Preliminary evaluation of a salivary urea test strip method for use in dogs Lucia Sanchini Cassia H.Z. Hare Olivier Restif Tim L. Williams Department of Veterinary Medicine, University of Cambridge, Cambridge, UK. Correspondence: Lucia Sanchini, Department of Veterinary Medicine, University of Cambridge, Cambridge, CB3 0ES, UK. lucia.sanchini@gmail.com Short title: Salivary urea test strip in dogs Abstract Background: Salivary urea concentrations correlate with serum urea concentrations in dogs and humans. Salivary urea concentrations can now be determined semi-quantitatively using a salivary urea test strip method that has been validated for use in humans. Objectives: Evaluation of the repeatability of a salivary urea test strip score and correlation between the salivary urea test strip score and serum urea concentration in dogs. Methods: Intra-run and inter-run variability were determined (n=10 in triplicate). Correlation between salivary urea test strip score and serum urea concentrations in dogs was assessed using Spearman’s correlation coefficient. Receiver operator curve analysis was used to evaluate diagnostic performance of the salivary urea test strip score to identify dogs with serum urea concentration >7.4 mmol/L (upper limit of laboratory reference interval). Results: Intra-run repeatability was good (28/30 concordant results) whereas inter-run repeatability was moderate (23/30 concordant results). Salivary and serum urea concentrations showed moderate positive correlation (rs= 0.63, n=33; P<0.0001). Salivary urea test strip score ≥4 was 57% sensitive and 96% specific for detecting a serum urea concentration >7.4 mmol/L. Conclusions: Uremia can be detected using salivary urea test strips in dogs. Based on our preliminary data, salivary urea test strip scores of 1 or 2 might exclude clinically relevant uremia in most cases, however it is recommended that the salivary urea test should be repeated in dogs with a test strip score of 3. Dogs with a salivary urea test strip score of >/=4 would likely require additional investigations. Introduction Saliva is a biological fluid produced predominantly in the salivary glands, with a negligible component derived from other sources, such as gastrointestinal reflux, the gingiva and oral mucosa (1). Various constituents enter saliva from the blood through passive diffusion or active transport and extracellular filtration (2). The mechanism of saliva production makes it functionally comparable to serum for detection and monitoring of a wide range of biomarkers, therefore in recent years, saliva analysis has become increasingly popular in human medicine, with chronic kidney disease (CKD) being an important field of application. Owning to ease of use and cost-effectiveness, salivary analysis has also become more popular in veterinary medicine. For example, recent publications (3,4) demonstrated a significant correlation between salivary and serum concentrations of urea (and creatinine) in dogs with CKD. In the aforementioned studies, salivary urea concentrations were determined using a biochemistry analyzer, however this methodology is expensive and therefore not widely available in general practice. In humans, salivary urea concentrations can be determined using a test-strip method (5) and a significant correlation was demonstrated between salivary and serum urea concentrations in patients with CKD (6). The advantage of this method is that it requires no laboratory equipment, as well as using only a small amount of saliva. Salivary urea test strips appeared to be useful as a screening tool for detection of kidney diseases in low-income countries (7) and in poorly compliant patients, such as infants, disabled or anxious patients (8). We hypothesized that salivary urea test strips could also be used in dogs. Therefore we aimed to evaluate the repeatability of the salivary urea test strip in dogs and to investigate the correlation between salivary and serum urea concentrations in this species. Materials and Methods Repeatability of the salivary urea test strip method Salivary samples were collected from healthy dogs owned by staff members of the University of Cambridge after obtaining written consent. To test the repeatability of the salivary urea test strip, the most likely error rate was estimated, based on a maximum likelihood computation. For intra-run repeatability, 10 different salivary samples were tested three consecutive times, at