C

Da

KRVa

b

3c

Pd

e

f

P

a

ARRA

KTRPR

cvFc

i

vP

0d

Resuscitation 81 (2010) 1544–1549

Contents lists available at ScienceDirect

Resuscitation

journa l homepage: www.e lsev ier .com/ locate / resusc i ta t ion

linical paper

oes change in thoracic impedance measured via defibrillator electrode padsccurately detect ventilation breaths in children?�

athryn Robertsa, Vijay Srinivasanb, Dana E. Nilesc,∗, Joar Eilevstjønnd, Lisa Tylere, Lori Boylea,am Bishnoi f, Susan Ferrye, Jon Nysætherd, Mette Stavlandd, Ronald S. Litmanb, Mark Helfaerb,inay Nadkarnib,c

Department of Nursing, The Children’s Hospital of Philadelphia, 34th Street and Civic Center Blvd., Philadelphia, PA 19104, United StatesDepartment of Anesthesiology, Critical Care and Pediatrics, The Children’s Hospital of Philadelphia, University of Pennsylvania School of Medicine,4th Street and Civic Center Blvd., Philadelphia, PA 19104, United StatesCenter for Simulation, Advanced Education and Innovation, Main Building, 8NW100, The Children’s Hospital of Philadelphia, 34th Street and Civic Center Blvd.,hiladelphia, PA 19104, United StatesLaerdal Medical, Tanke Svilandsgate 30, N-4002, Stavanger, NorwayDepartment of Respiratory Therapy, The Children’s Hospital of Philadelphia, 34th Street and Civic Center Blvd., Philadelphia, PA 19104, United StatesDepartment of Pediatrics, The Children’s Hospital of Philadelphia, University of Pennsylvania School of Medicine, 34th Street and Civic Center Blvd.,hiladelphia, PA 19104, United States

r t i c l e i n f o

rticle history:eceived 3 May 2010eceived in revised form 7 July 2010ccepted 19 July 2010

eywords:horacic impedanceesuscitationediatricescue ventilation

a b s t r a c t

Introduction: Resuscitation guidelines recommend rescue ventilations consist of tidal volumes7–10 ml/kg. Changes in thoracic impedance (�TI) measured using defibrillator electrode pads to detectand guide rescue ventilations have not been studied in children.Aim: We hypothesized that �TI measured via standard anterior–apical (AA) position can accurately detectventilations with volume >7 ml/kg in children. We also compared standard AA position with alternativeanterior–posterior (AP) position.Methods: IRB-approved, prospective, observational study of sedated, subjects (6 months to 17 years) onconventional mechanical ventilation. Thoracic impedance (TI) was obtained via Philips MRx defibrillatorwith standard electrode pads for 5 min each in AA and AP positions. Ventilations were simultaneouslymeasured by pneumotachometer (Novametrix CO2SMO Plus).Results: Twenty-eight subjects (median 4 years, IQR 1.7–9 years; median 16.3 kg, IQR 10.5–39 kg) wereenrolled. Data were available for 21 episodes in AA position and 22 episodes in AP position, with paired AAand AP data available for 18. For ventilations with volume <7 ml/kg, the defibrillator algorithm detected80.0% for both AA and AP (p = 0.99). For ventilations ≥7 ml/kg, detection was 95.1% for AA and 95.7% for

AP (p = 0.38).Conclusions: Changes in thoracic impedance obtained via defibrillator pads can accurately detect venti-lations above 7 ml/kg in stable, mechanically ventilated children, corresponding to rescue ventilationsrecommended during CPR. Both AA and AP pad positions were less sensitive to detect smaller volumes(<7 ml/kg) than higher volummissed. There were no differencommonly used alternative AP

Abbreviations: CPR, cardiopulmonary resuscitation; TI, thoracic impedance; �TI,hange in thoracic impedance; AA, anterior–apical; AP, anterior–posterior; Vt , tidalolume; Vts, tidal volume per kg bodyweight; TP, true positive; FP, false positive;N, false negative or undetected; ˛, impedance coefficient; ˛ω , specific impedanceoefficient.� “A Spanish translated version of the summary of this article appears as Appendixn the final online version at doi:10.1016/j.resuscitation.2010.07.010”.∗ Corresponding author at: Center for Simulation, Advanced Education and Inno-

ation, 8NW100, The Children’s Hospital of Philadelphia, 34th and Civic Center Blvd.,hiladelphia, PA 19104, United States. Tel.: +1 215 590 4039; fax: +1 267 426 2969.

