A matching-adjusted indirect comparison of results from REDUCE and RESPECT—two randomized trials on patent foramen ovale closure devices to prevent recurrent cryptogenic stroke

Abstract

Aims

Two randomized clinical trials, REDUCE and RESPECT, demonstrated that patent foramen ovale (PFO) closure in combination with antithrombotic therapy was more effective for the prevention of recurrent ischemic stroke compared with antithrombotic therapy alone. The aim of this study was to determine the relative efficacy and safety of the PFO closure devices used in REDUCE (HELEX and CARDIOFORM Septal Occluders) compared with the device used in RESPECT (Amplatzer PFO Occluder).

Methods

An unanchored matching-adjusted indirect comparison (MAIC) of the PFO closure arms of the REDUCE and RESPECT trials was performed using patient-level data from REDUCE weighted to match baseline characteristics from RESPECT. Comparisons of the following outcomes were made between the devices assessed in the trials: risk of recurrent ischemic stroke; recurrent ischemic stroke one year after randomization; any serious adverse event (SAE) related to the procedure or device; and atrial fibrillation or atrial flutter as an SAE related to the procedure or device.

Results

After conducting the MAIC, baseline characteristics were well-matched between the two trials. Compared to RESPECT, PFO closure using the devices from REDUCE resulted in a hazard ratio of 0.46 (95% confidence interval [CI] 0.15–1.43; p = 0.17) for the risk of recurrent stroke. For the recurrence of stroke after one year, SAE related to the procedure or device, and atrial fibrillation or atrial flutter as SAE related to the procedure or device, the MAIC resulted in a rate difference of −0.68 (95%CI −2.06 to 0.70; p = .34), −1.29 (95%CI −3.82 to 1.25; p = .32), and −0.19 (95%CI −1.16 to 0.78; p = .71), respectively. These findings were consistent across scenario analyses.

Conclusions

This MAIC analysis found no statistically significant differences in efficacy and safety outcomes between PFO closure with the HELEX and CARDIOFORM Septal Occluders versus the Amplatzer PFO Occluder, as used in the REDUCE and RESPECT trials.

PLAIN LANGUAGE SUMMARY

The individual efficacy and safety of medical devices used for patent foramen ovale (PFO) closure in patients with stroke of unknown origin has been demonstrated in two independent trials: REDUCE (using the HELEX Septal Occluder and the CARDIOFORM Septal Occluder) and RESPECT (using the Amplatzer PFO Occluder). In the absence of a direct head-to-head trial for these devices, indirect treatment comparisons offer an alternative to assess their relative efficacy and safety. This study used a matching-adjusted indirect comparison to demonstrate that there were no significant differences between the devices used for PFO closure in the REDUCE and RESPECT trials in terms of safety outcomes.

Keywords:

• Patent foramen ovale
• cryptogenic stroke
• PFO closure
• matching-adjusted indirect comparison
• comparative effectiveness

JEL Classification Codes:

• I10
• I1
• I
• C00
• C
• C10
• C1

Introduction

Stroke remains a leading cause of death and long-term disability worldwide, imposing a staggering human and economic toll¹. Globally, more than 12.2 million new strokes occur each year—almost two-thirds of strokes occur in people younger than 70 years²—and disease-related costs have been estimated at nearly half a trillion dollars². Approximately one quarter of strokes are classified as cryptogenic, or embolic strokes of undetermined source^1,3; an estimated 30% to 40% of strokes in young adults are of unknown origin⁴.

Several studies have demonstrated the association between cryptogenic strokes and the presence of patent foramen ovale (PFO), a part of the fetal circulation allowing for the passage of blood from the right atrium to the left atrium that fails to close after birth⁵. PFO is of clinical significance when the intra-atrial anatomical communication serves as a channel for paradoxical embolism that contributes toward the risk of ischemic strokes⁶.

Recent evidence confirmed that percutaneous PFO closure with transcatheter devices, in combination with antithrombotic therapy, reduces the risk of stroke compared to antithrombotic therapy alone in carefully selected patients ≤60 years of age^7–9. However, many transcatheter device options have entered the market in recent years, presenting a challenge for treatment decision-making due to the lack of direct comparisons available.

