Pre-injury neuro-psychiatric medication use, alone or in combination with cardiac medications, may affect outcomes in trauma patientsJR Wisler1, AN Springer2, K Hateley1, XM Mo3, DC Evans1, CH Cook1, AT Gerlach4, CV Murphy4, DS Eiferman1, SM Steinberg1, SD Bergese2, TJ Papadimos2, SP Stawicki5
1 Department of Surgery, Division of Trauma, Critical Care, and Burn, The Ohio State University College of Medicine, Columbus, Ohio, USA
2 Department of Anesthesiology, The Ohio State University College of Medicine, Columbus, Ohio, USA
3 Department of Biostatistics, The Ohio State University College of Medicine, Columbus, Ohio, USA
4 Department of Pharmacy, The Ohio State University College of Medicine, Columbus, Ohio, USA
5 Department of Research & Innovation, St Luke's University Health Network, Bethlehem, Pennsylvania, USA
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.143957
Source of Support: None, Conflict of Interest: None
Background: Recent review of older (≥45-years-old) patients admitted to our trauma center showed that more than one-third were using neuro-psychiatric medications (NPMs) prior to their injury-related admission. Previously published data suggests that use of NPMs may increase patients' risk and severity of injury. We sought to examine the impact of pre-injury NPM use on older trauma patients' morbidity and mortality. Materials and Methods: Retrospective record review included medication regimen characteristics and NPM use (antidepressants-AD, antipsychotics-AP, anxiolytics-AA). Hospital morbidity, mortality, and 90-day survival were examined. Comparisons included regimens involving NPMs, further focusing on their interactions with various cardiac medications (beta blocker - BB; angiotensin-converting enzyme inhibitor/angiotensin receptor blocker - ACE/ARB; calcium channel blocker - CCB). Results: 712 patient records were reviewed (399 males, mean age 63.5 years, median ISS 8). 245 patients were taking at least 1 NPM: AD (158), AP (35), or AA (108) before injury. There was no effect of NPM monotherapy on hospital mortality. Patients taking ≥3 NPMs had significantly lower 90-day survival compared to patients taking ≤2 NPMs (81% for 3 or more NPMs, 95% for no NPMs, and 89% 1-2 NPMs, P < 0.01). Several AD-cardiac medication (CM) combinations were associated with increased mortality compared to monotherapy with either agent (BB-AD 14.7% mortality versus 7.0% for AD monotherapy or 4.8% BB monotherapy, P < 0.05). Combinations of ACE/ARB-AA were associated with increased mortality compared to ACE/ARB monotherapy (11.5% vs 4.9, P = 0.04). Finally, ACE/ARB-AD co-administration had higher mortality than ACE/ARB monotherapy (13.5% vs 4.9%, P = 0.01). Conclusions: Large proportion of older trauma patients was using pre-injury NPMs. Several regimens involving NPMs and CMs were associated with increased in-hospital mortality. Additionally, use of ≥3 NPMs was associated with lower 90-day survival.
Keywords: Cardiac medications, medication interactions, mortality and morbidity, neuro-psychiatric medications, trauma outcomes
The average age of the U.S. population continues to increase, and physicians are recognizing the fast changing needs of the most active older patients. It is estimated that by the year 2020, the number of Americans older than 65 years will increase to over 80 million.  Following acute injury, older trauma victims are more likely to experience complications, mortality, and lower overall survival for any given level of injury severity when compared to younger patient groups. , This phenomenon is especially pronounced in the context of pre-existing medical conditions, disabilities, and polypharmacy.  A recent study found that >40% of trauma patients aged 45 years or older were receiving 5 or more medications at the time of their injury. 
While the impact of pre-injury polypharmacy on trauma outcomes is well-established,  we are only beginning to explore the intricacies of this complex issue. Our earlier work on polypharmacy in older trauma patients showed that 44% of patients 45 years and older were taking at least one neuro-psychiatric medication (NPM) and 42% of patients were using cardiac medications (CM) (coagulation modulators excluded).  Due to the frequent co-administration of both NPMs and CMs observed in our previous study, we sought to determine whether agents from these two intrinsically diverse medication classes had an association or impacted clinical outcomes in older trauma patients. Specifically, this study investigated the impact of medication regimens containing NPMs, CMs, or both on mortality and morbidity in older trauma patients.
