cpteibb

M

drps

MavMM

0d

Usefulness of Intravenously Administered Fluid Replenishmentfor Detection of Patent Foramen Ovale by

Transesophageal Echocardiography

Luis Afonso, MDa,c,*, Anupama Kottam, MDa,c, Ashutosh Niraj, MDa,c, Joya Ganguly, MDa,c,Pawan Hari, MD, MPHc, Mengistu Simegn, MDa,c, Rajeev Sudhakar, MDa, Sony Jacob, MDa,c,

Seemant Chaturvedi, MDb, Greg J. Ensing, MDd, and Theodore P. Abraham, MDe

Patent foramen ovale (PFO) is associated with cryptogenic stroke, migraine headache,decompression sickness, and platypnea-orthodeoxia syndrome. Patients undergoing trans-esophageal echocardiography are often hypovolemic from preprocedural fasting and mightnot demonstrate right to left shunting owing to insufficient right atrial pressure generation,despite provocative maneuvers. We hypothesized that volume replenishment with salineloading could potentially unmask a PFO by favorably modulating the interatrial pressuregradient. Our study sought to examine the role of pre- or intraprocedural intravenous fluidreplenishment on PFO detection during transesophageal echocardiography. A total of 103patients were enrolled. An initial series of bubble injections was performed unprovoked andthen with provocative maneuvers such as the Valsalva maneuver and coughing. Thepatients were then given a rapid 500 ml saline bolus, and the same sequence of bubbleinjections was repeated. The presence, type, and magnitude of the right to left shunts werenoted before and after the saline bolus. The detection rate of PFO increased from 10.6% to26.2% after saline loading without any provocative maneuvers. When combined withprovocative maneuvers (Valsalva or cough), saline loading improved the detection ratefrom 17.4% to 32.0%. Overall, from amongst the 103 enrolled patients, saline bolusingresulted in a de novo diagnosis of PFO in 15 patients, atrial septal aneurysm in 15, PFOcoexisting with an atrial septal aneurysm in 10, and pulmonary arteriovenous fistula in 5patients. In conclusion, saline infusion in appropriately selected patients during transesopha-geal echocardiography significantly enhances the detection of PFOs and pulmonary arterio-

venous fistulas. © 2010 Elsevier Inc. All rights reserved. (Am J Cardiol 2010;106:1054–1058)

osi

uApia

mitislgaa

a7t7

In an attempt to examine the effect of pre- or intrapro-edural intravenous fluid replenishment (saline bolus) onatent foramen ovale (PFO) detection rates, we prospec-ively enrolled 103 consecutive patients referred for trans-sophageal echocardiography (TEE) at a tertiary care med-cal center. All patients underwent a series of agitated salineubble injections (with and without provocative maneuvers)efore and after an intravenous saline bolus.

ethods

The present prospective experimental study was con-ucted from May 2008 to September 2009. The institutionaleview board reviewed and approved the protocol, in com-liance with HIPAA for investigator participation in thetudy before data use.

Divisions of aCardiology and bNeurology, cDepartment of Internaledicine, Wayne State University, Detroit, Michigan; dDivision of Pedi-

tric Cardiology, University of Michigan, Ann Arbor, Michigan; and eDi-ision of Cardiology, Johns Hopkins University, Baltimore, Maryland.anuscript received April 24, 2010; manuscript received and accepteday 18, 2010.

*Corresponding author: Tel: (313) 745-2647; fax: (313) 993-8627.

aE-mail address: lafonso@med.wayne.edu (L. Afonso).

002-9149/10/$ – see front matter © 2010 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2010.05.037

A total of 211 adult patients scheduled for TEE for a varietyf indications, as a part of their routine clinical care, werecreened for inclusion. Of these, 103 patients fulfilled thenclusion criteria and were considered in the final analysis.

The exclusion criteria (n � 108) included a left ventric-lar ejection fraction �30% (n � 32), New York Heartssociation class III-IV (n � 12), mechanically ventilatedatients (n � 20), end-stage renal disease (n � 36), knownntracardiac shunts (n � 4), inadequate opacification of righttrium with contrast (n � 2), and miscellaneous reasons (n � 2).