The efficacy and safety of PFO closure vs medical therapy for patients with cryptogenic stroke has been demonstrated in two pivotal, open-label, randomized controlled trials (RCTs)—REDUCE¹⁰ and RESPECT^11,12. The REDUCE trial (sponsored by W. L. Gore & Associates) examined the HELEX Septal Occluder and GORE CARDIOFORM Septal Occluder devices, hereafter referred to as the Gore devices. In REDUCE, closure reduced the risk of recurrent stroke by 77% (hazard ratio [HR] 0.23; 95% confidence interval [CI]: 0.09–0.62; p = .002). The RESPECT trial (sponsored by Abbott Laboratories) examined the Amplatzer PFO Occluder, hereafter referred to as the Amplatzer device. In RESPECT, closure reduced the risk of recurrent stroke by 45% (HR 0.55; 95%CI: 0.31–0.999; p = .046). Comparing these hazard ratios between studies is challenging due to inherent differences in the study populations. The aim of the current study was to conduct an unanchored matching-adjusted indirect comparison (MAIC) to determine the relative efficacy and safety of the Gore devices and the Amplatzer device used in the REDUCE and RESPECT trials, respectively.

Methods

Overview

The MAIC methodology uses individual patient data (IPD) from one study and re-weights the patient baseline characteristics to match those reported in the comparator trial. This method adjusts for potential biases due to differences in patient characteristics that could impact treatment outcomes, thus allowing for comparisons between balanced populations¹³. Comparisons made with an MAIC methodology can be anchored if there is a common comparator arm in each trial to form a connected network, or unanchored if no common comparator exists and the network is disconnected. Considering the different medical treatments given to patients in the comparator arms of the REDUCE and RESPECT trials, this study used an unanchored MAIC approach.

Clinical trial overview

The REDUCE study included 664 patients with a PFO who had experienced cryptogenic stroke. Patients were randomized in a 2:1 ratio to undergo transcatheter closure of the PFO using the HELEX Septal Occluder or CARDIOFORM Septal Occluder in addition to antiplatelet therapy (n = 441), or to receive antiplatelet therapy alone (n = 223). Patients in the REDUCE study were followed for up to 60 months¹⁰. The RESPECT trial randomized 980 patients in a 1:1 ratio to receive closure with the Amplatzer PFO Occluder plus an antiplatelet or anticoagulation therapy for five months (n = 499), or antithrombotic therapy alone, which could include antiplatelets or warfarin (n = 481). The median follow-up time in RESPECT was 5.9 years^11,12. REDUCE and RESPECT were assessed for differences in study methods, baseline characteristics, presence of potential treatment effect modifiers, and the availability of variables and outcomes of interest in the treatment comparisons.

Device overview

The HELEX Septal Occluder¹⁴ is an implantable prosthesis and a catheter delivery system. The Occluder is made from expanded polytetrafluoroethylene patch material with a hydrophilic coating, supported by a nickel-titanium, wire frame. The double-disc device stops the shunting of blood between the two atriums by spanning and eventually blocking the septal defect¹⁴. The HELEX device received approval from the United States Food and Drug Administration (FDA) in 2007; it was replaced in the market and in the REDUCE trial by a new version, the GORE CARDIOFORM Septal Occluder.

The GORE CARDIOFORM Septal Occluder¹⁵, which received FDA approval based on positive results from the REDUCE trial¹⁶, is an implantable closure device of a platinum-filled, nickel-titanium wire frame covered with expanded polytetrafluoroethylene on a delivery system. The double-disc configuration is intended to prevent the shunting of blood between the right and left atria¹⁵.

The Amplatzer PFO Occluder¹⁷ is a self-expandable device made from a nickel-titanium wire mesh with two discs joined by a short connecting waist to increase the closing ability of the device. The device, which received FDA approval following favorable results from its RESPECT trial, is designed to prevent the shunting of blood between the right and left atria.

Outcomes of interest

Two efficacy and two safety endpoints that were commonly reported in REDUCE and RESPECT were considered for the present analysis. These included the risk of recurrent ischemic stroke, recurrent ischemic stroke one year after randomization, any serious adverse event (SAE) related to the procedure or device, and atrial fibrillation or atrial flutter as an SAE related to the procedure or device.

Statistical methods

First, a naïve comparison was conducted for each endpoint of interest, directly comparing the treatment groups without any adjustments for differences between the trial populations. Next, an unanchored MAIC was conducted using IPD from REDUCE¹⁰, obtained from its sponsor (W. L. Gore & Associates), and the reported aggregate data from the RESPECT trial final publication¹². Since IPD from RESPECT were not available, reported Kaplan-Meier curves were used to derive pseudo-IPD from this study using a validated algorithm¹⁸.

