Title

Urine dipstick precision with standard visual and automated methods within a small animal teaching

hospital

Authors

Marisa da Fonseca Ferreira, IMVM DipECVIM-CA (Internal Medicine) MRCVS

Marta Garcia Arce, LV MRCVS

Ian Graham Handel, BVSc PhD MRCVS

Craig Robert Breheny, BVM&S MRCVS

Adam George Gow, BVM&S PhD DSAM DipECVIM-CA (Internal Medicine) FHEA MRCVS

Address for all authors

Hospital for Small Animals

The Royal (Dick) School of Veterinary Studies, The University of Edinburgh

Easter Bush Campus

Roslin

Midlothian

EH25 9RG

United Kingdom

Corresponding author’s e-mail address

marisa.ferreira@ed.ac.uk
Abstract

Urine dipstick results may vary between operators/methods. The magnitude of variation across the

veterinary field is currently unknown. The aim of this study was to compare the precision of urine

dipstick results between standard direct visual and automated reading methods, when performed by

several operators. Urine samples were pooled and divided into three aliquots: one plain, one with

glucose, and one with serum. Final year students, veterinary surgeons and veterinary nurses, blinded

to each sample, were then asked to perform dipstick analysis with direct visualization and an

automated analyser, and their technique was observed. A subsequent session was undertaken with

samples which had pH titrated to achieve an acidic, neutral or alkaline value. Sixty-four veterinary

students, 20 veterinary surgeons and seven veterinary nurses, performed the first (n=61) or second

(n=30) part of the study. Precision was greater using the automated reader. The most common

observed technique errors were; lack of sample mixing, for both visual and automated methods, and

not timing readings as per manufacturer instructions, when performing visual analysis. This study

suggests that in an environment with multiple operators, as is the case in veterinary teaching or

large private hospitals, automated urine dipstick reading improves precision of results.

Introduction

Urine dipstick analysis is one of the most frequently performed point-of-care (POC) tests, both for

diagnostic and monitoring purposes. It is considered part of a minimum database when assessing

not only urinary tract disorders, but also diseases in other body systems, for example, endocrine and

haematological diseases. Analysis is encouraged to be performed in-house, due to sample

degradation over time.1 Urine reagent strips developed for human medicine are often used in

veterinary medicine, given the lack of species specific products. Although the reagent strips are

designed to be a simple test, there is still the possibility of result variation between operators and/or

methods.2-4 The standard direct visual dipstick reading method involves comparing the colour of a

specific reagent impregnated pad against a reference chart provided by the manufacturer to obtain
the result. Therefore, this interpretation is subjective as colour perception varies between

individuals. There is also the possibility that individuals with colour vision deficiencies (CVD, (colour

blindness)) may record different results and that individuals may be unaware that they have CVD. 5 In

addition, visual analysis is liable to human error, specifically reading the wrong result by skipping a

pad in the chart, and not timing each reading as specified by the manufacturer, who generally

provides this information printed on the dipstick container.6 At present, the expected magnitude of

variation in results is unknown when obtained by veterinary clinical personnel. It is standard

practice in the authors’ institution that visual dipstick reading is performed without reference to a

standard operating procedure (SOP) and the authors would assert that this would be standard

practice in private veterinary surgeries. In the authors’ institution the technique is taught to

undergraduates as a “Day One Competence”.7

Automated urine dipstick analysis has been developed in an effort to improve accuracy, precision

and efficiency; aiming to overall enhance patient care.8-10 The results obtained are either available in

a printed form or directly updated into the patient’s medical record, likely reducing the risk of post-

analytical faults; namely manual transcription errors and loss of information. Additional advantages

might include an in-built memory for results and automatic error messages if the strip has been

exposed to humidity.11

The use of automated POC dipstick analysers has been validated in both dogs and cats and

documented through a few studies.12-1711 However, the visual readings undertaken in these

publications were performed by one or two experienced subjects. In human clinical practice, it has

been shown that when compared to visual reading, automated POC analysis of urine dipsticks may

be associated with higher sensitivity, higher reproducibility and higher agreement with a laboratory

reference method, being also user-friendly.18-20 Consequently, the automated method is currently

recommended21 and several human medical institutions routinely employ it.22; 23 However, these

human medical studies were limited in scope (one parameter investigated)18, number and

experience of subjects19; 20, as well as utilising different strips for visual and automated methods20. At
present, no studies have examined the potential for variation, and potential clinical impact of this

variation, in a large cohort who commonly perform this POC test, particularly in the veterinary field.

