a b s t r a c t

Objective: To compare tidal volume received during single rescuer ventilation with a modified bag-valve mask (BVM) with integrated internal handle vs standard BVM among healthy volunteers using a manikin model.

Methods: This study was a randomized crossover trial of adult healthcare provider volunteers performing venti- lation on a manikin. We randomized participants to perform single rescuer ventilation first using either a mod- ified BVM with integrated internal handle or a standard unmodified BVM. Participants were responsible for mask placement and delivery of 10 breaths per minute for 3 minutes as guided by a metronome. After a 3-minute rest period, they performed ventilation using the alternative device. The primary outcome measure was mean re- ceived tidal volume as measured by the manikin (IngMar RespiTrainer model). Secondary outcomes included participant reported device preference.

Results: Of 70 recruited participants, all completed the study. Mean received tidal volume was higher using the modified BVM with integrated internal handle vs standard BVM by 90 mL (95% confidence interval, 60-120 mL; Pb .0001). There were no significant differences in mean received tidal volume based on the order of study arm allocation. Eighty percent of participants reported preference for the modified BVM over the standard BVM (95% confidence interval, 70.0%-80.0%).

Conclusions: The modified BVM with integrated internal handle results in greater mean received tidal volume compared with standard BVM during single rescuer ventilation in a manikin model. This modified BVM design may prove a useful airway adjunct for ventilation.

Introduction

Background

Airway management is an important aspect of cardiopulmonary re- suscitation. Such management requires achieving adequate ventilation. The bag-valve mask (BVM) is a commonly used tool for ventilation as recommended by contemporary basic and advanced life support algo- rithms when patients have inadequate respiratory drive or airway ob- struction [1,2]. Indeed, prehospital research finds that BVM ventilation results in equivalent [3] or improved Neurologically favorable survival as compared with definitive airway management by endotracheal intu- bation [4]. This procedure entails external placement of a mask over a

? Trial Registration: NCT02792049.

?? Disclaimers: The view(s) expressed herein are those of the author(s) and do not re-

flect the official policy or position of Brooke Army Medical Center, the US Army Medical Department, the US Army Office of the Surgeon General, the Department of the Army, the Department of Defense or the US Government.

? Funding: None.

* Corresponding author at: 3551 Roger Brooke Dr., Fort Sam Houston, TX, 78216.

E-mail address: [email protected] (M.D. April).

patient’s face with adequate pressure to form an effective seal over the bridge of the nose, malar eminences, and the mandibular alveolar ridge.

Importance

Effective Ventilation and oxygenation in the emergency department (ED) and prehospital environment are of paramount importance given the association between hypoxia and hypercarbia with increased mor- bidity and mortality, particularly among patients with traumatic brain injury [5]. Despite the frequent need for this procedure during patient resuscitation, it is a challenging skill to master [6]. The principal chal- lenge relates to forming an effective mask seal over the patient’s face to deliver positive pressure breaths. The establishment of mask seal is particularly problematic for single-provider BVM ventilation when only 1 hand is available to form a seal [7-10].

Previous studies have examined the use of alternative mask designs to facilitate the formation of an effective mask seal. These studies sug- gest that novel mask designs may minimize air leak during ventilation. Although some previously studied devices include intraoral masks de- signed to bypass challenging facial anatomy [11], most masks used today are variations of the conventional external cuffed face mask

http://dx.doi.org/10.1016/j.ajem.2016.07.030 0735-6757/

with alternations designed to maximize seal over external facial features [12-15]. We sought to examine a modified BVM with an integrated solid internal handle situated inside the bag. The conceptual basis for the de- sign was the belief that one-handed mask seals suffer from the absence of bilateral downward pressure on the mask. The aim of the modified BVM is to allow the single rescuer to transmit downward pressure onto the face via the internal handle with the hand squeezing the bag.

Goals of this investigation

The aim of this manikin study is to compare mean received tidal vol- ume during ventilation with a standard unmodified BVM vs a modified BVM with integrated internal handle. We hypothesized that mean re- ceived tidal volume during ventilations performed using a standard BVM would be less than that during ventilations performed using the modified BVM.

