Adverse effects of trazodone in dogs on primary hemostasis and electrocardiogram: A single-blinded placebo-controlled crossover study

1 INTRODUCTION

Trazodone hydrochloride is a serotonin antagonist/reuptake inhibitor commonly used in small animals as an adjunctive treatment for behavioral disorders and management of anxiety during hospitalization and after operative confinement.^1-4 Trazodone is an effective sedative in cats.^{5, 6} Adverse effects associated with the oral administration of trazodone in small animals are generally mild. The most frequently reported adverse effects include gagging, nausea, vomiting, diarrhea, ataxia, sedation, increased appetite, painting, paradoxical excitement, and hypersalivation.^{3, 7, 8} In humans, commonly reported adverse effects include sedation, orthostatic hypotension, and priaprism.⁹

Selective serotonin reuptake inhibitors are associated with modifications of hemostasis markers and abnormal bleeding in humans.^10-15 Trazodone and other antidepressants are associated with an increased risk of QT interval prolongation of the electrocardiogram (ECG) in healthy humans and significantly increased arrhythmogenesis.^16-19 Of the adverse effects of trazodone, altered hemostasis and arrhythmogenesis are 2 effects that have the potential to create important dog morbidity. Anecdotal reports by veterinarians in our hospital of excessive bleeding in dogs receiving trazadone questioned whether trazodone could be associated with altered hemostasis.

The aim of the study was to evaluate the effects of standard dosages of trazodone vs placebo on primary hemostasis as well as ECG conduction times in apparently healthy dogs. Our hypothesis was that trazodone at standard therapeutic doses prolongs markers of primary hemostasis compared to placebo in apparently healthy dogs. Additionally, we hypothesized that trazodone at standard therapeutic doses prolongs QT interval in apparently healthy dogs.

2 MATERIALS AND METHODS

The study was a single-blinded randomized placebo-controlled crossover study. Apparently healthy, client-owned dogs between the ages of 1-7 years and with a bodyweight of less than 25 kg were recruited through a volunteer, informed consent basis. A complete medical history including current medications was collected. Dogs were excluded if receiving medications other than parasite preventatives or if a concurrent disease process known to be associated with altered cardiac rhythm, QT interval prolongation, impaired primary or secondary hemostasis was present. Screening for occult diseases, rhythm abnormalities as well as disorders in primary or secondary hemostasis was performed prior to inclusion by performing a complete physical examination, complete cell count (CBC) with blood smear evaluation for platelet aggregates, chemistry profile, urinalysis, prothrombin time (PT), partial prothrombin time (PTT) and a 2 minute 6-lead electrocardiogram (BeeCardia software supported by a Levono tablet; Haifa, Israel). Blood samples (~6 mL) for CBC, chemistry profile, PT/PTT were collected via routine venipuncture using either jugular vein. The 6 mL were often collected from a single attempt, however, some dogs needed additional attempts to collect the entire volume. Blood samples were placed in tubes with EDTA anticoagulant (for CBC) and sodium citrate anticoagulant (for PT/PTT). A urine sample was collected via free catch. The laboratory samples were analyzed in the Clinical Pathology Laboratory at Washington State University. Electrocardiograms were obtained with dogs in right lateral recumbency. All ECGs were digitally stored. Dogs were excluded from the study if there were clinically relevant CBC or biochemical abnormalities, abnormal coagulation variables, evidence of premature ectopy, pathologic AV nodal conduction disorders, or abnormal QT interval on the ECG. The study was approved by the Washington State University Institutional Animal Care and Use Committee (IACUC).