E-mail address: niles@email.chop.edu (D.E. Niles).

300-9572/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.resuscitation.2010.07.010

es (≥7 ml/kg), suggesting that shallow ventilations during CPR might beces in impedance measurements between standard AA pad position andpad position.

© 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Studies of adults and children experiencing cardiac arrestdemonstrate that providers frequently deliver poor qualitycardiopulmonary resuscitation (CPR) in both in-hospital and out-

of-hospital settings.1–6 Specifically, over-ventilation during CPRis common and related to poor outcomes from cardiac arrest.7,8

Published guidelines recommend that rescue ventilations consistof tidal volumes per kilogram bodyweight (Vts) of 7–10 ml/kg foroptimal CPR and emphasize the importance of avoiding over-

itation

vCtnC

mti(siv

uhiatdlotd

2

tDIfic

2

ctgCmopmob

2

wvvppetirtvtWc

K. Roberts et al. / Resusc

entilation during CPR.9,10 While there is evidence to suggest thatPR feedback/prompt devices to guide compressions and ventila-ions may improve skill acquisition and retention by providers, it isot known if such devices improve actual patient outcomes duringPR.11

Recent improvements in defibrillator and monitor technologyake it feasible to provide audiovisual feedback on quality of CPR

o providers in both adult and pediatric settings.12–15 While stud-es in adults have demonstrated that changes in thoracic impedance�TI) measured by defibrillator electrode pads are accurate in mea-uring presence and rate of ventilations delivered during CPR,16,17 its unknown if this technology is feasible to detect and guide rescueentilations in children.

In this study, we investigated the relationship between tidal vol-me (Vt) and changes in thoracic impedance (TI) in children. Weypothesized that impedance would be predictive of tidal volume

n a clinically useful way, and in particular, delta-TI could allowccurate detection of tidal volumes less than, equal to, or greaterhan 7 ml/kg. Thus, as our primary objective, we tested whetherelta-TI measured through standard anterior–apical (AA) defibril-

ator electrode pad placement can accurately detect ventilationsf >7 ml/kg in children. We also compared the AA position withhe commonly used alternative anterior–posterior (AP) position toetect similar ventilations.

. Methods

This prospective observational study was approved by the Insti-utional Review Board at the Children’s Hospital of Philadelphia.ata were collected in compliance with the guidelines of the Health

nsurance Portability and Accountability Act to ensure subject con-dentiality. Parental/guardian informed consent and, if capable,hild assent was obtained from all subjects.

.1. Subject enrollment

Children (ages 6 months to 17 years) who were hemodynami-ally stable on conventional mechanical ventilation and admittedo the pediatric intensive care unit or undergoing an elective sur-ical procedure in the operating room were screened for inclusion.hildren with indwelling chest tubes, obvious chest wall defor-ity, skin breakdown, or inability to place pads in standard AA

r AP positions were excluded. Children whose clinical conditionrecluded them from safely or comfortably laying supine duringeasurements were also excluded. Subjects were suitably sedated

r received general anesthesia with or without neuromuscularlockade per standard clinical practice.

.2. Data collection

Demographic data including age, gender, weight and diagnosesere obtained for all subjects. All subjects were mechanically

entilated via endotracheal tubes or tracheostomy tubes using con-entional mechanical ventilators in either volume controlled orressure controlled mode. Per patient criteria and standard clinicalrotocol, patients were intubated with either a cuffed or uncuffedndotracheal tube. In those patients with a cuffed endotrachealube, the cuff of the airway was inflated to eliminate leak so that thenhaled delivered tidal volume equaled the exhaled tidal volumeecorded by the monitor. In patients with uncuffed endotracheal

ubes, the size was chosen to accomplish similar goals. The tidalolume was measured by a pneumotachometer monitoring sys-em (Novametrix CO2SMO Plus, Novametrix Medical Systems, Inc.,

allingford, CT) attached directly to the proximal end of the tra-heal tube. The CO2SMO Plus was interfaced to a local laptop

