Animals.

All animal procedures were approved by the Centers for Disease Control and Prevention IACUC and conducted in accordance with the Guide for the Care and Use of Laboratory Animals at an AAALAC international-accredited facility. Healthy strain 13/N guinea pigs were selected from the Centers for Disease Control and Prevention breeding colony. Health monitoring of the colony includes serology (guinea pig basic Opti-Spot, IDEXX BioResearch, Columbia, MO), fecal examination, and bloodwork, including CBC and clinical chemistry analyses. All study animals were free of Clostridium piliforme, guinea pig parainfluenza virus 3, lymphocytic choriomeningitis virus, murine pneumonia virus, Sendai virus, and Encephalitozoon cuniculi, as determined by using serologic screening tests. Adult animals selected for study had no history of preexisting conditions (e.g., dental malocclusion, previous cystotomy, previous surgical mass removal, recurrent cystitis).

One 88-d-old female in the juvenile experimental group was delayed starting the study after developing circling behavior and reduced weight compared with conspecifics at the first timepoint. She was removed from study and treated empirically for otitis media/interna by our clinical veterinary staff with sulfamethoxazole and trimethoprim (Pharmaceutical Associates, Greenville, SC) 20 mg/kg PO twice daily for 18 d, and Critical Care Herbivore (Oxbow Animal Health, Omaha, NE) nutritional supplement PO twice daily for 18 d, and clinical signs resolved. After she recovered, she was reassigned to the control juvenile group since there were only 4 females available from the contemporary litters for this group. Otherwise, physical examinations and clinical chemistry were within normal limits for all study animals at baseline.

A total of 82 guinea pigs were divided by sex into cohorts of juveniles and adults: 37 females and 45 males; 35 juveniles and 47 adults (Table 1). Animals were socially housed in sex-segregated floor pens with paper (Techboard and Poly Pads, Shepherd Specialty Papers, Watertown, TN) and paper nesting material (EnviroDri, Fibercore, Cleveland, OH). Guinea pigs were provided unrestricted filtered municipal water via bottle with sipper tubes. Environmental enrichment was provided in the form of plastic huts (guinea pig hut, Bio-Serv, Flemington, NJ) and wood materials for gnawing (manzanita sticks and wood blocks, Bio-Serv). Environmental parameters were maintained within a temperature range of 68 to 79 °F (20.0 to 26.1 °C), 30% to 70% relative humidity, and a 12:12 h light:dark cycle. Age group guidelines at baseline were defined as 0- to 199-d-old for juveniles and 200- to 900-d-old for adults, adapted from previously established guidelines.¹⁵ Animals older than 900 d were considered geriatric. All animals were of intact reproductive status.

Adult animals were randomly selected groups of socially compatible animals. Juveniles were randomized by assigning animals to study as the next available litters were weaned. One juvenile assigned to the male cohort was diagnosed with a disorder of sexual development after the cessation of the study. Data from this animal were included in the male juvenile data set after no outlier analytes were identified. The first phase of the study was performed on the adult cohort and the second phase included procedures on the juvenile cohort and a 26-wk timepoint for the adult cohort. To facilitate blinding of investigators, all animals were assigned a unique study ID that did not indicate age, sex, or diet.

The juvenile cohort consisted of animals aged 88 to 197 d (median, 102 d) weighing 430 to 834 g (median, 633 g) at baseline. Juveniles were followed for a total of 14 wk, after which point some animals were bred. The adult cohort consisted of animals aged 254 to 865 d (median, 447 d) weighing 730 to 1,215 g (median, 962 g) at baseline. Adult animals were followed for a total of 26 wk, except for one male from the experimental group that was euthanized prior to week 26 for clinical reasons determined to be unrelated to this study. Longitudinal weight, body condition scores, blood (plasma chemistry), and urine (urinalysis) samples were collected at baseline and 4, 14, and 26 wk. Two males reached geriatric age (> 900 d) by week 14, and 3 males were geriatric by week 26. No females were geriatric by week 26 of study.

