This study was orally presented by the first author in the 20th ECVIM-CA Congress, Toulouse, France, September 9–11, 2010.
Serum Acute Phase Proteins as Clinical Phase Indicators and Outcome Predictors in Naturally Occurring Canine Monocytic Ehrlichiosis
Version of Record online: 12 MAY 2011
Copyright © 2011 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 25, Issue 4, pages 811–817, July/August 2011
How to Cite
Mylonakis, M.E., Ceron, J.J., Leontides, L., Siarkou, V.I., Martinez, S., Tvarijonaviciute, A., Koutinas, A.F. and Harrus, S. (2011), Serum Acute Phase Proteins as Clinical Phase Indicators and Outcome Predictors in Naturally Occurring Canine Monocytic Ehrlichiosis. Journal of Veterinary Internal Medicine, 25: 811–817. doi: 10.1111/j.1939-1676.2011.0728.x
- Issue online: 20 JUL 2011
- Version of Record online: 12 MAY 2011
- Submitted October 6, 2010; Revised January 13, 2011; Accepted March 17, 2011.
- Ehrlichia canis;
Background: Canine monocytic ehrlichiosis (CME), caused by Ehrlichia canis, is an important tick-borne disease of global importance. Currently, limited information is available on the diagnostic and prognostic value of acute phase proteins (APPs) in dogs naturally infected with E. canis.
Hypothesis: APPs may be useful indicators of the clinical phase of CME and predictive of the clinical outcome (death or survival).
Animals: Fifty-six dogs naturally infected with E. canis and 7 clinically healthy control dogs.
Methods: C-reactive protein (CRP), serum amyloid A (SAA), haptoglobin (Hp), and albumin concentrations determined on admission were retrospectively compared among 27 dogs with nonmyelosuppressive CME, 29 dogs with myelosuppressive CME and 7 healthy dogs. Diagnosis of CME was based on clinical and clinicopathological findings, seropositivity to E. canis, polymerase chain reaction amplification of E. canis-specific 16S rDNA, microscopic observation of Ehrlichia sp. morulae in blood monocytes or some combination of these.
Results: Mean concentrations of CRP, SAA, and Hp were significantly higher in the myelosuppressed dogs compared with the other groups, but no significant differences were found in the concentration of albumin. Survival analysis of the affected animals indicated that APP concentrations were not associated with clinical outcome; the latter was strongly associated with pancytopenia (odds ratio for death 22.7) and neutropenia (odds ratio for death 7.7).
Conclusions and Clinical Importance: CRP, SAA, and Hp serum concentrations on admission are useful indicators of the clinical phase of CME, but are not useful predictors of clinical outcome.
acute phase proteins
Companion Animal Clinic
canine monocytic ehrlichiosis
polymerase chain reaction
Ehrlichia canis is the major cause of canine monocytic ehrlichiosis (CME) worldwide.1 In experimental CME, the clinical course of the disease can be divided sequentially into acute, subclinical, and chronic phases, although this distinction is not obvious in the naturally occurring disease.1,2 Although complete clinical and clinicopathological recovery can be anticipated after treatment during the acute and subclinical phases, treating a dog with myelosuppressive CME may be cost-prohibitive and eventually ineffective.3,4 Therefore, the identification of indicators that might differentiate between dogs that are likely to respond to treatment would be beneficial. Markedly pancytopenic, leukopenic, or neutropenic dogs at presentation, especially of the German Shepherd breed, have been found to have a poor prognosis.3,5,6
Acute phase proteins (APPs) constitute a component of the acute phase response occurring shortly after tissue injury, most commonly of infectious, immunologic, neoplastic, or traumatic nature.7,8 There is accumulating evidence that these noninvasive markers may be helpful in determining disease severity and possibly the response to treatment or prognosis, on a disease-specific basis.9–12 Limited information is available on the diagnostic or prognostic value of APPs in dogs naturally infected with E. canis. In previous studies, increased C-reactive protein (CRP) and α1-acid glycoprotein concentrations have been demonstrated in dogs experimentally infected with E. canis and their concentrations decreased to preinoculation concentrations soon after the recovery from the acute phase of the disease.13,14 Moreover, in a group of 12 naturally infected dogs, CRP and α1-acid glycoprotein also were found to be increased, indicating that APPs may reflect ongoing tissue damage even after the acute phase of the disease.13 To our knowledge, the potential value of various APPs in discriminating between the different clinical phases of the disease or to predict the clinical outcome in naturally infected dogs has not been investigated to date.