The MAIC methodology in this study attempts to recreate the population characteristics from RESPECT using IPD from REDUCE by assigning weights to individual patients, such that the weighted mean baseline characteristics in REDUCE closely matched those reported in the RESPECT trial. Variables were adjusted by estimating a logistic propensity score model in which individuals were weighted by the inverse of their propensity score¹³.

Matching variables were selected for their clinical and statistical importance in explaining variability in the outcomes of interest and their demonstrated imbalance between the REDUCE and RESPECT populations. IPD from REDUCE were used to determine whether variables had a significant effect on the outcomes of interest by means of univariate Cox regressions that screen for prognostic indicators. Characteristics used in the matching process of the primary analysis of the current study included medical history variables such as history of diabetes, smoking status, any prior strokes, atrial septal aneurysm, PFO shunt size, and days from qualifying event to randomization.

Two MAIC sensitivity analyses were performed to determine how the results of the primary analysis were affected by changes in the matching of variables. The first sensitivity analysis excluded the variable shunt size to account for the inconsistency in how shunt size is measured in the trials, as well as in clinical practice¹⁹. The second sensitivity analysis introduced age to the list of matching variables because age is typically associated with clinical outcomes. The choice of variables selected for the sensitivity analyses was determined in collaboration with clinical experts.

After matching, the relative treatment effects of the PFO closure devices, in combination with antiplatelet therapy, used in the two trials were estimated. For determining the risk of recurrent ischemic stroke, a regression coefficient from the Cox regression model provided estimates of hazard ratios (HR) with two-sided 95% confidence intervals (CI) to compare treatments. In this case, the null hypothesis of no difference observed between the two treatment arms was tested with a log-rank test. The incidence of recurrent ischemic stroke one year after randomization and both safety outcomes were assessed by means of an absolute rate difference and odds ratios (OR), and in this case, the null hypothesis was tested with a two-sided Fisher’s exact test. Statistical significance was considered at the 5% level for all analyses. The analyses in this study were conducted using the R code from the National Institute for Health and Care Excellence²⁰.

The results of this analysis are presented in terms of point estimates for the relative treatment effects from comparing PFO closure with the Gore devices in REDUCE versus PFO closure with the Amplatzer device in RESPECT. The 95% CI that accompanies these values represent the possible range around the estimate and the p-values represent the probability that the outcome would have arisen by chance. HR and OR point estimates above 1.00 indicate higher risk with the Gore devices from REDUCE versus the Amplatzer device from RESPECT, while values below 1.00 suggest lower risk. HR and OR values equal to 1.00 signify no difference between treatments. Rate differences reflect the extra risk associated with PFO closure using the Gore devices, compared to the Amplatzer device.

Results

Study design and patient characteristics

Study design and patient characteristics (Tables 1 and 2, respectively) were generally comparable between the REDUCE and RESPECT trials. The most notable difference in study design between the two trials was the control arms. In REDUCE, the control arm consisted of antiplatelet therapy alone, whereas in RESPECT, the control arm consisted of antiplatelet therapy or anticoagulation therapy (warfarin), at the discretion of the treating physicians. As such, the control arms could not be considered a common treatment arm, as required for an anchored MAIC, which justified an unanchored MAIC approach. Prior to matching, the biggest observed differences between populations were the number of days from qualifying event to randomization, with an average of 100 days in REDUCE and 130 days in RESPECT. Also, patients in the RESPECT trial were more likely to have an atrial septal aneurysm (36.1% vs 20.4%) and a large or substantial right-to-left shunt size (49.5% vs 42.8%). After matching, these baseline characteristics were balanced between the two studies, and the distribution of the weights were deemed appropriate with no overly influential individuals, although this decreased the effective sample size in REDUCE from 441 to 252 patients in the primary analysis.

Table 1. Study characteristics.