Methods

Study design and setting

We conducted a prospective randomized, single-blind, crossover study comparing ventilation using a standard unmodified BVM vs mod- ified BVM with integrated internal handle. The study population com- prised healthcare provider volunteers from San Antonio Military Medical Center. This is an urban tertiary care hospital serving military active duty personnel, retirees, and beneficiaries in the greater San Antonio metropolitan area. The volunteers delivered respirations to a manikin model. The Brooke Army Medical Center institutional review board approved the study. The ClinicalTrials.gov registration number is NCT02792049.

Selection of participants

We recruited a convenience sample of healthy volunteers to partic- ipate in the study. Recruiting occurred by announcements to staff dur- ing teaching rounds for physicians, nurses, and medics. Inclusion criteria included healthcare providers 18 years of age or older assigned to San Antonio Military Medical Center. Eligible participants included attending physicians, residents, physician assistants, medical students, nurses, Respiratory therapists, and medics. Exclusion criteria included any physical impairments determined by either participant or investi- gator before study start to preclude the participant from performing conventional BVM ventilation.

We obtained verbal consent from all participants. No participants re- ceived any compensation for their participation. We maintained docu- mentation of participant flow in accordance with the Consolidated Standards for Reporting Trials Statement (Fig. 1) [16].

Interventions

This study compared the use of BVM devices for ventilation. For the control arm, participants used a standard unmodified Spur II BVM de- vice (Ambu, Ballerup, Denmark). For the intervention arm, participants used a Spur II BVM with the modification of an integrated solid internal handle within the bag (Fig. 2). There is a patent for this modified device with approval by the Federal Drug Administration pending. We used a randomization sequence with permuted blocks of 10. After screening, verbal consent, and enrollment, we allocated successive participants to an initial study arm. Study investigators were not privy to participant allocation (single-blinded).

Before the study started, participants completed a brief demographic questionnaire. Next, investigators obtained measurements of hand di- mensions and grip strengths. Participants then received brief refresher training on one-handed BVM ventilation technique in the form of a lec- ture lasting approximately 10 minutes. Investigators instructed

participants to use their nondominant hand to maintain head tilt and Chin lift and simultaneously achieve a mask seal using the E-C clamp technique [8,10]. Participants did not receive any specific instructions regarding how best to use the modified device.

After this course of instruction, investigators permitted participants the opportunity to practice their technique while administering breaths using both devices on the manikin model for approximately 1 minute. The model used was a RespiTrainer Advance manikin model (IngMar Medical, Pittsburgh, PA). Participants made any desired adjustments to the stretcher on which the manikin lay at this time. Investigators ad- vised participants to approach the manikin as a real patient and to ven- tilate by squeezing the bag over 1 second and then allowing complete bag inflation. The goal was visible chest rise with each breath delivered at a rate established by metronome (10 breaths per minute). Investiga- tors answered any remaining participant questions about the study or devices before study initiation.

Upon the start of the study, participants established a one-handed mask seal as described above. After establishment of a mask seal, inves- tigators started the metronome device. Participants ventilated the man- ikin at a rate of 10 breaths per minute over 3 minutes for a total of 30 breaths. Investigators permitted participants to adjust their grip and mask seals as needed during the study period. After a 3-minute washout period, participants then repeated these study procedures using the al- ternative mask. Upon study conclusion, participants completed a poststudy survey soliciting their opinions on the Alternative devices.

Methods of measurements

Investigators collected all data in a de-identified fashion. Investiga- tors measured participant demographics using hard copy, prestudy sur- veys completed by all participants. Data on these forms included job description (eg, attending physician), years of medical experience, ex- perience with BVM in real patients, hand dominance, and height. Next, investigators measured the grip strength of each participant’s nondominant hand (ie, the hand used to squeeze the bag to deliver breaths). Investigators used the mean of 3-hand grip measurements in pounds per square inch using a Jamar hydraulic hand dynamometer (Patterson Medical, Saint Paul, MN). Investigators also obtained several measurements of nondominant hand size on each participant including hand length, width, and span (Supplementary Figure).

Investigators blinded to participant intervention allocation mea- sured the primary outcome of mean received tidal volumes from the RespiTrainer Advance manikin model monitor output. Finally, investi- gators used hard copy, poststudy surveys completed by all participants to measure several secondary outcomes.