Ethics: The study protocol was approved by the Institutional Review Board and consent waiver was granted for retrospective analysis. Confidentiality of information was maintained.
Design and setting: A retrospective analysis of patients identified from an American College of Surgeons verified level I trauma center institutional database between January 2005 and January 2009.
Selection criteria: The patients selected were aged ≥45 years old. This was based on our previous research as this age cut-off represented the best estimate of a group that demonstrated an appreciable number of chronic health conditions commonly treated with long-term "maintenance" pharmacologic therapy. , Excluded were prisoners, pregnant patients, and patients <45 years of age.
Processes and procedures at the center: In accordance with our institutional pharmacy protocols, detailed medication reconciliation is performed for all newly admitted injured patients (with rare exceptions where such history is not obtainable). The list of medications that a patient had been taking prior to admission was designated as the "arrival medication list". Staff authorized to enter/edit the arrival medication list include: Licensed/registered nurses, nurse practitioners, attending physicians, fellow/resident physicians, medical students, pharmacists, and physician assistants. For patients arriving in the Emergency Department, the arrival medication list was entered into the centralized data system. For inpatients, the arrival medication list was likewise entered into the centralized data management system. For patients admitted from home, the arrival medication list was compiled from the medications the patient was taking at home. For patients transferred from another facility, the arrival medication list comprised both medications administered at the transferring facility and the medications the patient was taking at home. The overall process is further augmented by periodic review/reconciliation of medication lists. For the purposes of this study, only the original home medication list (and not the list of medications present on transfer) was used to determine outcome parameters.
Methodology: A detailed chart review was performed, including examination of demographics, pre-injury home medications, co-morbid conditions, injury characteristics (including injury severity score or ISS), Glasgow coma scale (GCS), morbidity, mortality, and 90-day survival. In addition, functional outcome measures at discharge for locomotion, self-feeding, and expression-communication were assessed. ,
Medication assessment and exclusions: The neuro-psychiatric medication (NPM) group encompassed the following medication classes:
The cardiac medication (CM) groups included in this study were:
Due to very small sample sizes, purely anti-arrhythmic agents (i.e., amiodarone), acetylcholinesterase inhibitors (i.e., donepezil), and anti-parkinsonian medications (i.e., levodopa-carbidopa) were excluded from analyses.
Statistical analysis: Data analysis included descriptive statistics, testing for normality, and subsequent group comparison testing. Non-normally distributed variables (ISS, number of injuries) were reported using median, range, inter-quartile range, and analyzed using Kruskal-Wallis test for multi-group comparisons. Normally distributed variables were reported as mean ± SD (or SEM) and tested using analysis-of-variance (ANOVA). Additional co-variate adjustments for age and ISS (ANCOVA) were made when comparing groups for differences in key outcome parameters. Categorical variables were analyzed using Chi-square or Fisher's exact test. Statistical significance was set at alpha = 0.05. Survival estimates were controlled for injury severity score. PASW Statistics 19 (SPSS, Inc., Chicago, IL, USA) software package was utilized.
Demographics: A total of 712 consecutive trauma patients aged 45 years and older were identified. The mean age was 63.5 ± 13.6 years and males represented 56.1% of the patient sample. Median ISS was 8 (range 1-75, interquartile range 4-11). Initial GCS and ISS were not significantly different between age groups [Table 1] and [Table 2]. Detailed demographic information, grouped by pre-injury medication usage, is provided in [Table 1]. Mortality: The overall in-hospital mortality was 5.8% with medication class details shown in [Figure 1]. The highest mortality figures were seen among patients with pre-injury antipsychotic (8.6%) and antiseizure medication (10%) use. The in-hospital mortalities for all NPM monotherapies however did not significantly differ from the mortality of the overall sample.
