Effects of Long-Term Carprofen and Omeprazole Administration in C57BL/6J Mice

Abstract

Osteoarthritis is the leading cause of disability in the United States and affects approximately half of adults over the age of 65. Many osteoarthritis patients take nonsteroidal anti-inflammatory drugs (NSAIDs) on a long-term basis, often concurrently with proton pump inhibitors (PPIs), such as omeprazole, to prevent gastric ulceration. Mice (Mus musculus) are a commonly used animal model of osteoarthritis. There are little data regarding long-term administration of NSAIDs or coadministration of PPIs and NSAIDs in mice. This study sought to determine if administration of carprofen, a commonly used veterinary NSAID, has adverse effects when administered for 21 days and if coadministration of omeprazole reduces the incidence of adverse effects. Four groups of C57BL/6J male (n = 5/group) and female (n = 5/group) mice were weighed daily and administered 10 mg/kg carprofen and 8.2 mg/kg omeprazole, 10 mg/kg carprofen, 8.2 mg/kg omeprazole, or control suspension once daily by oral gavage for 21 days. All mice were euthanized, and complete blood count (CBC), serum chemistry, fecal occult blood, and pyloric histopathology and gastritis scoring were conducted. All animals remained clinically healthy for the duration of the study. White blood cell counts (WBCs) and platelets were significantly lower in the carprofen and omeprazole group. Neutrophil counts were significantly lower in the carprofen and omeprazole and the carprofen groups. Compared with the control group, albumin was significantly higher in the carprofen group. Fecal occult blood tests were negative for all animals. No animals had pyloric mucosal ulceration, and gastritis scores were not significantly different between groups. Body weight significantly decreased for all groups over time, with no significant differences among treatment groups. Carprofen and omeprazole may be safely administered to C57BL/6J mice for 21 days but may induce significant changes in CBC and serum chemistry.

Introduction

Osteoarthritis (OA) is a chronic, progressive musculoskeletal disease resulting in joint dysfunction and ultimately leading to pain, mobility loss, and functional limitations.¹ OA is the leading cause of disability in the United States and affects approximately half of adults over the age of 65.² There are not currently any available disease-modifying treatments for OA. Until these treatments are developed, OA is primarily managed by treatment of symptoms, commonly with NSAIDs, until surgical joint replacement is deemed necessary.^3,4 Mice are a commonly used animal model of osteoarthritis.^5–7 OA in mice can be induced surgically, chemically through intraarticular injections, or via noninvasive trauma.⁷ For further utility as a translational research model, chronic NSAID use in mice should be investigated to model medical management of OA in humans.

Chronic NSAID use is associated with a variety of adverse effects in humans and in veterinary species. NSAIDs inhibit cyclooxygenase enzymes, thereby inhibiting prostaglandin production. Prostaglandin is needed for normal protection of the gastric mucosa, and inhibition of prostaglandin synthesis has been associated with toxic side effects of NSAIDs such as gastritis and pyloric ulceration.^8,9 Prostaglandins protect the gastric mucosa through several mechanisms including reduction of gastric acid secretion, maintaining gastric mucosal blood flow, and increasing mucosal bicarbonate formation.¹⁰ Chronic NSAID use has been shown to result in pyloric ulceration and gastritis in humans and in other species such as dogs.^9,11,12 Proton pump inhibitors (PPIs) are commonly used concurrently with NSAIDs to mitigate these effects on the stomach.^13,14 PPIs may provide gastroprotective effects in the face of chronic NSAID use by reducing gastric acid secretion.¹⁰ High intragastric acidity resulting from decreased prostaglandin synthesis may lead to mucosal injury and ulceration.¹⁵