The objective of this study was therefore to compare the precision of canine urine dipstick chemical

analysis results between standard direct visual and automated reading methods, when performed

by a large cohort of operators in veterinary clinical practice. It was hypothesised that inter-operator

reproducibility would be higher with automated analysis compared with visual reading, due to

decreased operator subjectivity/error (e.g. an automated analyser would reduce the potential for

operator related error by standardising reading timings, avoiding reading the incorrect pad against

the reference and removing operator subjectivity of colour).

Materials and methods

A prospective design was followed for this study and approval was obtained from both the

institution’s ethical review and education research committees.

Urine samples were obtained during routine clinical management of hospitalised canine patients in a

small animal referral teaching hospital. All the samples were frozen and prior to the study day,

thawed at 4oC, mixed and divided into three aliquots (850mL each): one plain, one mixed with 4.7mL

of a 50% glucose solution, and one mixed with 25mL of bovine serum containing 52g/L of total

solids. A fourth aliquot mixed with blood was initially planned, though not carried out, given the

presence of moderate to large amounts of blood already measurable by dipstick analysis in the

original pooled sample. Samples were kept refrigerated throughout the whole study period of 8

hours.

Subjects (n=61) were recruited throughout the study day and included final year veterinary students

on clinical rotations, as well as veterinary nurses and veterinary surgeons who were on clinical duties

during that day. A recruitment sheet was handed at the time containing details on the study’s

purpose, authors, participants’ rights and benefits, as well as confidentiality. Final year students had
had practical classes on urine analysis including the correct use of dipsticks, namely the importance

of read timings and mixing the sample prior analysis, and routinely performed this test in final year

rotations within the small animal teaching hospital and during extramural studies. Blinded to the

type of sample, provided in three numbered aliquots, each subject was requested to perform

dipstick analysis on every aliquot, first by direct visual reading, and then by using a POC automated

analyser based on reflectance photometry (Clinitek Status® Analyser, Siemens Healthcare

Diagnostics Inc.), which was previously validated for analysis of canine urine.13

The same type of commercial reagent strips was used for both visual and automated analysis

(Multistix® 10 SG Reagent Strips, Siemens Healthcare Diagnostics Inc.). Calibration of the automated

analyser was automatically performed each time a test would be undertaken, through an in-built

white calibration bar.

A self-confidence assessment was given and requested to be filled prior the analysis, with the

question “How confident are you in performing urine dipstick chemical analysis?”. The answer was

marked in a grade of 1 to 5, with 1 being “not confident at all” and 5 being “totally confident”. An

instruction and recording sheet was also provided to each subject (Supplementary File 1a). Subjects

were not provided with other (e.g. verbal) instructions .The written instruction and recording sheet

was designed to focus on the number of samples and task required, rather than describing the

technique of how to correctly perform urine dipstick analysis, as all subjects were all at the stage of

their career where they would be expected to perform full urine analysis competently.. However, it

was stated in the instruction and recording sheet that subjects should use the materials available to

perform the analysis. These included a clock with a second hand for timing readings, as well as

gloves, paper towel, a pen, hand sanitiser and a clinical waste bin. In addition, clear bullet point

operating procedure instructions were available next to the automated analyser, as this was

assumed to be an unfamiliar piece of equipment to all subjects (Supplementary File 1b). The

subjects’ technique was observed during the study period and recorded by three of the authors, who

sat in the centre of the room.

A subsequent session was undertaken with pH titrated urine samples. On the day of analysis, the

pooled sample was divided into three aliquots, each titrated with hydrochloric acid and sodium

hydroxide to achieve a visual urine dipstick pH reading of 6 (sample 1), 7 (sample 2) and 7.5 (sample

3) agreed between two of the authors. A calibrated reference benchtop pH meter (MP 225©,

Mettler Toledo, Leicester) was used to verify these results, and measurements of 5.44, 6.55 and 7.66

were respectively obtained. Subject recruitment (n=30) and analysis request was then undertaken in

the same manner as for the first part of the study.

Although all urine dipstick results were recorded, only the parameters considered to be useful and

reliable in small animal medicine were assessed, namely: bilirubin, blood, glucose, ketones, pH and

protein.24-27

Descriptive statistics were used for analysis of precision and for the self-confidence assessment.

