Impact of Prosthesis Size and Prosthesis-Patient Mismatch on Outcomes in Younger Female Patients Undergoing Aortic Valve Replacement

Accepted Manuscript

Impact of prosthesis size and prosthesis-patient mismatch on outcomes in Younger Female Patients Undergoing Aortic Valve Replacement Ahmed A. Kolkailah MD , Sameer A. Hirji MD , Julius I. Ejiofor MD, MPH , Fernando Ramirez Del Val MD, MPH , Ritam Chowdhury MBBS, MPH, PhD, SM , Siobhan McGurk BS , Jiyae Lee BS , Tsuyoshi Kaneko MD PII: DOI: Reference:

S1043-0679(18)30408-8 https://doi.org/10.1053/j.semtcvs.2019.01.005 YSTCS 1223

To appear in:

Seminars in Thoracic and Cardiovascular Surgery

Please cite this article as: Ahmed A. Kolkailah MD , Sameer A. Hirji MD , Julius I. Ejiofor MD, MPH , Fernando Ramirez Del Val MD, MPH , Ritam Chowdhury MBBS, MPH, PhD, SM , Siobhan McGurk BS , Jiyae Lee BS , Tsuyoshi Kaneko MD , Impact of prosthesis size and prosthesis-patient mismatch on outcomes in Younger Female Patients Undergoing Aortic Valve Replacement, Seminars in Thoracic and Cardiovascular Surgery (2019), doi: https://doi.org/10.1053/j.semtcvs.2019.01.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Impact of prosthesis size and prosthesis-patient mismatch on outcomes in Younger Female Patients Undergoing Aortic Valve Replacement

CR IP T

Ahmed A. Kolkailah MD1*, Sameer A. Hirji MD1*, Julius I. Ejiofor MD MPH1, Fernando Ramirez Del Val MD MPH1, Ritam Chowdhury MBBS MPH PhD SM2, Siobhan McGurk BS1, Jiyae Lee BS1, Tsuyoshi Kaneko MD1.

AN US

*Co-First Authors

1

Division of Cardiac Surgery, Brigham and Women’s Hospital, Harvard Medical School;

2

Harvard School of Public Health; Boston, MA.

M

Presented at the Western Thoracic Surgical Association 43rd Annual Meeting in Colorado Springs, Colorado – June 24, 2017

ED

Conflict of Interest: There are no potential conflicts that exist

PT

Funding: No funding was provided

CE

Corresponding author: Tsuyoshi Kaneko, MD

AC

Division of Cardiac Surgery Brigham and Women’s Hospital 75 Francis Street, Boston, MA Tel: 617-732-7678 Fax 617-732-6559 E-mail: [email protected]

ACCEPTED MANUSCRIPT ABSTRACT: Objectives: The effects of patient-prosthesis mismatch (PPM) after surgical aortic valve replacement (SAVR) suggest worse outcomes with smaller valves. We assessed clinical outcomes of younger females undergoing SAVR, using small and large prostheses, and the incremental risk of PPM.

CR IP T

Methods: Between January 2002 and June 2015, 451 younger (age ≤65 years) female patients underwent SAVR. Patients were stratified into small prostheses (SP)≤21-mm (n=256) and large prostheses (LP) ≥23-mm (n=195) groups. PPM was classified as moderate if iEOA 0.65 – 0.85 cm2/m2, or severe if iEOA<0.65 cm2/m2.

AN US

Results: Operative mortality was not statistically different between SP and LP groups (2.4% vs 0.5%; p=0.146). Unadjusted 10-year survival was 82% (95%CI:77%–87%), and was similar in both groups (p=0.210). When grouped by standard PPM thresholds, only severe

M

PPM was associated with significantly decreased survival (p=0.007). A significant survival decrease was detected in LP group with iEOA ≤0.75 cm2/m2(p<0.001). Among SP patients,

ED

iEOA ≤0.65 cm2/m2 was associated with increased mortality (p=0.075). After adjusting for potential confounders, Cox proportional hazard model identified iEOAs of ≤0.65 cm2/m2 (HR

PT

1.85;p=0.066) and ≤0.75 cm2/m2 (HR 2.3;p≤0.003) as predictors of decreased long-term

CE

survival, in SP and LP groups, respectively. Conclusions: Amongst younger females who underwent SAVR, postoperative complications

AC

and in-hospital outcomes were substantially similar between the SP and LP groups. However, patients who received LP were adversely affected at lesser degrees of PPM than those who received SP. While SP patients may tolerate until iEOA≤0.65 cm2/m2, our results suggest that moderate PPM of iEOA≤0.75 for LP patients should be avoided.

Word Count: 250/250

ACCEPTED MANUSCRIPT

GLOSSARY: BSA: Body surface area CABG: Coronary artery bypass graft CHF: Congestive heart failure

CR IP T

CVA: Cerebrovascular accident CVD: Cerebrovascular disease CKD: Chronic kidney disease DM: Diabetes mellitus

AN US

EOA: Effective orifice area iEOA: Indexed effective orifice area LOS: Length of stay

M

LP: Large prostheses LV: Left ventricular

ED

LVM: Left ventricular mass

PT

NYHA: New York Heart Association PH: Proportional hazards

CE

PVD: Peripheral vascular disease PPM: Patient-prosthesis mismatch

AC

SAVR: Surgical aortic valve replacement SP: Small prostheses TAVR: Transcatheter aortic valve replacement VIV: Valve-in-valve ARE: Aortic root enlargement

ACCEPTED MANUSCRIPT

ED

M

AN US

CR IP T

CENTRAL PICTURE:

AC

CE

PT

Adjusted Cox PH model for cumulative survival, stratified by prosthesis size and iEOA.

