Analgesia and Sedation Requirements in Mechanically Ventilated Trauma Patients With Acute, Preinjury Use of Cocaine and/or Amphetamines

Bridgette Kram, PharmD; Shawn J. Kram, PharmD; Michelle L. Sharpe, PharmD; Michael L. James, MD; Maragatha Kuchibhatla, PhD; Mark L. Shapiro, MD


Anesth Analg. 2017;124(3):782-788. 

In This Article


The study protocol was reviewed and approved by the institutional review board at Duke University Hospital with a waiver of informed consent. This retrospective cohort study identified patients from the Duke University Hospital Trauma Registry admitted between January 1, 2009, and June 15, 2013. Inclusion criteria were ≥16 years of age, an admitting diagnosis of trauma, admission to an adult ICU, a UDS documented within 24 hours of admission, and the need for mechanical ventilation for ≥24 hours. Trauma patients ≥16 years of age were included in this study because 16- and 17-year-old trauma patients are admitted and treated in an adult ICU at our institution. Patients were excluded if they were pregnant, received targeted temperature management, received continuous neuromuscular blockade, or did not receive any opioid analgesic or sedative because of declaration of brain death or withdrawal of care. Patients with long-acting opioids documented as a home medication were also excluded because they are expected to require higher analgesic doses due to tolerance and would confound the primary outcome. Long-acting opioids included fentanyl transdermal patch, hydromorphone extended release, methadone, morphine extended release, oxycodone extended release, and oxymorphone extended release.

Institutional policy mandates that a UDS is obtained for all level 1 and level 2 trauma patients and is optional for level 3 trauma admissions. For the purposes of this study, stimulant use was defined as having a positive UDS for cocaine and/or amphetamines. All patients with a UDS positive for stimulants, regardless of other positive results and who met the study criteria, were included in the stimulant-positive arm. The stimulant-negative group consisted of anyone meeting study eligibility who did not have a positive result for a stimulant, regardless of positive results for other agents. Because a larger sample of stimulant-negative patients was available in the registry, patients were organized chronologically by admission date, and every fifth patient was evaluated for enrollment eligibility to provide an even distribution of patients admitted during the study period.

The primary end point was the median daily dose of opioid analgesia required during mechanical ventilation, expressed in morphine equivalents. Secondary end points were: the median daily dose and infusion rate of propofol; the median daily dose and infusion rate of dexmedetomidine; the median benzodiazepine daily dose, expressed in lorazepam equivalents; duration of mechanical ventilation; ICU length of stay; hospital length of stay; and in-hospital mortality. Duration of mechanical ventilation was defined as the duration of endotracheal intubation until first extubation or tracheostomy. This end point was truncated at time of tracheostomy because sedative duration may be reduced for patients after the removal of the endotracheal tube and placement of the tracheostomy.[16]

During the study period, both pain and depth of sedation were assessed every 4 hours or within 30 to 60 minutes after intervention. The Numeric Rating Scale was used for those patients who could self-report pain. For patients unable to self-report pain, Face, Legs, Activity, Cry, Consolability[17] and Wong-Baker Faces or Pain Assessment in Advanced Dementia[18] were used at different points during the study period. Depth of sedation was evaluated using the Richmond Agitation Sedation Scale (RASS).

During the study period, a formal sedation protocol was not in place. Analgesic end points were defined by nursing policy (goal Numeric Rating Scale <4). Initial analgesic orders were placed at the discretion of the provider, and opioid infusions were initiated and titrated by provider order only based on pain assessment and intermittent opioid administration pursuant to nursing policy. Fentanyl and hydromorphone were the most commonly used IV opioids. RASS goal was specified in the administration instructions within the individual order for propofol and dexmedetomidine (default RASS goal -2) and titrated by bedside nurses. Benzodiazepine infusions were initiated and titrated by provider order only in response to RASS assessment and intermittent benzodiazepine use. The surgical ICU is a semi-enclosed unit staffed by: an attending physician (anesthesiology intensivist or trauma surgeon, 24 hour per day coverage); anesthesiology and/or surgical critical care fellows; anesthesiology and/or general surgery residents; advanced practice providers; and allied health staff, including a pharmacist, dietitian, respiratory therapist(s), and physical therapist(s). Multidisciplinary rounds occur 3 times each day, with routine evaluation of analgesic and sedative goals and requirements. Providers and nursing staff were aware of UDS results at the time of treatment.

Statistical Analysis

Based on the knowledge of patient availability, the study initially sought to collect information on approximately 50 patients in the stimulant-positive group. For the primary analysis, assuming 30% opioid dosage variability, a total sample size of 159 subjects (1:2) was needed to detect a difference of 15% in daily opiate dose with 80% power at a significance level of .05. In the absence of previously reported studies in a similar patient population, these values were selected by the authors to be clinically meaningful. The Wilcoxon rank sum test was used for continuous, nonparametric data, and Fisher exact test was used for categorical data where appropriate. Data are reported without correction for false-discovery rate.

A multivariate logistic regression with backward variable selection (P < .4) was used to predict increased opioid requirements, defined as greater than the median of 1.24 mg/kg/day. Variables evaluated in the regression analysis were: age; gender; Injury Severity Score; mechanism of injury (gunshot wound), a UDS positive for stimulants, a UDS positive for cannabinoids, and a positive blood alcohol level; Glasgow Coma Scale score; Abbreviated Injury Scale (AIS; head and neck) score; AIS (spine) score; and traumatic brain injury (TBI). In addition, a propensity score--adjusted model (adjusted for all of the above variables) was used to determine the significance of a UDS positive for stimulants on opioid requirements, expressed as morphine equivalents greater than the median of the cohort. First, a logistic regression was used to obtain probability predicted scores (propensity) of the above variables to predict a UDS positive for stimulants. In a separate logistic regression, a propensity score-adjusted model in which the propensity score was used as a covariate to examine whether a UDS positive for stimulants predicted opioid requirements equivalent to or greater than the median of the cohort. Model assumptions were examined, and goodness of fit was determined. Full details of the statistical model are available in the Supplemental Digital Content ( All analyses were conducted using JMP version 11 and SAS 9.4 (SAS Corporation, Cary, NC).