Transcript Recommending a Strategy

Evidence Review Workgroup
Advisory Committee on Heritable Disorders in
Newborns and Children
Report November 2008
James M. Perrin, MD
Professor of Pediatrics, Harvard Medical School
Director, MGH Center for Child and Adolescent
Health Policy
Current Progress and Activities
Pompe Disease – Final report submitted
SCID – Discussion today
Krabbe Disease – Review in process
SCID Report
Key authors:
– Ellen Lipstein, MD
– Alix Knapp, MS
Program director:
– James M. Perrin, MD
Staff:
– Marsha Browning, MD,
MPH
– Alex R. Kemper, MD,
MPH, MS
– Nancy Green, MD
– Lisa Prosser, PhD
– Denise Queally, JD
– Jennifer DeZarn, BA
– Sienna Vorono, BA
Severe Combined
Immunodeficiency (SCID)
Group of disorders characterized by absence of
both humoral and cellular immunity
Severe defect in T cell production and function,
with additional defects in B-lymphocytes and/or
NK cells depending on the mutated gene
At least 15 genes cause SCID when mutated
As protection from maternal antibodies wanes,
infants with SCID develop infections due to both
common and opportunistic pathogens
Rationale for Review
Without treatment SCID leads to death in early
childhood
Earlier treatment, particularly before the onset
of lung infection, may decrease mortality and
morbidity associated with SCID and with
treatment
Methods to screen infants for SCID using
quantitative PCR for T-cell receptor excision
circles (TREC) have been developed
Methods of Review
Systematic literature review, summarizing
the evidence available from published
studies
Assessment of critical unpublished data
and data presented at meetings from key
investigators
Key Review Questions
Incidence/prevalence
Natural history
– Timing of clinical onset
– Severity of disease and variations
– Genotype/phenotype
Screening
–
–
–
–
Methods of screening
Accuracy of screening; sensitivity/specificity
Methods of diagnosis
Risks and costs
Key Review Questions II
Treatment
– Methods
– Does treatment help?
– Does early treatment help?
– Availability
– Risks and costs
Critical information still needed
Materials Provided/Available
Draft review summary
Evidence table/abstracted articles
Bibliography of all identified articles
List of interviewees
Systematic Review
January 1988- October 2008: Medline, OVID In-Process
and Other Non-Indexed Citations database
– English language only
– Human studies only
– Eliminated: non-human data, reviews, editorials or other opinion
pieces, case-series of <4 patients, only contained adult subjects
or not addressing one or more of the key questions
– References also from nomination form, reviews
725 abstracts selected for initial review
60 articles selected for review and abstraction
Quality assessment directed toward study design
Papers Meeting Review Criteria
Study Design
Number of papers
Experimental intervention
0
Cohort study
11
Case-control study
8
Case series total
38
Sample size ≤ 10
11
Sample size 11 to 50
18
Sample size ≥ 51
9
Economic Evaluation
1
Other design
2*
Total
60
*Epidemiologic studies using retrospective record review (Jones et al. 1991)
and telephone survey (Boyle & Buckley, 2007)
Additional Expert Communication
Determined by literature review,
discussion with genetic experts
Included experts representing all major
issues related to SCID
Experts Contacted
Mei Baker - Department of Pediatrics, School of Medicine and Public Health, University of Wisconsin;
Newborn Screening Program, Wisconsin State Laboratory of Hygiene
Tony (Francisco) Bonilla - Harvard Medical School and Children’s Hospital, Boston, Mass
Rebecca Buckley -Duke University Medical Center, Durham, North Carolina
Anne Comeau -Deputy Director New England Newborn Screening Program University of Massachusetts
Lisa Filipovich -Cincinnati Children's Hospital Medical Center
Alain Fischer -Pediatric Immunology Department and the INSERM Research Unit
Alan P. Knutsen -Pediatric Clinical Immunology Laboratory, St Louis University Health Sciences Center
Ronald Laessig -Population Health Sciences and Pathology, University of Wisconsin
Edward McCabe -Clinical Biochemical Genetics, Department of Pediatrics, Mattel Children’s Hospital at
UCLA
Sean McGhee -Division of Pediatric Immunology, David Geffen School of Medicine at UCLA
Luigi Notarangelo -Division of Immunology, Children's Hospital Boston, Harvard Medical School
Hans Ochs -Pediatrics and Immunology, University of Washington
Sung-Yun Pai - Harvard Medical School and Children’s Hospital, Boston, Mass
Ken Pass -Molecular Medicine, New York State Department of Health, Albany, New York
Jennifer Puck -Department of Pediatrics, Institute for Human Genetics University of California, San
Francisco
Robert Vogt -Centers for Disease Control and Prevention, Newborn Screening Quality Assurance Program,
Atlanta, Georgia
Incidence
Chan/Puck 2005
– 1:105,000 live births
Extrapolation from XSCID samples sent to single
lab
Stephan et al., 1993
– 1:100,000 live births
Based on 5 years of referrals to specialized units in
France
Jones et al., 1991
– 52/100,000 births in Navajo families
Quality Assessment of Abstracted
Natural History Literature
Genotype/Phenotype Correlation
12
Data from retrospective screening studies in U.S. or similar
population.