ACCEPTED MANUSCRIPT CENTRAL MESSAGE: While younger females (≤65 years) undergoing SAVR with SP (≤21-mm) may tolerate until iEOA of ≤0.65 cm2/m2, moderate PPM of iEOA≤0.75 for LP patients should be avoided. Character count: 167/200 (with spaces)

CR IP T

PERSPECTIVE STATEMENT:

Younger female patients receiving LP were adversely affected at lesser degrees of PPM than those receiving SP. While SP patients may tolerate until iEOA≤0.65 cm2/m2, our results suggest that moderate PPM of iEOA≤0.75 for LP patients should be avoided. The optimal

consideration for future VIV-TAVR.

AC

CE

PT

ED

M

Character count: 391/405 (with spaces)

AN US

operative strategy and valve-size selection should be tailored on a case-by-case basis, with

ACCEPTED MANUSCRIPT INTRODUCTION: Patient-prosthesis mismatch (PPM) has long been a subject of controversy, ever since the concept was introduced by Rahimtoola in 1978.1 In his description, Rahimtoola defines PPM as an in-vivo prosthetic valve effective orifice area (EOA) that is less than that of a normal native valve. By this definition, most, if not all, prosthetic valves will have an EOA that is

CR IP T

less than the native valve, such that any patient undergoing surgical aortic valve replacement (SAVR) will develop PPM to some degree.1,2 This, in turn, leads to suboptimal decrease in trans-valvular gradients, inadequate left ventricular (LV) reverse remodeling, and lack of

AN US

significant improvement in overall functional status.3,4

Several groups have investigated the effects of PPM on short- and long-term outcomes, but with conflicting results, and have used various definitions for PPM severity

M

classification.2,4-6 Although some studies report that there is no significant influence of PPM,7-10 others suggest that PPM is detrimental to patients’ outcomes.11-14 Consequently,

PT

ED

cardiac surgeons strive to predict and avoid PPM, whenever possible.3,6,15,16

Various groups, including ours, have achieved favorable outcomes using small valves in

CE

the elderly. This may be explained by their smaller body surface area and relatively lower levels of physical activity compared to younger patients.17,18 However, the relatively larger

AC

body surface area and higher levels of physical activity, may be reasons to consider placing at least a 23-mm valve in younger, active patients, in whom PPM may impose a greater risk.19,20 Additionally, despite its increasing popularity, transcatheter aortic valve replacement (TAVR) in valve-in-valve (VIV) fashion has been associated with worse outcomes in valves 21-mm or less, not to mention its lack of feasibility with surgically implanted mechanical prostheses.21

ACCEPTED MANUSCRIPT

PPM is reported to be more common in females, with female gender being an independent risk factor for severe PPM.11 For this reason, we sought to assess clinical outcomes of younger female patients undergoing SAVR. We hypothesized that there would

validity of the current, standard PPM thresholds.

METHODS: Study Population

CR IP T

be no survival difference depending on implanted valve size. We further tested the prognostic

AN US

We retrospectively reviewed all electronic medical records (EMR) of female patients age ≤65 years at the time of surgery who underwent SAVR between January 2002 and June 2015. Patients undergoing concomitant CABG or aortic procedures were also included. Patients undergoing homografts, concomitant mitral or tricuspid, aortic dissections and ruptures, and

M

ventricular assist device procedures were excluded. 451 patients met study criteria. Patients

ED

were dichotomized into those with small prostheses (SP) ≤21-mm (n=256) or large prostheses (LP) ≥23-mm (n=195). Partners Healthcare Institutional Review Board approved

PT

this study with waived consent (protocol 2010p000292).

CE

Data Collection and Outcomes of Interest

AC

Patients’ characteristics, peri-operative data, laboratory test results, and in-hospital outcomes were recorded at the time of presentation and extracted from EMR. Variables were defined and coded according to the Society of Thoracic Surgeons’ (STS) Adult Cardiac Surgery database version 2.52 specifications, unless otherwise noted. Our primary endpoints were operative mortality and long-term survival. Secondary endpoints included postoperative complications, hospital length of stay (LOS) and echocardiographic findings at follow-up. Mortality data was obtained from our internal research data repository, routine patient follow-

ACCEPTED MANUSCRIPT up, and our state Department of Public Health. Follow-up time was calculated in months from the date of surgery to the date of death or June 30, 2016, if alive.

Echocardiographic data The EOAs were obtained from each device’s manufacturer’s information or data

CR IP T

available in the literature. Most EOA’s except for one were based on in-vivo measurements (Supplement Table e1). The EOA was indexed to body surface area (BSA), yielding a projected, indexed EOA (iEOA). PPM was classified according to standard criteria into three categories; absent (iEOA >0.85 cm2/m2), moderate (iEOA 0.65 – 0.85 cm2/m2), or severe

AN US

(iEOA <0.65 cm2/m2).6 Left ventricular mass (LVM) was calculated using Devereux and associates’ formula,22 and categories were assigned according to American Society of

M

Echocardiography definitions.23

All echocardiographic data were obtained by chart review or extracted from EMR.

ED

Preoperative echocardiographic data was recorded for 443 (98%) of patients. We obtained

PT

first follow-up echocardiographic reports that were done within one year of discharge, but beyond at least 3 months after surgery. Of the 438 patients who were alive at 90 days post-

CE

surgery, follow-up reports were available for 229 (52%) patients.

AC

Statistical Analysis

Normally distributed continuous variables are expressed as mean  standard deviation

(SD) and were compared using Student’s t-test with Levene’s test for homogeneity of variance. Paired t-test were used to compare pre- and post-operative mean and peak AV gradients. Non-normally distributed variables are expressed as median and interquartile range (IQR), and were compared using Mann-Whitney U tests. Categorical variables are presented

ACCEPTED MANUSCRIPT as number and percentage, and were compared using 2 or Fisher’s Exact test. Unadjusted survival was evaluated by Kaplan-Meier analyses, and groups were compared using log-rank tests.