Trial characteristics	Reduce	Respect
Study design	Multinational, prospective, randomized, controlled, open-label trial with blinded adjudication of outcome events	Multinational, prospective, randomized, controlled, open-label trial with blinded adjudication of outcome events
Patient enrollment and randomization	N = 664 in a 2:1 ratio	N = 980 in a 1:1 ratio
Treatment arms	PFO closure with either the GORE^® HELEX^® Septal Occluder or the GORE^® CARDIOFORM Septal Occluder within 90 days after randomization, in addition to antiplatelet therapy (n = 441)	PFO closure with the Abbott^® AMPLATZER PFO Occluder within 21 days of randomization plus antiplatelet therapy (n = 499)
Control arms	Antiplatelet therapy consisting of aspirin alone, a combination of aspirin and dipyridamole, or clopidogrel (n = 223)	Medical therapy consisting of either aspirin, warfarin, clopidogrel or aspirin combined with extended-release dipyridamole (n = 481)
Previous cryptogenic stroke eligibility criterion	Within 180 days before randomization	Within 270 days before randomization
Shunt size classification	According to the number of microbubbles detected in the left atrium within three cardiac cycles after appearance in the right atrium, as observed on transesophageal echocardiography while the patient was at rest or during Valsalva maneuver. The presence of 1–5 microbubbles was classified as small, 6–25 microbubbles as moderate, and more than 25 microbubbles as large.	According to the number of infused microbubbles in the left atrium within three cardiac cycles after their appearance in the right atrium at rest or during Valsalva release, as seen on a transesophageal echocardiogram. Grade 1 indicated 1–9 bubbles; grade 2, 10–20 bubbles; and grade 3, more than 20 bubbles.
Follow-up	5 years	5.9 years

Abbreviations: PFO, patent foramen ovale; SD, standard deviation.

Table 2. Baseline characteristics of patients in the PFO closure arm before and after matching.

	REDUCE – before matching, n = 441	REDUCE – after matching, n = 252^b	RESPECT n = 499
Mean age in years	45.4	46.3	45.7
Male sex	59.2%	58.3%	53.7%
Mean number of days from qualifying event to randomization^a	100	130	130
PFO shunt size – large/grade 3^a	42.8%	49.5%	49.5%
Atrial septal aneurysm – yes^a	20.4%	36.1%	36.1%
Medical history
Current smoker^a	14.3%	15.0%	15.0%
Diabetes mellitus^a	4.1%	6.6%	6.6%
Systemic hypertension	25.4%	30.7%	31.7%
Peripheral vascular disease	3.6%	5.4%	1.0%
Previous transient ischemic attack (prior to index event)	5.9%	5.6%	11.6%
Previous stroke (prior to index event)^a	9.5%	10.6%	10.6%

^aWeighted variables for the MAIC primary analysis.

^bEffective sample size.

Abbreviation: PFO, patent foramen ovale.

Comparison of efficacy outcomes

There were no statistically significant differences in the estimates of the two efficacy endpoints between the PFO closure group in REDUCE and RESPECT (Table 3). Overall, the matching adjustment resulted in similar outcomes as the naïve comparisons. The unadjusted HR for the time to recurrent ischemic stroke endpoint in REDUCE compared to RESPECT was 0.72 (CI 0.30, 1.74) and the MAIC resulted in an HR of 0.46 (CI 0.15, 1.43). Similarly, the unadjusted OR for the recurrence of stroke events after one year was 0.64 (CI 0.18, 2.21) with a rate difference of −0.50 (CI −1.86, 0.86), and the MAIC resulted in an OR of 0.51 (CI 0.12, 2.25) and a rate difference of −0.68 (CI −2.06, 0.70). These results did not achieve statistical significance (p > .05).

Table 3. Comparisons before and after matching (PFO closure from REDUCE versus PFO closure from RESPECT).