The purpose of the present study was to investigate whether CRP, serum amyloid A (SAA), haptoglobin (Hp), or albumin, measured at the time of admission, may be useful indicators of the clinical phase (myelosuppressive versus nonmyelosuppressive) of natural CME and may be predictive of the clinical outcome (death or survival) of these dogs.
Materials and Methods
Animals and Study Design
Medical records of dogs admitted to the Companion Animal Clinic (CAC) from January 1998 through July 2007 were examined and the medical records of 57 dogs with a diagnosis of CME were retrieved for further analysis. They included purebred (43/56, 76.8%) and mixed breed dogs (13/56, 23.2%), of which 39 (69.6%) were males and 17 (30.4%) females with an age range of 2 months to 10 years (mean, 3.0 years). Thirteen pure breeds were represented, including German Shepherd (n = 24), Collie (n = 3), Rottweiler (n = 3), Brittany Spaniel (n = 2), Saint Bernard (n = 2), Poodle (n = 2), and 1 each of Kurtzhard, West Highland White Terrier, Hellenic Hound, Pomeranian, Cocker Spaniel, Husky, and English Setter. Dogs were eligible to enter the study if the diagnosis of CME was confirmed, there was no evidence of concurrent vector-borne or noninfectious systemic diseases, they had not received antiehrlichial drugs before admission and a serum sample obtained upon admission to the CAC was available. The dogs were allocated into 2 groups. Group NME (n = 27) included dogs with nonmyelosuppressive (likely acute) CME, the majority of which had participated in previously published studies.15,16 Inclusion criteria included clinical and hematological findings compatible with CME (Table 1), seropositivity (indirect fluorescent antibody assay; cut-off titer, 1 : 100, in-office ELISAa or both) to E. canis, positive polymerase chain reaction (PCR) assay for E. canis 16S rDNA in bone marrow (BM) aspirates (23/23 dogs tested), observation of Ehrlichia sp. morulae in buffy coat smears (20/27 dogs, including those not tested by PCR), demonstration of BM normo- or hypercellularity in aspiration cytology smears, rapid clinical recovery after treatment with doxycyclineb or some combination of these. Group ME (n = 29) included dogs with myelosuppressive (likely chronic) CME, the majority of which had participated in previous studies.3,16 Inclusion criteria included clinical and hematological findings compatible with CME (Table 1), seropositivity to E. canis with or without positive PCR (10/17 dogs tested) for E. canis DNA in BM aspirates, profound BM hypoplasia documented by reviewing at least 4 BM aspiration smears and poor response (no clinical or hematological recovery including disease-related death or euthanasia) to doxycycline treatment with or without other supportive measures (eg, blood transfusions, bactericidal antibiotics). Group NME and ME dogs were negative on buffy coat (n = 56) cytology for Babesia sp. and Hepatozoon sp. and negative on lymph node (n = 50) and BM (n = 56) cytology for Leishmania amastigotes. All dogs tested were seronegative for Leishmania sp. (n = 50)c and Dirofilaria immitis (n = 43).d PCR on DNA extracted from BM aspirates performed in the Intracellular Pathogens Laboratory, College of Veterinary Medicine, North Carolina (n = 16 dogs), as described previously,3 or in the Acarus Laboratory (University of Bristol, UK) (n = 25 dogs) failed to amplify Anaplasma phagocytophilum (n = 41), Anaplasma platys (n = 41), Babesia sp. (n = 41), Bartonella sp. (n = 41), Borellia sp. (n = 25), and Rickettsia sp. (n = 16) DNA.
