critical care checklists, the keystone project, and the

Brief Reports

Critical care checklists, the Keystone Project, and the Office forHuman Research Protections: A case for streamlining the approvalprocess in quality-improvement research*

Richard H. Savel, MD, FCCM; Evan B. Goldstein, DO; Michael A. Gropper, MD, PhD, FCCP

Checklists—important remind-ers providing much-neededstructure in complex environ-ments—are one of the tools

recommended to improve patient safetyoutcomes in the intensive care unit (ICU)(1). Although checklists have been usedto enhance safety in other fields for sometime (most notably in the aerospace in-dustry and the military [2]), their pres-ence in health care has only recently be-come prominent. Implementation of ICUchecklists has been successfully docu-mented at the institutional and state lev-

els (3, 4); currently, these initiatives arebeginning to gain attention in the laypress (5, 6).

The purpose of this article is to review,discuss, and explain the distinction be-tween quality improvement (QI) and QIresearch, the controversies surroundingthe manner in which QI research is cur-rently conducted, and the role of the Of-fice for Human Research Protections(OHRP) in the structure of these trials—specifically presenting the recent case ofthe Keystone Project in Michigan, andthe temporary cessation of data collectionbecause of a perceived lack of regulatoryevaluation and proper informed consent(IC). In addition, we describe the strongreaction on the Internet to the termina-tion of this project and propose severalpossible solutions to this research conun-drum.

The Keystone Project

Dr. Peter Pronovost and his group atJohns Hopkins have been at the forefrontin the academic analysis of the use ofchecklists in the ICU in an attempt toimprove outcomes. Their landmark 2004study demonstrated that by addingchecklists to standard ICU practice, therate of catheter-related bloodstream in-

fections dropped significantly and re-mained low (3). These checklists were anattempt to enhance compliance with anapproach recommended by both the Cen-ters for Disease Control and Preventionas well as the Institute for HealthcareImprovement and included the following:hand washing, using full-barrier infec-tion precautions during catheter inser-tion, cleaning the patient’s skin withchlorhexidine, avoiding (whenever possi-ble) the femoral site for line placement,and removing unnecessary catheters in atimely fashion. This project was describedas a prospective cohort study with a con-current control ICU, and the InstitutionalReview Board (IRB) at the Johns HopkinsMedical Institutions approved the study,with a waiver for IC being granted (3).

In an analogous—albeit much larger—study, this same group worked with theMichigan Health and Hospital Associa-tion to implement checklists in ICUsthroughout that entire state (the Key-stone Project). Again, Pronovost was ableto demonstrate that the Institute forHealthcare Improvement/Centers for Dis-ease Control and Prevention checklistscould be implemented, and their use wasassociated with a persistent decrease incatheter-related bloodstream infections

*See also p. 791.From the Department of Critical Care Medicine

(RHS), Montefiore Medical Center, Bronx, NY; Depart-ment of Surgery (EBG), Maimonides Medical Center,Brooklyn, NY; and Department of Anesthesia/CriticalCare Medicine (MAG), University of California San Fran-cisco School of Medicine (UCSF), San Francisco, CA.

Dr. Gropper is the Director of Critical Care Medi-cine at UCSF, Vice Chairman of the Department ofAnesthesia for Clinical Quality, and a Professor ofAnesthesia and Physiology at UCSF.

The authors have not disclosed any potential con-flicts of interest.

For information regarding this article, E-mail:[email protected]

Copyright © 2009 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e31819541f8

Checklists have been recently promulgated as a method toenhance patient safety and improve outcomes for critically illpatients. Specifically, recent work performed by researchers fromthe Johns Hopkins Medical Institutions has demonstrated that theaddition of checklists to usual care in the intensive care unit isassociated with a decrease in the incidence of catheter-relatedbloodstream infections. Initially evaluated at the institutionallevel, this effort has been successfully expanded to the state levelas part of the Michigan Keystone Project. Although this work hasrecently received significant positive attention in the lay press,the Office for Human Research Protections—as they felt that thiswas a research project requiring Institutional Review Board ap-proval and informed consent—put the data collection on hold forlack of approval by the Institutional Review Board at the partic-

ipating hospitals in Michigan as well as for not having obtainedinformed consent from each patient and clinician involved in theproject. This article documents the recent events surrounding theKeystone Project and the response to the actions taken by theOffice for Human Research Protections in the lay press and thenew media (Internet and blogs), articulates how a determinationcan be made if a project is quality-improvement, human-subjectsresearch, or both, and proposes some solutions to create astructured approach to this kind of research in the future. (CritCare Med 2009; 37:725–728)

KEY WORDS: informed consent; quality improvement; humansubjects research; Keystone Project; catheter-related blood-stream infection; checklists; new media; blogs

725Crit Care Med 2009 Vol. 37, No. 2

(and, presumably, saved lives). The JohnsHopkins IRB made a determination thatthe study would be exempt from IRB eval-uation, and IC was waived (4). Of note,the project was performed without ap-proval from local IRBs (local approval wasnot sought) as well as in some smallhospitals where an IRB did not exist. Ad-ministrators in participating Michiganhospitals, however, were aware that thisproject was happening, as it was ap-proved by (and actually being per-formed with assistance from) the Mich-igan Health and Hospital Association.The decisions made by the Hopkins IRBto have the study be exempt from re-view and not require local IRB approvalwere (in retrospect) controversial andwill be discussed further below.

Checklists in the Lay Press

Approximately 1 year after the Key-stone Project results were published, AtulGawande, MD (a noted Harvard surgeon),wrote a piece entitled “The Checklist” inthe December 10, 2007, edition of TheNew Yorker, focusing on the important,often unrecognized, value of the multi-disciplinary critical care team; specifi-cally, he expounded on the manner inwhich the aforementioned checklists en-hance patient safety in the ICU (5). Thisarticle garnered significant discussion inboth the traditional and the new media(Internet and blogs) (7–11).