Precision was evaluated by visual analysis of distribution of results obtained for each dipstick

parameter, plotting both methods in perpendicular axes of the same graph. In addition,

nonparametric inferential statistics were employed to analyse the results of the subsequent

session’s pH values accuracy, using the Wilcoxon matched-pairs signed-ranks test. Statistical

significance level was set at P < 0.05. Data introduction, analysis and graphs were undertaken with

commercial software packages (Microsoft® Excel®, version 2016 MSO, ©Microsoft Corporation;

GraphPad InStat, version 3.10, GraphPad Software Inc.; and R, version 3.0.1, ©The R Foundation for

Statistical Computing).

Results

The first part of the study was completed by 61 subjects, comprising 45 veterinary students, 13

veterinary surgeons and three veterinary nurses. The results from the self-confidence assessment

questionnaire showed that 82% (n = 50) of the population were either very confident or totally

confident in performing urine dipstick chemical analysis.

Precision was higher overall using the automated analyser compared to visual reading, as depicted in

Figure 1. When assessing results obtained with the automated reader, glucose was always recorded

for the glycosuric sample, ketones were not detected in any sample, protein was present in all

samples, and pH was precise at 7. Conversely, when assessing results obtained by direct visual

reading, several discrepancies were noted as follows. The glucose containing sample was visually

classified as non-glycosuric by two individuals (one student and one veterinary surgeon). In addition,

ketones were stated to be present in the same sample by three individuals (two students and one

veterinary nurse) and in a non-glucose containing sample by one student. Furthermore, protein was

recorded to be absent in the glucose containing sample by one veterinary surgeon, whereas it was

recorded as present by all other participants. Moreover, two individuals, one student and one

veterinary surgeon, were responsible for variation in two parameters each, with one classifying the

glucose containing sample as non-glycosuric and the plain sample as positive for ketones, and the

other classifying the glucose containing sample as both non-glycosuric and negative for protein.

These individuals recorded that they were either totally or very confident in performing urine

dipstick chemical analysis prior to performing the tests. Finally, notable variation in pH was recorded

visually, with all three initial samples being classified as both acidic and alkaline (Figure 1).

The most common observed technique errors were lack of mixing the urine sample before both

visual and automated analysis (87%, n = 53), not timing readings as per strip manufacturer

instructions when performing visual analysis (52%, n = 32), and not blotting the strip after dipping it

into the urine sample (26%, n = 16). In addition, while undertaking visual analysis and using the

colour chart for guidance depicted in the dipstick strips bottle, five subjects held the strip

horizontally against the chart, instead of vertically, and one subject was noted to be holding the strip

upside down. The same two individuals recording the glucose containing sample as non-glycosuric

also didn’t time their visual readings. Furthermore, one of the subjects recording the glucose

containing sample as positive for ketones also held the strip horizontally against the chart.
For the second part of the study, with varying pH values, 30 subjects participated (19 veterinary

students, seven veterinary surgeons and four veterinary nurses). Precision was better using the

automated analyser compared to visual reading for every sample (Figure 2). Samples 2 and 3 were

visually reported by different individuals as both alkaline and acidic, whereas all results obtained

with the automated reader were recorded as either neutral or alkaline for both these samples.

Finally, considering the benchtop pH meter as gold standard, accuracy was higher for automated

compared to visual analysis for all the samples, with respective automated and visual median values

of; sample 1 with 6 and 6.5 (P < 0.0001), sample 2 with 7 and 7.5 (P = 0.0059), and sample 3 with 7.5

and 8 (P = 0.0003).

Discussion

To the authors’ knowledge, this is the first study to assess the impact of an automated urine dipstick

analyser when used in a small animal hospital with multiple operators. Considerable discordance of

results was seen between visual and automated urine dipstick reading methods in this study, with

improved precision seen with the latter.

Results visually read as non-glycosuric for sample 2, would potentially mean a diagnosis of diabetes

mellitus would be missed, with significant consequences for the patient. Conversely, the visual

ketonuria obtained, may lead to a false diagnosis of ketosis in diabetic patients or to a false suspicion

of proximal renal tubular disease28; 29. In addition, although the measurement of urine protein-to-

creatinine (UPC) ratio is a more objective and reliable assessment for proteinuria30, proteinuria

assessed by urine dipstick might still be used solely in several practices due to practical and financial

constraints. This technique can be of value if results are interpreted in conjunction with specific

gravity, also allowing decision making guidance towards requesting a UPC ratio only in selected

samples.31 Consequently, protein results visually read as negative or trace in this study, could have

led to a decision of not requesting a UPC ratio, when it would be indicated. Lastly, by classifying a
sample as acidic instead of alkaline or vice-versa, the management of urinary tract infections or

urolithiasis could be undermined.