81 (2010) 1544–1549 1545

computer using Analysis Plus! (version 5.0, Novametrix MedicalSystems, Inc., Wallingford, CT) for continuous recording of ven-tilation data. Data on TI was obtained via the Philips M3536AHeartStart MRx monitor/defibrillator with commercially availableself-adhesive defibrillator pads (HeartStart Pad, M3718A – AdultRadiotransparent Multifunction Electrode Pads and M3719A –Pediatric Radiotransparent Multifunction Electrode Pads, PhilipsMedical Systems, Seattle, WA). The manufacturer specified limitof TI detection with the internal MRx ventilation detection algo-rithm and appropriately placed pads was approximately 0.4 �. TheCO2SMO Plus pneumotachometer system was calibrated on roomair per manufacturer’s instructions prior to each data collectionsession and subsequently autocalibrated periodically during thedata collection sessions. Routine maintenance checks were alsoperformed by the institution’s Biomedical Engineering departmentthroughout the course of the study. Adult pads were selected forthose patients >10 kg and pediatric-size pads were selected forthose patients ≤10 kg as recommended by the manufacturer. Padswere placed in both standard AA and commonly used alternative APpositions unless pads touched or overlapped each other in the AApad position. If the adult pads overlapped in the AA position, pedi-atric pads were utilized as recommended by the manufacturer. Ifpatients could not clinically tolerate being rolled over to place theposterior pad for the AP position, only the AA position was recordedand analyzed for that patient. Electrode pads were secured firmly tothe skin and verified to have good surface-to-skin contact. There-after, ventilation (both mechanical, as dictated by the managingclinical team, and spontaneous) and impedance data were synchro-nized and continuously recorded for 5 min in each pad location.Patient weights were obtained by the unit healthcare providers perinstitutional standards using regularly maintained and calibratedinstitutional scales prior to enrollment.

2.3. Data analysis

CO2SMO Plus data was cleaned for delivered ventilations andtimestamps by the following criteria: (1) CO2SMO Plus did notdetect ventilation due to self-calibration; (2) ventilation detectedby CO2SMO Plus was clearly an artifact. In these two situations,the corresponding MRx ventilations were also removed prior tothe analysis. Additionally, since the CO2SMO Plus only suppliesdetection times in whole seconds, some ventilation timestampshad to be manually aligned with MRx data. After confirming syn-chronization of ventilation and impedance data, the MRx andCO2SMO Plus waveforms were examined with Q-CPR Review (ver-sion 2.1.0, Laerdal Medical AS, Stavanger, Norway) and Matlab (TheMathWorks, Natick, MA) to identify the ventilations. The CO2SMOdetection of tidal volume was considered the “gold standard” fordetection of ventilation and for tidal volume. Thus, a true posi-tive (TP), i.e. a true ventilation, was defined as being a ventilationdetected by both CO2SMO Plus and MRx; a false positive (FP), i.e. afalse ventilation, was defined as a ventilation identified by MRx, butnot confirmed by CO2SMO Plus; and a false negative (FN) ventila-tion was defined as a ventilation detected by CO2SMO Plus, but notby MRx. Sensitivity, defined as the number of ventilations simul-taneously detected by CO2SMO Plus and MRx divided by the sumof the total number of ventilations detected by CO2SMO Plus andMRx; and the total number of ventilations detected by CO2SMOPlus, but not by MRx, i.e. TP/(TP + FN), was calculated per patientin each pad position. Finally, the relationship between �TI and Vt

was derived for each subject in both AA and AP pad positions. This

was defined as �TI divided by Vt and expressed as the impedancecoefficient, ˛ (m�/ml). In order to standardize the results across arange of pediatric bodyweights, we divided the �TI by volume perkilogram body weight and defined this as the specific impedancecoefficient, ˛ω (m�/(ml/kg)).

1546 K. Roberts et al. / Resuscitation 81 (2010) 1544–1549

F al voli

dmnemorsc

3

isyubSefAtscpi

j

The linear relationship between volume and thoracic impedancechange is shown for one representative patient in Fig. 2. How-ever, the slope of this relationship (i.e. the impedance coefficient

ig. 1. One minute signal example of the impedance measured by the MRx and the tidn the impedance wave is the heart beats of the patient.

