Page 1 of 6

INTERMACS®

Interagency Registry for Mechanically Assisted Circulatory Support

CMS Report

June 23, 2006 – December 31, 2012

Prepared: April 21, 2013

Prepared by:

The Data Collection and Analysis Center

University of Alabama at Birmingham

For questions or comments contact:

James K. Kirklin, MD at

David C. Naftel, Ph.D at

Susan L. Myers at

INTERMACS Report to CMS

Statistical Analysis Plan for Evaluating Implant Volume

And Transplant Center Status

April 21, 2013

David C. Naftel, PhD

James K. Kirklin, MD

Susan L. Myers

1.  Introduction

  1. Purpose
  2. Cohort
  3. Outcomes for analysis
  4. Statistical methods
  5. Frequency of reports

2.  Patient Groups of Interest

  1. Age
  2. Implant volume
  3. Transplant center
  4. Implant strategy
  5. LVAD/BiVAD
  6. Combinations

3.  Results

  1. Pre-implant Characteristics
  2. Age
  3. Implant volume
  4. Transplant center/ DT center
  5. Implant strategy
  6. LVAD/BiVAD
  7. Combinations
  8. Survival
  9. Age
  10. Implant volume
  11. Transplant center/ DT center
  12. Implant strategy
  13. LVAD/BiVAD
  14. Combinations
  15. Competing Outcomes (death, transplant, recovery)
  16. Age
  17. DT center

4.  Risk Factor Analysis

5.  Conclusions

6.  References

1)  Introduction

During the spring of 2013, the Centers for Medicaid and Medicare Services (CMS) requested analyses from INTERMACS of several specific aspects of mechanical circulatory support device (MCSD) therapy. This document is in response to these requests. This report examines the effect of specific hospital characteristics on survival after device implant. Implant strategy (bridge to transplant, destination therapy, etc.) will also be examined. This report is actually a continuation and expansion of reports that INTERMACS has provided to CMS for several years.

a)  Purpose

The main purposes of this report include assessing the role of transplant center status and implant volume as possible correlates of survival post MCSD implant. The Medicare population will also be specifically examined. Pre-implant device strategy will be considered in the comparisons.

b)  Cohort

The inclusion criteria for a hospital to be enrolled in INTERAMACS are that the program has a viable MCSD implant program and the hospital is compliant with all regulatory requirements. (Please go to www.intermacs.org for a complete description of the enrollment process.) The inclusion criteria for patients to be enrolled into INTERMACS are that the patient receives an FDA approved, durable MCSD and the patient provides a signed informed consent. This report includes all primary implants from June, 2006 (initiation of INTERMACS) through December 31, 2012 and this report is restricted to continuous flow MCSDs which are essentially all devices since April 2008 (Figure 1). The cohort also excludes pediatric patients (< 19 years of age).

c)  Outcomes for analysis

This report focuses on survival after MCSD. All deaths that occurred while a device was place are included in the analyses. If a patient has a device exchange, then the patient is still followed for death. Patients are censored at time of transplant or device removal due to recovery of the heart/patient. In addition to death, time to transplant and time to recovery are also be examined.

d)  Statistical methods

Kaplan-Meier estimation for time-related events is the primary statistical method used in this report. Survival for subgroups is compared using the log rank statistic (this is the univariate version of Cox proportional hazard analysis). All confidence intervals presented in the figures are 70% confidence limits. 70% confidence limits are useful when visually comparing two Kaplan-Meier estimated curves. If the 70% confidence limits do not overlap, then the p-value for comparing the two lines will be approximately 0.05 or less. (Note: this is the same rational for presenting means along with standard error bars. The resultant interval approximately corresponds to 70% confidence limits.) The time related simultaneous occurrence of death, transplant and recovery will be estimated by competing outcomes methodology (reference 6b).

e)  Frequency of reports

Currently we provide quarterly reports to CMS. These reports include one report that summarizes the overall registry information and a second report that includes only patients who are 65 years old or older. Once this current report evolves to a report that is useful to CMS, we will have a discussion about what information/analyses would be useful to CMS on an ongoing basis and how frequently the analyses should be updated.

2)  Patient Groups of Interest

Several divisions of the total cohort have been created and compared.

a)  Age

The study population was divided according to patient age (<65 years and >= 65 years). Most results are analyzed separately for the two age groups. For some depictions, we have divided age into four groups in order to better understand the relationship between age and survival.

b)  Implant volume

We examined implant volume per hospital. We used the implant volume at each hospital over a two year period: 2011 and 2012. In addition to the patients who are enrolled in INTERMACS, the registry also tracks patients who are screened but are not enrolled. The implant totals for the two year period include patients who were eligible for INTERMACS (received a durable, FDA approved MCSD) but were not enrolled due to no informed consent. INTERMACS does not explicitly track implants of investigational devices. However, during the period of 2011 and 2012, there was very limited enrollment of investigational devices into FDA mandated studies. Therefore, device implants screened by INTERMACS should closely mimic the entire implant activity of durable devices at each hospital. We created four hospital groups according to quartiles based on the total implant count across the two years.

