In the United States Court of Federal Claims
OFFICE OF SPECIAL MASTERS
No. 10-398V
(To Be Published)
*************************
* Special Master Corcoran
SHEA KYLIE SULLIVAN, *
* Filed: February 13, 2015
Petitioner, *
* Entitlement Decision; Human
v. * Papillomavirus (“HPV”) Vaccine;
* Rheumatoid Arthritis (“RA”);
SECRETARY OF HEALTH AND * Molecular Mimicry; Homology;
HUMAN SERVICES, * Cumulative Effect
*
Respondent. *
*
*************************
Sarah McIntee, Nelson Mullins Riley & Scarborough, LLP, Washington, DC, for Petitioner.
Alexis Babcock, U.S. Dep’t of Justice, Washington, DC, for Respondent.
DECISION DENYING ENTITLEMENT1
On June 28, 2010, Shea Sullivan filed this action seeking compensation under the National
Vaccine Injury Compensation Program (the “Vaccine Program”),2 alleging that she incurred a variety of
arthritis-like injuries after receipt of the Human Papillomavirus (“HPV”) vaccine. Petition (“Pet.”) (ECF
No. 1) at 1, 3-4. After considering the record as a whole, and for the reasons explained below, I find that
Petitioner has failed to carry her burden establishing causation, and therefore has not demonstrated
entitlement to compensation under the Vaccine Program.
1
Because this decision contains a reasoned explanation for my action in this case, it will be posted on the United States
Court of Federal Claims’ website, in accordance with the E-Government Act of 2002, Pub. L. No. 107-347, § 205, 116 Stat.
2899, 2913 (codified as amended at 44 U.S.C. § 3501 note (2006)). As provided by 42 U.S.C § 300aa-12(d)(4)(B), however,
the parties may object to the decision’s inclusion of certain kinds of confidential information. To do so, Vaccine Rule 18(b)
permits each party fourteen (14) days within which to request redaction “of any information furnished by that party: (1) that
is a trade secret or commercial or financial in substance and is privileged or confidential; or (2) that includes medical files or
similar files, the disclosure of which would constitute a clearly unwarranted invasion of privacy.” Vaccine Rule 18(b).
Otherwise, the decision will be available to the public. Id.
2
The National Vaccine Injury Compensation Program comprises Part 2 of the National Childhood Vaccine Injury Act of
1986, Pub. L. No. 99-660, 100 Stat. 3755 (codified as amended, 42 U.S.C.A. § 300aa-10 to 34 (2006)) [hereinafter “Vaccine
Act” or “the Act”]. Individual section references hereafter will be to § 300aa of the Act.
I. Factual Background and Medical History
The record in this case consists of Ms. Sullivan’s medical records, a daily self-health log she
prepared, affidavits from her as well as her mother and father, the testimony of two experts, and medical
or scientific literature submitted by the parties in support of their respective positions. I have reviewed
the entire record as required by the Vaccine Act. In this ruling I address the sufficiency of Petitioner’s
evidence in support of an award of compensation.
A. Petitioner’s Medical History
Petitioner was nearly sixteen years old at the time she received her first Gardasil3 vaccination
(part of a series of three vaccinations) on June 27, 2007, during a wellness visit to the Pediatric and
Adult Medicine Clinic in Tustin, California. Pet’r’s Ex. 5 at 54-55; Pet’r’s Ex. 2 at 2 (birth certificate);
Pet’r’s Ex. 3 at 2 (vaccination record). Overall, Ms. Sullivan was assessed at the time as a “well
adolescent.” Pet’r’s Ex. 5 at 55.
On August 10, 2007, Ms. Sullivan’s mother contacted Valerie Kozak, M.D. (Petitioner’s
primary care physician) at Pediatric and Adult Medicine to report that Petitioner appeared to be
suffering from heart palpitations. Pet’r’s. Ex. 5 at 55. Ms. Sullivan was apparently seen again on August
29, 2007, when she received her second Gardasil vaccination. Pet’r’s Ex. 3 at 3. There are no
subsequent records of any medical visits between August 27, 2007, and the time Petitioner received her
third Gardasil vaccination on January 4, 2008. Pet’r’s Ex. 5 at 56.
On February 11, 2008, Ms. Sullivan’s mother called Dr. Kozak, this time to report that
Petitioner was suffering from a sore throat, headache, and abdominal pain, and Dr. Kozak prescribed an
antibiotic. Pet’r’s Ex. 5 at 56. Ms. Sullivan was not seen again by a physician until April 18, 2008
(fifteen weeks after receiving her third Gardasil vaccination), when she reported experiencing left knee
pain over the prior one to two weeks. Id. at 57. In the relevant medical history, it was noted that
Petitioner ran track at school. Id. On examination, Ms. Sullivan’s left knee proved to be tender medially
and in the patella (knee cap) region, but she had full range of motion and the knee was otherwise stable,
so she was merely referred to orthopedics for an evaluation (although the medical records do not
indicate whether she followed through with the evaluation). Id.
Months later, Ms. Sullivan went back to Dr. Kozak for a follow-up visit in October of 2008
regarding the knee pain that she had been experiencing as well as swelling in her hands, feet, and
forearms. Pet’r’s Ex. 5 at 56. The possibility of arthritis as an explanation for Petitioner’s symptoms is
recorded in the medical history from this visit. Id. Laboratory studies showed normal comprehensive
3
Gardasil is a quadrivalent recombinant vaccine (meaning it is made from genetically engineered material but does not
contain live viruses) manufactured from the L1 protein of four strains of HPV (strains 6, 11, 16, and 18). See generally
Gardasil Package Insert, available at http://www.fda.gov/BiologicsBloodVaccines/ Vaccines/
ApprovedProducts/UCM094042.
2
metabolic panel (“CMP”), complete blood count (“CBC”), and urinalysis results. Id. at 58-63. An
antinuclear antibody (“ANA”) screen, double–stranded DNA, smooth muscle antibody, C3
complement, cardiolipin antibodies, thyroid function, and direct antiglobulin tests were also negative
for abnormalities. Id.
Ms. Sullivan subsequently saw Eric Wei En Lee, M.D. at Orange Orthopedic Medical Group in
Orange, California on October 21, 2008, for an orthopedic evaluation. Pet’r’s Ex. 7 at 6-8. As the
treatment records from that visit indicate, Petitioner reported that her onset of symptoms had occurred
in April 2008 “while running track.” Id. at 6. She also related to Dr. Lee that she had experienced
another episode of pain while running on the beach during the same period, and characterized the pain
as constant and severe. Id. She specifically reported “objective instability, mechanical popping clicking,
weakness, and decreased range of motion.” Id. A physical examination performed on Petitioner
revealed “peripheral joint stiffness.” Id. at 7. An examination of her knee in particular revealed no
erythema, swelling, drainage, or sign of infection; other aspects of her exam also had normal results,
although there was a 1+ effusion and abnormalities with the patellar exam. Id. Dr. Lee diagnosed Ms.
Sullivan with a patellofemoral chondromalacia (damage to the patella) and malalignment, as well as
mild flexible pes planus (flat feet). Id.
Ms. Sullivan underwent an MRI of her knee on October 22, 2008, which showed intact menisci,
tendons, and ligaments. Pet’r’s Ex. 7 at 10. A moderate amount of joint effusion was observed
(although its etiology was unclear). Id. She returned to Dr. Lee’s clinic for follow-up on October 29,
2008, at which time (taking into account the MRI) she was again diagnosed with left knee
patellofemoral chondromalacia, leading Dr. Lee to recommend physical therapy. Id. at 11, 12.
On November 3, 2008, Ms. Sullivan presented to Scott Graham, M.D. at South County
Orthopedic Specialists in Laguna Woods, California for a second opinion regarding her left knee pain.
Pet’r’s Ex. 8 at 6-7. She again reported that her pain had existed since April of 2008 after “running on
the beach,” and although she could not remember if there had been a specific injury, she did recall
having pain and swelling the next day. Id. The medical records from that visit note that Petitioner’s
family history was significant for arthritis, diabetes, cancer, and stroke. Id. at 6. A physical examination
showed left knee swelling and tenderness, without warmth, but Petitioner nevertheless demonstrated a
full range of motion and no hip irritability. Id. Dr. Graham reviewed the imaging studies and agreed that
they were normal except for the presence of effusion. Id. He diagnosed Ms. Sullivan with a possible
occult medial meniscal tear, and gave her a cortisone injection. Id. at 7. There was no mention of
vaccination in Petitioner’s treatment records from this visit, nor was Gardasil, or any other vaccine,
identified as a possible cause of Ms. Sullivan’s injuries.
Petitioner had a follow-up visit with Dr. Graham on December 1, 2008. Pet’r’s Ex. 8 at 11.
Because she continued to report knee pain despite the cortisone injection she had received in November,
surgical intervention was recommended. Id. Ms. Sullivan subsequently underwent arthroscopic knee
surgery performed by Dr. Graham on December 16, 2008. Id. at 12-13. Her post-operative diagnoses
3
were: (1) medial compartmental synovitis scar tissue, (2) lateral tibial plateau grade III chondral fissure,
and (3) suprapatellar plica; follow-up physical therapy was recommended. Id. at 21-24. Ms. Sullivan
underwent physical therapy treatments at Knight Physical Therapy in Anaheim, California, but not
consistently, and she did not complete all of her sessions. See Pet’r’s Ex. 9 at 2-9. At a session on
February 2, 2009, she was noted to be doing well overall, although her knee was more swollen than
usual due to standing for an extended period of time at a recent concert, and she was urged to keep up
with her therapy. Id. at 11.
On February 11, 2009, Ms. Sullivan was seen by Dr. Graham for a post-surgery follow-up.
Pet’r’s Ex. 8 at 26. Petitioner now reported that although her knee pain had improved, she was
experiencing pain in multiple joints. Id. She thereafter went back to Dr. Kozak on April 6, 2009 (a year
after she claims to have first experienced knee pain) with complaints of joint pain in her arms and
hands, as well as a new kind of knee pain different from what she had previously experienced. Pet’r’s
Ex. 5 at 63. It was now noted (for the first time in the medical records) that Ms. Sullivan “associates
[her symptoms] with HPV vaccine.” Id. The contemporaneous medical records specifically note that
Petitioner had looked up information regarding the connection between the Gardasil vaccination and
rheumatoid arthritis (“RA”) on the internet and reported having found five cases associated with a class
action lawsuit regarding the vaccine. Id.
Ms. Sullivan returned to Dr. Kozak on July 13, 2009. Pet’r’s Ex. 5 at 64. By this time, Petitioner
reported that she had been diagnosed with “inflammatory” arthritis “per rheumatology” (id.),
presumably referring to the assessment of Thomas R. Powell, M.D., a rheumatologist in Orange,
California, who saw Petitioner that same month. See generally Pet’r’s Ex. 10. She later saw Dr. Kozak
again in August, at which time she repeated her concern that the Gardasil vaccine was the cause of her
arthritic condition. Pet’r’s Ex. 5 at 65. As a written phone record from August 3, 2009, indicates, Dr.
