Green Metaanalysis 2009

8/7/2019 Green Metaanalysis 2009

1/14

PEDIATRICS/ORIGINAL RESEARCH

Predictors of Emesis and Recovery Agitation With EmergencyDepartment Ketamine Sedation: An Individual-Patient Data

Meta-Analysis of 8,282 Children

Steven M. Green, MDMark G. Roback, MD

Baruch Krauss, MD, EdM

Lance Brown, MD, MPH

Ray G. McGlone, FCEM

Dewesh Agrawal, MD

Michele McKee, MD, MS

Markus Weiss, MD

Raymond D. Pitetti, MD, MPH

Mark A. Hostetler, MD, MPH

Joe E. Wathen, MD

Greg Treston, MBBS

Barbara M. Garcia Pena, MD

Andreas C. Gerber, MD

Joseph D. Losek, MD

For the Emergency Department

Ketamine Meta-Analysis

Study Group*

From the Department of Emergency Medicine, Loma Linda University Medical Center and ChildrensHospital, Loma Linda, CA (Green, Brown); the Department of Pediatrics, University of Minnesota,

Minneapolis, MN (Roback); the Division of Emergency Medicine, Childrens Hospital and Harvard

Medical School, Boston, MA (Krauss); the Royal Lancaster Infirmary, Lancaster, UK (McGlone); the

Division of Emergency Medicine, Childrens National Medical Center, Washington, DC (Agrawal);

the Division of Emergency Medicine, Boston Medical Center, Boston, MA (McKee); the Department

of Anaesthesia, University Childrens Hospital, Zurich, Switzerland (Weiss, Gerber); the Division of

Pediatric Emergency Medicine, Childrens Hospital of Pittsburgh, PA (Pitetti); the Department of

Pediatrics, University of Chicago, Chicago, IL (Hostetler); the Department of Pediatrics, University

of Colorado Health Sciences Center, Denver, CO (Wathen); the Emergency Department, Royal

Darwin Hospital, Darwin, Northern Territory, Australia (Treston); the Division of Emergency

Medicine, Miami Childrens Hospital, Miami, FL (Garcia Pena); and the Department of Pediatrics,

Medical University of South Carolina, Charleston, SC (Losek).

Study objective: Ketamine is widely used in emergency departments (EDs) to facilitate painful procedures; however,

existing descriptors of predictors of emesis and recovery agitation are derived from relatively small studies.

Methods: We pooled individual-patient data from 32 ED studies and performed multiple logistic regression todetermine which clinical variables would predict emesis and recovery agitation. The first phase of this study

similarly identified predictors of airway and respiratory adverse events.

Results: In 8,282 pediatric ketamine sedations, the overall incidence of emesis, any recovery agitation, and

clinically important recovery agitation was 8.4%, 7.6%, and 1.4%, respectively. The most important independent

predictors of emesis are unusually high intravenous (IV) dose (initial dose of2.5 mg/kg or a total dose of

5.0 mg/kg), intramuscular (IM) route, and increasing age (peak at 12 years). Similar risk factors for any

recovery agitation are low IM dose (3.0 mg/kg) and unusually high IV dose, with no such important risk factors

for clinically important recovery agitation.

Conclusion: Early adolescence is the peak age for ketamine-associated emesis, and its rate is higher with IM

administration and with unusually high IV doses. Recovery agitation is not age related to a clinically important

degree. When we interpreted it in conjunction with the separate airway adverse event phase of this analysis, we

found no apparent clinically important benefit or harm from coadministered anticholinergics and benzodiazepines

and no increase in adverse events with either oropharyngeal procedures or the presence of substantialunderlying illness. These and other results herein challenge many widely held views about ED ketamine

administration. [Ann Emerg Med. 2009;54:171-180.]

Provide feedback for this article at the journals Web site, www.annemergmed.com.

0196-0644/$-see front matter

Copyright 2009 by the American College of Emergency Physicians.

doi:10.1016/j.annemergmed.2009.04.004

*All members are listed in the Appendix.

Volume , . : August Annals of Emergency Medicine 171
http://www.annemergmed.com/http://www.lexisnexis.com/dpartner/process.asp?qs_id=4676


2/14

INTRODUCTION

BackgroundThe efficacy and safety of ketamine to facilitate painful

procedures for children in the emergency department (ED) have

been documented in 57 published series totaling nearly 10,000

patients.1 Airway and respiratory adverse events occur in 3.9%

of ketamine sedations, and their predictors have been previously

reported.2 Although less serious, vomiting with ketamine is a far

more common adverse event. In the largest 2 ED ketamine

series, the incidence of predischarge vomiting was 7% and

8%3,4; however, smaller reports have noted rates as high as

26%.5,6 Unpleasant recovery reactions, including nightmares,

can also occur.7

In a previous study assessing such predictors, Green et al8

analyzed 1,021 ketamine sedations and found that emesis was

not associated with dose or underlying illness, but was twice as

common in individuals aged 5 years or older. Recovery agitation

was not dose related, but was almost twice as common in

individuals younger than 5 years and in those with an

underlying medical condition. This study was limited by its

inclusion of just 2 clinical sites and only the intramuscular (IM)

route for the drug. Further, because all studied children received

atropine and essentially none received benzodiazepines, any

effect of these adjuncts could not be assessed.

ImportanceIf specific differences in ketamine technique or patient

variables are predictive of emesis or recovery agitation, thenemergency physicians may elect to modify their administrationtechnique or patient selection to minimize such adverse events.Also, patients and their families may be provided more accurateinformation about the risks and benefits of ketamine sedation.

Goals of This InvestigationWe pooled original data from all available series of ED

ketamine sedation in children to identify clinical predictors ofemesis and recovery agitation.

MATERIALS AND METHODSStudy Design

We performed a meta-analysis of all available original datafrom existing ketamine case series and have recently reported thefirst phase of this analysis that includes predictors of airway and

respiratory adverse events.2

This second preplanned phaseanalyzes predictors of emesis and recovery agitation. Allincluded trials had local ethics committee approval.

The methodology of our database assembly has been detailedin the first phase.2 Briefly, we used a literature search to identifyall previous ED ketamine series and contacted authors to obtaintheir original individual patient data. The final aggregatedatabase included 8,282 ketamine sedations from 32 reports.

Our 3 study outcomes were emesis, any recovery agitation,and the subset of recovery agitation deemed clinicallyimportant. We defined emesis as an episode of vomitingoccurring either during sedation or recovery but before ED

discharge (given that most studies did not assess vomiting afterdischarge). We defined recovery agitation as any combination ofagitation, crying, hallucinations, or nightmares, regardless ofseverity. We defined clinically important recovery agitation asoccurrences of abnormal behavioral responses (defined above)that either led to specific treatment or were specially describedby investigators as demonstrating substantial severity.

