Correlation between urine color and urine specific gravity in dogs: Can urine color be used to identify concentrated urine?

Abstract

This combined retrospective-prospective study evaluated the correlation between canine urine color (UC) and urine specific gravity (USG). Urine color was positively correlated with USG, but the relationship was not significantly improved by use of a UC chart. Urine color as an indicator of USG is limited as 20% of dark-yellow samples had a USG < 1.030.

The assessment of urine specific gravity (USG) by refractometry is a key component of urinalysis. Refractometry provides an estimation of USG, as it indicates how much solute is dissolved in urine and thus reflects its concentration. Urine specific gravity is typically ≥ 1.030 in dehydrated canine patients with normal renal function (1), and can be used to distinguish between pre-renal and renal azotemia.

It is commonly assumed that a deeper yellow color of urine is indicative of more concentrated urine. However, to the authors’ knowledge, there have been no studies in canine patients that have evaluated the correlation between urine color (UC) and USG, and whether UC can be used to estimate USG. If a relationship exists, it may help veterinary personnel or owners assess hydration status in dogs.

The relationship between UC and USG has been extensively evaluated in humans (2,3). Studies have investigated the use of urine color scales in different clinical environments, including the assessment of hydration status in healthy adults (2), nursing home patients (4), healthy children (5), and in acute stroke patients (6), with promising results unless co-morbidities are present (6). There has been an increasing use of color charts to standardize individual interpretation of color in medical literature. A tooth color guide has been developed and verified in dental literature (7), and a UC chart has been produced for naturally occurring urethral obstruction in male cats (8). The color chart used in the latter study was not designed to determine USG, but was instead designed to correlate gross urine color with other diagnostic findings in male cats with naturally occurring urethral obstruction.

We hypothesized that in canine patients UC would be positively correlated with USG and that the use of a UC chart would further increase the correlation between UC and USG. We also hypothesized that UC could be used to estimate USG. Thus, the goals of this study were to determine i) the degree of correlation between UC and USG, ii) if the use of a UC chart would have an effect on the correlation between UC and USG, and iii) whether dark yellow (UC score of 4) could be used to predict whether the USG is ≥ 1.030.

Medical records over an 18-month period were reviewed as part of the retrospective portion of this study. The search identified 1538 cases that met selection criteria. Urine samples from dogs of any breed that were > 1 y old were considered. Because hematuria and bilirubin crystals alter UC, samples with hematuria or bilirubin crystals were excluded from this study. Because the presence of 10 g/L of glucose or 10 g/L of protein in urine increases USG by 0.003 to 0.005 (9), samples with significant (> 1+) proteinuria or glucosuria were excluded from the study. A handheld refractometer (AO Scientific Instruments; Reichert Technologies, Buffalo, New York, USA) was used to assess USG. A total of 4 laboratory technicians performed the urinalyses.

Subsequently, 100 of the 1538 samples were randomly selected for statistical analysis using SAS for Windows 9.4 (SAS Institute, Cary, North Carolina, USA). The randomly selected retrospective subset was used to better match the sample size of the prospective study. These samples were analyzed without the use of a UC chart. However, color descriptors (clear, light yellow, yellow, dark yellow) were recorded, as this is part of a standard urinalysis performed at MSU-CVM.

A UC chart was then produced using canine urine samples submitted to the clinical pathology laboratory at MSU-CVM. Submitted samples were arranged based on visual assessment from clear to dark yellow. Four samples were then selected that were deemed to represent clear, light yellow, yellow, and dark yellow. The samples were then photographed using a Nikon D750 camera under identical lighting conditions against a white background. The photographs taken were then used to produce the UC chart.

The color chart assigned a UC score to each urine sample with a visual assessment of clear urine corresponding to a score of 1, light-yellow corresponding to a score of 2, yellow corresponding to a score of 3, and dark-yellow corresponding to a UC score of 4.

As part of the prospective portion of the study, 95 urine samples were analyzed over a 2-month period by the same 4 laboratory technicians used in the retrospective portion of the study. Although the time from sample collection to urinalysis was not recorded, it was typically 1 h or less after sample collection. Two steps were taken when analyzing the data;

1. The technicians first analyzed the color of the sample using the UC chart. Each sample was then assigned a score of 1 to 4 (Figure 1).

2. A standard urinalysis was performed, including the assessment of USG by refractometry. The same AO Scientific Instruments handheld refractometer (Reichert Instruments) was used to assess USG in both the retrospective and prospective studies.

Figure 1.

Urine color chart produced for this study. A urine color (UC) score of 1 corresponds to “clear,” a UC score of 2 corresponds to “light yellow,” a UC score of 3 corresponds to “yellow,” and a UC score of 4 corresponds to “dark yellow.”

Similar to the retrospective data, any urine samples that contained bilirubin crystals, significant proteinuria (> 1+), glucosuria, or hematuria were excluded from the study.