In addition to its effects on the stomach, chronic NSAID use can have detrimental effects on the kidneys and liver. Prostaglandins play a role in maintaining renal blood flow and glomerular filtration rate; thus, prostaglandin inhibition may lead to renal dysfunction. Increases in creatinine and BUN are observed in patients with NSAID-induced renal dysfunction.¹⁶ Hepatotoxicity is a rare adverse effect of NSAID use.¹⁷ NSAID-induced hepatotoxicity is generally considered to be an idiosyncratic reaction that occurs as a result of increased sensitivity of an individual patient, rather than direct cytotoxic effects of the drug. Affected patients may have elevated liver enzymes, including ALT and ALP.¹⁸

There are limited data on chronic carprofen administration in mice, necessitating a study to ensure the safety of long-term carprofen administration. One previous study¹⁹ investigated the administration of various NSAIDs to mice in feed for 6 months and saw no difference in survival between treatment groups or compared with the control group, but further toxicity evaluation was not performed. In another study, ibuprofen was administered to mice daily for 1 month with no reported adverse outcomes; however, no toxicity data were provided.²⁰ In the present study, we selected 21 days as the duration of medication administration. As there are limited data on the adverse effects of chronic NSAID use in mice, 21 days was considered a starting point to reveal potential adverse effects of chronic daily NSAID use in mice.

The aim of this study was to determine if mice would experience adverse outcomes after receiving 10 mg/kg carprofen and/or 8.2 mg/kg omeprazole daily for 21 days and therefore determine the feasibility of a murine model of chronic NSAID administration. No adverse effects were seen with the daily administration of 10 mg/kg carprofen, 8.2 mg/kg omeprazole, and carprofen with omeprazole during this study. This indicates that these dosing regimens are safe and therefore can be used in future studies requiring chronic NSAID administration in mice, such as OA research.

Ethical review.

All experimental procedures were approved by the Oklahoma Medical Research Foundation IACUC. Animals were housed in AAALAC International-accredited facilities in accordance with the Guide for the Care and Use of Laboratory Animals.²¹

Materials and Methods

Study design.

This study was performed to demonstrate the safety of long-term (21-day) administration of carprofen using male and female C57Bl/6J mice and to determine if concurrent omeprazole administration reduces potential adverse effects of carprofen administration in mice. The 4 experimental groups employed in the study were carprofen, omeprazole, carprofen + omeprazole, and control. Each individual mouse was considered the experimental unit. The carprofen dose was based on published literature that demonstrated no adverse effects in mice administered 10 mg/kg for short time periods.^22,23 The omeprazole dose was extrapolated from the dose (40 mg) used for treatment of NSAID-induced gastric ulcers in human medicine and converted to a mouse dose by calculating the animal equivalent dose based on body surface area.^24,25 All mice were euthanized, and all samples were collected after 21 days of dosing (Figure 1).

Figure 1.

Experimental Groups and Study Design.

Sample size.

Because this study investigated a novel dosing regimen, the effect size was unknown, thus a preexperiment power analysis could not be conducted. A recent paper evaluating the efficacy and toxicity of short-term carprofen administration used group sizes of 8-12 mice per group, with a mean of 10 per group.²² Therefore, 10 mice per group were used for the present study, resulting in a total sample size of 40 mice.

Inclusion and exclusion criteria.

Mice that were euthanized and did not complete the 21-day treatment course were excluded from analysis. CBC samples were excluded from analysis if blood became clotted in the sample tube or if there was insufficient sample volume. Chemistry samples were excluded from analysis if there was insufficient sample volume. Chemistry analytes were excluded from statistical analysis if the level of the analyte was below the threshold of the machine to determine an actual numerical value in more than one sample per group. Histology slides were excluded from gastritis scoring statistical analysis if the pylorus was not present on the slide (Table 1).

Table 1.

Number of Samples Analyzed per Treatment Group

	Female			Male
	CBC	Chemistry	Histopathology	CBC	Chemistry	Histopathology
Carprofen + omeprazole	4	4	4	3	4	5
Carprofen	5	5	4	5	5	4
Omeprazole	5	5	5	5	4	5
Control	5	5	5	4	5	4

Randomization.

Each sex and treatment group (5 mice/cage) was housed in separate cages to prevent confounding due to coprophagy between animals from different treatment groups. The cages were randomized to treatment groups using a random number generator (Random.org).