All participants would be expected to be competent to perform dipstick urine analysis. However,

despite a reported high level of confidence in performing urine dipstick analysis, several operator

technique errors were noted. Some of these, e.g. not respecting the advised visual reading times,

which was recorded in over half of the subjects, would be prevented using the automated method.

The discrepancy of results recorded by several individuals might have been explained by their

respective observed technique errors. Nonetheless, is not possible to exclude that these subjects

could have CVD. There is no routine testing for CVD and individuals may be unaware that they suffer

from it. Prospective testing could therefore promote self-awareness, aiming to provide appropriate

counselling and reassurance for future practising clinicians, including acknowledging the existence of

supporting equipment, namely automated analysers.32; 33 It is of note that omitting to mix the

sample prior to performing analysis was a common error, yet this would not explain the observed

variation, as lack of sample mixing also occurred for the automated analysis, nonetheless with much

less variation recorded.

The second part of this study was undertaken to assess if samples with pH other than 7 were still

read more precisely by the automated analyser compared to visual analysis. Although pH meters are

considered more accurate, urine dipstick pH analysis is more commonly used in clinical veterinary

practice due to method availability and expense, despite previously reported suboptimal agreement

results.16; 34; 35 The results obtained in this study served to establish that the precision of the

automated method was superior than visual reading for different pH values. Moreover, when using

the benchtop pH meter as a gold standard, accuracy was higher for automated versus visual

methods, as previously reported.15

It appeared that variation with the visual results was present in all the groups tested i.e. students,

veterinary surgeons and nurses however, further work could assess if increased experience would

reduce variation. Variation was still present using the automated method, but to a lesser degree.
This variation may still have been due to operator methods e.g. differences in dwell time in the

urine, blotting or not blotting the stick prior to insertion in the analyser tray, or may be that the true

value was between two pads.

Some limitations have been identified in this study. The presence of observers in the room, albeit

not immediately next to each subject, though recording their technique, could have led to bias by

introduction of an observer effect. Furthermore, there were no mechanisms in place to ensure that

subjects who had completed the study would not communicate with subjects who hadn’t. If

influencing final results, these two limitations would be thought to possibly lead to improvement of

precision in results obtained with the visual reading method. Nonetheless, precision was still inferior

with this method in comparison with the automated reader. An additional limitation is the fact that

the whole spectrum of possible results was not explored for each analyte. As an example, it is

unknown if a similar spread of variation would have occurred visually testing a sample with only a

small or trace amount of blood, as all the samples in this study contained moderate to large amounts

of blood read by the automated analyser. The initial pooled sample was collected from hospitalised

patients, some of which therefore must have had a degree of microscopic haematuria, which

precluded this assessment. . Moreover, no sample containing ketones was tested so it is unknown if

false negatives could be reported by either method.

In conclusion, this study suggested that when multiple operators are involved, as is the case in

veterinary teaching or large private hospitals, automated urine dipstick reading improves precision

of results. Further research is warranted to understand how this variation may impact clinical

decision making, as well as to understand the role of implemented quality assurance strategies, e.g.

written SOPs and training briefings, in improving precision for visual reading analysis and how would

this compare with results obtained through automated methods.

Acknowledgements
The authors would like to thank the institution’s hospital staff and students for participating in this

study.