Authors were queried about any missing data points andwere asked to recode their variables as needed to comply withour study definitions. One study in this sample did not assessrecovery agitation,9 and so its 44 otherwise eligible childrenwere omitted for these analyses.

Candidate predictor variables were selected according toprevious literature and on biological plausibility of associationwith the specific adverse events.

The ketamine technique variables chosen were route (codedas intravenous [IV] versus IM), initial dose (mg/kg), total dose(mg/kg), the presence or absence of coadministeredanticholinergics (eg, atropine, glycopyrrolate), and the presenceor absence of coadministered benzodiazepines (eg, midazolam,diazepam). We also included 2 specific dichotomizations ofketamine technique according to their previous independentprediction of airway and respiratory adverse events2: low IMdose (defined as a total dose of3.00 mg/kg) and unusually

Editors Capsule Summary

What is already known on this topic

Ketamine is commonly used for pediatric sedation.Although its complications are well known, theirfrequency and factors that predispose patients to

complications are not.What question this study addressed

This 8,282 individual-patient (32 reports) meta-analysis summarizes complications of ketamine useand their association with patient demographics andcharacteristics of the sedation procedure.

What this study adds to our knowledge

The study suggests that emesis is more common inearly adolescence and is more common when thedrug is given intramuscularly or in higher

intravenous doses. Recovery agitation was associatedwith dosage in patients given intravenous ketamine.

How this might change clinical practice

Although this meta-analysis will not systematicallychange practice, its findings may help physicianstailor their sedation practice to particular patientsand situations.

Ketamine Meta-analysis Green et al

172 Annals of Emergency Medicine Volume , . : August


3/14

high IV dose (defined as either an initial dose of2.5 mg/kg ora total dose of5.0 mg/kg).

The patient variables chosen were age, American Society ofAnesthesiologists (ASA) physical status,10 and oropharyngealprocedural indication (coded as present versus absent).

We examined the frequency distributions for the continuousvariables, and if their distributions were found to be bimodal,

the variables were dichotomized at logical thresholds. Weperformed separate multiple logistic regression analyses for eachof the 3 outcomes by using identical methodology to that of ourprevious analysis of airway and respiratory adverse events.2 Allsuch analyses were performed with Stata 9 software (StataCorp,College Station, TX).

Because of concerns about potential underreporting ofadverse events in retrospective research, as in the first phase2 wealso performed subset analyses with just the subset withprospectively obtained data. Our a priori intent was that if theprospective subset analyses disagreed from their overallcounterparts, then the prospective subsets would be deemed the

more reliable, given their stronger methodology. If theprospective subset analyses agreed with their overallcounterparts, then the overall analysis would be consideredreliable.

RESULTSCharacteristics of Study Subjects

The literature search, article processing, and demographics ofthe resulting database have been reported elsewhere.2 Emesiswas documented in 694 (8.4%) of 8,282 sedations, with noreports of aspiration. Recovery agitation was described in 630(7.6%) of the 8,238 sedations and was considered clinically

important in 115 (1.4%). The breakdown of adverse events bystudy is shown in Appendix E1 (available online at

http://www.annemergmed.com).When we examined the frequency distributions for theoutcomes stratified by age (Figure 1), we observed roughly log-linear distributions except for emesis, which increasedprogressively up until aged 12 years and then declinedthereafter. Despite this, we chose to not lose the enhancedpower of this continuous variable through dichotomization,given that only 5% of the subjects were older than 13 years.

The frequency distributions for outcomes stratified by initialand total dose (Figures 2 to 5) suggested a bimodal distributionfor initial and total IM dosing, which should already beaddressed by our inclusion of the low IM dose variable, and

Figure 1. Frequency distribution of adverse events by age. The total number analyzed for emesis was 8,282, with thefollowing number of subjects represented in each column: 131, 818, 1001, 860, 841, 754, 698, 579, 439, 430, 399,329, 310, 269, 206, 109, 109. The total number analyzed for recovery agitation was 8,238, with the following number ofsubjects represented in each column: 127, 805, 994, 859, 835, 752, 695, 579, 436, 428, 398, 329, 309, 269, 205,109, 109.

Figure 2. Frequency distribution of adverse events by initialIM dosing (n2,453). The number of subjects representedin each column is as follows: 475, 270, 148, 495, 917,51, 97.

Green et al Ketamine Meta-analysis

http://www.annemergmed.com/http://www.annemergmed.com/


4/14

for initial IV dosing, which should already be accounted for by

our high IV dose variable. Given an apparent progressive

increase in some adverse events with dose in Figures 3 and 5,

irrespective of these 2 thresholds, we also retained dose as a

continuous variable.

Initial ketamine dose was not collected in 5 studies

comprising 1,536 sedations (18.5%),2 and because as expectedthis variable was collinear with total dose (Spearmans

rho0.83), only the latter was retained in the multivariate

analyses. Given that typical IV and IM dosing

recommendations differ by a factor of approximately 3 (ie, 1.5

mg/kg IV versus 4 mg/kg IM),7,10 it was impossible to simply

combine observed doses from the 2 routes without adjustment.

Accordingly, we separately stratified IV and IM dose sets into

quintiles and used quintile rankings for the continuous variable

dose. Outcomes stratified by total dosing quintile are shown in

Figure 6.

Unadjusted comparisons of predictor variables by outcomeare shown in Appendices E2 to E4 (available online at http://www.annemergmed.com).

The final predictor variable list thus included 2 continuousvariables (age, total dose quintile) and 7 binary variables (lowIM dose, unusually high IV dose, route, oropharyngealprocedure, ASA score 3, anticholinergic, benzodiazepine).The number of total outcomes was sufficient to include allpredictor variables for all analyses.

The multiple logistic regression analyses are shown in Table1, demonstrating multiple significant independent predictors foreach outcome. In 3 circumstances, the prospective subsetanalyses overruled findings for the total sample.

To better interpret the clinical relevance of these findingsand those of the first phase study of airway and respiratoryadverse events,2 we ranked the significant predictors in Table 2and calculated a variation of the number needed to treatconcept that we call the number needed to cause. Specifically,this refers to the projected number of ketamine sedations withthe risk factor that would lead to 1 additional adverse event,relative to the same number of sedations without the risk factor.That is, if a clinician administers ketamine in the consistentpresence of the identified predictor, how many sedations arerequired to theoretically induce 1 additional related adverse

event? For example, as shown in Table 2, the routine use of anadjunctive benzodiazepine would be expected to result in onemore airway or respiratory adverse event every 33 to 321sedations (with 95% confidence), and one less occurrence ofvomiting every 19 to 107 sedations (with 95% confidence).

