Adjusted vs Total Body Weight–Based Dosing of Sedation and Analgesia Used in the Intensive Care Unit

Abstract

Background: The purpose of this study was to evaluate if dosing fentanyl, dexmedetomidine, and propofol based on ideal or adjusted vs actual weight in patients would decrease overall opioid and sedative use. Methods: This was a retrospective chart review comparing adjusted vs actual weight-based dosing protocol of mechanically ventilated (MV) intensive care unit (ICU) adult patients who required fentanyl and either propofol or dexmedetomidine. Results: A total of 261 patients were included in which 101 patients were in the actual weight group and 160 patients were in the adjusted weight group. Total doses per MV day of fentanyl was 1042 ± 1060 µg in the actual weight group vs 901 ± 1025 µg in the adjusted weight group (P = .13). Total doses per MV day of midazolam was 20 ± 19 mg in the actual group vs 15 ± 19 mg adjusted group (P = .02). Average MV days was 8.2 vs 7.1 days, ICU length of stay was 10.6 vs 9.4 days, and self-extubation rates were 17.8% vs 4.4% in the actual group and adjusted group, respectively. Conclusion: Total midazolam doses per MV day were lower in the adjusted group. No significant change was seen in MV days, ICU length of stay, or self-extubation rates.

Background

Propofol and dexmedetomidine are sedative agents used in the intensive care unit (ICU) for patients requiring mechanical ventilation along with fentanyl that has both analgesic and sedative properties.¹ The 2013 and 2018 guidelines of the management of Pain, Agitation, and Delirium suggest using nonbenzodiazepine sedatives such as propofol and dexmedetomidine to improve clinical outcomes of decreasing mechanical ventilation (MV) time and ICU days.^1,2 It has also been shown decreasing overall sedation by implementing daily interruption of sedative infusions decreased duration of MV and ICU length of stay (LOS).³ However, there is a lack of data regarding patient’s weight that may play an impact on dosing of sedation. As the rate of obesity increases across the nation to 31.2%, West Virginia remains to be the highest-ranking state at 35.1% of obese adults with body mass index (BMI) ≥ 30 in the United States based on 2016 data.⁴ Previous studies supporting these sedative and analgesic agents used in the ICU evaluated patients being nonobese (ie, 70-80 kg) which presents an issue when providing obese patients with adequate comfort but limiting oversedation practices.^5-9

In response to the increasing prevalence of obesity at our institution, a sedation protocol was implemented to dose fentanyl, propofol, and dexmedetomidine in MV patients based on an ideal body weight (IBW) or adjusted body weight, while midazolam remained standard as milligrams per hour.

Methods

This study was a retrospective chart review of all obese and nonobese patients admitted to the ICU (medical ICU or surgical ICU), who required fentanyl along with other sedatives (propofol, dexmedetomidine, or midazolam) from 2012 to 2016 at West Virginia University Medicine Ruby Memorial Hospital. Patients were included if they required MV for more than 48 hours and were greater than 18 years of age. Exclusion criteria included patients admitted to the cardiac ICU or cardiothoracic ICU units (units were not under adjusted weight protocol), patients who required deep sedation (status epilepticus, therapeutic coma, polysubstance abuse, hypothermia therapy, or paralytic therapy), and those under the alcohol withdrawal protocol.

The primary outcome was to measure total doses of continuous infusions of analgesic and sedatives given throughout MV time and ICU stay. Secondary outcomes included total MV time and ICU LOS. Dosing of these agents at our institution ranged and were weight-based as follows 0.5 to 3 µg/kg/h, 5 to 50 µg/kg/min, and 0.1 to 1.5 µg/kg/h for fentanyl, propofol, and dexmedetomidine, respectively. The adjusted weight-based protocol was implemented in 2014 and therefore patients admitted from January 2012 until December 2013 were dosed on actual body weight and were considered the actual weight group. Adjusted weight was calculated by IBW—calculated as 50 + (2.3 × Inches Over 5 ft Height) for males and 45.5 + (2.3 × Inches Over 5 ft Height) for females, subtracted from actual or total body weight (TBW), multiplied by 0.4, and added to IBW (IBW + (0.4 × [TBW – IBW])). If TBW was greater than 120% of IBW, then adjusted body weight would be used.¹⁰ This calculation was available on ordering of sedatives in the EPIC computer system. This implementation was in conjunction with an analgosedation nurse titration protocol at the institution. Patients dosed on either IBW or adjusted body weight between January 2014 and October 2016 were included in the adjusted weight protocol group. Dosing sedation with use of IBW or adjusted body weight could be left to the discretion of the prescriber.

