Share this article on:

Association of Selective Serotonin Reuptake Inhibitors with Transfusion in Surgical Patients

Sajan, Farrah MD; Conte, John V. MD; Tamargo, Rafael J. MD; Riley, Lee H. MD; Rock, Peter MD; Faraday, Nauder MD, MPH

doi: 10.1213/ANE.0000000000001319
Cardiovascular Anesthesiology: Research Report

BACKGROUND: The clinical relevance of chronic exposure to selective serotonin reuptake inhibitors (SSRIs) to transfusion in surgical patients is unclear.

METHODS: We conducted a prospective cohort study involving patients undergoing cardiac, vascular, spinal, and intracranial surgery at 2 academic medical centers. Medication use, demographics, comorbidities, and laboratory values were determined at baseline by patient interview and review of medical records. The primary outcome was transfusion of any hemostatic allogeneic blood product (i.e., fresh frozen plasma, platelets, and/or cryoprecipitate) through postoperative day 2.

RESULTS: The study sample consisted of 767 patients; 364 patients (47.5%) underwent cardiac surgery and the remainder underwent noncardiac surgery. Eighty-eight patients (11.5%) used SSRIs preoperatively. Among cardiac patients, the absolute number of allogeneic transfusions was higher for SSRI users than nonusers (2 [0–6] vs 0 [0–2], median [25%–75%], respectively, P = 0.008), and a similar trend was observed for noncardiac surgery. After adjusting for covariates using ordinal logistic regression, preoperative SSRI use was associated with an approximately 2-fold (odds ratio, 2.2; 95% confidence interval, 1.2–3.98) increase in odds of exposure to allogeneic hemostatic blood products; similar results were observed using propensity score adjustment (odds ratio, 1.85; 95% confidence interval, 1.11–3.07). A significant interaction between SSRI use and surgery type, age, sex, or concurrent antiplatelet therapy was not found; however, heterogeneity in magnitude of effect could not be excluded.

CONCLUSIONS: Preoperative use of SSRIs is associated with increased exposure to allogeneic hemostatic blood products in surgical patients at high risk for perioperative bleeding. Determining whether perioperative continuation or withdrawal of SSRIs produces a net clinical benefit requires randomized controlled trials.

Published ahead of print May 6, 2016

From the *Department of Anesthesiology/Critical Care Medicine, Department of Surgery, Division of Cardiac Surgery, Department of Neurosurgery, and §Department of Orthopedic Surgery, Johns Hopkins University School of Medicine (JHUSOM), Baltimore, Maryland; and Department of Anesthesiology, University of Maryland School of Medicine, Baltimore, Maryland.

Accepted for publication March 2, 2016.

Published ahead of print May 6, 2016

Funding: This work was supported by institutional funds from the Departments of Anesthesiology at the Johns Hopkins University School of Medicine and University of Maryland School of Medicine.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Nauder Faraday, MD, MPH, Department of Anesthesiology/Critical Care Medicine, Johns Hopkins University School of Medicine, Johns Hopkins Hospital, Meyer 298, 600 N. Wolfe St., Baltimore, MD 21287. Address e-mail to nfaraday@jhmi.edu.

Selective serotonin reuptake inhibitors (SSRIs) are commonly encountered in the perioperative period because they are widely used to treat a number of highly prevalent diseases, including depression, anxiety, panic disorder, obsessive compulsive disorder, bulimia nervosa, and neuropathic pain.1 SSRIs have the potential to cause an acquired hemostatic defect related to their ability to inhibit platelet function.2–4 Specifically, SSRIs inhibit the serotonin transporter on platelets, causing a decrease in intraplatelet serotonin and diminished feedback activation during dense granule secretion.5 Although an association between SSRIs and bleeding is well established in the medical literature,6–8 the clinical relevance of this association is not as well defined in surgical patients. A number of studies have reported an increase in the risk of perioperative packed red blood cell (PRBC) transfusion in patients receiving chronic SSRI therapy9–12; however, the reported risk is small, and some studies have reported no increase in transfusion risk.13,14 The data available from previous meta-analyses and primary reports suggest that SSRI treatment has no impact on perioperative PRBC transfusion in cardiac surgical patients, a group at particularly high risk for bleeding and perioperative transfusion.12,15