A detailed cross-sectional survey of paper and electronicedical records was undertaken for demographic and clin-

cal data. Upon obtaining informed consent from the pa-ients, local anesthesia and conscious sedation were admin-stered according to our laboratory routine. After intubation intandard manner, a minimum of 3 contrast injections (presalineoading) were performed as detailed in the protocol flow dia-ram (Figure 1). The patients’ vital signs were monitoredccording to the laboratory protocol throughout the procedurend during recovery.

All transesophageal echocardiographic images werecquired using a commercial ultrasound machine (Vivid, GE, Vingmed, Horten, Norway) with a GE 6 T, mul-iplane transesophageal echocardiographic probe (2.9 to.0 MHz) and images were digitally stored for off-line

nalysis. All machine settings were optimized before

www.ajconline.org

iisdvfppfli0itcbdaim

swmaotmrv

ccaaba

twsc

os

Pwao

wnblsnacgawPwe

tccps

lsp

Pn1ssdtccc

FP

Fu

1055Methods/Saline Bolus for PFO Detection

mage acquisition, and no further adjustments made dur-ng the course of the study. Digitized video loops ofufficient length (�15 beats) were acquired using stan-ard views and analyzed off-line by 2 independent re-iewers. The 90° mid-esophageal bicaval view was usedor all injections, with additional views to profile theulmonary veins as appropriate, in patients with sus-ected pulmonary arteriovenous fistulas (PAVFs). Con-icting results were adjudicated by consensus. All qual-

fying patients underwent a rapid fluid bolus of 500 ml of.9% normal saline within 25 to 30 minutes, deliveredntravenously by way of a forearm vein, with the assis-ance of a pressure bag (inflated to 300 mm Hg). Onompletion of the saline infusion, a repeat series of bub-le injections was performed in identical fashion, andigitized loops were similarly stored for comparativenalysis. Each set included a minimum of 3 injections (1njection without provocation and 2 with provocativeaneuvers, as outlined in Figure 1).For generation of saline contrast, we used two 10-ml

yringes, attached to a 3-way stopcock; 1 of the syringesas filled with 9.0 ml saline (0.9% NaCl) and 1 with 1.0l of air. After attaching the setup to the intravenous

ccess line with sterile precautions, the saline was vig-rously injected into the empty syringe back and forth 3

o 5 times to generate a uniform, opalescent suspension oficrobubbles, following which the contrast admixture was

apidly injected into the forearm vein (using a 21-gauge intra-enous catheter).

A PFO was diagnosed if passage of saline bubbleontrast was observed in the left atrium within 3 to 5ardiac cycles of complete opacification of the righttrium, if left atrial contrast was observed coincident withprovocative maneuver, or if bubble transit/exit was observedy way of the PFO tunnel, during the frame-by-frame, off-linenalysis.

Late appearance of bubble contrast in the left atrium (�6o 8 beats after right atrial opacification), or after contrastas no longer visualized in the right atrium was deemed

econdary to a PAVF. Attempts were made to independently

igure 1. Flowchart detailing “saline bubble contrast injection” protocolsed in present study.

onfirm bubble contrast emerging from the pulmonary vein S

stia in the subset of patients with suspected transpulmonaryhunting.

De novo PFO or PAVF was diagnosed when a PFO orAVF not evident in the baseline set of contrast injectionsas unmasked after fluid replenishment. An atrial septal

neurysm (ASA) was diagnosed if the maximum excursionf the mobile interatrial septum measured �1.0 cm.

The shunt size or degree of right-to-left shunting (RLS)as further semiquantitatively evaluated and stratified as 0,o shunt (�3 microbubbles); 1, small shunt (3 to 9 micro-ubbles); 2, moderate shunt (10 to 30 microbubbles); 3,arge shunt (�30 microbubbles).1 RLS was assessed in aimilar manner for both sets of bubble injections. Carefulote was again made of de novo diagnosed PFOs (detectedfter saline infusion), and the magnitude of shunting wasompared before and after saline loading. Echocardio-raphic images of patients from the study cohort weressigned to 2 certified echocardiographers (A.K. and M.S.),ho independently interpreted the presence or absence ofFO in a blinded manner. An interrater reliability analysisas performed to determine the consistency among observ-

rs using the � statistic.The mean and SDs of measurement differences for con-

inuous variables (before and after saline loading) werealculated, and a paired t test was used to test the signifi-ance of these differences. Similarly, between group com-arisons for categorical variables were made using the chi-quare test.

