Retrospective evaluation of hematological ratios in canine parvovirosis: 401 cases

1 INTRODUCTION

The availability of biomarkers to identify patients at high risk of death could assist in determining duration of hospitalization, severity of disease or predicting outcome.¹ Leukocyte ratios, especially the platelet-to-lymphocyte ratio (PLR) and the neutrophil-to-lymphocyte ratio (NLR), have been extensively studied in human medicine as biomarkers and have proven to be useful prognostic indicators in people with cancer, cardiovascular disease, pneumonia, sepsis and recently in patients with COVID-19.^2-7

In veterinary medicine, NLR is a useful marker of prognosis and disease severity in puppies with acute diarrhea and acute pancreatitis in dogs and cats.⁸ The PLR correlates with duration of ICU stay in dogs with severe hemorrhage, and PLR and NLR correlate with illness severity.⁹ In canine inflammatory bowel disease, PLR and NLR are markers of inflammation and could be useful in predicting treatment response.^10-13 Similarly to human medicine, both white blood cell counts and ratios have predictive and prognostic value in dogs and cats with cancer.^14-20 The NLR is a useful tool for differentiating between dogs with meningoencephalitis of unknown origin from other forebrain diseases.²¹ However, unlike in human medicine, the NLR does not precisely predict outcome in dogs with pneumonia and is not associated with hospitalization duration or death in dogs with septic peritonitis.^{22, 23}

Canine parvovirus (CPV) is a well-recognized cause of illness and death in young dogs because it is a predisposing factor for sepsis and promotes intestinal bacterial translocation and severe immunosuppression.^24-26 Although case fatality rates are low in dogs receiving aggressive inpatient treatment, dogs with CPV often require prolonged hospitalization with high associated costs, so having biomarkers that predict hospitalization times and prognosis would be useful in the context of emotional and financial owner expectations.²⁷ Prognostic biomarkers in CPV are identified but to the authors' knowledge hematological ratios have not been determined.^28-31

The aims of this retrospective study were to measure the NLR, PLR and monocyte-to-lymphocyte ratio (MLR) in dogs with CPV and determine whether these ratios were associated with duration of hospitalization and outcome. Our hypotheses were that the NLR at admission would be lower and PLR higher in non survivors than in survivors, and that the NLR could predict duration of hospitalization.

2 MATERIALS AND METHODS

This was a retrospective observational multicenter study using data from 4 university teaching hospitals and 3 private hospitals between 2016 and 2023. Medical records of these institutions were reviewed in order to identify dogs diagnosed with CPV. To be included in the study dogs had to have a CPV diagnosis with a positive parvovirus fecal antigen test (SNAP Canine parvovirus test kit [IDEXX Laboratories, USA] and Urano test Parvo-Corona [Uranovet, Spain]) and a CBC on the first day of hospitalization. CBCs were performed on 4 different automated hematology analyzers: CBC analyzer ADVIA 2120i, Siemens Healthcare GmbH, Munich, Germany (Veterinary Teaching hospital of the Catholic University of Valencia), Hematology Mindray BC-5000 Vet. Mindray Medical International Limited, Shenzhen, Guangdong (Puchol Veterinary Hospital), Procyte IDEXX laboratories, Hoofddorp, Netherlands (Veterinary Teaching hospital of the Universidad de Extremadura, IVC Evidensia Aúna Especialidades Veterinarias, AniCura Valencia Sur Veterinary Hospital, Veterinary Teaching hospital of the Alfonso X El Sabio University, Veterinary Teaching hospital of the Catholic University of Valencia and Veterinary Teaching hospital of the University of Leon) and Sysmex XN-1000-hematology-Analyzer, Sysmex Spain, Barcelona, Spain (Puchol Veterinary Hospital).

Data collected included signalment (age, sex, and breed), total neutrophil count, total monocyte count, total lymphocyte count, total platelet count, hematological ratios (NLR, MLR, and PLR), duration of hospital stay and outcome. The units used for all WBC and platelet counts were the same (cells 10⁹/L). The NLR was calculated by dividing the total neutrophil count by the total lymphocyte count, the MLR was calculated by dividing the total monocyte count by the total lymphocyte count and the PLR was calculated by dividing the total platelet count by the total lymphocyte count. Since a blood smear was not evaluated in all cases this information was not collected. Dogs were classified as survivors if they survived to discharge, or nonsurvivors if they died from natural causes or were euthanized.

