• Open Access

Prothrombotic and Inflammatory Effects of Intravenous Administration of Human Immunoglobulin G in Dogs


Corresponding author: Ryo Tsuchiya, DVM PhD, Laboratory of Internal Medicine II, School of Veterinary Medicine, Azabu University, 1-17-71 Fuchinobe, Sagamihara-shi, Kanagawa 229-8501, Japan; e-mail: tsutiyar@azabu-u.ac.jp.


Background: Intravenous administration of human immunoglobulin G (hIVIgG) has been suggested to potentiate thromboembolism in dogs, but supportive scientific reports are lacking.

Objectives: To determine if hIVIgG therapy promotes hypercoagulability and inflammation in dogs.

Animals: Twelve healthy Beagle dogs.

Methods: Prospective, experimental trial. An hIVIgG/saline solution was infused IV at 1 g/kg BW over 8 hours to 6 dogs, and physiological saline was infused to the other 6 dogs. Blood samples were drawn before, during, and after infusion for serial measurement of indicators of coagulation and inflammation. Data were analyzed by 2-way repeated measures analysis of variance.

Results: Dogs administered hIVIgG developed mildly decreased blood platelet concentrations without thrombocytopenia (median, 200 × 103/μL; range, 150–302 × 103/μL; P < .01), leukopenia (median, 3.5 × 103/μL; range, 20–62 × 103/μL; P < .001), and mildly increased plasma total protein concentrations (median, 6.3 g/dL; range, 5.6–6.7 g/dL; P < .001). Administration of hIVIgG was also associated with increases in fibrin/fibrinogen degradation products in all dogs (either 5 μg/mL or 10 μg/dL), thrombin-antithrombin III complexes (median, 7.2 ng/mL; range, 4.9–14.2 ng/mL; P < .001), and C-reactive protein concentrations (median, 2.5 mg/dL; range, 0.5–4.3 mg/dL; P < .01).

Conclusion and Clinical Importance: Administration of hIVIgG to dogs promotes hypercoagulability and an inflammatory state. This should be further evaluated and considered when using hIVIgG in dogs with IMHA or other prothrombotic conditions.


intrvaenous human immunoglobulin G


immune-mediated hemolytic anemia


plasma total protein


fibrinogen concentration


prothrombin time


activated partial thromboplastin time


fibrin/fibrinogen degradation products


C-reactive protein


thrombin-antithrombin III complexes

Intravenous administration of immunoglobulin G (IVIgG) has been applied to a wide range of human conditions including primary immunodeficiency diseases, idiopathic thrombocytopenic purpura, and Kawasaki syndrome, an idiopathic immune-mediated vasculitis occurring predominantly in young children.1 In the field of veterinary medicine, human IVIgG (hIVIgG) therapy was introduced in the 1990s for canine immune-mediated diseases such as primary immune-mediated hemolytic anemia (IMHA), immune-mediated thrombocytopenia, and pemphigus.2–7 It has been used in dogs with immune-mediated diseases and poor responses to other immunosuppressive therapy such as glucocorticoids, azathioprine, or cyclophosphamide administration.8

Adverse effects of hIVIgG therapy in humans are well recognized and include headache, fever, chills, and thromboembolism.9,10 There are also reports of thromboembolism developing in dogs with IMHA after treatment with hIVIgG therapy.7 However, the risk of thromboembolism with hIVIgG therapy itself has not been clearly established, in part because of the high incidence of thromboembolism and presence of a prothrombotic state in dogs with IMHA that have not received hIVIgG therapy.11,12

Dogs with IMHA also commonly mount an acute phase inflammatory reaction including increases in serum C-reactive protein (CRP) concentration.13 Although CRP did not predict risk of death or severity of disease in dogs with IMHA, a decreased concentration has been suggested as a marker for therapeutic response.13 If hIVIgG induces inflammation in dogs, CRP may have less utility as a therapeutic response marker in hIVIgG-treated dogs. The purpose of the present study was to determine if hIVIgG promotes hypercoagulability and an acute phase inflammatory response in dogs.

