Prescribing Patterns of Continuous Infusions in Nonobese versus Obese Children Admitted to the Pediatric Intensive Care Unit

Abstract

This retrospective study compared the continuous infusions prescribed for obese and nonobese children. Ninety-five (13.2%) received an infusion. A greater percentage of obese ( n  = 42/168) versus nonobese (53/552) children received infusions, p  < 0.01. No difference was noted in the median number of infusions between the obese and nonobese groups, 2 versus 2, p  = 0.975. The top 20 prescribed infusions included ten (50%) for sedation/analgesia or neuromuscular blockade and six (30%) for hemodynamic support. A literature search was performed for these 20 agents to determine pharmacokinetics, pharmacodynamics, and dosing in obese children and revealed six studies evaluating fentanyl ( n  = 2), midazolam ( n  = 1), and propofol ( n  = 3).

Introduction

Recent data from Ogden et al ¹ indicate that 17% of 2- to 19 year old in the United States are obese. However, Johnson et al ² found that approximately 19% of children 2 to 17 years of age admitted to two academic hospitals were obese. Some reports have described altered pharmacokinetics in obese adults indicating that traditional methods of weight-based dosing of medications in obese pediatric patients could result in adverse drug events and/or altered therapeutic concentrations. ^{2 3 4 5} There are a paucity of studies and case reports that have evaluated the pharmacokinetics and appropriate dosing of medications in obese children. ^{6 7 8 9 10 11 12 13 14} Most of these reports have evaluated either single-dose or intermittent administration of medications in overweight and/or obese children. When given as continuous infusion, medications with high lipophilicity and large volume of distribution may accumulate in the adipose tissue, and the redistribution among plasma and tissue compartments after terminating the infusion may differ significantly in obese children depending on the infusion duration. ¹⁵ As a result, the time to target steady-state concentration and apparent half-life of these drugs may be different in overweight and/or obese children. In the pediatric intensive care unit (PICU) setting, continuous infusions are routinely used in the management of critically ill patients. The purpose of this study was to evaluate the percent of obese and nonobese children receiving continuous infusions and provide recommendations regarding dosing and monitoring of the 20 most frequently prescribed (IV) continuous infusion medications received during the study period.

Methods

Study Design and Population

This was a descriptive, retrospective study conducted at a tertiary academic medical center with a 13-bed Medical PICU and a 12-bed cardiovascular PICU. Children 2 to 17 years were included if they were admitted to the PICU on or after January 1, 2011 and discharged before December 31, 2011. For the purpose of this study, nonobese children were classified as those with a body mass index (BMI) < 95th percentile for age and sex; obesity was defined as a BMI ≥ 95th percentile for age and gender, based on the Centers for Disease Control (CDC) and Prevention definition. ¹⁶ Patients were excluded for incomplete records.

Study Objectives and Data Collection

Following institutional review board approval, patients were identified via the electronic medical record, Meditech (Medical Information Technology, Inc., Westwood, Massachusetts, United States). Date of birth, height (cm), weight (kg), and gender were collected to determine the patient's BMI percentile using the CDC Pediatric BMI calculator. ¹⁷ In addition, the date of admission and discharge, race, and all medication regimens received during their admission were recorded.

The primary objective was to identify the number of obese and nonobese children receiving continuous infusions and to identify the 20 most frequently prescribed continuous infusions in obese and nonobese children. The secondary objectives were to compare the type of continuous infusions between groups. In addition, an in-depth literature evaluation was conducted to determine available data on the pharmacokinetic disposition and pharmacodynamic effects of the identified continuous infusions in obese children, if available. In situations in which data was lacking for pediatric patients, the literature search was expanded to include adults.