room temperature within five minutes of each other. Inter-run repeatability was evaluated by collecting 10 different salivary samples, each divided into three aliquots of 60uL. One aliquot was used immediately for the test, whereas the remaining two aliquots were stored at -80oC and tested on two consecutive days, after being thawed at room temperature for 30 minutes. Correlation of salivary and serum urea concentrations Adult client-owned dogs that were undergoing blood sampling for serum biochemistry (as part of medical work up or pre-anaesthetic screening) were recruited from the Queen's Veterinary School Hospital, Cambridge, UK between April and September 2018. The study was approved by the Ethics and Welfare Committee of the University of Cambridge (CR 271). Written consent was gained from pet owners prior to the collection of saliva. Blood samples were collected by jugular venipuncture and placed into plastic serum tubes. Following clot retraction and sample centrifugation the serum was extracted (usually within two hours of sample collection). Serum urea concentrations were measured by Central Diagnostic Services (Cambridge, UK) using an enzymatic method (Beckman Coulter, UK) on a Beckman Coulter AU480 biochemistry analyzer. Saliva collection and test strip analysis Saliva samples were obtained within two to six hours of blood sampling from dogs fasted for at least 6 hours. Saliva collection was performed by a single operator (LS) according to the following procedure. Three disposable sponge tipped makeup applicators (Essential Beauty Care, Amazon, UK, Supplementary Figure 1a) were inserted, one at a time, into the mouth and gently moved along the tongue and at the back of the oral cavity for approximately 30 seconds, in order to maximize the amount of saliva collected. Three miniature Eppendorf tubes (0.5 mL) were pre-prepared by drilling a small hole in the bottom of the tube with a 21 gauge needle (Figure 1). The three sponge tips containing saliva were each detached from the applicators and placed into an individual pre-prepared Eppendorf tube, which was positioned on top of a 5ml polypropylene centrifuge tube and secured using Parafilm® (Figure 2). These devices were centrifuged at 1431xg for 10 minutes (Denley BS400 centrifuge) in order to extract the saliva from the sponges. After centrifugation, 40uL of saliva was transferred (using a laboratory pipette) onto each pad (test pad and control pad in all cases) of the salivary test strip (Integrated Biomedical Technology, Elkhart, IN, USA), according to the manufacturer’s instructions. Saliva samples of insufficient volume to apply to the test and control pads were excluded. The salivary urea tests strip contains urease, which hydrolyses urea to carbon dioxide and ammonium hydroxide, and a pH indicator. This reaction results in an increase in pH, which is responsible for a visible color change of the pH indicator in the pad, and the increase in pH is proportional to the salivary urea concentration. Every strip contains a test pad and a control pad, the latter of which does not contain urease and thereby indicates the presence of interference in the color change due to high salivary pH. A change in the color of the control pad equivalent to the color block representing salivary urea test strip score 2 necessitates lowering of the test result by one level (-1). After one minute the color of the test pad was visually compared to the 6 color blocks provided by the manufacturer, which gives a semi-quantitative measure of the salivary urea concentration, corresponding to 6 different concentration ranges (Supplementary Figure 2). The same operator (LS) read all test strips and recorded the color score (see Table 2). In cases for which excess saliva was available, salivary pH was also determined independently using MColorpHastTM pH indicator strips. Statistical analysis The correlation between salivary and serum urea concentrations was assessed by the nonparametric Spearman rank test. Bland Altman plots were used to compare the difference between salivary and serum urea concentrations. In order to calculate the difference between the saliva and serum urea concentrations, each salivary urea test strip score was converted to a numerical value, which was the central value of the range of salivary urea concentrations that each score corresponded to. For example, for salivary urea test strip score of 1, the range of values is 1.8-5.2 