Outcomes

The primary outcome measure was mean received tidal volume for the 30 breaths delivered by the participants over 3 minutes as measured by the RespiTrainer Advance manikin model monitor output. Secondary outcomes included participant responses to 3 statements on a 1-5 Likert scale ranging from “strongly disagree” to “strongly agree.” The statements included “The modified device is easy to use.” “The modified device pro- vides a better mask seal than the standard device.” “I would be willing to use the modified device in an emergency situation.” Finally, participants selected either the modified or standard BVM in response to the question “Which device would you prefer to use in an emergency situation?”

Analysis

We based our sample size estimate on ? = .05 and ? = .20 with 2- sided testing. Based on previous research, we sought to identify a differ- ence in received tidal volume of approximately 100 mL [9] with a SD as high as 200 mL [11]. This resulted in a sample size estimate of 63 partic- ipants serving as their own controls.

Fig. 1. Consolidated Standards of Reporting Trials (CONSORT) diagram of participant enrollment, allocation, follow-up, and analysis.

Investigators double entered all hard copy case report forms into a secure Excel database (version 14, 2010, Microsoft, Redmond, WA). We subsequently exported all data to SPSS (version 22, IBM, Armonk, NY) for all statistical analyses. We compared the primary outcome of

Fig. 2. Image of Spur II BVM (Ambu, Ballerup, Denmark) modified by addition of integrated internal handle (circled).

mean received tidal volume (mL) between ventilations using the mod- ified vs standard BVMs. We planned parametric testing (paired sample t test) provided the data appeared normally distributed by both visual in- spection and Shapiro-Wilk testing.

Results

Characteristics of study participants

All 70 participants screened for study inclusion by the study investi- gators were eligible for participation. All 70 individuals agreed to partic- ipate. All enrolled participants underwent randomization to initial study arm: 35 to standard BVM and 35 to modified BVM. There were no losses to follow-up and all participants completed the study (Fig. 1).

Participant characteristics were comparable between participants allocated to the standard BVM first vs modified BVM first including sex and height. Similarly, hand characteristics were comparable be- tween both groups including all hand size measurements and grip strength. The distribution of participants by faculty positions was

Table 1

Participant characteristics for each study arm

Table 2

Study outcomes

Variables	Standard BVM	Modified BVM		Variables	Values
	first (35)	first (35)		Mean modified BVM received TV, mL (95% confidence interval)	681 (354-1025)
Sex, % female	25.7	31.4		Mean standard BVM received TV, mL (95% confidence interval)	591 (277-827)
Mean height, cm (SD)	67.5 (4.3)	68.3 (3.9)		Mean difference in received TV, mL (95% confidence interval)	90 (60-120)
Dominant hand, % right	88.6	85.7		Median modified BVM ease of use score, 1-5 Likert scale (IQR)	5 (4-5)
Mean hand length, cm (SD)	18.7 (1.2)	18.9 (1.2)		Median modified BVM better seal score, 1-5 Likert scale (IQR)	4 (3-4.75)
Mean hand width, cm (SD)	8.5 (0.7)	8.7 (0.7)		Median modified BVM willingness to use score, 1-5 Likert scale (IQR)	5 (4-5)
Mean hand span, cm (SD)	20.8 (2.1)	21.4 (2.3)		% preferring modified BVM	80
Mean hand grip, pounds per square inch (SD) 43.9 (12.0) 43.7 (11.5) Abbreviation: TV, tidal volume. Staff position, %

Physician assistant 11.4 22.9

Attending physician	5.7	8.6
Resident physician	34.3	8.6

Nurse 11.4 11.4

Medic 20 22.9

Other (eg, respiratory therapist) 17.1 25.7

innovation for increasing delivered tidal volumes by offering rescuers additional leverage with which to apply additional downward pressure with the nondominant hand during Bag-mask ventilation with the aim of maximizing mask seal. Our randomized crossover study found that

Median clinical experience, y (IQR)

Prior standard BVM use in real patients, %?

2 (1-5) 3.5 (0-6)

healthcare provider volunteers achieved higher mean received tidal volume using this modified device as compared with a standard BVM.

0-10 patients 28.6 45.7

11-30 patients 37.1 40

N30 patients 31.4 14.3

* Data missing for 1 participant in standard bag valve mask first group.

similar between both groups though there was a trend toward more years of clinical experience among participants allocated first to the modified BVM device. Previous experience with the standard BVM de- vice in real patients was similar (Table 1).