We first sought to investigate whether iEOA was associated with survival, and if this

CR IP T

association differed by implant size. To do so, we first tested unadjusted survival at different iEOA thresholds for the whole group, and stratified by valve size. At this point, we realized the PPM “penalty effect” was not the same in the two groups. Our previous work in the elderly patients has also indicated that PPM might not be a factor in survival for small valves

AN US

until severe (<=.65), so we hypothesized that there could be different thresholds for the patients receiving larger valves. In addition to using standard cutoffs of PPM, we also performed ROC analyses to identify the best cut-off point of iEOA predicting mortality in

M

each strata (Supplement Table e2).

ED

With only 61 events, initial univariate analyses were also conducted to determine factors

PT

most likely to be associated with all-cause mortality. These included age (years), ejection fraction (%), and preoperative creatinine (mg/dL) as continuous variables, and type of

CE

implanted valve (mechanical vs. bioprosthetic valves), BSA quartiles, congestive heart failure (CHF) within the last 2 weeks, New York Heart Association (NYHA) class III/IV, chronic

AC

kidney disease (CKD), concomitant CABG, concomitant aortic procedure, reoperative surgery, diabetes mellitus (DM), cerebrovascular accident (CVA), cerebrovascular disease (CVD), and peripheral vascular disease (PVD) as categorical variables. Variables found to be contributory using a retention threshold of p≤0.1, and those with known contribution to survival or deemed clinically important based on our experience, were then entered into a forward-stepwise Cox proportional hazards (PH) model to identify the strongest predictors of

ACCEPTED MANUSCRIPT all-cause mortality. A final enter-method Cox model was run with the resultant variables (for adjustment), in which we also included categorical variables for year of surgery and iEOA. All analyses were conducted using IBM SPSS Statistics version 23.0 (IBM Corporation, Armonk, NY) with p0.05 as the criterion for significance.

CR IP T

RESULTS:

Preoperative Characteristics and Echocardiographic Findings

Baseline patient characteristics and preoperative echocardiographic findings for both groups are summarized in Table 1. SP patients were older (56.7 ± 7.8 years vs 53.6 ± 10

AN US

years, p≤0.001) but LP patients had larger BSA (1.80 ± 0.22 m2 vs 1.87 ± 0.25 m2; p≤0.002). Although most patients presented with normal LVM, proportionally more LP patients had moderate or severe LVM enlargement at baseline (4.8% vs 11.9%; p=0.001). The groups

M

were otherwise similar in terms of comorbidities and cardiac function at baseline.

ED

Operative Characteristics

PT

Operative variables, types of implanted valves, and degree of PPM are presented in Table 2. SP patients were more likely to have had concomitant CABG (20.9% vs 12.4%,

CE

p=0.023), but fewer concomitant aortic procedures (6.3% vs 19.1%; p≤0.001). Although minimally utilized, aortic root enlargements (ARE) were performed with a similar frequency

AC

(5.5% vs 4.6%; p=0.830). Mean projected iEOA was 0.89 ± 0.22 in SP group and 1.04 ± 0.29 in LP group (p≤0.001). In SP group, 39.8% had moderate PPM and 10.6% had severe PPM. In LP group, the incidence of PPM was less i.e. 16% for moderate and 2.6% for severe PPM.

Postoperative Complications and In-Hospital Outcomes

ACCEPTED MANUSCRIPT Complications and in-hospital outcomes are highlighted in Table 3. Importantly, operative mortality was not statistically different between SP and LP patients (2.4% vs 0.5%; p=0.146). Postoperative stroke occurred in 3.9% of SP patients vs 1.5% of LP patients (p=0.163). Although ICU stays were substantially similar, SP patients had a longer LOS (6

CR IP T

[IQR 5, 8] days vs 6 [IQR 5, 7] days; p=0.022).

Echocardiographic Data at Follow-Up

Follow-up echocardiographic data are also presented in Table 3. A total of 229 patients had available qualifying studies. We conducted responder-bias analyses, and found that there

AN US

were no significant differences in age, comorbidities, preoperative cardiac status and echocardiographic findings, type of surgery or concomitant surgeries, or distributions of type and size of implant, between patients with and without follow-up (all p>0.29) (Supplement

M

Table e3) Median follow-up times were similar in SP and LP groups (23 [IQR 11, 44] months vs 20 [IQR 10, 48] months; p=0.486). Follow-up AV mean and peak gradients,

ED

respectively, were also similar in both groups (15.8 ± 9.2 mmHg vs 15.3 ± 12.8 mmHg;

CE

parameters.

PT

p=0.901, and 29.3 ± 15.6 mmHg vs 27.0 ± 21.2 mmHg; p=0.305) and within normal

PPM and Unadjusted Survival in SP and LP groups

AC

The median observation time was 6.1 (range 0–14.4) years, for a 2898 total patient-

years. There were 61 deaths during the observation period. We first tested unadjusted survival at different iEOA thresholds for the whole group, and stratified by valve size. Overall unadjusted survival at 1, 5, and 10 years were 95.8% (95% CI: 93.9–97.6), 89.5% (95% CI: 86.5–92.6) and 81.9% (95% CI: 77.2–86.6), and was similar between SP and LP groups (p=0.21) (Figure 1A). At this point, we realized the PPM “penalty effect” was not the

ACCEPTED MANUSCRIPT same in the two groups. When grouped by standard PPM thresholds (Table 2), only severe PPM was associated with significantly decreased survival (P=0.007). 5-year survival in this group was 78.5% (95% CI: 63.1–94.1) (Figure 1B, Supplementary table e4). We therefore examined iEOA with binary thresholds, also utilizing the results of our ROC analyses (Supplement table e1).