Upon inclusion into the study, dogs were randomly assigned by the pharmacist at the Washington State University Veterinary Teaching Hospital to either the AB or BA group using simple randomization. Dogs in the AB group received trazodone hydrochloride during the first week, followed by a placebo after a 1-week washout period. Dogs in the BA group received the placebo first, followed by trazodone after a 1-week washout period. Standard doses of trazodone (5-7.5 mg/kg) by mouth every 12 hours for 4 doses were administered by the owners using trazodone hydrochloride 50 mg tablets (NDC: 50111-0560-01, Lot#: 9009080, Exp Date: July 2022) before data collection. Given the elimination ½ life of 166 ± 47 minutes, the 4 doses would allow for therapeutic, steady-state plasma levels to be reached.⁸ Moreover, based on this elimination half-life, a 1-week washout period was considered sufficient. For placebo, capsules were formulated with methylcellulose (Lot#: Rx 951 866, Exp Date: June 07, 2021) by a local compounding pharmacy (SIDS Pharmacy, Pullman, Washington). The placebo was administered by the owners, every 12 hours, for 4 doses before data collection. Upon the morning of data collection, owners were instructed to give the treatment either on an empty stomach or with a small amount of food (meatball) and to be consistent for the following treatment administration (empty stomach or with a meatball for both placebo and trazodone). A total of 4 blood samples for primary hemostasis testing were collected during the study period (Day 0: first baseline, Day 2: 2 hours after trazodone or placebo administration, Day 9: baseline after washout period, and Day 11: 2 hours after trazodone or placebo).

Primary hemostasis was evaluated before and after trazodone and placebo administration via buccal mucosal bleeding times (BMBT), platelet count via machine count, platelet aggregation via Plateletworks, and platelet function testing using PFA-100 (Siemens Healthineers USA, Malvern, Pennsylvania, USA). Plateletworks were performed using ADP tubes (Helena Laboratories, Beaumont, Texas, USA), and platelet function testing via PFA-100 was performed using ADP cartridges (Siemens Healthineers USA, Malvern, Pennsylvania, USA). Blood collection and BMBT was performed within 1.5 to 2 hours after the last administration of either trazodone or placebo.

Blood samples (~6 mL) for Plateletworks, platelet count, and PFA-100 were collected via routine left or right jugular venipuncture by using a 6 mL syringe and a 20- or 22-gauge needle, depending upon the dog's size. EDTA and ADP Tubes for Plateletworks as well as citrate tubes (PFA-100) were filled within seconds after collection. The needle was removed from the syringe to fill the tubes in order to prevent sheer stress to the sample.

Plateletworks tubes were inverted 20 times as directed by the package insert and were ran promptly within 5-10 minutes after collection in the Clinical Pathology Laboratory at Washington State University as previously described.²⁰ Cytologic evaluation was performed by a clinical pathology staff member to evaluate for platelet aggregates. For PFA-100 analysis, the samples were inverted 20 times and were allowed to sit for at least 30 minutes before analysis. The samples were run between 30 minutes to 2 hours after phlebotomy. The ADP cartridges were stored at 4°C and allowed to warm to room temperature before usage. Immediately before analysis, the sample tube was inverted 3 times and 800 μL of whole blood were collected for analysis. Samples were run in duplicate and averaged as previously described.²¹ The PFA-100 was run by a single investigator (EB). If a sample error message was recorded, the sample was analyzed a third time. Determination of BMBT was performed by standard method^{21, 22} using a Surgicutt (Jorgensen Laboratories, Loveland, Colorado, USA) on the buccal mucosa avoiding any obvious superficial vessels. The BMBT was performed by a single investigator (EB).

A 2-minute, 6-lead electrocardiogram was performed a total of 3 times during the study period; 1 at Day 0 (baseline), after placebo, and after trazodone administration (Days 2 and 11). A rhythm diagnosis was determined by a single operator (EB) for all 3-time points. Three consecutive beats that had minimal baseline artifact were selected by a single observer (EB). The QT interval of the selected beats and the preceding instantaneous coupling intervals were measured by 3 investigators (EB, RB, ON) at a remote station to determine QT interval and corrected QT interval (cQT) using the van de Waters formula.²³ The investigators were blinded to the dog's treatment and other investigators' measurements. Interobserver variability was calculated.