Relevant articles were identified using MEDLINE (1946 to March 2019), EMBASE (1988 to March 2019), and International Pharmaceutical Abstracts (1970 to March 2019) using the drug name and the additional terms obesity and pharmacokinetics. Results were limited to studies involving children and published in the English language. Published abstracts were not evaluated because they lacked detailed results, and thus, the search was limited to published studies, letters to the editor, and case reports and case series. A two-step selection process was conducted for the published studies. The initial reports were screened by two authors (PNJ and JLM). The final selection process was determined by input from all authors. To be included for analysis, the published report had to evaluate one of the IV continuous infusion agents identified in this study and provide dosing, pharmacodynamics, and/or pharmacokinetic analysis in obese children. For those IV infusions in which no data was noted in children, a second search was conducted using the same criteria but expanded to include adult data. The reference list of articles included was searched for applicable studies. Articles were excluded from review if they evaluated intermittent dosing rather than IV continuous infusions.

Statistical Analysis

Descriptive statistics were used. Categorical data were presented in frequency (percent) and continuous data were presented in median interquartile range (IQR). Pearson chi-squared tests and Wilcoxon rank-sum tests were used to evaluate differences between demographic and clinical characteristics. Rankings of the 20 most frequently prescribed continuous infusions were made; however, no inferential statistics were used to compare differences. Data analysis was conducted using SAS software, Version 9.4 of the SAS System for Windows (SAS Institute Inc., Cary, North Carolina, United States), with the a priori set at p   <  0.05. Statistical analysis was not employed for the data obtained via the literature search for the secondary objective given the expected discrepancies in dosing and type of analysis (i.e., retrospective vs. prospective design).

Results

Table 1 provides a summary of demographics in patients receiving continuous infusions. During the study period, 720 patients were admitted to the PICU for a total of 834 admissions, including 168 obese children ( n  = 184 admissions) and 552 nonobese children ( n  = 650 admissions). Ninety-five patients received a continuous infusion on 102 different admissions; there was a significant difference in the number of patients receiving a continuous infusion between obese and nonobese children, 42 (25.0%) versus 53 (9.6%), p  < 0.01. Four nonobese children had one additional admission, for a total of 57 admissions, and two obese children had more than one admission (i.e., one patient had one extra admission, and the second patient had two subsequent admissions), for a total of 45 admissions. Table 1 provides an overview of the demographics for individual patients and overall admissions for those with continuous infusions. The median age was 7 to 8 years in both groups, and most were admitted for a medical diagnosis. There were no significant differences in the baseline demographics between groups.

Table 1. Demographic data of patients receiving continuous infusions.

Demographics by individual patients ( n  = 95)
Characteristics	Nonobese ( n  = 53)	Obese ( n  = 42)	p -Value
Number (%) or median (IQR)
Number of patients with multiple admissions	4 (7.5)	2 (4.8)	0.694
Demographics by admissions ( n  = 102)
Characteristics	Nonobese ( n  = 57)	Obese ( n  = 45)	p -Value
Males	28 (49.1)	26 (57.8)	0.385
Age (y)	8 (4–14)	7 (5–13)	0.489
BMI percentile class			<.0001
Underweight	46 (80.7)	0 (0.0)
Normal	11 (19.3)	0 (0.0)
Overweight	0 (0.0)	0 (0.0)
Obese (95.0–98.9%)	0 (0.0)	26 (57.8)
Extremely obese (≥99.0%)	0 (0.0)	19 (42.2)
Medical (versus surgical) diagnoses	36 (63.2)	27 (60.0)	0.745
Total number of infusions	224	177
Median (IQR) number of infusions, min, max	2 (1–4) 1, 40	2 (1–6) 1, 18	0.975

Abbreviations: BMI, body mass index; IQR, interquartile range.

There were 401 continuous infusions representing 35 individual medications administered during the study period. Thirty-four infusions were noted in nonobese children, whereas 26 infusions were noted in obese children. There was no significant difference in the median (IQR) number of continuous infusions between the obese versus nonobese groups, 2 (1–6) versus 2 (1–4), p  = 0.975.