Materials and Methods

Study Design

Twelve clinically healthy Beagle dogs, 3 intact males and 9 intact females, 5–10 years of age and weighing 7–13 kg, were used for the experiments. This number of dogs was determined by power analysis to be sufficient for detecting statistically and biologically important changes in relevant parameters. They were maintained in the Institute of Life Sciences or in the Veterinary Clinical Center, Azabu University, according to the Animal Experiment Guideline of the University. They had no history of venipuncture or treatment with any drug, including nonsteroidal anti-inflammatory agents, for at least 2 weeks before the study. They were allowed free access to water but were fasted for >12 hours before blood collection and were fed after the last sampling was complete on each experiment day.

A 100 mg/mL solution of hIVIgGa in saline was infused IV at 10 mL/kg (1 g/kg of body weight) over 8 hours to 6 dogs (treatment group; 1 male and 5 females) by using an automated infusion pump.b The dose and infusion rate were chosen according to published data6 (1 g/kg over 6–12 hours). Because fibrin/fibrinogen degradation products (FDP), thrombin-antithrombin III complexes (TAT), and CRP concentrations increased after hIVIgG infusion as described below, the hIVIgG product was tested for the presence of these components. This test was done to assess the possibility that increased blood concentrations could have been caused directly by infusion of gram amounts of product containing these analytes, but all were absent or present in negligible quantities. Physiological saline was similarly infused into the other 6 dogs (control group; 2 males and 4 females) at a rate of 10 mL/kg over 8 hours. The rate and duration of infusions were the same in each group. Clinical observation of the dogs was continued throughout the infusion and done again 3 hours after stopping the infusion. Clinical observations included a general physical examination with palpation and thoracic auscultation, as well as measurement of body temperature, arterial pulse, and respiratory rate. Particular attention was paid to evidence of acute immunologic reactions, including shock, because large quantities of foreign protein were infused.

Blood samples were drawn by jugular or cephalic venipuncture before hIVIgG infusion, 2 hours into the infusion, at the end of infusion (8 hours), and 1, 2, 4, and 7 days after the start of each infusion. CBCs were evaluated as a screen for any hematologic abnormalities. Plasma total protein (TP) concentration was measured because blood hyperviscosity induced by increased TP concentration is associated with thromboembolism in human patients receiving hIVIgG.14,15 Hemostasis was assessed by measuring platelet concentration (via CBCs), plasma fibrinogen concentration (Fbg), prothrombin time (PT), activated partial thromboplastin time (APTT), FDP, and TAT. Furthermore, because leukopenia was found in some of the study dogs receiving hIVIgG infusion, CRP was measured to assess for an inflammatory positive acute phase protein response. All parameters except TAT and CRP were measured for treated and control dogs at each sample time. Because of sample volume, time, or cost considerations, TAT and CRP were not measured in the day 4 and day 7 samples. In addition, TAT were not measured in the control dogs' 2-hour and day-2 samples because almost no changes were found even in the control samples at 8 hours, when the treatment group revealed the highest TAT concentrations. Furthermore, because results of our preliminary experiment suggested that the peak CRP increase would occur between 8 hours and 2 days, CRP was not measured in the 2-hour samples.

Sample Collection

Blood was collected into plastic syringes through 21-G winged butterfly needles. Aliquots of blood from the 1st syringe for each dog were transferred to tubes containing ethylenediaminetetraacetic acid for a CBC and to tubes containing thrombin and aprotinin for FDP measurement. A 2nd syringe containing 3.8% trisodium citrate (1 part anticoagulant to 9 parts blood) was attached to the infusion set and used to draw blood for the other analyses. Citrated blood from each dog was centrifuged and aliquots of plasma were stored in microtubes at −20°C for TP and CRP analyses and at −80°C for TAT analysis. Serum for FDP measurement was separated by centrifugation within 30 minutes of collection. The maximum total blood volume drawn at any time was 12.8 mL; volumes varied depending on the parameters measured.