LIMITATIONSThe principal limitation of this report is the heterogeneity of

the collated studies and the observational nature of the data. Asdiscussed in the first phase report,2 there is substantial variationin procedural indications, ages, doses, and other clinical

Figure 3. Frequency distribution of adverse events by total IM dosing (n2,605). The number of subjects represented ineach column is as follows: 425, 257, 173, 497, 857, 65, 149, 98, 84.

Figure 4. Frequency distribution of adverse events by initialIV dosing (n4,293). The number of subjects representedin each column is as follows: 784, 1,857, 601, 382, 192,477.


http://www.annemergmed.com/http://www.annemergmed.com/http://www.annemergmed.com/http://www.annemergmed.com/http://www.annemergmed.com/


5/14

variables, as might be expected from 32 studies coming frommultiple countries. The direction and magnitude of anyresulting bias, if present, cannot be ascertained. This samediversity could be argued as a major strength of the analysisbecause our findings are likely to have substantial externalvalidity, given the wide spectrum of collective input.

A limitation to our study of recovery agitation is thesubjective clinical interpretation of these responses. Variousauthors have used a spectrum of definitions and terms (eg,emergence phenomena) to describe them, and accordingly our

study definitions were subjective to accommodate this factor.Preprocedural agitation was found to correlate with recoveryagitation in one study,11 and we hope that future research willfocus on this issue.

Vomiting and recovery agitation can occur after EDdischarge; however, because these delayed adverse events wererecorded only in a minority of included studies, we were unableto study them.

Other limitations of this general analysis detailed in theprevious phase include the unavailability of other potentially

relevant clinical variables, the weaknesses inherent to multiplelogistic regression modeling, and potential differences inapplication of our study definitions by different authors.2

DISCUSSIONTo our knowledge, we report the largest ED ketamine

sample to date in this meta-analysis of original data of 8,282sedations collated from 32 previously published series. The firstphase of this study identified multiple statistically significantpredictors of airway and respiratory adverse events,2 and thissecond phase similarly found multiple predictors of emesis andrecovery agitation. Despite the statistical significance of this

array of predictors and their academic interest to researchers, webelieve that most of the identified associations are of amagnitude below the threshold of clinical importance. Perhapsthe most important finding of our 2-phase analysis is theabsence of clinical variables that have more than slight effect.Accordingly, we believe that our data effectively invalidate anumber of widely held ketamine suppositions (Figure 7) andshould therefore have broad clinical influence.

Our data confirm the findings of an earlier smaller study thatidentified a higher rate of vomiting in older children8 but addthat this higher propensity appears to decrease after adolescence(Figure 1). Regardless, the absolute age-based differences are

Figure 5. Frequency distribution of adverse events by total IV dosing. The total number analyzed for emesis was 5,677,with the following number of subjects represented in each column: 675, 2,068, 876, 593, 273, 462, 299, 184, 100,147. The total number analyzed for recovery agitation was 5,633, with the follow number of subjects represented in eachcolumn: 673, 2,063, 863, 580, 266, 458, 299, 184, 100, 147.

Figure 6. Frequency distribution of adverse events by totaldosing quintiles. The total number analyzed for emesiswas 8,282, with the following number of subjectsrepresented in each column: 1,288, 2,028, 1,441, 1,887,1,638. For emesis the break points between quintiles forIM dosing were 2.54, 3.81, 4.00, and 4.17 mg/kg, and forIV dosing were 1.00, 1.25, 1.83, and 3.01 mg/kg. Thetotal number analyzed for recovery agitation was 8,238,with the following number of subjects represented in eachcolumn: 1,097, 2,176, 1,441, 1,857, 1,667. For recoveryagitation the break points between quintiles for IM dosingwere 2.50, 3.81, 4.00, and 4.17 mg/kg, and for IV dosingwere 1.00, 1.25, 1.82, and 3.01 mg/kg.




6/14

small and are not likely to affect clinical practice. A recentrandomized controlled trial has shown that prophylacticondansetron modestly reduces the rate of emesis.12 Childrenat highest risk (ie, approximately 12 years old) could be

targeted with this intervention; however, such therapy has arelatively low yield, for even if the highest rate of vomiting(17% for 12-year-olds from Table 2) is reduced by twothirds,12 the number needed to treat to prevent a single

Table 1. Multiple logistic regression models of emesis, any recovery agitation, and clinically important recovery agitation.

Odds Ratio (95% CI)

Emesis Any Recovery Agitation

Clinically Important Recovery

Agitation

Variable Total Sample

Prospective

Subset Total Sample

Prospective

Subset Total Sample

Prospective

Subset

Age 1.15 (1.131.18) 1.15 (1.121.18) 1.03 (1.011.06) 1.04 (1.011.06) 1.08 (1.031.13) 1.07 (1.011.12)Low IM dose* 1.12 (0. 761.65) 1.04 ( 0. 691. 58) 2.88 (2.034.09) 1.96 (1.382.79) 3.71 (1.688.20) 2.20 (0.994.88)

Unusually high IV dose 3.42 (2.544.61) 2.78 (1.894.09) 1.87 (1.392.51) 2.13 (1.542.96) 0.97 (0.47197) 0.60 (0.201.80)

Total dose quintile 1.01 (0.931.10) 1.02 (0.931.13) 1.24 (1.131.35) 1.25 (1.141.37) 1.16 (0.961.41) 1.06 (0.871.30)

Oropharyngeal

procedure

1.32 (0.812.17) 1.43 (0.852.42) 0.55 (0.301.03) 0.46 (0.240.85) 0.65 (0.162.70) 0.73 (0.183.05)

ASA score 3 0.18 (0.040.74) 0.30 (0.071.29) 0.78 (0.341.80) 1.28 (0.533.08) 0.75 (0.105.51) 1.16 (0.168.63)

IV route (relative to IM) 0.41 (0.320.52) 0.41 (0.320.54) 0.69 (0.550.88) 0.52 (0.410.66) 1.10 (0.621.96) 0.89 (0.501.61)

Anticholinergic 0.69 (0.570.84) 0.78 (0.620.99) 1.07 (0.881.32) 0.87 (0.701.09) 1.18 (0.741.87) 0.94 (0.561.58)

Benzodiazepine 0.74 (0.610.90) 0.60 (0.480.76) 1.04 (0.861.27) 1.00 (0.821.22) 1.23 (0.801.87) 0.93 (0.581.48)

Area under ROC 0.695 0.685 0.637 0.642 0.624 0.609

Hosmer-Lemeshow P.588 P.078 P.997 P.952 P1.000 P1.000

Bolded results have confidence intervals that do not include 1.

