- Email: [email protected]
Anxiolytic medication is an independent risk factor for 30-day morbidity or mortality after surgery
Anxiolytic medication is an independent risk factor for 30-day morbidity or mortality after surgery Nicholas Ward, BSc,a J. Scott Roth, MD,b Clark C. Lester, MD,c Lori Mutiso, PhD,d Karen M. Lommel, DO,c and Daniel L. Davenport, PhD,b Lexington, KY
Background. This study examined the effects of the use of anxiolytic medications (AXM) and antidepressant medications (ADMs) on outcomes after noncardiac surgery. Study design. A single-center review of prospectively obtained, perioperative and 30-day outcome data, including AXM and ADM use at admission, as part of the National Surgery Quality Improvement Program. Results. Of the 1846 patients undergoing surgery, 380 (20.6%) were taking an ADM, 288 (15.6%) AXM, 124 (6.7%) were taking both, and 545 (29.5%) were taking either at the time of admission. Both ADM and AXM patients more often were female than nonusers, had a greater American Society of Anesthesiologists class and suffered more from hypertension, COPD, and dyspnea (all P < .005). AXM patients also were more often smokers. ADM patients had a greater mortality and a greater risk of an infective complication, but these effects did not remain after adjustment for procedure and comorbid risks. Patients taking AXM had greater duration of stay, as well as an increased incidence of return to the operating room, infections, wound occurrences, and cardiovascular or cerebrovascular events (all P < .005). After adjustment, AXM was associated with greater combined major morbidity or mortality (odds ratio 1.72, 95% confidence interval 1.08–2.73, P = .023). Conclusion. AXM was used by 16% of patients in our institution undergoing a noncardiac operation and was an independent risk factor for poorer short-term outcome after surgery. ADM was found to be used by 21% of patients but was not an independent risk factor for poor outcome. (Surgery 2015;158:420-7.) From the College of Medicine,a Department of Surgery,b Department of Psychiatry, c and College of Nursing,d University of Kentucky, Lexington, KY
THE
INCIDENCE OF PSYCHIATRIC ILLNESS IS INCREASING IN
UNITED STATES, with an estimated 9.6 million adult individuals (4.1% of all adults in the United States) suffering from serious mental illness as defined by the National Survey on Drug Use and Health.1,2 The lifetime morbid risk of any mood disorder, including major depression, is estimated to be 31%, and of any anxiety disorder to be as great as 41.7%.3 In addition to this increasing national incidence, the prevalence of depression and anxiety
THE
Presented at the 10th Annual Academic Surgical Congress in Las Vegas, NV, February 3–5, 2015. Accepted for publication March 16, 2015. Reprint requests: Daniel L. Davenport, PhD, Associate Professor, University of Kentucky, Department of Surgery, 800 Rose Street, Room MN278, Lexington, KY 40536-0298. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.03.050
420 SURGERY
has been shown to be substantially greater among cardiac surgery candidates than the general population. Indeed, after cardiac surgery, depression and anxiety are associated with greater durations of stay, greater rates of hospital readmission, reoperation, and short- and long-term mortality.4-9 Studies of the relationship between depression or anxiety and outcomes after noncardiac surgery are, as yet, rare. Therefore, the goal of this study was to assess the relationships between medically treated anxiety or depression and short-term outcomes after noncardiac surgery. METHODS The University of Kentucky Medical Institutional Review Board reviewed and approved this study and allowed a waiver of informed consent. This study was a retrospective review of the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) data gathered prospectively at the University of Kentucky A. B. Chandler Medical Center (UKMC) between
Surgery Volume 158, Number 2
Ward et al 421
Fig. The percentage of patients admitted for surgery using antidepressants or anxiolytics; stratified by the procedure performed. *‘‘Other’’ cases include a systematic sample of a broad range of general and vascular surgery cases performed under general anesthetic.
October 1, 2011, and September 30, 2012. All of the adult NSQIP procedures tracked at UKMC were included in the study and encompassed a broad range of general, vascular, urologic, and plastic surgery procedures (see the Fig).
Cardiothoracic, orthopedic, and neurosurgical cases were not included in the NSQIP protocol at the UKMC during this period. Two custom variables were added to the NSQIP protocol during that period at the time of
422 Ward et al
Surgery August 2015
Table I. List of antidepressant and anxiolytic drugs used to classify patients at admission Antidepressants, generic
Antidepressants, brand name
Anxiolytics, generic
Citalopram Escitalopram Paroxetine Fluoxetine Fluvoxamine Sertraline
Celexa Lexapro Paxil Prozac Luvox Zoloft
Duloxetine Venlafaxine Desvenlafaxine Buproprion Mirtazepine Trazodone Amitriptyline Imipramine Nortriptyline Protriptyline Perphenazine/Amitryptyline Tranylcypromine Phenelzine Selegiline Isocarbozxazid
Cymbalta Effexor XR Pristiq Wellbutrin/Zyban Remeron Desyrel Elavil Tofranil Pamelor Vivactil Triavil Parnate Nardil Emsam, Elderpyl, Zelapar Marplan
admission: antidepressant medications (ADMs) and anxiolytic medications (AXMs). The list of drugs used for each psychotropic designation is shown in Table I and was generated by a practicing clinical psychiatrist (K.L.). The data on medication usage were obtained from the UKMC medication review list, which indicated last dosage taken or ‘‘noncompliant’’ per the patient. The patients categorized as using the drugs in this study indicated they were taking these medications actively at the time of admission. At the UKMC, the NSQIP data are obtained prospectively and systematically by nurses with operative or intensive care unit experience who follow precise clinical definitions. The data include demographics, more than 30 preoperative clinical risk variables, perioperative processes, and 22 specific outcomes, including death up to 30 days after the operation. The validity of the data and the protocol have been described elsewhere.10,11 The procedures analyzed included all those included in the adult NSQIP program at the UKMC during the study period and are listed in the Fig. Each procedure had 100% sampling during the study period except for the ‘‘other’’ group, which included a broader systematic sample of other major general and vascular operations performed under general anesthetic.