Blinding.

The pathologist was blinded to the treatment groups of samples for scoring of pylorus histopathology. The individual who interpreted fecal occult blood tests was blinded to the treatment group for each test.

Outcome measures.

Outcome measures for this study include CBC, serum chemistry, gastritis histopathologic grade (Table 2), fecal occult blood test, and body weight. Parameters examined on CBC include WBC, lymphocyte, monocyte, neutrophil, and platelet counts. Parameters measured on serum chemistry included albumin, glucose, total protein, globulin, creatinine, BUN, ALT, and ALP.

Table 2.

Gastritis Histopathologic Grading Scale

0	No significant lesions
1	Mild gastritis (mild neutrophilic infiltrates)
2	Mild gastritis (mild neutrophilic infiltrates and minimal mucosal erosion)
3	Moderate gastritis with minimal mucosal erosion
4	Marked gastritis with mucosal erosion to ulceration

Statistical methods.

The Shapiro-Wilk test of normality was conducted on albumin, glucose, total protein, and globulin concentrations and WBC, lymphocyte, monocyte, neutrophil, and platelet counts. The Grubbs test to detect an outlier was conducted on these values. Significant outliers (P < 0.05) were excluded from data analysis. If multiple outliers from the same animal were detected or if multiple values were below the detection limits of the machine, the entire sample was removed from analysis (Table 1). For the CBC data, if samples were clotted, they were excluded from analysis. For the chemistry data, if there was an insufficient sample resulting in machine errors, the sample was excluded. One-way ANOVA was conducted for normally distributed CBC and chemistry data. Kruskal-Wallis was used for nonnormally distributed CBC and chemistry data and for gastritis scores. Post hoc analysis was conducted using the Dunn multiple comparisons test. Males and females were analyzed together for CBC and chemistry data analysis. Weights by treatment group were analyzed using 2-way ANOVAs, with male and female weights analyzed separately. All statistical analyses were conducted using GraphPad Prism for Windows version 10.2.1 (GraphPad, San Diego, CA).

Experimental animals.

Twelve-week-old male and female C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME) were used for this study. Following arrival from the vendor, mice were allowed to acclimate for 72 hours before starting experimental procedures. The mice were housed under specific pathogen-free conditions in AAALAC-accredited facilities in autoclaved individually ventilated Optimice cages (Animal Care Systems, Centennial, CO) on autoclaved corncob bedding (The Andersons, Maumee, OH) and provided with chlorinated water (0.8-1.6 ppm) ad libitum. Mice also received ad libitum γ-irradiated feed (LabDiet RMH 3000; LabDiet, Richmond, IN). The mice were maintained at a room temperature of 72 °F and 30%-70% humidity on a 14:10 light:dark cycle. Mouse hepatitis virus, mouse parvovirus, minute virus of mice, Theiler murine encephalomyelitis virus, mouse rotavirus, Sendai virus, pneumonia virus of mice, reovirus 3, lymphocytic choriomeningitis virus, ectromelia virus, mouse adenovirus 1, mouse adenovirus 2, mouse polyomavirus, Mycoplasma pulmonis, pinworms (Syphacia spp. and Aspicularis spp.), and fur mites (Radfordia affinis, Myobia musculi, and Myocoptes musculinis) were excluded from the facility. Helicobacter spp. and mouse norovirus are permitted in the room where the experimental mice were housed; however, the mice used in this study were free from these agents per vendor production health reports.

Experimental procedures.