All statistical analyses were performed with a commercially available statistical software program (Statistical Package for the Social Science, IBM; Microsoft Excel, Microsoft). The variable of closure time (CT) via PFA-100 was used for sample size calculation. Based on previous data²¹ reporting healthy dogs having a mean PFA CT of 67 seconds with a SD of 7.8 seconds and a range of 52-86 seconds. Using our control (baseline and after receiving placebo) with an expected mean of 67 seconds and our experimental group (received trazodone) having a PFA >86 seconds, our Cohen d effect size would be 2.43. Using an alpha level of 0.05 and a power of 0.8, a minimal sample size of 4 dogs was needed. In order to increase the power of the study and given the known inherent variability of coagulation testing, a total number of 15 dogs were included.

The PFA-100 CTs ran in duplicate were averaged. The 2 baseline values for PFA-100 CT, platelet counts, BMBT, and platelet aggregation % via Plateletworks were averaged to provide a single baseline measure. Descriptive statistics and a Shapiro Wilk test were performed to evaluate the distribution of the data. Normally distributed data are presented as sample mean and SD. Nonparametric data are presented as sample median and interquartile ranges. To evaluate for significant differences between variables of primary hemostasis and ECG variables (QT, cQT interval) between baseline data, before and after treatment with standard dosages of trazodone and placebo, a paired t test was performed for parametric data and a Wilcoxon signed-ranked test was performed for nonparametric data. The interobserver variability was evaluated by calculating a coefficient of variation. The coefficient of variation was calculated as follows: (within-subject SD/mean) × 100. A P-value of <.05 was considered statistically significant.

3 RESULTS

A total of 17 dogs were screened. Two dogs were not considered for the study given their poor compliance to venipuncture or BMBT procedure. A total of 15 dogs were included in the study consisting of 12 breeds; 2 Border Collies, 2 Australian Cattle dogs, 2 Pugs, and 1 of each: Australian Shepherd, Wire Fox Terrier, miniature Australian Shepherd, Beagle, Boston Terrier, Dachshund Mix, Chihuahua Mix, Border Collie mix, and Boxer mix. The mean age was 3.46 years (±1.92). The median weight was 13.1 kg (6.55-19.65). Ten dogs were female (8/10 spayed, 2/10 intact) and 5 dogs were male (3/5 neutered, 2/5 intact). Eight dogs were assigned to group AB, and 7 dogs were assigned to group BA. The average dose of trazodone was 6.20 mg/kg (±0.82 mg/kg). One dog became profoundly sedate after trazodone administration, but none of the dogs had to be removed from the study due to observed adverse effects from treatments reported by the owners or noted by the authors. None of the dogs had to be removed due to owner's lack of compliance with medication administration.

Summarized descriptive statistics for obtained ECG QT and cQT intervals, as well as hemostatic variables including platelet counts, BMBTs, PFA-100 CTs, and platelet aggregation % via Plateletworks at baseline and after trazodone and placebo administration, are presented in Table 1. In 3 individual dogs, a platelet aggregation via Plateletworks of <60% (33.4%, 48.7%, 59.7%) was observed during 1 of the baseline measurements which was thought to be lab error. Two of these values were seen during the first week of the study (Day 0) and 1 during the second week of the study (Day 9). Baseline platelet aggregation during the alternative week (either Day 0 or Day 9) resulted in significantly higher platelet aggregation (95.56%, 76.20% and 97.88%), as expected for a healthy dog. As such, the values <60% (n = 3) were excluded and the other baseline values were used for data evaluation.

No significant difference was noted in the platelet count, BMBTs, PFA-100 CTs, and platelet aggregation % between the 2 baseline values. No significant difference was noted in the platelet count, BMBTs, PFA-100 CTs, QT interval, and cQT interval after trazodone administration nor after placebo administration. No significant difference was noted on platelet aggregation after placebo administration. However, a significant decrease in platelet aggregation after trazodone administration was found. Given the high proportion of herding breeds in our study, platelet aggregation via Plateletworks was also evaluated between herding and non-herding breeds. Both groups had a statistically significant reduction in platelet aggregation with a P-value of .018 and .046 for herding and non-herding group, respectively.

