Transcript Use of Health Status Measures

Global and Disease-Specific Health
Status in Patients Undergoing
Continuous-Flow Left Ventricular
Assist Device Placement
Kelsey Flint, MD
University of Colorado School of Medicine
Division of Cardiology
General Cardiology Fellow
Survival after LVAD
Survival curve for all continuous-flow LVADs in a clinical database
Kirklin JHLT 2014
Risk Models for Survival After LVAD
RV Failure Score
Bilirubin
X
INR
MELD
MELD
excluding INR
X
HMRS
X
X
X
DTRS
X
X
X
X
X
Creatinine/BUN
X
Vasopressor use
X
X
AST
X
X
Age
X
Albumin
X
Center volume ≤ 15
X
X
Pulmonary artery pressure
X
Platelet count
X
Hematocrit
X
X
AUC derivation
0.73
--
--
0.89
0.77
AUC validation
0.61-0.66
0.66
--
0.60
0.64
Reproduced from Levy JACC 2013
Health Status and Heart Failure
Heart failure-specific health status is associated
with death and hospitalization in patients with
heart failure who were medically treated in the
outpatient setting.
Heidenreich JACC 2006
Health Status and CABG
• Health status is also associated with mortality
following cardiac surgery, such as CABG
The physical
component score
of the SF-36 had
greater impact on
6-month mortality
than creatinine or
smoking history
Rumsfeld JAMA 1999
Health Status and Outcomes After LVAD
Health Status and Outcomes After LVAD
Flint JHLT 2013
Health Status and Outcomes After LVAD
• Therefore we decided to further study health
status in the LVAD setting:
– Outside the clinical trial setting, as patients
entered in to the clinical trials had uniformly very
poor health status
– With a generic as well as heart failure-specific
health status measure
– Measured before and shortly (3 months) after
LVAD placement
Methods
• The INTERMACS (Interagency Registry for
Mechanically Assisted Circulatory Support)
database is a prospective, observational registry
of all FDA-approved LVADs placed at participating
centers (N-158)
– Health status data are collected pre-operatively and
after LVAD at 3 ,6 and 12 months, then yearly
– Hospitalization and mortality outcomes are recorded
as they occur and at each follow-up period
– The INTERMACS protocol was approved by each
institution’s IRB, and individual patient consent was
obtained when mandated by the local IRB
Methods – Statistics
• Heart failure-specific health status – KCCQ
• Global health status – Euroqol 5D visual analog scale (VAS)
• Based on the skewed distribution of health status scores
towards poor health status, KCCQ and VAS scores were
broken down into quartiles
• Inverse propensity weighting (IPW) was used to minimize
bias associated with missing health status scores (40%)
• Kaplan-Meier method assessed the association between
health status and mortality and hospitalization
• Incremental prognostic value of health status was assessed
using IPW-weighted Cox proportional hazards
– We used the variables included in the HeartMate II Risk Model
(age, albumin, creatinine, INR, center volume) as the base
model
Table 1
Age
19-20
30-39
40-49
50-59
60-69
70-79
≥ 80
Female
Overall
KCCQ Q1
KCCQ Q2
KCCQ Q3
KCCQ Q4
N=3836
N= 552
N= 558
N=556
N=559
P-value
0.292
150 (3.9)
242 (6.3)
513 (13.4)
972 (25.3)
1330 (34.7)
597 (15.6)
32 (0.8)
807 (21)
12 (2.2)
34 (6.2)
90 (16.3)
141 (25.5)