The detection rate of PFO on TEE before and after salineoading was calculated using the number of positive bubbletudies for each modality divided by the total number ofatients in the study cohort (n � 103).

The following estimates were made: (1) detection rate ofFO on TEE before saline loading was estimated as theumber of positive studies before saline loading divided by03 patients; and (2) the detection rate of PFO on TEE afteraline loading was estimated as the number of positivetudies after saline loading divided by 103 patients. Theetection rates of PFO on TEE with and without provoca-ive maneuvers, before and after the saline bolus, were alsoalculated in a similar manner. A 2-tailed p �0.05 wasonsidered statistically significant for all analyses. All statisti-al analyses were performed using the Statistical Package for

igure 2. Effect of saline loading on detection rate for PFO on TEE.rovoked � use of hepatic compression or Valsalva or cough maneuver.

ocial Sciences, version 15.0 (SPSS, Chicago, Illinois).

R

wtc(

sbcTit(fPp(bo

ts

sst(

bp

dssetba

ofOatutlo0

5dwtomp

btpc

D

aPn

Fostlsb“

Fn

1056 The American Journal of Cardiology (www.ajconline.org)

esults

The study cohort consisted of 103 participants (58.2%omen), with a mean age of 51.2 � 15 years. The indica-

ions for TEE included evaluations for PFO (43%), endo-arditis (32%), valvular heart disease (15%), thrombus9%), and a right atrial mass (1%).

A total of 33 patients were detected with a PFO in thistudy cohort. Bubble injections performed before the salineolus without and with provocation identified 11 and 18ases of PFO, respectively. The overall detection rate ofEE to identify PFO improved significantly after saline

nfusion (Figure 2 and Videos 1 and 2). Saline bolusing inhe absence of provocation identified a total of 27 PFOsdetection rate 26.2%); with provocation, this improvedurther to 33 cases (detection rate 32%). In all, 15 de novoFOs were identified using a 500-ml bolus in the enrolledopulation. Thus, compared to the conventional methodi.e., with provocation but without a saline bolus), a salineolus infusion with provocation improved the detection ratef PFO on TEE by 84% (from 17.4% to 32%).

Furthermore, in the small and moderate shunt category,he shunt size was exaggerated after the saline bolus in aignificant proportion of patients (Figure 3).

A similar effect of an improved detection rate was ob-erved when the maximum atrial septal excursion was mea-ured during TEE, before and after saline loading. Beforehe saline infusion, 11 patients were found to have an ASA

igure 3. Composite graph and table demonstrating effect of saline loadingn size of identified PFO. Number of PFO cases identified before and afteraline bolus cross-tabulated in rows and columns, respectively, accordingo size. De novo PFO cases and those with exaggerated shunts after salineoading shown in bold and underlined in table. First row of table lists “nohunt” category: 70 patients did not have identifiable shunts before fluidolus; after saline bolus, 6 “small shunts,” 2 “moderate shunts,” and 7large shunts” were diagnosed de novo.

maximum atrial septal excursion of �1.0 cm); after saline w

olus, de novo ASA was demonstrated in an additional 15atients, and a PFO was evident in 10 of these patients.

Before the saline bolus, 5 of the 103 patients had evi-ence of left to right shunting at the atrial level, as demon-trated by color flow Doppler. In the initial set of bubblestudies before saline infusion, except for 1 patient, none hadvidence of RLS. After the saline infusion, RLS was de-ected in 4 of the 5 patients. In the patient with RLS ataseline, the magnitude of right to left shunting increasedfter saline infusion.

Before the saline bolus, an excellent agreement wasbserved with a � of 0.88 (p �0.001) and 0.77 (p �0.001)or unprovoked and provoked bubble studies, respectively.n the postsaline bolus echocardiographic images, an

greement for the presence or absence of PFO was substan-ial with a � of 0.66 (p �0.001) and 0.76 (p �0.001) fornprovoked and provoked bubble study images, respec-ively. Finally, for all the study images analyzed, the overallevel of agreement for the presence of PFO between the 2bservers was very good, with a � of 0.80 (range 0.71 to.89, p �0.001).