A mere 3 weeks later, in the December30, 2007, edition of the New York Times,Dr. Gawande broke yet another story re-volving around checklists and patientsafety in the ICU: the OHRP had an-nounced the termination of any furtherdata collection involving the MichiganKeystone Project (12). OHRP claimedthat by not obtaining explicit IC fromeach patient and provider involved in theproject, the Hopkins researchers had vi-olated fundamental scientific and ethicalresearch regulations (13, 14).

With this simple statement, the OHRPsparked significant heated debate, withheels dug in firmly on both sides of thecontroversy. Numerous academic expertsin the field of QI research voiced theiropinions throughout the blogosphere,siding with regional and national hospitalassociations, strongly opining that thisruling was “absurd” and should bechanged immediately (7–11, 15). Clearly,numerous healthcare organizations weretroubled by the potential negative rami-fications of this federal action: specifi-

cally, concerns were raised that QIprojects (both current and future) mightbe construed as human-subjects research(HuSR) if patient data were collected andactions taken by the result of these data.The wording of a statement from theOHRP in response to numerous letters tothe organization did little to alleviatethese concerns:

“As stated above, the regulations donot apply when institutions are only im-plementing practices to improve thequality of care. At the same time, if insti-tutions are planning research activitiesexamining the effectiveness of interven-tions to improve the quality of care, thenthe regulatory protections are importantto protect the rights and welfare of hu-man research subjects . . . ” (11, 16).

Such vague statements continue tofuel the strongly worded, somewhat emo-tional, academic debate and discourse re-garding what, in fact, constitutes QI/HuSR and what sorts of regulatoryprotections will be required.

In mid-February 2008, after muchpressure from various medical societies,including a letter signed by five medicaland nursing organizations (17), theOHRP came out with another statement,in which their stand on the issue appearsto have changed:

“We do not want to stand in the way ofquality improvement activities that poseminimal risks to subjects. . . . [OHRP]regulations provide great flexibility andshould not have inhibited this activity. . . . Such research would likely havebeen eligible for both expedited IRB re-view and a waiver of the informed con-sent requirement.

. . . the Johns Hopkins project hasevolved to the point where the interven-tion, including the checklist, is now be-ing used at certain Michigan hospitalssolely for clinical purposes, not medicalresearch or experimentation. Conse-quently, the regulations that govern hu-man subjects research no longer applyand neither Johns Hopkins nor the Mich-igan hospitals need the approval of anIRB to conduct the current phase of theproject” (18).

Although this statement articulatedthat the current phase of the KeystoneProject could resume, the letter stoppedshort of clarifying the problems that ledto the research interruption in the firstplace and did little to help guide the ap-proval process in the future.

On the surface, the actions taken bythe OHRP appear to be understandable:

they received a letter of complaint alleg-ing that research was being performedthat potentially violated federal regula-tions; they investigated and discoveredthat research was, in fact, being per-formed without IC; finally, Johns Hop-kins postponed any further data collec-tion until a full investigation could becompleted. There are significant differ-ences, however, between this project andmore traditional HuSR, including: 1) therisks to the patient of being in the project(minimal to zero); 2) the risks of notimproving compliance with patient safetyguidelines (significant); 3) the impracti-cability of performing IC (both for thepatients and potentially the clinicians)given the number of hospitals involved,the necessity of these interventions, andthe difficulties of gaining IC in criticallyill patients or their surrogates, in general;and 4) the fact that the interventionsbeing performed could have been (somewould argue should have been) intro-duced into clinical practice without ei-ther IRB approval or IC. These distinc-tions are important, and are the primaryrationale as to why this kind of QI re-search should be handled in a differentregulatory fashion.

Distinguishing QI from HuSR

Fundamentally, QI is felt to be an in-tegral part of the ongoing management ofthe system for delivering clinical care,when compared with HuSR, which can bedefined as a knowledge-seeking enter-prise that is independent of routine med-ical care (19). The essential distinction isthat QI exists to ensure that patients arereceiving a standard of care, whereas theprimary focus on HuSR is to find gener-alizable knowledge and define a new stan-dard of care (20). Clinicians are under anobligation to perform QI, and it is widelyaccepted that IC is not necessary. Forexample, ICUs collect information re-garding infection rates, complications,and medication errors— data that areconsidered confidential and subject toHealth Insurance Portability and Ac-countability Act regulations—all withoutIC (although much of the data collectedin ICUs for QI is deidentified and/or pa-tient identifiers are not included in thedata collection process). In general,HuSR does require IC, as participation isvoluntary; patients must have the abilityto opt out of HuSR if they do not wish topartake.

726 Crit Care Med 2009 Vol. 37, No. 2

What is an institution to do when aproject has components of both QI andHuSR? Certain thought-leaders feel thatsuch projects should come under the ru-bric of HuSR and be treated as such—potentially requiring IC and evaluation byan IRB (19). Determining when a projectis more than just QI and has character-istics of HuSR can sometimes be difficult.Baily and coworkers (19, 20) have maderecommendations that a study may beQI/HuSR if it has any of the followingqualities: 1) randomization of patientsinto different intervention groups to en-hance confidence in differences obscuredby nonrandom selection; 2) testing of is-sues that are beyond current science andexperience, such as new treatments; 3)delayed or ineffective feedback of datafrom monitoring the implementation ofchanges; and 4) the involvement in keyproject roles of researchers who have noongoing commitment to improvement ofthe local care situation. It has also beenrecommended—given many of the differ-ences between traditional HuSR and QI/HuSR—that institutions consider form-ing QI-IRBs, with a focused interest inevaluating these sorts of projects (19).

