Evaluation of Serum L-phenylalanine Concentration as Indicator of Liver Disease in Dogs: A Pilot Study

Study Animals

In this study, 28 dogs from the Companion Animal Clinic, Institute of Veterinary Medicine, University of Goettingen, with different liver diseases were investigated. The diagnosis of liver disease was based on clinical signs, clinical chemistry values (i.e., alanine aminotransferase activity [ALT], aspartate aminotransferase activity [AST], glutamate dehydrogenase activity [GLDH], alkaline phosphatase activity [ALP], fasting bile acid concentrations, and albumin concentrations), hepatic ultrasonography, and histological examination of a liver biopsy. All dogs in this group showed clinical signs of liver disease, such as lethargy, polyuria, polydipsia, vomiting, or diarrhea. Values >2 times the upper limit of the normal reference range for ALT and AST, and values >1.5 times the upper limit of the normal reference range for GLDH and ALP were considered indicative of liver disease. Values for fasting bile acids and albumin were considered abnormal if they were higher or lower than the limit of the normal reference range.

Comparison groups included 28 clinically healthy dogs (group 2) and 13 dogs with nonhepatic diseases (group 3). The healthy dogs were free of clinical signs for any disease, and all clinical chemistries (i.e., calcium, phosphorus, sodium, potassium, amylase, lipase, creatinine, urea, creatine kinase, total bilirubin, ALT, AST, GLDH, ALP, fasting bile acids, cholesterol, total protein, albumin, glucose) were within the normal ranges. Dogs in group 3 had chronic renal failure (n=1), hypothyroidism (n=1), hypoadrenocorticism (n=1), hyperadrenocorticism (n=2), and diabetes mellitus (n=8). In all these dogs, liver diseases were excluded by clinical biochemical profiles and hepatic ultrasonography. Liver biopsies were performed in both cases of hyperadrenocorticism. The dogs with chronic renal failure, hypothyroidism, hypoadrenocorticism, and diabetes mellitus had normal blood parameters for liver diseases (e.g., ALT, AST, GLDH, ALP, bile acids) and showed no abnormalities on hepatic ultrasonography. Both dogs with hyperadrenocorticism had elevated serum ALT and ALP, but liver biopsies revealed only steroid-induced hepatopathy.

Laboratory Analysis

Following a 12-hour fast, blood samples were collected from the cephalic vein of dogs in all three groups. Shortly after collection, the blood was centrifuged (10 minutes, 3000 g) to obtain serum. Analyses of ALT, AST, GLDH, ALP, fasting bile acids, and albumin were done using an analyzer and commercial kits according to standardized procedures.^a

Serum L-phenylalanine was determined via high-performance liquid chromatography (HPLC) technique.^b L-phenylalanine was measured after reaction with o-phthalaldehyd/3 mercaptopropionic acid. For this procedure, 500 μL of serum was added to a tube with a filter membrane. Hydrochloric acid (50 μL of 1 mol/L) was added, and the tube was vortexed and centrifuged (15 minutes, 750 g). The filtrate was added into a HPLC column.^c The substrate was identified by comparing the chromatogram with a calibration solution. The concentration of the substrate was calculated by comparing the peak of the chromatogram with an internal standard.²⁰ The lower detection limit was 5 μmol/L, and imprecision was determined by intraassay and interassay coefficients of variation of 3.48% and 3.83%, respectively.

Reference Intervals for Serum L-phenylalanine

The reference interval for the L-phenylalanine serum concentration was established using a group of 28 healthy dogs. All dogs were investigated by clinical biochemical analyses and ultrasonography. The 28 dogs had normal values for serum ALT (range 7 to 52 U/L, reference range 0 to 55 U/L), AST (range 0 to 27 U/L, reference range 0 to 25 U/L), GLDH (range 1 to 7 U/L, reference range 0 to 6 U/L), ALP (range 30 to 120 U/L, reference range 0 to 110 U/L), albumin (range 32 to 40 g/L, reference range 25 to 40 g/L), and fasting bile acids (range 4 to 20 μmol/L, reference range 0 to 20 μmol/L). On ultrasonography, the livers of all 28 dogs were normal in size, shape, and echogenicity. The reference interval of L-phenylalanine was determined according to recommendations of the International Federation of Clinical Chemistry (IFCC), which recommends a 95% confidence interval.