CR IP T

Since there was a survival difference between moderate and severe PPM groups, our results suggested that there was an inflection point for decreased survival between iEOAs of 0.65 cm2/m2 and 0.85 cm2/m2. To explore this possibility, we created categorical variables for iEOAs in increments of 0.05 cm2/m2, testing them on the overall cohort and stratified by

AN US

prosthesis size. We observed a significant association with decreased survival at iEOA ≤0.75 cm2/m2 for LP patients (p<0.001; Figure 2A). For SP patients, a trend towards decreased survival did not appear until a threshold of < 65 (p=0.065). (Figure 2B, Supplementary

M

table e3).

ED

PPM and Adjusted Survival in SP and LP groups

PT

To examine whether the effect demonstrated above persisted in our adjusted analyses, we created a parsimonious Cox proportional hazard models using a categorical variable

CE

composed of 4 groups: SP, iEOA ≤0.65 cm2/m2; SP, iEOA >0.65 cm2/m2; LP, iEOA ≤0.75 cm2/m2; and LP, iEOA >0.75 cm2/m2 (the reference group). Variables in the model that were

AC

found to be contributory on exploratory analyses included age, CHF, NYHA class III/IV and CKD. Non-contributory variables which included concomitant procedures, reoperative status, DM, CVA, BSA, valve type, PVD and CVD (all p>0.18), were excluded. The results of the Cox model are presented in Figure 3. The HR associated with LP and iEOA ≤0.75 cm2/m2 was 2.62 (95% CI: 1.39–4.93; p=0.003) and for SP with iEOA ≤0.65 was 2.18 (95% CI: 0.95–5.00; p=0.066) (Table 4). Survival for SP and LP patients above the respective iEOA

ACCEPTED MANUSCRIPT thresholds was substantially similar. For comparative purposes, we compared the performance of this model (-2log likelihood [-2LL] =613.4, 2 = 82.3, df=20; p≤0.001) to the same model run using standard PPM thresholds, and using the same iEOA cutoff for each valve group. Our model out-performed both; PPM groups were not significantly predictive overall (p=0.07, model -2LL=790.1, 2=69.3, df=19) and an iEOA threshold of 0.75 cm2/m2

1.17–3.83, model -2LL=647.9, 2=80.5, df=20).

AN US

DISCUSSION:

CR IP T

was only significantly predictive for LP, in the second model (p=0.01; HR=2.11, 95% CI:

In this series of younger females undergoing SAVR, we observed several interesting and unique findings. First, there were no significant differences in postoperative complications and in-hospital outcomes among patients receiving SP and LP. AV gradients were reduced to

M

satisfactory values and significantly improved from baseline in both groups. Second, survival was adversely affected at differing iEOA thresholds, depending on valve size; the LP group

ED

with iEOA ≤0.75 cm2/m2 showed markedly reduced survival on Cox PH analyses, while

PT

iEOA was not associated with decreased survival in SP patients until the threshold of ≤0.65 cm2/m2, although this association did not meet our significance criterion, likely due to sample

CE

size. Our results suggest that valve size should not be the only clinical consideration, but rather patient-specific, iEOA projection (PPM severity). In light of our findings, LP patients

AC

are generally larger (higher BSA), often with increased cardiac demands, which may warrant upsizing to avoid an iEOA ≤0.75cm2/m2. Likewise, SP patients may tolerate until iEOA of ≤0.65 cm2/m2.

The ongoing debate remains regarding PPM, ever since the concept was introduced by Rahimtoola in 1978.1 PPM is defined as an in-vivo prosthetic valve EOA that is less than that

ACCEPTED MANUSCRIPT of a normal native valve. This is an inherent consequence to the fact that almost all prosthetic valves have an EOA that is less than that of a native valve, mostly from the stent which stabilizes the valve.1,2

Numerous groups have attempted to assess the true long-term effects of PPM. In a study

CR IP T

of 892 patients, Medalion and colleagues showed that valve size and moderate degrees of PPM did not adversely affect long-term survival, but rather patients’ baseline characteristics such as older age, family history of ischemic heart disease, and chronic lung or kidney disease.9 Jamieson et al. also examined the effect of PPM in 3,343 patients undergoing SAVR

AN US

and demonstrated no impact on overall, unadjusted 15-year survival.10 Likewise, Blackstone et al. conducted a large meta-analysis of individual patient data from nine representative sources which included a total of 13,258 SAVRs, and showed similar results; small valve size

M

and PPM were not associated with decreased long-term survival.24 In contrast, more recently, Dayan et al. showed in their largest meta-analysis of 382 studies (40,381 patients) that PPM

ED

increases perioperative and overall mortality proportionally to its severity. The impact of

PT

PPM was less pronounced in patients with larger BMI (>28 kg/m2) compared with those with lower index. Additionally, predictors of PPM included older age, female sex, hypertension,

CE

diabetes, renal failure, larger BSA, larger BMI, and the utilization of a bio-prosthesis. These findings conflict with our results which suggest that moderate to severe PPM for LP patients

AC

appears important for late survival. 16

Conversely, in a PARTNER trial sub-group analysis, severe PPM was associated with

decreased 2-year survival in the SAVR group (HR, 1.78; p=0.041).25 Additionally, a prospective registry of 1,266 patients found PPM to be an incremental predictor of 30-day mortality, particularly in patients with preoperative left ventricular dysfunction.12 Two of the

ACCEPTED MANUSCRIPT largest meta-analyses also revealed that PPM was strongly associated with decreased longterm survival, recommending the use of different strategies to avoid PPM.13,26 Several groups have also reported on the different effects of PPM on LVM regression; some report more LVM regression in patients with no PPM,4,27,28 while others report no significant influence of PPM on LV remodeling.29-31 The limited follow-up in our series, and relatively normal LVM

CR IP T

at baseline, precluded us from drawing conclusions in this regard.