Standard descriptive summaries were used for baselineemographic data. All data were reported as mean ± SD if nor-ally distributed or median with inter-quartile range (IQR) if

ot normally distributed. Ventilation and impedance param-ters were reported as the median (IQR) of each patientsedian value, unless otherwise noted. Paired data were tested

n the medians from each patient using Wilcoxon signedank test. Proportions (sensitivities) were tested using Chiquared test. P values >0.05 were considered statistically signifi-ant.

. Results

A total of 28 subjects met eligibility criteria and were enrolledn the study. Table 1 displays subject demographics and admis-ion diagnosis. Median age was 4 years (range 6 months to 17ears; IQR 1.7–9). Three subjects were excluded due to irreg-lar and dyssynchronous breathing patterns that resulted inoth CO2SMO + and MRx signals that were too noisy to analyze.imultaneous CO2SMO+ and MRx recordings (see Fig. 1 for anxample) were available for 21 subjects in the AA position andor 22 subjects in the AP position. Paired data in both AA andP positions were available for 18 subjects. There were difficul-

ies with placing pads in the AA position in one subject due tomall torso size and two subjects due to the location of theirentral venous catheters. Pads could not be placed in the AP

osition in one subject due to the location of a large spinal dress-

ng.In the AA position, aggregating all ventilations for the 21 sub-

ects, there were 1901 TP ventilations (median Vts 7.9 ml/kg (IQR,

Table 1Subject characteristics.

N = 28 subjects

Median age, years (IQR) 4 (1.7–9)Male gender, n (%) 17 (61)Median weight, kg (IQR) 16.3 (10.5–39)Diagnosis, n (%)

Respiratory 10 (36)Neurological 8 (29)Surgical 7 (35)Other 3 (10)

IQR = inter-quartile range.

ume reported by CO2SMO+. The small ripple wave superimposed on the ventilations

6.8–9.5 ml/kg)), 209 FN ventilations (median Vts 4.3 ml/kg (IQR,3.5–7.7 ml/kg)), and 0 FP ventilations. In the AP position, there were2016 TP ventilations (median Vts 8.0 ml/kg (IQR, 6.8–10.1 ml/kg)),205 FN ventilations (median Vts 4.6 ml/kg (IQR, 3.4–7.6 ml/kg)), and1 FP ventilation.

The overall sensitivity of ventilation detection by changes in tho-racic impedance in the AA position was 90.1% compared with 90.8%for the AP position (p = 0.45). Looking only at the ventilations withvolume per kilo bodyweight <7 ml/kg, the sensitivity of ventilationdetection by changes in thoracic impedance was 80.0% for both AAand AP (p = 0.99). For ventilations ≥7 ml/kg, the sensitivity of ven-tilation detection by changes in thoracic impedance was 95.1% forAA and 95.7% for AP (p = 0.38).

3.1. Relationship between volume and impedance

Fig. 2. Example of linear relationship in a patient between volume and impedancechange (AA pad position). The linear regression line (using a fixed zero intercept) isalso plotted. The slope of this line represents the specific impedance coefficient forthis patient.

K. Roberts et al. / Resuscitation 81 (2010) 1544–1549 1547

Fw(g

˛p

3

bpna

ig. 3. Relationship between specific tidal volume (milliliter per kilogram bodyeight) and change in thoracic impedance in (a) anterior–apical (AA) pad position

n = 21 patients) and (b) anterior–posterior (AP) pad position (n = 22). Each shade ofrey represents one patient.

) varies greatly from patient to patient as is shown for AA and APad position in Fig. 3(a) and (b).

.2. AA vs. AP pad position

Table 2 compares ventilation and impedance characteristicsetween the AA and AP pad positions. Eighteen patients with com-lete, paired data (i.e. both AA and AP) were included. There wereo significant differences between the two pad positions. Fig. 4(a)nd (b) shows that the impedance coefficient (˛) is correlated with

Table 2Comparison of ventilation and impedance characteristics by pad position. All tests c

AA (n = 18)

Ventilation characteristicsMedian Vt , ml (IQR) 207 (122–325)Median Vts, ml/kg (IQR) 7.5 (6.6–10.0)Median �TI, m� (IQR) 787 (568–1298)Median ˛, m�/ml (IQR) 3.5 (1.6–9.8)Median ˛ω , m�/(ml/kg) (IQR) 97.7 (74.2–126.7)

MRx detection sensitivityMedian (IQR) 0.97 (0.90–0.99)

AA, anterior–apical; AP, anterior–posterior; Vt , tidal volume; Vts, tidal volume per ki˛ω , specific impedance coefficient; IQR, inter-quartile range.