mmol/L, therefore the numerical value assigned to test strip score of 1 was 3.5 mmol/L. These assigned numerical values were then used in the calculation of the saliva-serum difference. Receiver operator characteristic (ROC) curve analysis was performed using GraphPad Prism. Results The intra-run experiment showed variability of the results in only two measurements out of 30. In these two cases, one measurement within each set of three measurements differed from the others by one unit. This corresponds to a 7% of chance that a single salivary urea test strip measurement would not be repeatable when performed on the same day. There was more variability in the inter-run experiment, with three sets of readings (out of 10) recording consistent results and seven showing variation of the results. For six of these seven data sets, one reading within each set differed from the other two by one unit. For one set of data, the difference between the three readings was two units. Overall, the inter-run experiment demonstrated a 23% of chance that a single reading would be not repeatable in the same sample on different days. Thirty three dogs were recruited into the study, 19 of which were female and 14 of which were male. The median age was 7 years (range 1 - 13 years). The population was comprised of five Springer Spaniels, four Labrador Retrievers, four crossbreed, two Cockerpoos, two Shi-tzus, two Golden Retrievers, plus one of each of the following breeds: Dachshund, German Shepherd, Wheaten Terrier, Miniature Schnauzer, Flat Coated Retriever, Beagle, Staffordshire Bull Terrier, Yorkshire Terrier, Border Terrier, Lhasa Apso, Bichon Frise and Cavalier King Charles Spaniel. Based on the upper limit of the laboratory reference interval for serum urea concentration (7.4 mmol/L), seven dogs were classified as uremic and twenty six dogs were non-uremic. The median and range serum urea and creatinine concentrations in the two groups are shown in Table 1. The median salivary pH was 8.7 (range 7.9-9.5, n=18). Salivary urea test strip scores are shown in Table 2. Salivary and serum urea concentrations were moderately positively correlated (rs= 0.63, n=33; P<0.0001, Figure 3). In general, the measured salivary urea concentration was lower than the serum urea concentrations. Bland Altman analysis indicated that there was a proportional error between the measured salivary urea concentration and the serum urea concentration. This error was relatively consistent, with salivary urea concentration estimated to be on average approximately 50% of the serum urea concentration across the range of serum urea concentrations that were evaluated (Figure 4). The diagnostic performance of the salivary urea test strip method for the detection of uremia was determined by ROC analysis. When using a serum urea concentration of >7.4 mmol/L (upper limit of the laboratory reference interval) as the diagnostic criterion, the area under the ROC curve was 0.91 (95% confidence interval [CI] 0.79-1.0, P=0.001, Figure 5a), indicating excellent diagnostic performance. A salivary urea test strip score of ≥3 had a sensitivity and specificity of 86% (95% CI 42-100%) and 82% (95% CI 62-94%) respectively to detect a serum urea concentration >7.4 mmol/L. A salivary urea test strip score of ≥4 had a sensitivity and specificity of 57% (95% CI 18-90%) and 96% (95% CI 81-100%) respectively to detect a serum urea concentration >7.4 mmol/L. When using a serum urea concentration of >15 mmol/L (the serum urea concentration regarded by some internists as the clinical decision limit above which there is clinical justification for medical intervention such as intravenous fluid therapy or gastroprotectants [Barbara Skelly, personal communication]) was used as the diagnostic criterion, the area under the ROC curve was 0.99 (95% CI 0.96-1.0, P=0.006, Figure 5b) which also indicates excellent diagnostic performance. A salivary urea test strip score of ≥3 had a sensitivity and specificity of 100% (95% CI 29-100%) and 93% (95% CI 78-99%) respectively to detect a serum urea concentration >15 mmol/L. A salivary urea test strip score of ≥4 had a sensitivity and specificity of 67% (95% CI 9-99%) and 100% (95% CI 88-100%) respectively to detect a serum urea concentration >15 mmol/L. Discussion Analysis of salivary biomarkers gained increasing popularity in recent years (9) particularly since the collection of saliva is easy and non-invasive. Salivary urea