Mice were weighed daily and then orally gavaged with 10 mg/kg carprofen (n = 5 per sex), 8.2 mg/kg omeprazole (n = 5 per sex), combination of carprofen and omeprazole (n = 5 per sex), or 5.3 mL/kg control solution (n = 5 per sex) once daily for 21 days using 20-gauge, 30-mm flexible plastic gavage needles (Instech Laboratories, Philadelphia, PA). The concentrations of the medications were 1.9 mg/mL carprofen, 1.5 mg/mL omeprazole, and 1.9 mg/mL carprofen + 1.5 mg/mL omeprazole. The control solution was the flavored vehicle liquid provided by the vendor. All medications were compounded by Wedgewood Pharmacy (Swedesboro, NJ). New gavage needles were used daily, and needles were changed between cages of mice or within the same cage if the gavage needle was bitten by the mice. Solution concentrations were chosen so that mice of the same weight would receive the same volume of each solution (0.005 mL/g) across all treatment groups. All medications contained tutti-frutti marshmallow flavoring to increase palatability and encourage the mice to voluntarily consume the medications from a pipette tip. However, not all mice were willing to voluntarily ingest the medication, and the amount of medication consumed by each mouse was variable. Due to not all mice receiving the full dose of medication on day 1, oral gavage was initiated on day 2.

On day 23, mice were euthanized via isoflurane overdose followed by cardiac exsanguination to collect blood for CBC and serum chemistry. CBCs were performed with the Hemavet (Drew Scientific, Plantation, FL), and serum chemistries were performed with the IDEXX Catalyst One (IDEXX Laboratories, Westbrook, ME) using the Chem10 clip. Both tests were performed according to the manufacturer’s instructions. Automated dilutions were performed on chemistry samples according to the manufacturer’s instructions. Five fresh fecal pellets were taken from each mouse cage and used for fecal occult blood testing (Hemoccult; Beckman Coulter, Brea, CA). Briefly, feces were mixed with 1.5 mL sterile phosphate-buffered saline and applied to the test card. The test cards were allowed to develop for 3 days. After 3 days, 2 drops of Hemoccult Developer were added to each test window, and the results were read within 60 seconds. Direct feces collection from individual mice was initially attempted but was unsuccessful. Collection of pellets from the cage was chosen to avoid blood contamination from postmortem collection from the rectum. The distal esophagus, stomach, and proximal duodenum were collected from each mouse, cut longitudinally along the greater curvature of the stomach, and fixed in 10% neutral buffered formalin. The tissues were paraffin embedded, sectioned longitudinally to reveal the pyloric-duodenal junction, placed on glass slides, and stained with hematoxylin and eosin. The slides were evaluated by a board-certified veterinary pathologist who was blinded to treatment groups and scored according to Table 2. The grading scale was adapted from a publication evaluating the toxicity of NSAIDs in mice and was revised to better reflect the milder lesions seen in our cohort.²⁶

Results

One female in the carprofen + omeprazole group was euthanized on day 1 due to respiratory distress following oral gavage. All other animals remained clinically healthy for the duration of the study. Two male samples in the carprofen + omeprazole group were excluded from CBC analysis due to clotting. One sample in the male control group was excluded from CBC analysis due to insufficient sample volume. There was one significant outlier in the female omeprazole group (P < 0.05) excluded from the WBC statistical analysis. Creatinine, BUN, ALT, and ALP were below the threshold of the chemistry machine to provide a definitive numeric value in one or more samples in at least 2 groups. All creatinine values were less than 0.8 mg/dL. All BUN values were less than 20 mg/dL. All ALT values were less than 80 U/L. All ALP values were less than 149 U/L. Because these were not definitive numeric values, they could not be analyzed statistically. There was one significant outlier in the female carprofen + omeprazole group (P < 0.05) excluded from albumin statistical analysis. One sample in the male omeprazole group and one sample in the male carprofen + omeprazole group were excluded from chemistry data analysis due to insufficient sample volume. All animals were negative for fecal occult blood. Three samples were excluded from gastritis histopathology scoring due to slides not containing the pylorus (Table 1).