The purpose of this study was to clarify the relation between SSRI use and perioperative transfusion in surgical patients at high risk of bleeding, with particular emphasis on patients undergoing cardiac surgery. Because SSRIs cause an acquired hemostatic defect, we reasoned that the transfusion of blood products intended to improve hemostasis would be a better measure of the effect of SSRIs than transfusion of PRBCs. Furthermore, because the cause of intraoperative bleeding is often unclear and treatment is empiric, we hypothesized that patients exposed to SSRIs preoperatively would have a higher risk of exposure to any allogeneic hemostatic blood product, that is, fresh frozen plasma (FFP), platelets, or cryoprecipitate, within the first 48 hours of surgery than patients not exposed to SSRIs. To test our hypothesis, we conducted a prospective cohort study at 2 academic medical centers in patients undergoing cardiac surgery, vascular surgery, spinal surgery, and intracranial surgery.

Back to Top | Article Outline

METHODS

Study Subjects and Design

We conducted a prospective cohort study of patients undergoing elective surgical procedures at the Johns Hopkins Hospital and University of Maryland Medical Center to identify novel determinants of bleeding and thrombotic outcomes after surgery. At the time of analysis of this study, follow-up data were complete from subjects enrolled from February 1, 2007, through January 1, 2013. Eligible patients included those aged 40 years or older who underwent any of the following elective surgical procedures associated with blood transfusion: cardiac surgery (coronary artery bypass, aortic or mitral valve repair/replacement, or ascending aortic aneurysm repair), vascular surgery (open abdominal aortic aneurysm repair or lower extremity revascularization), multilevel spinal fusion, or craniotomy (for tumor or aneurysm). To limit the selection of patients across procedures to those at highest risk, eligibility criteria included planned postoperative admission to the intensive care unit and/or insertion of intraoperative invasive hemodynamic monitors (arterial catheter or central venous catheter). Exclusion criteria included chronic liver disease, congenital disorder in hemostasis, chronic autoimmune disease, human immunodeficiency virus, chronic and/or perioperative steroids or other immunosuppressive therapy, circulatory arrest as part of planned surgical procedure, circulatory assist device placement, or were unable/unwilling to sign written informed consent. This study was approved by The Johns Hopkins Medical Institutions and University of Maryland Medical Center IRBs, and all subjects signed written informed consent.

Back to Top | Article Outline

Data Collection and Main Exposure

Potential study subjects were first screened for eligibility using the elective operative schedules at the Johns Hopkins Hospital and the University of Maryland Medical Center. In total, 3609 patients were screened and 2079 patients were found to be potentially eligible. Of these, 1071 patients were approached for consent and 767 patients agreed to participate.

Baseline information was obtained through patient interviews and medical chart reviews. The following baseline characteristics were recorded as covariates: age, sex, race, educational level, height, weight, diabetes (defined by medical history or use of oral hypoglycemics or insulin preoperatively), coronary artery disease (defined by medical history of myocardial infarction or ≥50% stenosis of one or more coronary arteries on previous cardiac catheterization), congestive heart failure (defined by medical history or ejection fraction ≤30% on previous echocardiogram or catheterization), chronic pulmonary disease (defined as use of inhaled bronchodilator medications or use of home oxygen therapy), medications, and type of surgical procedure (dichotomized as cardiac and noncardiac surgery). In addition, preoperative laboratory values including hemoglobin, platelet count, and creatinine were collected. Duration of surgery (defined as incision to end of surgery) was also recorded from intraoperative medical records.