0
Data from systematic studies other than whole population
screening.
5
Estimated from the known clinical features of the condition as
described for individual cases or short series.
7
Incidence (cases per 100,000), average within the U.S.
4
Data obtained from whole-population screening or comprehensive
national surveys of clinically detected cases.
1
Ia. As in I but more limited in geographical coverage or
methodology.
2
Extrapolated from class I data for non-U.S. populations.
0
Estimated from number of cases clinically diagnosed in U.S.
1
Adapted from Pandor et al. 2004, Pollitt et al. 1997
Natural History
Except for children diagnosed early in life,
because of affected sibling, most children are
diagnosed after recurrent (pulmonary) infections
All SCID subtypes exhibit infection with
opportunistic organisms, although timing of
onset may vary
Without treatment of underlying
immunodeficiency, children with SCID die in
early childhood from infection
Known phenotype/genotype differences do not
affect main findings related to infection and
death
Quality Assessment of Abstracted Screening
Test Characteristic Literature
Overall sensitivity and specificity of screening & false-positive rate
3
Data obtained from screening programs in U.S. population or similar.
0
Data from systematic studies other than from whole population screening.
3
Estimated from the known biochemistry of the condition.
0
Repeat specimen rate
0
Data obtained from screening programs in U.S. population or similar.
0
Data from systematic studies other than whole population screening.
0
Estimated from the known biochemistry of the condition.
0
Second-tier testing
1
Data obtained from screening programs in US population or similar.
0
Data from systematic studies other than whole population screening.
1
Estimated from the known biochemistry of the condition.
0
Adapted from Pandor et al. 2004, Pollitt et al. 1997
Main Screening Methods
Whole blood lymphocyte counts
Quantitative polymerase chain reaction
Enzyme-linked immunosorbent assay
(ELISA) of dried blood spots
Screening Test
Study
Population
Screening Methods
Accuracy of Screen;
Sens/Spec.
Chan,
Puck 2005
23 children with SCID
2 children with non-SCID
immunodeficiencies
242 anonymized
newborn screening cards
DNA amplification of TREC from dried blood spot
Among children known to have SCID, none had
detectable levels of TREC and all had detectable
β-actin.
2 children with non-SCID immunodeficiency had
detectable TREC.
*False positive rate: 1.5%
from routine nurseries;
5% from special-care
nurseries.
^Sensitivity: 84%-100%
^Specificity: 97-97.1%
Hague et
al. 1994
135 total children:
45 with SCID; 90 without
First available lymphocyte count; cut-off of
2.8x109/l indicates SCID
Children with SCID had significantly lower levels
of lymphocytes. Unlike the 5 control children with
low lymphocyte count, low lymphocyte count
persisted in children with SCID
*False-positive rate: 8%
^Sensitivity: 86.3%, and
^Specificity: 94.4%
Hennewig
et al. 2007
36 children with rotavirus
gastroenteritis;
18 with SCID
18 without SCID
Lymphocyte study
SCID children more likely to have
low WBC count:10/18 vs. 0/18,
eosinophilia: 12/18 vs. 0/18 relative lymphopenia:
17/18 vs. 10/18
absolute lymphopenia:16/18 vs. 4/18
^Sensitivity: 55.6% to 94.4%
^Specificity: 44.4% to 100%.
