Paraoxonase‐1 activity evaluation as a diagnostic and prognostic marker in horses and foals

Abstract Background In several species, paraoxonase‐1 (PON‐1) decreases during inflammation, because of the presence of oxidative stress; its measurement recently has been validated in horses, but its role as a clinical biomarker is unknown. Objectives To evaluate sensitivity, specificity and likelihood ratio of PON‐1 activity to identify systemic inflammatory response syndrome (SIRS)‐positive horses or horses with a poor prognosis. Animals One hundred seventy‐two blood samples from 58 sick horses from 3 different veterinary hospitals. Methods In a cross‐sectional study, PON‐1 activity was measured upon admission and at 24‐hour intervals until discharge or death, and results were analyzed based on SIRS status and outcome. Results No statistically significant difference was found in median PON‐1 activity between SIRS and non‐SIRS cases or between survivors and non‐survivors except for mares, in which PON‐1 activity was significantly lower in SIRS‐positive horses (P = .05). The sensitivity of PON‐1 activity in identifying horses with SIRS or negative outcome was low (0.0%‐46.2% depending on the examined group) but its specificity was high (87.0%‐100.0%). However, when PON‐1 is low, the likelihood of death is 2.40‐3.89 times higher than the likelihood of survival. Repeated measurement of PON‐1 after treatment does not predict outcome. Conclusions and Clinical Importance Evaluation of PON‐1 activity in horses with inflammation might be advisable in the future, but only low activity at admission may be relevant in predicting SIRS or negative outcome.


| INTRODUCTION
Paraoxonase-1 (PON-1) is an enzyme associated with high-density lipoproteins (HDL) that protects low-density lipoproteins (LDL) and HDL from peroxidation. 1 It is mainly synthesized in the liver and transported in the plasma bound to HDL. Serum activity of PON-1 decreases during inflammation. During the acute phase response, in both laboratory animals 2,3 and people, 4

changes in HDL composition
and structure inactivate the enzyme PON-1 and hepatic gene expression of PON-1 is inhibited. 2,5 For these reasons, PON-1 is considered a negative acute phase protein (APP). 2,5 Few studies have investigated changes in PON-1 activity in sick animals (cattle, 6 cats, 7 swine, 8 and dogs 9,10 ). These studies also supported the role of PON-1 as a negative APP in animals, but decreases of PON-1 appear to be relevant only when inflammation is particularly severe and potentially associated with oxidative phenomena. Activity of PON-1 can be evaluated using different substrates, but the paraoxon-based method to measure serum PON-1 activity has been validated only in dogs, 9 cattle, 11 and horses. 12 Recently, the term systemic inflammatory response syndrome (SIRS), rather than endotoxemia, has been suggested to describe the clinical status of endotoxemic horses. Diseases that have been associated with SIRS in adult horses include especially those involving the gastrointestinal tract, such as the inflammatory intestinal diseases and strangulating obstructions. 13 An early diagnosis should be the goal in management of SIRS patients, allowing starting adequate treatment in an early stage. 14,15 Our aim was to compare PON-1 activity in sick horses classified as SIRS-positive or SIRS-negative, as proposed by others, 16 and to investigate the performances of PON-1 activity in terms of sensitivity, specificity, and likelihood ratio in identifying SIRS-positive and negative horses.
Finally, the possible prognostic role of PON-1 activity was assessed by evaluating PON-1 activity in sick horses based on clinical outcome.

| Samples and study design
The study was performed on 172 blood samples from 58 horses (36 mares, 17 geldings, and 5 stallions). Sick animals were referred to 3 different hospitals providing secondary health care.
Horses were classified as sick on the basis of clinical examination and ancillary tests (routine serum biochemistry and hematology, radiographs, ultrasound examination, and cytological and bacteriological evaluation of synovial fluid or bronchoalveolar lavage fluid).
The following data were recorded in order to classify and divide the sick horses in SIRS-positive or SIRS-negative groups 16  Horses with 0 or 1 abnormal criterion were included in the SIRS-negative group, whereas horses with ≥2 abnormal criteria were included in the SIRS-positive group. 16 Retrospectively, sick horses also were divided into survivors and non-survivors. Animals were considered survivors if they were discharged from the hospital, whereas they were considered nonsurvivors if they died or were humanely euthanized because of severe medical prognosis rather than for economic reasons.
Although all the sick horses included in the study were admitted to the hospitals within 30 hours after the onset of the disease, it was not possible to standardize the time elapsed from the onset of clinical signs and the first sampling or to collect a complete history regarding the pre-admission period. After the first sampling (T0), all of the horses received appropriate treatment based on the actual diagnosis and clinical presentation. Blood samples were collected from the jugular vein for the determination of CBC and PON-1 activity using a sterile syringe and 16G needle. Each blood sample was divided in 2 aliquots: a 1-mL aliquot was collected in potassium ethylene diamine tetra-acetic (K 2 EDTA) and analyzed using a cell counter (ProCyte Dx, IDEXX, Westbrook, Maine) within 5 minutes after collection. A second 2.5-mL aliquot was collected in plain tubes and centrifuged at 2100 relative centrifugal force for 10 minutes within 4 hours of collection. The harvested serum was placed in sterile tubes, frozen at −20 C to be transported to the laboratory, where serum PON-1 activity was measured in a single batch as described below.
The study was performed during the regular course of hospitalization of the animals and with the owner's written consent.