In clinical terms, of the 103 enrolled patients, a single00-ml saline bolus enabled us to establish a de novoiagnosis of PFO in 15 patients, ASA in 15, PFO coexistingith an ASA in 10, and PAVF in 5 patients (Figure 4). Of

he 44 patients undergoing TEE specifically for evaluationf an intracardiac shunt, a de novo diagnosis of PFO wasade in 8, ASA in 5, and PFO coexisting with ASA in 3

atients.None of the enrolled patients encountered a complication

ecause of saline loading or multiple saline contrast injec-ions, suggesting that saline loading can be safely accom-lished in a preselected population without foreseeableomplications.

iscussion

The key findings of the present study can be summarizeds follows. First, a saline bolus remarkably improved theFO detection rates by 84%, accounting for an overall deovo diagnosis in 15 patients. Second, in the 18 patients

igure 4. Bar graph summarizing utility of saline loading in diagnosis of deovo PFO and assorted clinical entities.

ith detectable PFOs at baseline, the magnitude of RLS

iin

PtAdtrcf

sphfatdtcvcru

imupowDt

Tttmsdvtt

ccs�eelsaRhatflt

tdocdcamtecedm

soicitrHusmpnmwioigui

AMo

S

Sf0

1057Methods/Saline Bolus for PFO Detection

ncreased in 13 (72%). Finally, saline bolusing also resultedn the de novo diagnosis of PAVF and ASA in a substantialumber of patients (5 and 15, respectively).

The underlying explanation for the enhanced detection ofFO in our study appears to rest on a favorable alteration of

he interatrial pressure gradient ensuing from fluid repletion.lthough the precise mechanisms for the enhanced PFOetection were not ascertained, it is tempting to speculatehat rapid fluid bolusing likely led to a subtle change in theight-to-left atrial pressure differential, increasing the pro-livity of achieving the critical threshold pressure requiredor bubble crossover.2

Several studies have suggested that PFO size (assessedemiquantitatively by the severity of shunting), stronglyredicts the stroke risk.3,4 Our findings demonstrate thatypovolemia is the proximate cause of the observed high-alse negative rate of PFO detection and also responsible forserious underestimation of shunt size when using conven-

ional TEE protocols. Of the provocative maneuvers useduring TEE, coughing has been demonstrated to be superioro the Valsalva maneuver for delineating a PFO duringontrast TEE.5 However, performing provocative maneu-ers in sedated, intubated patients with compromised glottislosure is challenging6,7; thus, our observation that fluidepletion improves the diagnostic yield for PFOs, even withnprovoked injections, is particularly relevant.

It should be remembered that although saline contrastnjections are optimally suited for RLS detection; color flowapping is of complementary value and should be routinely

sed to rule out left to right shunting from a stretched PFO,articularly in the setting of enlarged atria.8–10 The resultsf our study have also shown that the interatrial gradientith predominantly left to right shunting on color flowoppler can be overcome and bidirectional shunting rela-

ively easily detected after saline loading.A recently published descriptive study using contrast

EE (n � 30) suggested that the sensitivity for PFO detec-ion increased with an increasing number of contrast injec-ions.11 Although they offered useful insights in terms of the

ost helpful provocative maneuvers, the unique study de-ign and numerous provocative maneuvers used, made itifficult to discriminate the effect of the individual maneu-ers (because a prespecified algorithm was used each pa-ient) versus the number of contrast injections used versushe cumulative volume administered on the diagnostic yield.

Data have suggested that ASA and PFO are significantlyorrelated,12 and the combination has been associated withryptogenic stroke1,4; thus, an assiduous search for a PFOhould be undertaken when the atrial septal excursion is10.0 mm.13 Although TEE has been historically consid-

red the reference standard for PFO detection, transthoracicchocardiography with second harmonic imaging has excel-ent diagnostic accuracy,14,15 with some studies suggestinguperiority compared to TEE.16,17 However, the sensitivityppears to be largely contingent on the image quality.15,18

eal-time 3-dimensional transthoracic echocardiographyas shown promise and recently made its debut in thisrena.19 In contrast, the data on the utility of contrast-ranscranial Doppler for PFO detection have been con-icting, with one study suggesting a complementary role

