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


This study enrolled 150 patients: 49 with a UDS positive for cocaine and/or amphetamines, and 101 with a UDS negative for these substances (Figure). Enrolled patients were equally distributed in each group over time, thus they were subjected to similar sedation strategies. Baseline characteristics are reported in Table 1. In the stimulant-positive group, 10 patients (20.4%) were positive for amphetamines, 39 patients (79.6%) were positive for cocaine, and 2 patients (4.1%) were positive for both substances on the UDS. Of note, 2 of the 10 amphetamine-positive patients (20%) had amphetamine/dextroamphetamine documented as a home medication: one was admitted because of a fall after an overdose of opioids and benzodiazepines, and the second after a motor vehicle collision that involved alcohol intoxication. The nature of stimulant use was evaluated by provider review and social work progress notes documenting patient- and family-reported preinjury use. Binge use was noted in 7 patients (14.2%), chronic ingestion for 23 patients (46.9%), and 19 patients (38.8%) denied use or had an unknown history of use.


Subject selection.

An opioid was administered to 148 patients (98.7%) in the cohort. Two patients who did not receive an opioid were stimulant positive on admission, met inclusion criteria and were successfully extubated, thus they were included in the analysis. The use of a fentanyl infusion, a hydromorphone infusion, or the sequential administration of a fentanyl and hydromorphone infusion were as follows: 57.1% vs 64.4%, P = .47; 14.3% vs 5%, P = .06; and 18.4% vs 19.8%, P = 1.0, respectively. Oxycodone and methadone use in the stimulant-positive compared with the stimulant-negative arm were 6.1% vs 8.9%, P = .75 and 4.1% vs 1%, P = .25, respectively. The median number of IV opioid bolus doses per day of mechanical ventilation was 2.82 (interquartile range [IQR], 0.94–4.99) in the stimulant-positive arm and 3.07 (IQR, 1.5–5.48) in the stimulant-negative arm (P = .36). The primary end point and median daily opioid requirements, expressed as morphine equivalents, are reported in Table 2. Different pain scales were used during the study period and are not comparable, thus these data were not recorded.

Sedative administration and assessments are reported in Table 2. Propofol was the most commonly used sedative, followed by benzodiazepines and dexmedetomidine. For patients receiving a benzodiazepine infusion, midazolam infusions were used most commonly (18.4% vs 23.8%, P = .53), followed by lorazepam infusions (6.1% vs 10.9%, P = .55) or the sequential use of a midazolam and lorazepam infusion (4.1% vs 5.9%, P = 1.0).

To further characterize the medication requirements for these patients, administration of adjunctive sedative and analgesic medications was also collected (Table 3). For all 9 patients who received ketamine, ketamine was initiated in the setting of multimodal analgesia for uncontrolled pain despite traditional IV opioid therapy and was initiated between days 1 and 3 of intubation. All patients had lidocaine transdermal patches placed, 8 received acetaminophen, 3 received nonsteroidal anti-inflammatory drugs, 4 received adjuncts for neuropathic pain, and 2 patients had epidural analgesia. One patient in the stimulant-positive arm had documentation that ketamine was initiated specifically because of a history of polysubstance abuse. The number of visits to the operating room during mechanical ventilation was a median of 1 in each group. Clinical outcomes including the duration of mechanical ventilation, ICU length of stay, hospital length of stay, and in-hospital mortality are reported in Table 4.

Self-extubation was reported for 1 patient in each arm (2.04% vs 0.99%, P = .55), and extubation requiring reintubation within the following 24 hours was not different between groups (6.12% vs 5.94%, P = 1.0). The proportion of patients who received a tracheostomy and the time to tracheostomy were 22.5% vs 27.7%, P = .55 and 5.33 vs 6.84 days, P = .46, respectively. Two patients (18.1%) in the stimulant-positive arm and 6 patients (21.4%) in the stimulant-negative arm received an early tracheostomy (day 2 or 3) because of facial fractures or spinal cord injury. The median Injury Severity Score for patients who received a tracheostomy was 25 (IQR, 22–34) vs 24.5 (IQR, 14.8–37.3, P = .46), and the majority presented with TBI (72.7% vs 57.1%, P = .47), with a median Glasgow Coma Scale score in each group of 7.

Predictors of opioid requirements, defined as greater than the median of 1.24 mg/kg/day of morphine equivalents, were evaluated using a multivariate logistic regression analysis with backward variable selection. In the final model, increasing age and AIS (head and neck) were associated with decreased daily opioid requirements: OR, 0.95 (95% CI, 0.93–0.97) and OR, 0.71 (95% CI, 0.58–0.87), respectively, and preinjury stimulant use was not predictive of opioid requirements (OR, 0.88, 95% CI, 0.40–1.90).

In a second multivariate analysis of a propensity score--adjusted model, the same predictors as those used in the above logistic regression were used to first predict stimulant use. The predicted scores from this model were then used to adjust the stimulant use and predict opioid requirements. In this propensity score--adjusted model, opioid requirements were not significantly affected by preinjury stimulant use (OR, 0.97, 95% CI, 0.44–2.11). In a third logistic model adjusted by quintiles of propensity scores, stimulant use was not a significant predictor of opioid requirements (OR, 0.89, 95% CI, 0.41–1.94).

A post hoc analysis of patients presenting with TBI was performed to determine additional factors influencing the primary and secondary outcomes of the cohort. Baseline characteristics and opioid and propofol requirements in this subgroup are reported in Table 5.