Table 1: Hospital quartiles based on total implants during 2011 and 2012 of durable FDA

approved MCSDs

Quartile Number of hospitals Range of two year implant totals

1 34 1-12

2 35 13-32

3 33 33-56

4 33 57-149

Total 135 1-149

This analysis provided additional information about the enrollment of patients into INTERMACS. During 2001-2012, the INTERMACS hospitals implanted 5083 FDA approved durable devices. The INTERMACS hospital enrolled 4592(90.3%) of the 5083. Figure 2 is a scattergram comparing the enrolled devices vs the implanted devices at each INTERAMCS hospital during 20011 and 2012.

c)  Transplant center/DT center

Currently, 149 hospitals are actively enrolled in INTERMACS (a few of the recently enrolled hospitals are yet to enroll patients). Of the 149 hospitals, 116 are Destination Therapy (DT) hospitals and 95 of these DT hospitals are also Transplant hospitals. Of the 149 hospitals, 33 are Non-DT hospitals and 18 of these Non-DT hospitals are Transplant hospitals. We plan to analyze 4 groups of hospitals: DT/TXPL (n=95), DT/Non-TXPL (n=21), Non-DT/TXPL (n=18) and Non-DT/Non-TXPL (n=15). (Note: of the 149 hospitals in INTERMACS, 135 have patients who are included in this report. This reduces the number of hospitals in each group.)

d)  Implant strategy

INTERMACS records categories of pre-implant device strategy: bridge-to-transplant listed, bridge-to-transplant likely to be listed, bridge-to-transplant moderately likely to be listed, bridge-to-transplant unlikely to be listed, destination therapy, rescue therapy, and bridge-to-recovery. In many of the previous INTERMACS analyses examining device strategy, we have combined the three categories of bridge-to-transplant where the patient is not actually listed at time of implant into one category labeled “bridge-to-candidacy”. We have used this designation for this report. We have also combined “rescue therapy”, “bridge-to-recovery” and a small number of “others” into a single category labeled “other”. These device strategies were compared.

e)  LVAD/BiVAD

Survival after LVAD was compared to survival after BiVAD (placement of both LVAD and RVAD in the same operation). Patients with BiVADs are included in all other depictions

f)  Combinations

In addition to the “main effects” of age, volume, transplant center and implant strategy, selected combinations (interactions) are examined.

3)  Results

a)  Pre-implant Characteristics

The patient groups may be associated with different pre-implant characteristics. For example, transplant centers may implant patients who have different characteristics than non-transplant centers. Understanding these differences may help to understand the role of transplant center in the application of MCSD therapy. The following patient groups were examined for pre-implant characteristics:

(1)  Age (Table 2)

(2)  Implant volume (Table 3)

(3)  Transplant center/DT center (Tables 4 and 5)

(4)  Implant strategy (Table 6)

(5)  LVAD/Bi-VAD (Table 7)

b)  Survival

The patient groups were examined for survival after MCSD implant. Survival at 30 days and at 1 year post implant was specifically estimated for each patient group listed below:

(1)  Age (Figures 3 and 4)

(2)  Implant volume (Figures 5, 6, 7 and 8)

(3)  Transplant center/DT center (Figures 9, 10 and 11)

(4)  Implant strategy (Figures 12 and 13)

(5)  LVAD/BiVAD (Figures 14 and 15)

(6)  Combinations (“interactions”) Many of the figures examine specific combinations of the groups

c)  Competing Outcomes (death, transplant, recovery)

Competing outcome methodology was applied to several patient sub-groups. This estimates a time related depiction of the “flow” of patients into death, transplant and recovery. These depictions may be especially useful in comparing transplant centers with non-transplant centers. Sub-groups of the following characteristics will be examined:

(1)  Age (Figures 16a and 16b)

(2)  DT Center (Figures 17a and 17b)

4)  Risk Factor Analysis

We have initiated a risk factor analysis for the event death in this cohort. We are examining all of the factors examined in this report plus a number of the factors that we have identified in earlier risk factor analyses. The purpose is to estimate the effect of age, volume, DT center, transplant center, device side (LVAD vs BiVAD), and implant strategy after adjustment for other risk factors. We are running a separate analysis for patients above 64 years old at time of implant. We will present the results of these analyses during our conference call to discuss this report.

5)  Conclusions

The INTERMACS leadership would like to collaborate with CMS to draw conclusions from the information presented in this report.

6)  References

a)  INTERMACS 5th Annual Report; Feb 2013, JHLT

b)  McGiffin D, Naftel D, et al. Predicting Outcome Following Listing for Cardiac Transplantation in Children: Comparison of Kaplan-Meier and Parametric Competing Risk Analysis. JHLT. 1997; 16: 713-22