Diet.

Two closed-formula diets were chosen for adherence to guidelines published by the National Research Council (NRC), commercial availability, specific formulation for guinea pigs, and difference in primary ingredient, but otherwise had similar nutrient composition. The control group was maintained on the colony diet: AHP diet ad libitum (guinea pig diet 5025, LabDiet) and a daily allotment of ½ flake of loose timothy hay (sourced from Michigan) per 12 animals. The experimental group was fed 1 oz daily of a THP diet (timothy-based guinea pig diet, formula number 5664, Mazuri) with ad libitum loose timothy hay replenished throughout the day and given a second ½ flake per 12 animals in the afternoon. The ad libitum loose timothy hay was given to increase available roughage, and correspondingly decrease the overall calcium by providing an ad libitum food option that is lower in calcium than pelleted feed. Timothy hay calcium content can range from 0.32% to 0.46% depending on when in the bloom cycle it is harvested.¹⁶ Edible enrichment was provided in the same manner as described below for the control diet. Animals in experimental groups were transitioned from the control diet to the experimental diet incrementally during a period of 7 d. Both pellet diets have calcium contractions (1.10%) and phosphorus (0.6%) above what is recommended by the NRC for growing animals (0.8% and 0.4%, respectively). The NRC guidelines recommend magnesium concentrations of 0.1%; however, the experimental diet contains 0.4% and the control diet contains 0.35% magnesium. Finally, the NRC recommend 0.5% potassium, and the experimental diet and control diet contain 1.20% and 1.42%, respectively.³¹ These concentrations are calculated by the manufacturer based on most recent nutrient analysis information. Manufacturer guaranteed analysis and selected nutrient/chemical composition are provided (Tables 2 and 3).^24,28

Prior to the study, colony management practices included providing edible enrichment in the form of fresh vegetables 3 times a week; these vegetables were selected for vitamin C content, palatability, and availability. During the first phase of the study, the enrichment items given to the adult animals were restricted to carrots, Swiss chard, and bell peppers (red, yellow, or green). This combination of items was selected based on historic availability from vendor, historic animal preference, low calcium content (1% to 2% daily value based on human daily value requirements), or high vitamin C content, especially from bell pepper (190% daily value) and Swiss chard (18% daily value). In the second phase (juvenile animals), edible enrichment was not restricted by item to reduce the risk of development of food-related neophobia.^7,12 Enrichment selection was also widened for the adult cohort during weeks 15 to 26 to capture effects on broader enrichment selection in this group. The selection of enrichment items during this phase included carrots and bell peppers, given most often; kale, celery, mustard greens and snow peas, given 2 to 5 times during this period; and other items such as broccoli, parsley, lettuce, Swiss chard, Brussels sprouts, turnip greens, bok choy, spinach, and cilantro, only given once to a given group during this period. The trends of mixed enrichment were consistent across all groups, represent the routine selection of enrichment for our colony, and let us evaluate the impact of the diet modifications in the context of our routine practice.

Weight, body condition, blood collection, and plasma chemistry.

Weights were collected using a pediatric scale (Scale-Tronix 4802, Welch Allyn) calibrated annually and checked for accuracy prior to each group of weight collections with a calibrated weight. Body condition scores were assessed by one investigator at all timepoints and based on a modification of previously published scoring systems (Table 4).^36,46 Weights were collected weekly for the first 4 wk of the adult cohort. When no problems with weight maintenance were seen, weights were collected again at weeks 14 and 26. For the juvenile cohort weights were collected at weeks 0, 4, and 14.

Up to 1 mL of blood was collected from the cranial vena cava under general anesthesia (3% to 4% isoflurane via induction box, then nose cone) into a 1.3-mL lithium heparin Microtainer tube (Sarstedt, Nümbrecht, Germany). Samples were placed on ice, then plasma was separated and frozen at −80 °C in 2-mL microtubes (PCR-PT, Sarstedt, Nümbrecht, Germany) for later analyses. Plasma chemistry was analyzed using the comprehensive metabolic panel on the Piccolo Xpress analyzer (Abaxis). Samples were coded at time of collection and stored frozen 4 to 66 d prior to analysis.