Kozak informed Ms. Sullivan’s mother that there were no scientific articles linking the Gardasil vaccine
to arthritis. Id. Laboratory work performed by Dr. Powell showed that Ms. Sullivan’s cyclic
citrullinated peptide antibody (“CCP”) (the presence of which is strongly associated with RA) was
elevated, while serum protein electrophoresis, rheumatoid factor (“RF”), ANA, CBC, and CMP results
were normal. Pet’r’s Ex. 10 at 3-9.
On September 9, 2009, Petitioner was seen by William Shiel, M.D., FACP, FACR, at Mission
Internal Medicine Group, Inc. in Mission Viejo, California for a rheumatologic evaluation. Pet’r’s Ex.
11 at 1-3. A physical examination was significant for hyperhydrosis of the feet, swollen proximal
interphalangeal (“PIP”) joints with decreased flexion, swollen wrists, trace swelling of the knees with
decreased flexion to 120 degrees, and tender metatarsophalangeal (“MTP”) joints. Id. at 2-3. Ms.
Sullivan’s erythrocyte sedimentation rate (a test used to measure the degree of inflammation present)
was slightly elevated at thirty-two. Id. 5-7. RF and ANA tests were negative, while her anti-CCP
antibody was strongly positive. Id. Petitioner was diagnosed with symmetric polyarthritis, based on
such test results and the generally-observed loss of range of motion of the wrists and PIP joints. Id. at 4.
4
Dr. Shiel ultimately felt that Ms. Sullivan’s symptoms were very suggestive of RA, and recommended a
number of medications to treat her symptoms. Id.
At a follow–up visit with Dr. Shiel on September 25, 2009, Ms. Sullivan complained of “[p]ain
all over the joints for [two] years” and was started on medication to treat her symptoms. Pet’r’s Ex. 11
at 10. Imaging studies included a normal bilateral wrist x-ray series, hand x-rays that showed mild
periarticular osteopenia, and foot x-rays that showed periarticular osteopenia. Id. at 11-14. Petitioner
was thereafter formally diagnosed by Dr. Kozak as suffering from Juvenile Rheumatoid Arthritis
(“JRA”).4 Pet’r’s Ex. 5 at 68. An addendum from Ms. Sullivan’s primary care physician on May 19,
2010, stated that she continued to have achy joints and shoulders, and swollen fingers, but her
medications had alleviated some of her symptoms. Id. at 69. Those same records note concerns about
why Petitioner was “ill” so frequently but offer no explanation. Id.
B. Expert Testimony
1. Dr. Richard Roseff – Petitioner’s expert, Richard Roseff, M.D., graduated from
Boston University School of Medicine in 1980 (after completing his undergraduate degree at Amherst
College). Tr. at 5; ECF No. 12-1 at 3. Dr. Roseff went on to complete his residency at Boston Medical
Center, followed by a two-year fellowship in rheumatology at Massachusetts General Hospital. Tr. at 5.
Dr. Roseff is board certified in internal medicine and rheumatology but not in pediatric rheumatology.
Id. at 6, 43. He is currently a member of the American College of Rheumatology, and he has a
rheumatology sub-specialty private practice in Connecticut. Id. at 7. Dr. Roseff has lectured on RA, but
he has not published any articles on this topic. Id. Dr. Roseff is also not an immunologist,
epidemiologist, or toxicologist, and he has no personal expertise in the subject of vaccine causation. Id.
at 80, 136-38.
Dr. Roseff’s opinion is based on his review of Petitioner’s medical records, affidavits (from
Petitioner as well as her parents), and medical or scientific literature (including articles on vaccine-
induced autoimmune illness as well as articles on the latency period5 between vaccination and the
development of autoimmune diseases). Tr. at 12. It is not part of Dr. Roseff’s regular practice to see
individuals who are under eighteen years of age (as Ms. Sullivan was at the time of her diagnosis). Id. at
43. Additionally, Dr. Roseff could not recall having diagnosed anyone with a vaccine-related arthritic
condition as part of his clinical practice. Id. at 76.
4
As both parties’ experts agreed, rheumatoid conditions that begin when a patient is sixteen are typically not classified as
juvenile, according to relevant diagnostic criteria. Tr. at 88, 92, 94. Therefore, although Dr. Kozak formally diagnosed Ms.
Sullivan with JRA (presumably due to her young age at the time of onset of her symptoms), there is no dispute in this case
that RA was the proper diagnosis (see, e.g., Tr. at 66, 94), and I accept it as such for purposes of this decision.
5
Dr. Roseff defined latency period to be the asymptomatic period following exposure. Tr. at 133.
5
Dr. Roseff’s opinion began with testimony about RA, which he characterized as an
“inflammatory polyarthritis of unclear etiology.” Pet’r’s Expert Report dated Feb. 2, 2011 at 3
[hereinafter “Roseff Report”]; see also Tr. at 10. As Dr. Roseff explained, RA has several patterns of
onset, but classically results in a symmetric (meaning bilateral) inflammation of joints associated with
stiffness and pain that, if left untreated or treated inadequately, can result in disability and even
mortality. Tr. at 10. The majority of individuals who develop RA fall into the same demographic as
Petitioner (i.e., young adult females), although Dr. Roseff later clarified that he believed Ms. Sullivan
fell on the younger end of that spectrum. Id. at 10-11.
Dr. Roseff acknowledged that certain hereditary and environmental factors may place an
individual at higher risk for developing RA (Tr. at 2), and that Ms. Sullivan’s illness could have been
caused by such non-vaccine related factors. Addendum to Pet’r’s Expert Report, dated Feb. 9, 2011
(ECF No. 12-3) at 2 [hereinafter “Roseff Amended Report”]. However, Dr. Roseff opined that it was
unlikely that Petitioner’s RA could be attributed to this type of random occurrence. Tr. at 17. He
initially reached this conclusion after reading various reports from the Gardasil Vaccine Adverse Event
Reporting (“VAERS”) database concerning the experiences of other individuals who developed RA
after receiving the Gardasil vaccine – suggesting to him a correlation between the two. Id. at 18. He
later admitted, however, that the VAERS database is a passive reporting system permitting anyone to
report an adverse event, regardless of whether a medical professional has concluded that the adverse
event can be linked to vaccination. Id. at 46.
As additional support for his causation opinion, Dr. Roseff pointed to findings from clinical
studies included in the Gardasil vaccine package insert, which indicated that of the 9,412 individuals
receiving placebo, two developed RA, while of 10,706 individuals who had received Gardasil, six
developed this condition. Roseff Report at 3. Dr. Roseff characterized such numbers as a “disturbing
trend,” while admitting that they were nevertheless small and not otherwise corroborated by any
statistical evidence from any specific scientific or medical studies regarding the Gardasil vaccine (or
even the HPV vaccine more generally). Id. at 3, 8.
Dr. Roseff opined that because various viral antigens have been implicated in the development
of RA, “[b]y extension, it is reasonable to suppose that immunizations can also be potential
environmental triggers for RA development.” Roseff Report at 3. As evidence of this concept, Dr.
Roseff asserted that “[i]t had been known for years that exposure to the rubella vaccine represents a
causal link to a chronic arthritis resembling RA in a small percentage of patients.” Id. (emphasis added).
Dr. Roseff also cited an article6 involving an attempt to link human parvovirus B19 to development of
RA and acute inflammatory arthritis in genetically predisposed individuals. Tr. at 21-22. Although no
studies have linked the wild HPV virus to the development of RA, Dr. Roseff testified that the vaccine
6
B. J. Cohen, M. M. Buckley, J. P. Clewley, V. E. Jones, A. H. Puttick & R. K. Jacoby, Human Parvovirus Infection in
Early Rheumatoid and Inflammatory Arthritis, 45 Annals of Rheumatic Diseases 832 (1986) (Pet’r’s Ex. 37).
6
was different because an individual is exposed to four strains of HPV all at once – meaning that the
vaccinated individual is “getting a pretty heavy dose of the papilloma virus.” Id. at 77-78.
For the precise mechanism by which the Gardasil vaccine could result in the development of
RA, Dr. Roseff proposed molecular mimicry. Tr. at 76. Under Dr. Roseff’s theory, (a) components of
the Gardasil vaccine present to the body an antigen,7 (b) resulting in an immune response (the
development of B cell lymphocytes), (c) which in turn produce antibodies to the viral sequences in the
vaccination that will in the future protect the body from a wild HPV infection. The body, however, may
also have homologous protein sequences that will also be attacked by the same antibodies, thereby
resulting in the onset of autoimmunity (where the body is attacking its own tissue).8 Id. at 76-77. In
reaching this conclusion that the Gardasil vaccination could cause cross-reactions leading to the onset of
autoimmunity, Dr. Roseff relied on Kanduc, D., Penta- and Hexapeptide Sharing Between HPV16 and
Homo Sapiens Proteomes, 1 Int’l J. Med. and Med. Sci. 383, 386-87 (Sept. 2009) (ECF No. 65-3)
(Pet’r’s Ex. 36) [hereinafter “Kanduc”], which examined the cross-reactivity potential of HPV-16 (one
of the strains of HPV included in the Gardasil vaccination) and finding that it shares thousands of
identical peptide motifs with human proteomes.9 Kanduc at 386. Dr. Roseff opined that, based on
Kanduc, there is likely sufficient homology between the viral components of the Gardasil vaccine and
human proteins for the development of autoimmune illness. Tr. at 19-20.
Dr. Roseff also proposed an alternative mechanism for how the Gardasil vaccine could produce
an autoimmune response leading to the development of RA. He testified that aluminum contained in the
Gardasil vaccine10 could be implicated in the development of RA, opining that as aluminum is “gobbled
up” by white blood cell macrophages, the process can stimulate an immunostimulatory cascade. Tr. at
77. Dr. Roseff acknowledged, however, that this aspect of his opinion was based solely on a single
article11 published by Judicial Watch, an advocacy organization, rather than a verifiable scientific
7
An antigen is defined as “any substance capable, under appropriate circumstances, of inducing a specific immune response
and of reacting with the products of that response, that is, with specific antibody or specifically sensitized T lymphocytes, or
both.” Dorland’s Illustrated Medical Dictionary 103 (32d ed. 2012) [hereinafter Dorland’s].
8
Dr. Roseff admitted that he did not know what particular amino acid chain was relevant in triggering the autoimmune
response that he alleges could have caused Ms. Sullivan’s RA. Tr. at 80.