Parameters to evaluate level of sedation included obtaining average Riker Sedation-Agitation Scales (SAS) which was used prior to implementation of body weight dosing protocol along with Richmond Agitation-Sedation Scale (RASS) used primarily after protocol implementation. Riker Sedation-Agitation Scale evaluates level of sedation and agitation in a 1 to 7 scoring tool with a score of 4 representing calm, awakens easily, and follows command. Scores that are greater than 4 in the SAS scale represent agitation and less than 4 represent sedation and unable to arouse.^11-13 The RASS is a 10-point scale with 0 representing calm and alert, +1 to +4 representing combative state, and −1 to −5 representing unarousable state^14,15; SAS or RASS scores were evaluated by the reviewers of this study and averaging highest and lowest recorded sedation scores were collected while patients were receiving analgosedation. Self-extubation rates were also collected by way of note documentation when available to further assess level of sedation. Other data points included hospital admission diagnosis, oral benzodiazepine use, duration and dose of continuous midazolam infusions, and Pao₂/Fio₂ (partial arterial oxygen/fraction of inspired oxygen) ratio to assess severity of sickness and pulmonary dysfunction. Wilcoxon 2-sample test was performed for continuous nonparametric data and a chi-square test was used for categorical nominal data. A P value of <.05 was determined to be statistically significant. Descriptive statistics was applied to all other data. Power analysis could not be conducted given large standard deviation within the sample size of dosing ranges of each drug.

Results

A total of 833 patients were screened in which 261 were included in the evaluation. Most patients were excluded due to mechanical ventilation time being less than 48 hours. Of the 261 patients included, a total of 101 patients were included in the actual group and 160 patients in the adjusted group. Baseline demographics are provided in Table 1 with average weight as 87.5 ± 22.7 kg in the actual group and 96.4 ± 32.6 kg in the adjusted group. Body mass index was 29.6 ± 7.6 kg/m² in the actual group and 30.7 ± 10.7 kg/m² in the adjusted group.

Table 1.

Baseline Demographics.

	Actual group (n = 101)	Adjusted group (n = 160)
Age (average years + SD)	53.7 ± 15.7	55.4 ± 15.8
Male, No. (%)	61 (60.4)	97 (60.6)
Weight (average kg + SD)	87.5 ± 22.7	96.4 ± 32.6
BMI (average kg/m2 + SD)	29.6 ± 7.6	30.7 ± 10.7
Pao2/Fio2 ratio (average + SD)	141 ± 87	161 ± 95

Note. Table represents the average age in years plus standard deviation, number of males and percentage, average weight in kilograms plus standard deviation, average body mass index (BMI) in kilograms per meter squared plus standard deviation and the average partial arterial oxygen (Pao₂) and fraction of inspired oxygen (Fio₂) ratio plus stand deviation.

Patients’ admission diagnosis was included in the analysis and further characterized by a broadened classifications including hemorrhage, infection, oncology, respiratory, trauma, surgery, neurology, or other causes included in Figure 1.

Figure 1.

Admission diagnosis. Hemorrhage: gastrointestinal bleed, cerebral bleeds; infection: abscess, cellulitis, febrile neutropenia, meningitis, osteomyelitis, pneumonia, sepsis, urosepsis; oncology; respiratory: pulmonary embolism, pneumothorax, respiratory distress, trauma; surgery: abdominal surgery; neurology: altered mental status, seizures, stroke; other: angioedema, cardiac arrest, cirrhosis, diabetic ketoacidosis, hyperkalemia, pancreatitis, congestive heart failure exacerbation.