Urine collection.

Urine was collected from each animal by natural, conscious urination or urination induced by anesthesia. If a sample was not collected at a scheduled timepoint, the sample was collected up to 10 d after. All urine samples were collected from a sanitized transport box (urination while conscious), anesthesia induction chamber, or free catch with stainless steel collection cup while under general anesthesia. Prior to sampling, each animal was individually held in a sanitized clear filter top static cage. Animals were monitored and urine was collected as soon as it was seen in the cage by aspirating it into a 3-mL syringe (Henke-Sass, Wolf, Tuttlingen, Germany); urine was then placed in a 2-mL microtube (PCR-PT, Sarstedt, Nümbrecht, Germany) and immediately placed on ice.

Urination was often stimulated with the induction of anesthesia and collected from the induction box. When urine samples were collected during induction of anesthesia, the induction box was cleaned with Peroxigard (Virox Technologies, Oakville, ON, Canada) between each animal and dried with a paper towel. Animals were monitored during induction, and, when produced, urine was immediately aspirated into a 3-mL syringe (Henke-Sass, Wolf, Tuttlingen, Germany), then placed in a 2-mL microtube (PCR-PT, Sarstedt, Nümbrecht, Germany) and immediately placed on ice.

After the first adult baseline sampling, a collection modification was adopted: use of a disposable Elizabethan-style collar to prevent contamination of the urine sample with lacrimal and/or salivary secretions. An Elizabethan-style collar was developed using an 8-in.-diameter paper coffee filter (WinCo Foods, Boise, ID). A 3-cm ‘X’ was cut in the center of the filter and slipped over the head and ears for use during anesthesia to catch secretions. Each disposable filter was used once per animal.

If an animal did not urinate in either the transport box or the induction box, a stainless-steel cup was positioned near the prepuce or vulva to catch urine if the animal urinated during the blood draw procedure. Only 3 urine samples were collected this way. Of the 292 urine collections, 205 occurred in the induction box (70%), 84 occurred in the transport boxes (28%), and 3 were collected in the stainless-steel cup (1%). Males and juveniles were more likely to produce a urine sample at anesthesia induction than while conscious in a transport box.

Urinalysis.

Urine samples were evaluated for turbidity, urine specific gravity (USG), chemistry, and sediment within 2 h of collection or placed in refrigerator (4 °C) and analyzed within 24 h of collection. Methods and workflow are summarized in Figure 1. Most samples were read within 6 h of collection. Refrigerated samples were allowed to equilibrate to room temperature for at least 20 min before analysis. To assess turbidity, settled sediment was resuspended using a VWR VM-3000 mini vortexer (Henry Troemner).

View Figure

• Download as Powerpoint
• Download Figure

Some sources state that an accurate urinalysis should be run on uncentrifuged urine samples and then repeated if blood or turbidity impedes the results.⁵ Thus, during the first timepoint, USG, urinalysis test strips, and sediment were read on samples both before and after centrifugation to determine the method that would prove reliable and most time efficient. While chemistry test strip results were generally consistent between precentrifuged urine and postcentrifuged supernatant, notable differences were observed in the USG and sediment analyses. For USG determination, urine samples that had moderate to marked turbidity blurred the shadow line, making interpretation difficult. Spinning down the sample and reading USG from the supernatant consistently corrected this issue. Sediment analysis was also impacted, as centrifugation concentrates the material for microscopic examination. We elected to always read the sediment after centrifugation to capture all crystals in a sample, rather than reading part of the urinalysis on precentrifuged urine and part on postcentrifuged urine.