9
A proteome is defined as “the complete set of proteins produced from the information encoded in a genome.” Dorland’s at
1535.
10
“Gardasil itself contains 225mcg of aluminum.” Pet’r’s Ex. 35 at 8. Aluminum acts as an adjuvant in the Gardasil vaccine
(similar to many other common vaccines), meaning that it is a substance that enhances the body’s immune response to an
antigen. See Dorland’s at 32 (defining adjuvant).
11
Judicial Watch Special Report: Examining the FDA’s HPV Vaccine Records – Detailing the Approval Process, Side-
Effects, Safety Concerns and Marketing Practices of a Large-Scale Public Health Experiment (June 30, 2008) (ECF No. 65-
2) (Pet’r’s Ex. 35) [hereinafter “Judicial Watch”]. The Judicial Watch article notes that unspecified testing reports showed
that Merck had tested the Gardasil vaccine against an aluminum-containing placebo, and the authors of the article argue that
“[u]sing a reactive aluminum-containing placebo instead of a non-reactive saline base can make vaccines seem safer than
they may actually be.” Judicial Watch at 8. The Judicial Watch article further indicated that the Merck study had identified a
difference in terms of initial side effects (such as injection site reaction) among the ten percent of individuals who received
7
study.12 Id. at 71-72, 78. Dr. Roseff cited no other literature supporting his assertion that aluminum in
vaccines could cause injury. Id. at 79-80.
One factual issue that Dr. Roseff’s opinion attempted to address was the lengthy time gap
between when Ms. Sullivan received the three separate Gardasil vaccines (between June 27, 2007, and
January 4, 2008) and when her symptoms actually began. (As noted below, after an onset hearing held
by the prior special master presiding over this case, a fact ruling was issued determining that onset of
Ms. Sullivan’s illness occurred sometime in “late March or early April 2008, in the few weeks prior to
April 18, 2008.” See Sullivan v. Sec’y of Health & Human Servs., No. 10-398V, 2013 WL 4011056, at
*16 (Fed. Cl. Spec. Mstr. June 30, 2013) [hereinafter “Ruling Regarding Finding of Fact”]). Dr. Roseff
acknowledged that the first time Petitioner sought care for her knee problem was on April 18, 2008, at
which time she reported experiencing knee problems for only a week or two. Tr. at 44-45.
Dr. Roseff admitted that he could not distinguish which of three Gardasil vaccinations that
Petitioner received was the primary cause of her RA, but argued that the build-up of all three
vaccinations in her system could have been the instigating factor. Tr. at 69. However, he acknowledged
that he could identify no particular study to support the conclusion that the series of Gardasil vaccines
that Petitioner received could have had such a “cumulative effect.” Id. at 85. Dr. Roseff also
acknowledged that if the first vaccine that Ms. Sullivan received in June 2007 represented the beginning
of this proposed build-up, the timeframe between her initial vaccination and onset of her alleged injury
in April 2008 was greater than two months (Id. at 54), but he maintained that even the ten months
between the first vaccination to onset would not be so long as to exonerate the Gardasil vaccination as
having caused Ms. Sullivan’s illness. Id. at 73.
To support his opinion regarding the lengthy temporal gap between vaccination and onset in this
case, Dr. Roseff attempted to analogize Ms. Sullivan’s circumstances to studies of other diseases
featuring long latency periods, such as Lyme arthritis. Tr. at 23.13 In the case of Lyme arthritis,
exposure to an antigen (caused by a tick bite that introduces a bacterium into the body) can result in
inflammatory arthritis with onset months after initial exposure. Id. The bacterium replicates in the
what Dr. Roseff characterized as the “true placebo” (saline) versus other individuals who received the aluminum-containing
vaccine. Tr. at 79. “By its own account, Judicial Watch is a ‘non-profit, non-partisan, tax-exempt 501(c)(3) organization
which as a public interest law firm specializes in deterring, monitoring, uncovering, and addressing public corruption in
government.’” Judicial Watch, Inc. v. U.S. Dep't of Justice, 185 F. Supp. 2d 54, 57 (D.D.C. 2002).
12
Indeed, Dr. Roseff admitted that he did not know whether the Judicial Watch article was authored by a doctor, or whether
its conclusions or methodology were peer-reviewed by the scientific community. Tr. at 51. The authors of the Judicial Watch
article themselves acknowledge that “Judicial Watch was not and is not interested in proving causality,” indicating that
“[o]nly science can do that.” Judicial Watch at 20.
13
For this aspect of his opinion, Dr. Roseff relied upon an article by Allen C. Steere & Lisa Glickstein, Elucidation of Lyme
Arthritis, 4 Nature Reviews Immunology 143 (2004) (ECF No. 59-3) (Pet’r’s Ex. 31). But when asked whether this study
provided him with merely a hypothesis, or if in fact he believed that it was more likely than not that this is what happened to
Ms. Sullivan, Dr. Roseff admitted that he did not know because the matter had not been studied adequately to answer the
question. Tr. at 59.
8
individual’s system, ultimately finding its way to the synovium (or the joint capsule) and causing an
inflammatory arthritis which may persist despite successful antibiotic treatment of the infection. Id. at
23, 59. Molecular mimicry has been offered to explain the process by which this arthritic reaction
occurs. Pet’r’s Ex. 31 at 5. Dr. Roseff acknowledged, however, that this analogy was limited in its
application. For individuals who develop Lyme arthritis, the bacterial infection has been found to be
present in the joint capsule, and thus at the locus of the arthritis. Tr. at 24. By contrast, Dr. Roseff was
unaware as to whether synovial biopsies had ever been conducted on individuals who had received the
Gardasil vaccine and then developed RA, to see if, in fact, there was evidence of any components of the
vaccine similarly present, acknowledging that the failure to find them present would greatly weaken his
argument. Id.
Dr. Roseff went on to discuss another article addressing a case study that considered from a
retrospective standpoint the records of ten lupus patients to evaluate the length of time in which each
individual developed lupus after receipt of the Hepatitis B vaccine. See N. Agmon-Levin, Y. Zafrir, Z.
Paz, T. Shilton, G. Zandman-Goddard & Y. Shoenfeld, Ten Cases of Systemic Lupus Erythematosus
Related to Hepatitis B Vaccine, 18 Lupus 1192, 1194-96 (2009) (ECF No. 39-4) (Pet’r’s Ex. 32)
[hereinafter “Agmon-Levin”]. The researchers found that there was variable timing for onset of disease
following vaccination, ranging from a few days to as long as a year, with an average time of onset
around two months post-vaccination. Id. at 1193; Tr. at 27. Dr. Roseff opined that based on such results,
Ms. Sullivan’s initial development of knee symptoms about two months after exposure to her last
Gardasil vaccine was a reasonable timeframe. Tr. at 28.
In addition, Dr. Roseff also cited an article evaluating a rise in an autoimmune form of diabetes
(Type 1, insulin dependent diabetes mellitus (“IDDM”)) in children in Finland following the
introduction of the Hemophilus influenza B vaccine, finding an increase in the diagnosis clustering in a
period starting at about the thirty-eighth month mark from the date of vaccination and lasting
approximately six months. Tr. at 32; see John Barthelow Classen & David C. Classen, Clustering of
Cases of Insulin Dependent Diabetes (IDDM) Occurring Three Years After Hemophilus Influenza B
(HiB) Immunization Support Causal Relationship Between Immunization and IDDM, 35 Autoimmunity
247, 250-52 (2002) (ECF No. 59-2) (Pet’r’s Ex. 30) [hereinafter “Classen”]. He also referred to a
prospective cohort study which attempted to evaluate the possibility of an autoimmune response
following annual influenza (“flu”) vaccination in healthy adults by looking at the change in antibodies
post-vaccination, using blood samples taken prior to vaccination and then at one month and six months
post-vaccination. Tr. at 29-30; see N. Toplak, T. Kveder, A. Trampus-Bakija, V. Subelj, S. Cucnik, &
T. Avcin, Autoimmune Response Following Annual Influenza Vaccination in 92 Apparently Healthy
Adults, 8 Autoimmunity Reviews 134, 137-38 (2008) (ECF No. 59-1) (Pet’r’s Ex. 29) [hereinafter
“Toplak”]. While the flu vaccination generally did not alter the percentage of healthy adults with
positive autoantibodies, the Toplak researchers found that six months post-vaccination, approximately
thirteen percent of patients either had higher titers of antibodies or developed new antibodies. Tr. at 30;
Toplak at 2. Dr. Roseff focused on the presence of antibodies and postulated that six months post-
9
exposure to a vaccination could actually be too soon to look for manifestations of clinical disease. Tr. at
30.
Dr. Roseff admitted that none of these studies involved the HPV vaccine generally (or the
Gardasil vaccine more specifically), but justified his reliance on such studies in the absence of relevant,
non-industry-funded, independent studies addressing the specific issue of RA following vaccination
with Gardasil. Tr. at 33-34. Dr. Roseff also acknowledged that he knew of no persuasive epidemiologic
or biologic evidence directly supporting his theory, but nevertheless opined “that it may just be too
early to establish a statistical link.” Roseff Report at 3. He maintained overall that there is “a strong
possibility, knowing what we know about immunizations, and Gardasil specifically, that” the
vaccinations that Ms. Sullivan received caused her RA. Id.; Tr. at 17.
In distinguishing reliable from unreliable studies, Dr. Roseff expressed the general opinion that
studies sponsored by the pharmaceutical industry are inherently suspect due to bias. Tr. at 67
(referencing articles that he had read regarding bias in studies sponsored by the pharmaceutical
industry); see also Joel Lexchin, Lisa A. Bero, Benjamin Djulbegovic & Otavio Clark, Pharmaceutical
Industry Sponsorship and Research Outcome and Quality: Systematic Review, 326 BMJ 1167 (2003)
(ECF No. 30) (Pet’r’s Ex. 33) [hereinafter “Lexchin”]. He specifically questioned the scientific
methodology of certain such studies. See, e.g., Tr. at 68. Thus, he opined that early studies sponsored by
Merck, which did not observe harmful effects associated with the Gardasil vaccination, were tainted –
both generally by Merck’s involvement, as well as more specifically by the failure to use what he
referred to as a “true placebo” group. Id. at 73. He testified that there were six initial studies conducted
by Merck that purported to compare individuals who received the Gardasil vaccination to a placebo
group, but only one study used a saline placebo (which Dr. Roseff deemed as the most scientifically
reliable control) while most of the others used an aluminum-containing placebo (which he viewed as
not a true placebo). Id. at 25. Dr. Roseff thus expressed the opinion that this group of studies was
ultimately not reliable. Id. at 73.