References

1. FRY, M. M. (2011) Urinalysis. In Nephrology and Urology of Small Animals. Eds J. BARTGES, D.
J. POLZIN. West Sussex, Wiley-Blackwell. pp 46-57
2. ASSA, S. (1977) Evaluation of urinalysis methods used in 35 Israeli laboratories. Clin Chem
23, 126-128
3. PROPP, D. A., WEBER, D. & CIESLA, M. L. (1989) Reliability of a urine dipstick in emergency
department patients. Ann Emerg Med 18, 560-563
4. SIMPSON, E. & THOMPSON, D. (1978) An assessment of hospital routine urinalysis. Ann Clin
Biochem 15, 241-242
5. SPALDING, J. A., COLE, B. L. & MIR, F. A. (2010) Advice for medical students and practitioners
with colour vision deficiency: a website resource. Clin Exp Optom 93, 39-41
6. BRERETON, D. M., SONTROP, M. E. & FRASER, C. G. (1978) Timing of urinalysis reactions
when reagent strips are used. Clin Chem 24, 1420-1421
7. ROYAL COLLEGE OF VETERINARY SURGEONS (2014) Day One Competences.
https://www.rcvs.org.uk/document-library/day-one-competences/. Accessed November 22, 2017
8. PEELE, J. D., JR., GADSDEN, R. H. & CREWS, R. (1977) Semi-automated vs. visual reading of
urinalysis dipsticks. Clin Chem 23, 2242-2246
9. TIGHE, P. (2000) Urine dry reagent strip "error" rates using different reading methods.
Accreditation and Quality Assurance 5, 488-490
10. WAUGH, J. J., BELL, S. C., KILBY, M. D., BLACKWELL, C. N., SEED, P., SHENNAN, A. H. &
HALLIGAN, A. W. (2005) Optimal bedside urinalysis for the detection of proteinuria in hypertensive
pregnancy: a study of diagnostic accuracy. BJOG 112, 412-417
11. SIEMENS HEALTHCARE DIAGNOSTICS INC. (2014) Clinitek Status+ Analyzer.
https://static.healthcare.siemens.com/siemens_hwem-hwem_ssxa_websites-context-root/wcm/
idc/groups/public/@global/@lab/@poc/documents/download/mda1/ndk2/~edisp/121674-
gc1_clinitek_status_plus_epd_brch_fnl-02442734.pdf Accessed November 22, 2017
12. ALDRIDGE, C., BEHREND, E. N., SMITH, J., WELLES, E. G. & LEE, H. P. (2016) Comparison of
visual and automated interpretation of urinary dipsticks with glucose:creatinine ratio and glucose
concentration. In Proceedings of the 2016 ACVIM Forum. Denver, USA. pp 1488-1488
13. BAUER, N., RETTIG, S. & MORITZ, A. (2008) Evaluation the Clinitek status automated dipstick
analysis device for semiquantitative testing of canine urine. Res Vet Sci 85, 467-472
14. BEHREND, E. N., TAPIA, J., WELLES, E. G. & SUDDETH, J. (2008) Evaluation of a conventional
urine glucose test strip method for detection of glucosuria in dogs and cats. In Proceedings of the
26th Annual ACVIM Forum. San Antonio, USA. pp 790-790
15. DEFONTIS, M., BAUER, N., FAILING, K. & MORITZ, A. (2013) Automated and visual analysis of
commercial urinary dipsticks in dogs, cats and cattle. Res Vet Sci 94, 440-445
16. PAQUIGNON, A., TRAN, G. & PROVOST, J. P. (1993) Evaluation of the Clinitek 200 urinary
test-strip reader in the analysis of dog and rat urines in pre-clinical toxicology studies. Lab Anim 27,
240-246
17. WELLES, E. G., WHATLEY, E. M., HALL, A. S. & WRIGHT, J. C. (2006) Comparison of Multistix
PRO dipsticks with other biochemical assays for determining urine protein (UP), urine creatinine (UC)
and UP:UC ratio in dogs and cats. Vet Clin Pathol 35, 31-36
18. CORREA, M. E., COTE, A. M., DE SILVA, D. A., WANG, L., PACKIANATHAN, P., VON
DADELSZEN, P. & MAGEE, L. A. (2017) Visual or automated dipstick testing for proteinuria in
pregnancy? Pregnancy Hypertens 7, 50-53
19. RUMLEY, A. (2000) Urine dipstick testing: comparison of results obtained by visual reading
and with the Bayer CLINITEK 50. Ann Clin Biochem 37 ( Pt 2), 220-221
20. VAN DELFT, S., GOEDHART, A., SPIGT, M., VAN PINXTEREN, B., DE WIT, N. & HOPSTAKEN, R.
(2016) Prospective, observational study comparing automated and visual point-of-care urinalysis in
general practice. BMJ Open 6, e011230
21. EUROPEAN CONFEDERATION OF LABORATORY MEDICINE (2000) European urinalysis
guidelines. Scand J Clin Lab Invest Suppl 231, 1-86
22. BARNSLEY HOSPITAL (2013) Clinitek Status Urinalysis Training plan for Ward Staff.