Anxiolytics, brand name
Alprazolam Diazepam Clonazepam Lorazepam Chlordiazepoxide
Xanax Valium Klonopin Ativan Librium
Hydroxyzine
Atarax/Vistaril
Specific outcomes analyzed included duration of hospital stay, death, and major morbidity occurring up to 30-day after the operation including: return to the operating room for any reason, surgical-site infection (superficial, deep, or organ/ space) or wound dehiscence, sepsis or septic shock, urinary tract infection, renal insufficiency or failure, deep venous thrombosis or pulmonary embolism, stroke with deficit, cardiac arrest, and acute myocardial infarction. Univariate comparisons of all preoperative, intraoperative and outcome variables were performed between patients using AXM or ADM and nonusers using t-tests, Mann-Whiney U, v2, or Fisher exact tests as appropriate. Given the number of comparisons performed, we increased the threshold for significance to P < .005 for the univariate analyses. Two multivariable logistic regressions were performed. The first analyzed the effects of AXM or ADM use (with their interaction term) on combined mortality or major morbidity with adjustment for only the procedure performed. The second regression had 4 stages that started with the comorbid conditions that differed significantly in the univariate analyses fixed in the model; then other comorbid risks were added in a forward fashion with P for entry <.05, for exit >.10; third, intraoperative variables were added in a similar forward manner; and finally AXM use, ADM use and
Ward et al 423
Surgery Volume 158, Number 2
Table II. Patient characteristics that differed in antidepressant or anxiolytic groups Patient characteristic No. patients Preoperative characteristics Female, % ASA class I or II III IV or V Smoker, % Severe COPD, % Dyspnea, % Treated hypertension, % Intraoperative characteristics Duration of operation, min, mean ± SD
No antidepressant use 1,466
Antidepressant use 380 (20.6%)
52.2 44.1 49.2 6.6 32.7 5.9 14.1 49.7 169 ± 133
P value
1,557 <.001 <.001
68.2 32.4 60.3 7.4 33.4 12.4 21.8 57.9
No anxiolytic use
.802 <.001 .001 .005
183 ± 123
.076
54.6 44.5 49.2 6.3 30.5 5.9 13.7 49.7 166 ± 126
Anxiolytic use
P value
288 (15.6%) 60.2 26.3 64.4 9.3 45.7 14.9 26.3 60.9
.079 <.001
203 ± 153
<.001
<.001 <.001 <.001 <.001
ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease.
their interaction term were forced into the model. The regression models were assessed using the Hosmer-Lemeshow test and c-index. RESULTS A total of 1,846 patients underwent a noncardiac operation during the study period and had NSQIP data collected. The procedures are listed with their counts and rate of patient usage of ADM and AXM in the Fig. Procedure counts ranged from 20 abdominoplasties and 21 cystectomies to 230 ventral hernia repairs and 241 thyroidectomies. Overall, 380 (20.6%) patients were admitted on ADM, 288 (15.6%) on AXM, and 124 (6.7%) on both. Combining groups showed 29.5% of patients were on either anxiolytics or antidepressants. The percent of patients on ADM or AXM varied by the procedure performed (v2, P < .001); ranging from 12% of patients undergoing appendectomy to 55% of patients undergoing abdominoplasty. Use of ADM was greater in patients undergoing abdominoplasty (45%), breast reduction (41%), or breast reconstruction (35%) than the rest of the procedures (KruskalWallis subset analysis P < .05). Use of AXM was less in patients undergoing appendectomy (4%) than the rest of the procedures (Kruskal-Wallis subset analysis P < .05). Patient demographic and clinical characteristics compared between medication groups. Uses of AMD or AXM were both associated with a greater American Society of Anesthesiologists (ASA) class, a history of severe chronic obstructive pulmonary disease or dyspnea, and treated hypertension (Table II). In addition, patients on AXM
more often were smokers. No associations were found between use of AXM or ADM and age, race, body mass index, functional status, transfer status, history of diabetes, congestive heart failure, renal failure or dialysis, steroid treatment for a chronic condition, recent weight loss, or disseminated cancer. No serum biochemical values were associated with use of AXM or ADM (including hematocrit, white blood cell or platelet count, bilirubin, blood urea nitrogen, glutamic oxaloacetic transaminase, sodium, creatinine, the international normalized ratio, or albumin). Intraoperative variables compared between medication groups. Two thirds of the procedures were performed during inpatient admissions; 79% were same-day elective operations and 14.5% were performed emergently. None of these parameters varied by medication group. Wound class (49.2% clean) and perioperative transfusion rates (0.9%) did not vary by use of ADM or AXM. Patients on AXM had longer operations (Table II); this increase remained after multivariable adjustment for the procedure performed, emergency case, same-day elective case, inpatient versus outpatient status, wound class, need for transfusion, and ASA class (adjusted increase of 19 minutes, 95% confidence interval [95% CI] 5 to 33, P = .008). Thirty-day outcomes compared between medication groups. Patients on ADM had greater unadjusted mortality (3.9% vs 1.6% in nonusers, P = .004) and a greater composite rate of infection (14.5% vs 9.5%, P = .005, Table III). Neither of these differences remained after adjustment for
424 Ward et al
Surgery August 2015
Table III. Outcomes that differed in antidepressant or anxiolytic groups 30-day outcome No. patients All cause mortality, % Duration of stay, median (interquartile range) Composite death or major morbidity, % All cause return to operating room, % Composite infection: SSI, sepsis/shock, UTI, or pneumonia, % Wound occurrence: superficial, deep, or organ/space SSI, dehiscence, % Cardiac arrest, acute MI, or CVA with deficit, %
No antidepressant Antidepressant P No anxiolytic use use value use 1,466 1.6 2 (1–5.25)
Anxiolytic use
P value
380 (20.6%) 3.9 3 (1–7)
.004 .048
1,557 1.8 2 (1–5)
289 (15.6%) 3.5 .067 3 (1–7) <.001
15.6 4.4 9.5
19.3 5.0 14.5
.102 .591 .005
14.9 3.9 9.4
24.4 8.0 16.3
<.001 .002 .001
9.1
10.3
.503
8.5
13.8
.004
2.0
3.2
.244
1.8
4.8
.004
CVA, Cerebrovascular accident; MI, myocardial infarction; SSI, surgical-site infection; UTI, urinary tract infection.