2. Dr. Carlos Rosé – Respondent’s expert, Daniel Carlos Rosé, M.D., graduated in
1977 from the University of Buenos Aires School of Medicine in Argentina, completing his residency
in internal medicine at the University’s hospital. Tr. at 87; Resp’t’s Ex. B. He then went on to an adult
rheumatology fellowship in the National Institute of Rehabilitation; Department of Medicine,
Rheumatology Division, Buenos Aires. Tr. at 87. In 1987, Dr. Rosé finished a pediatric residency at
Thomas Jefferson University in Philadelphia, Pennsylvania, which was followed by a fellowship in
pediatric rheumatology at Children’s Hospital of Philadelphia. Id. Dr. Rosé is board-certified in
pediatrics as well as pediatric and adult rheumatology. Id. He is currently a Professor of Pediatrics at
Thomas Jefferson University where he teaches pediatric rheumatology to medical students, residents,
adult fellows, and pediatric fellows, and he also lectures and publishes papers on the topic of RA. Id. at
89. He is a reviewer for several journals and on the editorial board for Rheumatology International. Id.
at 89-90. He also serves on various Data Safety Monitoring Committees – also known as Data Safety
10
Monitoring Boards (which are committees of independent experts responsible for monitoring patient
safety and treatment efficacy data during ongoing clinical trials).14 Id. at 90.
Dr. Rosé has been treating rheumatology patients for over thirty years. Tr. at 88. Since 1989 he
has practiced at the duPont Hospital for Children in Wilmington, Delaware, where he is currently the
Head of Pediatric Rheumatology with supervisory responsibility over four pediatric rheumatologists.
Tr. at 86-88. As part of his rheumatology practice, Dr. Rosé sees patients at least twice a week, ranging
from infants to young adults. Id. at 88. Dr. Rosé diagnoses both RA and JRA because he sees patients
up to eighteen years of age, and the relevant diagnostic criteria require a diagnosis of RA (rather than
JRA) if onset occurred after the patient was sixteen years of age.15 Id. at 90-91. On average, Dr. Rosé
estimated that he sees approximately one hundred and fifty rheumatology patients per month. Id. at 88.
Relying on his experience treating patients with rheumatologic disease, Dr. Rosé formulated his
opinion in this case after reviewing Petitioner’s medical records, affidavits from Petitioner as well as
her parents, and pertinent medical or scientific literature. Resp’t’s Expert Opinion Report dated Mar. 9,
2011 (ECF No. 19-1) [hereinafter “Rosé Report”] at 1; Resp’t’s Expert Opinion Supplementary Report
dated Apr. 30, 2011 (ECF No. 19-1) [hereinafter “Rosé Supplementary Report”]; Tr. at 91-92. Based on
this review, Dr. Rosé formed an opinion that it is more likely than not that the series of Gardasil
vaccinations that Ms. Sullivan received were unrelated to the onset and development of her RA. Tr. at
92, 94.
Dr. Rosé began by providing an overview of arthritis generally, and RA more specifically. Dr.
Rosé explained that RA is a disease of unknown etiology, with one hundred percent of cases being
idiopathic. Tr. at 95. He indicated that RA is a uniform disease (although many rheumatologists
distinguish between individuals with RA who are seronegative versus those who are seropositive).16 Id.
at 93. Dr. Rosé expressed the view that while infections can cause arthritis, most recent research
identifies genetic factors, rather than environmental factors, as the most likely mechanisms associated
with the development of RA (although he acknowledged that these findings may be influenced by the
relative ease of studying such genetic factors as opposed to environmental factors). Id. at 96.
Dr. Rosé opined that there is no scientific or medical support for an association between the
Gardasil vaccination and rheumatic disease. Rosé Report at 5-6. He indicated that although several
14
Dr. Rosé is currently involved in two such committees, and he has experience dealing with both committees responsible
for overseeing clinical trials funded by the pharmaceutical industry and committees responsible for overseeing non-industry
sponsored clinical trials. Tr. at 90-91.
15
Dr. Rosé noted an age distinction between RA and JRA – he indicated that the relevant diagnostic criteria put the age of
onset of JRA up to sixteen years of age. Tr. at 88. Accordingly, Dr. Rosé indicated that rheumatoid conditions that begin
when a patient is sixteen are not classified as juvenile. Id.
16
Dr. Rosé indicated that seropositive individuals have rheumatoid factor or anti-CCP antibodies in the serum. Tr. at 94. Of
those individuals who meet the diagnostic criteria for RA, eighty-five percent are seropositive, as was the case with
Petitioner, and fifteen percent are seronegative. Id.
11
human infections have arthritis as part of their clinical presentation or are associated with the
development of a rheumatoid-like disease, there are no known viral causes of RA. Tr. at 96. Moreover,
the wild-type human papillomavirus is not a known cause of acute or chronic arthritis in humans –
making it even less likely in his estimation that the Gardasil vaccine could be so associated. Rosé
Report at 5-6; Tr. at 98. Although he acknowledged that the Gardasil vaccine contains viral particles
from multiple strains of HPV rather than a single strain from the wild type infection, he maintained his
view that it was unlikely that the Gardasil vaccine can cause arthritic conditions. Tr. at 99.
Dr. Rosé further testified that his view was supported by scientific and medical literature (or at
least the absence of negating literature). Dr. Rosé opined that he could not find a single case of RA
established to have been caused by the Gardasil vaccination. Tr. at 98-99. Moreover, he expressed the
view that if such an association existed, he would expect to have seen it in his clinical practice, but
never has. Id. at 98. Indeed, clinical trial data collected to date (from a safety database built with data
from spontaneous reporting by those patients who participated in clinical trials and their treating
physicians) on the development of autoimmune disease following receipt of a different formulation of
the HPV vaccination17 (as well as other vaccinations) illustrated that reporting rates of overall
autoimmune events (including RA) did not differ between the vaccinated and control groups. Rosé
Report at 3; Tr. at 99 (citing Verstraeten at 6631).18 Dr. Rosé expressed the view that this is among the
best evidence available regarding the vaccine’s capacity (or lack thereof) to produce RA, as there
currently are not (and may never be) any other case-control prospective epidemiological studies
addressing this specific issue. Rosé Report at 3; Tr. at 99-101.19
Dr. Rosé went on to attempt to rebut Dr. Roseff’s proposed mechanism by which RA could
result after receipt of Gardasil. Although he agreed with Dr. Roseff’s general characterization of the
concept of molecular mimicry, he opined that there was no evidence that this mechanism is implicated
in the development of RA; in fact, he indicated that it has not even been discussed in serious scientific
or medical literature as a potential mechanism for the development of RA in at least the last ten years,
17
This study involved a collection of “[a]ll completed or ongoing controlled, randomized studies of AS04 adjuvanted HPV-
16/18, HSV and HBV vaccines conducted by GSK Biologics [GlaxoSmithKline, the manufacturer of those vaccines] or
collaborators,” with one exception. Thomas Verstraeten, et al., Analysis of Adverse Events of Potential Autoimmune
Aetiology in a Large Integrated Safety Database of AS04 Adjuvanted Vaccines, 26 Vaccine 6630, 6631 (2008) (Resp’t’s Ex.
A, No. 5) (ECF No. 19-6) [hereinafter “Verstraeten”]. This study collected data on vaccinations containing the AS04
adjuvant; Cervarix (as opposed to Gardasil) is the human papillomavirus-16/18 (HPV-16/18) vaccine from GlaxoSmithKline
Biologics that contains this adjuvant. Id. at 6630. Thus, the research referenced in this article involves a different HPV
vaccine than that at issue in this case.
18
The “objective of this integrated analysis was to assess safety of AS04 adjuvanted vaccines with regard to adverse events
(AEs) of potential autoimmune aetiology, particularly in adolescents and young adults.” Verstraeten at 6630. It thus
attempted to address the potential problem with smaller studies, which may not by themselves detect rare events. Id. at 6631.
19
“Reporting rates of overall autoimmune events were around 0.5% and did not differ between the vaccinated and control
groups. The relative risk (vaccine/control) was 0.92 [] in the HPV-16/18. Relative risks calculated overall, for disease
category or for individual events were closer to 1, and all confidence intervals around the relative risk included 1, indicating
no statistically significant difference in event rates between vaccine and control groups.” Rosé Report at 3 (citing
Verstraeten).
12
even though it was previously considered, a fact he attributed mainly to an evolving understanding of
RA. Tr. at 104-05. He also emphasized that (as Dr. Roseff acknowledged in his expert report) the
relationship between the development of anti-CCP antibodies associated with RA and their reactivity
with host tissues is unclear (thus underscoring that the biological mechanism by which RA develops is
not fully understood). Rosé Report at 4; Roseff Report at 3.
Dr. Rosé specifically took issue with the assumption that the existence of some homology
between peptide chains in human tissue and the components of vaccines logically leads to the
conclusion that an autoimmune process will occur. Rosé Report at 4-5; Tr. at 106-07. Dr. Rosé
indicated that short peptide homology, referenced in the Kanduc article relied upon by Dr. Roseff in
formulating his opinion, “is at most a hypothesis generation step towards a theory of molecular
mimicry.” Rosé Report at 4. Moreover, he noted that because there are only a limited number of amino
acids in nature that can be assembled in a limited number of ways to make proteins, it is not uncommon
to find five amino acids that are homologous merely by chance. Tr. at 107. Dr. Rosé further testified
that the Kanduc article made what he termed “extremely far-fetched statements” regarding potential
side effects associated with receipt of the HPV vaccine, including setting forth (based on peptide
homology) “a whole list of diseases that bear limited relationship with immune responses,” yet even the
authors had not suggested or concluded that receipt of the HPV vaccine would result in the development
of RA. Id.
Dr. Rosé also questioned the logic of Dr. Roseff’s opinion that the series of Gardasil vaccines
that Ms. Sullivan received could have cumulatively resulted in the onset and development of her RA.
Tr. at 108. Dr. Rosé noted that the vaccine contains only microdoses of aluminum insufficient in
volume to have a negative effect in the body, while the protein components of the vaccine would
themselves not accumulate in similar fashion (if at all). Id. at 108-09. Alternatively, Dr. Rosé asserted,
if Dr. Roseff meant to argue that “more challenge”20 posed by the additional Gardasil vaccinations
served to overstimulate Ms. Sullivan’s immune system, then her medical history did not support such a
theory because there was no evidence in the treatment record of any instances in which Ms. Sullivan
experienced a physiologic response (e.g., a fever) following receipt of vaccination, as would be
expected if the vaccinations were affecting her in this manner. Id. at 109.
Dr. Rosé then pointed out what he saw as deficiencies in Dr. Roseff’s explanations for the lapse
of time between Petitioner’s receipt of the Gardasil vaccinations and her development of RA. Tr. at 97-
98. As a general matter, Dr. Rosé acknowledged that there could be a variable period of time between a
viral or bacterial infection (such as Lyme disease) and the development of some arthritic conditions. Id.