|No. of Dogs with the Abnormality (%)|
|Group NME (n = 27)||Group ME (n = 29)|
|Fever (≥39.5°C)||24 (88.9)||8 (27.6)|
|Hypothermia (<37.8°C)||0||2 (6.9)|
|Depression or lethargy||25 (92.6)||26 (89.7)|
|Anorexia||24 (88.9)||25 (86.2)|
|Weight loss||19 (70.4)||24 (82.8)|
|Lymphadenomegaly||16 (59.3)||9 (31)|
|Ocular discharge||14 (51.9)||5 (17.2)|
|Mucosal pallor||7 (25.9)||27 (93.1)|
|Palpable splenomegaly||4 (14.8)||7 (24.1)|
|Dyspnea||5 (18.5)||7 (24.1)|
|Bleeding tendency||4 (14.8)||29 (100)|
|Tick infestation||13 (48.1)||6 (20.7)|
|HematologyaNo. with abnormality/No. examined (%)|
|Low HCT (<37%)||23/27 (85.2)||29/29 (100)|
|Thrombocytopenia (<200,000 cells/μL)||26/27 (96.3)||29/29 (100)|
|Leukopenia (<6,000 cells/μL)||6/27 (22.2)||26/29 (89.7)|
|Neutropenia (<3,000 cells/μL)||1/26 (3.8)||21/26 (80.8)|
|Neutrophilia (>11,000 cells/μL)||3/26 (11.5)||0/26|
|Lymphopenia (<1,000 cells/μL)||7/26 (26.9)||23/26 (88.5)|
|Monocytosis (>1,350 cells/μL)||3/26 (11.5)||0/26|
|Monocytopenia (<150 cells/μL)||0/26||11/26 (42.3)|
|Pancytopeniab||4/27 (14.8)||26/29 (89.7)|
|BiochemistryaNo. with abnormality/No. examined (%)|
|Hyperproteinemia (>8 g/dL)||1/26 (3.8)||5/29 (17.2)|
|Hypoproteinemia (<6 g/dL)||8/26 (30.8)||7/29 (24.1)|
|High ALP (>124 or 210 U/L)c||11/20 (55)||12/23 (52.2)|
|High ALT (>38 U/L)||8/20 (40)||14/25 (56)|
Serum samples from 7 clinically healthy Beagle dogs, 5 males and 2 females, with an age range of 0.75–1 years (mean: 0.9 years), unremarkable clinicopathological evaluation and negative serology to E. canis, Leishmania infantum, and D. immitis also were included in the study as controls (group C). These dogs participated in a research project in the CAC in 1999 that had official institutional approval, were handled in compliance with the guidelines of the European Committee for Care of Animals for Scientific Purposes, and serum samples were taken after a proper acclimatization period.
Three positive APPs (CRP, SAA, and Hp) and 1 negative APP (albumin) were measured in serum aliquots that had been stored at −25°C for several years. These serum samples were frozen-packaged and delivered to the Laboratory of Clinical Pathology, Faculty of Veterinary Medicine, University of Murcia, Spain, for APPs measurement.
CRP concentration was measured using a immunoturbidometric assaye designed for human samples that showed a correlation of 0.98 with a specific canine ELISA assayf that has been validated for use in dogs.17 A pooled canine serum sample with high concentration of CRP measured by the specific canine ELISA assay was used as standard. Hp and SAA concentrations were measured by commercially available methodsg,h that were validated previously for use in dogs.17 Albumin concentration was determined by a colorimetric assay with a commercially available bromocresol green reagent.i All samples were analyzed in the same batch to avoid the high interrun coefficient of variation reported for the SAA assay.17 All biochemical methods used in this study had intrarun coefficients of variation <10%.
Differences in the distribution of age, breed (purebreds versus mixed breeds), and sex between groups NME and ME were assessed by t-test (age) or Pearson's χ2. The former test also was used to compare the duration of serum storage time between groups NME and ME. Differences in serum APP concentrations among dogs of groups NME, ME, and C were assessed by use of Pearson's χ2 or Fisher's exact test (frequency of increased or decreased concentrations) or by use of ANOVA (mean concentrations). For the latter test, when needed, APP concentrations were log-transformed before analysis to meet the assumption of normality. Where statistical significance was identified, a Tukey's test was used to assess for differences among the groups of dogs. The associations between outcome (death or survival; at least up to the end of the doxycycline treatment), the APPs concentrations, the frequency of selected clinical (fever and bleeding tendency) and laboratory (anemia, leukopenia, neutropenia, thrombocytopenia, and pancytopenia) findings upon admission were evaluated by multivariable logistic regression analysis. Initially, the significance of each of these variables was evaluated in univariable models. Those with P≤ .25 were subjected to a multivariable model. Also, bivariable interactions between the APPs found significant during the univariable screening were created and subjected one-by-one to the multivariable model. The model was reduced, in a stepwise fashion, until the variables were significant at P < .05. The fit of the final model to the data was assessed by the Hosmer-Lemeshow χ2 test. The STATA ver. 9.2 statistical software j was used for all analyses.