Fig. 4. (a) Relationship of impedance coefficient and patient weight inanterior–apical (AA) and anterior–posterior (AP) pad positions. (b) Logarithmic-relationship of impedance coefficient and patient weight with regression lines forAA and AP pad position.

patient bodyweight, but with no significant difference between AAvs. AP positions.

In order to determine the lowest volume per kg bodyweightthat we can likely detect using a specific impedance threshold,we calculated for each patient a “volume detection limit” for each

patient by dividing a hypothetical impedance threshold by the spe-cific impedance coefficient. Fig. 5 shows the spread across patientsof volume per kg bodyweight needed to generate an adequateimpedance change to be detected by a defibrillator for variousthreshold impedances in both AA and AP pad positions (using the

ompare AA vs. AP pad position and are paired.

AP (n = 18) p value

220 (139–322) 0.327.4 (6.5–9.9) 0.41773 (579–1316) 0.454.2 (1.8–8.2) 0.56109.5 (83.1–128.4) 0.71

0.95 (0.93–1.00) 0.72

lo bodyweight; �TI, change in thoracic impedance; ˛, impedance coefficient;

1548 K. Roberts et al. / Resuscitation

Fta

pbbip

4

cttpidotis(bmsHt

ig. 5. Boxplot (n = 18 patients) showing volume per kg body weight neededo satisfy various hypothetical detection algorithm impedance thresholds in (a)nterior–apical (AA) and (b) anterior–posterior (AP) pad positions.

aired data, n = 18). The upper whisker of the boxplot can thuse interpreted as the lowest volume per kg body weight that cane detected for the worst case patient using a specific threshold

mpedance. There was no significant difference between AA vs. APositions (p = 0.88).

. Discussion

Our study is the first to examine the feasibility and accuracy ofhanges in thoracic impedance (using standard defibrillator elec-rode pads) to detect ventilations and quantify tidal volumes inhe pediatric age group. We also examined the impact of differentad positions (AA vs. AP) to determine if there were differences

n estimation of ventilation volumes. In addition to verifying thatefibrillator electrode pads could accurately detect guideline rec-mmended ventilations in the 7–10 ml/kg range, we also observedhat there were no significant differences in ventilation detectionn AA vs. AP pad position. Both pad positions were, however, lessensitive to detect smaller volumes (<7 ml/kg) than higher volumes≥7 ml/kg), suggesting that shallow ventilations during CPR might

e missed using an impedance detection threshold of approxi-ately 0.4 �. A ventilation algorithm more sensitive to detect

maller volumes might need to be considered for pediatric use.owever, this may compromise the ability to detect true ventila-

ions versus the ability to reject false positives during CPR. The idea

81 (2010) 1544–1549

of a more sensitive pediatric ventilation algorithm is also supportedby the specific impedance coefficient being lower for children (AAmedian 98 m�/(ml/kg)) than the corresponding published adultmean 152 m�/(ml/kg).14

It is important that any technology which guides adequate ven-tilations in children is also sensitive to small but meaningful tidalvolumes (e.g. 2–6 ml/kg). Excessive rate or tidal volume (hyper-ventilation) will cause an increase in intrathoracic pressure.7,8

In a low-flow state, this increased intrathoracic pressure fromover-ventilation can reduce venous return and cardiac output inchildren and adults.8,18 In a pig cardiac arrest model, hyperventi-lation increased the mean intrathoracic pressure, decreased rightheart filling, and worsened hemodynamics and survival outcomes.8

The observed variation in thoracic impedance changes acrosssubjects was a significant challenge to quantifying tidal volumes.Although each patient had a linear relationship between volumeand thoracic impedance change, the slope of this relationship (i.e.the impedance coefficient ˛) varied greatly across the subjects forboth AA and AP pad position (Fig. 3(a) and (b)). This implies thatit is not feasible to derive tidal volume from thoracic impedanceusing a single average impedance coefficient for the pediatric pop-ulation as a whole. Unless some feasible means of calibration perpatient can be found, impedance is likely not useful as a substi-tute for highly accurate volume measurement in the setting of CPR.Similar to earlier studies,16,19 we observed an inverse relationshipbetween impedance coefficient and body weight in both AA and APpad position.