Once a decision has been made that aproject qualifies as QI/HuSR, a significantstep is determination of the necessity ofIC. Under current regulations, IC can bewaived if four conditions are present: 1)the research cannot “practicably be car-ried out” without the waiver; 2) the sub-jects’ rights and welfare will not be ad-versely affected; 3) the research involvesno more than minimal risk; and 4) sub-jects will be provided with additional per-tinent information after participating. Itis beyond the scope of this article to de-termine when it is “not practicable” toperform a research project without awaiver, and it should be beyond the scopeof the principal investigator of a projectto define what constitutes “minimal risk”(19). Federal agencies, such as OHRP,should provide guidance and assist re-searchers and IRBs with these determina-tions, in the form of clear, easy-to-understand, well-structured guidelines.

The crux of the ethical argument hasbeen discussed in two articles recentlypublished in the New England Journal ofMedicine by prominent ethicists from theNational Institutes of Health, and the Hast-ings Center (21, 22): that in the name ofprotecting patients, the actions of theOHRP may have had the potential to domore harm than good. Both groups cameto similar conclusions and provide a ratio-

nal, logical approach to the ethical and reg-ulatory issues surrounding this case. Thefirst issue that is addressed is as follows:should the Johns Hopkins IRB havedeemed the research to be exempt fromIRB evaluation. The consensus is that, infact, it should have undergone review.Both sets of ethicists believe (and it isthe opinion of the authors), however,that it would have been more than ac-ceptable from a regulatory and ethicalstandpoint to have it evaluated in anexpedited fashion and that approvalfrom a central IRB such as this wouldhave been sufficient: it was not deemednecessary to get approval from each andevery hospital for this particular kind ofproject.

The second, and clearly more impor-tant, issue in this case is the role of IC.Again, with consensus, these prominentbioethicists state that the way to bestanswer this question is to focus on thepurpose of IC: to protect subjects fromwhatever risks may be inherent in theresearch project itself. The use of check-lists to enhance compliance with agreed-upon patient-safety best practices posedno threat to subjects. One author goes sofar as to say that this is not to be consid-ered any kind of human subjects researchat all; rather, it is to be viewed as a com-bination of QI and research of organiza-tional function (22). Regardless of howthis kind of research is named (and as theOHRP itself eventually concluded), awaiver of IC appears to be appropriatefrom an ethical and regulatory stand-point, without putting humans at anyrisk. Perhaps, if the Hopkins IRB hadevaluated the study and provided such awaiver, the OHRP might not have inter-rupted the Keystone Project. These regu-lations, however, can often be difficult tointerpret. As stated by Baily, “The factthat a sophisticated IRB interpreted theregulations differently from the OHRP isa bad sign in itself. You know you are inthe presence of dysfunctional regulationswhen people cannot easily tell what theyare supposed to do” (22).

Although the risks to the subjectswere little to none (and it has been ar-gued that there may have been more of arisk of not being involved in the trial)(21), the presence of minimal risk shouldby no means minimize the value and ne-cessity of such research. This field ofstudy—although it is new and still in theprocess of having the optimal regulatoryapproaches worked out—is crucial to en-sure that national quality-improvement

projects have a rigorous, scientific, andevidence-based underpinning.

Conclusion andRecommendations

It is imperative that a middle-groundapproach be found for approval of large-scale QI/HuSR to prevent laudableprojects, such as the Keystone, from be-ing shut down in the future. It is oftendifficult to generate institutional supportfor local QI projects, and actions such asthose taken by the OHRP will have achilling effect on these important efforts.Our recommendations include the fol-lowing: 1) the OHRP must develop newways to help institutions streamline theapproval process of QI/HuSR by design-ing and developing structured criteria asto what constitutes “impracticability” ofprotocols without waivers of consent and“minimal risk” to the subjects, as well asproviding some guidance for multicenterQI/HuSR projects like Keystone (includ-ing an explicit policy for central IRB ap-proval of such projects); 2) more institu-tions should have formal processeswithin their IRBs to rapidly and safelyevaluate QI/HuSR, with the goal of im-proving patient safety, making the ap-proval process less onerous and more ef-ficient, while at all times protecting therights of the individual patient; and 3)hospitals too small to have their ownIRBs should be allowed to use IRBs frompredetermined designated nearby re-gional centers of excellence.

Given the intense and impassionedreaction (in the lay press, as well as onblogs throughout the Internet) both tothe impressive results shown by Dr.Pronovost’s checklist research, as wellas the termination of his importantproject by the OHRP, there is clearly areal demand for large-scale QI/HuSR toproceed— creating new, better stan-dards of care for our hospitalized pa-tients. With a sense of urgency and fo-cused determination, leaders inacademia and government must meetto create a structured approach to thiskind of research, allowing much-needed innovation to flourish unim-peded while simultaneously protectingthe rights of individual patients.

Hear Dr. Savel interview Dr. Prono-vost as part of the 100th icritical carepodcast at http://www.sccm.org/podcast


REFERENCES

1. Pronovost P, Berenholtz S, Dorman T, et al:Improving communication in the ICU usingdaily goals. J Crit Care 2003; 18:71–75

2. Cause of the Crash of the Boeing Bomber.National Museum of the United States AirForce. November 15, 1935. Available at: http://www.nationalmuseum.af.mil/factsheets/factsheet.asp?id�2478. Accessed January 23,2008

3. Berenholtz SM, Pronovost PJ, Lipsett PA, etal: Eliminating catheter-related bloodstreaminfections in the intensive care unit. CritCare Med 2004; 32:2014–2020

4. Pronovost P, Needham D, Berenholtz S, et al:An intervention to decrease catheter-relatedbloodstream infections in the ICU. N EnglJ Med 2006; 355:2725–2732

5. Gawande A: The Checklist. The New Yorker.December 10, 2007. Available at: http://www.newyorker.com/reporting/2007/12/10/071210fa_fact_gawande. Accessed January23, 2008