Table 2 provides an overview of the 20 most frequently prescribed continuous infusions. Ten (50.0%) were administered for sedation and analgesia or neuromuscular blockade. Six (30.0%) were vasodilators, inotropes, or vasopressors. The remaining medications in the 20 most frequently prescribed list included insulin, aminocaproic acid, furosemide, and bumetanide. Although they were not included in the overall top 20 list, other continuous infusions utilized in one or more obese patients included dobutamine ( n  = 2;1.1%), naloxone ( n  = 2;1.1%), aminophylline ( n  = 1;0.5%), and amiodarone ( n  = 1;0.5%). The two groups were similar for the overall top 20, except no obese patients received vecuronium during the study period. To further illustrate differences in the IV continuous infusions between groups, Table 2 also provides the percent differences in use between groups for the top 20 infusions. Overall, the percent differences in use was not clinically significant.

Table 2. Twenty most frequently prescribed continuous infusions in obese and nonobese admissions in the pediatric intensive care unit.

Overall rank a	Medication name ( n  = orders)	Obese rank	Number of obese admissions ( n  = 184) n (%)	Nonobese rank	Number nonobese of admissions ( n  = 650) n (%)	Percentage points difference (obese—nonobese)
1	Fentanyl ( n  = 92)	2	13 (7.1)	1	79 (12.2)	−5.1
2	Insulin regular ( n  = 92)	1	15 (8.2)	2	77 (11.8)	−3.6
3	Midazolam ( n  = 86)	4	10 (5.4)	3	76 (11.7)	−6.3
4	Epinephrine ( n  = 65)	5	9 (4.9)	4	56 (8.6)	−3.7
5	Milrinone ( n  = 59)	3	11 (6.0)	5	48 (7.4)	−1.4
6	Dopamine ( n  = 41)	10	4 (2.2)	6	37 (5.7)	−3.5
7	Remifentanil ( n  = 39)	6	7 (3.8)	9	32 (4.9)	−1.1
8	Propofol ( n  = 37)	8	5 (2.7)	8	32 (4.9)	−2.2
9	Dexmedetomidine ( n  = 36)	9	4 (2.2)	7	32 (4.9)	−2.7
10	Cisatracurium ( n  = 30)	11	3 (1.6)	10	27 (4.2)	−2.6
11	Aminocaproic acid ( n  = 19)	7	5 (2.7)	12	14 (2.2)	0.5
12	Morphine ( n  = 19)	21	1 (0.5)	11	18 (2.8)	−2.3
13	Norepinephrine ( n  = 16)	14	3 (1.6)	13	13 (2.0)	−0.4
14	Hydromorphone ( n  = 15)	13	3 (1.6)	14	12 (1.8)	−0.2
15	Furosemide ( n  = 13)	12	3 (1.6)	15	10 (1.5)	0.1
16	Sodium nitroprusside ( n  = 11)	17	2 (1.1)	16	9 (1.4)	−0.3
17	Vasopressin ( n  = 11)	15	2 (1.1)	17	9 (1.4)	−0.3
18	Ketamine ( n  = 7)	20	1 (0.5)	19	6 (0.9)	−0.4
19	Vecuronium ( n  = 7)	NR	–	18	7 (1.1)	−1.1
20	Bumetanide ( n  = 6)	19	1 (0.5)	20	5 (0.8)	−0.3

Abbreviation: NR, not ranked.

Table 3 provides a summary of the published studies on dosing, pharmacodynamics, and/or pharmacokinetics of medications listed in the 20 most frequently prescribed continuous infusions in obese children. ^{18 19 20 21 22 23} Only six studies representing three continuous infusion agents (15.0%) and a total of 115 obese children were identified. All of these studies focused on sedatives and analgesics.

Table 3. Summary of published studies in obese patients receiving continuous infusions ^{18 19 20 21 22 23} .