| Measurement of serum PON-1 activity
Serum PON-1 activity was measured spectrophotometrically using an automated analyzer (Cobas Mira, Roche diagnostic, Basel, Switzerland), and an enzymatic method already validated in horses. 12 Briefly, 6 μL samples were incubated at 37 C with 89 μL of distilled water and 100 μL of reaction buffer (glycine buffer 0.05 mM, pH 10.5 containing 1 mM of paraoxon-ethyl, purity >90% [Sigma-Aldrich, Saint Louis, Missouri], and 1 mM of CaCl 2 ). The rate of hydrolysis of paraoxon to p-nitrophenol was measured by monitoring the increase in absorbance at 504 nm using a molar extinction coefficient of 18.050 L/mol/cm −1 as previously suggested. 2 The unit of PON-1 activity expressed as U/mL is defined as 1 nmol of p-nitrophenol formed per minute under the assay conditions.

| Statistical analysis
Statistical analysis was performed using an Excel spreadsheet and specific software (Analyse-it, Analyse-it Software Ltd, Leeds, UK).
Results for PON-1 activity upon admission to the hospital obtained from SIRS-positive horses 16 were compared with results from the SIRSnegative horses, using the Mann-Withney U test. The same test was used to compare the results obtained upon admission between survivors versus non-survivors. A Fisher exact test was used to verify the association between categorized PON-1 results (within versus below the RI) and SIRS-positive or negative classification or outcome (survivors versus non-survivors). Statistical differences were set for P < .05.
To assess diagnostic performance of PON-1 activity in detecting SIRS-positive horses, the numbers of samples from SIRS-positive and SIRS-negative subjects that upon admission had PON-1 activity within or below the RI established in the previous study 12 were counted.
Data then were classified as follows: The same classification scheme was used to assess the diagnostic performance of PON-1 activity in detecting animals with a poor prognosis, considering as "positive" the non-survivors and as "negative" the survivors.
In both instances, TP, TN, FP, and FN results were used to calculate sensitivity, specificity and likelihood ratios using standard formulas. 17,18 Results from sequential samplings collected after treatment were not statistically compared to each other because of the low number of samples that had comparable follow-up. Hence, the analysis of results collected during the follow-up was limited to a visual observation of the trend recorded in animals that survived or died despite treatment.
As noted in Figure 1A, horses with PON-1 activity lower than the RI were found only in the SIRS-positive group, that, however, also included several horses with normal PON-1 activity. Conversely, either horses with low PON-1 activity or horses with normal PON-1 activity were found in both groups on the basis of outcome ( Figure 1B). The number of FP, FN, TP, and TN recorded in each group, and specificity, sensitivity or positive likelihood ratio calculated are reported in Table 1. Table 1, low PON-1 activity has an absolute specificity for the diagnosis of SIRS (no FP were found, independent of group and corresponding cutoff). Conversely, PON-1 activity has a variably low sensitivity, because the number of FN (normal PON-1 activity in SIRS-positive horses) was high, especially in male adults.

As shown in
When animals are grouped based on outcome, specificity, although high, is not absolute, and FP results may occur. However, when PON-1 activity is low, the likelihood to have a poor prognosis is 2.40 to 3.89 times higher than the likelihood to have a positive response to treatment.

| DISCUSSION
In several species, the antioxidant enzyme PON-1 decreases in association with oxidative stress that characterizes sepsis. [5][6][7][8][9][10] Our study was designed to assess whether the activity of the antioxidant enzyme PON-1 may serve as a diagnostic or prognostic marker in horses. To this aim, a paraoxon-based method recently validated in horses 12 was used. In other species, decreased PON-1 activity is a negative prognostic marker and may predict outcome. 10,11,20 This does not seem to be true in horses, because the specificity of decreased PON-1 was not absolute, possibly because of the same factors described above. However, comparison of results from horses with positive or negative outcome also may have been biased by the low number of observations, on one hand, and by the wide distribution of results in the non-survivors on the other hand. It is possible that, with increasing numbers of horses with a negative prognosis, these differences would become significant. Additionally, our study failed to provide information on the possible utility of sequential measurement of PON-1 activity after administration of treatment to achieve prognostic information. The number of horses that died and were repeatedly sampled was too low to draw any conclusions, as was the number of horses that had low PON-1 activity upon admission and survived. Some of the horses with normal PON-1 activity at admission that died during follow-up had transient decreases of PON-1 activity below the lower limit of the RI, but this also happened in some of the survivors. In other species, a rapid increase of PON-1 activity was recorded in animals that had abnormal activity at admission and responded to treatment. 9 Therefore, it would be advisable in the future to increase the number of animals, and especially the number of non-survivors with repeated samplings.
Finally, although a previous study 12 demonstrated that the RI varies among horses of different breeds or use, we did not have the opportunity to apply breed-related RIs because the number of horses with low PON-1 activity per breed was too low to allow a reliable statistical comparison. Therefore, it also may be advisable to investigate, by analyzing a larger caseload, whether diagnostic performance could improve by comparing the results of sick horses with breedspecific RIs.
In conclusion, our study suggests that horses with low PON-1 activity likely have SIRS and may have a negative prognosis, whereas

ACKNOWLEDGMENT
The authors thank private veterinarians for collaboration in the collection of samples.