Ideally, surgeons should avoid PPM of any degree. However, this is not always possible, and some advocate just avoiding severe PPM, based on existing literature (above) that

AN US

demonstrate no association between moderate PPM and patients’ outcomes.9,10 It is possible that there may be confounders which could justify the conclusions from these studies. Nonetheless, in our series, this held true for patients receiving SP, where an association with

M

decreased survival only appeared at an iEOA threshold of 0.65 cm2/m2. However, patients who received LP were significantly affected at a higher threshold of 0.75 cm2/m2. This sheds

ED

some doubts on the traditional practice of surgeons’ satisfaction if they can implant a 23-mm

PT

or larger valve. For example, if you are in the operating room and has a young female patient that fits a 23mm valve, one would probably will feel comfortable rather than placing a 21mm.

CE

However, if the iEOA is 0.73 with 23mm valve, one should perform a root enlargement and place a 25mm to avoid iEOA of 0.75. When 23mm or larger valve is placed, you have to pay

AC

more attention to the iEOA. Additionally, patients receiving LP are generally larger and usually have increased cardiac demands due to their BSA, which warrant upsizing to an even larger prosthesis to avoid an iEOA of 0.75 or less, based on preoperative projections. Furthermore, it was interesting that 90% of the patients with a small prosthesis have iEOA <65, and about 90% of the patients with a large prosthesis have an iEOA >75. That also suggests that our chosen valve size cutoffs did matter.

ACCEPTED MANUSCRIPT

Although VIV-TAVR is increasingly popular and a viable alternative to reoperative SAVR, Dvir et al. however, described worse outcomes associated with VIV-TAVR in smaller, surgically implanted valves (21-mm or less), in terms of higher postoperative gradients and lower 1-year survival, as opposed to larger valves.21 Hence, there seems to be a

CR IP T

momentum towards implanting a prosthesis large enough to avoid PPM and which allows for future VIV-TAVR.

Several options could be utilized to overcome the PPM problem. Several groups have

AN US

demonstrated the safety and feasibility of ARE in situations when surgeons believe the patient requires a LP. 32-34 However, many physicians are reluctant to perform this procedure, as it is associated with longer bypass and cross clamp times, thereby increasing the operative

M

complexity.32-34 Ogawa and colleagues further suggest that its proper utilization could be agedependent, where younger patients may receive a greater benefit from the procedure, but

ED

older, less active patients may not.18 In their series, they achieved satisfactory results with 17-

PT

mm valves, without the need for ARE.18 Our institution has previously reported favorable outcomes using 19-mm valves in older patients, which was also an impetus for us to look at

CE

a younger population in our center.17 We consider AREs on a case-by-case basis to avoid severe PPM but often the decision to perform ARE is surgeon dependent. However, given

AC

our observed findings, as well as poor outcomes experienced with VIV-TAVR in smaller valves, we are starting to perform more ARE procedures although we acknowledge that further robust survival data is needed to assess the utility of ARE in younger females. Other studies suggest using stentless bioprostheses, or homografts, as they have been linked to a reduced incidence and/or severity of PPM.2,3,25,35

ACCEPTED MANUSCRIPT Our results are subject to all limitations inherent to a retrospective, observational study. Also, as a single center study, our findings may not be readily generalizable to other populations. Precisely, surgical selection and referral, clinical preferences and practices, in addition to the small number of events in our series, are potential contributors to our observations. Specifically, the number of patients in the LP 0.75 group was the smallest of

CR IP T

the cohorts, and numbers at risk beyond 5-years was very small. Likewise, the number of patients in specific age groups were small. Being an academic, tertiary center, patient followup commonly takes places outside of our institution. Moreover, echocardiographic findings were presumably mostly from surviving patients, and unavailable for most patients who

AN US

ultimately died or followed at outside institution. The EOAs were based on data provided from manufacturers or manufacturer-sponsored studies, so it is possible that our iEOAs are biased. Likewise, the n was small beyond 5 years for SP patients with iEOA > 65, and for LP

M

patients with iEOA < 75. Ideally, the threshold should be check in a validation cohort but given the sample size, it was difficult to split the cohort into the separate derivation and

ED

validation cohort. There could have been other confounders other than valve size, such as bad

PT

endocarditis, very poor LV function, or perhaps mechanical valve thrombosis, which could have explained the observed high mortality in the LP < 0.75. We were also unable to evaluate

CE

other clinical data of interest including cardiac cause of death or impact of PPM on LV

AC

remodeling. Our findings should be interpreted with these cautions.

CONCLUSION In this series, postoperative complications and in-hospital outcomes were substantially similar between the SP and LP groups. However, patients who received LP were adversely affected at lesser degrees of PPM than those who received SP. Thus, we believe that valve size should not be the only clinical consideration, but rather patient-specific, iEOA projection

ACCEPTED MANUSCRIPT (PPM severity). While SP patients may tolerate until iEOA of ≤0.65 cm2/m2, our results suggest that moderate PPM of iEOA≤0.75 for LP patients should be avoided. The optimal operative strategy and valve-size selection should be tailored on a case-by-case basis, with

AC

CE

PT

ED

M

AN US

CR IP T

special consideration for future VIV-TAVR.

ACCEPTED MANUSCRIPT References

7.

8.

9.

10.

11.

12.

AC

13.

CR IP T

6.

AN US

5.

M

4.

ED

3.

PT

2.