189 (34.2)
80 (14.5)
6 (1.1)
26.8%
INTERMACS profile
1 (Critical cardiogenic shock)
13 (2.3)
36 (6.5)
70 (12.5)
131 (23.5)
207 (37.1)
97 (17.4)
4 (0.7)
23.8%
16 (2.9)
36 (6.5)
65 (11.7)
145 (26.1)
177 (31.8)
111 (20)
6 (1.1)
17.4%
20 (3.6)
28 (5)
68 (12.2)
126 (22.5)
212 (37.9)
101 (18.1)
4 (0.7)
14.7%
<0.001
Overall
KCCQ Q1
KCCQ Q2
KCCQ Q3
KCCQ Q4
N=3836
N= 552
N= 558
N=556
N=559
P-value
<0.001
500 (13)
66 (12)
27 (4.8)
23 (4.1)
26 (4.7)
2 (Progressive decline)
3 (Stable but inotrope dependent)
1361 (35.5)
1193 (31.1)
237 (42.9)
162 (29.3)
197 (34.2)
201 (36)
188 (33.8)
202 (36.3)
195 (34.9)
213 (38.1)
4 (Resting symptoms)
5 (Exertion intolerant)
6 (Exertion limited)
7 (Advanced NYHA III)
608 (15.8)
112 (2.9)
33 (0.9)
29 (0.8)
73 (12.2)
7 (1.2)
2 (0.4)
5 (0.9)
117 (21)
14 (2.5)
3 (0.5)
5 (0.9)
110 (19.8)
25 (4.5)
5 (0.9)
3 (0.5)
101 (18.1)
13 (2.3)
5 (0.9)
6 (1.1)
Table 1, continued
Overall
KCCQ Q1
KCCQ Q2
KCCQ Q3
KCCQ Q4
N=3836
N= 552
N= 558
N=556
N=559
Device strategy
Bridge to
transplant
Bridge to decision
P-value
0.110
788 (20.5)
97 (17.6)
112 (20.1)
102 (18.3)
141 (25.2)
1,329 (34.6)
169 (30.6)
178 (32)
183 (32.9)
167 (29.9)
1,694 (44.2%)
282 (51.1)
266 (47.7)
269 (48.4)
248 (44.4)
25 (0.7)
4 (0.7)
2 (0.4)
2 (0.4)
3 (0.6)
Dialysis within 48
hrs
Mechanical
ventilation
IABP
40 (1)
9 (1.6)
1 (0.2)
1 (0.2)
2 (0.4)
0.005
210 (5.5)
20 (3.6)
10 (1.8)
10 (1.8)
12 (2.1)
0.127
918 (23.9)
146 (26.4)
89 (15.9)
86 (15.5)
71 (12.7)
<0.001
Inotropes
3097 (80.7)
461 (83.5%)
449 (80.5)
429 (77)
446 (79.8)
0.116
Creatinine (mg/dl)
1.4 ± 0.6
1.4 ± 0.6
1.4 ± 0.8
1.4 ± 0.6
1.4 ± 0.6
0.880
Hemoglobin (g/dl)
11.4 ± 2.1
11.2 ± 2.1
11.6 ± 2.0
11.8 ± 2.0
11.8 ±2.1
<0.001
Sodium (mmol/L)
135.0 ± 4.7
134.1 ± 4.9
134.9 ± 4.8
135.3 ± 4.3
135.5 ± 4.1
<0.001
Albumin (g/dl)
3.4 ± 0.7
3.4 ± 0.6
3.5 ± 0.7
3.5 ± 0.6
3.6 ± 0.6
<0.001
INR
1.3 ± 0.4
1.3 ± 0.5
1.3 ± 0.4
1.3 ± 0.5
1.3 ± 0.3
0.165
Destination
therapy
Other
Baseline and 3-month Health Status
and Survival
Only 3-month
KCCQ was
associated
with 2-year
mortality
Baseline and 3-month Health Status
and Rehospitalization
3-month KCCQ
and VAS score
quartiles were
associated with
24-month
rehospitalizaiton
rate
Incremental Prognostic Value of
Baseline and 3-month Health Status
C-statistic
(base clinical
score only)
C-statistic
(base clinical +
health status
measure)
Outcome:
24-month mortality
C-statistic
(base clinical
score only)
C-statistic
(base clinical +
health status
measure)
Outcome:
24-month rehospitalization
Baseline KCCQ
0.60
0.61
0.51
0.50
(N=2225)
Baseline EQ-5D
VAS
0.60
0.60
0.51
0.52
(N=2205)
3-month KCCQ
0.60
0.66
0.52
0.55
(N=2060)
3-month EQ-5D
VAS
0.59
0.60
0.52
0.54
(N=2005)
Base clinical score was comprised of the variables included in the HeartMate II Risk
Score (Cowger JACC 2013) – age, albumin, creatinine, center volume, INR
Limitations
• Nearly 40% of patients were missing health status
data, introducing significant selection bias
– We attempted to account for this bias using IPW
• We were not able to characterize causes of death
and hospitalization. This information would be
useful as device-related complications leading to
death or rehospitalization would not be expected
to correlate with patient-reported outcomes.