Both the detection sensitivity and the impedance responses(median �TI, ˛, ˛ω) were similar between AA and AP pad position.In fact, we found no statistically significant difference between AAvs. AP pad positions for any parameter assessed.

There were several limitations to this study. Subjects in thisstudy were not in cardiac arrest, nor receiving active chest com-pressions during measurement of thoracic impedance. This mayhave resulted in an overall “cleaner” signal than if they had beenundergoing active chest compressions. Additionally, ventilationsduring CPR may have fundamentally different impedance charac-teristics than ventilations when the chest is not being compressed.Potential sampling bias may have arisen from the inclusion of sta-ble subjects who may have had thoracic impedance values differentfrom sicker subjects. Patients with respiratory diagnoses were notevaluated for association of baseline chest impedance with extra-cellular lung water measures. This may have affected results in avery small number of subjects. The use of bag-valve-mask ventila-tion which is more widespread than invasive ventilation was notconsidered in this study. However, we believe that there would bea high correlation between the use of bag-valve-mask ventilationand mechanical ventilation as thoracic impedance is affected bychanges in lung volume. Patient weights were obtained by the unithealthcare providers per institutional standards using regularlymaintained and calibrated institutional scales. However, patientswere not re-weighed by the study team immediately prior to datacollection. There were also a number of different devices were usedto weigh patients (e.g. infant scale, bed scales, etc.) which may havecontributed to an unknown variability in measurement techniques.Finally, the investigators did not perform additional calibrationtests to the defibrillator or pneumotachometer equipment aboveand beyond manufacturer’s standard performance verification pro-cedures.

Nevertheless, we believe that the findings from this study arevery useful to further develop technologies based on thoracic

impedance to detect and guide rescue ventilations in childrenundergoing CPR. This technology lends itself to simplicity and easeof use due to its ability to be incorporated into commonly avail-able defibrillator electrode pads that are universally used duringCPR. Future studies should examine this technology both in simu-

itation

lf

midattpnv

5

ottrcaAadsTta

C

orMe

A

Retm

1

1

1

1

1

1

1

1

arrest. Resuscitation 2008;79:432–7.18. Zuercher M, Hilwig RW, Ranger-Moore J, et al. Leaning during chest compres-

K. Roberts et al. / Resusc

ated conditions as well as actual cardiac arrest to provide real-timeeedback to guide rate of rescue ventilations in children.

Finally, future defibrillator impedance measurement systemsay also be optimized for detecting ventilations. The defibrillator

mpedance measurements were originally designed for detection ofefibrillator pad connections, with detection of ventilations only assecondary development. Therefore, to obtain a more robust signal,

he impedance measurement system, including electronics (exci-ation frequency and current), defibrillation pads (size, type andosition), and the signal usage and processing (e.g. absolute versusormalized impedance change) could be optimized for detection ofentilations as a primary function.

. Conclusions

This study demonstrates that changes in thoracic impedancebtained via defibrillator pads can accurately detect and guide ven-ilations in stable, mechanically ventilated children, correspondingo rescue ventilations recommended during cardiopulmonaryesuscitation. The observed variability in the thoracic impedanceoefficient between subjects precludes the use of a single aver-ge impedance coefficient to accurately estimate tidal volumes.s lower tidal volume provides less impedance change in both AAnd AP pad positions, the study defibrillator was less sensitive toetect smaller volumes (<7 ml/kg) than higher volumes (≥7 ml/kg),uggesting that shallow ventilations during CPR might be missed.here were no differences in impedance measurements betweenhe standard anterior–apical pad position and the commonly usedlternative anterior–posterior pad position.

onflict of interest

The authors acknowledge the following potential conflictsf interest. Dana Niles and Vinay Nadkarni receive unrestrictedesearch grant support from the Laerdal Foundation for Acute Careedicine. Mette Stavland, Joar Eilevstjønn and Jon Nysæther were

mployed by Laerdal Medical during this work.

cknowledgments

We wish to thank Stephanie Tuttle MBA, Raymond MatthewsRT, and the staff of the pediatric intensive care unit and periop-rative complex at CHOP for their support and contributions tohis study. We also thank Robert Berg MD for his support on the

anuscript.