	HR/OR (95% CI)	P-value	Rate difference (95% CI)	P-value
Time to recurrent ischemic stroke
Naïve comparison	HR 0.72 (0.30, 1.74)	0.46	–	–
MAIC, primary analysis	HR 0.46 (0.15, 1.43)	0.17	–	–
MAIC, SA 1	HR 0.50 (0.17, 1.51)	0.21	–	–
MAIC, SA 2	HR 0.47 (0.15, 1.45)	0.18	–	–
Recurrent stroke events one year after randomization
Naïve comparison	OR 0.64 (0.19, 2.21)	0.55	−0.50 (−1.86, 0.86)	0.48
MAIC, primary analysis	OR 0.51 (0.12, 2.25)	0.50	−0.68 (−2.06, 0.70)	0.34
MAIC, SA 1	OR 0.57 (0.14, 2.36)	0.75	−0.60 (−2.01, 0.81)	0.42
MAIC, SA 2	OR 0.52 (0.12, 2.26)	0.49	−0.68 (−2.06, 0.71)	0.34
Serious adverse event related to the procedure or device
Naïve comparison	OR 0.86 (0.44, 1.66)	0.74	−0.58 (−3.06, 1.90)	0.66
MAIC, primary analysis	OR 0.68 (0.31, 1.49)	0.46	−1.29 (−3.82, 1.25)	0.32
MAIC, SA 1	OR 0.65 (0.29, 1.43)	0.34	−1.44 (−3.93, 1.06)	0.26
MAIC, SA 2	OR 0.72 (0.33, 1.55)	0.46	−1.14 (−3.71, 1.43)	0.39
Atrial fibrillation or atrial flutter as a serious adverse event related to the procedure or device
Naïve comparison	OR 0.75 (0.12, 4.53)	1.00	−0.15 (−1.07, 0.78)	0.77
MAIC, primary analysis	OR 0.68 (0.09, 5.23)	1.00	−0.19 (−1.16, 0.78)	0.71
MAIC, SA 1	OR 0.57 (0.06, 4.96)	1.00	−0.26 (−1.18, 0.67)	0.59
MAIC, SA 2	OR 0.70 (0.09, 5.30)	1.00	−0.18 (−1.16, 0.80)	0.73

SA1 excludes shunt size and SA2 includes age as a matching variable.

Abbreviations: CI, confidence interval; HR, hazard ratio; MAIC, matching-adjusted indirect comparison; OR, odds ratio; SA, sensitivity analysis.

Comparison of safety outcomes

There was no statistically significant difference in the proportion of patients with an SAE related to the procedure or device among the PFO closure group in the RESPECT trial (21 patients) compared with the PFO closure group in REDUCE (16 patients). The results from this comparison yielded an OR of 0.86 (CI 0.44, 1.66) and a rate difference of −0.58 (CI −3.06, 1.90) for the unadjusted comparison, and an OR of 0.68 (CI 0.31, 1.49) and rate difference of −1.29 (CI −3.82, 1.25) for the primary MAIC (Table 3). Further, the incidence rate of atrial fibrillation or atrial flutter as an SAE related to the procedure or device was comparable between RESPECT (3 events) and REDUCE (2 events). This comparison yielded an OR of 0.75 (CI 0.12, 4.53) and rate difference of −0.15 (CI −1.07, 0.78) for the unadjusted comparison, and an OR of 0.68 (CI 0.09, 5.23) and rate difference of −0.19 (CI −1.16, 0.78) for the primary MAIC (Table 3). These results did not achieve statistical significance (p > .05).

Sensitivity analyses

The exclusion of shunt size and the inclusion of age from the list of matching variables did not impact the results compared with the primary MAIC. The point estimates from the sensitivity analyses did not reveal a statistically significant difference between the PFO closure groups from REDUCE and RESPECT (Table 3).

Discussion

PFO closure reduces the risk of recurrent stroke in patients with a prior cryptogenic stroke⁵. In the absence of direct comparisons between devices for PFO closure, an unanchored MAIC analysis was used in the current study to determine the relative efficacy and safety of the Gore devices from the REDUCE trial compared with the Amplatzer device used in RESPECT. This was the first population-adjusted comparison between REDUCE and RESPECT to compare the relative efficacy and safety of transcatheter devices used for PFO closure in these trials, providing a more equitable comparison than the unadjusted trial-level hazard ratios. The MAIC methodology allowed for the adjustment of differences in the baseline characteristics across the patient populations in the two trials by combining IPD from REDUCE and published aggregate data from RESPECT, thus attenuating any potential biases caused by imbalances in the patient characteristics.

The results from this study indicated that there are no statistically significant differences between the REDUCE and RESPECT PFO closure arms, both in terms of efficacy and safety. Adjustment for population differences did not substantially alter the crude cross-trial comparison. Safety profiles were also similar with respect to SAEs related to the procedure or device, and atrial fibrillation or atrial flutter as an SAE related to the procedure or device. These findings were consistent across the scenario analyses, which either included or excluded one clinically relevant variable from the matching criteria.