the procedure. Patients on AXM had a day greater median duration of stay (3 vs 2 days, P < .001), increased composite mortality or major morbidity (24.4% vs 14.9%, P < .001), increased return to the operating room (8.0% vs 3.9%, P = .002), composite infections (16.3% vs 9.4%, P = .001), wound occurrences (13.8% vs 8.5%, P = .004), and cardiovascular or cerebrovascular events (4.8% vs 1.8%, P = .004). Neither AXM nor ADM was associated with increased venous thrombolytic events, renal failure or insufficiency, or respiratory failure (unplanned intubation or prolonged ventilation). With adjustment for the procedure performed, use of AXM was a predictor of composite mortality or major morbidity (odds ratio [OR] 1.82, 95% CI 1.30–2.56, P = .001). After further adjustment for 15 preoperative and perioperative risk factors (Table IV), AXM use conferred a greater risk for combined mortality or major morbidity (OR 1.72, 95% CI 1.08–2.73, P = .023). Use of ADM was not an independent predictor of poor outcome, nor did use of both drug types confer increased risk. The HosmerLemeshow test P = .798 indicated good fit in the multivariable model and the c-index = 0.848 indicated very good model discrimination of morbid/mortal events. DISCUSSION Thirty percent of patients presenting for a broad range of noncardiac surgeries at this large, tertiary referral center were actively taking AXM, ADM, or both. One in five was on ADM, almost double the national rate for US adults,12 whereas 1 in 6 was on AXM, again, at least double the rate of general anxiety disorder in the United States.13 This observation
confirms in a more general operative population what has been shown in the cardiac surgery literature, namely that the use of ADM and AXM are more common in the surgery population than the overall adult population.4-6 The percent of patients on AXM or ADM varied by the procedure performed, with appendectomy patients having the least incidence. This difference is likely attributable to the younger presentation of these patients. The top 4 procedures in terms of ADM use were primarily elective plastic surgery procedures performed on women. Women are known to have a greater incidence of depression14 and anxiety15 than men. One large study of young women undergoing cosmetic surgery showed greater levels than their peer controls of depression and anxiety before surgery.16 Our study confirms these relationships in regards to ADM use. Operative patients using AXM or ADM presented with a greater ASA class and were more likely to suffer from COPD, dyspnea, and medically treated hypertension. AXM users also were more likely smokers. The literature has shown that affective disorders are associated with obesity, hypertension, hyperlipidemia,17 physical inactivity,18 and regular smoking.19 Our results confirm the relationship of mood disorders with hypertension and smoking in patients undergoing surgery while adding the novel comorbid factors COPD and dyspnea. Body mass index was not significantly greater in our patients using AXM or ADM, although this lack of association could be related to the effect of many operative diseases on body weight. These comorbid relationships could be attributable to sicker patients being more often depressed or anxious and requiring these mood altering
Ward et al 425
Surgery Volume 158, Number 2
Table IV. Multivariable logistic regression results for 30-day composite mortality or major morbidity after a broad range of noncardiac operative procedures Variable Antidepressant use Anxiolytic use Interaction term (both) With adjustment for: Age, y Female sex Preoperative SIRS or sepsis Preoperative functional status vs independent Partially dependent Fully dependent ASA class vs I–II ASA III ASA IV–V Current smoker Severe COPD Dyspnea w/moderate exertion At rest Medically treated hypertension Disseminated cancer Known serum albumin <3.3 g/dL Procedure group Outpatient surgery Emergent surgery Operative duration, min Wound class vs 1. Clean 2. Clean contaminated 3. Contaminated 4. Dirty/infected Transfusion PRBCs during or within 72 h of operation
Incidence, n,% or mean ± SD
Odds ratio (95% CI) for major morbidity or mortality within 30 d
377, 20.6% 286, 15.6% 123, 6.7%
1.14 (0.74–1.77) 1.72 (1.08–2.73) 0.72 (0.32–1.61)
.553 .023 .419
52.7 ± 15.8 1019, 55.6% 123, 6.7% 1760, 96.0%
1.00 (0.99–1.01) 0.79 (0.57–1.09) 1.96 (1.05–3.65)
.766 .143 .035
49, 2.7% 24, 1.3% 763, 41.6% 948, 51.7% 122, 6.7% 603, 32.9% 132, 7.2%
2.51 (1.19–5.29) 4.35 (1.38–13.69)
.015 .012
0.98 2.11 0.98 1.69
.932 .022 .885 .040
(0.65–1.48) (1.22–3.98) (0.71–1.35) (1.03–2.78)
240, 13.1% 1.23 (0.78–1.91) 48, 2.6% 1.13 (0.49–2.58) 944, 51.5% 1.10 (0.78–1.56) 80, 4.4% 1.32 (0.71–2.45) 247, 3.5% 1.17 (0.77–1.79) 20 procedure groups not reported for clarity are available from authors See the Fig for group listing. 615, 33.6% 0.35 (0.20–0.60) 260, 14.2% 1.70 (0.99–2.92) 171 ± 131 1.001 (1.000–1.003) 907, 49.5% 668, 36.4% 1.30 (0.82–2.07) 151, 8.2% 2.12 (1.18–3.80) 107, 5.8% 2.34 (1.20–4.58) 194, 10.6% 2.23 (1.46–3.41)
P value in final model
.367 .780 .580 .379 .461 <.001 <.001 .054 .016 .270 .012 .013 <.001
ASA, American Society of Anesthesiologists; CI, confidence interval; COPD, chronic obstructive pulmonary disease; PRBCs, packed red blood cells; SIRS, systemic inflammatory response syndrome.
medications, or they could be attributable to psychiatric illness and treatment predisposing one toward comorbid disease and behaviors. This study does not allow us to comment. It is of note that impaired functional status, diabetes, disseminated cancer, and chronic conditions treated with steroids were not associated with increased AXM or ADM use in our study. With adjustment for comorbid condition and procedure performed, AXM use at admission predicted poorer outcomes, whereas ADM use did not. To our knowledge, this is the first study of its kind demonstrating worse outcomes in patients using AXM after noncardiac surgery. The increase in wound complications is consistent with
studies that show that chronic wound healing is delayed in patients with depression or anxiety.20,21 This study should serve to heighten awareness of the potential impact of the increasing usage of mood-altering medications, particularly AXM, on operative outcomes. Because of the design of this study, it is not possible to determine whether these outcomes were attributable to a side effect of the AXM or to physiologic mechanisms resulting from chronic anxiety. It is plausible that physiologic changes secondary to chronic anxiety are responsible for the increase in morbidity. People with anxiety disorders experience chronic stress, which stimulates the adrenal overproduction of stress
426 Ward et al
hormones, including cortisol.22 Cortisol is involved in a wide range of physiologic processes, including the immune response and regulation of inflammation among others.23 The increased morbidity observed in patients using AXM may result from the effects of cortisol on inflammation and/or the immune response. Other biologic mechanisms have been proposed to explain the impact of psychosocial factors on outcomes in the cardiac surgery literature, including poor compliance with recovery regimens,24 dysregulation of the hypothalamicpituitary-adrenal axis,25 heart rate variability,26 altered serotonergic pathways, and altered platelet aggregability.27 The incidence of use of moodaltering medications and the relationship with poor outcomes in this study supports the need for further study of their relationship. There are several limitations to this study. As mentioned previously, the preoperative use of ADM and AXM were proxies for chronic anxiety and depression, so the actual incidence of anxiety and depressive disorders in our study population is not known. Also, we cannot presume that all of the AXM tracked for this study were prescribed for chronic anxiety, nor that patient compliance was universal. Similar to all non-controlled studies, we cannot comment on specific mechanisms. Finally, the study has unknown treatment bias given our sample of patients presenting for surgery, however we adjusted for several procedure-related variables in the analyses to mitigate bias. In summary, AXM is an independent risk factor for major morbidity or mortality after surgery. The current study cannot determine whether this is related to the medication or pathophysiological changes due to chronic anxiety. ADM was not an independent risk factor for poor outcome after surgery. The authors acknowledge the diligent and knowledgeable work of our NSQIP data coordinators Devauna Riley, RN, and Roseanna Adair, RN, in the collection of the data used in this study.
