View
216
Download
0
Category
Preview:
Citation preview
SRTR
Transplant Benefit-BasedLiver Allocation
Robert M. Merion, MD, FACS
OPTN/UNOS Liver ForumAtlanta, GA
April 12, 2010
SRTR
Outline
• Introduction: organ allocation schemes• Transplant Benefit Score• Comparison to MELD Score• Simulation results: LSAM
– Comparison of deaths, life-years saved, and distance• Benefit score vs. MELD• Benefit score vs. geographic modifications of MELD
• Conclusions
SRTR
Background
• Currently, chronic liver failure patients waitlisted for deceased-donor liver transplantation are prioritized with respect to medical urgency in decreasing order of Model for End-stage Liver Disease (MELD) score
• Shortfall in donor livers increases pressure to make the best possible use of available organs
• It has been suggested that post-transplant survival should play a role in liver allocation
• Transplant benefit has been a central component of the allocation system since May, 2005.
SRTR
Organ Allocation Schemes
SRTR
Organ Allocation Policy
• Possible bases for organ allocation:
1. URGENCY: future wait-list lifetime
2. UTILITY: post-transplant lifetime
3. BENEFIT: combines (1) and (2)
SRTR
Organ Allocation: Example Expected Lifetimes
PatientID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example Expected Lifetimes
• Q: An organ is procured. To which patient should it be allocated? Patient
ID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example Expected Lifetimes
• Q: An organ is procured. To which patient should it be allocated?
• A: depends on the allocation rules …..
PatientID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example Expected Lifetimes
• Minimum WL lifetime:– Allocate to: #2
PatientID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example Expected Lifetimes
• Minimum WL lifetime:– Allocate to: #2
• Maximum post-LT lifetime:– Allocate to: #1
PatientID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example Expected Lifetimes
• Minimum WL lifetime:– Allocate to: #2
• Maximum post-LT lifetime:– Allocate to: #1
• Benefit (LT - WL):– Allocate to: #3
PatientID #
PredictedWL
Lifetime
Predictedpost-LTLifetime
1 7 10
2 2 3
3 5 9
SRTR
Organ Allocation: Example
ID#
Future Lifetime (patient)
WL LT Benefit =LT-WL
1 7 10 3
2 2 3 1
3 5 9 4
SRTR
Transplant Survival
Benefit Score
SRTR
Liver Transplant Survival Benefit
• Concept developed in collaboration with OPTN/UNOS Liver Committee since 2006
• Patient-specific and donor-specific• Separate models for waitlist and post-transplant lifetimes• Uses available factors other than MELD components• Difference in area under 5-year survival curves
– Predicted post-transplant lifetime minus predicted future waitlist lifetime
– Reflects life years gained through liver transplantation
SRTR
Waitlist Survival Model: Covariates
• Not just MELD• Creatinine, bilirubin, INR, albumin, sodium, dialysis,
age, BMI, diagnosis, HCC, diabetes, hospitalization status, prior malignancy, growth failure, previous time on waitlist, rate of change: creatinine, bilirubin, albumin
SRTR
Post-Transplant Survival Model: Covariates• Not just MELD• Recipient factors:
– Creatinine, albumin, age, diagnosis, diabetes, dialysis, hospitalization status, previous liver transplant, life support, portal vein thrombosis, previous abdominal surgery, hepatitis C, growth failure
• Donor factors: – Age, race, cause of death, donation after cardiac death
SRTR
Waitlist Survival
SRTR
Waitlist 5-Year Expected Lifetime
SRTR
Post-Transplant Survival
SRTR
Post-Transplant 5-Year Expected Lifetime
SRTR
Mean 5-Year Future Lifetime by MELD
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
MELD
Life
expec
tancy
(ye
ars)
-
Transplant
Benefit
post-LT
WL
SRTR
Transplant Benefit by MELD
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
MELD
5-ye
ar B
enef
it (ye
ars)
-
SRTR
Transplant Benefit vs. MELD• Q: If average transplant benefit increases with MELD, does
allocation by benefit amount to allocation by MELD?