o TEE.20

PAVFs are relatively infrequent compared to PFOs. Al-hough mostly asymptomatic, they can manifest with oxygenesaturation (on pulse oximetry) and abnormal radiologicpacities on chest radiography. Less widely recognized is thelinical importance of considering PAVFs in the differentialiagnosis as the underlying cause of paradoxic embolism anderebrovascular accidents.21–23 As such, the accurate diagnosisnd differentiation from PFO are imperative, because treat-ent options differ.24 The significant rate of de novo detec-

ion of PAVFs in our study (2 of 5 patients were undergoingvaluation for cryptogenic stroke) further emphasizes thelinical relevance of a fluid bolus in patients undergoingvaluation for stroke etiology. It might be that saline loadingilates the microvascular bed to a degree, allowing theicrobubbles to pass.The interobserver variability analyses in the present

tudy revealed excellent agreement between the 2 blindedbservers. A few limitations that could have potentiallynfluenced our study findings should be acknowledged, in-luding variable operator assistance and patient effort dur-ng provocative maneuvers. Greater participation by pa-ients in the second set of contrast injections owing to partialecovery from sedation could not be entirely ruled out.owever, the detection rate improved remarkably even withnprovoked contrast injections after fluid replenishment,uggesting that patient effort was not the predominantechanism for the enhanced detection rates. None of our

atients with a PFO detected before saline infusion hadegative findings with the postbolus sequence of injections,inimizing the probability that the improved detection rateas a result of chance alone or dependent on the number of

njections used. Although the overall prevalence of PFO inur unselected study population was 32%, it should not benterpreted as representative of the PFO prevalence in theeneral population. Finally, the present study did not eval-ate the clinical outcomes or events, and the potential clin-cal implications of these findings merit future study.

cknowledgment: We are indebted to Brian Ference, MD,Phil, for his valuable suggestions and critical appraisal of

ur report.

upplementary Data

upplementary data associated with this article can beound, in the online version, at doi:10.1016/j.amjcard.2010.5.037.

1. Mas JL, Arquizan C, Lamy C, Zuber M, Cabanes L, Derumeaux G,Coste J. Recurrent cerebrovascular events associated with patent fo-ramen ovale, atrial septal aneurysm, or both. N Engl J Med 2001;345:1740–1746.

2. De Freitas FM, Faraco EZ, De Azevedo DF, Zaduchliver J, Lewin I.Behavior of normal pulmonary circulation during changes of totalblood volume in man. J Clin Invest 1965;44:366–378.

3. Hausmann D, Mugge A, Daniel WG. Identification of patent foramenovale permitting paradoxic embolism. J Am Coll Cardiol 1995;26:1030–1038.

4. Goel SS, Tuzcu EM, Shishehbor MH, de Oliveira EI, Borek PP,Krasuski RA, Rodriguez LL, Kapadia SR. Morphology of the patent

foramen ovale in asymptomatic versus symptomatic (stroke or tran-sient ischemic attack) patients. Am J Cardiol 2009;103:124–129.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

1058 The American Journal of Cardiology (www.ajconline.org)

5. Stoddard MF, Keedy DL, Dawkins PR. The cough test is superior tothe Valsalva maneuver in the delineation of right-to-left shuntingthrough a patent foramen ovale during contrast transesophageal echo-cardiography. Am Heart J 1993;125:185–189.

6. Gorgulu S, Eksik A, Eren M, Celik S, Uslu N, Yildirim A, DagdevirenB, Tezel T. Assessment of the effects of various maneuvers on bothatrial pressure changes. Int J Cardiol 2003;92:241–245.

7. Thomas JD. Patent foramen ovale with right-to-left shunting: echo-cardiographic alternatives. Eur J Echocardiogr 2001;2:74–75.

8. Schneider B, Zienkiewicz T, Jansen V, Hofmann T, Noltenius H,Meinertz T. Diagnosis of patent foramen ovale by transesophagealechocardiography and correlation with autopsy findings. Am J Cardiol1996;77:1202–1209.

9. de Belder MA, Tourikis L, Griffith M, Leech G, Camm AJ. Trans-esophageal contrast echocardiography and color flow mapping: meth-ods of choice for the detection of shunts at the atrial level? Am Heart J1992;124:1545–1550.