In reference to the primary outcome, doses of medication administered to patients were not significantly different as seen in Table 2. Total fentanyl dose given per MV day was 1042 ± 1060 µg in the actual group compared with 901 ± 1025 µg in the adjusted group (P = .1312). Total propofol dose received was 935 ± 906 mg per MV day in the actual group (n = 94) compared with 1129 ± 1214 mg per MV day in the adjusted group (n = 116; P = .7817). Total dexmedetomidine dose received was 329 ± 348 µg per MV day in the actual group (n = 33) compared with 341 ± 508 µg per MV day in the adjusted group (n = 92; P = .3933).

Table 2.

Total Analgesic/Sedative.

Drug (total admission)	Actual group (n = 101)	Adjusted group (n = 160)	P value
Fentanyl (µg/d ± SD)	1042 ± 1060	901 ± 1025	.1312
Propofol (mg/d ± SD)	935 ± 906	1129 ± 1214	.7817
Dexmedetomidine (µg/d ± SD)	329 ± 348	341 ± 508	.3933

Note. Total drug given throughout hospital admission.

Midazolam infusions were used by 49 (49%) in the actual group and 57 (36%) in the adjusted group as seeen in Table 3. Total midazolam dose received throughout hospital admission was 20 ± 19 mg per MV day in the actual group and 15 ± 19 mg per MV day in the adjusted group (P = .0230) despite both groups having similar average number of days on these infusions—4.0 days (interquartile range [IQR] = 2-5) vs 3.9 days (IQR = 1-6) in the actual vs adjusted group, respectively. Percentage of oral benzodiazepine given for further sedation was 23% in the actual group and 10% in the adjusted group.

Table 3.

Benzodiazepine Use.

	Actual group (n = 101)	Adjusted group (n = 160)	P value
Midazolam infusions, No. (%)	49 (49)	57 (36)	—
Total Midazolam dose (mg/d ± SD)	20 ± 19	15 ± 19	.0230
Days on midazolam infusion (IQR)	4.0 (2-5)	3.9 (1-6)	—
Oral Benzodiazepine use, No. (%)	23 (23)	16 (10)	—

Note. Benzodiazepine use between pre- and postgroup as number of midazolam infusions, total midazolam doses given throughout hospital stay, number of days patients receiving midazolam infusions, and if oral benzodiazepines were given for sedation purposes. IQR = interquartile range.

Figure 2 provides additional secondary endpoints of ICU LOS and MV days throughout the hospital course. Intensive care unit LOS was 10.6 days (IQR = 6-14 days) in the actual group vs 9.4 days (IQR = 5-12 days) in the adjusted group (P = .1241). Mechanical ventilation days was 8.2 (IQR = 4-10 days) in the actual group and 7.1 (IQR = 4-9 days) in the adjusted group (P = .1435).

Figure 2.

ICU LOS and MV days.

Note. ICU = intensive care unit; LOS = length of stay; MV = mechanical ventilation.

Sedation scoring with use of SAS and RASS were similar between groups (average SAS score range 2.1–4.5 vs 2.2–4.8, average RASS score range −3.1 to 0.8 vs −2.2 to 0.6 in the actual group and adjusted group, respectively) as seen in Table 4. Self-extubation rates were higher in the actual group as compared with the adjusted group (18 vs 7 days).

Table 4.

Sedation Scores and Self-Extubation Rates.

	Actual group	Adjusted group
Average SAS Score range	2.1 to 4.5	2.2 to 4.8
Average RASS Score range	−3.1 to 0.8	−2.2 to 0.6
Self-extubation rates, No. (%)	18 (17.8)	7 (4.3)

Note. SAS = Riker Sedation-Agitation Scale; RASS = Richmond Agitation-Sedation Scale.

Discussion

The management of sedation and agitation has progressed throughout the past few decades with the discovery of neuropsychological sequelae in patients associated with higher sedation and longer LOS in the ICU. Therefore, current sedation-agitation guidelines recommend light sedation practices and avoiding benzodiazepines in hopes to decrease rates of ICU-delirium in our patients.^1,2