Laboratory Methods

CBCs were done using an automated blood cell analyzer,c and differentiation of 200 leukocytes was done by microscopy of Giemsa-stained blood smears. Plasma TP concentration was measured by the biuret reaction using an automated blood chemical analyzer.d The effects of citrate on measurement of TP concentration in bovine samples has been assessed previously, and differences from serum could be accounted for by dilution.16 The same was confirmed for the present study by the addition of 1/5 volume of citrate to three canine sera and an IVIgG preparation. Data were not corrected for the dilution effect. Assays for Fbg, PT, and APTT were done using an automated coagulometere and reagent kitsf according to the manufacturers' instructions.e Serum FDP concentration was measured using a latex agglutination human FDP kit using a rabbit anti-(human fibrinogen) polyclonal antibody.g Although validation studies for use of this kit in dogs have not been published, other human FDP kits with polyclonal anti-(human fibrinogen or FDP) antibodies appear useful in dogs,17,18 and this FDP kit is widely used for dogs in Japan where its clinical utility was reported in dogs developing malignancy-associated disseminated intravascular coagulation.19 In our preliminary studies with this kit, increasing concentrations of commercially purified canine fibrinogenh yielded increasing values, supporting cross-reactivity of the antibody with canine fibrinogen (data not shown) and therefore likely with canine FDP. FDP concentrations were semiquantitated by applying the positive dilution ratio to the calculation formula described in the manufacturer's technical manual for measuring human FDPs. Although the kit's FDP calculations are based on human samples and may not be accurate for dogs, comparisons between treated and control dogs should reflect differences in FDP concentrations. Samples for TAT measurement were sent to a commercial clinical laboratoryi and measured by a human TAT ELISAj which has been validated for canine samples.20 CRP was measured according to the manufacturer's instructions by using a laser nephelometric immunoassay system for dogs.k The assay validation, reference interval, and clinical usefulness were reported by Onishi et al.21 and domestic marketing of the canine-specific system is permitted by the Japanese Ministry of Agriculture, Forestry, and Fisheries.

Statistical Analysis

All data were analyzed using the same statistical software.1 Normality was confirmed by the Kolmogorov-Smirnovl test. Concentrations of platelets, total leukocytes, TP, Fbg, PT, APTT, TAT, and CRP were analyzed by 2-way repeated measures analysis of variance (ANOVA) to assess for differences between groups, difference over time, and group-time interactions. When significant differences (P < .05) were detected, differences between baseline (before starting the infusion) and each other time point of each group, and differences between control and treatment groups at each time point were analyzed with the Bonferroni test. Two-way ANOVA was performed only for 3 time points in the TAT data because concentrations of TAT were not measured in the control group at other 2 time points (2 hours and day 2). TAT and CRP values below analytical detection limits were considered to be 0. Because FDP data are not continuous, ANOVA was not performed for FDP. Because there were no significant changes in total-leukocyte concentrations in the control group, specific leukocytes were enumerated only for the treatment group; consequently, a 1-way ANOVA and Dunnett's posthoc test were used to analyze specific leukocyte concentrations, including differences between baseline values and each other time.


No clinical signs were observed in any dog during or after hIVIgG or saline infusion. However, there were significant changes over time and significant interactions between treatment group and time for concentrations of total leukocytes, TP, TAT, and CRP. Additionally, there were significant group differences for concentrations of total leukocytes and TATs.

Statistically significant decrements in platelet and leukocyte concentrations occurred with hIVIgG infusion (Fig 1). Platelet concentrations decreased only mildly, but leukopenia was more severe, occurred by 2 hours, and persisted through day 4. Differential leukocyte counts by microscopy revealed significant decreases in neutrophil and lymphocyte concentrations that were severe enough to cause neutropenia and lymphopenia in the treatment group (Fig 2). No significant changes were found in other leukocyte concentrations (data not shown).

Figure 1.

 Box plot (box representing interquartile ranges and whiskers representing minimum to maximum) of platelet concentrations (A) and total leukocyte concentrations (B) in dogs for the 1st 7 days after intravenous administration of human immunoglobulin G (hIVIgG) (1 g/kg; 10 mL/kg) over 8 hours (n=6; black box) or isotonic saline (n=6; white box). Symbols * and ‡ indicate a significant difference from baseline within each group and a significant difference between the groups at each time point, respectively. One symbol (eg, * or ‡) indicates .01 ≤P < .05, duplicate symbols indicate .001 ≤P < .01, and triplicate symbols indicate P < .001 (same in the figures below).

Figure 2.