Table 1

Table 1

The main exposure was self-reported preoperative use of SSRIs at baseline assessment. This was expressed dichotomously as the presence/absence of exposure. Antidepressants were classified into 3 categories: SSRIs, serotonin norepinephrine reuptake inhibitors (SNRIs), and all other antidepressants excluding SSRIs and SNRIs (Table 1). The SSRI categories were defined to distinguish serotonergic antidepressants by the strength of their inhibitory effect on serotonin reuptake.4

Back to Top | Article Outline

Study Outcomes

The primary outcome was the proportion of patients transfused with any hemostatic allogeneic blood product, that is, FFP, platelets, and/or cryoprecipitate, intraoperatively and postoperatively through day 2. Transfusion of all hemostasis-related blood products was combined into a single outcome because the treatment of nonsurgical bleeding in the operating room is most often empiric, with choice of agent (FFP, platelets, cryoprecipitate) according to preferences of the clinical team, rather than directed by a specific diagnostic test. Transfusion was also reported as the absolute number of allogeneic exposures; for example, 1 unit of FFP equaled 1 exposure, 1 apheresis platelet transfusion equaled 1 exposure. Secondary transfusion outcomes included transfusion of allogeneic PRBCs (each unit equal to 1 exposure) and a composite of all allogeneic exposures, that is, combined total number of allogeneic exposures, including FFP, platelets, cryoprecipitate, and PRBCs. Additional secondary outcomes were time to hospital discharge and mortality at 32 days. All outcomes were determined by the investigative team by review of intraoperative and postoperative medical records. Study participants and/or their designated contact and the participant’s medical providers were contacted by the study team after hospital discharge for additional follow-up. Mortality was verified from death certificates and inpatient death notes.

Back to Top | Article Outline

Statistical Analysis

Continuous variables were described using means ± SDs or median with 25 to 75 percentiles, and categorical variables were described as proportions. Each covariate was evaluated for association with exposure (SSRI use) and the primary outcome (allogeneic hemostatic transfusion) using t tests or χ2 tests, as appropriate.

In the unadjusted crude analysis, χ2 was used to determine the relation of SSRI exposure to the proportion transfused for the whole group and for groups stratified by surgical procedure. The relation between SSRI use and the absolute number of allogeneic exposures was estimated without adjustment using the nonparametric Wilcoxon-Mann-Whitney test.

We used multivariable ordinal logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (95% CIs) for the relation between SSRI use and the primary outcome adjusted for surgical procedure, with and without adjustment for other covariates. Ordinal logistic regression (proportional odds regression) is an extension of logistic regression that is useful when the number of categorical outcomes is >2 and the categories follow in sequence. For our analyses, the primary outcome was described by 3 sequential categories: 0, 1 to 2, and ≥3 allogeneic exposures because of the clinical relevance of increasing categorical levels of exposure. Brant testing verified that the proportional odds assumption was not violated by our models. In additional analyses, we used propensity score instead of individual covariates to adjust ordinal logistic regression models. A propensity score for SSRI exposure was generated for each subject using multivariable logistic regression. Propensity scores were generated using the “pscore” routine as implemented in STATA (StataCorp, College Station, TX). The propensity score was generated from a regression model that included all demographic, comorbidity, laboratory, and medication variables in this study, except for surgical procedure. Surgical procedure was included as a separate covariate in ordinal logistic regression of transfusion outcome. The pscore algorithm stratified subjects by propensity score into 5 well-balanced blocks, which was used as a categorical variable, along with surgical procedure, in the ordinal logistic regression model to evaluate the relation between SSRI exposure and transfusion outcome. We also examined the relation between SSRI exposure and the primary outcome expressed as a simple dichotomous category, no hemostatic allogeneic exposure versus any hemostatic allogeneic exposure, using standard logistic regression. Heterogeneity of effect and interactions were sought in specific subgroups in ordinal logistic regression models. The impact of SSRI use on time to hospital discharge was estimated using the log-rank test; the impact on mortality at 32 days was estimated using the Fisher exact test.