Use of Neuropsychiatric medication: Neuropsychiatric medication (NPM) use was prevalent in the study group with 34.4% (245/712) of patients taking at least one NPM. The mean number of medications per patient was 4.95 ± 4.51, and the mean number of NPMs was 2.68 ± 1.93 per patient (median, 2). Although total medication use increased with age [Table 2], the use of NPM did not follow the same age related trend [Table 3]. Comorbidity-polypharmacy score (CPS) and mortality increased with advancing age.
Antidepressants were the most common NPMs used (22.2%, 158/712), followed by anxiolytics (15.2%, 108/712), antipsychotics (4.9%, 35/712), and antiepileptics (2.8%, 20/712). The most common cardiac medication groups included beta blockers (11.7%, 83/712), followed by angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (ACEI/ARB, 11.5%, 82/712), and calcium channel blockers (6.4%, 46/712). The incidence of combination NPM therapy was also high in the study group. When grouped by the number of NPMs co-administered, there were no differences in mean age, gender or median ISS. When controlled for ISS there was lower 90-day survival seen among patients taking 1-2 NPMs compared to patients not taking NPMs (95% vs 89%, P = 0.009). This observation was even more pronounced in patients taking 3 or more NPMs [95% vs 81%, P = 0.01, [Figure 2].
Sub group analysis: Because of the high incidence of co-administration of NPMs and CMs, combination therapies involving these two broad medication groups were analyzed. The most common combinations of CMs and NPMs included beta-blockers/antidepressants (14%), followed by angiotensin system modulator/antidepressant combinations (12%). When compared with monotherapy consisting of either CMs or NPMs, several combination therapy regimens were associated with significantly higher in-hospital mortality [Figure 3]. Beta blocker plus antidepressant combinations were associated with a mortality of 14.7% compared with 7.0% and 4.82% for antidepressant or beta-blocker monotherapy respectively (P < 0.05 for both). The combination of ACE-I/antidepressant was associated with greater mortality when compared to ACE-I monotherapy (13.5% vs. 4.9%, P = 0.01). Likewise, ACE-I/anxiolytic combination therapy was associated with greater mortality when compared to ACE-I monotherapy (11.5% vs. 4.9%, P = 0.04).
When compared to no pharmacologic therapy, the effect of combination therapy of cardiac medications and NPMs was even more pronounced. Combinations of beta blockers and antipsychotics had a mortality of 25.0% compared to 3.8% for patients not on any CMs or NPMs (P < 0.002). Additional combinations with significant increases in mortality compared to patients not on any CMs or NPMs included beta blocker-antidepressant combinations (14.7%, P < 0.01), calcium channel blocker-anxiolytic (18.2%, P<0.02), calcium channel blocker-antidepressant (14.3%, P = 0.02), ACE inhibitor-anxiolytic (14.3%, P = 0.02), and ACE inhibitor-antidepressant (16.1%, P < 0.01) [Figure 3]. Combinations with antiseizure medications were excluded from analysis due to the small sample size.
Length of stay, discharge destination, and functional outcomes: There were no statistically significant differences among NPM or CM monotherapy groups in terms of SICU lengths of stay or total hospital lengths of stay. Among patients who survived to discharge, there were no statistically significant differences in functional outcome measures (i.e., self-feeding, expression-communication, locomotion) between NPM and cardiac monotherapy groups, except for beta blocker and ACE-I/ARB monotherapy groups. In both of these groups, statistically lower self-feeding scores (3.74 vs 3.91 for beta blockers, 3.75 vs 3.90 for ACE-I/ARB, both P < 0.01) were noted, but the difference does not seem significant in clinical terms. Patients taking antidepressants were significantly more likely to require discharge to an institution (70/148 or 47.3% versus 199/535 or 37.2%, P < 0.03). When compared to patients receiving specific CMs to those not receiving those particular agents, the use of beta blockers (95/186 or 51.1% versus 174/499 or 34.9%, P < 0.01), calcium channel blockers (47/97 or 48.5% versus 222/588 or 37.8%, P < 0.05), and ACE-I/ARB (52.2% or 94/180 versus 34.7% or 175/505, P < 0.01) were all significantly associated with discharge to a long-term care/rehabilitation facility. No differences in discharge destination were observed for patients taking other NPM (anti-seizure, anti-psychotic, anxiolytic) or CM monotherapies. Among combination therapy regimens, synergistic effect was present when antidepressants were co-administered with beta blockers (53.7% discharged to institution for combined therapy, 47.3% for monotherapy, and 32.0% for neither, P < 0.01) and with ACE-I/ARB (56.5% for combination therapy, 47.5% for monotherapy, and 32.6% for neither, P < 0.01). Similar effect was observed for combined antipsychotic and ACE-I/ARB therapy (55% for combined therapy, 50% for monotherapy, and 34.8% for neither, P < 0.01).