Main results

The mean received tidal volume was 681 mL for the modified BVM with integrated internal handle compared with 591 mL for the standard BVM. These data were normally distributed by both visual inspection and Shapiro-Wilk testing. The difference in means was 90 mL with 95% confidence interval 60-120 mL (Pb .0001 by paired sample t test, Table 2). The median received tidal volume was 685 mL (interquartile range [IQR], 605-771 mL) for the modified device vs 586 mL (IQR, 507.5-691 mL) for the standard device. Variation in received tidal vol- umes was comparable between the modified device (SD, 142 mL) and the standard BVM (SD, 148 mL, Fig. 3). There were no significant differ- ences in received tidal volumes between participants allocated to the modified BVM first vs patients allocated to the standard BVM first.

Sixteen of 70 participants (22.9%) achieved higher mean tidal volumes with the standard BVM. These 16 participants did not differ significantly from the other participants in terms of any baseline charac- teristics. The only other outcome measure for which this group differed from the other participants was lower mean received tidal volume by 136 (95% confidence interval, 58-215).

Participant responses to poststudy surveys yielded favorable median

5-point Likert scale scores regarding modified BVM ease of use, better seal compared with standard BVM, and participant willingness to use the modified device on a real patient (Table 2). Ultimately, 80% of partic- ipants reported a preference for the modified BVM design over a stan- dard design. There were no significant differences in participant poststudy survey responses and order of device allocation.

Discussion

Overview of findings

Bag-mask ventilation is a critical component of patient resuscitation. Effective ventilation is challenging and requires establishment of an ef- fective mask seal to minimize air leak and delivery of effective tidal volumes. Devices and strategies to facilitate formation of effective tidal volume delivery would be useful to prehospital and ED providers. A modified BVM with integrated internal handle is one potential

The modified BVM offers a simple intervention which may facilitate ventilation during resuscitation. It adds no size and minimal additional weight to the standard BVM design. Moreover, our study suggests effec- tive use of the device requires minimal training. Our study entailed a short period of didactic instruction lasting approximately 10 minutes followed by a 1-minute time period to allow hands on use and experi- mentation. We did not provide participants with any specific tutorial re- garding how best to optimize ventilation using the modified design. Despite this minimal experience and practice, our participants achieved significantly higher tidal volumes with this device compared with a standard BVM with which they had received previous training.

The intraoral mask design overcomes several limitations in the stan- dard BVM. The standard BVM relies entirely upon the dominant hand to establish a mask seal, whereas the only role of the nondominant hand is to squeeze the bag to deliver breaths. The literature clearly highlights the limitations and difficulties of a one-handed mask seal [7-10]. The modified device we studied seeks to augment the one-handed mask seal by allowing additional contribution to the mask seal by the contra- lateral hand via downward pressure transmitted through the integrated internal handle. Our study finds that this design is superior for deliver- ing effective ventilation in a manikin model. We suspect that this find- ing is due to improved mask seal though our data can make only a limited case for this explanation.

Of note, 16 (22.9%) of our participants achieved higher mean re- ceived tidal volume using the standard BVM. Although mean received tidal volumes delivered by standard BVM were similar between this

Fig. 3. Received mean volumes during single rescuer ventilation using a manikin model (30 breaths delivered over 3 minutes). The vertical axis represents mean received tidal volume (mL) as measured by by the RespiTrainer Advance manikin model. The horizontal axis represents individual study participants (n = 70). The participant numbers reflect as- cending mean received tidal volumes with the standard BVM and do not reflect chrono- logical order of study participation. Each pair of bars represents data for an individual participant. The gray bars represent mean received tidal volumes using a standard BVM. The black bars represent mean received tidal volumes using an intraoral mask.

group and the other participants, this group achieved significantly lower mean received tidal volumes with the modified device. Our ob- servation was that these participants did not leverage downward pres- sure with their dominant hand during the study. Without such pressure, the integrated internal handle provided no benefit in terms of mask seal and in fact has the potential to compromise ventilation by interfering with bag squeezing. We suspect these participants represent a subset of our study population who would most benefit from specific instruc- tion regarding how best to leverage the integrated internal handle to maximize ventilation.