Reference	Type of report	N	Age (y)	Results
Fentanyl
Gish (2010)	Retrospective, descriptive	15 (obese and overweight)	Mean ± SD, 8.5 ± 2.8	• No significant difference in number of dosage changes per day between overweight/obese vs. normal weight group, 0.92 vs. 0.69 ( p  = 0.16)  • Numerical difference in sedative/analgesic boluses between overweight/obese vs. normal weight group, 45.6 ± 62.0 vs. 22.9 ± 25.2 ( p  = 0.19)
Fentanyl
Johnson et al (2016) 2	Retrospective, descriptive	18	Median (IQR), 6.3 (3.8–6.3)	• For every increase in 10 µg/h, nonobese children were 19% less likely to achieve goal sedation at any point in time, accounting for potential confounding factors than obese children (adjusted HR = 0.81, 95% CI: 0.70–0.93)  • Fentanyl initial dosing was not associated with time to goal sedation in obese children (adjusted HR = 1.09, 95% CI: 0.93–1.28, p  > 0.05)
Midazolam
van Rongen et al (2015) 20	Prospective, pharmacokinetic	16 (obese and overweight)	Mean ± SD, 15.6 ± 1.6	• Clearance was not affected by body weight  • Peripheral Vd increased with increased body weight  • Suggested fixed rate (mg/h) should be used rather than mg/kg/h  • Obese adolescents may need a higher initial infusion rate at initiation of midazolam with increased Vd
Propofol
Diepstraten et al (2012) 22	Prospective, pharmacokinetic	20	Mean SD, 16 ± 2	• Total body weight was most predictive covariate for determining propofol clearance  • Propofol dosing should be based on an ABW
Propofol
Chidambaran et al (2013) 22	Prospective, pharmacokinetic	20	Mean SD, 15.8 ± 2.2	• BIS values were below 40 for 93% of maintenance dosing  • Exhibited delayed emergence for 25.8 ± 22 minutes, increased somnolence (RSS > 4) in first 30 minutes of recovery from anesthesia  • 30% incidence of postoperative respiratory adverse events
Chidambaran et al (2015) 23	Prospective, pharmacokinetic	26	Mean SD, 15.7 ± 2.1	• TBW was most predictive variable for propofol clearance  • Maintenance dosing should utilize ABW with a co-factor of 0.75 times TBW

Abbreviations: ABW, adjusted body weight; BIS, bispectral index; CI, confidence interval; HR, hazard ratio; IQR, interquartile range; RSS, Ramsey Sedation Score; SD, standard deviation; TBW, total body weight; Vd, volume of distribution.

Discussion

This is the first study to evaluate the use of continuous infusions in obese and nonobese children. Ninety-five children received a continuous infusion, including 25% of obese children, and there was a significant difference when compared with the percentage of nonobese children receiving continuous infusions. These findings are clinically meaningful in that a recent study noted that one-fifth of all admissions to the PICU involved obese children. ²⁴ Several studies have evaluated the frequency of medication use in obese children, but it is difficult to compare their findings versus our study since they included oral and IV intermittent medications versus a more focused study on only IV continuous infusions. ^{2 6 7}

There were 401 continuous infusions representing 35 distinct agents administered during the study period. Of the 20 most frequently prescribed agents, 10 (50%) were sedatives and analgesics or neuromuscular blockers, 6 (30%) were for hemodynamic support, and the remaining 4 (20%) were for various indications. A literature search was conducted to identify studies that provided pharmacodynamic and pharmacokinetic data in obese children. The goal of this literature search was to provide the best approach for dosing of the infusion for each specific agent. There are two general approaches for continuous infusion dosing, fixed-dose (e.g., mg/h), and weight-based dosing (e.g., mg/kg/h). Weight-based dosing could be further broken down into four different strategies: total body weight (TBW), ideal body weight (IBW), adjusted body weight (ABW), and lean body weight (LBW). ^{2 7 25 26} LBW has been most recently described and focuses on an estimation of lean tissue mass (e.g., skeleton, organs). Janmahasatian et al ²⁶ have suggested that LBW may be a better predictor of medication dosing in obese patients. However, the LBW calculation is very complex and difficult to perform at the bedside. The most appropriate weight-based dosing strategy depends on the individual pharmacokinetics of specific agents; there are no randomized studies comparing these different approaches.