13 children with SCID
183 anonymized dried
blood spots, presumed to
be from children without
SCID
Dried blood spot study
Evaluated 2-tiered screening approach first
measuring IL-7 and then TREC only those with
elevated IL-7
*Combined specificity of
100% (confidence interval,
97-100%) *Combined
sensitivity of at least 85%
McGhee
et al. 2005
(Second-tier testing)
*Calculation stated in article
^Our calculation using data provided in article
Wisconsin Screening Experience
Number Screened:
47,250 (01/01/2008-08/31/2008)
Abnormal Results:
20
o Premature (<37 weeks)
11 (0.023%)
o Full term
9 (0.019%)
Inconclusive Results
76
o Premature (<37 weeks)
57 (0.121%)
o Full term
19 (0.040%)
Abnormal Results:
Inconclusive Results:
1 DiGeorge Syndrome
1 Downs Syndrome with sepsis at birth
1 Idiopathic T-cell lymphopenia
1 Leukocyte migration defect
4 normal Flow Cytometry results
9 normal results on repeated newborn
screening
2 pending cases
1 expired case
1 DiGeorge Syndrome
59 normal results on repeated newborn
screening
2 pending cases
14 expired cases
Courtesy of Dr. Mei Baker, presented at the Newborn Screening Symposium, November 2008
Treatment Methods
Over the last twenty years, three modes of
treatment for SCID have been
investigated:
– Allogeneic hematopoietic stem cell transplant
(HSCT) including bone marrow transplant
(BMT)
– Enzyme replacement therapy (ERT) for some
patients with ADA-deficient SCID
– Small trials of gene therapy
Quality Assessment of Abstracted
Treatment Literature
Effectiveness of treatment
47
I. Well-designed RCTs.
0
II-1. Well-designed controlled trials with pseudorandomization or no
randomization.
0
II-2. Well-designed cohort studies:
8
A. prospective with concurrent controls
0
B. prospective with historical control
1
C. retrospective with concurrent controls.
7
II-3. Well-designed case-control (retrospective) studies.
0
III. Large differences from comparisons between times and/or places with and
without intervention
4
IV. Opinions of respected authorities based on clinical experience, descriptive
studies and reports of expert committees.
35
Adapted from Pandor et al. 2004, Pollitt et al. 1997
Does Treatment for SCID Help?
Efficacy of HSCT over time: Large case-series
Study
Population
Significant Findings
Quality of
Evidence
Buckley et
al. 1999
Case series
89 children treated
with HSCT between
1982 and 1998
72 (81%) survived to follow-up (median follow-up 5.6 years)
HLA-identical transplant from a related donor: 12/12 (100%) survived
T-cell depleted haplo-identical: 60/77 (78%) survived
Survival not statistically related to genotype
Survival varied by race (white more likely to survive) and gender (all girls
survived)
36 children developed GVHD; most cases of GVHD required no
treatment.
IV
Stephan et
al. 1993*
Case series
117 patients treated
for SCID between
1970 and 1992; 85
with bone marrow
transplant
HLA-identical transplant from a related donor 21/25 (84%) survived)
Pheno-identical transplant (HLA genotypically haplo-identical) from
related donor 2/5 (40%) survived
HLA haplo-identical transplant without T-cell depletion 0/5 (0%) survived
T-cell depleted haplo-identical transplant 28/50 (56%) survived
22 children did not receive any type of transplant and died
10 received fetal liver transplant, 9 died post transplant
IV
van
Leeuwen et
al. 1994 *
Case series
31 patients between
the ages of 1-94
months; BMT
HLA-identical related: 6/10 (60%) survived
HLA haplo-identical related: 9/19 (47%) survived
HLA-matched unrelated: 0/2 (0%) survived
IV
* Potential patient overlap with Stephan et al. 1993 and van Leeuwen et al. 1994
Does Treatment for SCID Help?