6. Brody JE: A basic hospital to-do list saveslives. New York Times. January 22, 2008.Available at: http://www.nytimes.com/2008/01/22/health/22brod.html?_r�1&scp�1&sq�basic�hospital�to-do�list&st�nyt&oref�slogin. Accessed January 23, 2008

7. Lockshin MD: About Research Using Patients.Michael Lockshin’s Blog;December 31, 2007.Available at: http://www.michaellockshin.com/category/1-current-comments/medical-research. Accessed January 23, 2008

8. Mahar M: Checklists Update–AdministrationStops Program. Health Beat [blog];December 30,2007. Available at: http://www.healthbeatblog.org/2007/12/checklists-upda.html. Accessed Jan-uary 23, 2008

9. Poses RM: Why was this quality improve-ment research project shut down? HealthCare Renewal [blog];December 31, 2007.

Available at: http://hcrenewal.blogspot.com/2007/12/why-was-this-quality-improvement.html. Accessed January 23, 2008

10. Sabin J: Research Ethics and Quality Im-provement. Health Care Organizational Eth-ics [blog];December 31, 2007. Available at:http://healthcareorganizationalethics.blogspot.com/2007/12/research-ethics-and-quality-improvement.html. AccessedJanuary 23, 2008

11. Wachter RM: Did I Violate Federal Regula-tions Today? (I Hope So). Wachter’s World[blog];January 16, 2008. Available at: http://www.the-hospitalist.org/blogs/wachters_world/archive/2008/01/16/checklists-redux-did-i-violate-federal-regulations-today-i-hope-so.aspx. Accessed January 23, 2008

12. Gawande A: A Lifesaving Checklist. New YorkTimes;December 30, 2007. Available at: http://www.nytimes.com/2007/12/30/opinion/30gawande.html?_r�2&oref�slogin&oref�slogin. Accessed January 23, 2008

13. Borror KC: Letter to Johns Hopkins fromOffice of Human Research Protections. Hu-man Research Protections Under Federal-wide Assurances FWA-5752, FWA-287, andFWA-3834;July 19, 2007. Available at: http://www.hhs.gov/ohrp/detrm_letrs/YR07/jul07d.pdf. Accessed January 23, 2008

14. Borror KC: Letter to Johns Hopkins fromOHRP. Human Research Protections UnderFederalwide Assurances FWA-5752, FWA-287,and FWA-3834;November 6, 2007. Available at:http://www.hhs.gov/ohrp/detrm_letrs/YR07/nov07c.pdf. Accessed January 23, 2008

15. Sisto D: HANYS blasts federal action prohib-iting use of medical checklists. Press Release:Hospital Association of New York State. Jan-uary 3, 2008. Available at: http://www.hanys.org/communications/pr/2008/upload/HANYS-Blasts-Federal-Action-Prohibiting-use-of- Med-ical-Checklists.pdf. Accessed January 23, 2008

16. Carome M: OHRP Statement Regarding The

New York Times Op-Ed Entitled “A Lifesav-ing Checklist”. Office for Human ResearchProtections-news [official website];January15, 2008. Available at: http://www.hhs.gov/ohrp/news/recentnews.html#20080115. Ac-cessed January 23, 2008

17. Hanson MD, Thomas A, Ingbar DH, et al:Concerns about a recent ruling by the Officefor Human Research Protections. [letter] bythe Society of Hospital Medicine, the Amer-ican College of Chest Physicians, the Societyof Critical Care Medicine, the American As-sociation of Critical Care Nurses, and theAmerican Thoracic Society. Letter to Mi-chael Leavitt, Secretary, US Department ofHealth and Human Services. February 5,2008. Available at: http://www.hospitalmedicine.org/AM/Template.cfm?Section�Advocacy_Policy&Template�/CM/ContentDisplay.cfm&ContentID�16830. Accessed April 21,2008

18. Pritchard I: OHRP Concludes Case Regard-ing Johns Hopkins University Research onHospital Infections. Office for Human Re-search Protections-news [official website];February 15, 2008. Available at: http://www.hhs.gov/ohrp/news/recentnews.html#20080215. Accessed April 21, 2008

19. Baily MA, Bottrell M, Lynn J, et al. The ethicsof using QI methods to improve health carequality and safety. Hastings Cent Rep 2006;36:S1–40

20. Lynn J, Baily MA, Bottrell M, et al. The ethicsof using quality improvement methods inhealth care. Ann Intern Med 2007; 146:666–673

21. Miller FG, Emanuel EJ. Quality-improve-ment research and informed consent. N EnglJ Med 2008; 358:765–767

22. Baily MA. Harming through protection?N Engl J Med 2008; 358:768–769


Unexplained hypotension: The spectrum of dynamic leftventricular outflow tract obstruction in critical care settings*

Anand Chockalingam, MD; Smrita Dorairajan, MD; Meenakshi Bhalla, MD; Kevin C. Dellsperger, MD, PhD

Dynamic left ventricular out-flow tract obstruction (LVOTO)is typically associated withhypertrophic cardiomyopa-

thy (1, 2). As systole progresses, the an-terior mitral leaflet (AML) is drawn to-ward the hypertrophied basal ventricularseptum. This leads to subaortic stenosisthat increases throughout systole pro-ducing a characteristic late peaking echoDoppler profile. Independent of hypertro-phic cardiomyopathy, left ventricular(LV) hypertrophy and increased cardiaccontractility can cause dynamic LVOTO(3–7). Besides being a coincidental phe-nomenon of academic curiosity, the fol-lowing situations illustrate the clinicallycentral role played by dynamic LVOTO in

critically ill patients. Our institutional re-view board waived the need for approvalfor this study.