*Low IM dose refers to those receiving less than 3.00 mg/kg IM. High IV dose refers to those receiving either an initial dose of greater than or equal to 2.5 mg/kg or

a total dose of greater than or equal to 5.0 mg/kg. The reference group is children without low IM or high IV dosing.

Table 2. Summary of independent predictors of adverse events identified by this and the first phase of this study.

Adverse Event

Incidence in

Samp le Overall, % Ind epe nden t Pred icto rs Odd s Ratio

Number Needed to Cause 1 Additional

Adverse Event, Range*

All airway and respiratory

adverse events

3.9 Avoidance of low IM dose

Age 13 y

Unusually high IV dose

Age 2 y

Addition of anticholinergic

Addition of benzodiazepine

2.9

2.7

2.2

2.0

1.8

1.4

581

926

1343

1555

1871

33321

Laryngospasm 0.3 Unusually high IV dose 3.8 341,190

Apnea 0.8 Unusually high IV dose

Age 13 y

Addition of benzodiazepine

5.1

2.9

2.3

1568

26291

39568

Emesis 8.4 Unusually high IV dose

IM route

Omitting anticholinergic

Omitting benzodiazepine

Increased age

3.4

2.4

1.4

1.4

1.2/year

3 8

613

1663

19107

6692

Any recovery agitation 7.6 Low IM dose

No oropharyngeal procedure

Unusually high IV dose

IM route

Increased total dose

Increased age

2.9

2.2

2.1

1.4

1.2/quintile

1.02/year

413

475

934

1696

38101

2191,316

Clinically importantrecovery agitation

1.4 Increased age 1.1 5492,381

*This refers to the projected number of ketamine sedations with the risk factor that would lead to 1 additional adverse event, relative to the same number of seda-

tions without the risk factor. For example, the routine use of an anticholinergic would be expected to result in 1 more airway or respiratory adverse event every 18 to

71 sedations and 1 less occurrence of vomiting every 16 to 63 sedations. These figures are calculated according to the overall incidence of adverse events in the

sample and the confidence intervals of the odds ratios.




7/14

occurrence calculates to 9. This yield will be less favorable atother age categories.

We also found that increased age was significantly associatedwith recovery agitation (including clinically important recoveryagitation), but at levels that we regard as too trivial for clinicalconsideration. In contrast to traditional thinking,13 adolescents

were not at substantially higher risk of clinically importantrecovery agitation (Figure 1). Our sample included 693 childrenaged 13 years or older and should be robust on this issue.

Thus, our data show no evidence that clinicians should avoidor restrict ketamine use in any specific age group included,

except that we studied too few infants younger than 3 monthsto counter this traditional contraindication.7 Our earlier phase

identified a greater risk of airway and respiratory events inchildren younger than 2 years or aged 13 years or older;however, we believe that the corresponding values of number

needed to cause (Table 2) are insufficient to dissuade cliniciansfrom administering ketamine when otherwise indicated.

Age between 3 and 12 months has been described as arelative contraindication to ketamine7; however, our sampleincludes 168 such children and supports the general safety ofketamine in this age range.

The first phase of our study revealed that unusually high IVdoses of ketamine (initial dose 2.5 mg/kg or total dose 5.0mg/kg) increased the risk of airway and respiratory adverse

events, primarily through an increase in apnea. This secondstudy phase finds that such dosing also increases the risk of

emesis and any recovery agitation. Although these findingssupport the advisability of avoiding such high IV dosingwhenever feasible, any clinical importance focuses on emesis(Table 2; number needed to cause 3 to 8) because the effect onairwayand respiratory adverse events (13 to 43) and recoveryagitation (9 to 34) is not large.

Some clinicians advocate unusually low IM doses (eg, 2 mg/kg), with the goal of minimizing adverse events.14,15 Suchsubdissociative dosing is inadequate for painful procedures suchas fracture reduction or abscess drainage but can provide

analgesia and anxiolysis for minor procedures such as lacerationrepair with local anesthetic. The first phase of our study foundthat low IM doses (3.0 mg/kg) led to fewer overall airway andrespiratory adverse events (number needed to cause 5 to 81);however, this second phase shows that this benefit of suchsubdissociative dosing is countered by a correspondingly higherrate of recovery agitation (number needed to cause 4 to 13).

An inverse relationship between dose and recovery agitationmay seem paradoxic but has been observed in the past.16-18 Anunproven but possible explanation might be that subdissociativedoses only partially impair sensory input. As a result, during the

procedure children may be feeling pain or become frightened byauditory and visual stimuli that are distorted andincomprehensible.

The adverse event associations observed with low IM dosingwere not similarly suggested with low IV dosing, and anexplanation for this different effect by route is not apparent.

Other than the 2 dosing subgroups identified above (high IVdose, low IM dose), the only further dosing association observedin either phase of our study is the prediction of any recoveryagitation at a trivial level (number needed to cause 38 to 101per dosing quintile). Within the range of clinically useddissociative doses short of the unusually high IV category

Patient selection Early adolescence is the peak age for vomiting, with

lesser risk in younger and older children. Adolescents and teenagers are not at higher risk for

clinically important recovery agitation than youngerchildren.

Ketamine appears safe in children 3 to 12 months ofage.

Typical ED oropharyngeal procedures as described inthe reports herein pose no extra risk of adverse events.

Unlike other sedatives, the presence of substantialunderlying illness (ASA class 3) does not appear topresent any greater risk of adverse events.

Route The IM and IV routes display similar risk of airway and

respiratory adverse events, and of clinically importantrecovery agitation. However, the IM route is associatedwith a higher rate of vomiting, and IV administration

appears preferred in settings where venous access can beobtained with minimal upset to the child.

Dose Unusually high IV doses of ketamine (initial dose 2.5

mg/kg or total dose 5.0 mg/kg) are associated with anincreased risk of vomiting, and a slight increase inapnea and recovery agitation.

Excluding the unusually high IV doses described above,there is no apparent dose-response with adverse eventsover the range of clinically common doses. Thus, thereis no apparent benefit to using 1 mg/kg IV rather than2 mg/kg IV, or 3 mg/kg IM rather than 4-5 mg/kg IM.

Subdissociative ketamine (3 mg/kg IM) appears tohave the least airway and respiratory adverse effects ofany form of ketamine administration, but is unsuitablefor more painful procedures and has a higher risk ofrecovery agitation.

Co-administered medications Anticholinergics do not decrease airway adverse events,

and there is no compelling basis for their routine use. Benzodiazepines do not decrease recovery agitation, and

there is no compelling basis for their routine use asprophylaxis. Instead, they should be reserved for pre-

procedural anxiolysis and treatment of unpleasantrecovery reactions.