Post hoc power was estimated for regression analyses of the primary and secondary outcomes, that is, the proportion receiving hemostatic allogeneic transfusion and the proportion receiving PRBC transfusion, respectively. Based on the observed rates for these outcomes and our actual sample size, power to detect a significant difference by SSRI exposure was >90% and 50% for hemostatic transfusion and PRBC transfusion, respectively, at 2-tailed α of 0.05. All data were analyzed using STATA 11.

Back to Top | Article Outline

RESULTS

The study sample consisted of 767 patients; the average age of study participants was 62.8 (11.4) years, 42.6% were women, and 10.8% were African American. Three hundred sixty-four patients (47.5%) underwent cardiac surgery, and the remaining 403 (52.5%) underwent noncardiac surgery. In total, 88 study participants (11.5%) reported current SSRI use at the time of the preoperative visit. Characteristics of subjects exposed and not exposed to SSRIs preoperatively are shown in Table 2. Follow-up was complete for all 767 patients. Two hundred twenty-three subjects (29.1%) were transfused allogeneic hemostatic blood products. Characteristics of subjects who were transfused and not transfused allogeneic hemostatic blood products are shown in Table 3.

Table 2

Table 2

Table 3

Table 3

In the unadjusted analysis of all study participants, hemostatic transfusion occurred in 31 of 88 participants (35.2%) taking SSRIs compared with 192 of 679 participants (28.3%) not taking SSRIs (χ2P = 0.177). Because of the substantial difference in the rate of transfusion between cardiac and noncardiac surgery and the association between type of surgical procedure and SSRI exposure, all subsequent analyses were stratified by surgical procedure or included surgical procedure as a covariate. In the unadjusted analysis of the cardiac surgical strata, the proportion receiving hemostatic transfusion was 62.5% for SSRI users and 44.9% for nonusers (P = 0.056). When hemostatic transfusion was expressed as the absolute number of allogeneic exposures, the number of exposures (median, 25%–75%) was significantly greater among cardiac surgical patients taking SSRIs than those not taking SSRIs (2 [0–6] vs 0 [0–2], for SSRI and no SSRI, respectively; P = 0.008) (Table 4). Overall transfusion rates were lower for the noncardiac surgical strata, with hemostatic transfusion rates of 19.6% for SSRI users and 12.4% for nonusers (P = 0.139) (Table 4).

Table 4

Table 4

Ordinal logistic regression was used to determine the relation between preoperative SSRI use and the proportion receiving hemostatic transfusion (0, 1–2, and ≥3 allogeneic exposures). In the multivariable model adjusted for type of surgery only (cardiac versus noncardiac), preoperative SSRI use was associated with an approximately 2-fold increase in odds of exposure to hemostatic transfusion (OR, 1.96; 95% CI, 1.20– 3.20) (Table 5). After multivariable adjustment for all covariates, those treated with SSRIs remained at 2-fold increased risk of exposure to allogeneic hemostatic transfusion (OR, 2.2; 95% CI, 1.2–3.98) (Table 5). Treatment with SNRIs and other antidepressant agents was not associated with the risk of hemostatic transfusion. As expected, the risk of hemostatic transfusion was much greater for cardiac surgery than noncardiac surgery (OR, 7.19; 95% CI, 4.16–12.4), and the odds of exposure to hemostatic transfusion increased with duration of the surgical procedure (Table 5). Other covariates associated with transfusion included older age, smaller body mass index, no history of diabetes, and lower preoperative hemoglobin level and platelet count. Interestingly, the associations of recent exposures to aspirin, nonsteroidal anti-inflammatory drugs, or warfarin to hemostatic transfusion were not significant (Table 5).

Table 5

Table 5

The use of propensity block scores to adjust the ordinal regression model instead of multivariable adjustment for individual covariates yielded very similar results, with approximately 2-fold (OR, 1.85; 95% CI, 1.11–3.07; P = 0.019) increase in risk of hemostatic transfusion associated with preoperative SSRI use. Similarly, when the primary outcome was analyzed as a simple dichotomous variable, preoperative SSRI use was associated with an approximately 2-fold increase in risk of hemostatic transfusion (OR, 2.10; 95% CI, 1.13–3.91; P = 0.019).