20
Challenge-rechallenge is “a paradigm for exploring whether one substance caused an adverse reaction. Under this model,
an individual who has had an adverse reaction to the initial vaccine dose (the challenge event) suffers a worsening of
symptoms after a second or third injection (the rechallenge event).” Viscontini v. Sec'y of Health & Human Servs., No. 98-
619V, 2011 WL 5842577, at *22 (Fed. Cl. Oct. 21, 2011) (quoting Doe/70 v. Sec'y of Health & Human Servs., 95 Fed. Cl.
598, 603 (2010) (quotations omitted)).
13
But he contested Dr. Roseff’s assertion that such conditions were analogous to RA.21 Id. at 97.
Moreover, in the case of Ms. Sullivan, he opined that the approximately three-month span between the
onset of her first symptoms (around April 2008) and the completion of her Gardasil series (January 4,
2008) seemed too lengthy not to allow for the possibility of some intervening event as the real cause.
Rosé Report at 2, 6. In Dr. Rosé’s view, a relationship between the Gardasil vaccine and Ms. Sullivan’s
RA would be conceivable only if the period were shorter – two to four weeks between onset and last
vaccination. Id. at 6.
C. The Chao Study
The parties debated the evidentiary significance of an epidemiological study cited by
Respondent in rebuttal of Petitioner’s claim. C. Chao, et al., Surveillance of Autoimmune Conditions
Following Routine Use of Quadrivalent Human Papillomavirus Vaccine, 271 J. Intern. Med. 193 (2012)
(ECF No. 49-4) (Resp’t’s Ex. F) [hereinafter “Chao Study”]. This peer-reviewed observational study22
analyzed a database comprised of the medical histories of approximately 189,000 California women
(members of two of Kaiser Permanente’s managed care organizations in the State) to determine whether
the studied population developed a variety of autoimmune conditions23 after receiving the Gardasil
vaccine. Chao Study at 194. The study monitored individuals for 180 days (six months) after receipt of
vaccine one, vaccine two, and vaccine three. Tr. at 100-01; Chao Study at 194-95. As vaccine three was
given six months after vaccine one, the total duration of the study was twelve months. Tr. at 100-01;
Chao Study at 194-95. The researchers compared the results of the studied vaccinated population with
unvaccinated, similarly-situated individuals also enrolled with Kaiser Permanente in Southern
California, in order to compare incident rate ratios for the identified autoimmune conditions. Chao
Study at 194-95. Based upon the data reviewed, the researchers did not observe an increased risk of
21
Dr. Rosé particularly distinguished Lyme arthritis from RA, noting that Lyme arthritis is a single joint disease with a
living microorganism (a bacterium) present in a patient’s joint, whereas in a case such as this a vaccine administered
remotely in an arm allegedly produces inflammation in multiple joints throughout the body. Id. at 108. Dr. Rosé
acknowledged that in the case of Lyme arthritis, approximately ten percent of children and approximately thirteen percent of
adults experience arthritic inflammation despite no evidence of persistent bacterial infection. Id. at 131-32. But he
characterized such inflammation not as a dormant/latent condition that hides or takes time to occur, but as a persistent
condition that is treatment-resistant, existing even after the initial infection is successfully resolved with antibiotics. Id. at
133-34.
22
In an observational epidemiologic study, researchers analyze groups of individuals who were exposed to a test agent,
comparing them with groups not so exposed. Michael D. Green, et al., “Reference Guide on Epidemiology,” in Reference
Manual on Scientific Evidence 549, 555-56 (3d ed. 2011). The Chao Study investigators drew inferences about the side
effects of Gardasil based on historic data, where the assignment of subjects into a treatment group (women who received the
Gardasil vaccine) versus a control group (women who did not) was outside of the investigators’ control. See Chao Study at
194.
23
The diseases monitored in the study included rheumatologic / autoimmune disorders, including immune
thrombocytopenia, autoimmune hemolytic anemia, systemic lupus erythematosus, RA, and JRA. Chao Study at 194.
14
developing RA (or other autoimmune conditions) following receipt of the Gardasil vaccine.24 Id. at 196,
201; Tr. at 101.
While acknowledging that he did not have any special expertise in identifying conflicts of
interest in scientific studies (Tr. at 62, 84-85),25 Dr. Roseff asserted that the Chao Study was biased
because it had been sponsored by a pharmaceutical company (Merck). Id. at 35-36. In support of his
view, Dr. Roseff pointed to articles like Lexchin that in his opinion raised legitimate questions about the
impartiality of such pharmaceutical industry-sponsored studies; the article cited selection of an
inappropriate comparator to the product being investigated and publication bias as potential
explanations for this bias. With regards to the Chao Study, Dr. Roseff specifically questioned the
control groups utilized in the study, arguing that he could not ascertain whether it was a valid
comparison to the group that received the Gardasil vaccine. Tr. at 35, 38-39, 84-85. He further noted
that the rheumatologists who were reviewing the cases of patients who had received the vaccine were
not blinded as to that fact – a discrepancy that he proposed could have introduced bias into the study’s
results. Id. at 36, 65. At the same time, however, Dr. Roseff acknowledged that the Chao Study set forth
many steps that the researchers had taken to help ensure the independence of its results from its
sponsor.26 But in his opinion, a prospective, placebo-controlled, double-blinded study would provide far
more persuasive evidence regarding the safety of the Gardasil vaccination. Id. at 85.
In response, Dr. Rosé argued that the factors cited in the article referenced by Dr. Roseff –
inappropriate comparator to the product being investigated and publication bias – were not applicable to
the Chao Study. For instance, he argued that the requirement that Merck publish the Chao Study’s
findings regardless of outcome lent credibility to its results. Tr. at 110-13, 117. Dr. Rosé also disputed
the validity of Dr. Roseff’s concerns regarding the study’s control group. He noted that it would likely
have been approved by the Food and Drug Administration (“FDA”) in advance, and in fact appeared to
him (from his own review of the Chao Study) that the control group had been carefully selected. Id. at
113-14. Further, Dr. Rosé contested Dr. Roseff’s argument that the only possible persuasive study
24
The database reviewed in the Chao Study identified only four cases of RA within the included vaccinated population, with
a frequency of incidence calculated at 4.6 per 100,000 patients a year. Chao Study at 199 (Table 3). By contrast, there were
39 cases observed among the unvaccinated, with an incidence rate of 7.0 for 100,000 patients a year, yielding an incidence
rate ratio of 0.71. Id.; see also Tr. at 102. Dr. Rosé explained that this ratio, which is below 1.0, suggests that there is no
increase relative to incidence – meaning that the number of vaccinated individuals in the database population experiencing
the subsequent onset of RA was statistically insignificant. Tr. at 102. Moreover, it appeared to Dr. Rosé from review of the
Chao Study that the researchers had double-checked their results to ensure they had captured all cases of RA or JRA, further
supporting the reliability of the findings. Id. at 103.
25
In fact, Dr. Roseff admitted that he lacked the professional background and expertise necessary to intelligently critique the
methodology of the study. See, e.g., Tr. at 84-85.
26
Thus, as part of the post-marketing surveillance of the Gardasil vaccine, Merck was required by the Food and Drug
Administration (“FDA”) to report the results of the study regardless of outcome. Tr. at 62-63; see also Chao Study at 194.
Moreover, the study design and methodology were pre-specified, with the protocol specifically approved by the FDA as well
as an independent board, and all of the data was collected and analyzed at the managed care consortium Kaiser Permanente
rather than by Merck. Tr. at 63-65; see also Chao Study at 202-03. And after its results were written up, it went through a
scientific peer review process in accordance with the protocol set up by the Journal of Internal Medicine (the journal in
which the results of the study were published). Tr. at 65.
15
regarding the safety of the Gardasil vaccine (in terms of autoimmune side effects, such as RA) would be
a prospective, placebo-controlled study using saline for the control group. Id. at 114. Because
autoimmune diseases occur so infrequently, such a study would require depriving thousands of control
group participants of a vaccine that has been demonstrated to be beneficial – an outcome Dr. Rosé
deemed unethical. Id.
D. Other Evidence
Ms. Sullivan submitted four affidavits regarding the onset of her symptoms, including two from
herself. See Pet’r’s Exs. 1, 21. One of Petitioner’s own affidavits indicated that prior to receipt of the
Gardasil vaccinations she was “a healthy and active teenager,” but after receiving the vaccinations she
“began to experience chronic joint pain, joint inflammation, severe systemic body pain, problems
concentrating, memory difficulties, headaches, and changes in [her] personality and temperament.” Id.
at 1. Petitioner’s affidavits were supplemented by a daily self-health log (essentially a journal) in which
she documented her daily symptoms beginning in January of 2010. Id. at 4-52 (Pet’r’s Ex. A). Petitioner
also submitted affidavits from her parents, Sandy and Dennis Sullivan, which included information
regarding their understanding of the onset of her symptoms. Pet’r’s Exs. 23, 24.
Both sides also offered substantial scientific and medical literature. Ms. Sullivan submitted
eleven articles relied upon by Dr. Roseff in formulating an opinion in this case. Respondent cites certain
literature submitted by Petitioner, as well as eleven additional articles, relied upon by her expert in
formulating his opinion or rebutting Petitioner’s expert.
II. Procedural History
Ms. Sullivan filed her Petition on June 28, 2010. Pet. at 1. In it, she specifically alleged that she
suffers from chronic fatigue, severe joint and body pain, joint inflammation, cognitive impairment,
headaches, numbness and tingling in her extremities, and irregular or heavy menstrual cycles which
were all caused in fact by the HPV vaccine. Id. at 3-4. Since the filing of the Petition, Ms. Sullivan has
alleged more specifically (and offered testimony during the hearing to this end) that her vaccination
caused her to develop RA. See, e.g., Pet’r’s Pre-Hr’g Filing (ECF No. 54); Tr. at 17.
On September 27, 2010, Respondent filed her Rule 4(c) report denying that Ms. Sullivan was
entitled to compensation. ECF No. 7. In the ensuing twelve months, Petitioner filed medical records and
both sides submitted expert reports. Thereafter, in September of 2011, the special master previously
responsible for this matter scheduled a hearing for January 24, 2012, to resolve factual issues regarding
the onset of Petitioner’s condition. ECF No. 26.
The onset fact hearing was held as scheduled, with both sides filing pre- and post-hearing
memoranda. That hearing included testimony from Petitioner plus her mother and father. See Ruling
16
Regarding Finding of Fact at 7. After completion of the hearing and the passage of additional time, the
special master issued a Ruling Regarding Finding of Fact on June 30, 2013, determining as follows:
Petitioner’s onset of joint pain following her receipt of HPV vaccinations on June
27, 2007, August 26, 2007, and January 4, 2008, occurred somewhere between
March and April 2008, in the few weeks prior to April 18, 2008, when she first
visited her doctor complaining, inter alia, of pain to her left knee.