SRTR
Transplant Benefit vs. MELD• Q: If average transplant benefit increases with MELD, does
allocation by benefit amount to allocation by MELD?• A: NO!
SRTR
Transplant Benefit vs. MELD• Q: If average transplant benefit increases with MELD, does
allocation by benefit amount to allocation by MELD?• A: NO!
– e.g., two patients could have similar (or equal) MELD scores, but very different benefit scores
SRTR
Transplant Benefit vs. MELD• Q: If average transplant benefit increases with MELD, does
allocation by benefit amount to allocation by MELD?• A: NO!
– e.g., two patients could have similar (or equal) MELD scores, but very different benefit scores
– MELD is not the only factor predicting waitlist and post-transplant survival
SRTR
5-Year Transplant Benefit by MELDBox-Whisker Plots
Rank correlation between MELD and transplant benefit score = 0.67
SRTR
Calculation of MELD and Benefit Scores
SRTR
MELD: CalculationVariable Coefficient
Intercept 0.6431
Not on dialysis:log [min{max(Creatinine,1),4}]
0.957
log (max{Bilirubin,1}) 0.378
log (max{INR,1}) 1.120
On dialysis 1.326
• Notes: – Multiply the score by 10 and round to the nearest whole number.– MELD score is bounded by 6 and 40 – HCC T2 candidates receive an exception score of 22.
SRTR
PELD: Calculation
• Notes: – Multiply the score by 10 and round to the nearest whole number. – Laboratory values less than 1.0 are set to 1.0 for the purposes of the
PELD score calculation.
Variable Coefficient
log (Bilirubin) 0.480
log (INR) 1.857
log (Albumin) -0.687
Patient is less than 1 year old 0.436
Growth failure 0.667
SRTR
Benefit Score: Calculation
• Calculation of benefit score by linear approximation is analogous to MELD score– more terms
• Comparison to statistically rigorous calculation– almost perfect correlation (rank correlation 0.99)– used in simulations to reduce computing time
• Linear approximation would be utilized by OPTN/UNOS if incorporated into a future allocation system, just as MELD is currently
SRTR
Microsimulation to Compare Allocation
Systems
SRTR
Requests by OPTN/UNOS Liver Committee
• LSAM Simulations Study Population – Data from waitlist candidates and donors during 2006 were
used for the simulations.• Compared multiple broader distribution, MELD-based
and transplant benefit-based allocation systems to the current allocation system
• For each run, we recorded: number of transplants, mean transplant benefit, life-years saved, total deaths, median distance traveled, percent shared, and percent distance traveled >100 nautical miles.
• Results averaged over 10 iterations
SRTR
InitialWaitlist
NewCandidates
DonatedOrgans
Wait List Post-graft
Survival
RemovalFrom
Waitlist
OrganAllocation/ Placement
WaitlistMortality
Relisting
Post-Transplant Mortality
Object Pascal SAS
LSAM Event-Sequenced Modeling
Waitlisted Patients’ histories
SRTR
LSAM: Number of Deaths
Deaths MELD/PELD System
Regional Sharing
Transplant Benefit
Waitlist 1,6601,602(-58)
1,519(-83)
Post-transplant 609
607(-2)
601(-6)
Post-removal 407397(-10)
384(-13)
Total 2,6752,606(-69)
2,504(-102)
8%
6%
(n) Indicates difference vs. prior column; n% indicates difference vs. MELD/PELD system
SRTR
LSAM: Life Years Gained by Transplant
MELD/PELD System
Regional Sharing
Transplant Benefit
Mean Extra Years From Transplant
1.561.63
(+0.07)2.01
(+0.38)
∑ Life Years Gained by Transplant
9,87510,225(+350)
12,448(+2,223) 23%