Heterogeneity of effect for the relation between SSRI exposure and hemostatic transfusion outcome was evaluated for several subgroups of interest (Table 6). Point estimates for risk were all >1.00, suggesting increased risk of hemostatic transfusion related to SSRI exposure for all subgroups, with risk approximately 2-fold or more for most strata. Although statistically significant interactions were not found by subgroup, heterogeneity in the magnitude of effect for SSRIs by subgroup cannot be excluded, particularly for the interactions of SSRI use with surgical procedure and preoperative aspirin exposure (Table 6).

Table 6

Table 6

In contrast to results with hemostatic transfusion, we did not find a significant association between preoperative SSRI use and perioperative exposure to allogeneic PRBCs (OR, 1.33; 95% CI, 0.80–2.21; P = 0.269). However, preoperative SSRI use was associated with an approximately 2-fold increase in odds (OR, 1.78; 95% CI, 1.09–2.90; P = 0.021) of the composite outcome of all allogeneic exposures (FFP, platelets, cryoprecipitate, and PRBCs).

There was no significant difference in time to hospital discharge between SSRI users and nonusers. There was also no difference in mortality between SSRI users and nonusers.

Back to Top | Article Outline

DISCUSSION

The main finding of this study was that the preoperative use of SSRIs was associated with an increase in odds of transfusion of allogeneic hemostatic blood products in surgical patients at high risk of perioperative bleeding. This risk was present for the surgical cohort as a whole, which included patients undergoing cardiac, vascular, spinal, and intracranial surgery, and was significant for the cardiac subgroup. In the subgroup undergoing noncardiac surgery, which experienced a much lower transfusion rate than in cardiac surgery, the risk associated with SSRIs was lower and not significant; however, trends were similar. To our knowledge, this is the first study to focus on the relation between SSRIs and perioperative transfusion of hemostatic blood products as a primary outcome and the first to demonstrate a clinically relevant effect of SSRIs on transfusion outcomes in cardiac surgery.

Our study included both cardiac and noncardiac surgical patients because previous studies suggested different effects of SSRIs on perioperative bleeding outcomes in cardiac and noncardiac surgery,10–12,14 and we wished to clarify these effects for hemostatic transfusions, which we reasoned would be an outcome more directly related to the impact of SSRIs. We combined all hemostasis-related blood products into a single outcome because the treatment of nonsurgical bleeding in the operating room is most often empiric, with choice of agent (FFP, platelets, cryoprecipitate) at the discretion of the clinical team. Laboratory-based treatment algorithms were not in place at our institutions to correct intraoperative bleeding that was considered nonsurgical. As expected, the rate of transfusion was much higher for cardiac surgery than noncardiac surgery, which necessitated stratified analyses and adjustments for surgical type in regression models. In the unadjusted stratified analyses, we found trends for a relation between SSRI use and the proportion of patients requiring hemostatic transfusion in the cardiac subgroup. When outcome was expressed as the absolute number of allogeneic exposures in unadjusted analyses, SSRI use was associated with a significant increase in exposures in cardiac surgery with a similar trend evident in noncardiac surgery. Regression models, whether adjusted for surgery only, all covariates individually, or propensity score, all yielded the same result: an approximate 2-fold increase in the risk of exposure to allogeneic hemostatic blood products related to preoperative use of SSRIs. These results are consistent with previous reports that suggested SSRIs increase risk of perioperative bleeding outcomes.9–12

The specificity of effect of SSRIs was examined by concurrently evaluating the relation between other antidepressants and hemostatic transfusion outcome. SNRIs, which block the serotonin transporter less effectively than SSRIs,4,16 and other classes of antidepressant drugs did not show any relation with hemostatic transfusion outcome. The relation of covariates to hemostatic transfusion was largely in line with expectation. We found it interesting to note that in contrast to findings with SSRIs, we did not find a significant relation of hemostatic transfusion with preoperative aspirin (within 5 days of surgery) or warfarin use. Point estimates for the risk of hemostatic transfusion related to SSRI exposure were uniformly increased across all subgroups examined. Although we found no statistical evidence for heterogeneity of effect from SSRI exposure, we could not exclude the potential for interactions, particularly between SSRIs and surgical procedure and preoperative aspirin treatment.