Ruling Regarding Finding of Fact at 19.
In light of the results of this ruling, the parties were directed to file supplemental expert reports
incorporating the Court’s factual determinations and evaluating to what extent, if any, their experts’
conclusions were altered as a result. ECF No. 45. Those supplemental reports were submitted in the fall
of 2013. The report submitted by Dr. Roseff indicated that the timing outlined in the Ruling Regarding
Finding of Fact did not change his opinion regarding the role that Gardasil played in Ms. Sullivan’s
ultimate development of rheumatic disease (indicating that there are “precedents in the rheumatic
diseases, or in the literature, that support a delay of several months (as opposed to weeks) between
exposure to an offending antigen, and onset of disease”). Supplemental Expert Report by Dr. Roseff
dated Sept. 16, 2013 (ECF No. 48-1) (Pet’r’s Ex. 28) at 1.
In January of 2014, I was assigned to this matter, and I scheduled an evidentiary hearing for July
10, 2014. ECF No. 53. The parties made additional pre-hearing filings and then participated in the
hearing as scheduled, concluding the proceeding in a single day, and filing no post-trial briefs. The
matter is now ripe for resolution.
III. Applicable Legal Standards
To receive compensation under the Vaccine Program, a petitioner must prove either: (1) that
she suffered a “Table Injury” – i.e., an injury falling within the Vaccine Injury Table – corresponding to
one of the vaccinations in question, or (2) that her illness was actually caused by a vaccine (a category
of claim often generically referred to as a “non-Table Injury”). See §§ 300aa-13(a)(1)(A),11(c)(1); §
300aa-14(a), as amended by 42 C.F.R. § 100.3; § 300aa-11(c)(1)(C)(ii)(I); see also Moberly v. Sec’y of
Health & Human Servs., 592 F.3d 1315, 1321 (Fed. Cir. 2010); Capizzano v. Sec’y of Health & Human
Servs., 440 F.3d 1317, 1320 (Fed. Cir. 2006).27 No Table Injury is alleged in this case, so Ms. Sullivan
must prove causation-in-fact.
27
Decisions of special masters (some of which I reference in this ruling) constitute persuasive but not binding authority.
Hanlon v. Sec’y of Health & Human Servs., 40 Fed. Cl. 625, 630 (1998). By contrast, Federal Circuit decisions are binding
on special masters. Guillory v. Sec’y of Health & Human Servs., 59 Fed. Cl. 121, 124 (2003), aff’d, 104 F. App’x 712 (Fed.
Cir. 2004); see also Spooner v. Sec’y of Health & Human Servs., No. 13-159V, 2014 WL 504728, at *7 n.12 (Fed. Cl. Spec.
Mstr. Jan. 16, 2014).
17
Petitioners bear the burden of demonstrating actual causation by preponderant evidence. Cedillo
v. Sec’y of Health & Human Servs., 592 F.3d 1315, 1321 (Fed. Cir. 2010); § 300aa-13(a)(1). To do so, a
petitioner must provide: “(1) a medical theory causally connecting the vaccination and the injury; (2) a
logical sequence of cause and effect showing that the vaccination was the reason for the injury; and (3)
a showing of a proximate temporal relationship between vaccination and injury.” Althen v. Sec’y of
Health & Human Servs., 418 F.3d 1274, 1278 (Fed. Cir. 2005). The preponderance standard requires a
petitioner to demonstrate that it is “more likely than not” that the vaccine at issue caused her injury.
Moberly, 592 F.3d at 1322 n.2. Proof of medical certainty is not required. Bunting v. Sec’y of Health &
Human Servs., 931 F.2d 867, 873 (Fed. Cir. 1991). In particular, a petitioner must demonstrate that the
vaccine was “not only [the] but-for cause of the injury but also a substantial factor in bringing about the
injury.” Moberly, 592 F.3d at 1321 (quoting Shyface v. Sec’y of Health & Human Servs., 165 F.3d
1344, 1352-53 (Fed. Cir. 1999)); Pafford v. Sec’y of Health & Human Servs., 451 F.3d 1352, 1355
(Fed. Cir. 2006). To determine if the petitioner has carried her burden, I must assess “the record as a
whole” and may not make an entitlement decision in her favor based solely on her own claims
“unsubstantiated by medical records or by medical opinion.” § 300aa-13(a)(1).
Each of the Althen prongs requires a different showing (although the preponderant evidence
standard applies to each). Under Althen prong one, petitioners must provide a “reputable medical
theory,” demonstrating that the vaccine received can cause the type of injury alleged. Pafford, 451 F.3d
at 1355-56 (citations omitted). To satisfy this prong, petitioner’s theory must be based on a “sound and
reliable medical or scientific explanation.” Knudsen v. Sec’y of Health & Human Servs., 35 F.3d 543,
548 (Fed. Cir. 1994). Such a theory must only be “legally probable, not medically or scientifically
certain.” Knudsen, 35 F.3d at 549.
Petitioners may satisfy the first Althen prong without resort to medical literature,
epidemiological studies, demonstration of a specific mechanism, or a generally accepted medical
theory. Andreu v. Sec’y of Health & Human Servs., 569 F.3d 1367, 1378-79 (Fed. Cir. 2009) (citing
Capizzano, 440 F.3d at 1325-26). Special masters, despite their expertise, are not empowered by statute
to conclusively resolve what are essentially thorny scientific and medical questions, and thus scientific
evidence offered to establish Althen prong one is viewed “not through the lens of the laboratorian, but
instead from the vantage point of the Vaccine Act’s preponderant evidence standard.” Andreu, 569 F.3d
at 1380; W.C. v. Sec’y of Health & Human Servs., 704 F.3d 1352, 1356 (Fed. Cir. 2013) (a petitioner
“must do more than demonstrate a ‘plausible’ or ‘possible’ causal link between the vaccination and the
injury; he must prove his case by a preponderance of evidence”) (citations omitted).
Often, however, establishing a sound and reliable medical theory requires that the parties present
expert testimony in support of their claims. Lampe v. Sec’y of Health & Human Servs., 219 F.3d 1357,
1361 (Fed. Cir. 2000). Vaccine Program expert testimony is usually evaluated according to the factors
for analyzing scientific reliability set forth in Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 594-
96 (1993). Cedillo, 617 F.3d at 1339 (citing Terran v. Sec’y of Health & Human Servs., 195 F.3d 1302,
1316 (Fed. Cir. 1999)). “The Daubert factors for analyzing the reliability of testimony are: (1) whether
18
a theory or technique can be (and has been) tested; (2) whether the theory or technique has been
subjected to peer review and publication; (3) whether there is a known or potential rate of error and
whether there are standards for controlling the error; and (4) whether the theory or technique enjoys
general acceptance within a relevant scientific community.” Terran, 195 F.3d at 1316 n.2 (citing
Daubert, 509 U.S. at 592-95).
In other federal judicial fora (such as the district courts), the Daubert factors are employed by
judges (in the performance of their evidentiary gatekeeper roles) to exclude evidence that is unreliable
and/or could confuse a jury. In Vaccine Program cases, by contrast, these factors are used in the
weighing of the reliability of scientific evidence proffered. Davis v. Sec’y of Health & Human Servs., 94
Fed. Cl. 53, 66-67 (2010) (“uniquely in this Circuit, the Daubert factors have been employed also as an
acceptable evidentiary-gauging tool with respect to persuasiveness of expert testimony already
admitted”). The flexible use of the Daubert factors to determine the persuasiveness of expert testimony
has routinely been upheld. See, e.g., Snyder v. Sec’y of Health & Human Servs., 88 Fed. Cl. 706, 742-45
(2009). In this matter (as in numerous other Vaccine Program cases), Daubert has not been employed at
the threshold, to determine what evidence should be admitted, but instead is employed to determine
whether expert testimony offered is reliable and/or persuasive.
Where both sides offer expert testimony, a special master’s decision may be “based on the
credibility of the experts and the relative persuasiveness of their competing theories.” Broekelschen v.
Sec’y of Health & Human Servs., 618 F.3d 1339, 1347 (Fed. Cir. 2010) (citing Lampe, 219 F.3d at
1362). However, nothing requires the acceptance of an expert’s conclusion “connected to existing data
only by the ipse dixit of the expert,” especially if “there is simply too great an analytical gap between
the data and the opinion proffered.” Snyder, 88 Fed. Cl. at 743 (quoting Gen. Elec. Co. v. Joiner, 522
U.S. 146 (1997)). Weighing the relative persuasiveness of competing expert testimony, based on a
particular expert’s credibility, is part of the overall reliability analysis to which special masters must
subject expert testimony in Vaccine Program cases. Moberly, 592 F.3d at 1325-26 (“[a]ssessments as to
the reliability of expert testimony often turn on credibility determinations”); see also Porter v. Sec’y of
Health & Human Servs., 663 F.3d 1242, 1250 (Fed. Cir. 2011) (“this court has unambiguously
explained that special masters are expected to consider the credibility of expert witnesses in evaluating
petitions for compensation under the Vaccine Act”).
The second Althen prong requires proof of a logical sequence of cause and effect, usually
supported by facts derived from a petitioner’s medical records. Althen, 418 F.3d at 1278; Andreu, 569
F.3d at 1375-77; Capizzano, 440 F.3d at 1326; Grant v. Sec’y of Health & Human Servs., 956 F.2d
1144, 1148 (Fed. Cir. 1992). In this regard, the opinions and views of the injured party’s treating
physicians are entitled to some weight. Andreu, 569 F.3d at 1367; Capizzano, 440 F.3d at 1326
(“medical records and medical opinion testimony are favored in vaccine cases, as treating physicians
are likely to be in the best position to determine whether a ‘logical sequence of cause and effect show[s]
that the vaccination was the reason for the injury’”) (quoting Althen, 418 F.3d at 1280). Medical records
are generally viewed as trustworthy evidence, since they are created contemporaneously with the
19
treatment of the patient. Cucuras v. Sec’y of Health & Human Servs., 993 F.2d 1525, 1528 (Fed. Cir.
1993).
However, medical records setting forth a treating physician’s views do not per se bind the
special master to adopt the conclusions of such an individual, even if they must be considered and
carefully evaluated. § 300aa–13(b)(1) (providing that “[a]ny such diagnosis, conclusion, judgment, test
result, report, or summary shall not be binding on the special master or court”); Snyder, 88 Fed. Cl. at
746 n.67 (“there is nothing . . . that mandates that the testimony of a treating physician is sacrosanct—
that it must be accepted in its entirety and cannot be rebutted”). Rather, as with expert testimony offered
to establish a theory of causation, the opinions or diagnoses of treating physicians are only as
trustworthy as the reasonableness of their suppositions or bases. The views of treating physicians should
also be weighed against other, contrary evidence also present in the record – including conflicting
opinions among such individuals. Hibbard v. Sec'y of Health & Human Servs., 100 Fed. Cl. 742, 749
(2011) (not arbitrary or capricious for special master to weigh competing treating physicians’
conclusions against each other), aff'd, 698 F.3d 1355 (Fed. Cir. 2012); Caves v. Sec'y of Dep't of Health
& Human Servs., 100 Fed. Cl. 119, 136 (2011), aff'd, 463 F. App'x 932 (Fed. Cir. 2012); Veryzer v.