and creatinine concentrations (determined on a biochemistry analyzer) are significantly increased in dogs with chronic kidney disease, and therefore these parameters can be used as markers of renal failure (4). The salivary urea test strips that were validated in the present study require minimum laboratory equipment, therefore these test strips could be useful for home monitoring of animals by owners or for veterinarians without access to biochemistry analyzers due to technical or financial constraints. In the present study, we demonstrated a significant positive correlation between the salivary urea test strip score and the serum urea concentration. This correlation is similar to that reported in human by Raimann and others (rs=0.63). In the aforementioned study, elevated serum urea concentrations were diagnosed with high accuracy by measurement of salivary urea concentrations (determined on a biochemistry analyzer). Using ROC analysis, the salivary urea test strip also demonstrated excellent diagnostic performance for the diagnosis of animals with a serum urea concentration above the upper limit of the laboratory reference interval (7.4 mmol/L) or above a serum urea concentration at which medical intervention might be justified (15 mmol/L). Salivary urea test strip values of 4 and 5 were associated with a high or high-normal serum urea concentration (based on the laboratory reference interval), whereas patients with salivary urea test strip values of 1 and 2 were very unlikely to be uremic. In a clinical context, where 15mmol/L is considered the threshold for clinically relevant uremia, a salivary urea test strip result of 1, 2 or 3 could probably be used to exclude clinically relevant uremia, based on our preliminary data. However, some patients with a salivary urea test strip value of 4 did not have a serum urea concentration >15 mmol/L, therefore evaluation of the serum urea concentration would be necessary to confirm the presence of uremia in these cases. Unfortunately, very few patients with very high salivary urea test strip values (5 or 6) were recruited in the present study, however in the two cases that were identified (both with a salivary urea test strip value of 5), the presence of a moderately elevated serum urea concentrations (>20 mmol/L) was confirmed. The sensitivity and specificity of the salivary test strip method for the diagnosis of uremia may have been influenced by differences in the salivary pH of dogs relative to humans. The salivary urea test strip method is based on changes in the test-pad colors due to the release of hydroxyl ions (and subsequent alteration of the pH) during the cleavage reaction of the salivary urea nitrogen by urease present in the test pad. A recent study (10) found that the average pH of canine saliva was 7.93, whereas in humans the salivary pH ranges from 6.2-7.6, with a mean pH of 6.7 (11). In our study, the median salivary pH was 8.7, more alkaline than previously reported (10), although this was only assessed in 18/33 dogs due to an insufficient volume of saliva being obtained. This higher salivary pH may falsely increase the salivary test strip value since the test pad detects alterations in the pH, although the results were corrected using the control pad according to the manufacturer’s instructions, which should have accounted for these differences. Many factors can affect oral pH, including the presence of normal resident urease-producing bacterial flora (12). Therefore patients with significant dental disease may have had falsely decreased salivary urea test strip scores. Complete oral examination was not conducted in our patients prior to the saliva collection therefore correlation of our results with the presence of dental disease was not possible. However, further investigation of the effect of dental disease on the salivary urea test strip is warranted. Salivary clearance of urea is also highly dependent on salivary flow rate under stimulated conditions (13). We did not stimulate salivary excretion in our patients, however it is possible that salivary flow rate of each dog could have been affected by clinical conditions (e.g. nausea), or medications, given that drugs can influence saliva flow rate in healthy humans (14). Water was also not always withdrawn prior to sampling, which could have affected the salivary flow rate in some cases. Although the salivary urea test strips demonstrate good diagnostic performance in