Continuous variables were tested for normality using a Kolmogorov-Smirnov test. It was determined that all values were not normally distributed despite log transformation and were described as medians and 25th and 75th percentiles. A Mann-Whitney U-test was performed to analyze the associations between NLR, MLR, PLR, total leukocyte counts and outcome. A receiver operating characteristic (ROC) curve analysis was used to identify cutoff levels of sensitivity and specificity for the hematological ratios that showed a statistically significant difference between survivors and nonsurvivors. Correlations between hematological ratios and duration of hospitalization were evaluated using Spearman's rank correlation. The statistical analysis was performed using a commercially available statistical software (IBM Statistical Package for the Social Sciences version 27, IBM Corporation United Kingdom Limited, PO Box 41, North Harbour, Portsmouth, Hampshire, PO6 3AU). A P-value <.05 was considered statistically significant.

3 RESULTS

A total of 463 dogs were found in the database with a diagnosis of CPV. Forty-five dogs were excluded because a CBC was not performed at admission time and 17 were excluded for not having a positive fecal antigen test upon admission. Four hundred one dogs fulfilled the inclusion criteria (Veterinary Teaching hospital of the Universidad de Extremadura n = 134, IVC Evidensia Aúna Especialidades Veterinarias n = 73, Veterinary Hospital Puchol n = 58, AniCura Valencia Sur Veterinary Hospital n = 41, Veterinary Teaching hospital of the Catholic University of Valencia n = 38, Veterinary Teaching hospital of the University of Leon n = 38 and Veterinary Teaching hospital of the Alfonso X El Sabio University n = 19) (Figure 1). The study cohort was composed of 183 females (45.6%) and 218 males (54.4%). The median age was 121 (84-210) days. One hundred thirty-three were mixed breed dogs and 268 were purebred. There were 65 different breeds, the most common were German shepherd 19 (4.7%), Labrador retriever 19 (4.7%), Border collie 17 (4.2%) and greyhound 16 (4%). Sixty-five of the 401 dogs died or were euthanized before discharge, representing an in-hospital case fatality rate of 16.2% and constitute the nonsurviving group. Three hundred thirty-six dogs survived to discharge. This group included 4 dogs who were discharged voluntarily because of financial reasons. Median duration of hospitalization was significantly lower in non survivors (3 [2-5] days) than in survivors (6 [4-8] days) (P < .001).

The PLR was significantly higher in nonsurvivors in comparison with survivors (Table 1, Figure 2). The area under the ROC curve for prediction of nonsurvival in dogs with CPV using the PLR was 0.615 (95% CI [0.593-0.691], P = .003) (Figure 3). A cut off of 700 showed a 21.5% sensitivity and 90% specificity for nonsurvival. There were no statistically significant associations between outcome and neither the MLR nor the NLR (Table 1). The total neutrophil count and the total platelet count in nonsurvivors did not differ significantly from survivors. However, the total monocyte and total lymphocyte count differed between groups (Table 1). There were no significant correlations between duration of hospitalization and the different hematological ratios or WBC.

4 DISCUSSION

In this study, dogs with CPV that died had higher PLR than surviving dogs. The results of this preliminary study suggest that the PLR could be used as a prognosis tool in dogs with CPV. However, MLR and NLR showed limited utility.

The PLR, calculated as a simple ratio between the platelet and lymphocyte counts measured in peripheral blood, is a biomarker which reflects the balance between 2 aspects of the immune/inflammatory system: immunomodulatory and inflammatory processes (as indicated by platelet count) and immunity response (lymphocyte count).^{25, 32} However, many other factors that affect lymphocyte count, such as neoplasia or corticosteroid administration, or platelet count, such as traumatic bleeding or disseminated intravascular coagulation, could alter the PLR. Our study has found that dogs with CPV and higher PLRs at admission had higher case fatality rates. PLR might increase, on the other hand, because of lymphocyte depletion.