No change in the QT interval or ECG rhythm was noted after trazodone or placebo administration. Sinus arrhythmia was the most common rhythm recorded. Mobitz type 1 second degree atrioventricular block was noted in 1 dog (a Pug), during all 3 of the electrocardiograms. The AV block was relatively more frequent after trazodone administration. A left axis deviation supportive of a left anterior fascicular block was observed in 2 dogs (Beagle and Australian Cattle dog), during all 3 of the recorded electrocardiograms.

The interobserver variability for the measured QT interval and heart rate was 2.285% (1.85%) and 0.249% (0.22%), respectively. The maximum coefficient of variation for the measured QT interval and heart rate was 8.86% and 3.84%, respectively.

4 DISCUSSION

This study found that trazodone administration, at standardly recommended dosages, statistically significantly decreases platelet aggregation via Plateletworks in healthy dogs where other markers of hemostasis were unaffected. This finding might be clinically relevant in dogs with disorders of primary hemostasis, after trauma, or unexpected surgical bleeding, as trazodone could be a factor contributing toward the dog's impaired primary hemostasis and worsening of bleeding.

Trazodone is a serotonin 2A receptor antagonist and reuptake inhibitor. Serotonin enhances platelet aggregation in humans and dogs by activation of the serotonin 2A receptors on the platelet surface membrane, thus antagonism of these receptors decreases platelet aggregation.^24-28 Moreover, platelet serotonin is derived predominantly via uptake from the enterochromaffin system and its storage is substantially reduced via uptake inhibition by serotonin reuptake inhibitors.^{13, 29, 30} It is likely that trazodone causes impaired platelet aggregation as a class effect as previously noted with other serotonin 2A antagonists and reuptake inhibitors.^{27, 29, 30}

The reasons for the lack of significant changes seen in the other measured markers of primary hemostasis in this study are unknown. Previous studies in dogs and cats have demonstrated similar findings in that platelet aggregation by Plateletworks was affected or affected to a greater degree by drug treatments compared to the other markers utilized (PFA 100-200, impedance aggregometry).²⁰ Plateletworks using ADP as the agonist might be a more sensitive test to detect the effects of medications on primary hemostasis in dogs.

Different markers of platelet function are based on either the measurement of platelet compounds, in vivo evaluation of primary hemostasis (BMBT), evaluating platelet aggregation (Plateletworks), platelet adhesion under sheer stress (PFA-100), whereas others are based on evaluating global hemostasis (thromboelastography).³¹ As such, particular tests can be considered superior over others when evaluating specific processes of platelet function (platelet aggregation vs adhesion), considering the tests strengths and limitations.³¹ Plateletworks is a rapid test that does not require manipulation of the sample, however, it is an indirect assay that needs to be performed within minutes and platelets do not undergo shear forces.^{31, 32} BMBT is a quick in-vivo test of platelet function, but relatively invasive and poorly standardized whereas PFA-100 is considered an in-vitro, more sensitive, “standarized” BMBT.³¹ Nevertheless, factors such as high hematocrit, fibrinogen and platelet count have been reported to shorten PFA-100 CT, albeit using epinephrine as the agonist.^{33, 34} In human medicine, PFA-100 is characterized by a high negative predictive value and not sensitive to platelet secretion defects.^{31, 35} Considering the mechanism of action of trazodone and theoretic reduction of platelet serotonin storage and decreased secretion upon platelet activation in dogs, this could explain why a lack of change in PFA-100 CTs was observed in our study. Moreover, we administered a relatively conservative dose of trazodone for a brief period. In humans, intraplatelet serotonin concentration were noted to decrease by 83% and 81% following 14 days of paroxitene treatment and 12 weeks of fluoxetine treatment, respectively.^{13, 36} However, sooner evaluation (eg, 2 days after starting treatment) was not evaluated. It is unknown if protracted administration and/or higher dosages of trazodone could lead to more profound effects on platelet aggregation or on other markers of primary hemostasis in dogs.