Clinical Implications
• Very poor pre-operative health status should
not preclude LVAD implantation
• 3-month KCCQ score was associated with
long-term mortality, therefore serial
assessments of heart failure-specific health
status may help inform prognosis and goals of
care discussions
Implications for Future Study
• 24-month rehospitalization is poorly predicted by
the HeartMate II Risk Score or health status
– Further study is needed to better characterize, and
eventually predict rehospitalization in this population
beyond the existing single-center descriptions
• 24-month mortality is moderately-well predicted
by the HeartMate II Risk Score and health status
– Further study is needed to further characterize
patient, device and institution-related factors
associated with mortality
Summary
• Pre-operative heart failure-specific and global
health status are not predictive of mortality or
rehospitalization after LVAD
– Poor health status does not necessarily preclude LVAD
• 3-month KCCQ adds incremental prognostic value
to an established risk model for predicting 24month mortality after LVAD
– Serial health status measurements after LVAD may be
clinically useful
Thank you!
• INTERMACS DAAP
• Kathy Grady
• MAHI for analytic support and guidance
– John Spertus, MD
– Fengming Tang, MS
• Mentors and colleagues
– Larry Allen, MD, MHS
– Timothy Fendler, MD
Questions?
Reference slides
Reference material
Survival after DT HMII during the
the clinical trial vs. post-approval
Reference material
• Retrospective study examining the predictive
value of the HMII Risk Score in all patients
implanted with a CF-LVAD at Columbia
University Medical Center from 3/2004 to
9/2012 (N=201).
• The HMII RS had a c-stat of 0.56 for the
outcome of 90-day mortality in the Columbia
population
Thomas JHLT 2014
Reference
• The HMII RS was validated at
Barnes-Jewish Hospital in 269
consecutive patients receiving the
HMII (June 2005 – June 2013).
• The HMII RS had a c-stat of 0.70 for
90-day mortality
Adamo JACC HF 2015
Reference
Article
Population
Health Status
Measure
Result
Soto Circ 2004
1516 patients in
the EPHESUS
trial
KCCQ
KCCQ 1 month after MI complicated by
HF was associated with 1-year
mortality and hospitalization
Kosiborod Circ
2007
1358 patients in
the EPHESUS
trial
KCCQ
Change in KCCQ from 1 to 3 months
after MI complicated by HF was
associated with long-term all-cause
mortality and rehospitalization
Kato Circ J
2011
114 outpatients
with HF
MLWHF
Worse health status was associated
with increased risk of cardiac death or
hospitalization for HF and all-cause
mortality
Konstam Am J
Card 1996
6797 patients in
the SOLVD trial
HRQOL measure
created from
established sources
HRQOL was associated with
subsequent mortality and HF
hospitalizations
Additional studies examining the prognostic value of health
status in medically treated patients with heart failure.
Reference
Article
Population
Health Status
Measure
Result
Curtis Medical
Care 2002
1,778 patients
undergoing CABG
under usual care
SF-36
Worse PCS score was associated
with increased in-hospital mortality
and prolonged length of stay
Lindsay Heart
2001
183 patients
undergoing CABG
SF-36
Worse pre-operative health status
was associated with angina 10
months post-CABG
Koch Circ 2007
6,305 patients who
underwent CABG
Duke Activity
Status Index
Worse pre-operative and 6 or 12
month post-operative DASI were
associated with long-term mortality
Additional studies examining the prognostic value of health
status in patients undergoing CABG.
Continuous-Flow LVAD
•
There are continuous-flow LVAD’s approved by the FDA – the HeartMate II and the
Heartware
•
Blood is propelled from the LV to the aorta by a pump with a rotor that rotates in
response to the electromotive forces of the motor.