REFERENCES 1. National Institutes of Mental Health. Use of mental health services and treatment among adults. Available from: http://www.nimh.nih.gov/health/statistics/prevalence/useof-mental-health-services-and-treatment-among-adults.shtml. 2. Kessler RC, Angermeyer M, Anthony JC, de Graaf R, Demyttenaere K, Gasquet I, et al. Lifetime prevalence and age-ofonset distributions of mental disorders in the World Health Organization’s World Mental Health Survey Initiative. World Psychiatry 2007;6:168-76.
Surgery August 2015
3. Kessler RC, Petukhova M, Sampson N, Zaslavsky A, Wittchen H. Twelve-month and lifetime prevalence and lifetime morbid risk of anxiety and mood disorders in the United States. Int J Methods Psychiatr Res 2012;21:169-84. 4. Ho MP, Pignay-Demaria V, Lesperance F, et al. Depression and anxiety and outcomes of coronary artery bypass surgery. Ann Thorac Surg 2003;75:314-21. 5. Blumenthal JA, Lett HS, Babyak MA, et al. Depression as a risk factor for mortality after coronary artery bypass surgery. Lancet 2003;362:604-9. 6. Ho PM, Masoudi FA, Spertus JA, Peterson PN, Shroyer AL, McCarthy M Jr, et al. Depression predicts mortality following cardiac valve surgery. Ann Thorac Surg 2005;79: 1255-9. 7. Stengrevics S, Sirois C, Schwartz CE, Friedman R, Domar A. The prediction of cardiac surgery outcome based upon preoperative psychological factors. Psychol Health 1996;11:471-7. 8. Tully PJ, Baker RA, Knight JL. Anxiety and depression as risk factors for mortality after coronary artery bypass surgery. J Psychosom Res 2008;64:285-90. 9. Tully PJ, Pedersen SS, Winefield HR, et al. Cardiac morbidity risk and depression and anxiety: a disorder, symptom and trait analysis among cardiac surgery patients. Psychol Health Med 2011;16:333-45. 10. Khuri SF, Daley J, Henderson W, Hur K, Demakis J, Aust JB, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peercontrolled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg 1998;228: 491-507. 11. Fink AS, Campbell DA Jr, Mentzer RM Jr, Henderson WG, Daley J, Bannister J, et al. The National Surgical Quality Improvement Program in non-veterans administration hospitals: initial demonstration of feasibility. Ann Surg 2002; 236:344-53; discussion 353–354. 12. LA Pratt, DJ Brody, and Q Gu, ‘‘Antidepressant Use in Persons Aged 12 and Over: United States, 2005–2008,’’ NCHS Data Brief No. 76, 2011 Oct. Available from: http://www.cdc.gov/nchs/data/databriefs/db76.pdf. Accessed December 20, 2014. 13. National Institutes of Mental Health. Generalized Anxiety Disorder Among Adults. Available from: http://www.nimh. nih.gov/health/statistics/prevalence/generalized-anxietydisorder-among-adults.shtml. Accessed December 20, 2014. 14. Weissman MM, Klerman GL. Gender and depression. Trends Neurosci 1985;8:416-20. 15. McLean CP, Asnaani A, Litz BT, Hofmann SG. Gender differences in anxiety disorders: prevalence, course of illness, comorbidity and burden of illness. J Psychiatr Res 2011; 45:1027-35. 16. von Soest T, Kvalem IL, Skolleborg KC, Roald HE. Psychosocial changes after cosmetic surgery: a 5-year follow-up study. Plast Reconstr Surg 2011;128:765-72. 17. Barger SD, Sydeman SJ. Does generalized anxiety disorder predict coronary heart disease risk factors independently of major depressive disorder? J Affect Disord 2005;88:87-91. 18. Goodwin RD. Association between physical activity and mental disorders among adults in the United States. Prev Med 2003;36:698-703. 19. Davydow DS. Symptoms of depression and anxiety after delirium. Psychosomatics 2009;50:309-16. 20. Cole-King A, Harding KG. Psychological factors and delayed healing in chronic wounds. Psychosom Med 2001; 63:216-20.
Surgery Volume 158, Number 2
21. Kiecolt-Glaser JK, Marucha PT, Malarkey WB, Mercado AM, Glaser R. Slowing of wound healing by psychological stress. Lancet 1995;346:1194-6. 22. Heim C, Ehlert U, Hellhammer DH. The potential role of hypocortisolism in the pathophysiology of stressrelated bodily disorders. Psychoneuroendocrinology 2000; 25:1-35. 23. Koliantzaki I, Zakynthinos SG, Mentzelopoulos SD. The potential contribution of corticosteroids to positive cardiac arrest outcomes. In: Gullo A, Ristagno, editors. Resuscitation. Milan: Springer Milan; 2014. p. 143-55. 24. Kuhl EA, Fauerbach JA, Bush DE, et al. Relation of anxiety and adherence to risk-reducing recommendations
Ward et al 427
following myocardial infarction. Am J Cardiol 2009;103: 1629-34. 25. Tsigos C, Chrousos GP. Hypothalamic-pituitary-adrenal axis, neuroendocrine factors and stress. J Psychosom Res 2002; 53:865-71. 26. Tully PJ, Bennetts JS, Baker RA, et al. Anxiety, depression, and stress as risk factors for atrial fibrillation after cardiac surgery. Heart Lung 2011;40:4-11. 27. Tully PJ, Cardinal T, Bennetts JS, Baker RA. Selective 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 Circ 2012;21:206-14.