Based on these findings, the authors recommend exercising caution when administering trazodone to dogs with known or suspected impaired primary hemostasis. Additional research is needed on the combined effects of trazadone and other medications that affect primary hemostasis such as aspirin or clopidogrel before clear recommendations can be made regarding their use.^{37, 38} In dogs with known or suspected primary hemostasis already receiving trazodone, discontinuation of the medication should be considered, especially if no response to therapy is observed or clinical signs worsen. Additionally, caution might also be exercised in certain interventional procedures such as occlusion of patent ductus arteriosus and intrahepatic shunts, as these require normal primary hemostasis for the implanted devices (Amplatz Canine Ductal Occluders, vascular plugs, or endovascular coils) to be effective.

No effect on cardiac rhythm and cQT interval was noted on the ECG after trazodone administration. In humans, trazodone is associated with a dose dependent prolongation of QT interval and ventricular arrhythmias.^39-41 The QT interval prolongation occurs secondary to a dose dependent blockage of the hERG channel current.^{42, 43} Further studies might elucidate QT interval prolongation occurring with higher dosages of trazodone or in dogs receiving concurrent treatment with medications known to prolong QT interval such as class 1a or 3 antiarrhythmics.⁴⁴

There are limitations to this study. First, the medications were administered by the clients. Even though most clients were veterinary students or members of the veterinary community, we cannot be fully certain that the administration of trazodone and placebo was performed at the recommended dose and time intervals. Additionally, as previously mentioned, the dose of trazodone was relatively conservative and administered for a brief period of time.

Plasma concentration of trazodone hydrochloride were not investigated during the study. Given previous studies on trazodone pharmacokinetics and oral bioavailability, we expect to have had achieved therapeutic trazodone plasma concentration levels at the time of data collection (1.5-2 hours) after oral administration.⁸ However, even though our population consisted of apparently healthy dogs, we cannot rule out occult gastrointestinal disease and decreased drug absorption in some of the dogs. Even though this study was designed as a single blinded study, a few dogs were obviously sedated given the nature of the medication, and thus the authors performing data collection might have become “unblinded.”

Almost 50% of our study cohort (7/15) was herding breeds, prone to MDR1 mutation. While evaluating the decrease in platelet aggregation via Plateletworks between herding breeds and non-herding breeds, both groups had a statistically significant reduction in platelet aggregation, indicating the decrease in platelet aggregation is not only in herding breeds. Even though trazodone is not considered a drug that would negatively affect dogs with MDR1 mutation at higher dosages, genetic testing was not performed and it could be considered a limitation.

Venipuncture was performed by several technicians and veterinarians over the course of the study which might increase variability in measured hemostasis variables. In 3 dogs, platelet aggregation measured via Plateletworks resulted in <60% during baseline and these 3 sample values were removed from data analysis given the high likelihood for a spurious result. It is possible that the venipuncture technique by different technicians/veterinarians could have factored in these results.

The study was performed in healthy dogs and not in dogs receiving other drugs capable of affecting primary hemostasis. As such, our results might not be comparable to these dogs. In addition, only 1 agonist (ADP) was used for both Plateletworks and PFA-100 analysis. It is possible that a different effect could have been seen by using a different agonist such as collagen, epinephrine or arachidonic acid.

Heart rhythm was evaluated via 2-minute ECG, which is an insensitive test to detect intermittent arrhythmias when compared to a Holter monitor.⁴⁵

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

Approved by the IACUC at the Washington State University (protocol ASAF #6819).

HUMAN ETHICS APPROVAL DECLARATION

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