In analyses of secondary outcomes, we did not find a significant relation of SSRI use with transfusion of PRBCs; however, our study was underpowered (only 50% power) to detect a small effect that has been reported in prior meta-analyses and administrative database studies.9,12 The OR of 1.33 (95% CI, 0.80–2.21) we observed is very consistent with the modest 10% to 20% increase in risk of PRBC transfusion reported in previous studies. We also did not find a relation of preoperative SSRI exposure with the length of hospital stay or short-term mortality, which has been reported to be increased in surgical patients taking SSRIs.9,12 A link between allogeneic blood exposure and mortality has been established by other investigators.17,18 Increased exposure to allogeneic blood products is a plausible cause for the excess mortality reported from perioperative use of SSRIs; however, our study cannot confirm or refute this hypothesis.

Our study has several limitations, many of which are inherent in the observational nature of its design. Although we considered numerous covariates and analytic methods for the control of those covariates, we cannot exclude the possibility of residual confounding. We did not record the indication for antidepressant medication use and, therefore, could not control for its specific indication in each subject. Of particular concern is the potential for confounding by indication; however, data in Table 2 suggest that patients exposed to SSRIs preoperatively were less likely to undergo the highest risk procedure for transfusion (cardiac) and had fewer comorbidities and medication treatments that might necessitate transfusion than those not taking SSRIs. We cannot exclude the possibility that nonblinded clinicians administered more hemostatic transfusions to SSRI users because they were aware of the potential for bleeding. However, a causal link between preoperative SSRI treatment and surgical bleeding is not currently established and, therefore, unlikely to have influenced clinical decision making. In addition, we were unable to establish a dose-response relationship between preoperative SSRI exposure and hemostatic transfusions because of inadequate data on SSRI dosage, duration of usage, and history of compliance with SSRI medications.

This study included several different surgical procedures with differing clinical practices within and across surgical subgroups that were not standardized, and the resultant heterogeneity creates several important limitations. Cardiac and noncardiac procedures were both included to allow the study to address conflicting findings from previous reports about the effects of SSRIs on transfusion. To account for the unique effects of cardiopulmonary bypass on transfusion, we dichotomized surgical procedure (cardiac versus noncardiac) in regression models and performed analyses stratified by surgical procedure. Point estimates for the risk of transfusion were in the same direction for both cardiac and noncardiac surgery; however, risk estimates were higher, and power was only sufficient to detect a significant effect of SSRIs in the cardiac subgroup. Thus, it is possible that SSRIs only affect transfusion outcome under the altered hemostatic environment of cardiopulmonary bypass; results for noncardiac surgery remain inconclusive at this time. Dexter et al.19 reported the potential for bias in risk estimates when surgical specialty is used in regression models instead of the specific surgical procedure. We attempted to limit heterogeneity among the procedures included in this study by restricting eligibility, a prior, not only by specialty but by complexity, as defined by the scheduled need for postoperative intensive care unit care and/or intraoperative use of invasive hemodynamic monitors. Indeed, the mean (SD) duration of surgery was similar for the cardiac (263 ± 90 minutes), vascular (283 ± 138 minutes), spinal (278 ± 130 minutes), and intracranial (247 ± 96 minutes) cases included in this study. Furthermore, in sensitivity analyses, risk estimates for transfusion related to SSRI exposure and each covariate were similar whether surgical procedure was modeled dichotomously (cardiac and noncardiac) or categorically (cardiac, vascular, spinal, and intracranial) (data not shown). Results from this study can only be considered relevant to complex surgical procedures in which blood transfusion is common, and results need to be confirmed by larger data sets and in specific subsets of surgical patients.