Rahimtoola SH. The problem of valve prosthesis-patient mismatch. Circulation. 1978;58(1):20-24. Daneshvar SA, Rahimtoola SH. Valve prosthesis-patient mismatch (VP-PM): a longterm perspective. Journal of the American College of Cardiology. 2012;60(13):11231135. Pibarot P, Dumesnil JG. Hemodynamic and clinical impact of prosthesis-patient mismatch in the aortic valve position and its prevention. Journal of the American College of Cardiology. 2000;36(4):1131-1141. Kandler K, Moller CH, Hassager C, Olsen PS, Lilleor N, Steinbruchel DA. Patientprosthesis mismatch and reduction in left ventricular mass after aortic valve replacement. The Annals of thoracic surgery. 2013;96(1):66-71. Pibarot P, Dumesnil JG. Prosthetic heart valves: selection of the optimal prosthesis and long-term management. Circulation. 2009;119(7):1034-1048. Pibarot P, Dumesnil JG. Prosthesis-patient mismatch: definition, clinical impact, and prevention. Heart (British Cardiac Society). 2006;92(8):1022-1029. Rubio Alvarez J, Sierra Quiroga J, Vega Fernandez M, et al. Up to twenty-five-year survival after aortic valve replacement with size 19 mm valves. Interactive cardiovascular and thoracic surgery. 2010;10(1):32-35. Hanayama N, Christakis GT, Mallidi HR, et al. Patient prosthesis mismatch is rare after aortic valve replacement: valve size may be irrelevant. The Annals of thoracic surgery. 2002;73(6):1822-1829; discussion 1829. Medalion B, Blackstone EH, Lytle BW, White J, Arnold JH, Cosgrove DM. Aortic valve replacement: is valve size important? The Journal of thoracic and cardiovascular surgery. 2000;119(5):963-974. Jamieson WR, Ye J, Higgins J, et al. Effect of prosthesis-patient mismatch on longterm survival with aortic valve replacement: assessment to 15 years. The Annals of thoracic surgery. 2010;89(1):51-58; discussion 59. Astudillo LM, Santana O, Urbandt PA, et al. Clinical predictors of prosthesis-patient mismatch after aortic valve replacement for aortic stenosis. Clinics. 2012;67(1):5560. Blais C, Dumesnil JG, Baillot R, Simard S, Doyle D, Pibarot P. Impact of valve prosthesis-patient mismatch on short-term mortality after aortic valve replacement. Circulation. 2003;108(8):983-988. Head SJ, Mokhles MM, Osnabrugge RL, et al. The impact of prosthesis-patient mismatch on long-term survival after aortic valve replacement: a systematic review and meta-analysis of 34 observational studies comprising 27 186 patients with 133 141 patient-years. European heart journal. 2012;33(12):1518-1529. Hong S, Yi G, Youn YN, Lee S, Yoo KJ, Chang BC. Effect of the prosthesis-patient mismatch on long-term clinical outcomes after isolated aortic valve replacement for aortic stenosis: a prospective observational study. The Journal of thoracic and cardiovascular surgery. 2013;146(5):1098-1104. Pibarot P, Dumesnil JG, Cartier PC, Metras J, Lemieux MD. Patient-prosthesis mismatch can be predicted at the time of operation. The Annals of thoracic surgery. 2001;71(5 Suppl):S265-268.

CE

1.

14.

15.

ACCEPTED MANUSCRIPT

21. 22.

23.

24.

25.

AC

CE

26.

CR IP T

20.

AN US

19.

M

18.

ED

17.

Dayan V, Vignolo G, Soca G, Paganini JJ, Brusich D, Pibarot P. Predictors and Outcomes of Prosthesis-Patient Mismatch After Aortic Valve Replacement. JACC Cardiovasc Imaging. 2016;9(8):924-933. Khalpey Z, Myers PO, McGurk S, et al. Nineteen-Millimeter Bioprosthetic Aortic Valves Are Safe and Effective for Elderly Patients With Aortic Stenosis. The Annals of thoracic surgery. 2016;101(2):650-657. Ogawa T, Onoe M, Moriwaki S, et al. Aortic valve replacement with a 17-mm mechanical prosthesis in octogenarian or older patients. The Journal of thoracic and cardiovascular surgery. 2016;152(1):112-117. Noorani A, Radia R, Bapat V. Challenges in valve-in-valve therapy. Journal of Thoracic Disease. 2015;7(9):1501-1508. Forcillo J, Tsai LL, Thourani VH. Will transcatheter valve-in-valve become the standard for replacing the degenerative aortic valve bioprosthesis? Hold your horses, the jury is still out. The Journal of thoracic and cardiovascular surgery. 2015;150(6):1568-1569. Dvir D, Webb JG, Bleiziffer S, et al. Transcatheter aortic valve implantation in failed bioprosthetic surgical valves. Jama. 2014;312(2):162-170. Ganau A, Devereux RB, Roman MJ, et al. Patterns of left ventricular hypertrophy and geometric remodeling in essential hypertension. Journal of the American College of Cardiology. 1992;19(7):1550-1558. Marwick TH, Gillebert TC, Aurigemma G, et al. Recommendations on the Use of Echocardiography in Adult Hypertension: A Report from the European Association of Cardiovascular Imaging (EACVI) and the American Society of Echocardiography (ASE). Journal of the American Society of Echocardiography : official publication of the American Society of Echocardiography. 2015;28(7):727-754. Blackstone EH, Cosgrove DM, Jamieson WR, et al. Prosthesis size and long-term survival after aortic valve replacement. The Journal of thoracic and cardiovascular surgery. 2003;126(3):783-796. Pibarot P, Weissman NJ, Stewart WJ, et al. Incidence and sequelae of prosthesispatient mismatch in transcatheter versus surgical valve replacement in high-risk patients with severe aortic stenosis: a PARTNER trial cohort--a analysis. Journal of the American College of Cardiology. 2014;64(13):1323-1334. Chen J, Lin Y, Kang B, Wang Z. Indexed effective orifice area is a significant predictor of higher mid- and long-term mortality rates following aortic valve replacement in patients with prosthesis-patient mismatch. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2014;45(2):234-240. Tasca G, Brunelli F, Cirillo M, et al. Impact of valve prosthesis-patient mismatch on left ventricular mass regression following aortic valve replacement. The Annals of thoracic surgery. 2005;79(2):505-510. Mannacio VA, Mannacio L, Antignano A, et al. Impact of different values of prosthesis-patient mismatch on outcome in male patients with aortic valve replacement. Journal of cardiovascular medicine (Hagerstown, Md). 2017;18(5):366373. Roscitano A, Benedetto U, Sciangula A, et al. Indexed effective orifice area after mechanical aortic valve replacement does not affect left ventricular mass regression

PT

16.