Survival after HSCT: Long-term follow-up
Study
Population
Significant Findings
Quality of
Evidence
Antoine et
al. 2003;
Fischer et
al. 1990
Cohort
study
475 patients (566
transplants) from 37
European centers;
1968 to 1999
Three-year survival with sustained engraftment was 77% for HLAidentical and 54% for HLA-non-identical transplants
Survival has improved over time for both HLA-identical and HLA-nonidentical transplant recipients
SCID phenotype was not associated with difference in survival
Lung infection before HSCT and absence of a protective environment
significantly affected outcome
II-2 C,
IV
Cavazzan
a-Calvo et
al. 2007
Case
series
31 children 10-27
years post-transplant:
focus on factors
associated with longterm T-cell
reconstitution
Myeloablation patients more likely to have evidence of donor-derived
granulocytes and persistent naïve T-cells, as measured by TREC.
60% of TREC+ and 45% of TREC- had no clinical manifestations, at
average follow-up in the TREC+ group was 13 years and in the
TREC- was 16 years
IV
Friedrich,
Honig &
Muller
2007
Cohort
study
7 children with SCID
receiving HLA-identical
transplant, 25 with
HLA-haploidentical
transplant, all at least
10 years out from
transplant
3 patients’ T-cell numbers were decreased
4 patients’ PHA responses were decreased (all in HLA-haploidentical
group and no chemotherapy)
HLA-haploidentical with no conditioning had lower levels of naïve CD4+
cells and impaired B cell functioning
IV
Does Treatment for SCID Help?
Additional Long-term follow-up studies
Study
Population
Significant Findings
Quality of
Evidence
Gennery et al.
2001
Case series
19 children treated
from 1987-1998
using CAMPATH1MT depleted BMT
for SCID
19/30 children survived longer than 1 year post-BMT
Most deaths attributed to pre-existing infection
17 had normal immune function following transplantation
IV
Haddad et al.
1998
Case series
193 patients from
18 European
centers between
1982 and 1993:
focus on immune
reconstitution
116 alive with evidence of engraftment 5 months after BMT. 24 later
died (20%)
T-cell function improved during the 2 years after BMT and continued to
be better than B-cell function
Poor outcomes associated with: absence of T-cell reconstitution,
presence chronic GVHD 6 months after transplant, B- SCID
(multivariate analysis)
At last follow up (median, 6 years after transplant), 93% had normal Tcell function and 68% had normal B-cell function
IV
Slatter et al.
2008
Cohort study
36 children treated
with BMT with
depletion of T-cells
from non-identical
donor
No significant survival difference between children receiving
transplants depleted using anti-CD52 or anti-CD34 antibodies
5 patients in the anti-CD52 group and 2 in anti-CD34 group had GvHD
II-2
C
Does Early Treatment Help?
Efficacy of HSCT in neonates/infants
Study
Quality of
Evidence
Population
Significant Findings
21/22 (95%) infants alive at follow-up 51/67 (76%) who received
transplants at 3.5 months or older survived to follow-up
Median follow-up 5.6 years (range 3 months -16.5 years)
IV
Case series
22 infants transplanted
prior to 3.5 months of life
67 infants transplanted
later than 3.5 months of life
Kane et al.
2001
13 children transplanted
between 7 and 68 days-old
All patients alive and well 0.5-11.5 years post-transplant (median 3
years).
children required more than one transplant
All children achieved neutrophil engraftment and normal levels of
IgA, 7 have normal IgG, 12 have normal IgM
10/12 maintained normal development, of the other two: 1/12 has
developmental problems and 1/12 has motor delay
IV
21 children transplanted
prior to 28 days of life
(early treatment)
96 children transplanted at
a median age of 190 days
(range 45-516) (late
treatment)
20/21 (95%) early treatment children survived
One ADA-deficient patient died of CMV despite successful
engraftment
71/96 (74%) late treatment children survived
Mean time to significant T-cell function in all early treatment was 33
days and to normal T-cell function was 103 days
Mean TREC value peaked earlier post-transplant for early treatment
recipients but the 2 groups were indistinguishable by 5 years
II-2 C
Buckley et
al. 1999*
Case series
Myers et al.