Background. This study examined the effects of the use of anxiolytic medications (AXM) and antidepressant medications (ADMs) on outcomes after noncardiac surgery. Study design. A single-center review of prospectively obtained, perioperative and 30-day outcome data, including AXM and ADM use at admission, as part of the National Surgery Quality Improvement Program. Results. Of the 1846 patients undergoing surgery, 380 (20.6%) were taking an ADM, 288 (15.6%) AXM, 124 (6.7%) were taking both, and 545 (29.5%) were taking either at the time of admission. Both ADM and AXM patients more often were female than nonusers, had a greater American Society of Anesthesiologists class and suffered more from hypertension, COPD, and dyspnea (all P < .005). AXM patients also were more often smokers. ADM patients had a greater mortality and a greater risk of an infective complication, but these effects did not remain after adjustment for procedure and comorbid risks. Patients taking AXM had greater duration of stay, as well as an increased incidence of return to the operating room, infections, wound occurrences, and cardiovascular or cerebrovascular events (all P < .005). After adjustment, AXM was associated with greater combined major morbidity or mortality (odds ratio 1.72, 95% confidence interval 1.08–2.73, P = .023). Conclusion. AXM was used by 16% of patients in our institution undergoing a noncardiac operation and was an independent risk factor for poorer short-term outcome after surgery. ADM was found to be used by 21% of patients but was not an independent risk factor for poor outcome. (Surgery 2015;158:420-7.) From the College of Medicine,a Department of Surgery,b Department of Psychiatry, c and College of Nursing,d University of Kentucky, Lexington, KY
THE
INCIDENCE OF PSYCHIATRIC ILLNESS IS INCREASING IN
UNITED STATES, with an estimated 9.6 million adult individuals (4.1% of all adults in the United States) suffering from serious mental illness as defined by the National Survey on Drug Use and Health.1,2 The lifetime morbid risk of any mood disorder, including major depression, is estimated to be 31%, and of any anxiety disorder to be as great as 41.7%.3 In addition to this increasing national incidence, the prevalence of depression and anxiety
THE
Presented at the 10th Annual Academic Surgical Congress in Las Vegas, NV, February 3–5, 2015. Accepted for publication March 16, 2015. Reprint requests: Daniel L. Davenport, PhD, Associate Professor, University of Kentucky, Department of Surgery, 800 Rose Street, Room MN278, Lexington, KY 40536-0298. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.03.050
420 SURGERY
has been shown to be substantially greater among cardiac surgery candidates than the general population. Indeed, after cardiac surgery, depression and anxiety are associated with greater durations of stay, greater rates of hospital readmission, reoperation, and short- and long-term mortality.4-9 Studies of the relationship between depression or anxiety and outcomes after noncardiac surgery are, as yet, rare. Therefore, the goal of this study was to assess the relationships between medically treated anxiety or depression and short-term outcomes after noncardiac surgery. METHODS The University of Kentucky Medical Institutional Review Board reviewed and approved this study and allowed a waiver of informed consent. This study was a retrospective review of the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) data gathered prospectively at the University of Kentucky A. B. Chandler Medical Center (UKMC) between
Surgery Volume 158, Number 2
Ward et al 421
Fig. The percentage of patients admitted for surgery using antidepressants or anxiolytics; stratified by the procedure performed. *‘‘Other’’ cases include a systematic sample of a broad range of general and vascular surgery cases performed under general anesthetic.
October 1, 2011, and September 30, 2012. All of the adult NSQIP procedures tracked at UKMC were included in the study and encompassed a broad range of general, vascular, urologic, and plastic surgery procedures (see the Fig).
Cardiothoracic, orthopedic, and neurosurgical cases were not included in the NSQIP protocol at the UKMC during this period. Two custom variables were added to the NSQIP protocol during that period at the time of
422 Ward et al
Surgery August 2015
Table I. List of antidepressant and anxiolytic drugs used to classify patients at admission Antidepressants, generic
Antidepressants, brand name
Anxiolytics, generic
Citalopram Escitalopram Paroxetine Fluoxetine Fluvoxamine Sertraline
Celexa Lexapro Paxil Prozac Luvox Zoloft
Duloxetine Venlafaxine Desvenlafaxine Buproprion Mirtazepine Trazodone Amitriptyline Imipramine Nortriptyline Protriptyline Perphenazine/Amitryptyline Tranylcypromine Phenelzine Selegiline Isocarbozxazid
Cymbalta Effexor XR Pristiq Wellbutrin/Zyban Remeron Desyrel Elavil Tofranil Pamelor Vivactil Triavil Parnate Nardil Emsam, Elderpyl, Zelapar Marplan
admission: antidepressant medications (ADMs) and anxiolytic medications (AXMs). The list of drugs used for each psychotropic designation is shown in Table I and was generated by a practicing clinical psychiatrist (K.L.). The data on medication usage were obtained from the UKMC medication review list, which indicated last dosage taken or ‘‘noncompliant’’ per the patient. The patients categorized as using the drugs in this study indicated they were taking these medications actively at the time of admission. At the UKMC, the NSQIP data are obtained prospectively and systematically by nurses with operative or intensive care unit experience who follow precise clinical definitions. The data include demographics, more than 30 preoperative clinical risk variables, perioperative processes, and 22 specific outcomes, including death up to 30 days after the operation. The validity of the data and the protocol have been described elsewhere.10,11 The procedures analyzed included all those included in the adult NSQIP program at the UKMC during the study period and are listed in the Fig. Each procedure had 100% sampling during the study period except for the ‘‘other’’ group, which included a broader systematic sample of other major general and vascular operations performed under general anesthetic.