Ross et al ⁷ also conducted a literature search of the top 100 medications in obese children from the Pediatric Health Information System database and attempted to provided weight-based dosing strategies for each of these medications. However, their literature search focused on studies providing pharmacodynamic and pharmacokinetic data in medications administered intermittently and via continuous infusion in both adult and pediatric patients. They did not specify if the recommended weight-based dosing strategy was based on single-dose administration versus continuous infusion. In addition, they provided recommendations based on adult and pediatric data. Considering the significant pharmacokinetic differences in adults and children, there may be fallacies in this approach. The median age in obese group was 7 years. It is well known that patients in this age range have increased hepatic and renal clearance compared with older adolescents and adults, so adult dosing based on pharmacokinetic data should not be extrapolated to children. ²⁷

Similar to Ross et al, ⁷ the purpose of our literature review was to determine the dosing strategy, but we wanted to focus specifically on continuous infusions using data only from pediatric patients. We noted that 3 (15.0%) of the 20 most frequently prescribed continuous infusions in our study had pharmacokinetic and pharmacodynamic outcomes reported in obese children ( Table 3 ). Fentanyl was the most common continuous infusion in all PICU patients and second most common when specifically looking at obese patients. Two of the identified studies from the literature search evaluated dosing and pharmacodynamics. ^{18 19} Gish et al ¹⁸ retrospectively evaluated the number of fentanyl infusion dosage changes in 15 overweight and obese children versus 16 normal-weight controls, with 18 and 16 individual infusions identified. They found a numerical but not a statistically significant difference in the number of dosage changes between the overweight and obese children versus controls, 0.92 ± 0.60 versus 0.69 ± 0.34, p  = 0.16. This study is limited in that it lacked a pharmacodynamic target, such as a validated sedation score. Johnson et al ¹⁹ conducted a retrospective study to examine the association between initial fentanyl infusion rate and time to goal sedation, defined as achieving three consecutive State Behavioral Scale scores from 0 to −1, and dosing characteristics among 57 nonobese and 18 obese children. They noted that for every 10 µg/h increase in initial fentanyl dose in nonobese children there was 19% lower odds of achieving goal sedation at any point in time compared with obese children. In addition, the investigators also found that initial dosing was not associated with time to goal sedation in obese children (adjusted hazard ratio =1.09, 95% confidence interval: 0.93–1.28). These reports provide some indirect evidence that there may be different outcomes between obese and nonobese children receiving fentanyl infusions. However, it is difficult to draw significant conclusions considering the retrospective design and lack of pharmacokinetic analysis. Data in adults suggest that an ABW may be more appropriate than other weight-based dosing strategies. ^{28 29} Because children do not have the same fat disposition as adults, the use of an ABW cannot be recommended until further studies are conducted.

Midazolam was the third most commonly prescribed continuous infusion in the PICU and the fourth most prevalent infusion noted in obese children. Based on the literature search, there is only one study that has evaluated the pharmacokinetics and dosing of midazolam infusions in adolescents. van Rongen et al ²¹ prospectively assessed the pharmacokinetics in 16 overweight and obese adolescents receiving midazolam infusions. They noted that the peripheral volume of distribution increased with increasing excess body weight associated with overweight/obesity but not the body weight related to growth. Based on these findings, they proposed that obese children may need a higher initial infusion rate or loading dose to account for this alteration. Further, van Rongen et al ²¹ suggested that clinicians should utilize a fixed-dosing regimen in mg/h versus mg/kg/h. It should be noted that they did not assess dosing based on a pharmacodynamic marker, such as sedation scores. These findings are limited to the adolescent population, and at this time, they cannot be utilized to guide dosing in obese children.