0. Sun JP, Stewart WJ, Hanna J, Thomas JD. Diagnosis of patent foramenovale by contrast versus color Doppler by transesophageal echocardi-ography: relation to atrial size. Am Heart J 1996;131:239–244.

1. Johansson MC, Helgason H, Dellborg M, Eriksson P. Sensitivity fordetection of patent foramen ovale increased with increasing number ofcontrast injections: a descriptive study with contrast transesophagealechocardiography. J Am Soc Echocardiogr 2008;21:419–424.

2. Fisher DC, Fisher EA, Budd JH, Rosen SE, Goldman ME. The inci-dence of patent foramen ovale in 1,000 consecutive patients: a contrasttransesophageal echocardiography study. Chest 1995;107:1504–1509.

3. Cabanes L, Mas JL, Cohen A, Amarenco P, Cabanes PA, Oubary P,Chedru F, Guerin F, Bousser MG, de Recondo J. Atrial septal aneu-rysm and patent foramen ovale as risk factors for cryptogenic stroke inpatients less than 55 years of age: a study using transesophagealechocardiography. Stroke 1993;24:1865–1873.

4. Ha JW, Shin MS, Kang S, Pyun WB, Jang KJ, Byun KH, Rim SJ, HuhJ, Lee BI, Chung N. Enhanced detection of right-to-left shunt throughpatent foramen ovale by transthoracic contrast echocardiography usingharmonic imaging. Am J Cardiol 2001;87:669–671, A611.

5. Clarke NR, Timperley J, Kelion AD, Banning AP. Transthoracic

echocardiography using second harmonic imaging with Valsalva ma-

noeuvre for the detection of right to left shunts. Eur J Echocardiogr2004;5:176–181.

6. Madala D, Zaroff JG, Hourigan L, Foster E. Harmonic imaging im-proves sensitivity at the expense of specificity in the detection of patentforamen ovale. Echocardiography 2004;21:33–36.

7. Thanigaraj S, Valika A, Zajarias A, Lasala JM, Perez JE. Comparisonof transthoracic versus transesophageal echocardiography for detectionof right-to-left atrial shunting using agitated saline contrast. Am JCardiol 2005;96:1007–1010.

8. Di Tullio M, Sacco RL, Venketasubramanian N, Sherman D, Mohr JP,Homma S. Comparison of diagnostic techniques for the detection of apatent foramen ovale in stroke patients. Stroke 1993;24:1020–1024.

9. Monte I, Grasso S, Licciardi S, Badano LP. Head-to-head comparisonof real-time three-dimensional transthoracic echocardiography withtransthoracic and transesophageal two-dimensional contrast echocar-diography for the detection of patent foramen ovale. Eur J Echocar-diogr 2010;11:245–249.

0. Devuyst G, Despland PA, Bogousslavsky J, Jeanrenaud X. Comple-mentarity of contrast transcranial Doppler and contrast transesopha-geal echocardiography for the detection of patent foramen ovale instroke patients. Eur Neurol 1997;38:21–25.

1. Li QZ, Yao F, Yang LX, Xu ZY. Paradoxical brain embolism causedby pulmonary arteriovenous fistula. J Thorac Cardiovasc Surg 2009;57:373–375.

2. Miyabe H, Ohte N, Kurokawa K, Goto T, Takada N, Yajima K, AkitaS, Mukai S, Kimura G. Paradoxical brain embolism caused by pul-monary arteriovenous fistula and coincident pulmonary embolism—acase report. Angiology 2004;55:577–581.

3. Oliveira GH, Seward JB, Stanson AW, Swanson JW. Paradoxicalcerebrovascular embolism associated with pulmonary arteriovenousfistula: contrast transoesophageal echocardiographic diagnosis. Eur JEchocardiogr 2001;2:207–211.

4. Chessa M, Drago M, Krantunkov P, Bini MR, Butera G, De Rosa G,Bossone E, Carminati M. Differential diagnosis between patent fora-men ovale and pulmonary arteriovenous fistula in two patients withprevious cryptogenic stroke caused by presumed paradoxical embo-

lism. J Am Soc Echocardiogr 2002;15:845–846.