Compared with the control group, neutrophil counts were significantly lower in the carprofen and omeprazole group and in the carprofen group (carprofen + omeprazole: 0.37 ± 0.3; P = 0.04; carprofen: 0.41 ± 0.2; P = 0.03; control: 0.9 ± 0.7; median ± IQR) (Figure 2). WBC and platelet counts were significantly lower in the carprofen + omeprazole group compared with the control group (WBC: carprofen + omeprazole: 1.6 ± 0.6; control: 3.2 ± 1.7; P = 0.04; platelet: carprofen + omeprazole: 794.7 ± 265; control: 1,093 ± 278; P = 0.05; mean ± SD) (Figure 2). Compared with the control group, albumin was significantly higher in the carprofen group (carprofen: 2.5 ± 0.3; control: 2.3 ± 0.2; P = 0.005; median ± IQR) (Figure 2). All other comparisons had no significant differences between groups (P > 0.05 for all comparisons). No animals had pyloric mucosal ulceration, and gastritis scores were not significantly different between groups (P > 0.05) (Figures 3 and 4). There was no difference in weight gain between groups over time (P > 0.05). There was a statistically significant decrease in weight over time for all groups in both sexes (P < 0.05) (Figure 5). The source of weight variation originated from time as a factor (female: P = 0.05; male: P = 0.005), rather than from the treatment group (P = 0.48).

Figure 2.

Significant CBC and chemistry changes by treatment group. (A) Serum albumin by treatment group. (B) WBC count by treatment group. (C) Neutrophil count by treatment group. (D) Platelet count by treatment group. Summary data are represented as mean ± SD for platelets and WBCs. Data are represented as median ± IQR for albumin and neutrophils. c, carprofen; c+o, carprofen + omeprazole; o, omeprazole. *P ≤ 0.05 and ⁺P ≤ 0.005.

Figure 3.

Pylorus Histopathology and Gastritis Scoring (H&E stain). (A) Gastritis score 0: sagittal section of normal pyloric region with pyloroduodenal junction visible; 4× magnification. (B) Gastritis score 1: mild neutrophilic infiltrate in lamina propria (circle); 40× magnification. (C) Gastritis score 2: mild surface erosion (arrow); minimal neutrophils within lamina propria (circle); 40× magnification. (D) Gastritis score 3: mild surface erosion (arrow); small to moderate numbers of neutrophils in the lamina propria (circle) and submucosa (star); 20× magnification.

Figure 4.

Gastritis Scores by Treatment Group. Data represented as median ± IQR. c, carprofen; c+o, carprofen + omeprazole; o, omeprazole.

Figure 5.

Weight over time by sex and treatment group. (A) Mean male weight ± SD by treatment group. (B) Mean female weight ± SD by treatment group. Weights by treatment group were analyzed using 2-way ANOVA, with male and female weights analyzed separately.

Discussion

The results of this study indicate that 10 mg/kg carprofen, 8.2 mg/kg omeprazole, and carprofen + omeprazole may be safely administered to C57BL/6J mice for 21 days but may introduce significant changes on CBC and serum chemistry.

Mean absolute lymphocyte counts for all groups (range: 0.86-1.93 × 10³/μL) and subsequently mean total WBC counts (range: 1.5-3.2 × 10³/μL) were lower than published reference ranges for males and females for all groups (published range for lymphocytes 2.88-11.15 × 10³/μL, WBCs 3.9-13.9613 × 10³/μL). Mean absolute neutrophil, monocyte, and platelet counts and mean hematocrits were within published reference ranges for male and female C57BL/6 mice.²⁷ Mean glucose, total protein, and albumin concentrations for all groups were within the published reference ranges for mice.²⁸ There was no available reference interval for globulin in mice. The published reference ranges used for comparison were the same CBC and chemistry analyzer equipment that was used in the present study. Information about which strain of mice was used to determine the chemistry reference ranges was not available for the chemistry reference range data.²⁸

Neutrophil, WBC, and platelet counts were significantly lower in the carprofen + omeprazole group compared with the control group. In addition, neutrophil counts were significantly lower in the carprofen group compared with the control group. It is possible that coadministration of carprofen and omeprazole had a synergistic or cross-reactive effect resulting in decreased neutrophil, WBC, and platelet counts. Both NSAIDs and omeprazole have been reported to rarely cause neutropenia and thrombocytopenia in humans, but this has not been a published finding in mice.^29–32 The mean neutrophil and platelet counts for the carprofen + omeprazole group remained within published reference ranges for C57BL/6 mice, indicating that these decreases are unlikely to be clinically significant. Further, there were no clinical signs or sequelae of neutropenia or thrombocytopenia on routine observation or gross necropsy.