Lymphopenia is one of the most common hematological abnormalities in CPV.²⁵ However, all the hematological ratios measured in this study include lymphocyte count, only PLR was significantly different between survivors and non survivors so it is likely PLR did not increase in non survivors solely because of lymphopenia. Similarly to our study, an increased PLR has been associated with a higher risk of death in septic humans and dogs, and has been used as a prognostic marker in people with cardiovascular diseases.^{3, 6, 32} Critically ill dogs with nonseptic processes that die have higher median PLR values than those that survive.⁹

Lymphocytes play a central role in modulating the immune response in sepsis because of their capacity to interact with the innate and adaptative immune systems and their ability to coordinate, amplify and attenuate the inflammatory response.³³ The median total lymphocyte count was below reference interval in all dogs and was significantly lower in dogs which died (Table 1). In dogs with CPV, lymphopenia may be attributed to cortisol release and the tropism that the virus has for lymphoid tissue leading to atrophy and destruction of lymphoid tissue.³¹ Total lymphocyte count was not associated with duration of hospitalization. Dogs with CPV and lymphopenia (<1.0 × 10⁶/L) at admission remained hospitalized 1.9 days longer compared to those which did not present this abnormality.³⁴ A lymphocyte count above 1.0 × 10⁶/L at 24 and 48 hours after admission has a positive predictive value of 100% for survival.³¹

In our study the median total neutrophil count at admission was below reference range in nonsurvivors (Table 1). Destruction of cellular precursors and secondary bacterial infections because of intestinal translocation play an important role in neutropenia in CPV infected dogs and is associated with endotoxemia and the development of sepsis.³² However, we found no statistically significant differences in total neutrophil count at admission between survivors and nonsurvivors, and there was no association with hospitalization times either. Although some authors have found a significantly lower total neutrophil count at admission in nonsurvivors,²⁸ in other studies severe neutropenia was not a consistent prognostic indicator in CPV³¹ and was not associated with duration of hospitalization in others.³⁴

We did not observe any significant associations between the NLR and outcome or between NLR and duration of hospital stay. In dogs with chronic enteropathy some studies have shown the utility of the NLR and PLR as markers of inflammation, disease severity and response to treatment.^10-12 It could be that the NLR is not a good biomarker in CPV because of the acute nature of the disease. In people with sepsis a higher NLR often indicates an unfavorable prognosis, and it is associated with a higher risk of neonatal sepsis.^{2-4, 6} However, in veterinary medicine there is still controversy regarding the use of these ratios. Although some studies support an association between NLR and survival and hospitalization time, others did not find that it predicted outcome in septic peritonitis.^{23, 32, 35}

In our study sample, although the MLR was not related to either survival or duration of stay, dogs which died had higher monocytes median total counts. An imbalance between the compensatory anti-inflammatory response syndrome and the systemic inflammatory response syndrome leading to excessive tissue inflammation appreciated often in sepsis could explain this difference.³³ The MLR is not a biomarker widely used in human sepsis, although some authors have suggested that it could be useful in the prognosis of septic dogs.³²

The main limitations of our study are its retrospective nature and the lack of a control group. Although databases are searched by computer, these studies often have a risk of selection bias. Moreover, there are always some unknown confounding factors that remain unrecognized. Although our study is a multicenter study and this can introduce some variation in the treatments prescribed, it also improves its reproducibility and generalizability. No data was collected about reproductive status of the dogs, and this can be a source of variability. Dogs diagnosed with CPV were included regardless of their age and vaccination status. Puppies and unvaccinated dogs might present with a more severe clinical picture and have a worse prognosis because of an immature immune system or because of the absence of antibodies against the virus, respectively.³⁶ A more homogenous cohort of dogs with similar age and the same vaccinate status could have yielded different results. Complete blood counts were performed on various analyzers at the different centers and all WBC were performed on an automated hematology analyzer without a differential WBC performed to confirm the differential leukocyte counts and to determine the presence of platelet aggregates. The blood smear examination was performed based on clinical criteria, so not all cases had a leukocyte count obtained, and the presence of platelet aggregates was ruled out. This is one of the main limitations of this study. Most hematological analyzers can produce overlapping results in cases of the presence of band neutrophils or activated monocytes, as occurs in CPV.

The nonsurvivor group included dogs which died naturally and those which were euthanized because of poor prognosis or financial considerations. In addition, 4 dogs whose owners requested voluntary discharge were included in the study as survivors without determining their outcome. Those dogs could have died after discharge.

In conclusion, in dogs with CPV, higher PLR on admission was associated with a worse prognosis. However, this biomarker cannot be considered as excellent for prediction of outcome in CPV since the AUC of the ROC curve was 0.615. Hematologic ratios should be examined with caution when making prognostic judgments.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION

Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.