Pharmacokinetic (PK) evaluation of propofol found the volume of distribution can drastically affect sedation in morbidly obese patients based on the lipophilic nature of the drug that may cause oversedation and risk of drug accumulation.⁷ Cortinez et al evaluated blood samples from 20 obese patients receiving propofol during elective laparoscopic bariatric surgery. The average propofol dose used was 6.53 ± 1.11 mg/kg/h to achieve 4 µg/mL plasma concentrations and a bispectral index (BIS) was measured to assess hypnotic effect (goal 40-60). Total body weight and adjusted body weight models were performed to assess PK and pharmacodynamic (PD) parameters and when using adjusted body weight dosing, a reduction in predictive errors and improved PK and PD performances in this patient population was found.^16-18

Dexmedetomidine has given practitioners the flexibility of not only sedating patients while on MV but also continuing this agent during extubation to ease agitation without compromising respiratory status. Dexmedetomidine has limited PK data but has been shown to have highly lipophilic properties for rapid absorption in cerebrospinal fluid.¹⁹ Based on its lipophilic properties, one can infer that the drug’s volume of distribution may also be impacted by obesity. Tufanogullari et al reviewed 3 different dosing strategies (saline/placebo, 0.2 µg/kg/h, 0.4 µg/kg/h, and 0.8 µg/kg/h) in 80 morbidly obese patients undergoing laparoscopic bariatric surgery. As a result, patients receiving the lower dosing range (0.2 µg/kg/h) had earlier recovery while minimizing adverse perioperative cardiovascular side effects. Postoperative pain management was similar among treatment groups which may give insight that lower doses kept patients sedated without compromising pain and comfort.²⁰

Fentanyl is one of the most common opioids used for analgesia in the ICU due to its sedative properties, lesser hypotension risk compared with morphine, and quick onset/offset duration of action. Fentanyl is highly lipophilic with rapid and extensive redistribution into highly vascular tissues such as muscle and adipose.²¹ A derived dosing weight strategy for fentanyl was reported by Shibutani et al as a “pharmacokinetic mass (PK mass)” which was highly correlated to lean body mass. Authors found that plasma fentanyl concentration exponentially increases as weight increases when comparing fentanyl blood samples in 37 normal weight patients (BMI < 30) and 33 obese patients undergoing surgery.²²

As the available evidence provides some insight into PK and PD parameters of obese patients receiving fentanyl, propofol, and dexmedetomidine, this data only reflects the intraoperative settings and not prolonged infusions for >24 hours in ICU patients. This study evaluates MV patients receiving continuous sedative and analgesic infusions for at least 48 hours dosed based on IBW/adjusted body weight compared with actual body weight. Midazolam dose was found to be statistically significantly lower in the adjusted group vs actual group. Dexmedetomidine and fentanyl total doses appeared lowered in the adjusted group and propofol use increased in the adjusted group when compared with the actual group, but this was not found to be statistically significant. It is difficult to discern whether patients truly required less midazolam infusions or if there was an inherent decline in their use based off the recommendations from the 2013 Peripheral Artery Disease (PAD) Guidelines.¹

Intensive care unit LOS and MV days was were not found to be statistically significant between the groups. Self-extubation rates were decreased in the adjusted group; however, the retrospective nature of this evaluation makes this data difficult to assess as rates were reported in notes. Richmond Agitation-Sedation Scale and SAS evaluations were assessed by collecting the average highest and lowest RASS/SAS score while being sedated in the ICU. It should be noted that the average lowest RASS score was lower in the actual group giving rise to the possibility that those patients remained more sedated due to increased midazolam dosing. However, this point of reference again was difficult to assess by retrospective chart review. Statistical tests were not evaluated on these endpoints, therefore cannot conclude actual differences.

Limitations of this study include the retrospective nature of the study. This was a single center review at an academic medical center and delirium assessment was not included due to limited validated evaluation and difficulty with chart review process. The evaluation of daily sedation discontinuation (drug holiday) and awakening trials was inconsistent and therefore not collected in this review. As Acute Physiology and Chronic Health Evaluation (APACHE) II scores or Sequential Organ Failure Assessment (SOFA) scores were not evaluated, the only reference to severity of disease was left to the P:F ratio.

In conclusion, it was found that dosing fentanyl, propofol, and dexmedetomidine on IBW or adjusted body weight resulted in less overall benzodiazepine usage. Despite the changes in dosing and increased propofol use, there remained to be a trend in shorter ICU stay and MV length.