Figure 7. Key findings of the 2 phases of this analysis.




8/14

described, we found no evidence that any adverse events weredose related to any magnitude approaching clinical importance,which is in marked distinction to other parenteral proceduralsedation agents (eg, benzodiazepines, propofol, opioids) in

which proportional dose-related increases in such events areevident throughout the full spectrum of doses administered.10

This suggests that, as long as excessively high IV doses are

avoided, emergency physicians should not avoid higher dosesfor fear of adverse events but rather select doses that will bemore consistently effective. For example, our data suggest no

overall disadvantage to using 2 mg/kg IV rather than 1 mg/kgor 4 to 5 mg/kg IM rather than 3 mg/kg.

The first phase of this analysis failed to identifyoropharyngeal procedures as an independent predictor of airwayand respiratory adverse events, and this second phase finds onlythat such procedures predict less recovery agitation. The formerfinding should be interpreted with caution because the ED

procedures in this sample likely involved substantially less throatstimulation than other settings in which higher risk is

apparent.7,19 Nevertheless, our data support the general safety ofusing ketamine with minor oropharyngeal procedures such asthose described in the included reports (eg, intraoral lacerationrepair, dental procedures, removal of oropharyngeal andesophageal foreign bodies). The latter finding of less recoveryagitation is not intuitive but may reflect the fact that EDoropharyngeal procedures are, on average, shorter and lessinvolved than more common indications such as fracturereduction and abscess drainage.

With all nondissociative agents, the risk of adverse events isthought to be proportional to the degree of underlying physicalillness, as is typically quantified by using the ASA physical statusclassification.10 Our analysis included 92 children with ASAscore greater than or equal to 3 and confirms previousobservations that ketamine lacks such an association.8,19,20 Thecardiopulmonary support characteristic of this drug may make itpreferable to other sedatives in children with substantialunderlying illness.7

The first phase of our study found that the choice of route(IV versus IM) did not predict airway or respiratory adverseevents; however, this second phase found the IM route to beindependently associated with emesis at a rate that appearsclinically important (number needed to cause 6 to 13). Thisconcurs with a randomized trial by Roback et al,5 in which theIM route led to twice the rate of vomiting. No explanatoryrationale for why the IM route should independently influenceemesis is apparent to us. Because we observed no dose response(except at extremes) when controlling for route, we think itunlikely to result from the generally higher IM doses. Perhapsthe IM route instead represents a surrogate marker for longerduration of dissociation, and this variable, unavailable to us forstudy, is the true independent predictor of emesis.

The randomized trial by Roback et al5 also found

significantly shorter recoveries with the IV route (median 80versus 129 minutes). Accordingly, it appears that the IV route is

preferred in settings in which venous access can be obtainedwith minimal disturbance to the child, such as pediatric EDs inwhich nurses primarily caring for children are skilled in thisprocedure. General EDs lacking such expertise may still find theIM route preferable.21

An unexpected finding of the first phase of our analysis wasthat overall rates of airway and respiratory adverse events were

significantly highernot lowerin the group receivingconcurrent anticholinergics. The second phase finds lessvomiting with these adjuncts and no effect on recovery

agitation. Despite the statistical significance of the airway andemesis findings, however, their clinical importance appears oflittle or no significance (numbers needed to cause 18 to 71 and16 to 63, respectively). Given this lack of apparent benefit orharm, our data do not support the routine use of

coadministered anticholinergics.The principal intent of coadministering anticholinergics is to

minimize hypersalivation and its risk of laryngospasm or airwayobstruction7,22; however, our data suggest that hypersalivation-

based adverse events are rare in the manner in which ketamine isadministered in the ED. Some may argue that becauseanticholinergics are inexpensive and pose no apparent harm,there is no real risk and perhaps rare benefit from their routineuse. Our data, however, suggest that airway adverse events aremore likely to be increased than decreased, and given that 2,602children in our sample did not receive anticholinergics with nodescribed problem, we believe that our data do not support theroutine use of these adjuncts.

The first phase of our analysis showed that coadministeredbenzodiazepines increased the odds of airway adverse events,primarily through an increase in apnea. As withanticholinergics, the second phase finds less vomiting withbenzodiazepines and no effect on recovery agitation. The effecton apnea and emesis is of unlikely clinical importance, however,because the numbers needed to cause are 39 to 568 and 19 to107, respectively. Thus, there appears to be no apparent

clinically important benefit or harm from coadministeredbenzodiazepines, and our data do not support their routine use.

Benzodiazepines showed no effect on recovery agitation,their primary intended purpose. A limitation of this specificanalysis is that clinicians may have preferentially administeredbenzodiazepines to those perceived at higher risk (eg, greaterpreprocedural anxiety) and that this selection bias may haveobscured a true treatment effect. However, our findings supportthose of 2 randomized trials that found no decrease in recoveryagitation with benzodiazepines.11,23 Despite the results of thesetrials, it has been postulated that there are subsets of childrenwho will benefit from coadministered benzodiazepines24;however, this meta-analysis and other research fail to identifyany such patient groups. Our meta-analysis strongly supportsthe safety of ketamine as a sole agent, and our interpretation ofthe evidence is that the only effective role for benzodiazepines isas treatment for unpleasant recovery reactions or significantpreprocedural anxiety rather than as prophylaxis.




9/14

Figure 7 outlines the principal findings of our 2-phaseanalysis. Several of these conclusions run counter to widely heldsuppositions about ketamine and anecdotal practice perceptions.Regardless, to our knowledge our data sample is the largestavailable, and our findings are statistically robust.

In summary, this second phase of our analysis shows that themost important independent predictors of emesis are unusually

high IV dose, IM route, and increasing age (peak at 12 years).Similar risk factors for any recovery agitation are low IM doseand unusually high IV dose, with no such important risk factorsfor clinically important recovery agitation. Taken together withthe first phase of this analysis, we found no apparent benefit orharm from coadministered anticholinergics and benzodiazepinesand no increase in incidence of adverse events with eitheroropharyngeal procedures or the presence of substantialunderlying illness.

Supervising editors: Kathy N. Shaw, MD, MSCE; Michael L.

Callaham, MD

Dr. Shaw and Dr. Callaham were the supervising editors on

this article. Dr. Green did not participate in the editorial review

or decision to publish this article.

Author contributions: SMG conceived and designed the study.

The methodology was critiqued and revised with extensive

input from MGR, BK, LB, DA, RDP, JEW, and GT. All authors

reviewed and recoded their data to comply with study

definitions, and before data analysis the study protocol was

critiqued and refined by all authors. SMG performed the data

analysis, and a writing committee composed of SMG, MGR,

and BK then created the article. All authors critiqued the draft

and there were substantial revisions. SMG takes responsibility

for the paper as a whole.