http://www.barnsleyhospital.nhs.uk/pathology/files/2014/01/Clinitek-Status-Urinalysis-Training-
Plan.pdf. Accessed November 22, 2017
23. MASSACHUSETTS GENERAL HOSPITAL (2012) Clinitek Status Connect Plus Procedure.
http://www.massgeneral.org/pathology/assets/poct/clinitekstatusconnectltr9242.pdf. Accessed
November 22, 2017
24. DOSSIN, O., GERMAIN, C. & BRAUN, J. P. (2003) Comparison of the techniques of evaluation
of urine dilution/concentration in the dog. J Vet Med A Physiol Pathol Clin Med 50, 322-325
25. KLAUSNER, J. S., OSBORNE, C. A. & STEVENS, J. B. (1976) Clinical evaluation of commercial
reagent strips for detection of significant bacteriuria in dogs and cats. Am J Vet Res 37, 719-722
26. LEVY, M., LESTER, R. & LEVINSKY, N. G. (1968) Renal excretion of urobilinogen in the dog. J
Clin Invest 47, 2117-2124
27. VAIL, D. M., ALLEN, T. A. & WEISER, G. (1986) Applicability of leukocyte esterase test strip in
detection of canine pyuria. J Am Vet Med Assoc 189, 1451-1453
28. THOMPSON, M. F., FLEEMAN, L. M., KESSELL, A. E., STEENHARD, L. A. & FOSTER, S. F. (2013)
Acquired proximal renal tubulopathy in dogs exposed to a common dried chicken treat:
retrospective study of 108 cases (2007-2009). Aust Vet J 91, 368-373
29. LANGLOIS, D. K., SMEDLEY, R. C., SCHALL, W. D. & KRUGER, J. M. (2013) Acquired proximal
renal tubular dysfunction in 9 Labrador Retrievers with copper-associated hepatitis (2006-2012). J
Vet Intern Med 27, 491-499
30. LEES, G. E., BROWN, S. A., ELLIOTT, J., GRAUER, G. E., VADEN, S. L. & AMERICAN COLLEGE OF
VETERINARY INTERNAL, M. (2005) Assessment and management of proteinuria in dogs and cats:
2004 ACVIM Forum Consensus Statement (small animal). J Vet Intern Med 19, 377-385
31. ZATELLI, A., PALTRINIERI, S., NIZI, F., ROURA, X. & ZINI, E. (2010) Evaluation of a urine
dipstick test for confirmation or exclusion of proteinuria in dogs. Am J Vet Res 71, 235-240
32. BALASUNDARAM, R. & REDDY, S. C. (2006) Prevalence Of Colour Vision Deficiency Among
Medical Students And Health Personnel. Malays Fam Physician 1, 52-53
33. LONG, J. A., HONSON, V., KATALINIC, P. & DAIN, S. J. (2015) Re.: Is screening for congenital
colour vision deficiency in school students worthwhile? A review. Clin Exp Optom 98, 192
34. HEUTER, K. J., BUFFINGTON, C. A. & CHEW, D. J. (1998) Agreement between two methods
for measuring urine pH in cats and dogs. J Am Vet Med Assoc 213, 996-998
35. JOHNSON, K. Y., LULICH, J. P. & OSBORNE, C. A. (2007) Evaluation of the reproducibility and
accuracy of pH-determining devices used to measure urine pH in dogs. J Am Vet Med Assoc 230,
364-369
Figure 1 – Precision of visual (x axis) and automated (y axis) urine dipstick analysis results when

performed within a small animal teaching hospital (n = 61), using three different urine samples (1 –

plain, 2 – added glucose solution, 3 – added bovine serum). The circle diameter in each area

represents the number of results obtained for each. Values lying within the 45-degree dashed line

represent the agreement of results between visual and automated methods. Categories from each

parameter are divided according to the reagent strips manufacturer’s criteria. Units are in mg/dL for

glucose, protein and ketone. Categories for blood correspond to the following approximate

concentrations (erythrocytes/µL): 10 for trace, 25 for small, 80 for moderate and 200 for large. For

bilirubin, correspondent values for each category are not available from the manufacturer (reported

sensitivity is 0.4-0.8mg/dL).

Figure 2 – Precision of visual (x axis) and automated (y axis) urine dipstick pH results when

performed within a small animal teaching hospital (n = 30), using urine samples with three different

pH values (1 – 5.44, 2 – 6.55, 3 – 7.66). The circle diameter in each area represents the number of

results obtained for each. Values lying within the 45-degree dashed line represent the agreement of

results between visual and automated methods.