Introduction

Although copper is essential to a variety of metabolic processes, the ability of this transitional metal to participate in the Haber-Weiss reaction can lead to significant oxidative insults.¹ Once injected or ingested, the majority of copper accumulates in the liver and can contribute to a constellation of pathology, referred to as copper-associated hepatopathy (CuAH).^2,3 The incidence of increased amounts of copper and CuAH in canines has reportedly increased in recent years.² This may be due to breed disposition, excessive dietary copper intake, cholestasis and hepatitis, or a combination of these and other factors.^4–6

Recent work has shown that the ability to detect hepatic copper content in cytologic preparations is greatly increased by specific staining with rhodanine.⁷ Similar to histologic copper (hCu) grade, cytologic copper (cCu) grading of rhodanine-stained hepatic cytology samples correlates well with the gold standard of quantitative copper (qCu) measurement using atomic absorption spectroscopy.^7,8 Using a recently published cytologic copper grading scale, cCu ≥5 correlated with toxic levels of qCu (≥1500 ppm dry weight). Correlation of the cytologic grading scale with a histologic copper grade and histologic findings commonly found in copper-associated hepatopathy has not been evaluated.

This study was designed to investigate the prevalence of increased cCu in routine cytologic samples, identify clinical data associated with increased cCu grade, and evaluate the correlation of cCu grade with hCu grade and other histologic findings

Materials and Methods

To produce a random sampling appropriate for determination of prevalence, the electronic database of Colorado State University was searched for cases of canine hepatic aspirates accessioned between 2011 and 2013. All cases with cytology reports that described low cellularity were excluded. Of the remaining cases, a random number generator was utilized to select 200 cases. The Wright-Giemsa^a–stained slides from each case were retrieved; cases that did not have two highly cellular slides composed of predominantly hepatocytes were further excluded. One Wright-Giemsa–stained slide from each case was rhodanine stained and assigned a cCu grade using a previously described protocol.⁷ A control slide made from fresh liver samples with known quantitative copper levels was included with each rhodanine stain batch. Prevalence of cCu ≥5 was determined from this group.

Because of the low prevalence of cCu ≥5 in the randomly collected group, 35 cases from a previous report were included into the study population, resulting in a total of 198 cases. Clinical data from all 198 cases was collected from the electronic medical record; this included signalment, a biochemistry profile concurrent to hepatic aspiration, and the finding of hepatic nodules on abdominal ultrasound reports. Quantitative copper data and hepatic histologic samples were included if available and within 4 mo after the cytologic sample collection. For the purposes of this study, alanine transaminase (ALT) >180, aspartate transaminase (AST) >90, alkaline phosphatase >700, γ-glutamyl transpeptidase >18 IU/L, or total bilirubin >0.2 mg/dL were considered elevated.

For cases with available archived histologic slides or formalin-fixed tissue, a rhodanine or hematoxylin and eosin–stained histology slide was reviewed by a board-certified veterinary pathologist (B.P.) based on published grading schemes.² Specific metrics obtained for this evaluation included hCu and scores for interface hepatitis, portal inflammation, lobular activity, zonal necrosis, and fibrosis.

A D’Agostino-Pearson test was used to test for normality. A Fisher exact test was used to investigate the differences in proportion of clinical findings between cases with and without increased cCu. Pearson’s correlation coefficient, r, was calculated to evaluate the correlation of cCu and hCu or histologic scores and was interpreted according to published recommendations.⁹ Statistical significance was set at P ≤ .05 for all statistical tests. Statistical analysis was performed using statistical software^b,c,d. To help facilitate communication of the data, principles outlined in the Standards for Reporting of Diagnostic Accuracy were followed, including production of a flowchart of the sample processing and evaluation (Figure 1).¹⁰

View Figure

• Download as Powerpoint
• Download Figure

Results

The 163 cases used to determine prevalence included 88 females, 75 males, and 56 breeds. The additional 35 cases included 19 females and 16 males and added 4 breeds. A list of breeds represented by three or more individuals is included as Table 1.

TABLE 1 Table Listing the Breeds Represented by 3 or More Individuals in the Complete Study Population (All Cases) and the 163 Randomly Collected Members of the Study Population Used to Determine Prevalence

View Fullscreen

Two of 163 randomly collected samples had cCu ≥5, producing a prevalence of 1.23% (95% confidence interval, 0.36–2.09%).

In the complete set of 198 cases, 9 cases had cCu ≥5. Data for clinical findings, stratified by cCu grade and the result of a Fisher exact test, is presented in Table 2. Cases with cCu ≥5 were more likely to have elevated ALT (P = .04) or AST (P = .02). Although none of the cytologic reports indicated a suspicion of copper-associated hepatopathy, cases with cCu ≥5 were more likely to have qCu measurement subsequent to cytologic evaluation (P < .0001). The relationships of the other investigated clinical findings and cCu ≥5 were not statistically significant.