Anxiolytics, brand name
Alprazolam Diazepam Clonazepam Lorazepam Chlordiazepoxide
Xanax Valium Klonopin Ativan Librium
Hydroxyzine
Atarax/Vistaril
Specific outcomes analyzed included duration of hospital stay, death, and major morbidity occurring up to 30-day after the operation including: return to the operating room for any reason, surgical-site infection (superficial, deep, or organ/ space) or wound dehiscence, sepsis or septic shock, urinary tract infection, renal insufficiency or failure, deep venous thrombosis or pulmonary embolism, stroke with deficit, cardiac arrest, and acute myocardial infarction. Univariate comparisons of all preoperative, intraoperative and outcome variables were performed between patients using AXM or ADM and nonusers using t-tests, Mann-Whiney U, v2, or Fisher exact tests as appropriate. Given the number of comparisons performed, we increased the threshold for significance to P < .005 for the univariate analyses. Two multivariable logistic regressions were performed. The first analyzed the effects of AXM or ADM use (with their interaction term) on combined mortality or major morbidity with adjustment for only the procedure performed. The second regression had 4 stages that started with the comorbid conditions that differed significantly in the univariate analyses fixed in the model; then other comorbid risks were added in a forward fashion with P for entry <.05, for exit >.10; third, intraoperative variables were added in a similar forward manner; and finally AXM use, ADM use and
Ward et al 423
Surgery Volume 158, Number 2
Table II. Patient characteristics that differed in antidepressant or anxiolytic groups Patient characteristic No. patients Preoperative characteristics Female, % ASA class I or II III IV or V Smoker, % Severe COPD, % Dyspnea, % Treated hypertension, % Intraoperative characteristics Duration of operation, min, mean ± SD
No antidepressant use 1,466
Antidepressant use 380 (20.6%)
52.2 44.1 49.2 6.6 32.7 5.9 14.1 49.7 169 ± 133
P value
1,557 <.001 <.001
68.2 32.4 60.3 7.4 33.4 12.4 21.8 57.9
No anxiolytic use
.802 <.001 .001 .005
183 ± 123
.076
54.6 44.5 49.2 6.3 30.5 5.9 13.7 49.7 166 ± 126
Anxiolytic use
P value
288 (15.6%) 60.2 26.3 64.4 9.3 45.7 14.9 26.3 60.9
.079 <.001
203 ± 153
<.001
<.001 <.001 <.001 <.001
ASA, American Society of Anesthesiologists; COPD, chronic obstructive pulmonary disease.
their interaction term were forced into the model. The regression models were assessed using the Hosmer-Lemeshow test and c-index. RESULTS A total of 1,846 patients underwent a noncardiac operation during the study period and had NSQIP data collected. The procedures are listed with their counts and rate of patient usage of ADM and AXM in the Fig. Procedure counts ranged from 20 abdominoplasties and 21 cystectomies to 230 ventral hernia repairs and 241 thyroidectomies. Overall, 380 (20.6%) patients were admitted on ADM, 288 (15.6%) on AXM, and 124 (6.7%) on both. Combining groups showed 29.5% of patients were on either anxiolytics or antidepressants. The percent of patients on ADM or AXM varied by the procedure performed (v2, P < .001); ranging from 12% of patients undergoing appendectomy to 55% of patients undergoing abdominoplasty. Use of ADM was greater in patients undergoing abdominoplasty (45%), breast reduction (41%), or breast reconstruction (35%) than the rest of the procedures (KruskalWallis subset analysis P < .05). Use of AXM was less in patients undergoing appendectomy (4%) than the rest of the procedures (Kruskal-Wallis subset analysis P < .05). Patient demographic and clinical characteristics compared between medication groups. Uses of AMD or AXM were both associated with a greater American Society of Anesthesiologists (ASA) class, a history of severe chronic obstructive pulmonary disease or dyspnea, and treated hypertension (Table II). In addition, patients on AXM
more often were smokers. No associations were found between use of AXM or ADM and age, race, body mass index, functional status, transfer status, history of diabetes, congestive heart failure, renal failure or dialysis, steroid treatment for a chronic condition, recent weight loss, or disseminated cancer. No serum biochemical values were associated with use of AXM or ADM (including hematocrit, white blood cell or platelet count, bilirubin, blood urea nitrogen, glutamic oxaloacetic transaminase, sodium, creatinine, the international normalized ratio, or albumin). Intraoperative variables compared between medication groups. Two thirds of the procedures were performed during inpatient admissions; 79% were same-day elective operations and 14.5% were performed emergently. None of these parameters varied by medication group. Wound class (49.2% clean) and perioperative transfusion rates (0.9%) did not vary by use of ADM or AXM. Patients on AXM had longer operations (Table II); this increase remained after multivariable adjustment for the procedure performed, emergency case, same-day elective case, inpatient versus outpatient status, wound class, need for transfusion, and ASA class (adjusted increase of 19 minutes, 95% confidence interval [95% CI] 5 to 33, P = .008). Thirty-day outcomes compared between medication groups. Patients on ADM had greater unadjusted mortality (3.9% vs 1.6% in nonusers, P = .004) and a greater composite rate of infection (14.5% vs 9.5%, P = .005, Table III). Neither of these differences remained after adjustment for
424 Ward et al
Surgery August 2015
Table III. Outcomes that differed in antidepressant or anxiolytic groups 30-day outcome No. patients All cause mortality, % Duration of stay, median (interquartile range) Composite death or major morbidity, % All cause return to operating room, % Composite infection: SSI, sepsis/shock, UTI, or pneumonia, % Wound occurrence: superficial, deep, or organ/space SSI, dehiscence, % Cardiac arrest, acute MI, or CVA with deficit, %
No antidepressant Antidepressant P No anxiolytic use use value use 1,466 1.6 2 (1–5.25)
Anxiolytic use
P value
380 (20.6%) 3.9 3 (1–7)
.004 .048
1,557 1.8 2 (1–5)
289 (15.6%) 3.5 .067 3 (1–7) <.001
15.6 4.4 9.5
19.3 5.0 14.5
.102 .591 .005
14.9 3.9 9.4
24.4 8.0 16.3
<.001 .002 .001
9.1
10.3
.503
8.5
13.8
.004
2.0
3.2
.244
1.8
4.8
.004
CVA, Cerebrovascular accident; MI, myocardial infarction; SSI, surgical-site infection; UTI, urinary tract infection.