The correlation between UC and USG was assessed for both the prospective and retrospective data sets separately with Spearman rank correlation using PROC CORR in SAS for Windows 9.4 (SAS Institute). A 2-way analysis of variance (ANOVA) was used to compare the correlation between USG and UC score, with and without the use of a UC chart. The explanatory variables were method (use of a UC chart, or not), color scale (with UC scores of 1, 2, 3, and 4), and the interaction term of these 2 variables. If the interaction was not significant, it was removed from the model. Differences in least squares means with Tukey adjustment for multiple comparisons, between the 4 UC scores (i.e., 1, 2, 3, and 4), were determined for outcomes with significant effects. The 25th, 50th, and 75th percentiles for USG for each UC score, using samples from both the retrospective and prospective study, were determined by using PROC MEANS in SAS for Windows 9.4. Diagnostic plots of the residuals were evaluated to ensure the assumptions of the statistical model had been met. An alpha level of 0.05 was used to determine statistical significance for all methods.

The results revealed that UC and USG were correlated in both the retrospective (r_s = 0.44, P < 0.0001) and prospective data sets (r_s = 0.63, P < 0.0001).

A 2-way ANOVA revealed that the relationship between USG and UC score was not significantly affected by the use of a UC chart (P = 0.5030). There were significant differences in the mean USG at the different scores on the UC scale (P < 0.0001).

All samples with a UC score of 1 had a USG < 1.030. Forty out of 50 samples (80%) with a UC score of 4 had a USG ≥ 1.030 (Table 1). The 25th percentile for samples classified with a UC score of 4 was 1.031. The median USG values for UC scores of 1, 2, and 3 were all < 1.030 (1.013, 1.016, 1.029, respectively); the median USG for a sample with a UC score of 4 was 1.041. This is demonstrated graphically in the form of a box and whisker plot of USG by UC score (Figure 2).

Table 1.

Comparison between samples with a urine specific gravity (USG) < 1.030 or ≥ 1.030 at each urine color (UC) score.

	USG
	< 1.030	≥ 1.030
UC Score
1	100.0% (5)	0.0% (5)
2	84.9% (28)	15.2% (5)
3	59.4% (63)	40.6% (43)
4	20.0% (10)	80.0% (40)

Numbers in parentheses represent the numbers of samples.

100% of samples with a UC score of 1 had a USG < 1.030. Although 80% of the samples with a UC score of 4 had a USG ≥ 1.030, it is important to note that 20% of the samples with a UC score of 4 had a USG < 1.030.

Figure 2.

Box and whisker plot of urine specific gravity by urine color (UC) score for the combined retrospective and prospective data. The box represents the interquartile range from the 25th to 75th percentile, the horizontal bar through the box represents the median, the diamond in the box represents the mean, and the whiskers represent the 5th and 95th percentile values. The circles represent outliers. The median USG for a UC score of 1 and 2 are below 1.030 (1.013, 1.016, respectively). The median USG for a UC score of 3 is 1.029, and the median USG for a UC of 4 is 1.041. 80% of data points with a UC of 4 had a USG ≥ 1.030.

This study is the first investigation into the correlation between UC and USG in dogs. The results of this study support the hypothesis that UC on a numerical scale is positively correlated with USG. The degree of correlation is considered moderate. Although the Spearman rank correlation coefficient for the prospective data (r = 0.63) was greater than that for the retrospective (r = 0.44) data, the lack of significant interaction term indicated that the use of a UC chart did not significantly affect the relationship between USG and the UC score. It is therefore unlikely to be of clinical use in veterinary practice; however, urine color charts may improve assessment for inexperienced users.

The Spearman rank correlation coefficients reported in this study (both retrospective and prospective) were similar to those reported by Kavouras et al (5), who reported an R² = 0.45 in healthy children. However, the Spearman rank correlation coefficient is lower than that reported by Armstrong et al (2) (r = 0.80) who investigated the relationship between UC and USG in adult humans.

The results of the study also agree, in part, with the hypothesis that UC can be used to estimate USG. A UC score of 1 or 2 is unlikely to be concentrated above 1.030; in this study, 100% of samples with a UC score of 1 had a USG < 1.030 and ~85% of samples with a UC score of 2 had a USG < 1.030. Thus, if a dog’s urine is clear (UC score of 1) or light yellow (UC score of 2) and the dog is clinically dehydrated, then there is likely inadequate urine concentrating ability. The data also revealed that 80% of patients with a UC score of 4 had a USG ≥ 1.030, which would indicate adequate urine concentrating ability in a clinically dehydrated patient. However, it is important to also consider that 20% of dogs with a UC score of 4 had a USG < 1.030. Thus, dark yellow urine (UC score of 4) does not always indicate adequate urine concentrating ability.

Caution should be taken when extrapolating the results of this study for use in clinical patients, as this study included only patients with no evidence of abnormalities on urinalysis. This involved the exclusion of samples with bilirubinuria or hematuria, which would make the urine a darker color. Rowat et al (6) showed that in humans, co-morbidities can have a significant negative effect on the relationship between UC and USG. Thus, further research is required to determine the effect of co-morbidities on the correlation between UC and USG in canine patients. CVJ