Limitations

This study had several limitations. We used a manikin model to sim- ulate ventilation in living patients. This design allowed us to test our hy- pothesis in a highly controlled environment and mimics the approach taken by other studies in the airway literature [9,17-20]. Nevertheless, it is unclear whether our results are generalizable to living patients. In particular, our study cannot address the utility of the modified BVM de- vice in patients with difficult airway characteristics such as redundant soft tissue due to obesity or craniofacial abnormalities.

A related limitation is our controlled study environment. Our study setting does not reflect the stresses and challenges of performing venti- lation during resuscitation. Our participants maintained mask seals in a quiet room with excellent lighting free of distractions. We did not at- tempt to simulate those environments where ventilation may occur in clinical practice (eg, back of an ambulance). Moreover, our study ran only for 3 minutes, whereas real-life resuscitations may last much lon- ger, resulting in more fatigue than our participants likely experienced. Whether such fatigue might impact the increase in received tidal vol- ume achieved by the modified BVM is unclear from our results.

Finally, our results provide limited insight into the mechanisms un- derlying our findings. The modified device design seeks to maximize mask seal by offering rescuers a means to provide additional contralat- eral downward pressure to supplement that provided by the dominant hand forming a mask seal. We did not measure delivered tidal volumes

[11] and so cannot definitively speak to whether the increased tidal vol- umes measured in this study are the result of decreased mask leak or some other mechanism. However, we accounted for many potential sources of variation in delivered tidal volumes to include controlling re- spiratory rate by metronome and controlling for differences in hand size and grip strength by randomized Crossover design. Although we could not control for the potential of fatigue as the study continued, the ab- sence of any significant difference in received tidal volumes based on order of device allocation suggests that this was not a major contributor to our results.

Future research

This study suggests several areas for future research. Replication of our study with different participant population (eg, paramedics, ED physicians) and in various simulated scenarios (eg, ambulance, busy ED) would help to establish external validity. Such research might also examine the impact that education on the use of the novel device has on outcomes. Most importantly, future research would benefit from data comparing actual patient outcomes when performing ventilation using the modified device vs a standard BVM.

Another area for potential future research is replication of the study

on alternative models with various airway characteristics (eg, edentu- lous, facial trauma). Although the modified design yielded higher re- ceived tidal volumes, it is ultimately still subject to many of the same limitations and challenges associated with obtaining a mask seal as a standard BVM. Future studies should examine the device in manikins and cadavers with facial features known to complicate mask seal forma- tion (eg, excessive facial hair, obesity, and superficial facial trauma) [21]. Such studies would clarify those anatomic features for which the

modified BVM device is likely to be most beneficial. In addition, study- ing the use of bags with integrated internal handles with alternative mask designs (eg, intraoral masks [11]) could potentially identify device combinations with greater synergy for optimizing ventilation.

Conclusions

Ventilation using a modified BVM with integrated internal handle resulted in higher mean received tidal volumes compared with a stan- dard BVM in our study of healthcare providers ventilating a manikin model. Participants reported by poststudy survey that the device was generally easy to use, formed a better mask seal than the standard BVM, and was a device that most of them were willing to use in real pa- tients. Indeed, 80% of participants reported their preference for the modified device over a standard BVM with which they had previous ex- perience. These findings followed a minimal amount of training and ed- ucation. The modified BVM may provide a feasible alternative to the standard BVM for airway management during resuscitation.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2016.07.030.

Acknowledgments

The authors acknowledge the contributions of Craig Navarijo in con- ception of the study idea and provision of study materials.