the procedure. Patients on AXM had a day greater median duration of stay (3 vs 2 days, P < .001), increased composite mortality or major morbidity (24.4% vs 14.9%, P < .001), increased return to the operating room (8.0% vs 3.9%, P = .002), composite infections (16.3% vs 9.4%, P = .001), wound occurrences (13.8% vs 8.5%, P = .004), and cardiovascular or cerebrovascular events (4.8% vs 1.8%, P = .004). Neither AXM nor ADM was associated with increased venous thrombolytic events, renal failure or insufficiency, or respiratory failure (unplanned intubation or prolonged ventilation). With adjustment for the procedure performed, use of AXM was a predictor of composite mortality or major morbidity (odds ratio [OR] 1.82, 95% CI 1.30–2.56, P = .001). After further adjustment for 15 preoperative and perioperative risk factors (Table IV), AXM use conferred a greater risk for combined mortality or major morbidity (OR 1.72, 95% CI 1.08–2.73, P = .023). Use of ADM was not an independent predictor of poor outcome, nor did use of both drug types confer increased risk. The HosmerLemeshow test P = .798 indicated good fit in the multivariable model and the c-index = 0.848 indicated very good model discrimination of morbid/mortal events. DISCUSSION Thirty percent of patients presenting for a broad range of noncardiac surgeries at this large, tertiary referral center were actively taking AXM, ADM, or both. One in five was on ADM, almost double the national rate for US adults,12 whereas 1 in 6 was on AXM, again, at least double the rate of general anxiety disorder in the United States.13 This observation
confirms in a more general operative population what has been shown in the cardiac surgery literature, namely that the use of ADM and AXM are more common in the surgery population than the overall adult population.4-6 The percent of patients on AXM or ADM varied by the procedure performed, with appendectomy patients having the least incidence. This difference is likely attributable to the younger presentation of these patients. The top 4 procedures in terms of ADM use were primarily elective plastic surgery procedures performed on women. Women are known to have a greater incidence of depression14 and anxiety15 than men. One large study of young women undergoing cosmetic surgery showed greater levels than their peer controls of depression and anxiety before surgery.16 Our study confirms these relationships in regards to ADM use. Operative patients using AXM or ADM presented with a greater ASA class and were more likely to suffer from COPD, dyspnea, and medically treated hypertension. AXM users also were more likely smokers. The literature has shown that affective disorders are associated with obesity, hypertension, hyperlipidemia,17 physical inactivity,18 and regular smoking.19 Our results confirm the relationship of mood disorders with hypertension and smoking in patients undergoing surgery while adding the novel comorbid factors COPD and dyspnea. Body mass index was not significantly greater in our patients using AXM or ADM, although this lack of association could be related to the effect of many operative diseases on body weight. These comorbid relationships could be attributable to sicker patients being more often depressed or anxious and requiring these mood altering
Ward et al 425
Surgery Volume 158, Number 2
Table IV. Multivariable logistic regression results for 30-day composite mortality or major morbidity after a broad range of noncardiac operative procedures Variable Antidepressant use Anxiolytic use Interaction term (both) With adjustment for: Age, y Female sex Preoperative SIRS or sepsis Preoperative functional status vs independent Partially dependent Fully dependent ASA class vs I–II ASA III ASA IV–V Current smoker Severe COPD Dyspnea w/moderate exertion At rest Medically treated hypertension Disseminated cancer Known serum albumin <3.3 g/dL Procedure group Outpatient surgery Emergent surgery Operative duration, min Wound class vs 1. Clean 2. Clean contaminated 3. Contaminated 4. Dirty/infected Transfusion PRBCs during or within 72 h of operation
Incidence, n,% or mean ± SD
Odds ratio (95% CI) for major morbidity or mortality within 30 d
377, 20.6% 286, 15.6% 123, 6.7%
1.14 (0.74–1.77) 1.72 (1.08–2.73) 0.72 (0.32–1.61)
.553 .023 .419
52.7 ± 15.8 1019, 55.6% 123, 6.7% 1760, 96.0%
1.00 (0.99–1.01) 0.79 (0.57–1.09) 1.96 (1.05–3.65)
.766 .143 .035
49, 2.7% 24, 1.3% 763, 41.6% 948, 51.7% 122, 6.7% 603, 32.9% 132, 7.2%
2.51 (1.19–5.29) 4.35 (1.38–13.69)
.015 .012
0.98 2.11 0.98 1.69
.932 .022 .885 .040
(0.65–1.48) (1.22–3.98) (0.71–1.35) (1.03–2.78)
240, 13.1% 1.23 (0.78–1.91) 48, 2.6% 1.13 (0.49–2.58) 944, 51.5% 1.10 (0.78–1.56) 80, 4.4% 1.32 (0.71–2.45) 247, 3.5% 1.17 (0.77–1.79) 20 procedure groups not reported for clarity are available from authors See the Fig for group listing. 615, 33.6% 0.35 (0.20–0.60) 260, 14.2% 1.70 (0.99–2.92) 171 ± 131 1.001 (1.000–1.003) 907, 49.5% 668, 36.4% 1.30 (0.82–2.07) 151, 8.2% 2.12 (1.18–3.80) 107, 5.8% 2.34 (1.20–4.58) 194, 10.6% 2.23 (1.46–3.41)
P value in final model
.367 .780 .580 .379 .461 <.001 <.001 .054 .016 .270 .012 .013 <.001
ASA, American Society of Anesthesiologists; CI, confidence interval; COPD, chronic obstructive pulmonary disease; PRBCs, packed red blood cells; SIRS, systemic inflammatory response syndrome.
medications, or they could be attributable to psychiatric illness and treatment predisposing one toward comorbid disease and behaviors. This study does not allow us to comment. It is of note that impaired functional status, diabetes, disseminated cancer, and chronic conditions treated with steroids were not associated with increased AXM or ADM use in our study. With adjustment for comorbid condition and procedure performed, AXM use at admission predicted poorer outcomes, whereas ADM use did not. To our knowledge, this is the first study of its kind demonstrating worse outcomes in patients using AXM after noncardiac surgery. The increase in wound complications is consistent with
studies that show that chronic wound healing is delayed in patients with depression or anxiety.20,21 This study should serve to heighten awareness of the potential impact of the increasing usage of mood-altering medications, particularly AXM, on operative outcomes. Because of the design of this study, it is not possible to determine whether these outcomes were attributable to a side effect of the AXM or to physiologic mechanisms resulting from chronic anxiety. It is plausible that physiologic changes secondary to chronic anxiety are responsible for the increase in morbidity. People with anxiety disorders experience chronic stress, which stimulates the adrenal overproduction of stress
426 Ward et al
hormones, including cortisol.22 Cortisol is involved in a wide range of physiologic processes, including the immune response and regulation of inflammation among others.23 The increased morbidity observed in patients using AXM may result from the effects of cortisol on inflammation and/or the immune response. Other biologic mechanisms have been proposed to explain the impact of psychosocial factors on outcomes in the cardiac surgery literature, including poor compliance with recovery regimens,24 dysregulation of the hypothalamicpituitary-adrenal axis,25 heart rate variability,26 altered serotonergic pathways, and altered platelet aggregability.27 The incidence of use of moodaltering medications and the relationship with poor outcomes in this study supports the need for further study of their relationship. There are several limitations to this study. As mentioned previously, the preoperative use of ADM and AXM were proxies for chronic anxiety and depression, so the actual incidence of anxiety and depressive disorders in our study population is not known. Also, we cannot presume that all of the AXM tracked for this study were prescribed for chronic anxiety, nor that patient compliance was universal. Similar to all non-controlled studies, we cannot comment on specific mechanisms. Finally, the study has unknown treatment bias given our sample of patients presenting for surgery, however we adjusted for several procedure-related variables in the analyses to mitigate bias. In summary, AXM is an independent risk factor for major morbidity or mortality after surgery. The current study cannot determine whether this is related to the medication or pathophysiological changes due to chronic anxiety. ADM was not an independent risk factor for poor outcome after surgery. The authors acknowledge the diligent and knowledgeable work of our NSQIP data coordinators Devauna Riley, RN, and Roseanna Adair, RN, in the collection of the data used in this study.