Report of Patients

Patient 1. A 68-year-old white womanwith hypertension, asthma, and rheuma-toid arthritis presented to the emergencyroom with fatigue and dizziness. Dobut-amine (5 �g/kg/min) was started for hy-potension of 80/50 mm Hg. Electrocar-diogram (ECG) revealed anterior ST-Tchanges suggestive of ischemia and tro-ponin I elevation to a peak of 5.1 ng/mL.A grade 3/6 ejection systolic murmur inthe left third intercostal space and suspi-cion of cardiogenic shock lead to a bed-side echocardiogram in the emergencyroom. Two-dimensional imaging revealedoverall preserved systolic function withan ejection fraction of 55%, normal LVwall thickness without septal hypertro-phy, and systolic anterior motion (SAM)of the chordal apparatus of the AML. Dis-continuation of dobutamine, a saline in-travenous bolus, and closely supervisedadministration of 15 mg metoprolol in-travenously within 15 minutes reducedthe murmur to 1/6 intensity and theLVOTO from 150 to 40 mm Hg. Bloodpressure improved to 100/70 mm Hg andcardiac catheterization with coronary an-

giography confirmed patent coronarieswithin the hour. An echocardiogram 3months later showed persistent chordalSAM with peak LVOT gradients of 50 mmHg. The patient did well clinically onatenolol 100 mg/day with arthritis-related functional limitations only. A yearlater, she underwent a lengthy surgeryfor vertebral fracture stabilization andonce again became hypotensive. ECGconfirmed LVOTO up to 120 mm Hg andthis time she had LV dysfunction, ejec-tion fraction 30% with akinesia of thedistal and apical LV segments. She re-sponded to medical management withfluids and intravenous metoprolol onceagain. The latest echocardiogram 3months later confirmed normalization ofLV function (Fig. 1). On atenolol 100 mgdaily, her resting and stress (dobutamine50 mcg/kg/min) gradients were 30 and 50mm Hg, respectively.

Patient 2. A 79-year-old white womanwith hypertension and arthritis presentedwith exertional chest tightness and dys-pnea. Vitals were stable and a 3/6 ejectionsystolic murmur was noted at the leftsternal border. ECG showed LV hypertro-phy with T inversions and peak troponinI of 0.44 ng/mL (normal �0.05). Coro-nary angiography showed luminal irreg-ularities and an 80% lesion in mid-

*See also p. 793.From the Division of Cardiovascular Medicine, De-

partment of Internal Medicine (AC, SD, MB, KCD),University of Missouri School of Medicine, Columbia,MO; and Cardiology Section (AC), Harry S Truman VAMedical Center, Columbia, MO.

Supported, in part, by Department of VeteransAffairs, VISN 15 Research Award.

The authors have not disclosed any potential con-flicts of interest.

For information regarding this article, E-mail:[email protected]

Copyright © 2009 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e3181958710

Objective: To illustrate the clinical and hemodynamic abnor-malities caused by dynamic left ventricular outflow tract obstruc-tion (LVOTO) in critical care setting.

Design: We reviewed cases referred to Cardiology with echo-cardiographic evidence of LVOTO and their clinical presentations.We present those cases where LVOTO can transiently occur with-out hypertrophic cardiomyopathy when inotropic agents are usedfor hypotension.

Measurements and Main Results: Five women in the 50–70age range and prior history of hypertension presented with vari-ous symptoms like chest discomfort, fatigue, dizziness, atrialfibrillation, and hypotension. An ejection systolic murmur wasnoted most often in the left third intercostal space and ECGrevealed ST-T wave abnormalities. LVOTO caused by mitral sys-tolic anterior motion was detected by echocardiography and

catheterization excluded acute coronary disease. In critical caresetting, LVOTO can occur due to apical ballooning syndrome,coronary disease, medications, volume depletion, and valvularabnormalities. Because this condition mimics acute coronarysyndrome or other etiologies of hypotension in medical and sur-gical intensive care units, appropriate treatment can be delayed.Nonhypertrophic cardiomyopathy LVOTO usually responds well tofluid replacement, beta blockers, and medication changes.

Conclusions: LVOTO should be suspected especially in womenpresenting with hypotension and systolic murmur in critical caresettings. Clinical acumen and timely echocardiography are re-quired to effectively counter this transient but potentially lethalproblem. (Crit Care Med 2009; 37:729–734)

KEY WORDS: left ventricular outflow tract obstruction; acutemyocardial infarction; hypotension; shock


posterior descending artery. Catheterpullback and echocardiogram confirmedapical ballooning syndrome (ABS) withLVOTO and a peak gradient of 60 mmHg. We believe in this patient LV basalhypercontractility as a compensatorymechanism for the ABS-related severeapical dysfunction resulted in transientSAM and development of LVOTO. Metopro-lol 50 mg twice daily was initiated and re-peat echocardiogram 3 months later re-vealed normalization of LV function and noevidence for LVOT obstruction.

Patient 3. A 52-yr-old white womanwith history of hypertension and depres-sion presented with anginal discomfortthat started 18 hours earlier. Vitals werestable and a grade 2/6 ejection systolicmurmur was noted at the left sternalborder. ECG showed anterior deep T in-versions and peak troponin I of 0.95ng/mL (normal �0.05). Coronary angiog-raphy showed normal coronaries with se-

vere LV dysfunction suggestive of ABSwith pullback LVOT gradient of 70 mmHg. Echocardiogram showed basal hyper-contractility and SAM of the AML withsevere LVOTO. She was discharged thenext day on metoprolol, and repeat echo-cardiogram in 2 weeks showed normal LVfunction and absent LVOTO.