Sec'y of Health & Human Servs., No. 06-522V, 2011 WL 1935813, at *17 (Fed. Cl. Spec. Mstr. Apr.
29, 2011).
The third Althen prong requires establishing a “proximate temporal relationship” between the
vaccination and the injury alleged. Althen, 418 F.3d at 1281. That term has been equated to the phrase
“medically-acceptable temporal relationship.” Id. A petitioner must offer “preponderant proof that the
onset of symptoms occurred within a timeframe which, given the medical understanding of the
disorder’s etiology, it is medically acceptable to infer causation.” Bazan v. Sec'y of Health & Human
Servs., 539 F.3d 1347, 1352 (Fed. Cir. 2008). The explanation for what is a medically acceptable
timeframe must also coincide with the theory of how the relevant vaccine can cause an injury (Althen
prong one’s requirement). Id. at 1352; Shapiro v. Sec’y of Health & Human Servs., 101 Fed. Cl. 532,
542 (2011), recons. denied after remand, 105 Fed. Cl. 353 (2012), aff’d mem., 2013 WL 1896173 (Fed.
Cir. 2013); Koehn v. Sec'y of Health & Human Servs., No. 11-355V, 2013 WL 3214877 (Fed. Cl. Spec.
Mstr. May 30, 2013), mot. for review denied (Fed. Cl. Dec. 3, 2013), aff’d, 773 F.3d 1239 (Fed. Cir.
2014).
IV. Analysis
I find, based on the existing record and testimony at hearing, that Ms. Sullivan has not met her
burden of proof with respect to any of the Althen prongs.
A. Althen Prong One
At the outset, I note that Dr. Roseff’s testimony on the “can cause” Althen prong one required
him to go beyond his immediate professional expertise in the field of rheumatology. By his own
20
admission, Dr. Roseff is not an immunologist, and has no demonstrated experience evaluating the
effects of any vaccines in causing injury, in his own medical practice or elsewhere.28 And he did not
ground his proposed theory for how the vaccine could have resulted in RA with any reference to his
own professional experience (such as a study or experiment he participated in, or article he authored,
relevant to such testimony).
I have heard and considered Dr. Roseff’s testimony despite the above. However, I may take into
account an expert’s overall competence in the field upon which he testifies as part of my weighing of
the evidence. See, e.g., Walton v. Sec’y of Health & Human Servs., No. 04-503V, 2007 WL 1467307, at
*17-18 (Fed. Cl. Spec. Mstr. Apr. 30, 2007) (otolaryngologist not well suited to testify about disciplines
other than her own specialty). Dr. Roseff’s lack of expertise on the topic of the capacity of vaccines
such as Gardasil to cause injury leads me to give his testimony less weight than I might give the
testimony of an expert with greater demonstrated experience in the subject matter. See, e.g., King v.
Sec’y of Health & Human Servs., No. 03-584V, 2010 WL 892296, at *78-79 (Fed. Cl. Spec. Mstr. Mar.
12, 2010) (petitioner’s expert far less qualified to offer opinion on general causation issues pertaining to
autism than specific issues pertaining to the petitioner’s actual medical history, given the nature of the
expert’s qualifications).
Putting that aside, I find that the substance of Dr. Roseff’s opinion was unpersuasive and not
supported by reliable evidence. He referenced no studies in which the Gardasil vaccine, any other HPV
vaccination, or even a single strain of HPV wild virus, has caused RA (or any disease comparable to
RA). Instead, he attempted to develop his theory by way of analogy to studies involving arthritic
conditions arising after exposure to different viruses (such as the rubella virus or human parvovirus), or
studies (like Toplak) observing autoimmune responses after vaccination. See supra Section IB1
(outlining Dr. Roseff’s testimony). In effect, Dr. Roseff argues, (a) other viruses can produce arthritic
conditions, and (b) vaccines can cause autoimmunity, therefore (c) it is reasonable to conclude by
means of extrapolation from such evidence that the Gardasil vaccine could cause RA.
Although there is logic to such reasoning, I do not find that the evidence upon which it is based
is sufficiently reliable to bulwark Petitioner’s causation theory with the needed evidentiary ballast. As
another special master noted in considering the Althen prong one analysis, “[t]he weight to be given an
expert’s opinion is based in part on the size of the gap between the science and the opinion proffered.”
Isaac v. Sec’y of Health & Human Servs., No. 08-601V, 2012 WL 3609993, at *17 (Fed. Cl. Spec.
Mstr. July 30, 2012), mot. for review den’d, 108 Fed. Cl. 743 (2013), aff’d, 540 Fed. App’x 999 (Fed.
Cir. 2013) (citing Cedillo, 617 at 1339). Here, that gap is too great. The fit between the literature Dr.
Roseff cites and the theory he proposes is poor, based on inapposite comparisons involving different
28
Admittedly, Dr. Rosé similarly lacked such qualifications. Because Petitioner bears the burden of proof, however, it is
especially important that an expert testifying on her behalf possess sufficient credentials and expertise upon which to base
his opinion. Daubert v. Merrell Dow Pharm., Inc., 43 F.3d 1311, 1317 (9th Cir. 1995) (indicating that “[o]ne very significant
fact to consider is whether the experts are proposing to testify about matters growing naturally and directly out of research
they have conducted independent of the litigation, or whether they have developed their opinions expressly for purposes of
testifying”).
21
vaccines and different illnesses or untestable premises. Dr. Roseff assumes, without demonstrating, that
other viruses that have been scientifically observed to produce arthritic conditions are comparable not
only to the HPV wild virus but to the Gardasil vaccine itself (which contains four HPV strains). He also
conclusorily reasons that a generalized reaction to one kind of vaccine is comparable to the type of
reaction that one would expect to see from an entirely different vaccine. And in some cases,
(specifically, with regard to the Classen article purporting to identify a causal link between other
vaccines and IDDM), he relies on literature directly relating to theories that have been routinely rejected
when offered to establish causation involving the tested vaccine. See, e.g., Meyers v. Sec’y of Health &
Human Servs., No. 04-1771V, 2006 WL 1593947, at *5 (Fed. Cl. Spec. Mstr. May 22, 2006) (“[t]his
court has previously considered and discredited the theories advanced by Dr. Classen”) (discussing
Baker v. Sec’y of Health & Human Servs., No. 99-653V, 2003 WL 22416622, at *33 (Fed. Cl. Spec.
Mstr. May 22, 2006)).29
Ms. Sullivan’s overall theory regarding how the Gardasil vaccination could cause RA fails to
hold up when analyzed under the Daubert framework used to evaluate the reliability of expert
testimony in Vaccine Program cases. It has not been tested, subjected to peer review, or shown to be
generally accepted in the relevant medical community. Terran, 195 F.3d at 1316 n.2 (citing Daubert,
509 U.S. at 592-95). While some of the individual pieces of literature cited by Petitioner may meet
those criteria, that does not mean that Petitioner’s overall theory (in support of which they are cited) is
similarly reliable. In addition, (as noted above) nothing about Petitioner’s theory is derived from Dr.
Roseff’s personal expertise, further diminishing its reliability (since he is the theory’s author). And in
many instances Dr. Roseff (who appeared to me an honest witness) undercut his opinion with
admissions that Ms. Sullivan’s illness may just as likely be of unknown origin. See, e.g., Roseff Report
at 3 (indicating that “we cannot exclude the possibility that her development of disease was a random
event,” and admitting that “how [molecular mimicry] could result in our patient’s CCP positivity is
unclear”).
The same considerations apply to Dr. Roseff’s proposed molecular mimicry mechanism. A
petitioner may successfully establish causation without proving the mechanism of injury. Knudsen, 35
F.3d at 548-49 (citations omitted). And there are Vaccine Program decisions in which molecular
mimicry has been found to be a scientifically acceptable explanation sufficient to meet the
preponderant, “more likely than not” evidentiary standard. See, e.g., Tompkins v. Sec’y of Health &
Human Servs., No. 10-261V, 2013 WL 3498652, at *22 (Fed. Cl. Spec. Mstr. June 21, 2013) (in the
specific context of establishing causation of Guillain-Barré syndrome after vaccination, “[t]he
molecular mimicry theory is the one most widely accepted for the agents most frequently accepted as
causal”), mot. for review denied, 117 Fed. Cl. 713 (2014); but see Wirt v. Sec’y of Health & Human
Servs., No. 11-118V, 2014 WL 1911421, at *9-10 (Fed. Cl. Spec. Mstr. Apr. 18, 2014) (petitioner failed
29
The special master’s decision in Baker specifically discussed the weaknesses of the Classen article cited by Dr. Roseff
herein. See Baker v. Sec’y of Health & Human Servs., No. 99-653V, 2003 WL 22416622, at *11 (Fed. Cl. Spec. Mstr. May
22, 2006).
22
to satisfy Althen prong one in case in which expert proposed molecular mimicry as a mechanism by
which the HPV vaccine could cause RA).
A petitioner cannot, however, simply intone the phrase “molecular mimicry” and thereby be
deemed to have satisfied her Althen prong one burden. Hennessey v. Sec'y of Health & Human Servs.,
91 Fed. Cl. 126, 134-35 (2010) (noting expert’s overly broad application of the molecular mimicry
theory made it meaningless). Here, and for the same reasons stated above, Petitioner has not sufficiently
connected this aspect of her causation theory to the Gardasil vaccine or her RA to be legally
persuasive,30 because she has not offered sufficient reliable proof supporting the concept in this
particular factual context.
Petitioner offers a few alternative mechanisms, but they are incompletely sketched out at best.
Dr. Roseff proposed that aluminum contained in the Gardasil vaccine (as an adjuvant) could be
responsible for stimulating an individual’s immune system in a destructive manner (and he ties this
argument in with his criticisms of a Merck study that used a placebo containing aluminum). See, e.g.,
Tr. at 25. Yet he admitted that this theory was highly speculative and based on very little scientific
literature, if any. Id. at 60. Indeed, it appears from my review of the record that the only piece of filed
literature even mentioning aluminum as a vaccine adjuvant, and the potentially negative impact it could
have in causing a reactive disease or condition, is the Judicial Watch article – an advocacy piece with
little scientific reliability, and which is conclusory in making this assertion as well. Joiner, 522 U.S. at
146 (“[t]rained experts commonly extrapolate from existing data[,] but nothing in either Daubert or the
Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to
existing data only by the ipse dixit of the expert.”).