Hepatic Biopsy

All dogs in the hepatic disease group had their liver evaluated by ultrasonography.^d Three or four tissue samples were taken via ultrasound-guided biopsy from areas of structural alterations using a Tru-cut biopsy needle.^e In six dogs with liver disease, tissue samples were taken during laparotomy. All tissue samples were fixed in 10% formaldehyde solution and stained with hematoxylin and eosin. Histological diagnoses were established by a certified pathologist. In addition to a histological diagnosis, inflammatory hepatopathies were semi-quantitatively classified as moderate or severe. The hepatitis was classified as moderate if 10% to 50% of the liver cells had pathological changes and moderate infiltration with inflammatory cells was present. If >50% of liver cells had pathological changes and severe inflammation was present, then the hepatitis was classified as severe.²¹

Statistical Analysis

For statistical analyses, three study groups were formed. Group 1 consisted of dogs (n=28) with liver disease; group 2 consisted of clinically healthy dogs (n=28); and group 3 consisted of dogs (n=13) with nonhepatic diseases. The data were checked for normal distribution with the D’Agostino- Pearson test. The serum L-phenylalanine concentrations of all groups were compared by the Mann-Whitney U test.^f The dogs of group 1 were also divided into those with hepatitis and those with liver tumors, and these two subgroups were statistically compared to each other. Dogs with moderate and severe hepatitis were combined and compared with dogs having liver tumors. The upper reference limit for serum L-phenylalanine concentration (87 μmol/L) was used as the cutoff value for the calculations of diagnostic sensitivity, diagnostic specificity, positive and negative predictive values, and positive and negative likelihood ratios. The diagnostic properties of L-phenylalanine concentrations were compared with those of fasting bile acid concentrations (upper limit of 20 μmol/L). Statistical significance was set at a P value <0.05.

Dogs With Liver Disease

The 28 dogs with different liver diseases (group 1) consisted of 16 females (five spayed) and 12 males. Ages ranged from 7 to 17 years (mean±standard deviation [SD] 11±3.2 years). Twenty-one dogs were purebred and seven dogs were mixed breeds [Table 1]. All 28 dogs had one or more clinical signs, including lethargy (n=26), diarrhea (n=19), vomiting (n=14), polyuria/polydipsia (n=8), abdominal pain (n=5), weight loss (n=2), pruritus (n=2), and icterus (n=1).

Abnormalities in serum chemistries were detected in all 28 dogs, including elevations in ALT (n=26, range 50 to 1472 U/L), AST (n=23, range 29 to 617 U/L), GLDH (n=20, range 4 to 1302 U/L), ALP (n=27, range 157 to 4167 U/L), and decreased albumin (n=5, range 12 to 48 g/L). Ten dogs had elevated fasting bile acids [Table 1]. All 28 dogs showed abnormalities on ultrasonography of the liver. The abnormalities were differentiated into diffuse hypoechoic areas (n=6), focal hypoechoic areas (n=8), focal masses (n=4), and mixed hypoechoic and hyperechoic areas (n=10).

The liver diseases were classified according to the literature. ^17–19 Sixteen dogs in the group had hepatitis, with reactive hepatitis being the most frequent diagnosis (n=11). Three dogs had chronic-active hepatitis, and one dog had an acute hepatitis. Twelve dogs had liver neoplasia, with lymphoma being the most frequent. Hepatocellular carcinoma, diagnosed in three cases, was the next common liver tumor [Table 1].

The reference interval of the L-phenylalanine serum concentration in this study was 30 to 87 μmol/L. In group 1, the L-phenylalanine serum concentration was not distributed normally. In groups 2 and 3, the L-phenylalanine serum concentration was distributed normally. Compared with clinically healthy dogs (group 2) and dogs with nonhepatic diseases (group 3), the L-phenylalanine concentration in dogs with hepatopathies was significantly higher (P<0.001 and P<0.01, respectively). Sixteen (57%) dogs of group 1 had inflammatory liver disease, and 12 (43%) dogs had tumors in the liver (11 of which were malignant). The dogs with hepatitis were separated into those with moderate histological changes (n=8) and those with severe histological changes (n=8). Compared with healthy dogs (group 2), dogs with moderate hepatitis had statistically different (P<0.01) L-phenylalanine concentrations. The differences between the dogs with severe hepatitis and the healthy dogs were also significant (P<0.01). Values in dogs with moderate hepatitis were not significantly different when compared to values in dogs with severe hepatitis (P>0.05). The L-phenylalanine concentrations in dogs with tumors, moderate hepatitis, and severe hepatitis were compared to values for healthy dogs. The differences between the dogs with liver tumors and the healthy dogs were significant (P<0.01). Values in dogs with tumors were not significantly different from values in dogs with hepatitis (P>0.05) [Table 2, Figure 1]. These data demonstrated that dogs with different forms of liver disease had significantly higher serum concentrations of L-phenylalanine than healthy dogs and dogs with nonhepatic diseases.

To estimate the value of L-phenylalanine for diagnosing liver disease, serum concentrations were compared to fasting bile acid concentrations in the dogs with liver disease. The diagnostic sensitivities were 0.54 and 0.46 for L-phenylalanine concentration and fasting bile acid concentration, respectively. The diagnostic specificity of L-phenylalanine was 0.96 compared to 0.89 for fasting bile acid. The negative predictive value of L-phenylalanine was 0.68 compared to 0.63 for fasting bile acid. The positive predictive value of L-phenylalanine was 0.94 compared to 0.81 for fasting bile acid [Table 3]. The positive and negative likelihood ratios were also calculated. The positive likelihood ratio of L-phenylalanine in this study was 13.5, and the negative likelihood ratio was 0.48.