JACC Vol. 63, No. 17, 2014 CorrespondenceMay 6, 2014:1808–13

1809

period and another patient declined repeat invasive assessment dueto advanced age; therefore, data were available for 8 patients. Allpatients had a mean atrioventricular gradient >40 mm Hg andreceived CoreValve (Medtronic, Inc., Minneapolis, Minnesota)implantation by the transfemoral approach. Mean aortic valvepressure gradients as assessed by transthoracic echocardiographydecreased from 56.3 mm Hg (95% CI: 48.7 to 63.9 mm Hg) to11.8 mm Hg (95% CI: 8.6 to 14.9 mm Hg; p < 0.01) at 12-monthfollow-up.Using a cutoff value for the CFR ratio of 2.5 (the expected

normal), all patients had impaired CFR prior to TAVI (mean CFR1.53; 95% CI: 1.27 to 1.8). At follow-up invasive assessment ata mean of 376 days, there was a significant increase in CFR amongthe 8 patients (mean absolute increase in CFR at follow-up vs. pre-TAVI of 0.65; 95 CI: 0.36 to 0.93; p < 0.01) with a mean CFRof 2.18 (95% CI: 1.88 to 2.47). Although there was a significantchange from baseline pre-TAVI measurements to follow-up, therewas no significant improvement in CFR immediately post TAVI(mean change in CFR immediately post-TAVI vs. pre-TAVI of0.045; 95% CI: �0.4 to 0.49; p ¼ 0.41). These data are alsodemonstrated in Figures 1A and 1B, and Figure 1C demonstratesthe raw APV data pre- and post-hyperemia induction at each timepoint. There were no adverse events as a result of the invasive CFRevaluation. The vessel interrogated with the saphenous vein graftto the left circumflex artery in 6 patients (native vessels too diseasedupstream) and the native left circumflex artery in 2 patients. Twocardiologists trained in CFR assessment reviewed the raw tracesand made CFR measurements with the results compared forinterobserver variability. There were no statistically significantdifferences between the 2 sets of measurements, and both sets ofmeasurements independently (and when averaged) demonstratedthe same positive primary outcome.Our data showed that treatment of AS by relieving the me-

chanical obstruction with TAVI led to a significant improvementin CFR at medium-term follow-up. This is the first study todemonstrate this serially using invasive assessments, although it hasbeen assessed indirectly using noninvasive techniques in 2 nonin-vasive studies on patients undergoing conventional aortic valvereplacement (3,4).Limitations of this study include the small sample size. However,

the difficulty of performing the current study (serial invasivemeasurements sometimes requiring guide catheter cannulationbetween CoreValve struts) may explain why it has not been donepreviously. However, the magnitude of the change in the principalvariable is consistent with that in other previous noninvasivestudies.In conclusion, we report that the treatment of severe AS with

TAVI led to a significant improvement in coronary flow dynamicsat 12-month follow-up as measured by CFR. This potentiallyrepresents one of the key physiological pathways that is responsiblefor symptomatic and prognostic improvement of patients with ASwho are treated with TAVI.

Anthony C. Camuglia, MBBS (Hons)Jaffer Syed, MD, MEdPallav Garg, MBBS, MScBob Kiaii, MDMichael W. Chu, MD, MEdPhilip M. Jones, MD, MScDaniel Bainbridge, MD*Patrick J. Teefy, MD

*Department of Interventional CardiologyLondon Health Sciences Centre339 Windermere RoadLondon, Ontario, N6A 5A5CanadaE-mail: patrick.teefy@lhsc.on.ca

http://dx.doi.org/10.1016/j.jacc.2013.11.040

From the London Health Sciences Centre and University ofWestern Ontario, London, Ontario, Canada

Please note: Drs. Kiaii and Chu have received consulting fees from Medtronic. Dr.

Teefy received consulting fees and honoraria for lectures and served on advisory boards

for Medtronic. All other authors have reported that they have no relationships relevant

to the contents of this paper to disclose.