No differences were found between group NME and ME dogs regarding breed and sex distribution, but group ME dogs were older (mean age, 4.4 [SD 1.7] versus 1.6 [SD 2.7] years, P < .0001). The median storage time of serum samples was 10 (range, 6–11 years) and 10 (range, 2–11 years) years for groups NME and ME, respectively, whereas all group C aliquots had been stored for 11 years; no difference was found between groups NME and ME (group C samples were not considered for comparison because there was no storage time variability). The frequency of the increases (Hp, CRP, SAA) or decreases (albumin) and the median or mean values of the APPs concentrations in the 3 groups of dogs are summarized in Table 2 and Figures 1 and 2. Overall, between groups NME and ME, Hp tended to increase more frequently in group ME, and the mean concentrations of Hp, CRP, and SAA were significantly higher in the same group. In dogs with CME, Hp more frequently increased in group ME compared with group NME dogs (P= .0036), as opposed to CRP, SAA, and albumin which did not differ significantly in the number of dogs with increased or decreased concentrations in each group. Group NME dogs more frequently demonstrated increased CRP (P= .0016), SAA (P= .0262) and decreased albumin (P < .0001) concentrations compared with group C dogs, but Hp concentrations did not differ significantly. Group ME dogs more frequently developed increased CRP (P= .0003), SAA (P= .00041), and Hp (P= .0365) and decreased albumin (P < .0001) concentrations compared with group C dogs. Mean concentrations of Hp differed among the 3 groups (P= .0005); it was higher in group ME compared with group NME and group C dogs, but there was no difference between group NME and C dogs. The mean concentrations of CRP differed significantly among the 3 groups (P < .0001); it was higher in group ME compared with group NME and group C dogs, and in group NME compared with group C dogs. The mean concentrations of SAA differed significantly among the 3 groups (P= .0003); it was higher in group ME compared with group NME and group C dogs, but there was no difference between group NME and C dogs. The mean concentrations of albumin differed significantly among the 3 groups (P < .0001); it was lower in groups NME and ME compared with group C dogs, but there was no difference between NME and ME groups.
|APPs (Reference Interval)d||Group NME||Group ME||Group C|
|Increased concentrations/dogs examined (%)|
|Hp (<3 mg/L)a,c||6/26 (23)||18/29 (62)||1/7 (14.3)|
|CRP (<20 mg/L)b,c||24/26 (92.3)||28/29 (96.6)||2/7 (28.6)|
|SAA (<5 mg/L)b,c||14/27 (51.9)||22/29 (75.9)||0/7 (0)|
|Decreased concentrations/dogs examined (%)|
|Albumin (>2.4 g/dL)b,c||22/26 (84.6)||26/29 (89.7)||0/7 (0)|
|Hp (mg/L)a,c||1.96 (0.09–8.1)||4.15 (0.58–22.2)||1.63 (0.78–4.17)|
|CRP (mg/L)a,b,c||144.2 (14.4–261)||237.3 (19.8–572.1)||14.4 (5–31.5)|
|SAA (mg/L)a,c||7.14 (0.4–160)e||70.78 (0.84–160)e||1.92(1.33–2.75)|
|Albumin (g/dL)b,c||1.8 (0.8–3)||1.8 (0.9–2.9)||3.0 (2.5–3.2)|
Twenty-three group NME dogs survived (follow-up period ranged from 3 months to 1 year; median, 6 months) and 4 were lost to follow-up because of client noncompliance (although they had experienced clinical recovery during treatment). On the other hand, 18 group ME dogs died (day 1 to 30 postadmission), 4 were euthanized (within 3 days postadmission), 4 dogs survived (follow-up period ranged from 2 to 8 months; median, 5 months) and 3 were lost to follow-up (but demonstrated no clinical improvement during treatment). Of the 4 APPs assessed with the univariable model, only Hp (P= .01), CRP (P= .0012), and SAA (P < .0001) were found to be suitable for the multivariable model; the bivariable interactions between CRP, SAA, and Hp also were subjected to the model. Of the clinical and hematologic variables, all except thrombocytopenia were found to fit the multivariate model (data not shown). None of the APPs evaluated was significant and were all dropped from the final model. In contrast, pancytopenia was found to increase the odds of death 21.8 times (95% CI, 2.2–212.6) and neutropenia increased the odds of death by 7.7 times (95% CI, 1.2–50.9).