Funding and support: By Annals policy, all authors are required

to disclose any and all commercial, financial, and other

relationships in any way related to the subject of this article

that might create any potential conflict of interest. The authors

have stated that no such relationships exist. See the

Manuscript Submission Agreement in this issue for examples

of specific conflicts covered by this statement.

Publication dates: Received for publication January 27, 2009.

Revision received March 13, 2009. Accepted for publication

April 1, 2009. Available online June 6, 2009.

Reprints not available from the authors.

Address for correspondence: Steven M. Green, MD,

Department of Emergency Medicine, Loma Linda University

Medical Center, 11234 Anderson St, Loma Linda, CA 92354;

805-969-2144, fax 775-307-4121; E-mail

[email protected].

REFERENCES1. Green SM, Cote CJ. Ketamine and neurotoxicity: clinical

perspectives and implications for emergency medicine. Ann Emerg

Med. 2009;54:181-190.

2. Green SM, Roback MG, Krauss B, et al. Predictors of airway and

respiratory adverse events with ketamine sedation in the

emergency department: an individual-patient data meta-analysis of

8,282 children. Ann Emerg Med. 2009;54:158-168.

3. Green SM, Rothrock SG, Lynch EL, et al. Intramuscular ketamine

for pediatric sedation in the emergency department: safety profile

with 1,022 cases. Ann Emerg Med. 1998;31:688-697.

4. Roback MG, Bajaj L, Wathen JE, et al. Preprocedural fasting and

adverse events in procedural sedation and analgesia in a

pediatric emergency department: are they related? Ann Emerg

Med. 2004;44:454-459.5. Roback MG, Wathen JE, MacKenzie T, et al. A randomized,

controlled trial of IV versus IM ketamine for sedation of pediatric

patients receiving emergency department orthopedic procedures.

Ann Emerg Med. 2006;48:605-612.

6. Heinz P, Geelhoed GC, Wee C, et al. Is atropine needed with

ketamine sedation? a prospective, randomized, double blind

study. Emerg Med J. 2006;23:206-209.

7. Green SM, Krauss B. Clinical practice guideline for emergency

department ketamine dissociative sedation in children. Ann Emerg

Med. 2004;44:460-471.

8. Green SM, Kuppermann N, Rothrock SG, et al. Predictors of

adverse events with ketamine sedation in children. Ann Emerg

Med. 2000;35:35-42.

9. Hostetler MA, Barnard JA. Removal of esophageal foreign bodiesin the pediatric ED: is ketamine an option? Am J Emerg Med.

2002;20:96-98.

10. Krauss B, Green SM. Procedural sedation and analgesia in

children. Lancet. 2006;367:766-780.

11. Sherwin TS, Green SM, Khan A, et al. Does adjunctive midazolam

reduce recovery agitation after ketamine sedation for pediatric

procedures? a randomized, double-blind, placebo-controlled trial.

Ann Emerg Med. 2000;35:239-244.

12. Langston WT, Wathen JE, Roback MG, et al. Effect of

ondansetron on the incidence of vomiting associated with

ketamine sedation in children: a double-blind, randomized,

placebo-controlled trial. Ann Emerg Med. 2008;52:30-34.

13. Green SM, Sherwin T. Incidence and severity of recovery agitation

following ketamine sedation in young adults. Am J Emerg Med.

2005;23:142-144.

14. McGlone RG, Fleet T, Durham S, et al. A comparison of

intramuscular ketamine with high dose intramuscular midazolam

with and without intranasal flumazenil in children before suturing.

Emerg Med J. 2001;18:34-38.

15. McGlone RG, Howes MC, Joshi M. The Lancaster experience of

2.0 to 2.5 mg/kg intramuscular ketamine for paediatric sedation:

501 cases and analysis. Emerg Med J. 2004;21:290-295.

16. Szappanyos GG, Gemperle M, Rifat K. Selective indications for

ketamine anaesthesia. Proc R Soc Med. 1971;64:1156-1159.

17. Faithfull NS, Haider R. Ketamine for cardiac catheterisation.

Anaesthesia. 1971;26:318-323.

18. Phillips LA, Seruvatu SG, Rika PN. Anaesthesia for the surgeon-

anaesthetist in difficult situations. Anaesthesia. 1970;25:36-45.

19. Green SM, Klooster M, Harris T, et al. Ketamine sedation for

pediatric gastroenterology procedures. J Pediatr Gastroent Nutr.

2001;32:26-33.

20. Green SM, Denmark TK, Cline J, et al. Ketamine sedation for

pediatric critical care procedures. Pediatr Emerg Care. 2001;17:

244-248.

21. Green SM, Krauss B. Should I give ketamine IV or IM [editorial]?

Ann Emerg Med. 2006;48:613-614.

22. Brown L, Christian-Kopp S, Sherwin TS, et al. Adjunctive atropine

is unnecessary during ketamine sedation in children. Acad Emerg

Med. 2008;15:314-318.

23. Wathen JE, Roback MG, Mackenzie T, et al. Does midazolam

alter the clinical effects of intravenous ketamine sedation in


mailto:[email protected]:[email protected]


10/14

children? a double-blind, randomized, controlled emergency

department trial. Ann Emerg Med. 2000;36:579-588.

24. Kennedy RM, McAllister JD. Midazolam with ketamine: who

benefits [editorial]? Ann Emerg Med. 2000;35:297-299.

APPENDIX:In addition to the authors listed at the beginning of the article,

the following investigators and institutions participated in thisstudy.

Department of Emergency Medicine, Akdeniz UniversitySchool of Medicine, Antalya, Turkey: Cem Oktay; Departmentof Emergency Medicine, Queens Elizabeth II Hospital, Hertford-shire, UK: J. P. Saetta, MD, Victoria Holloway, MD; EmergencyDepartment, Princess Margaret Hospital for Children, Perth,

Australia: Peter Heinz, MD; Department of Pediatrics, Universityof Texas Southwestern Medical Center, Dallas, TX: Alan H.Bleiberg, MD; Department of Paediatrics, Starship ChildrensHospital, Auckland, New Zealand: David Herd, BSc, MBChB;Division of Pediatric Emergency Medicine, LeBonheur Chil-drens Medical Center, Memphis, TN: Sandip A. Godambe, MD,PhD, Jay Pershad, MD; Division of Emergency Medicine, St.