•
•
HeartMate II is an axial-flow pump – cylindrical rotor with helical blades, causing blood to accelerate in
the direction of the rotor’s axis
Heartware is a centrifugal-flow pump – rotors are shaped to accelerate blood circumferentially
(towards the outer rim of the pump). Unlike the HeartMate II, the Heartware has no mechanical
bearings or points of contact between the impeller and the pump housing
Quantifying and Qualifying Morbidity
in LVAD Candidates
Flint Circ HF 2012
Heart failure-specific health status, as measured by the KCCQ, may help
capture LVAD-responsive frailty, which likely does not influence outcomes postLVAD because LVAD-responsive components of a patient’s condition should be
largely reversed.
Methods – Health Status Measures
• Kansas City Cardiomyopathy Questionnaire (KCCQ)
– Measures heart failure-specific health status in 5 domains: physical
limitation, heart failure symptoms, social limitation, self-efficacy,
health-related quality of life
– Answers to questions are converted into a scale of 0-100, with higher
scores indicating worse health status
• Euroqol 5-Dimensions (EQ-5D)
– Measures global health status with the EQ-5D index and the visual
analog scale (VAS)
– The EQ-5D index is weighted to societal-based utilities to calculate
quality adjusted life years
– The VAS asks patients to indicate on a 100mm line how they would
rate their overall health from 0-100, with 0 being the worst health
imaginable
Methods:
Final patient
cohort
selection
Patients in the
INTERMACS Registry
N=11,162
Continuous-flow LVAD
N=9,690
Eligible for analysis:
N=3,836
Missing:
Did not receive
continuous-flow
LVAD
N=1,472
Enrolled prior to
mandatory health
status reporting
N=5,854
Baseline KCCQ
N=2,225
Baseline VAS
N=2,205
3-month KCCQ
N=2,060
3-month VAS
N=2,005
N=1,611
29% too sick, 29%
enrolled too late
N=1,631
33% too sick, 33%
enrolled too late
N=1,776
63% coordinator too
busy, 32% unspecified
N=1,831
48% coordinator too
busy; 21% unspecified
Results
Quartile 1
Quartile 2
Quartile 3
Quartile 4
Total
KCCQ score,
pre-operative
14.3 (9.6, 18.2) 28.1 (25.3, 31.5) 41.1 (37.5, 45.6) 63.8 (56.3, 74.5) 34.6 (21.4, 50.5)
VAS score,
pre-operative
10.0 (5.0, 20.0) 30.5 (30.0, 40.0) 50.0 (50.0, 60.0) 75.0 (70.0, 84.0) 43.0 (25.0, 65.0)
KCCQ score,
3-month
43.4 (34.8, 49.0) 62.2 (58.3, 65.6) 76.0 (72.9, 79.2) 89.6 (85.7, 93.8) 69.3 (54.2, 82.3)
VAS score,
3 month
40.0 (9.0, 50.0) 70.0 (62.0, 70.0) 80.0 (75.0, 80.0) 90.0 (88.0, 95.0) 75.0 (60.0, 85.0)
Median (IQR) of the health status scores in each quartile range
Results
Discussion
• Pre-operative heart failure-specific and global health status
were not associated with 24- month mortality or
rehospitalization following LVAD placement
• 3-month KCCQ score was associated with 24-month
mortality, and added incremental prognostic value when
added to a previously validated base clinical model
(Heartmate II Risk Model)
• 3-month KCCQ and VAS scores were associated with 24month rehospitalization but did not add prognostic value to
the Heartmate II Risk Model
• In general, the predictive value of the Heartmate II Risk
Model and health status were moderate for the outcome of
mortality, and poor for the outcome of hospitalization
Discussion
• LVAD is a significant intervention aimed at
reversing the hemodynamic effects of
advanced heart failure
– Unlike medical management of heart failure or
CABG, LVAD may potentially reverse the adverse
effects of heart failure
– Therefore, pre-operative health status may no
longer be relevant
Discussion
• Heart failure-specific
health status may
reflect LVADresponsive frailty.
• Therefore low heart
failure-specific health
status by 3 months
after LVAD (when
most patients will
have recovered from
surgery) portends a
poor prognosis by
signifying lack of the
expected benefit from
the device.
Flint Circ HF 2012