All groups had mean absolute lymphocyte counts and mean total WBC counts below published reference ranges for mice. Mice are lymphocyte redominant; therefore, the lymphopenia likely resulted in the observed leukopenia.³³ Chronic stress is a likely differential for lymphopenia in rodents.³⁴ A possible underlying etiology of the lymphopenia seen in all groups is stress secondary to daily restraint and oral gavage.³⁵ Oral gavage has been shown to significantly increase fecal corticosterone in mice, suggesting that this procedure causes stress in mice.³⁶ Other hematologic changes associated with chronic stress in mice include neutrophilia and eosinopenia.^34,37 Neutrophilia was not observed in any of the animals in this study. Because eosinophils are normally present in very low numbers in the blood (0.01%-0.50%), it is unclear if eosinopenia occurred in the experimental mice. In addition, all groups had decreased body weight over time. This finding was statistically significant but likely clinically irrelevant, as all mice maintained normal body condition scores throughout the study. Repeated oral gavage has been shown to result in decreased food intake and weight loss in some strains of mice, likely secondary to stress.³⁸ Weight loss in mice may be associated with stress.^39,40

In addition, albumin was significantly higher in the carprofen group compared with the control group. However, adverse effects of NSAIDs in veterinary species include gastrointestinal hemorrhage and acute kidney injury, which generally result in hypoalbuminemia.^41,42 The unexpected finding of decreased albumin in the carprofen group compared with the control group is unlikely to be clinically significant, as the mean albumin for all groups remained within the published reference range for C57BL6 mice.²⁷

The possible effects of oral gavage on body weight, lymphocyte counts, and total WBC counts should be considered as possible confounding factors when conducting studies in mice. Refinements in dosing, such as training the mice to take oral medications or providing the medication in water or food, should be used whenever possible as a refinement to animal welfare and to reduce possible confounding effects. Previous publications^43,44 have shown that carprofen can successfully be administered to mice in drinking water at 10 and 25 mg/kg. Providing the medications in water or food was not possible for this study because data regarding stability and tissue levels of omeprazole in feed or drinking water are lacking. It was necessary for the same method of administration to be used across all groups to prevent confounding effects. However, in studies where prolonged carprofen administration is needed, providing carprofen in drinking water represents a refinement to animal welfare and is likely more practical and efficient for laboratory staff. Further work should investigate the stability and pharmacokinetics of omeprazole when provided in food or water to mice.

Chronic NSAID use is commonly associated with pyloric ulcers and gastritis in human and veterinary patients.^9,11,12 Meloxicam, a commonly used NSAID in veterinary species, has been shown to cause pyloric ulcers and gastritis in mice when administered at 20 mg/kg subcutaneously for 3 and 7 days. When administered subcutaneously to mice at 20 mg/kg for 3 and 7 days, carprofen did not lead to gastric toxicity. No clinically significant changes were found on serum chemistries of mice treated with either carprofen or meloxicam with these dosing regimens.²⁶ However, significant gastric toxicity has been observed in mice administered 50 mg/kg carprofen subcutaneously for 3 days.^45,46 In the present study, carprofen administered at 10 mg/kg for 21 days did not result in pyloric ulceration, clinically relevant gastritis, or clinically relevant changes on serum chemistry. Therefore, carprofen can likely be safely administered to mice at this dosing regimen and duration.