This retrospective study examined 2 groups of dogs that largely satisfied the diagnostic criteria of acute and severe chronic CME.18 Because the distinction between the acute and chronic CME is not straightforward in the clinical setting, the terms “nonmyelosuppressive” and “myelosuppressive” were used, which correlate to acute and chronic CME, respectively.16 The results indicate that in naturally occurring CME, several APPs tend to increase (although not with the same frequency or to the same extent) or decrease. In previous studies, increased CRP and α1-acid glycoprotein and decreased albumin concentrations have been documented in experimentally or naturally E. canis-infected dogs.13,14,19 However, this study represents the first time that novel APPs, such as SAA and Hp, were investigated in different clinical stages of CME. Most importantly, exaggerated APP activity was documented in dogs with severe BM suppression, indicating that the so-called “acute phase” proteins are expressed and secreted in stages beyond the acute phase, in chronic or relapsing conditions.8,10,20–22 Therefore, selected APPs such as CRP, SAA, and Hp may be useful in assessing the clinical severity of CME, ideally in association with standard clinicopathological tests (eg, CBC, serum biochemistry, BM examination). Presumably, the reason for the higher concentrations of certain APPs in chronic CME relates to the fact that tissue damage and inflammatory reaction in this phase is more severe compared with the acute phase because of the frequent occurrence of severe anemia, bleeding tendency, and bacterial septicemia.3 In the present study, albumin concentrations did not appear to be of value in staging the disease, as their concentration did not differ between groups NME and ME, a finding that was in accordance with a previous study.19 Unexpectedly, mildly increased Hp and CRP concentrations were found in 1 and 2 healthy dogs, respectively. Despite the fact that every effort was made to recruit healthy dogs in group C, the possibility that some of them had unnoticed subclinical inflammatory conditions cannot entirely be ruled out.
Dogs in the subclinical phase of the disease were not included in this study because previous experimental evidence indicated that CRP and α1-acid glycoprotein return to baseline after the recovery from acute CME.13 Provided that concurrent diseases are ruled out, periodic measurements of APPs after recovery from the acute disease are indicated, because their resurgence may herald the emergence of chronic CME. Of comparative interest, APPs have been evaluated in several canine infectious and noninfectious diseases as measures of disease activity. CRP, but not ceruloplasmin or Hp concentrations, was useful in distinguishing symptomatic from asymptomatic dogs with leishmaniasis.20 Also, CRP and SAA were significantly higher in dogs with babesiosis compared with healthy dogs,23 CRP, SAA, and α1-acid glycoprotein increased significantly in dogs with clinical parvovirus enteritis7,24 and CRP and Hp were significantly increased in dogs with clinical leptospirosis.25 CRP also was increased in dogs with acute pancreatitis26 and systemic inflammatory response syndrome or sepsis,27 whereas CRP and α1-acid glycoprotein correlated well with disease activity in canine autoimmune hemolytic anemia.28
In the present study, we investigated the potential prognostic value of APPs in relation to an array of conventional clinical and hematologic variables that are routinely assessed in everyday practice, including fever, bleeding tendency, anemia, pancytopenia, leukopenia, and neutropenia. When subjected to multivariable analysis, none of the single or paired APPs evaluated was predictive of survival. In contrast, pancytopenia and neutropenia documented on admission were strongly associated with a fatal outcome, as has been established previously.5,6 It appears, therefore, that cross-sectionally measured APPs on admission may have limited prognostic value in dogs with CME over traditional hematologic evaluation. Similarly, APPs measurements (CRP and α1-acid glycoprotein) on admission were not predictive of mortality in canine autoimmune hemolytic anemia.28 This is in contrast to recent work in which APPs on admission were of prognostic value in dogs with parvovirus enteritis.29 Additional studies may be warranted to investigate if serially measured APPs in CME could be more meaningful as outcome predictors. Notably, serial CRP measurements were predictive of outcome in systemic inflammatory response syndrome27 and acute pancreatitis.26
