was shipped on dry ice to the Gastrointestinal Laboratory at Texas A&M University. Serum gastrin concentrations were measured by an automated chemiluminescence, enzyme-labeled immunometric assay (Immulite 2000, Siemens Healthcare Diagnostics, Malvern, Pennsylvania), as previously described.²²

2.2 pH capsules

All dogs were lightly sedated with butorphanol (0.3 mg/kg IV; Torbugesic 10 mg/mL injection; Fort Dodge Animal Health, Fort Dodge, Iowa) and dexmedetomidine (1 mcg/kg IV [CKD] or 2 mcg/kg IV [healthy]; Dexdomitor 0.5 mg/mL injection; Orion Pharma, Espoo, Finland). A Bravo pH calibration free capsule (Medtronic, Minneapolis, Minnesota) was passed orally using a delivery device attached to the capsule. Once positioning of the capsule in the stomach was confirmed radiographically, mucosal attachment of the pH capsule was achieved as previously reported.²³ Before use, all pH capsules and receivers were linked as previously described according to the manufacturer's instructions.²⁴ After the procedure, the sedation was reversed with an equal volume of atipamezole (0.05-0.1 mg/kg IM; Antisedan 5 mg/mL injection; Orion Pharma, Espoo, Finland). Dogs were allowed to eat once fully recovered and in the care of the clients. Neither the amount of food fed nor the timing of feeding was standardized. Clients of dogs with CKD were instructed to feed as they previously had, and clients of healthy dogs were given specific amounts to feed of the renal diet that was roughly equivalent to what they have been feeding with the dog's maintenance diet. Gastric pH recordings were obtained telemetrically at 6-second sampling intervals for 48 hours after pH capsule administration. Receivers were maintained in a vest worn by the dog during the 48-hour monitoring period. The pH data were uploaded to the computer by proprietary software (Reflux Software v6.1, Medtronic) after the monitoring period. Early capsule detachment from the stomach was defined by a rapid and persistent rise in pH >4. Only data from the first 48 hours of recording in the stomach were included in the analysis. Mean pH and mean percentage of time (MPT) that gastric pH was 0 to 1 and 1 to 2 were calculated using data from the proprietary software supplied by the device manufacturer.

2.3 Statistical analysis

A longitudinal matched case-control study design and corresponding mixed model analysis of variance (ANOVA) was performed to evaluate mean intragastric pH and MPT that intragastric pH was ≤1 and ≤2. Each outcome was evaluated for status (CKD/healthy), time (day 1/day 2), and status-by-time differences. Dogs were individually paired based on the closest similar age of dogs with CKD and healthy dogs. Each matched pair, and status nested within pair were considered random effects.²⁵ A Shapiro-Wilk test and QQ plots were used to evaluate normality of ANOVA residuals for each outcome. Levene's equality of variances test was used to evaluate equality of treatment variances. Box-and-whisker plots and studentized residual diagnostics were performed to evaluate each mixed model for the presence of outliers. A log transformation of MPT ≤1 was required to meet underlying statistical assumptions. After applying a log transformation for MPT ≤1, all statistical assumptions were met.

Both 48-hour mean gastric pH and serum creatinine measures were evaluated for normality using the Shapiro-Wilk Test. Since healthy dog creatinine was not normally distributed, Spearman's correlations analysis was performed to evaluate whether a significant relationship was present between pH and creatinine for both CKD and healthy dogs.

Serum gastrin values below the detectable limit were conservatively assumed to be at the limit of detection of 10 ng/dL. Serum gastrin concentration differences were evaluated between age-matched CKD and healthy dogs using a Shapiro-Wilk test and determined to not be normally distributed. Subsequently, a Wilcoxon Signed Rank Test was performed to evaluate serum gastrin for mean rank status differences. Statistical significance was defined as P < .05. Statistical analysis was performed using commercial software (SAS software, version 9.4, Cary, North Carolina, Release TS1M7 and IBM SPSS Statistics software, Version 29, Armonk, New York).

3.1 Dogs

Thirteen dogs with CKD and 11 healthy dogs were enrolled. One healthy dog was determined to have an SDMA (18 μg/dL) outside of the reference range following inclusion and was subsequently excluded. One dog with CKD underwent pH monitoring and passed the capsule after 24 hours; a second capsule was not administered and only the first 24 hours were included in the analysis. Demographic information is summarized in Table 1, including number of dogs per group, breeds, age, sex, weight, body condition score, and muscle condition score. There was no significant difference in age between groups, and 2 dogs with CKD did not have a muscle condition score recorded.

In total, 3 dogs with IRIS stage I, 3 dogs with IRIS stage II, 2 dogs with IRIS stage III, and 5 dogs with IRIS stage IV CKD were enrolled. All healthy dogs had hematocrits within the reference range. Seven out of 13 dogs with CKD were anemic (median, range hematocrit: 35, 33-50; reference interval: 40-61). Out of these 7 dogs, 2 were microcytic and 2 were macrocytic.

Dogs with CKD were receiving the following medications or treatments for CKD: anti-proteinuric or anti-hypertensive medications (telmisartan, n = 2 or enalapril, n = 1 or amlodipine, n = 1), as needed maropitant (n = 2), as needed mirtazapine (n = 1), calcitriol (n = 2), daily subcutaneous fluids (n = 1), and phosphate binders (n = 3; all aluminum hydroxide). Three dogs were receiving acid suppressants (omeprazole n = 2, famotidine n = 1) before discontinuation for study enrollment. Eleven dogs with CKD were eating a therapeutic renal diet at the time of pH monitoring (Hill's k/d, Hill's Pet Nutrition, Inc., Topeka, KS [n = 6]; Royal Canin Renal, Royal Canin USA Inc., St. Charles, Missouri [n = 4]; Royal Canin Feline Renal, Royal Canin USA Inc., St. Charles, Missouri [n = 1]; Blue Buffalo Kidney Support, Blue Buffalo Co., Wilton, Connecticut [n = 1]; home-prepared diet [n = 1]; laboratory chow diet [n = 1]). All healthy dogs accepted and were fed the therapeutic renal diet before and during the study period.

3.2 pH monitoring and serum gastrin

Mean gastric pH and the percentage of time that gastric pH was strongly acidic (pH <1 and <2) during day 1 (first 24-hour period) and day 2 (second 24-hour period) were used for comparative analyses and are presented in Table 2. No significant difference was observed in status (CKD/healthy), time (day 1/day 2), or the status by day interaction for mean pH, or mean percentage time that gastric pH was strongly acidic (P > .05, for all). Spearman's correlation for serum creatinine and gastric pH for both the CKD (ρ = −0.25, P = .43) and healthy (ρ = −0.39, P = .26) dogs were not significant (Figures 1 and 2).

All dogs with CKD (median [range], 10.5 ng/dL [<10-17.1]) and healthy dogs (median [range], 10.9 ng/dL [<10-15]) had serum gastrin concentrations within the reference range. Two dogs with CKD did not have gastrin data available. Serum gastrin concentrations were not significantly different between CKD and healthy dogs (P > .05).