Patient 4. A 76-year-old woman wasadmitted to the intensive care unit withischemic cerebrovascular accident andintubation for airway protection. Cardiol-ogy was consulted on the 3rd hospital dayfor new onset tachycardia and hypoten-sion. She was found to be in rapid atrialfibrillation with ventricular rate of about150/min, blood pressure of 80/50 mm Hgwith a 3/6 ejection systolic murmurnoted in the left 3rd intercostal space.Troponin was borderline at 0.10 ng/mL(normal �0.05). Bedside echocardiogramshowed SAM of the AML with LVOTO anda peak gradient of 145 mm Hg. Digoxin

was discontinued because theoreticallyits inotropic action could aggravateLVOTO. Intravenous normal saline wasinitiated totaling 3–4 L/day over the next3 days along with metoprolol 5 mg intra-venously every 4–6 hours. Hypotensionresolved in 1 hour, atrial fibrillation re-verted to sinus rhythm spontaneously in3 hours. LVOTO murmur persisted fortwo more days. Repeat echocardiogramafter 2 weeks showed no evidence forLVOTO.

Patient 5. A 79-year-old white womanwas evaluated for new onset systolic mur-mur with hypotension in the surgical in-tensive care unit. She had undergone re-peated laparotomies for colon cancer andwas intubated for airway support on day21 of hospitalization. Echocardiogramshowed a small LV chamber, mild con-centric LV hypertrophy, hyperdynamicLV contractility with ejection fractionover 85%. Color Doppler suggested mid-

Figure 1. Top: Patient 1 Doppler echo image showing peak left ventricular outflow tract (LVOT) gradient of 112 mm Hg at initial presentation in November,2006 with normal left ventricular (LV) function (left). Repeat Doppler study 30 minutes later with reduction of peak gradient to 34 mm Hg of Dobutamine,following metoprolol intravenously and saline intravenous bolus (right). Bottom: Same patient in December, 2007, 2D echocardiography in apical long-axisview end systolic frame showing chordal systolic anterior motion (SAM) and anteroapical akinesia (left). Repeat study 3 months later documentingnormalization of systolic function (right).


cavity level LV outflow obstruction andpulse wave Doppler confirmed peak gra-dient of about 40 mm Hg at the mid-cavitylevel. She responded to intravenous fluidsinitially. Three days later, she was startedon dopamine for hypotension, and murmurincreased without clinical improvement.Carefully initiating beta blockers and con-tinued fluid hydration resolved the cardiacissues over the next 3 days.

DISCUSSION

These patients represent only thosecases we recognized with dynamicLVOTO developing in the critical care en-vironment (emergency room, surgical, ormedical intensive care units). We sharethese illustrated cases in belief that pa-tients with similar presentations, if rec-ognized early and treated, can have rapid

reversal of their hypotension if dynamicLVOTO is the underlying cause. Figure 2outlines the various underlying condi-tions and characteristics that may play arole in initiating dynamic LVOTO.

Trauma, bleeding, diuretics-relatedvolume depletion, or any condition thatresults in a reduction of LV volume couldbe sufficient to reduce the LVOT area. Inpatients 4 and 5, prior unrecognized hy-povolemia might have contributed to arelative reduction in LV chamber sizethereby promoting the Venturi effect,SAM, and increasing the likelihood ofLVOTO. Severe emotional stress has beenshown to increase serum catecholaminelevels substantially in ABS (8, 9). Approx-imately, 20% of patients with ABS dem-onstrate LVOTO (7). Although the mentalor physical stress causing ABS is obscurein many instances, the resultant basalhypercontractility likely caused theLVOTO in patients 2 and 3. Acute coro-nary syndromes involving left anteriordescending artery could lead to similarsignificant dysfunction of the apical seg-ments (5, 10, 11). Basal hypercontractil-ity tries to compensate, and this maycause dynamic LVOTO in some patientswith anteroapical infarcts.

Nearly all the commonly used vaso-pressors for hypotension have significantinotropic effects and possible direct tox-icity leading to myocyte damage (12–14).When initiated before ensuring adequatevolume repletion, this could provide a

Figure 2. Patient—clinical factors that predispose and precipitate dynamic left ventricular outflowtract obstruction (LVOTO) arranged in pyramidal figure to depict prevalence.

Table 1. Summary of background issues, cardiac testing, and final diagnosis among our patients with LVOTO

Patient/Date of InitialPresentation

November 1, 2006;December, 2007 May 2, 2007 July 3, 2007 November 4, 2007 March 5, 2008

Age/sex 68/F 79/F 52/F 76/F 74/FSymptoms/initial

impressionFatigue anddizziness/ACS

Angina/ACS Angina/ACS Rapid atrial fibrillation,intubated for

intracranial bleed

Rapid atrial fibrillation,hypotension

Medical history Hypertension (HTN),rheumatoid

arthritis, asthma

HTN, arthritis HTN, depression HTN without priorvalve disease

HTN, emphysema,diabetes, and obesity

Hypotension Yes No No Yes YesPeak troponin I ng/mL 5.1 0.44 0.95 0.10 1.24Peak BNP level pg/mL 1383/2401 817 — — 3864MR grade 0–4/4 (initial/

follow-up)2/1 1/1 1/0 2/0 1/—

SEM/peak LVOTgradient mm Hg

�/150 �/60 �/70 �/145 Absent/35

RWMA (initial/follow-up) No/no (2006); Yes/no(2007)

Yes/none Yes/none None/none None/none

Final diagnosis Chordal SAM-relatedLVOTO withepisodic ABS

ABS with singlevessel CAD

ABS Stress-related isolatedtransient LVOTO

Mid-cavity obstructionwithout SAM

LVOTO, left ventricular outflow obstruction; MR, mitral regurgitation; ACS, acute coronary syndrome; SAM, systolic anterior motion; HCM,hypertrophic cardiomyopathy; ABS, apical ballooning syndrome; CAD, coronary artery disease; CHF, congestive heart failure.