1

81 (2010) 1544–1549 1549

References

1. Abella BS, Alvarado JP, Myklebust H, et al. Quality of cardiopul-monary resuscitation during in-hospital cardiac arrest. JAMA 2005;293:305–10.

2. Wik L, Kramer-Johansen J, Myklebust H, et al. Quality of cardiopulmonary resus-citation during out-of-hospital cardiac arrest. JAMA 2005;293:299–304.

3. Abella BS, Sandbo N, Vassilatos P, et al. Chest compression rates duringcardiopulmonary resuscitation are suboptimal: a prospective study during in-hospital cardiac arrest. Circulation 2005;111:428–34.

4. Olasveengen TM, Wik L, Kramer-Johansen J, Sunde K, Pytte M, Steen PA. IsCPR quality improving? A retrospective study of out-of-hospital cardiac arrest.Resuscitation 2007;75:260–6.

5. Olasveengen TM, Vik E, Kuzovlev A, Sunde K. Effect of implementation of newresuscitation guidelines on quality of cardiopulmonary resuscitation and sur-vival. Resuscitation 2009;80:407–11.

6. Sutton RM, Maltese MR, Niles D, et al. Quantitative analysis of chest compressioninterruptions during in-hospital resuscitation of older children and adolescents.Resuscitation 2009;80:1259–63.

7. Aufderheide TP, Lurie KG. Death by hyperventilation: a common and life-threatening problem during cardiopulmonary resuscitation. Crit Care Med2004;32:S345–51.

8. Aufderheide TP, Sifurdsson G, Pirrallo RG, et al. Hyperventilation-induced hypotension during cardiopulmonary resuscitation. Circulation2004;109:160–5.

9. American Heart Association Guidelines for Cardiopulmonary. Resuscitation andemergency cardiovascular care. Circulation 2005;112:IV-1–203.

0. International Liaison Committee on Resuscitation: 2005. International con-sensus on cardiopulmonary resuscitation and emergency cardiovascu-lar care science with treatment recommendations. Circulation 2005;112:III-1–136.

1. Yeung J, Meeks R, Edelson D, Gao F, Soar J, Perkins GD. The use of CPR feed-back/prompt devices during training and CPR performance: a systematic review.Resuscitation 2009;80:743–51.

2. Sutton RM, Niles DE, Nysaether JB, et al. Quantitative analysis of CPR qualityduring in-hospital pediatric resuscitations. Pediatrics 2009;124:494–9.

3. Niles D, Nysaether J, Sutton R, et al. Leaning is common during in-hospital pedi-atric CPR, and decreased with automated corrective feedback. Resuscitation2009;80:553–7.

4. Kramer-Johansen J, Myklebust H, Wik L, et al. Quality of out-of-hospital car-diopulmonary resuscitation with real time automated feedback: a prospectiveinterventional study. Resuscitation 2006;71:283–92.

5. Abella BS, Edelson DP, Kim S, et al. CPR quality improvement during in-hospitalcardiac arrest using a real-time audiovisual feedback system. Resuscitation2007;73:54–61.

6. Losert H, Risdal M, Sterz F, et al. Thoracic impedance changes measuredvia defibrillator pads can monitor ventilation in critically ill patientsand during cardiopulmonary resuscitation. Crit Care Med 2006;34:2399–405.

7. Stecher FS, Olsen JA, Stickney RE, Wik L. Transthoracic impedance used to evalu-ate performance of cardiopulmonary resuscitation during out of hospital cardiac

sions impairs cardiac output and left ventricular myocardial blood flow in pigletcardiac arrest. Crit Care Med 2010;38:1141–6.

9. Valentinuzzi ME, Geddes LA, Baker LE. The law of impedance pneumography.Med Biol Eng 1971;9:157–63.