 Box plot of neutrophil (A) and lymphocyte (B) concentrations in dogs receiving intravenous administration of human immunoglobulin G (hIVIgG) for the 1st 7 days after administration of hIVIgG. Two hundred leukocytes were differentiated by microscopic examination of Giemsa-stained blood smears. See Figure 1 for explanation of symbols.

Plasma TP concentration was significantly increased by an average of approximately 0.6 g/dL at the end of IVIgG infusion (8 hours), both in comparison with baseline and in comparison with the control group (Fig 3). A statistically significant difference between groups was also present at 2 hours. Although not statistically increased, mean TP concentration in the treatment group remained above baseline through day 7, decreasing from an 8 hours maximum of 6.3 to 5.9 g/dL on day 7, a 67% decrease toward baseline.

Figure 3.

 Box plot of plasma total protein concentration in dogs for the 1st 7 days after intravenous administration of human immunoglobulin G. Trisodium citrated plasma was used, and the data were not corrected for dilution. See Figure 1 for explanation of symbols.

Fbg concentration did not show significant changes over time or between groups (data not shown). APTT and PT were not changed in either group (data not shown). Transient and significant increases in FDP and TAT concentrations were present after hIVIgG infusion (Fig 4). All samples of the control group and pre-hIVIgG samples of the treatment group yielded negative results for FDPs (<2.5 μg/mL). In contrast, samples from dogs given hIVIgG had mild to moderate increases in FDP concentrations 2 hours after starting the infusions and through day 2. The peak FDP increase was found at 24 hours. Moderate increases in TAT occurred during the hIVIgG infusion, reaching statistical significance at 8 hours, but the change had disappeared by 24 hours.

Figure 4.

 Dot plot of fibrin/fibrinogen degradation products (A) and box plot of thrombin/antithrombin complex (B) concentration in dogs for the 1st 7 days (A) and for 2 days (B) after intravenous administration of human-immunoglobulin G, respectively. ND, not done. See Figure 1 legend for explanation of symbols.

CRP concentrations increased moderately after hIVIgG (Fig 5). The peak CRP increase was found the day after hIVIgG infusion, at which time there was a significant difference from the control group. In the treatment group, CRP concentrations were significantly increased compared with baseline from 8 hours after IVIgG infusion through day 2. Slight increases were found in the control group after saline infusion, but the changes were not statistically significant.

Figure 5.

 Box plot of plasma C-reactive protein concentration in dogs for the 1st 2 days after intravenous administration of human immunoglobulin G. Fewer time points were assessed based on preliminary studies. See Figure 1 legend for explanation of symbols.


Although headache, fever, and chills are frequent adverse effects of hIVIgG therapy, thromboembolism has been recognized as the most serious complication of this therapy in human patients.9,10 In this study, hIVIgG decreased platelet concentrations in dogs, consistent with a previous study.6 The cause is unknown, but it might relate to accelerated consumption in association with a prothrombotic state, or to accelerated destruction related to the presence of immunoglobulin. The increased FDP and TAT concentrations in this study suggest that hIVIgG somehow stimulates coagulation and fibrinolysis in dogs. Decreased TAT and FDP clearance via competition of IgG for cellular uptake is an unlikely explanation for the increased serum concentrations given that IgG is removed via Fc receptors on leukocytes, while TAT and FDP each bind to their own distinct receptors, many of which are on hepatocytes rather than macrophages.22,23

Hyperviscosity of the circulating plasma due to massive infusion of the plasma protein preparation (hIVIgG) is thought to be a major contributor to thromboembolism in human patients receiving IVIgG therapy.14,15 Approximately a 1.8 g/dL increase in plasma protein concentration has been reported to occur with a 2 g/kg infusion of hIVIgG over 2–5 days, and in that study, hyperproteinemia was a significant predictor of plasma hyperviscosity.15 Others have reported plasma IgG concentrations to increase to 4 times baseline after hIVIgG therapy, and the increase was strongly related to blood viscosity.24 To prevent the abrupt increase of plasma viscosity, slow hIVIgG infusion (<400 mg/kg over 8 hours within one 24-hour period) is recommended in people.9,25