Complications: There were no significant differences in incidence of complications for patients taking NPM or CM monotherapy compared to patients not taking the corresponding agent for seizure medications, antidepressants, antipsychotics, anxiolytics, and calcium channel blockers (all, P > 0.05). Patients on pre-hospital beta blocker and ACE-I/ARB monotherapies had greater overall incidence of complications than patients not taking these agents. For the beta blocker group, incidence of any complication was 37.2% versus 24.0% for the non-beta blocker group (P < 0.01). In the ACE-I/ARB group, 35.3% of patients experienced at least one complication compared to 23.2% in the non-ACE-I/ARB group (P < 0.01). In terms of mean number of complications per patient, there were no significant differences when comparing monotherapy groups with respective controls for calcium channel blockers, antiseizure, anxiolytics, and antipsychotics (all, P > 0.05). Patients taking antidepressants had, on average, 0.91 complications during their hospitalization compared to 0.51 complications among the non-antidepressant group (P < 0.01). In the ACE-I/ARB group, there were 0.81 mean complications per patient, versus 0.52 complications in the non-ACE-I/ARB group (P < 0.01). Likewise, patients in the beta blocker group had an average of 0.94 complications compared to 0.46 complications in the non-beta blocker group (P < 0.01). Significantly greater morbidity was noted for combinations of antidepressants and beta blockers (28/60 ; 47%) when compared to either antidepressants or beta blocker monotherapy (76/236 or 32%, P < 0.05). No other statistically significant interactions were observed in terms of patient morbidity when examining combination NPM-CM pharmacotherapies.
As the world's population grows older, the management of chronic health conditions continues to evolve. Inherent to pharmacotherapy of elderly is polypharmacy, or the co-administration of multiple medical regimens that facilitate management of the associated medical conditions. , Along with the success of pharmacologic interventions in chronic illness come the added risks.  The present study identified several medication co-administration patterns that appear to be associated with increased morbidity and mortality in older trauma patients.
It is well established that some of the most common drug-drug combinations can have significant unintended side effects.  Specifically, the two medication classes examined in this study have the potential to adversely interact when co-administered.  Following traumatic injury, numerous acute hard-wired processes contribute to multi-faceted physiologic and psychological patient responses. , Various medications are known to alter/modulate these physiologic processes at various levels and to different degrees. ,,, Beta blockers, for example, may suppress the natural catecholamine response associated with trauma.  In patients with brain injury, beta-blocker use was associated with lower mortality (5.1%) when compared to patients who were not exposed to beta blockers (10.8%) despite the fact that patients taking beta-blockers were older and had greater injury severity scores. 
Although existing literature indicates that beta-blocker therapy may be associated with lower mortality in trauma, our data suggests that when co-administered with certain neuropsychiatric medications this trend may actually be reversed. Previous studies have demonstrated that selective-serotonin reuptake inhibitors (SSRI) decrease the metabolism of beta blockers thus resulting in higher serum beta blocker concentrations. , In addition, newer antidepressants, including SSRI and SNRI (serotonin-norepinephrine reuptake inhibitors), have been shown to have clinically significant hemodynamic effects, albeit less than the older tricyclic antidepressants.  While the clinical consequences of these side effects may not be life-threatening in an otherwise healthy patient, little is known about such interactions in the setting of acute traumatic injury. Although the observation of increased mortality in patients simultaneously taking certain NPM-CM combinations [Figure 3] is largely phenomenologic in nature, it is important nevertheless because these two medication classes represent some of the most commonly used pharmaceutical agents in our study population.