Louis Childrens Hospital, Washington University, St. Louis,MO: Jan D. Luhmann, MD, Robert M. Kennedy, MD; Depart-ment of Emergency Medicine, Ellis Hospital, Schenectady, NY:Robert J. Dachs, MD; Sunshine Hospital, Melbourne, Australia:Stephen J. Priestley, MD; Department of Emergency Medicine,Royal Childrens Hospital, Brisbane, Australia: Jason P. Acworth,MD.

IMAGES IN EMERGENCY MEDICINE(continued from p. 155)

DIAGNOSIS:Purpura fulminans. Purpura fulminans, a widespread ecchymosis and gangrene of the extremities, is a clinical

manifestation of acute meningococcal septicemia.1 Meningococcemia is characterized by an abrupt onset of fever

and rash. The rash is identified by petechiae, which are the most common sign, occurring in 50% to 60% ofpatients with meningococcemia.2 Petechiae are most often located on the extremities and trunk but may involveany body part. As the disease progresses, pustules, bullae, and hemorrhagic lesions with central necrosis maydevelop. The presence of stellate purpura with a central gunmetal-gray hue is highly suggestive ofmeningococcemia (Figure 1).3 Large purpuric lesions with necrotic areas are associated with disseminatedintravascular coagulation and are characteristic of purpura fulminans, which is often associated with the rapidonset of hypotension, acute adrenal hemorrhage (Waterhouse-Friderichsen syndrome), and multiorgan failure.1

Vascular complications can lead to the loss of digits or limbs (Figure 2), leaving survivors severely handicapped.1,2

The diagnosis of meningococcemia is often made clinically; however, it can be confirmed by isolating Neisseriameningitidisfrom blood cultures. Management includes supportive care and prompt administration of antibioticsonce the diagnosis is suspected.4 Patients should be cared for in the ICU. Despite antimicrobial therapy, themortality approaches 20% to 25%, with most deaths occurring within the first 48 hou rs.5 The patient in our

scenario demonstrated positive blood culture for Neisseria meningitidis, underwent a prolonged and complicatedhospital course in the pediatric ICU, and continues to receive hemodialysis.

REFERENCES1. Rosenstein N, Perkins B. Meningococcemia. N Engl J Med. 2001;344:1378-1388.

2. Thompson M, Ninis N. Clinical recognition of meningococcal disease in children and adolescents. Lancet. 2006;367:397-

403.

3. Darmstadt GL. Acute infectious purpura fulminans: pathogenesis and medical management. Pediatr Dermatol. 1998;15:

169-183.

4. Kaplan S. Multicenter surveillance of invasive meningococcal infection in children. Pediatrics. 2006;118:e979-e984.

5. Leclerc F, Leteurtre S, et al. Do new strategies in meningococcemia produce better outcomes? Crit Care Med.

2000;28:S60-S63.




11/14

Appendix E1. Included ED ketamine studies (n32).

Study Format Sedations Emesis (%)

Any Recovery

Agitation (%)

Clinically Important

Recovery Agitation (%)

1997 Dachs Prospective 30 2 (6.7) 4 (13.3) 0

1998 Green Mixed 1,020 68 (6.7) 89 (8.7) 9 (0.9)

1998 Green Retrospective 142 4 (2.8) 2 (1.4) 1 (0.7)1999 Pena Prospective 219 1 (0.5) 0 0

2000 Holloway Retrospective 78 9 (11.5) 0 0

2000 Sherwin Prospective 104 7 (6.7) 28 (26.9) 5 (4.8)

2001 Acworth Prospective 26 2 (7.7) 0 0

2001 Gloor Prospective 200 38 (19.0) 65 (32.5) 0

2001 Priestley Prospective 28 3 (10.7) 1 (3.6) 1 (3.6)

2002 Hostetler Prospective 290 4 (1.4) 19 (6.6) 2 (0.7)

2002 Hostetler Retrospective 44 1 (2.3) n/a n/a

2003 Agrawal Prospective 468 11 (2.4) 3 (0.6) 0

2003 Godambe Prospective 54 2 (3.7) 3 (5.6) 0

2003 Kim Prospective 20 (0) 0 0

2003 Pitetti Prospective 351 9 (2.6) 1 (0.3) 0

2004 Ellis Prospective 89 8 (9.0) 5 (5.6) 0

2004 Imak Prospective 26 6 (23.1) 11 (42.3) 2 (7.7)2004 McGlone Prospective 507 53 (10.5) 96 (18.9) 20 (3.9)

2004 Roback Prospective 1,509 145 (9.6) 37 (2.5) 6 (0.4)

2004 Treston Prospective 263 34 (12.9) 6 (2.3) 0

2005 Green Prospective 26 2 (7.7) 2 (7.7) 0

2005 Oktay Prospective 141 9 (6.4) 25 (17.7) 0

2006 Heinz Prospective 82 14 (17.1) 1 (1.2) 0

2006 Kriwanek Prospective 21 0 0 0

2006 Losek Retrospective 143 3 (2.1) 1 (0.7) 0

2006 Luhmann Prospective 52 11 (21.2) 27 (51.9) 11 (21.2)

2006 Roback Prospective 208 39 (18.8) 76 (36.5) 23 (11.1)

2006 Wissler Retrospective 453 100 (22.1) 41 (9.1) 14 (3.1)

2007 Bleiberg Retrospective 72 3 (4.2) 1 (1.4) 0

2007 Herd Prospective 60 12 (20) 5 (8.3) 1 (1.7)

2008 Brown Prospective 1,085 77 (7.1) 79 (7.3) 20 (1.8)

2008 McKee Retrospective 471 17 (3.6) 2 (0.4) 0Total* 8,282 694 (8.4) 630 (7.6) 115 (1.4)

n/a; Not applicable.

*The denominator for recovery agitation is 8,238.

Volume , . : August Annals of Emergency Medicine 180.e1


12/14

Appendix E2. Unadjusted association of clinical variables with emesis (n8,282).