References

1. Kleinman ME, Brennan EE, Goldberger ZD, Swor RA, Terry M, Bobrow BJ, et al. Part 5: adult basic life support and cardiopulmonary Resuscitation quality: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emer- gency cardiovascular care. Circulation 2015;132:S414-35.
2. Link MS, Berkow LC, Kudenchuk PJ, Halperin HR, Hess EP, Moitra VK, et al. Part 7: adult advanced cardiovascular life support: 2015 American Heart Association guide- lines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2015;132:S444-64.
3. Gausche M, Lewis RJ, Stratton SJ, Haynes BE, Gunter CS, Goodrich SM, et al. Effect of out-of-hospital pediatric endotracheal intubation on survival and neurological out- come: a controlled clinical trial. JAMA 2000;283:783-90.
4. Hasegawa K, Hiraide A, Chang Y, Brown DF. Association of prehospital advanced air- way management with neurologic outcome and survival in patients with out-of- hospital cardiac arrest. JAMA 2013;309:257-66.
5. Chi JH, Knudson MM, Vassar MJ, McCarthy MC, Shapiro MB, Mallet S, et al. Prehospital hypoxia affects outcome in patients with traumatic brain injury: a pro- spective multicenter study. J Trauma 2006;61:1134-41.
6. Komatsu R, Kasuya Y, Yogo H, Sessler DI, Mascha E, Yang D, et al. Learning curves for bag-and-mask ventilation and orotracheal intubation: an application of the cumula- tive sum method. Anesthesiology 2010;112:1525-31.
7. Davidovic L, LaCovey D, Pitetti RD. Comparison of 1- versus 2-person bag-valve- mask techniques for manikin ventilation of infants and children. Ann Emerg Med 2005;46:37-42.
8. Joffe AM, Hetzel S, Liew EC. A two-handed jaw-thrust technique is superior to the one-handed “EC-clamp” technique for mask ventilation in the apneic unconscious person. Anesthesiology 2010;113:873-9.
9. Otten D, Liao MM, Wolken R, Douglas IS, Mishra R, Kao A, et al. Comparison of bag- valve-mask hand-sealing techniques in a Simulated model. Ann Emerg Med 2014; 63:6-12 [e3].
10. Hart D, Reardon R, Ward C, Miner J. Face mask ventilation: a comparison of three techniques. J Emerg Med 2013;44:1028-33.
11. Amack AJ, Barber GA, Ng PC, Smith TB, April MD. Comparison of ventilation with one-handed mask seal with an intraoral mask versus conventional cuffed face mask in a cadaver model: a randomized crossover trial. Ann Emerg Med 2016.
12. Lee HY, Jeung KW, Lee BK, Lee SJ, Jung YH, Lee GS, et al. The performances of stan- dard and ResMed masks during bag-valve-mask ventilation. Prehosp Emerg Care 2013;17:235-40.
13. Redfern D, Rassam S, Stacey MR, Mecklenburgh JS. Comparison of face masks in the bag-mask ventilation of a manikin. Eur J Anaesthesiol 2006;23:169-72.
14. Stewart RD, Kaplan R, Pennock B, Thompson F. Influence of mask design on bag- mask ventilation. Ann Emerg Med 1985;14:403-6.
15. Wagner-Berger HG, Wenzel V, Stallinger A, Voelckel WG, Rheinberger K, Stadlbauer KH, et al. Decreasing peak flow rate with a new bag-valve-mask device: effects on respiratory mechanics, and gas distribution in a bench model of an unprotected air- way. Resuscitation 2003;57:193-9.
16. Schulz KF, Altman DG, Moher D, Group C. CONSORT 2010 statement: updated guide- lines for reporting parallel group randomised trials. PLoS Med 2010;7, e1000251.
17. Menegazzi JJ, Winslow HJ. In-vitro comparison of bag-valve-mask and the manually triggered oxygen-powered breathing device. Acad Emerg Med 1994;1:29-33.
18. Updike G, Mosesso Jr VN, Auble TE, Delgado E. Comparison of bag-valve-mask, man- ually triggered ventilator, and automated ventilator devices used while ventilating a nonintubated mannikin model. Prehosp Emerg Care 1998;2:52-5.
19. von Goedecke A, Wagner-Berger HG, Stadlbauer KH, Krismer AC, Jakubaszko J, Bratschke C, et al. Effects of decreasing peak flow rate on stomach inflation during bag-valve-mask ventilation. Resuscitation 2004;63:131-6.
20. Zecha-Stallinger A, Wenzel V, Wagner-Berger HG, von Goedecke A, Lindner KH, Hormann C. A strategy to optimise the performance of the mouth-to-bag resuscita- tor using small tidal volumes: effects on lung and gastric ventilation in a bench model of an unprotected airway. Resuscitation 2004;61:69-74.
21. Kheterpal S, Martin L, Shanks AM, Tremper KK. Prediction and outcomes of impossible mask ventilation: a review of 50,000 anesthetics. Anesthesiology 2009;110:891-7.