REFERENCES

1. Marcus ML, Doty DB, Hiratzka LF, et al. Decreased coronary reserve:a mechanism for angina pectoris in patients with aortic stenosis andnormal coronary arteries. N Engl J Med 1982;307:1362–6.

2. Garcia D, Camici PG, Durand L-G, et al. Impairment of coronary flowreserve in aortic stenosis. J Appl Physiol 2009;106:113–21.

3. Rajappan K, Rimoldi OE, Camici PG, et al. Functional changes incoronary microcirculation after valve replacement in patients with aorticstenosis. Circulation 2003;107:3170–5.

4. Hildick-Smith DJ, Shapiro LM. Coronary flow reserve improves afteraortic valve replacement for aortic stenosis: an adenosine transthoracicechocardiography study. J Am Coll Cardiol 2000;36:1889–96.

Letters to the Editor

Cardiac HepatopathyVersus End-StageLiver Disease

Two Different Entities

Kim et al. (1) have recently reported that the assessment of liverfunction may be a useful tool in the prediction of prognosis inpatients with chronic heart failure (HF). Liver dysfunction is afrequent finding in HF, especially if signs of accompanying right-sided HF are present (2). Indeed, impairment of hepatic functionis typically caused by venous congestion as a result of backwardHF (2). Congestion of the liver and intestinal wall may promoteworse outcomes due to increased intestinal translocation anddecreased hepatic clearance of bacterial endotoxin resulting inprognostically important activation of systemic inflammation (3,4).Kim et al. (1) have used the MELD (Model of End Stage Liver

Disease) scoring system to assess the liver function in patients withHF. The components of the MELD score, total bilirubin, serumcreatinine, and the international normalized ratio reflect, however,not only liver dysfunction but also the prognostically relevant renalfailure and atrial fibrillation (5–7); the last if patients are adequatelytreated with oral anticoagulants. The use of the MELD score inpatients with chronic HF should, in our opinion, be understood asassessment of all 3 comorbidities.

Correspondence JACC Vol. 63, No. 17, 2014May 6, 2014:1808–13

1810

The use of total bilirubin as a component of the MELD score iscomprehensible if applying the score to patients with liver cirrhosis,as the complex distortion of hepatic lobule results in elevation ofboth indirect and direct bilirubin. However, liver cirrhosis as anend-stage organ failure is a rare finding in patients with HFnowadays (8). The underlying mechanism responsible for abnormalliver function tests in patients with HF is primarily liver congestiondue to backward right HF. This results predominantly in mildelevation of cholestatic markers (2). Thus, direct bilirubin isprobably a more specific indicator of liver dysfunction in patientswith chronic HF than total bilirubin is, especially as total bilirubinmight be elevated due to other reasons than chronic HF. Therefore,we suggest modifying the MELD score for HF patients by usingdirect bilirubin instead of total bilirubin.As the investigators themselves state, there cannot be established

any causality between a higher MELD score and mortality and/or need for cardiac transplantation. Beyond the reported associationbetween markers of right heart function and the MELD score, itwould be helpful to provide information about the medication ofpatients with a higher MELD score. What if the score reflectedsicker patients with more medication and a variety of subsequentdrug-liver interactions and only to a lesser extent the suggestedcongestion due to HF?In accordance with Eisen (9), we think that it remains to be

elucidated whether optimization of medical therapy can lower theMELD score. Especially interventions targeting splanchniccongestion (e.g., intensified diuretic therapy) are eagerly awaited andshould be evaluated with regard to the stability of the MELD score.

*Miroslava Valentova, MDStephan von Haehling, MD, PhDStefan D. Anker, MD, PhDAnja Sandek, MD

*Applied Cachexia ResearchDepartment of CardiologyCharité Medical SchoolCampus Virchow-KlinikumAugustenburger Platz 1D-13353 BerlinGermanyE-mail: miroslava.valentova@charite.de

http://dx.doi.org/10.1016/j.jacc.2013.08.1654

REFERENCES

1. Kim MS, Kato TS, Farr M, et al. Hepatic dysfunction in ambulatorypatients with heart failure: application of the MELD scoring system foroutcome prediction. J Am Coll Cardiol 2013;61:2253–61.