Urine samples were coded at the time of collection, urinalysis strips performed by 5 investigators, and turbidity and sediment were evaluated by one investigator. Urine turbidity was noted first (Figure 1), after which samples were centrifuged (5810, Eppendorf) at 580 × g for 5 min. After centrifugation, supernatants were poured off into a separate tube and the sediment was set aside. Supernatants were used for USG and chemistry readings. USG was read by a handheld refractometer (Schuco 5711 to 2020). Calibration was checked with distilled water each day that urinalyses were run. If the shadow line did not read 1.000 with distilled water, the calibration was adjusted so that the shadow line matched the 1.000 line. Urine chemistry was performed with VetStix 11 test strips (Vet One, Boise, ID) by placing a drop of urine on each test pad using a 200-µL pipette and waiting the recommended time interval before reading (30 to 60 s for bilirubin, urobilinogen, ketone, protein, nitrite, glucose [GLU], pH, occult blood, USG, and ascorbic acid; 90 to 120 s for leukocytes). For sediment analysis, supernatant was either removed or added so that the total volume in the sediment tube equaled 0.25 mL. The sediment was then resuspended by gently stirring the puck of material with the pipette tip until a homogeneous mixture was obtained. Three drops of resuspended sediment were placed on a glass slide, covered with a glass coverslip, and viewed with light microscopy at 20× at the center of the coverslip. All sediment samples were read by one investigator.

A crystal load scoring system was developed to categorize the findings on sediment analysis. Crystal load was graded as 1 to 4 (Figure 2): 1) thin layer with ample negative space between crystals and no large clumps of crystals; 2) moderate amount of crystals with reduced negative space, or frequent large clumps of crystals; 3) thick layer with no negative space visible between crystals or large areas of clumped crystals, but crystal morphology can be readily determined; and 4) transillumination obscured, creating a dark and blurry field, and crystal morphology could only be determined by looking at the margin of the coverslip where crystal accumulation was thinnest. Red blood cells were also noted as present or not present.

View Figure

• Download as Powerpoint
• Download Figure

Statistical methods.

We performed initial data analysis for the 25 analytes to include overall mean and SD across all inputs. ANOVA was conducted for each of the 25 analytes, using sex, diet, age, and sample timepoint as the independent variables. To statistically evaluate effects of diet, all parameters were consolidated independently of sample timepoint to determine whether the input variables produced statistically significant models based on time, sex, age, and diet. We evaluated the residuals from the models and found the ANOVA model appropriate. We used the Benjamini–Hochberg technique of controlling the false discovery rate to balance the power and type-1 error of our analysis with an α = 0.05.¹⁵ Further analysis was performed on the 19 analytes found to be statistically significant to identify interactions between diet with sex and age, including analysis of first-order interactions between diet and age (adult × diet) and diet and sex (sex × diet); this allowed us to determine the effects of diet on subpopulations (age and/or sex). Stepwise regression was used for each significant model to find the best model for each analyte, excluding insignificant variables at α = 0.05. To confirm adequate group sizes, a power analysis was conducted based on the aim to assess differences between the 2 diets, males and females, and adults and juveniles with the following parameters: α = 0.05, β = 0.1, and 1 SD distinction; results indicated that for each combination of diet/sex/age there should be a minimum of 5 animals.

Body weight, body condition, and blood chemistry.

Twenty-five analytes were selected for ANOVA using sex, diet, and sample timepoint as the independent variables. In this input variable testing, we found 19 analytes that were statistically significant overall (Table 5): weight, ALP, sediment calcium oxalate, plasma potassium, plasma calcium, overall sediment crystal burden, plasma albumin (ALB), ALT, GLU, urine turbidity, urine protein, urine USG, urine pH, plasma total protein (TP), plasma chloride, plasma BUN, body condition score, plasma total bilirubin (TBIL), and plasma creatinine (CRE). In addition, AST and sediment struvite had P values below 0.05 but were not significant under the criteria for multiple comparisons.