“double hit”—namely hypercontractilityon an already small LV chamber—initiating LVOTO.

Another common combination is theoveruse of loop diuretics along with beta

agonists in critical care settings to relieveheart failure and bronchospasm (15). Inour experience, a history of hypertensionwas noted in nearly all the subjects withsome concentric hypertrophy. In this

substrate, any one of the above insultsmay be sufficient, but commonly wefound several contributors to dynamicLVOTO.

Clinical Recognition. Dyspnea, chestdiscomfort, and dizziness are the com-monest presenting symptoms. Tachycar-dia is usually present due to the adrener-gic excess. Hypotension and ST-T changes inthe ECG may suggest acute coronary pro-cess (ACS). The only finding that wasconsistently recognized was the systolicejection murmur (Table 1). The murmuris late peaking and maximally heard inthe left 3rd intercostal space. Valsalvausually augments this murmur, whereashandgrip and squatting may reduce themurmur intensity. Figure 3 lists thesigns and symptoms that are attributabledirectly to progressively increasing sever-ity of dynamic LVOTO.

Echocardiography is ideally suited todefine the severity of AML SAM, andDoppler analysis quantifies the LVOTO,Figure 4. Care must be taken to isolatethe LVOT Doppler signal from the con-tamination of a commonly present mitralregurgitation jet. Adjusting the gain set-tings may at times separate the lowervelocity, late peaking dynamic LVOTODoppler signal from the symmetrical mi-tral regurgitation jet. Being a variableprocess, the clinician must ensure echoimaging is done at the time of hearingthe murmur. Also, from the systolicblood pressures, peak LV subendocardialwall stress can be indirectly estimated asthe sum of the systolic blood pressureand the LVOT gradient. Markedly ele-vated wall stress may unfavorably shiftthe oxygen supply–demand curve even inthe presence of normal coronary arteries.This acute subendocardial wall stressalong with microvascular endothelialdysfunction caused by hypertension (withor without LV hypertrophy), diabetes, hy-percholesterolemia, and tobacco use mayaccount for the mild to moderate tropo-nin elevations that we observed.

Cardiac catheterization may still berequired to exclude ACS. Catheter pull-back from the LV would identify the sub-aortic location and severity of LVOTO.For ACS to produce enough mid and api-cal segment dysfunction and compensa-tory basal hypercontractility to causeLVOTO, the culprit lesion has to be in theproximal-mid left anterior descendingcoronary artery territory. If significantobstructive coronary artery disease isfound, these lesions should be treated

Figure 3. Progression of clinical presentations based on increasing severity of left ventricular outflowtract obstruction (LVOTO). The pyramidal arrangement depicts typically recognized clinical signs andsymptoms we observed in our patients as depicted in bold (center). Milder forms of LVOTO may bemore frequent yet under diagnosed because of nonspecific symptoms (bottom), whereas severe LVOTOmay be rare but could cause life-threatening complications (top). VT-VF, ventricular tachycardia,ventricular fibrillation; EMD, electro mechanical dissociation; EF, ejection fraction; MR, mitralregurgitation; RWMA, regional wall motion abnormalities; SVT, supraventricular tachycardia.

Figure 4. Role of echocardiography in clarifying cardiac hemodynamics in critically ill patients.


aggressively to reduce ischemia and ulti-mately reverse the dynamic LVOTO.

Management Options. Significant LVdysfunction with an ejection fraction inthe 30% range either due to ACS or dueto ABS is common in subjects withLVOTO. In this setting, some of the ther-apies that are commonly initiated for he-modynamic instability like inotropes andintra-aortic balloon may initiate or aggra-vate LVOTO. In many instances, it maynot be possible to differentiate betweenLVOTO that is caused by apical dysfunc-tion from the coincidental LVOTO that iswell known to occur while on inotropes.Frequent auscultation is a cornerstonefor diagnosis and treatment as well ascorrelating echo LVOTO gradients withthe clinical situation and reviewingclosely the management steps, becausepresentation may favor one vs. the otheras the primary issue. Mitral regurgita-tion, focal and global LV dysfunction, pul-

monary edema, hypotension, and comor-bidities interact with LVOTO and each ofthese needs to be independently ad-dressed to optimize overall outcome. Ta-ble 2 outlines the interventions and theirmechanism of reducing the LVOTO. Longterm, patients should be counseled toavoid dehydration and medical recordsshould list inotropes in allergy section tominimize future indiscriminate use.

Long-Term Prognosis. Dobutaminestress echocardiography literature doesnot suggest a higher cardiovascular eventrate for “incidental” occurrence ofLVOTO in the absence of CAD. ABS alsoseems to have reasonably good prognosis.During dobutamine stress testing, an-teroapical akinesia occurred with devel-opment of SAM and LVOTO in one in-stance with normalization of wall motionafter SAM resolved. Coronaries were nor-mal and blood pressure response to stresswas not hypertensive. Thus, the likely

mechanism to explain the transient myo-cardial “stunning” was sudden increase inwall stress caused by LVOTO. Most of thecommon cardiac screening tools, cathe-terization, and even autopsy may missthis diagnosis. We suspect “silent” dy-namic LVOTO could account for some ofthe excess unstable angina, ACS, andeven sudden cardiac deaths that occur inwomen with normal coronaries (5, 10,11, 15). Maintenance on long-term betablockers or calcium channel blockersmay help reduce recurrence of this prob-lem. Regular exercise may improve vagaltone and blood pressure managementwith agents that are known to regress LVhypertrophy may prevent future LVOTOrecurrence.