2. Valentová M, von Haehling S, Doehner W, Murín J, Anker SD,Sandek A. Liver dysfunction and its nutritional implications in heartfailure. Nutrition 2013;29:370–8.

3. Sandek A, Bjarnason I, Volk HD, et al. Studies on bacterial endotoxinand intestinal absorption function in patients with chronic heart failure.Int J Cardiol 2010;57:80–5.

4. Sandek A, Bauditz J, Swidsinski A, et al. Altered intestinal functionin patients with chronic heart failure. J Am Coll Cardiol 2007;50:1561–9.

5. Allen LA, Felker GM, Pocock S, et al., for the CHARM Investigators.Liver function abnormalities and outcome in patients with chronic heartfailure: data from the Candesartan in Heart Failure: Assessment ofReduction in Mortality and Morbidity (CHARM) program. Eur J HeartFail 2009;11:170–7.

6. McAlister FA, Ezekowitz J, Tonelli M, Armstrong PW. Renal insuf-ficiency and heart failure prognostic and therapeutic implications from aprospective cohort study. Circulation 2004;109:1004–9.

7. Wasywich CA, Pope AJ, Somaratne J, Poppe KK, Whalley GA,Doughty RN. Atrial fibrillation and the risk of death in patients withheart failure: a literature-based meta-analysis. Intern Med J 2010;40:347–56.

8. Myers RP, Cerini R, Sayegh R, et al. Cardiac hepatopathy: clinical,hemodynamic, and histologic characteristics and correlations. Hepatol-ogy 2003;37:393–400.

9. Eisen HJ. The MELD scoring system and the prediction of outcomes inheart failure patients: what we have learned from the hepatologists. J AmColl Cardiol 2013;61:2262–3.

Reply

Cardiac Hepatopathy VersusEnd-Stage Liver Disease

Two Different Entities

We thank Dr. Valentova and colleagues for their insightfulcomments on our recent study (1) of liver function abnormalitiescharacterized by various MELD (Model of End Stage LiverDisease) scoring systems in patients evaluated for heart trans-plantation (HTx). On the basis of our analysis, utilization ofMELD scoring systems may improve our risk prediction on thebasis of a composite evaluation of end-organ dysfunction includingliver and renal functional abnormalities. Of note, the use of thestandard MELD score is limited to patients not on oral anti-coagulation due to the impact of vitamin K antagonists on theinternational normalized ratio (INR), which is an essentialcomponent of the standard MELD score. For that reason, weanalyzed both patients on and off oral anticoagulants separatelyand included the modified MELD-XI score, which allows char-acterization of patients on vitamin K antagonists.Dr. Valentova and colleagues also state that in patients with

liver cirrhosis, the use of total bilirubin is most appropriate andthat liver cirrhosis is a rare finding in patients with HF. This doesnot reflect our experience, as the degree of liver cirrhosis is afrequent concern in our population of patients evaluated for HTx;in particular, in patients with longstanding right and biventricularheart failure, patients on left ventricular assist device support andpatients with congenital heart disease after, for example, Fontanrepair. At our center, we performed liver biopsies in more than 5%of all transplant candidates evaluated between January 1, 2000and April 1, 2013 due to the suspicion of structural liver damagesuch as liver fibrosis secondary to congestive hepatopathy, previoushepatitis or irreversible liver function abnormalities, and concom-itant imaging abnormalities. Unfortunately, our current analysisfocused on the utilization of MELD scores that included only totalbilirubin, and we did not include direct bilirubin into our analysis(1). It would be interesting, however, to study this parameter in asubsequent analysis. Nevertheless, the impact of bilirubin and itsspecific statistical weight within the composite MELD scores hasonly been established for total bilirubin, and it will require ad-justments if direct bilirubin will be used instead of total bilirubin.We agree with Dr. Valentova and colleagues that elevated

MELD scores might reflect more severe comorbidities andmore advanced HF. However, in our cohort, we did not detect