Of the 11 analytes found to be significant by feed (Table 6), 5 were unaffected by either sex or age (potassium, ALB, ALT, GLU, BUN). Four were unaffected by sex (ALB, GLU, chloride, TBIL), whereas 7 were found to have interactions with sex (ALP, potassium, calcium, ALT, TP, BUN, CRE). Similarly, 2 were unaffected by age (GLU, CRE), and 9 were found to have interactions with age (ALP, potassium, calcium, ALB, ALT, TP, CL, BUN, TBIL). The experimental diet had no impact on chloride, TBIL, and CRE; overall, values for study animals remained within published ranges.^9,37,45 The experimental diet affected 8 of the analytes. It reduced potassium (average difference of 0.6 mmol/L; in adults only), calcium (average difference of 0.43 mg/dL; males only), and ALT (average difference of 2.8 U/L) and GLU (average difference of 8.4mg/dL), both independent of sex or age. The experimental diet increased ALP (average difference of 18.8 U/L) and ALB (average difference of 0.34 g/dL) in juveniles only, BUN (average difference of 2.4 mg/dL for males and average difference of 4.0 mg/dL for juveniles), and TP (average difference of 0.14 g/dL, independently of sex or age).

Weights were significantly different between diet groups for the male cohort and juvenile cohort. The males fed the experimental diet had an average weight of 0.891 kg compared with the males fed the control diet with an average weight of 0.992 kg (101 g difference between averages). The juveniles fed the experimental diet had an average weight of 0.701 kg compared with the control diet with an average weight of 0.781 kg (80 g difference between averages). These differences were mirrored by significant reductions in body condition score for males and juveniles fed the experimental diet; however, these changes were clinically minimal, equaling a fraction of a point on the body condition scale with all animals between ideal weight and overweight, but not obese. Despite male and overall juvenile weights being lower, there was not a significant difference in female weight or body condition scores between diet groups.

Urinary parameters.

After urinalysis results were reviewed, we elected to include 9 urine parameters in statistical models based on suspected relevance to in vivo urolith formation: sediment crystal burden of sample, occurrence of calcium oxalate crystals, occurrence of struvite crystals, occurrence of calcium carbonate crystals, occurrence of calcium phosphate crystals, urine turbidity, urine protein, USG, and urine pH. Six analytes were statistically significant using sex, diet, and sample time as the independent variables (sediment crystal burden, occurrence of calcium oxalate crystals, urine turbidity, urine protein, USG, and urine pH). The experimental diet affected only 2 of the urine analytes (Table 7): it reduced urine turbidity (in females only) and increased USG (in adult males only). Overall, the experimental diet had no impact on total crystal estimate in sediment, calcium oxalate crystals in sediment, or urine protein. Urine pH was observed to significantly decrease over time with no significant interactions with age, sex, or diet. Sediment crystal burden, occurrence of calcium oxalate crystals, urine protein, and average USG interacted with age, but overall, males showed on average more concentrated urine on the experimental diet compared with the control (average of 1.024 and 1.019, respectively). Some changes were observed independently of diet. Males had a higher crystal load than females; juveniles had a higher crystal load than adults; and calcium oxalate crystals increased with age among males. Four types of crystals were seen on sediment analysis: calcium carbonate, calcium oxalate monohydrate, calcium phosphate, and struvite. Overall, 34% percent of juveniles (12 of 35) had only one type of crystal at all time points (calcium carbonate). All adults and 51% of juveniles (18 of 35) had calcium oxalate crystals at one or more time points. Calcium oxalate crystals are subclassified as monohydrate, dihydrate, or trihydrate; the morphology most consistently seen in these animals was monohydrate. Calcium phosphate and struvite crystals were noted less frequently. Only 11% of adults (5 of 47) and 29% of juveniles (10 of 35) had struvite crystals observed at one or more timepoints. Eleven percent of juveniles (4 of 35) had calcium phosphate crystals at least once, but none of the adults had this crystal type. A statistical analysis was performed to assess the impact of different urine collection methods, on the sediment crystal burden. Between the most frequent methods used, we found urine collected from the induction box (mean = 3.17, SD = 0.73, n = 205) compared to the transport box (mean = 2.21, SD = 0.79, n = 84) had a 0.96 higher sediment crystal burden grade (p-value = 1.85 E-19).