Previous comparisons of trials have established the benefit of PFO closure combined with antiplatelet treatment over medical therapy alone in reducing the risk of stroke recurrence, particularly in younger patients and patients with an inter-atrial shunt of substantial size^8,9,21. However, it has been found that patients undergoing a PFO closure may be at a higher risk of postprocedural atrial fibrillation regardless of the device used^22,23, which prompted our inclusion of this safety outcome in our study. In a meta-analysis on the device-related risk of atrial fibrillation, the authors were unable to demonstrate a statistically significant difference in the incidence of serious atrial fibrillation between the Amplatzer and Gore devices which was consistent with the results of the current study²⁴. In a network meta-analysis²⁵, the authors concluded that Amplatzer and Gore devices were superior to antithrombotic therapy alone but results also demonstrated an increased risk of atrial fibrillation for all the devices. However, a real-world evidence study found that PFO closure did not substantially increase the long-term risk of atrial fibrillation versus PFO diagnosed patients without PFO closure, suggesting that increased risk is more prominent during, or limited to, the postprocedural period²⁶. Further research is warranted to weigh the risks and benefits of treatment with specific PFO devices.

This study should be considered with the following limitations. First, there was no clear evidence of superiority from the results, and it was not feasible to demonstrate non-inferiority between the devices given the difficulty in defining a clinically meaningful non-inferiority margin from the low number of occurring events in these studies. Second, the planned medical regimen of anticoagulant therapy in the RESPECT trial was deemed a potential treatment effect modifier that could not be adjusted for because REDUCE did not allow anticoagulant therapy as concomitant treatment; this could pose a confounding factor in the comparison of outcomes. Limited evidence suggests that PFO closure has a larger benefit compared to antiplatelet therapy than to anticoagulant therapy²⁷, and therefore the inclusion of anticoagulation in RESPECT may have underestimated the relative effect of PFO closure. Notably, within the RESPECT trial, closure reduced stroke by 62% (HR 0.38; 95%CI: 0.18–0.79) compared to antiplatelet therapy but without a benefit compared to anticoagulation (HR 1.32; 95%CI: 0.43–4.03; p for interaction of .07). The adjustments in this analysis may not account for this potentially important treatment factor. However, this difference in the medical arms between trials should not confound estimates of atrial fibrillation or flutter after PFO closure. Third, only the baseline characteristics that were reported in the RESPECT trial publication could be matched; therefore, differences in characteristics that were unmeasured or unreported could have allowed for residual bias due to unmeasured confounders. Further, the adjustment required a considerable reduction in the sample size, which was attributed to the combined effect of including various covariates in the reweighting of patient characteristics employed in the MAIC, thereby reducing the precision of the estimates. Ultimately, we cannot determine if the comparisons between trials reflect the devices or the trial designs. Despite these limitations, the population-adjusted approach in this study substantially reduced the risk of bias caused by patient heterogeneity across trials compared with traditional indirect comparisons. Future research could consider an alternative method for comparison, such as propensity score matching; however, this would require complete IPD from both REDUCE and RESPECT PFO closure groups.

Conclusions

This study offers further evidence to confirm that percutaneous PFO closure with transcatheter devices, in combination with antithrombotic therapy, reduces the risk of stroke compared to antithrombotic therapy alone. Based on results from this MAIC, there were no statistically significant differences in terms of clinical outcomes between the devices used for PFO closure in patients with cryptogenic stroke in the REDUCE and RESPECT trials. These results can inform clinicians when making evidence-based decisions when opting for one device over another in clinical practice.

Transparency

Author contributions

MN, MH and MGM participated in the design of the study; MH and MGM performed the statistical analyses; LS, EL and SEK provided critical feedback on the scientific content of the study; MN was responsible for project oversight; MGM developed the first draft of the manuscript; and all authors participated in the review and approval of the manuscript.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgements

Editorial support was provided by Colleen Dumont, an employee of Cytel.

Declaration of financial/other relationships

SEK and LS participated as the National Principal Investigators for the REDUCE Clinical Study. SEK has received grant support from W. L. Gore & Associates and consulting fees from Abbott. LS is Chief Medical Officer and Divisional Vice President for Medical Affairs, Abbott Structural Heart. MN and EL are Associates at W. L. Gore & Associates. MGM and MH are employees of Cytel and have received consulting fees from W. L. Gore & Associates.

Additional information

Funding

The analyses and associated manuscript were completed by Cytel and were funded by W. L. Gore and Associates.