(n) Indicates difference vs. prior column; n% indicates difference vs. MELD/PELD system
SRTR
Equity: Age of Transplanted Patients
SRTR
Equity: Gender of Transplanted Patients
SRTR
Equity: Race/Ethnicity of Transplanted Patients
SRTR
Comparison of Allocation Rules for Adult Deceased Donor Livers
Yellow = UnchangedBlue = Current SystemGreen = Share Positive Benefit
Current SystemLocal – Status 1ARegional – Status 1ALocal – Status 1BRegional – Status 1BLocal – MELD/PELD 15Regional – MELD/PELD 15Local – MELD/PELD < 15Regional – MELD/PELD < 15National – Status 1ANational – Status 1BNational – MELD/PELD
Share Positive BenefitLocal – Status 1ARegional – Status 1ALocal – Status 1BRegional – Status 1BLocal – Transplant benefit score > 0Regional – Transplant benefit score > 0Local – Transplant benefit score < 0Regional – Transplant benefit score < 0National – Status 1ANational – Status 1BNational – Transplant benefit score
SRTR
Decrease in Total Deaths (vs. Current)
51
0
51 46
163
61
112
147
0
25
50
75
100
125
150
175
200
Share PositiveBenefit
Current Share 25L Share 25R RegionalSharingBenefit
RegionalSharing
MELD/PELD
Concentric 25 Five Zones
Num
ber
of T
otal
Dea
ths
-
SRTR
Increase in Life Years Saved via Transplant (vs. Current)
0
500
1,000
1,500
2,000
2,500
3,000
SharePositiveBenefit
Current Share 25L Share 25R RegionalSharingBenefit
RegionalSharing
MELD/PELD
Concentric25
Five Zones
Additio
nal
Life
Yea
rs S
aved
-
SRTR
Median Distance Between Donor Hospital and Transplant Center
0
50
100
150
200
250
SharePositiveBenefit
Current Share 25L Share 25R RegionalSharingBenefit
RegionalSharing
MELD/PELD
Concentric 25 Five Zones
Dis
tance
(nau
tica
l m
iles
) -
SRTR
Median Distance vs. Decrease in Total Deaths
0
50
100
150
200
0 50 100 150 200 250
Median Distance (nautical miles)
Dec
reas
e in
tota
l dea
ths
-
current
Shr Pos Benefit
Five Zones
Concentric 25
Reg Shr Benefit
Reg Shr MELDShr 25L
Shr 25R
SRTR
Median Distance vs. Decrease in Total Deaths
Current
Shr Pos Benefit
Five Zones
Concentric 35
Reg Shr Benefit
Reg Shr MELD
Shr35R, National 15
250_25250_35 250_22
250_29National
15/3515/32
15/29 15/2515/22
Concentric
Concentric
250_1535
250_1532 250_1529250_1525
250_1522MELDNaRegShr
MELDNa
SRTR
Percent Shared vs. Decrease in Total Deaths
Current
Shr Pos Benefit
Five Zones
Concentric 35
Reg Shr Benefit
Reg Shr MELD
Shr35R, National 15
15/3517/35
20/35
Share
250_25250_35 250_22250_29
Concentric
MELDNaRegShr
MELDNa
SRTR
Benefit Scores, Life-Years Saved, and Distance Traveled
Allocation SystemMean
BenefitScore
Life-YearsSaved in 1
Year
Average Distance
Regional Sharing Benefit 2.01 12,448 124
Share Positive Benefit 1.87 11,794 59
Regional Sharing MELD/PELD 1.63 10,217 127
Current 1.57 9,929 65
SRTR
Benefit Score: Reducing Complexity
• Can a modified benefit score be calculated with fewer factors?
• This work is ongoing, and decisions about the tradeoffs can be made on clinical and statistical grounds by examining the effect on the rankings of candidates under various scenarios with fewer factors
SRTR
Summary
• Incorporation of a measure of post-transplant survival into the allocation system is being actively explored
• No specific policies or proposals have been put forward
• The role of transplant benefit score as a criterion for liver allocation will require further discussion
• Input from the transplant community is extremely important
Recommended