27.

28.

29.

ACCEPTED MANUSCRIPT

32.

33.

34.

AC

CE

PT

ED

M

35.

CR IP T

31.

AN US

30.

in elderly. European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2006;29(2):139-143. Tao K, Sakata R, Iguro Y, et al. Impact of valve prosthesis-patient mismatch on intermediate-term outcome and regression of left ventricular mass following aortic valve replacement with mechanical prosthesis. Journal of cardiac surgery. 2007;22(6):486-492. Hanayama N, Christakis GT, Mallidi HR, et al. Determinants of incomplete left ventricular mass regression following aortic valve replacement for aortic stenosis. Journal of cardiac surgery. 2005;20(4):307-313. Coutinho GF, Correia PM, Pauperio G, de Oliveira F, Antunes MJ. Aortic root enlargement does not increase the surgical risk and short-term patient outcome? European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2011;40(2):441-447. Dhareshwar J, Sundt TM, 3rd, Dearani JA, Schaff HV, Cook DJ, Orszulak TA. Aortic root enlargement: what are the operative risks? The Journal of thoracic and cardiovascular surgery. 2007;134(4):916-924. Castro LJ, Arcidi JM, Jr., Fisher AL, Gaudiani VA. Routine enlargement of the small aortic root: a preventive strategy to minimize mismatch. The Annals of thoracic surgery. 2002;74(1):31-36; discussion 36. Rodes-Cabau J, Pibarot P, Suri RM, et al. Impact of aortic annulus size on valve hemodynamics and clinical outcomes after transcatheter and surgical aortic valve replacement: insights from the PARTNER Trial. Circulation Cardiovascular interventions. 2014;7(5):701-711.

ACCEPTED MANUSCRIPT TABLES Table 1: Baseline Characteristics

ED

PT

CE

AC

± 7.8 ± 7.6 ± 0.22 (18.5) (2.8) ± 0.7 (55.9) (54.3) (6.7) (7.1) (3.1) (2.4) (34.6) (33.9) (3.5) (0.4) (10.6)

53.6 30.7 1.87 37 8 0.9 93 88 22 13 9 6 61 67 3 0 12

AN US

56.7 29.7 1.80 47 7 0.9 142 138 17 18 8 6 88 86 9 1 27

M

Age (years) Body mass index (kg/m2) Body surface area (m2) Diabetes Chronic kidney disease Preoperative creatinine Hypercholesterolemia Hypertension Peripheral vascular disease Cerebrovascular disease Previous stroke History of atrial fibrillation Congestive heart failure NYHA class III/IV Active endocarditis Preoperative intra-aortic balloon pump Previous cardiac surgery AV disease Stenosis Insufficiency Mixed STS-PROM (%) AV mean gradient (mmHg) AV peak gradient (mmHg) Ejection fraction (%) ASE LVM (grams) Degree of LVM enlargement Unknown None/mild Moderate Severe

LP (N=194)

157 31 67 2.33 44 73.1 62 102.9 4 238 5 7

(61.8) (12.1) (26.2) ± 2.36 ± 17 ± 27.2 (60, 65) ± 35.9 (1.4) (93.7) (2) (2.8)

± 10 ± 8.7 ± 0.25 (19.1) (4.1) ± 0.6 (47.9) (45.4) (11.3) (6.7) (4.6) (3.1) (31.4) (34.4) (1.5) (0) (6.2)

P-value 0.001 0.232 0.002 0.903 0.440 0.739 0.105 0.070 0.092 1.000 0.460 0.770 0.481 1.000 0.246 1.000 0.127 0.189

CR IP T

SP (N=254)

111 39 44 2.40 47.5 77.5 60 111.5 3 168 16 7

(57.4) (20.0) (22.6) ± 5.14 ± 19.5 ± 31.3 (55, 65) ± 30.7

0.857 0.079 0.147 0.007 0.024 0.001

(1.5) (86.6) (8.3) (3.6)

All values are expressed as mean ± SD, n (%), or median (IQR). NYHA: New York Heart Association; AV: Aortic Valve; ASE: American Society of Echocardiography; LVM: Left Ventricular Mass; STSPROM: Society of Thoracic Surgeons’ Predicted Risk of Mortality.

ACCEPTED MANUSCRIPT

Table 2: Operative outcomes

ED

PT

CE

AC

LP (N=194) 3 (1.5)

P-value ---

24 (12.4)

0.023

16 (6.3)

37 (19.1)

0.001

9 (4.6)

0.830

14 (5.5) 59.4 (2.3) (48.4) (0.8) (5.1) (2.8) 40.6 (3.5) (7.1) (11.1) (17.3) (1.6) (0.8) (81, 146) (61, 110) (27.2) (1, 3) ± 0.37 ± 0.22

127 (0) 121 2 3 1 67 9 10 29 10 7 3 105 76 32 2 1.94 1.04

AN US

151 6 123 2 13 7 103 9 18 28 44 4 2 107 81 69 2 1.58 0.89

CR IP T

53 (20.9)

M

Emergent status Concomitant procedures CABG Aortic (excluding dissections or ruptures) Annular enlargement Valves implanted Bioprostheses Biocor Porcine CE Magna Mosaic Sorin Mitroflow St. Jude Trifecta Mechanical Carbomedics Top Hat On-X St. Jude Mechanical St. Jude Regent St. Jude Valve Conduit Intraoperative IABP placement Perfusion time (minutes) Cross-clamp time (minutes) Patients transfused with pRBCs Units/patient EOA Projected iEOA Postoperative PPM None (iEOA >0.85) Moderate (iEOA 0.65 – 0.85) Severe (iEOA <0.65)