While our study was rigorously controlled and conducted to maximize reproducibility, there were several potential weaknesses. Due to the death of one animal and 2 clotted CBC samples, the sample size for the carprofen + omeprazole group was 7, which was decreased compared with the original sample size of 10. The decreased sample size could theoretically result in a decreased ability to detect significant differences; however, multiple significant differences were found between the carprofen + omeprazole group and the control. Therefore, this sample size reduction does not appear to have affected the results in a meaningful way. Fecal samples for fecal occult blood testing were pooled at the cage level. This introduces the possibility of false negative results if there were positive animals whose feces were not included in the pooled sample. Due to concerns of inadvertently introducing blood into the fecal samples, we did not collect feces from the rectum postmortem. Future studies should seek to ensure that feces are collected from each individual animal using noninvasive methods, such as a temporary separation of the animals until a fecal pellet is produced. Three samples were excluded from gastritis histopathology scoring due to slides not containing pylorus, which was due to errors with sectioning resulting in the pylorus being inadvertently removed or not included in the final stained section. These 3 samples were each in a different treatment group; therefore, this is unlikely to have affected statistical significance. This could have resulted in falsely decreased mean gastritis scores if animals with gastritis scores of 3 or 4 were among the excluded samples. In addition, it is possible that the duration of dosing for this study (21 days) was not a sufficient time period to observe adverse effects at the 10-mg/kg carprofen dose. Longer dosing regimens may reveal adverse effects such as pyloric ulceration or renal dysfunction. Future studies should examine longer dosing regimens to determine if longer durations of 10 mg/kg carprofen administration result in an increased incidence of adverse effects.

In addition, creatinine, BUN, ALT, and ALP values were below the detection limit of the chemistry analyzer to provide a definitive numeric value. This was likely due to automated dilutions performed by the analyzer (IDEXX, personal communication via phone call, July 16, 2025). The detection limits for all values for all creatinine, BUN, ALT, and ALP were within or below the published reference ranges for mice.²⁸ Clinically significant kidney and liver dysfunction caused by chronic NSAID use would be expected to cause an increase above the upper end of the reference range for these analytes.¹⁷ However, given that kidney and liver dysfunction may be adverse effects of chronic NSAID use, it is important to precisely determine the values of these analytes to detect potential subtle changes that may occur. Therefore, future work should use methodologies that are able to consistently determine exact numeric values for these analytes in murine blood samples.

While the present study demonstrates the lack of toxic effects of 10 mg/kg carprofen administered for 21 days, this dose may be insufficient to achieve analgesia for some murine experimental models. A previous study showed that 20 mg/kg of carprofen was insufficient to achieve analgesia in female CD1 mice postovariectomy.²² Recent publications^45,46 showed that 25 mg/kg of carprofen in male and female C57BL/6J and NSG mice provided effective postoperative mechanical and thermal analgesia in incisional pain models, but carprofen at 5 mg/kg was not effective. Further work is needed to determine if carprofen dosed at 10 mg/kg is sufficient to achieve analgesia in mouse models of chronic pain, such as OA models.

In conclusion, the present study demonstrates the feasibility of chronic NSAID and PPI administration in a murine model. This model has potential utility in the field of OA research; however, further work is needed to determine the effects of these drugs in mice with induced OA. Potential confounding effects, including reduced serum albumin level and reduced WBC, neutrophil, and platelet counts, should be considered when using carprofen and carprofen + omeprazole, respectively, for research purposes in C57BL6 mice. Finally, all groups demonstrated lymphopenia, leukopenia, and decreases in body weight over time, which may be associated with the stress of repeated oral gavage. Dosing methods such as medicated food or drinking water may be less stressful than oral gavage and should be used when feasible. The use of refined dosing methods has the potential to reduce the confounding effects of stress on experiments and to improve research animal welfare.

Conflict of Interest

The authors have no conflicts of interest to declare.

Funding

Funding for this study was provided by Congressionally Directed Medical Research Program/Peer Reviewed Medical Research Program (CDMRP/PRMRP) PR220910 and Veterans Affairs (VA) Merit BX004882.

Protocol registration

A protocol including the research question, key design features, and analysis plan was prepared prior to initiation of the study. The protocol was not registered.

Data availability

All study data are available. Interested parties may contact the corresponding author.