Characteristic Emesis (n694) No Emesis (n7,588) Difference (95% CI)

Age, y Mean 8.2

Median 7.9

IQR 5.1, 11.4

Mean 6.2

Median 5.3

IQR 3.0, 8.9

2.0 (1.7 to 2.3)

Age group, y (%)

2 13 (1.4) 936 (98.6)2 to 13 595 (9.0) 6,044 (91.0)

13 86 (12.4) 608 (87.6)

IV route (%) 463 (66.7) 5,214 (68.7) 2.0 (1.6 to 5.7)

Initial IM dose* Mean 3.5

Median 4.0

IQR 2.5, 4.0

Mean 3.5

Median 3.9

IQR 2.5, 4.0

0 (0.2 to 0.1)

Total IM dose

Mean 3.9

Median 4.0

IQR 3.0, 4.0

Mean 3.8

Median 4.0

IQR 2.9, 4.1

0.1 (3 to 0.1)

Initial IV dose

Mean 1.9

Median 1.5IQR 1.0, 2.9

Mean 1.5

Median 1.1IQR 1.0, 1.9

0.4 (0.3 to 0.5)

Total IV dose

Mean 2.7

Median 1.8

IQR 1.0, 3.7

Mean 2.1

Median 1.5

IQR 1.0, 2.4

0.6 (0.4 to 0.8)

Total dose quintiles (%)

1 95 (7.4) 1,193 (92.6)

2 152 (7.5) 1,876 (92.5)

3 95 (6.6) 1,346 (93.4)

4 163 (8.6) 1,724 (91.4)

5 189 (11.5) 1,449 (88.5)

Oropharyngeal procedure (%) 19 (2.7) 249 (3.3) 0.6 (1.8 to 0.7)

ASA score 3 (%) 2 (0.3) 90 (1.2) 0.9 (-0.4 to 1.4)Anticholinergic (%) 397 (57.2) 5,283 (69.6) 12.4 (16.2 to 8.6)

Benzodiazepine (%) 219 (31.6) 2,521 (33.2) 1.6 (5.3 to 1.9)

CI, Confidence interval.

*Includes the 2,453 IM sedations with documented initial doses, of which there were 228 occurrences of emesis.Includes all 2,605 IM sedations, of which there were 231 occurrences of emesis.Includes the 4,291 IV sedations with documented initial doses, of which there were 399 occurrences of emesis.Includes all 5,677 IV sedations, of which there were 463 occurrences of emesis.

180.e2 Annals of Emergency Medicine Volume , . : August


13/14

Appendix E3. Unadjusted association of clinical variables with any recovery agitation (n8,282).

Characteristic

Any Recovery

Agitation (n630)

No Recovery

Agitation

(n7,608) Difference (95% CI)

Age, y Mean 6.4

Median 5.6

IQR 2.8, 9.7

Mean 6.4

Median 5.6

IQR 3.0, 9.1

0 (0.3 to 0.3)

Age group, y (%)

2 91 (9.8) 841 (90.2)

2 to 13 479 (7.2) 6,135 (92.8)

13 60 (8.7) 632 (91.3)

IV route (%) 375 (59.5) 5,258 (69.1) 9.6 (13.6 to 5.6)


Median 3.6

IQR 2.5, 4.0

Mean 3.5

Median 3.9

IQR 2.5, 4.0

0.3 (0.5 to 0.2)

Total IM dose

Mean 3.6

Median 3.9

IQR 2.5, 4.1

Mean 3.8

Median .0

IQR 3.0, 4.1

0.2 (0.4 to 0)

Initial IV dose

Mean 1.8

Median 1.5

IQR 1.0, 2.8

Mean 1.6

Median 1.1

IQR 1.0, 1.9

0.2 (0.1 to 0.4)

Total IV dose

Mean 2.9

Median 2.0

IQR 1.2, 4.0

Mean 2.1

Median 1.5

IQR 1.0, 2.4

0.8 (0.6 to 0.9)


1 82 (7.5) 1,015 (92.5)2 133 (6.1) 2,043 (93.9)

3 82 (5.7) 1,359 (94.3)

4 140 (7.5) 1,717 (92.5)

5 193 (11.6) 1,474 (88.4)

Oropharyngeal procedure (%) 11 (1.7) 213 (2.8) 1.1 (2.1 to 0)

ASA score 3 (%) 6 (1.0) 86 (1.1) 0.1 (1.0 to 0.6)

Anticholinergic (%) 434 (68.9) 5,202 (68.4) 0.5 (3.2 to 4.3)


*Includes the 2,453 IM sedations with documented initial doses, of which there were 254 occurrences of any recovery agitation.Includes all 2,604 IM sedations, of which there were 255 occurrences of any recovery agitation.Includes the 4,293 IV sedations with documented initial doses, of which there were 346 occurrences of any recovery agitation.Includes all 5,634 IV sedations with documented recovery status, of which there were 375 occurrences of any recovery agitation.

Volume , . : August Annals of Emergency Medicine 180.e3


14/14

Appendix E4. Unadjusted association of clinical variables with clinically important recovery agitation (n8,282).

Clinically Important Recovery

Agitation

Characteristic Yes (n115) No (n8,123) Difference (95% CI)

Age, y Mean 7.3

Median 6.8

IQR 3.8, 11.8

Mean 6.4

Median 5.6

IQR 3.0, 9.1

0.9 (0.1 to 1.6)

Age group, y (%)

2 16 (1.7) 916 (98.3)

2 to 13 81 (1.2) 6,553 (98.8)

13 18 (2.6) 674 (97.4)

IV route (%) 76 (66.1) 5,557 (68.4) 2.3 (11.0 to 6.4)


Median 2.5

IQR 2.3, 4.0

Mean 3.5

Median 3.9

IQR 2.5, 4.0

0.5 (0.8 to 0.1)

Total IM dose

Mean 3.3

Median 3.5

IQR 2.5, 4.0

Mean 3.8

Median 4.0

IQR 3.0, 4.1

0.5 (0.9 to 0.1)

Initial IV dose

Mean 1.7

Median 1.3

IQR 1.0, 2.0

Mean 1.6

Median 1.1

IQR 1.0, 2.0

0.1 (0.3 to 0.1)

Total IV dose

Mean 2.3

Median 1.6

IQR 1.0, 3.1

Mean 2.1

Median 1.5

IQR 1.0, 2.6

0.2 (0.5 to 0.2)


1 17 (1.5) 1,080 (98.5)2 30 (1.4) 2,146 (98.6)

3 19 (1.3) 1,422 (98.7)

4 24 (1.3) 1,833 (98.7)

5 25 (1.5) 1,642 (98.5)

Oropharyngeal procedure (%) 2 (1.7) 222 (2.7) 1.0 (3.4 to 1.4)

ASA score 3 (%) 1 (0.9) 91 (1.1) 0.2 (2.0 to 1.5)

Anticholinergic (%) 81 (70.4) 5,555 (68.4) 2.0 (6.4 to 10.5)


*Includes the 2,453 IM sedations with documented initial doses, of which there were 39 occurrences of clinically important recovery agitation.Includes all 2,604 IM sedations, of which there were 39 occurrences of clinically important recovery agitation.Includes the 4,293 IV sedations with documented initial doses, of which there were 74 occurrences of clinically important recovery agitation.Includes all 5,634 IV sedations with documented recovery status, of which there were 76 occurrences of clinically important recovery agitation.

180.e4 Annals of Emergency Medicine Volume , . : August

Documents

Green Metaanalysis 2009