SP (N=254) 8 (3.1)

126 (49.6) 101 (39.8) 27 (10.6)

(65.5) 0.0 (62.4) (1) (1.5) (0.5) (34.5) (4.6) (5.2) (14.9) (5.2) (3.6) (1.5) (79, 157) (57, 118) (16.5) (2, 3) ± 0.59 ± 0.29

0.203

0.204

0.656 0.852 0.516 0.009 0.576 0.001 0.001 0.001

158 (81.4) 31 (16) 5 (2.6)

All values are expressed as mean ± SD, n (%), or median (IQR). CABG: Coronary Artery Bypass Grafting; CE: Carpentier-Edwards; pRBCs: Packed Red Blood Cells; EOA: Effective Orifice Area; iEOA: indexed EOA; PPM: Patient-Prosthesis Mismatch; iEOA unit: cm2/m2; IABP: Intra-Aortic Balloon Pump

ACCEPTED MANUSCRIPT

Table 3: Postoperative outcomes and echocardiographic follow-up SP (N=254) 3 10 2 0 56 2

(1.2) (3.9) (0.8) (0) (22) (1, 3)

In-Hospital outcomes Operative mortality Ventilation time (hours) Ventilation >24 hours Total ICU stay (hours) Postoperative LOS (days) 30-day readmission

6 5.6 13 30 6 16

(2.4) (3.5, 9.0) (5.1) (23, 68) (5, 8) (6.3)

AN US M

Echocardiographic follow-up

CE

PT

ED

Time to postoperative echo (months) Preoperative AV mean gradient (mmHg) Postoperative AV mean gradient (mmHg)† Preoperative AV peak gradient (mmHg) Postoperative AV peak gradient (mmHg)†

2

AC

Postoperative LVM (g/m ) Absolute reduction in LVM (g/m2 ) Percent reduction in LVM

SP (n=133) 23 (11, 44) 43.2 ± 16.4 15.8 ± 9.2 72.3 ± 27.2 29.3 ± 15.6

(n=51) 90.9 ± 29.2 -21.5 ± 41.0 13.3 ± 25.6

Pvalue

3 3 3 0 43 2

(1.5) (1.5) (1.5) (0) (22.2) (1, 3)

1.000 0.163 0.656 -1.000 0.527

1 5.3 12 38 6 15

(0.5) (3.6, 9.7) (6.2) (24, 61) (5, 7) (7.7)

0.146 0.650 0.680 0.847 0.022 0.557

CR IP T

Postoperative complications Reoperation required Permanent stroke New onset renal failure Dialysis requirement Patients transfused with pRBCs Units/patient

LP (N=194)

LP (n=96) 20 (10, 48) 46.7 ± 19.2 15.3 ± 12.8 72.8 ± 30.6 27.0 ± 21.2 (n=42) 95.3 ± 30.2 -22.4 ± 32.6 16.6 ± 24.1

0.486 0.140 0.901 0.897 0.305

0.417 0.903 0.518

†p≤0.001 vs. preoperative values All values are expressed as mean ± SD, n (%), or median (IQR). AV: Aortic Valve; Echo: Echocardiography. pRBCs: Packed Red Blood Cells; ICU: Intensive Care Unit; LOS: Length of Stay; LVM: Left Ventricular Mass;

ACCEPTED MANUSCRIPT Table 4: Multivariable analysis using Cox modeling for cumulative survival HR

0.412 0.066 0.003 0.027 0.001 0.002 0.001

1.318 2.179 2.618 1.833 2.919 1.080 3.759

95% CI Lower Upper 0.682 0.949 1.389 1.070 1.696 1.033 1.740

2.545 5.003 4.926 3.141 5.025 1.128 8.117

CR IP T

LP, iEOA >0.75 cm2/m2 (reference) SP, iEOA >0.65 cm2/m2 SP, iEOA ≤0.65 cm2/m2 LP, iEOA ≤0.75 cm2/m2 NYHA III/IV CHF Age (years) CKD

P-value

AC

CE

PT

ED

M

AN US

iEOA: indexed EOA; NYHA: New York Heart Association class; CHF: congestive heart failure; CKD: chronic kidney disease; LP: large prosthesis; SP: small prosthesis

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

CR IP T

FIGURE LEGENDS:

Figure 1. (A). Kaplan-Meier analysis of cumulative survival for the overall group, stratified by prosthesis size. There was no survival difference between SP and LP groups (p=0.21). (B). Kaplan-Meier analysis of cumulative survival for the overall group, stratified by standard PPM thresholds. Only severe PPM (p=0.007) was significantly associated with decreased survival.

AC

CE

PT

ED

M

AN US

CR IP T

ACCEPTED MANUSCRIPT

Figure 2. (A) Kaplan-Meier analysis of cumulative survival with binary iEOA thresholds for LP. A significant association with decreased survival existed at ≤ 0.75 cm2/m2 for LP patients (p<0.001) (B). Kaplan-Meier analysis of cumulative survival with binary iEOA thresholds for SP. A significant association with decreased survival existed at an iEOA threshold of 0.65 cm2/m2 (p=0.075)

CR IP T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

Figure 3. Cox PH model for cumulative survival, adjusted for various factors but stratified by prosthesis size and iEOA. Adjusted survival was most reduced for LP, iEOA ≤0.75 cm2/m2 (p=0.008) and SP, iEOA ≤0.65 cm2/m2 ; survival for SP and LP patients above the respective iEOA thresholds was substantially similar.

VIDEO LEGEND: The relevance and importance of this paper