In summary, our data suggest that SSRI use is associated with increased risk of perioperative hemostatic transfusion. Although the true magnitude of clinical risk is not possible to determine from our observational study, after controlling for potential biologic confounders in our regression models, the risk of transfusion associated with SSRIs appeared to be of potential clinical relevance, with magnitude at least as high as other antiplatelet medications (e.g., aspirin, nonsteroidal anti-inflammatory drugs) that are routinely withheld before surgery. Like all medications that incur bleeding risk, the decision to continue or withhold SSRIs perioperatively is largely an individualized risk-benefit analysis. Withholding SSRIs is associated with adverse effects, such as serotonin withdrawal syndrome,20 worsening depression,21 and suicide.22 Furthermore, if withdrawal of SSRIs is considered, a gradual taper is recommended to mitigate the development of discontinuation symptoms.23 These risks must be weighed against the risks of transfusion and associated morbidity for each patient and each specific surgical procedure. Results from our cohort study are insufficient to determine this risk/benefit ratio and should be independently confirmed by other investigators. We recommend caution in implementing a practice of tapering preoperative SSRIs on the basis of this report alone. Randomized controlled trials in high-risk patients are needed to accurately assess the risk/benefit ratio of continued use versus withdrawal of SSRIs in the perioperative period.

Back to Top | Article Outline

DISCLOSURES

Name: Farrah Sajan, MD.

Contribution: This author helped with study design, data analysis and interpretation, and manuscript preparation.

Attestation: Farrah Sajan approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: John V. Conte, MD.

Contribution: This author helped with data acquisition and manuscript preparation.

Attestation: John V. Conte approved the final manuscript.

Name: Rafael J. Tamargo, MD.

Contribution: This author helped with data acquisition and manuscript preparation.

Attestation: Rafael J. Tamargo approved the final manuscript.

Name: Lee H. Riley, MD.

Contribution: This author helped with data acquisition and manuscript preparation.

Attestation: Lee H. Riley approved the final manuscript.

Name: Peter Rock, MD.

Contribution: This author helped with study design, data acquisition and interpretation, and manuscript preparation.

Attestation: Peter Rock approved the final manuscript.

Name: Nauder Faraday, MD, MPH.

Contribution: This author helped with study design, data analysis and interpretation, and manuscript preparation.

Attestation: Nauder Faraday approved the final manuscript, attests to the integrity of the original data and the analysis reported in this manuscript, and is the archival author.

This manuscript was handled by: Steven L. Shafer, MD.