TABLE 2 Observed Clinical Findings and Results of Correlation Analysis of Those Clinical Findings with Cytologic Copper Grade

View Fullscreen

Tissue was available from 37 of 198 cases for evaluation of hCu on a rhodanine-stained slide, whereas a hematoxylin and eosin–stained slide was available for evaluation of histologic scores from 32 of 198 cases. Of the cases with concurrent histologic and cytologic samples, the cCu, hCu, and interface hepatitis data sets were not normally distributed. Zonal necrosis was not present in any of the samples and was not evaluated further. There was a significant moderate positive correlation between cCu and hCu grade, portal inflammation, interface hepatitis, or fibrosis and significant moderate positive correlation between hCu grade and portal inflammation, interface hepatitis, or lobular activity (Table 3, Figure 2).

TABLE 3 Correlation (Pearson’s r), 95% CI of r, and P Value of the Relationship Between the Cytologic Copper Grade or Histologic Copper Grade and Various Histologic Findings in Canine Hepatic Tissue

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Discussion

The observed prevalence of increased canine hepatic cytologic copper content in this random sampling of routine hepatic cytology cases is very low. This would suggest that routine rhodanine staining of all canine hepatic cytology samples would have a low diagnostic yield and cannot be recommended.

Because routine cytologic staining methods have been demonstrated to have very poor sensitivity for detection of hepatic copper content, clinicians and cytologists must rely on clinical findings as an indication to grade hepatic cytology samples for copper. Cases of CuAH often have more profound increases in hepatocellular leakage enzymes (AST, ALT) than enzymatic evidence of cholestasis (alkaline phosphatase, γ-glutamyl transpeptidase).^11,12 The expected correlation between biochemical findings and cCu was found in this study; evidence of hepatocellular leakage is supported as an appropriate indication grading of copper in cytologic samples.

An association between qualitative copper results and cCu grade has been established.⁷ This study found an association between having quantitative copper analysis ordered and an elevated cCu grade. As clinicians are not expected to order qCu testing without a high index of suspicion for elevated copper within the liver, one could use a request to perform this test as an indicator of clinical suspicion. The association between performance of quantitative copper analysis and cCu grade may suggest that clinical suspicion of increased hepatic copper may be an appropriate indication to perform grading of copper in cytologic samples.

Previous work has found that CuAH has been associated with several canine breeds, including Bedlington terriers, Doberman pinschers, Labrador retrievers, Skye terriers, West Highland white terriers, and Dalmatians, among others.^4,8,13–16 Additionally, the presence of nodular hepatic changes on abdominal ultrasound has been associated with CuAH.⁵ A relationship between breed and the appearance of nodular changes on ultrasonography with cCu was not demonstrated in this study; this may be because of the low number of individuals, and further studies of these associations may be appropriate.

Histologic and cytologic copper grades had significant and moderate correlations. It should be recognized that cytologic and histologic samples evaluated in the study were collected at different times and, presumably, are from different areas within the liver. The decreased correlation between histologic and cytologic grades noted in this study, is at least partially, due to the sample collection from different locations within the hepatic lobes and lobules and may be associated with true changes in hepatic copper content. It is interesting to note that most discordant results reflect a failure to detect copper on cytology (cCu = 1) in a patient who had copper present when a larger amount of tissue was subsequently reviewed histologically. Evaluation of histologic samples has documented patchy distribution of copper throughout larger, wedge biopsy-type samples. The smaller sample size associated with fine needle aspiration of a liver may adversely affect the correlation of cCu and hCu grades more significantly than it affects the correlation of cCu and qCu.⁷ The correlation found in this population likely mirrors the clinical setting where less invasive cytologic techniques such as cytology are used prior to more invasive techniques such as biopsy for histopathology or quantitative copper assessment. Evaluation of the correlation of the cCu and hCu should be performed in concurrently collected samples.

The modest correlations between cCu and the various histologic scores are intriguing. Previous work demonstrated a significant correlation between histologic copper grade and interface hepatitis or portal inflammation but did not provide correlation coefficients to allow full evaluation of that relationship.² In this study, both cCu and hCu demonstrated a statistically significant relationship between interface hepatitis and portal inflammation, and the correlation coefficient indicates a relatively similar degree of relationship. In contrast to the previous work, a statistically significant relationship was found between lobular activity and hCu.² This correlation was not present between cCu and lobular activity, but there was a relationship between cCu and fibrosis, which was not present in this data set or the previous study. Although a strong correlation was not found between either cCu or hCu and other histologic findings, and correlation does not imply causation, this data may still be clinically useful. For example, excess copper both causes and is caused by inflammation, and, as shown in this study, the presence of increased copper on a cytologic sample modestly correlates with interface hepatitis and portal inflammation.^3,4,17 Other work has shown a various but poor ability of cytology to detect inflammation in the liver; reported sensitivity is as low as 31%.¹⁸ A known correlation between increased cytologic copper and hepatic inflammation may help bolster a diagnosis of inflammation when there is suspicion of inflammation based on other cytologic findings in the preparation.

Conclusion

The presented data on prevalence and associated clinical findings can be used to help determine when cytologic grading for copper on hepatic cytology samples is warranted.