REFERENCES 1. National Institutes of Mental Health. Use of mental health services and treatment among adults. Available from: http://www.nimh.nih.gov/health/statistics/prevalence/useof-mental-health-services-and-treatment-among-adults.shtml. 2. Kessler RC, Angermeyer M, Anthony JC, de Graaf R, Demyttenaere K, Gasquet I, et al. Lifetime prevalence and age-ofonset distributions of mental disorders in the World Health Organization’s World Mental Health Survey Initiative. World Psychiatry 2007;6:168-76.
Surgery August 2015
3. Kessler RC, Petukhova M, Sampson N, Zaslavsky A, Wittchen H. Twelve-month and lifetime prevalence and lifetime morbid risk of anxiety and mood disorders in the United States. Int J Methods Psychiatr Res 2012;21:169-84. 4. Ho MP, Pignay-Demaria V, Lesperance F, et al. Depression and anxiety and outcomes of coronary artery bypass surgery. Ann Thorac Surg 2003;75:314-21. 5. Blumenthal JA, Lett HS, Babyak MA, et al. Depression as a risk factor for mortality after coronary artery bypass surgery. Lancet 2003;362:604-9. 6. Ho PM, Masoudi FA, Spertus JA, Peterson PN, Shroyer AL, McCarthy M Jr, et al. Depression predicts mortality following cardiac valve surgery. Ann Thorac Surg 2005;79: 1255-9. 7. Stengrevics S, Sirois C, Schwartz CE, Friedman R, Domar A. The prediction of cardiac surgery outcome based upon preoperative psychological factors. Psychol Health 1996;11:471-7. 8. Tully PJ, Baker RA, Knight JL. Anxiety and depression as risk factors for mortality after coronary artery bypass surgery. J Psychosom Res 2008;64:285-90. 9. Tully PJ, Pedersen SS, Winefield HR, et al. Cardiac morbidity risk and depression and anxiety: a disorder, symptom and trait analysis among cardiac surgery patients. Psychol Health Med 2011;16:333-45. 10. Khuri SF, Daley J, Henderson W, Hur K, Demakis J, Aust JB, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peercontrolled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg 1998;228: 491-507. 11. Fink AS, Campbell DA Jr, Mentzer RM Jr, Henderson WG, Daley J, Bannister J, et al. The National Surgical Quality Improvement Program in non-veterans administration hospitals: initial demonstration of feasibility. Ann Surg 2002; 236:344-53; discussion 353–354. 12. LA Pratt, DJ Brody, and Q Gu, ‘‘Antidepressant Use in Persons Aged 12 and Over: United States, 2005–2008,’’ NCHS Data Brief No. 76, 2011 Oct. Available from: http://www.cdc.gov/nchs/data/databriefs/db76.pdf. Accessed December 20, 2014. 13. National Institutes of Mental Health. Generalized Anxiety Disorder Among Adults. Available from: http://www.nimh. nih.gov/health/statistics/prevalence/generalized-anxietydisorder-among-adults.shtml. Accessed December 20, 2014. 14. Weissman MM, Klerman GL. Gender and depression. Trends Neurosci 1985;8:416-20. 15. McLean CP, Asnaani A, Litz BT, Hofmann SG. Gender differences in anxiety disorders: prevalence, course of illness, comorbidity and burden of illness. J Psychiatr Res 2011; 45:1027-35. 16. von Soest T, Kvalem IL, Skolleborg KC, Roald HE. Psychosocial changes after cosmetic surgery: a 5-year follow-up study. Plast Reconstr Surg 2011;128:765-72. 17. Barger SD, Sydeman SJ. Does generalized anxiety disorder predict coronary heart disease risk factors independently of major depressive disorder? J Affect Disord 2005;88:87-91. 18. Goodwin RD. Association between physical activity and mental disorders among adults in the United States. Prev Med 2003;36:698-703. 19. Davydow DS. Symptoms of depression and anxiety after delirium. Psychosomatics 2009;50:309-16. 20. Cole-King A, Harding KG. Psychological factors and delayed healing in chronic wounds. Psychosom Med 2001; 63:216-20.
Surgery Volume 158, Number 2
21. Kiecolt-Glaser JK, Marucha PT, Malarkey WB, Mercado AM, Glaser R. Slowing of wound healing by psychological stress. Lancet 1995;346:1194-6. 22. Heim C, Ehlert U, Hellhammer DH. The potential role of hypocortisolism in the pathophysiology of stressrelated bodily disorders. Psychoneuroendocrinology 2000; 25:1-35. 23. Koliantzaki I, Zakynthinos SG, Mentzelopoulos SD. The potential contribution of corticosteroids to positive cardiac arrest outcomes. In: Gullo A, Ristagno, editors. Resuscitation. Milan: Springer Milan; 2014. p. 143-55. 24. Kuhl EA, Fauerbach JA, Bush DE, et al. Relation of anxiety and adherence to risk-reducing recommendations
Ward et al 427
following myocardial infarction. Am J Cardiol 2009;103: 1629-34. 25. Tsigos C, Chrousos GP. Hypothalamic-pituitary-adrenal axis, neuroendocrine factors and stress. J Psychosom Res 2002; 53:865-71. 26. Tully PJ, Bennetts JS, Baker RA, et al. Anxiety, depression, and stress as risk factors for atrial fibrillation after cardiac surgery. Heart Lung 2011;40:4-11. 27. Tully PJ, Cardinal T, Bennetts JS, Baker RA. Selective 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 Circ 2012;21:206-14.