Back to Top | Article Outline

REFERENCES

1. Masand PS, Gupta SSelective serotonin-reuptake inhibitors: an update.Harv Rev Psychiatry1999769–84
2. Ross SB, Aperia B, Beck-Friis J, Jansa S, Wetterberg L, Aberg AInhibition of 5-hydroxytryptamine uptake in human platelets by antidepressant agents in vivo.Psychopharmacology (Berl)1980671–7
3. Hergovich N, Aigner M, Eichler HG, Entlicher J, Drucker C, Jilma BParoxetine decreases platelet serotonin storage and platelet function in human beings.Clin Pharmacol Ther200068435–42
4. Meijer WE, Heerdink ER, Nolen WA, Herings RM, Leufkens HG, Egberts ACAssociation of risk of abnormal bleeding with degree of serotonin reuptake inhibition by antidepressants.Arch Intern Med20041642367–70
5. De Clerck FThe role of serotonin in thrombogenesis.Clin Physiol Biochem19908suppl 340–9
6. Dalton SO, Johansen C, Mellemkjaer L, Nørgård B, Sørensen HT, Olsen JHUse of selective serotonin reuptake inhibitors and risk of upper gastrointestinal tract bleeding: a population-based cohort study.Arch Intern Med200316359–64
7. van Walraven C, Mamdani MM, Wells PS, Williams JIInhibition of serotonin reuptake by antidepressants and upper gastrointestinal bleeding in elderly patients: retrospective cohort study.BMJ2001323655–8
8. Targownik LE, Bolton JM, Metge CJ, Leung S, Sareen JSelective serotonin reuptake inhibitors are associated with a modest increase in the risk of upper gastrointestinal bleeding.Am J Gastroenterol20091041475–82
9. Auerbach AD, Vittinghoff E, Maselli J, Pekow PS, Young JQ, Lindenauer PKPerioperative use of selective serotonin reuptake inhibitors and risks for adverse outcomes of surgery.JAMA Intern Med20131731075–81
10. Movig KL, Janssen MW, de Waal Malefijt J, Kabel PJ, Leufkens HG, Egberts ACRelationship of serotonergic antidepressants and need for blood transfusion in orthopedic surgical patients.Arch Intern Med20031632354–8
11. van Haelst IM, Egberts TC, Doodeman HJ, Traast HS, Burger BJ, Kalkman CJ, van Klei WAUse of serotonergic antidepressants and bleeding risk in orthopedic patients.Anesthesiology2010112631–6
12. Singh I, Achuthan S, Chakrabarti A, Rajagopalan S, Srinivasan A, Hota DInfluence of pre-operative use of serotonergic antidepressants (SADs) on the risk of bleeding in patients undergoing different surgical interventions: a meta-analysis.Pharmacoepidemiol Drug Saf201524237–45
13. Seitz DP, Bell CM, Gill SS, Reimer CL, Herrmann N, Anderson GM, Newman A, Rochon PARisk of perioperative blood transfusions and postoperative complications associated with serotonergic antidepressants in older adults undergoing hip fracture surgery.J Clin Psychopharmacol201333790–8
14. Tully PJ, Cardinal T, Bennetts JS, Baker RASelective serotonin reuptake inhibitors, venlafaxine and duloxetine are associated with in hospital morbidity but not bleeding or late mortality after coronary artery bypass graft surgery.Heart Lung Circ201221206–14
15. Xiong GL, Jiang W, Clare RM, Shaw LK, Smith PK, O’Connor CM, Ranga K, Krishnan R, Kristin Newby LSafety of selective serotonin reuptake inhibitor use prior to coronary artery bypass grafting.Clin Cardiol201033E94–8
16. Tatsumi M, Groshan K, Blakely RD, Richelson EPharmacological profile of antidepressants and related compounds at human monoamine transporters.Eur J Pharmacol1997340249–58
17. Spiess BD, Royston D, Levy JH, Fitch J, Dietrich W, Body S, Murkin J, Nadel APlatelet transfusions during coronary artery bypass graft surgery are associated with serious adverse outcomes.Transfusion2004441143–8
18. Murphy GJ, Reeves BC, Rogers CA, Rizvi SI, Culliford L, Angelini GDIncreased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery.Circulation20071162544–52
19. Dexter F, Dexter EU, Ledolter JStatistical grand rounds: importance of appropriately modeling procedure and duration in logistic regression studies of perioperative morbidity and mortality.Anesth Analg20111131197–201
20. Rosenbaum JF, Fava M, Hoog SL, Ascroft RC, Krebs WBSelective serotonin reuptake inhibitor discontinuation syndrome: a randomized clinical trial.Biol Psychiatry19984477–87
21. Zajecka J, Tracy KA, Mitchell SDiscontinuation symptoms after treatment with serotonin reuptake inhibitors: a literature review.J Clin Psychiatry199758291–7
22. Tint A, Haddad PM, Anderson IMThe effect of rate of antidepressant tapering on the incidence of discontinuation symptoms: a randomised study.J Psychopharmacol200822330–2
23. Anderson IM, Nutt DJ, Deakin JFEvidence-based guidelines for treating depressive disorders with antidepressants: a revision of the 1993 British Association for Psychopharmacology guidelines. British Association for Psychopharmacology.J Psychopharmacol2000143–20
© 2016 International Anesthesia Research Society