2002*
Cohort study
* Potential patient overlap of Myers et al. 2002, Buckley et al. 1999
Early Treatment for SCID
161 SCID infants transplanted over the past 26
years, overall survival rate of 125/161 (78%)
Graph 1B. Transplanted after First 3.5 Months of
Life
Kaplan Meier Plot of 113 SCIDs Transplanted at Duke University
Graph 1A. Transplanted in First 3.5 Months of Life
Medical Center after the First 3.5 Months of Life
100
100
90
90
80
80
70
70
Percent
Surviving
Surviving
Percent
Percent Surviving
Kaplan Meier Plot of 48 SCIDs Transplanted at Duke University
Medical Center in the First 3.5 Months of Life
60
50
40
30
60
50
40
30
20
20
survival rate 96%
10
survival rate 71%
10
0
0
2
4
6
8
10
12
14
16
18
Years Post-Transplantation
Years Post-Transplantation
20
22
24
26
0
0
2
4
6
8
10
12
14
16
Years Post-Transplantation
Years Post-Transplantation
The Kaplan-Meier graphs from Dr. Buckley (with permission)
18
20
22
24
26
Treatment for SCID: Availability
From SCID expert interviews:
– Dr. Buckley reports there are fifteen major and 34
minor centers in the U.S. and Canada currently
performing stem cell transplantation for SCID.
– Dr. Notarangelo, Bonilla, and Pai stated that an
informal survey performed under the auspices of the
NIAID/Rare Diseases workshop identified 34 centers
in the United States and Canada that currently
perform HCT for SCID
Harms and Cost-effectiveness
Screening and Diagnosis
– No studies identified
Diagnosis
– None identified
Treatment (2 studies)
– 8/41 children undergoing HSCT developed auto-immune
hemolytic anemia; 3 died from complications
– 4 children (of the 9/10 who had successful gene therapy)
developed leukemia between 30 and 68 months after gene
therapy. 3/4 were successfully treated with chemotherapy and
regained poly-clonal T-cell populations
Cost-effectiveness (1 study)
– Using a deterministic decision-tree model, comparing universal
and targeted screening approaches, the authors assessed the
thresholds at which screening would be cost-effective from a
health care system perspective. Study reported, at a threshold
of $100,000 per quality adjusted life year, an 86% likelihood of
screening being cost-effective
Summary
Key findings:
– SCID affects at least 1/100,000 newborns in the US
– Several population-based screening trials are underway
(Wisconsin) or planned; to date no population-based
screening trial has been completed
– Without curative treatment, newborns develop severe,
often opportunistic, infections which lead to early death
– Treatment, most commonly with hematopoietic stem cell
transplant, decreases morbidity and mortality associated
with SCID
– Some evidence supports the notion that earlier treatment
may lead to better outcomes
Critical Needed Information
Prevalence of SCID
– A systematic method of case finding is needed in order to
determine prevalence accurately. The new consortium of
treatment centers (USIDNET) may serve as a method of more
systematic case-finding.
Accuracy of Screening
– Initial pilot screening data from Wisconsin suggest that a
relatively low false-positive rate. However, current data are
limited.
– Data regarding the accuracy of other screening methods in
population-based protocols, are not available
Feasibility of Screening
– Wisconsin’s experience provides at least pilot data re feasibility.
Data are needed regarding the ability of other newborn
screening laboratories to offer SCID screening
Critical Needed Information
Acceptability of Screening
– No data describe consumer or physician acceptance of newborn
screening for SCID
Is the evidence for early treatment enough?
Cost-effectiveness (of screening and treatment)
– Cost-effectiveness analyses utilizing measured costs and
utilities, as well as applicable sensitivity analyses, are needed
Adequacy of available treatment centers
– No current data address variation in treatment success among
centers or the number of centers in the United States and their
capacity to provide treatment for SCID. Data from USIDNET and
CIBMTR may, in the future, provide evidence on this topic
Thank you