Although plasma TP concentrations increased significantly during hIVIgG infusion in dogs, the increase (approximately 0.6 g/dL) was less than half of that reported with hIVIgG therapy in human patients and less than half of what was expected. Based on a circulating blood volume of 88 mL/kg and a mean hematocrit of 45% in dogs (plasma volume is 55%), the 1 g/kg protein infusions of this study were estimated to yield approximately 2 g/dL increases in TP concentration. In human patients receiving 2 g/kg IVIgG infusion over 24 hours, serum TP concentration increased significantly from pretreatment level (6.32 g/dL) to 8.15 g/dL at the end of infusion, and tended to remain high on day 10 (7.10 g/dL).15 Given the smaller TP increases and apparent faster clearance of hIVIgG in dogs, a mechanism unrelated to hyperviscosity may have promoted coagulation in the dogs receiving hIVIgG. However, the 0.6 g/dL increment may have been enough to promote coagulation, particularly because hIVIgG was infused at >2 times the rate recommended in people, thereby producing a more abrupt change in plasma viscosity.

Leukopenia and an increased plasma CRP concentration in this study suggest an acute inflammatory reaction. Transient neutropenia is reported in human patients receiving IV immunoglobulin, but the mechanism is not clear.26,27 IVIgG-F(ab′)2 and intact IVIgG preparations have been shown to contain antineutrophil antibodies that acted on TNFα-primed neutrophils to trigger an oxidative burst reaction.28 This neutrophil activation was Fc-receptor independent and might have lead to shortened neutrophil survival time. The same process might occur in dogs given hIVIgG, and neutrophil activation might be associated with the CRP response. Another possibility is that immune complexes associated with hIVIgG administration might induce neutropenia by Fc-mediated phagocytosis, despite inhibition of this process by the hIVIgG product. Metalloproteinase-9 released by activated neutrophils plays a key role in the acute activation of coagulation observed during severe heatstroke, and the same mechanism was suggested for patients with sepsis.29 The mechanism of hIVIgG-induced lymphopenia is not clear but could relate to the observation that hIVIgG binds to canine lymphocytes.30

Although the adverse effects of hIVIgG have been considered minimal in dogs, the possibility of thromboembolism caused by hIVIgG therapy has been mentioned previously, especially in the context of IMHA.6,31 The results of this study show that hIVIgG administration promotes hypercoagulability and inflammation, and therefore can increase the risk of thromboembolism. This should be considered before instituting hIVIgG therapy, especially in dogs with prothrombotic states. If CRP is used to assess therapeutic response in IMHA dogs treated with hIVIgG, the effects of hIVIgG on CRP concentrations should be considered. Consideration should also be given to carefully evaluate hemostatic parameters before hIVIgG therapy, and to coadminister antithrombotic therapy such as heparin, antiplatelet drugs such as low-dose aspirin, or both. Further studies are needed to characterize the contributions of hIVIgG to thromboembolism or hypercoagulability in canine patient populations, and to elucidate the mechanisms responsible for the hematologic changes reported herein.


aGammmagard, Baxter, Lessines, Belgium

bTelfusion TE-331, Terumo Company, Tokyo, Japan

cSysmex F-800, Sysmex Corporation, Kobe, Japan

dClinical Chemistry Analyzer Hitachi 9000, Hitachi High-Technologies Corporation, Tokyo, Japan

eSysmex CA-100, Sysmex Corporation

fThromboplastin-C plus, Dade Actin and Dade Thrombin reagents, Siemens Healthcare Diagnostics Inc, IL

gFDPL test, Kyowa Medex Co, Ltd, Tokyo, Japan

hFibrinogen, Fraction 1 from dog plasma, Sigma Chemical Company, St Louis, MO

iMitsubishi Chemical Medience Corporation, Tokyo, Japan

jDiagnost TAT micro, Siemens Healthcare Diagnostics Inc

kCanine CRP measurement system CRP-2, Arrows Co Ltd, Osaka, Japan

lPrism, Version 5.0, Graphpad Software, San Diego, CA


This research was partially supported by The Promotion and Mutual Aid Corporation for Private Schools of Japan, Grant-in-Aid for Matching Fund Subsidy for private universities. The authors thank Dr Madhu P. Sirivelu for assistance with statistical analysis of the data.