The possibility some of the APPs differences between group NME and ME dogs were because of demographic heterogenicity is rather low. Breed and sex composition of the groups did not differ, and the older age of group ME dogs most likely reflected the fact that myelosuppression usually occurs several months to years after the acute disease.3 Moreover, age-related differences have not been found regarding the APPs response.9 Although APPs seem to be very stable in serum, changes in their concentrations can occur in long-term storage, possibly because of evaporation. However, reliability of long-term storage for APPs has been shown recently in studies of humans, in which CRP showed a high correlation before and after storage for 13.8 years30 and Hp and albumin deterioration was not observed after storage for a period ranging from 8 to 11 years.31 In this study, storage time would not be expected to impact group differences, because there were no significant differences in the duration of storage time among the 3 groups of dogs. In light of the present data, the previous suggestions that stability (eg, approximately 2–3 months for CRP at −10 to −20°C) seems rather conservative.32
In conclusion, results of this study indicate that CRP, SAA, and Hp concentrations in CME are significantly higher in dogs with severe myelosuppression compared with those with uncomplicated CME, but none of the APPs cross-sectionally evaluated in this study were predictive of clinical outcome.
The authors thank Dr E.B. Breitschwerdt (Intracellular Pathogens Laboratory, College of Veterinary Medicine, NC) and Dr S. Shaw (Acarus Laboratory, University of Bristol, UK) for performance of the complementary PCR assays, as specified in the manuscript, in a subset of the dogs participated in this study.
This project was financially supported by and performed in the CAC, Veterinary Faculty, Aristotle University of Thessaloniki, Thessaloniki, Greece and the Department of Animal Medicine and Surgery, Faculty of Veterinary Medicine, University of Murcia, Spain. All authors declare no conflict of interest.
a ImmunoComb, Biogal-Galed, Kibbutz Galed, Israel
b Ronaxan, Merial, Lyon, France
c Snap Leishmania, IDEXX, Westbrook, ME
d Snap Canine Heartworm PF, IDEXX
e CRP OSR 6147 Olympus Life and Material Science Europe GmbH, Lismeehan, O'Callaghan's Mills Co, Clare, Ireland
f Tridelta Phase range canine CRP kit, Tridelta Development Ltd, Brey, Ireland
g Tridelta Phase range haptoglobin kit, Tridelta Development Ltd
h Tridelta Phase range SAA kit, Tridelta Development Ltd
i Albumin OSR 6102 Olympus Life and Material Science Europe GmbH, Lismeehan, O'Callaghan's Mills Co
j StataCorp. 2006, Stata Statistical Software: Release 9.2, Stata Corp LP, College Station, TX
- 3Chronic canine ehrlichiosis (Ehrlichia canis): A retrospective study of 19 natural cases. J Am Anim Hosp Assoc 2004;40:174–184., , , et al.
- 4Myelosuppressive canine monocytic ehrlichiosis (Ehrlichia canis): An update on the pathogenesis, diagnosis and management. Israel J Vet Med 2010;65:129–135., ,
- 18Canine monocytotropic ehrlichiosis and neorickettsiosis (E. canis, E. chaffeensis, E. ruminantium, N. sennetsu, and N. risticii infections). In: GreenCE, ed. Infectious Diseases of the Dog and Cat. St Louis, MO: Saunders Elsevier; 2006:203–216.,
- 21Evaluation of acute phase protein indexes in dogs with leishmaniasis at diagnosis, during and after short-term treatment. Vet Med-Chech 2005;50:39–46.,
- 22Proteins. In: StockhamSL, ScottMA, eds. Fundamentals of Veterinary Clinical Pathology. Ames, IA: Blackwell Publishing; 2008:369–413.,
- 25Clinicopathologic features and outcome predictors of Leptospira interogans Australis serogroup infection in dogs: A retrospective study of 20 cases (2001–2004). J Vet Intern Med 2007;21:3–10., , , et al.