The higher mortality rate in patients taking antipsychotics in combination with both beta blockers and ACE-I/ARBs is likely to be multifactorial. Previous research has shown an increased mortality and morbidity in patients with dementia taking antipsychotics. ,, However, few direct interactions have been identified between most antipsychotics and beta blockers or ACE-I/ARBs. A review of the physiologic effects of many of the newer antipsychotics shows that their use may have associated hemodynamic side effects.  In fact, some atypical antipsychotics have significant alpha-1 adrenergic-blocking properties  which when combined with the antihypertensive effects of beta blockers or ACE-I/ARBs in the setting of trauma may contribute to unintended hemodynamic changes.
The association between co-administration of calcium channel blocker/anxiolytic and increased mortality is also likely multifactorial. Previous studies show that non-dihydropyridine calcium channel blockers decrease the elimination of benzodiazepines that are metabolized by oxidation and can significantly prolong their effects. ,, In addition, a relationship between benzodiazepine use and increased fall risk in elderly patients is well described and may be associated with orthostatic hypotension.  Moreover, multiple studies have shown that the combination of anxiolytics with antihypertensives such as dihydropyridine calcium channel blockers and ACE-ARBs may unfavorably alter hemodynamic status and result in physiologic perturbations poorly tolerated by acute trauma patients. , Due to the complexity of these interactions , we can only speculate as to the exact mechanism or mechanisms that result in higher mortality in this population.
Another factor to consider is the still poorly defined or understood phenomenon of "polypharmacy withdrawal" on mortality in our patient population.  Our study is limited by the fact that only pre-injury medications were studied, without elucidating whether those medications were continued during the patients' hospital stays. Previous studies have shown that withdrawal from home therapy of both cardiac and neuropsychiatric drugs worsens mortality in surgical patients, and that this may be further potentiated by withdrawal from multiple agents. , Although every effort is made to restart home pharmacotherapy, certain medications may be discontinued for a variety of reasons, whether intentionally or otherwise. This situation may be further compounded by a patient's inability to take oral medications, especially when no acceptable intravenous substitutions are available. An additional factor that may influence patient outcomes in this context is the risk inherent to converting an oral agent to another oral or intravenous agent of equivalent potency during acute hospitalization.
It should be noted that several psychiatric medications can induce adverse reactions that have significant cardiac implications. For example, the serotonin syndrome  and the neuroleptic malignant syndrome  are rare side effects of several of the psychiatric agent classes examined in our study. Both syndromes can produce wide hemodynamic fluctuations and are potentially life-threatening on their own. Because of their idiosyncratic nature, analyzing the interaction between these conditions and the cardiac medications covered in our study would be very difficult. Unfortunately, our study was not designed to identify either of these conditions in our population, but this is certainly an area that merits further study in the future.
This study has several important limitations. The small number of patients taking certain classes of medications in our sample limited the amount of subgroup analysis we could perform. Further sampling would likely help our analysis, for example, by looking at conventional and atypical antipsychotics separately. Also, our analysis did not specifically include other pre-existing medical conditions, making it difficult to assess their impact on patient mortality. Additionally, we did not perform multiple comparison analysis on the data so as to include as many significant relationships as possible. This may result is erroneous identification of significant drug-drug interactions. Many of these relationships were highly significant (P< 0.01) and may withstand the increased scrutiny of multiple comparison analysis; however, the study design of this initial manuscript was meant to be as inclusive as possible so as to identify all possible relationships for further study. The small sample sizes seen in combination therapies also limited the ability to perform multivariate analysis as this study is underpowered to make statistical conclusions from those tests. This study should however provide a starting point for further research into this area.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3]