human patients and in dogs with uremia (based on the preliminary results of the present study), the lowest urea test strip category encompasses the lower limit of the laboratory reference interval for serum urea, and so is unlikely to be useful for the detection of low serum urea concentrations. The main limitation of this study is the low number of uremic dogs that were tested, which reflects the relatively low prevalence of CKD in the canine population referred to the QVSH. Evaluation of the sensitivity of the salivary urea test strips in a larger number of uremic patients is warranted. Unfortunately, many uremic patients that presented to QVSH during the study were dehydrated, and collection of sufficient saliva to permit our analysis was not possible. This also reflects a possible limitation of this method, in that it may not be useful in animals with dehydration and tacky mucous membranes. When we assessed the inter- and intra-run repeatability of the test, some variability was detected, mainly when the test was repeated on different days. This may have been an effect of sample storage and thawing, however, to the authors’ knowledge, no published data regarding storage effects on human or canine saliva urea are currently available. In light of the variability detected, dogs with a salivary urea test strip result of 3 or 4 may warrant repeat testing to confirm the categorization of uremia prior to further investigations. In a clinical context, the inter-observer variability could also reduce the accuracy of the salivary urea test strip as a marker of uremia, however the inter-observer co-efficient of variation of the test strip was ~5% in a previous study (when evaluating human salivary urea), therefore the degree of error attributable to inter-observer variability is likely to be negligible. A further limitation of this study was that we did not measure the salivary urea concentration using a biochemistry analyzer. This was due to the small amount of saliva that was collected and also was not deemed to be clinically relevant since we were ultimately interested in the correlation between the salivary urea test strip score and the serum urea concentration. In summary, the salivary urea test strip is a rapid, non-invasive and convenient test which shows moderate correlation with serum urea concentrations in dogs. Only a small amount of saliva is required for the test, which can be convenient in cases of distressed or poorly compliant patients, or where access to biochemical analyzers is limited. Based on our preliminary evaluation, dogs with a salivary urea test strip result of 3 would warrant repeat testing and those with a result of 4 or greater would likely require additional investigations to confirm and further characterize the cause of uremia, however larger studies are required to establish the sensitivity and specificity of the urea test strip test to detect uremia in dogs. Acknowledgments All the members of the staff that volunteered their own dogs for the method validation are acknowledged. We are grateful to the clinicians of the QVSH (and in particular to Andrea Mosca) for helping with the recruitment of the patients. Funding No funding was received for this study. Salivary urea test strips were provided free of charge by the manufacturer (Integrated Biomedical technology). We acknowledge Wen Wu for his assistance with the study. Conflict of interested None declared. References 1. Nagler RM. Saliva analysis for monitoring dialysis and renal function. Clin Chem. 2008;54(9):1415-1417. 2. Javaid MA, Ahmed AS, Durand R, Tran SD. Saliva as a diagnostic tool for oral and systemic diseases. J Oral Biol Craniofac Res. 2016;6(1):66-75. 3. Raja M, Suganya G, Leela V, Balagangathare T. Estimation and correlation of Urea and Creatinine Levels in Saliva and Serum of Dogs with Renal Failure. International Journal of Science, Environment and Technology. 2017;6(5):2744-2751. 4. Tvarijonaviciute A, Pardo-Marin L, Tecles F, et al. Measurement of urea and creatinine in saliva of dogs: a pilot study. BMC Vet Res. 2018;14(1):223. 5. Kaczor-Urbanowicz KE, Martin Carreras-Presas C, Aro K, Tu M, Garcia-Godoy F, Wong DT. Saliva diagnostics - Current views and directions. Exp Biol Med (Maywood). 2017;242(5):459-472. 6. Raimann JG, Kirisits W, Gebetsroither E, et al. Saliva urea dipstick test: application in chronic kidney disease. Clin Nephrol. 2011;76(1):23-28. 