Propofol was the eighth most common continuous infusion in all PICU and obese PICU patients. We noted three studies evaluating the pharmacokinetics and/or outcomes in 66 children receiving propofol infusions. ^{21 22 23} Diepstraten et al ²¹ conducted a prospective, pharmacokinetic study in 20 obese children and adolescents. They utilized a two-compartment pharmacokinetic model and determined that TBW was the most predictive covariate for determining propofol clearance compared with LBW and IBW. In addition, they recommended that future dosing should be based on TBW using an allometric function. Chidambaran et al ²² conducted a prospective study evaluating pharmacodynamic outcomes such as bispectral index (BIS), delayed emergence, somnolence, and respiratory adverse events (RAE) in 20 obese children receiving propofol infusions during laparoscopic surgery. They noted patients had below target BIS values, delayed emergence, and increased somnolence. In addition, they noted a 30% increase in RAE, with a 14% increase in odds for each one unit increase in BMI. As a follow-up to this study, Chidambaran et al ²³ performed a prospective pharmacokinetic, pharmacodynamic study in 26 obese children, 9 to 18 years, receiving propofol infusions during surgery. They noted that TBW was the most predictive variable for propofol clearance and suggested that maintenance dosing should utilize TBW in an allometric function with an exponent of 0.75. The investigators indicated that their proposed dosing regimen was appropriate when targeting desired propofol concentrations and BIS scores. These studies provide some promising results, but were mainly focused on adolescents rather than younger children.

There was insufficient data for 17 (85.0%) of the 20 most frequently prescribed continuous infusions ( Table 2 ). A second literature search was conducted to evaluate dosing, pharmacodynamics, and/or pharmacokinetics of these agents in adults from studies evaluating continuous infusions. This data was used to provide a framework for dosing recommendations in obese children until further studies are conducted and are not meant to provide specific recommendations. Studies were noted for 7 (35%) of the top 20 medications. These studies recommend fixed-dosing for norepinephrine (µg/min) ^{30 31} and vasopressin (units/min), ^{32 33 34} LBW should be used for remifentanil, ^{35 36 37} IBW should be used for cisatracurium ^{38 39} and vecuronium, ^{40 41} TBW for morphine, ⁴² and ABW should be used for dexmedetomidine. ^{43 44} No studies in adults provide dosing guidance for 10 (50%) of the remaining top 20 continuous infusions.

Until further studies can be conducted, it seems prudent to utilize weight-based dosing with TBW when calculating dosing. Clinicians should utilize a pharmacodynamic target such as sedation score, train of four, blood pressure or mean arterial pressure, or other tools as appropriate, and titrate the agent based on its therapeutic effect. If possible, clinicians can also consider monitoring drug concentration for adjusting the medication if available. In the future, researchers should utilize a more sophisticated analysis in obese children like a population pharmacokinetic analysis with large sample size, or a physiologically based pharmacokinetic (PBPK) model. ⁴⁵ These models are able to account for the heterogeneity in pharmacokinetic parameters across different age-groups in children and account for the impact of obesity on age-specific pharmacokinetic parameters.

This study has several limitations. First, the data represent the percentage of obese PICU admissions and number of continuous infusions from a single-center. Our findings may not be reflective of other institutions across the United States. Second, these data were obtained from a sample of patients admitted to our PICUs in 2011, and we acknowledge that there have been advances in pediatric critical care and that the distribution of medications may be different now versus then. However, we believe that the sample list of medications would be similar to those seen in hospitals with both medical and cardiovascular PICUs. Last, our literature search was limited to studies in the English language and may not reflect studies in other countries.

Conclusions

Continuous infusions were administered in 42 (25.0%) of obese children, which was statistically significant compared with nonobese children. The literature guiding the dosing of the 20 most frequently prescribed continuous infusions in children was limited to six studies, representing just three medications. Future studies evaluating the pharmacokinetics of obese children should employ a PBPK model to account for the heterogeneity of pharmacokinetic parameters across the continuum of age and the effect of obesity on age-related pharmacokinetic changes.