CONCLUSIONS

In the emergency room and in criticalcare units, dynamic LVOTO occurs more

Table 2. Various immediate and long-term measures to counter dynamic LVOTO

Measures for Immediate Relief Long-Term Options Comments

IV beta blockers Maintenance oral beta blockers Directly reduce left ventricular outflow tractgradients by

1. Reduced inotropy2. Slowing of heart rate3. Reduced Venturi forces and systolic

anterior motion4. Blunting of catecholamine over activity

(in apical ballooning syndrome)IV Verapamil or diltiazem Oral Verapamil or Diltiazem Effects similar to beta blockers and may be

better tolerated in COPD and asthmasubjects

IV fluids Ensure adequate oral hydrationalways

1. Replenishes intravascular volume2. Increase in volume reduces LVOTO

Discontinue inotropes Avoid digoxin and cardiac stimulants Directly affects inotropy and LVOTODiscontinue � agonist levalbuterol only

if other measures fail to controlbronchospasm

Minimize � agonist use inbronchospasm/COPD

Directly affects inotropy and LVOTO

Discontinue diuretics and nitrates Minimize to as needed use ofdiuretics and nitrates

Increases filling and cavity size

Thrombolytics/percutaneous coronaryintervention/coronary artery bypassgraft

Optimize coronary artery diseasesecondary prevention

Will relieve basal hypercontractility byaugmenting apical function in coronaryartery disease

Disopyrimide or amiodarone Disopyrimide or amiodarone 1. Control tachyarrhythmias—like atrialfibrillation

2. May improve LVOTO—based onhypertrophic obstructive cardiomyopathyliterature

Phenylephrine infusion Blood pressure improvement by selective �agonist without increasing inotropy

Mitral valve repair/replacement Rarely indicated—for persistent or recurrentsymptoms

Hypertension—aggressive control�130/80 mm Hg

Maximize left ventricular hypertrophyregression

Regular exercise Improves vagal toneStress reduction, meditation, and

counselingReduce catecholamine surge

Heart rate reduction �60 bpm Reduces susceptibility for LVOTO

LVOTO, left ventricular outflow tract obstruction; COPD, chronic obstructive pulmonary disease; IV, intravenous.


often than is recognized. Although it maysometimes be an “innocent bystander,” itmay generate severe aortic stenosis typephysiology in ventricles that have hith-erto not been exposed to such acute pre-load elevations. Early recognition and ap-propriate medical management areessential to optimize the clinical outcomein these settings.

REFERENCES

1. Braunwald E, Lambrew CT, Rockoff SD, et al:Idiopathic hypertrophic subaortic stenosis. I.A description of the disease based upon ananalysis of 64 patients. Circulation 1964;30(Suppl 4): 3–119

2. Maron MS, Olivotto I, Betocchi S, et al: Effectof left ventricular outflow tract obstructionon clinical outcome in hypertrophic cardio-myopathy. N Engl J Med 2003; 348:295–303

3. Pellikka PA, Oh JK, Bailey KR, et al: Dynamicintraventricular obstruction during dobut-amine stress echocardiography: A new obser-vation. Circulation 1992; 86:1429–1432

4. Haley JH, Sinak LJ, Tajik AJ, et al: Dynamicleft ventricular outflow tract obstruction inacute coronary syndromes: An importantcause of new systolic murmur and cardio-

genic shock. Mayo Clin Proc 1999; 74:901–906

5. Hrovatin E, Piazza R, Pavan D, et al: Dy-namic left ventricular outflow tract obstruc-tion in the setting of acute anterior myocar-dial infarction: A serious and potentially fatalcomplication? Echocardiography 2002; 19:449–455

6. Chockalingam A, Tejwani L, Aggarwal K, etal: Dynamic left ventricular outflow tract ob-struction in acute myocardial infarction withshock: Cause, effect, and coincidence. Circu-lation 2007; 116:e110–e113

7. Tsuchihashi K, Ueshima K, Uchida T, et al:Transient left ventricular apical ballooningwithout coronary artery stenosis: A novelheart syndrome mimicking acute myocardialinfarction. Angina Pectoris-Myocardial In-farction Investigations in Japan. J Am CollCardiol 2001; 38:11–18

8. Wittstein IS, Thiemann DR, Lima JA, et al:Neurohumoral features of myocardial stun-ning due to sudden emotional stress. N EnglJ Med 2005; 352:539–548

9. Sharkey SW, Lesser JR, Zenovich AG, et al:Acute and reversible cardiomyopathy pro-voked by stress in women from the UnitedStates. Circulation 2005; 111:472–479

10. Bartunek J, Vanderheyden M, de Bruyne B:Dynamic left ventricular outflow tract ob-struction after anterior myocardial infarc-

tion. A potential mechanism of myocardialrupture. Eur Heart J 1995; 16:1439–1442

11. Joffe II, Riley MF, Katz SE, et al: Acquireddynamic left ventricular outflow tract ob-struction complicating acute anterior myo-cardial infarction: Serial echocardiographicand clinical evaluation. J Am Soc Echocar-diogr 1997; 10:717–721

12. Luria D, Klutstein MW, Rosenmann D, et al:Prevalence and significance of left ventricu-lar outflow gradient during dobutamineechocardiography. Eur Heart J 1999; 20:386–392

13. Fisher AA, Davis MW, McGill DA: Acutemyocardial infarction associated with albu-terol. Ann Pharmacother 2004; 38:2045–2049

14. Cohen R, Rivagorda J, Elhadad S: Asymmet-ric septal hypertrophy complicated by dy-namic left ventricular obstruction after intra-aortic balloon counterpulsation placement inthe setting of anterior myocardial infarction.J Invasive Cardiol 2006; 18:E207–E208

15. Hochman JS, Tamis JE, Thompson TD, et al:Sex, clinical presentation, and outcome inpatients with acute coronary syndromes.Global Use of Strategies to Open OccludedCoronary Arteries in Acute Coronary Syn-dromes IIb Investigators. N Engl J Med 1999;341:226–232


critical care checklists, the keystone project, and the

Documents