7. Evans R, Calice-Silva V, Raimann JG, et al. Diagnostic Performance of a Saliva Urea Nitrogen Dipstick to Detect Kidney Disease in Malawi. Kidney Int Rep. 2017;2(2):219-227. 8. Liu J, Duan Y. Saliva: a potential media for disease diagnostics and monitoring. Oral Oncol. 2012;48(7):569-577. 9. Tvarijonaviciute A, Barranco T, Rubio M, et al. Measurement of Creatine kinase and Aspartate aminotransferase in saliva of dogs: a pilot study. BMC Vet Res. 2017;13(1):168. 10. Iacopetti I, Perazzi A, Badon T, Bedin S, Contiero B, Ricci R. Salivary pH, calcium, phosphorus and selected enzymes in healthy dogs: a pilot study. BMC Vet Res. 2017;13(1):330. 11. Baliga S, Muglikar S, Kale R. Salivary pH: A diagnostic biomarker. J Indian Soc Periodontol. 2013;17(4):461-465. 12. Grant Maxie M. Jubb, Kennedy and Palmer's Pathology of Domestic Animals. Vol 2nd. 6th ed. Missouri: Elsevier; 2016. 13. Watanabe J, Mizuno S, Masuda N, et al. Salivary excretion of urea in dogs. J Pharmacobiodyn. 1984;7(5):294-303. 14. Dawes C. Physiological factors affecting salivary flow rate, oral sugar clearance, and the sensation of dry mouth in man. J Dent Res. 1987;66 Spec No:648-653. Table 1. Serum urea and creatinine concentrations in dogs with and without uremia. Uremic (n=7) Non-uremic (n=26) Laboratory Reference Interval Median (range) serum urea concentration (mmol/L) 13.9 (7.8-24.4) 5.3 (1.4-7.1) 2.5-7.4 Median (range) serum creatinine concentration (mmol/L) 108 (71-263) 84 (28-110) 34-136 Table 2. Table showing possible results for salivary urea test strip, the observed color change, the salivary urea concentration range for each category and the number the patients which were classified in each category. Salivary Urea Test Strip Value (range in mg/dL) Corresponding range in mmol/L Color of the test pad No of animals with salivary urea concentration in this category 1 (5-14) 1.8-5.2 11 2 (15-24) 5.2-8.8 11 3 (25-34) 8.8-12.3 5 4 (35-54) 12.3-19.4 4 5 (55-74) 19.4-26.6 2 6 (>75) >26.6 0 Figure 1 Picture showing a mini Eppendorf tube (0.5 mL) pre-prepared by drilling a small hole in the bottom of the tube with a 21 gauge needle. Figure 2 Picture showing a mini Eppendorf tube secured to the centrifuge tube using Parafilm®. Figure 3 Correlation between salivary urea test strip score (1-6) and serum urea concentration. The left dotted line represents the upper limit of the laboratory reference interval for serum urea concentration (7.4 mmol/L) and the right dotted line represents the serum urea concentration above which some clinicians would consider medical intervention (15 mmol/L). Figure 4 Bland Altman plot showing the difference between salivary urea concentration and serum urea concentration as a percentage of the serum urea concentration. In order to calculate the difference between the saliva and serum urea concentrations, each salivary urea test strip score was converted to a numerical value, which was the central value of the range of salivary urea concentrations that each score corresponded to. These assigned numerical values were then used in the calculation of the saliva-serum difference.. The salivary urea test strip underestimated the serum urea concentration in most (28/33) samples. Figure 5a and 5b Figure 5a. Receiver operating characteristic curve for diagnosis of uremia (serum urea concentration >7.4 mmol/L [the upper limit of laboratory reference interval]) in 33 dogs (7 of which were uremic) by use of a salivary urea test strip. The area under the curve was 0.91 (95% confidence interval 0.79-1.0, P=0.001). The red line indicates the line of identity. Figure 5b. Receiver operating characteristic curve for diagnosis of uremia (serum urea concentration >15.0 mmol/L [serum urea concentration considered clinically significant by some clinicians]) in 33 dogs (3 of which were uremic) by use of a salivary urea test strip. The area under the curve was 0.99 (95% confidence interval 0.96-1.0, P=0.006). The red line indicates the line of identity. Supplementary material Figure 1 Picture showing the size of the disposable sponge tipped make up applicator (Essential Beauty Care, Amazon UK) used for saliva collection. Figure 2 Picture showing a salivary urea test strip. CONTROL PAD PAPPADPAD TEST PAD PAPPADPAD image4.emf 10 20 30 -100 -50 0 50 100 150 Serum urea concentration (mmol/L) S a l i v a r y - s e r u m u r e a c o n c e n t r a t i o n ( m m o l / L ) % e r r o r image5.tiff image6.jpeg image7.jpeg image1.jpeg image2.jpeg image3.tiff