Overall, Dr. Roseff has (by his own admission) done no more than opine that it is intellectually
conceivable that administration of the Gardasil vaccine could result in the onset and development of
RA. But this is insufficient to meet Petitioner’s burden of proof, which (while not requiring scientific
certainty) does obligate Petitioner to provide a “legally probable” explanation. Moberly, 592 F.3d at
1322 (quoting Knudsen, 35 F.3d at 548-49). Dr. Roseff’s opinion, and the evidence offered in support
of it, do not rise to that level of preponderant proof. See also Wirt, 2014 WL 1911421, at *9
(determining that “[t]here are simply too many unknowns, too many gaps in the analytical process of
the theory, [] to conclude that Petitioner has proven a medical theory causally connecting the [HPV]
vaccination and the injury [RA].”) (citation omitted).
30
There are also deficiencies with Dr. Roseff’s explanation of homology within the context of the molecular mimicry
process he proposed. Dr. Roseff relies on Kanduc for the proposition that one HPV strain of the four contained in the
Gardasil vaccine shares peptide sequences with a variety of human proteomes, and therefore the cross-reactivity that would
result in RA could conceivably occur. See Tr. at 19-20. But as Respondent pointed out, that HPV strain shares numerous
sequences with the human body, perhaps too many to reach the conclusion that molecular mimicry would in fact occur
resulting in RA. See Id. at 107. While the law does not require Petitioner to “prove” homology in a Program case, I take note
of the fact that Respondent’s challenges to this part of Petitioner’s theory were not rebutted. See, e.g., Id. at 80; Rosé Report
at 4; Roseff Report at 3.
23
Also relevant to my analysis is the fact that Respondent offered credible and persuasive
epidemiologic evidence – the Chao Study – undermining Petitioner’s theory. Unquestionably, a
petitioner need not offer epidemiologic proof to establish a reasonable and scientifically-reliable theory
under Althen prong one. Capizzano, 440 F.3d at 1325. However, I may properly weigh such evidence
against Petitioner’s proof in evaluating whether she has carried her overall burden as to this first Althen
prong. Koehn, 2013 WL 3214877, at *25 (“[t]he Federal Circuit has endorsed consideration of
epidemiological studies as one factor in the special master’s analysis”);31 see also C.K. v. Sec’y of
Health & Human Servs., 113 Fed. Cl. 757, 770 (2013) (a special master may evaluate contradictory
evidence offered by Respondent).
The Chao Study is directly relevant to Ms. Sullivan’s claim, and is persuasive evidence
contradicting her causation theory. The same study has been determined in other Vaccine Program
cases (in which it was invoked by Respondent) to be a valid epidemiologic study. See, e.g., C.K., 113
Fed. Cl. at 770; Godfrey v. Sec’y of Health & Human Servs., No. 10-565V, 2014 WL 3058353, at *17
(Fed. Cl. Spec. Mstr. June 11, 2014); Harris v. Sec’y of Health & Human Servs., No. 10-322V, 2014
WL 3159377, at *13-14 (Fed. Cl. Spec. Mstr. June 10, 2014), mot. for review denied, __ Fed. Cl. __
(Sept. 23, 2014). Although the Chao Study is not proof positive (from a legal standpoint) that Gardasil
does not “cause” RA (and indeed as a special master I am not empowered to make such a scientific
determination), the Chao Study undermines Petitioner’s case that “more likely than not” the vaccine
could have that effect.
Ms. Sullivan endeavored to call into question the study’s credibility by impugning its source.
The argument that a given study’s authorship might impinge on the reliability of its findings is
reasonable. See, e.g., Harris, 2014 WL 3159377, at *13 (“the source for Dr. Chao’s funding opens a
potential (if ultimately unresolvable) argument that her conclusions are not valid”). But the Supreme
Court in Daubert, as well as the cases following it, has provided analytical tools for evaluating the
existence of alleged errors in a study’s methodology or analysis that, if present, would illustrate bias
better than the unsubstantiated argument that an interested sponsor automatically casts doubt on the
honesty of the study’s findings. Daubert, 509 U.S. at 592-95.
Therefore, in order to put flesh on the bones of her argument that the Chao Study was biased,
Petitioner needed to point out specific elements of the study demonstrating its unreliability. She failed
to do so. The Chao Study simply looked at what actually happened to a large group of individuals who
received the Gardasil vaccine versus those who did not, performing some subsequent statistical analyses
to evaluate the significance of the incidence of autoimmune diseases after vaccination. An argument
could be made that a larger sample size was needed to produce more reliable results, but Petitioner does
31
On appeal, the Federal Circuit upheld the special master’s finding in Koehn (which specifically considered the Chao
Study) but found that he had misapplied the law in determining that the petitioner had not met her burden under the first
Althen prong. Koehn, 773 F.3d at 1244 n.1. In so doing, however, the Federal Circuit did not call into question the special
master’s earlier statement on the evidentiary significance of epidemiologic evidence (when offered) in evaluating this prong
of the Althen test, nor did it find that the manner in which he had evaluated the weight to be afforded the Chao Study was an
element of the error it ruled he had otherwise committed.
24
not make that assertion. Hart v. Sec’y of Dep't of Health & Human Servs., 60 Fed. Cl. 598, 608 (2004)
(quoting In re Norplant Contraceptive Prods. Liab. Litig., 215 F. Supp. 2d 795, 830 (E.D. Tex. 2002))
(“epidemiological data that is not statistically significant cannot provide a scientific basis for an opinion
of causation”). And the low incidence rate ratio for vaccinated individuals who subsequently developed
RA (0.71) supports (at least from a statistical standpoint) the Chao Study’s conclusion that Gardasil is
highly unlikely to cause RA. See Daubert, 43 F.3d at 1321 (citing DeLuca v. Merrell Dow Pharm., Inc.,
911 F.2d 941, 958 (3rd Cir. 1990)) (“[f]or an epidemiological study to show causation under a
preponderance standard, ‘the relative risk of ... [the defect or injury] arising from the epidemiological
data . . . will, at a minimum, have to exceed ‘2’”).32
Those criticisms of the Chao Study Dr. Roseff did raise, such as the adequacy of its
methodological control group, were not backed up by reliable evidence. Dr. Roseff’s complaint that a
truly independent study of vaccine safety would have created a blinded control group to receive a “true
placebo” (saline) amounts to the assertion that only a randomized control study is trustworthy, as
opposed to the observational study set forth in the Chao Study. It may be true that a more scientifically
certain (and therefore persuasive) study is conceivable, and it is also the case that observational studies
have their own inherent flaws. Dwyer v. Sec’y of Health & Human Servs., No. 03-1202V, 2010 WL
892250, at *64 (Fed. Cl. Spec. Mstr. Mar. 12, 2010) (“[e]very observational epidemiological study has
some weaknesses because such studies examine the world as it is”). But (operating from the well-worn
maxim that “the perfect is the enemy of the good”)33 the possibility of a better study is not an effective
critique of an existing, otherwise valid study. See also Harris, 2014 WL 3159377, at *13 (“a perfect
scientific study is not required by the relevant legal standards”).
The only other proof that Ms. Sullivan offers to support her causation theories are VAERS
reports and data from the Gardasil package insert – none of which are particularly persuasive evidence.
Because of their passive nature and unverified claims, VAERS reports are too anecdotal and
unscientific to have much probative value in establishing a causation theory.34 Statements contained in
32
The Chao Study’s findings regarding the safety of the Gardasil vaccination are bolstered by the findings from Verstraeten,
which compiled data from smaller clinical trials in an effort to identify a correlation between receipt of the HPV vaccination
and the subsequent development of autoimmune disease when such data was aggregated (but finding no such association).
See Verstraeten at 6630.
33
This phase is attributed to Voltaire and is literally translated as “the best is the enemy of the good.” Bartlett’s Familiar
Quotations 343 (125th ed. 1980).
34
As another special master has commented in connection with the evidentiary value of VAERS reports,
VAERS is a stocked pond. It only contains reports (many of which are unverified or incomplete) of adverse events
after vaccinations. VAERS contains no reports or data about the relative rate of these same events in individuals
who have not been vaccinated. Thus, the number of specific adverse events, such as GBS, reported after any
vaccine, is meaningless without information about the background rate of that adverse event and information about
the number of vaccines administered.
Tompkins v. Sec’y of Health and Human Servs., No. 10-261V, 2013 WL 3498652, at *16 (Fed. Cl. Spec. Mstr. June 21,
2013), review denied sub nom., Tompkins v. United States, 117 Fed. Cl. 713 (2014).
25
vaccine package inserts do not constitute reliable proof of causation, and cannot be deemed admissions
that the vaccines in question have the capacity to harm a particular petitioner in a specific manner. See
Werderitsh v. Sec'y of Health & Human Servs., No. 99–319V, 2005 WL 3320041, at *8 (Fed. Cl. Spec.
Mstr. Nov. 10, 2005) (quoting 21 C.F.R. § 600.80(l) as saying “[a] report or information submitted by a
licensed manufacturer . . . does not necessarily reflect a conclusion by the licensed manufacturer or
FDA that the report or information constitutes an admission that the biological product caused or
contributed to an adverse effect”).35 The record is therefore insufficient to establish by a preponderance
of the evidence Petitioner’s theory that the Gardasil vaccine can cause RA.
B. Althen Prong Two
There is a notable lack of record support for the second Althen prong – that the Gardasil vaccine
“did” cause Ms. Sullivan’s injuries. Petitioner’s contemporaneous medical records provide no evidence
– such as a treating physician’s statement, or a test result – supporting the view that the Gardasil series
she had received was connected to her RA. There is only the fact that Ms. Sullivan received her last
dose of the vaccine in January 2008, and then began to complain of knee-related pain no earlier than
late March to early April of 2008 (as the Ruling Regarding Finding of Fact establishes).36 Mere
temporal association, without more, is insufficient to establish causation. See Moberly, 592 F.3d at
1323. Here, however, that is all the Petitioner can offer.
35
In any event, as Dr. Rosé noted, the numbers reported on the Gardasil package insert which Dr. Roseff characterized as
evidencing a “disturbing trend” merely represented raw data gathered through a self-reporting mechanism after the trial, and
do not suggest anything regarding causality - a fact which Dr. Rosé illustrated by referencing other numbers in the Gardasil
package insert regarding trials of the vaccine. Based on Dr. Roseff’s same logic, such data would seem to suggest that there
were more adverse events following receipt of a saline placebo than there were following receipt of the Gardasil vaccination.
Rosé Report at 3; See also Roseff Report at 3.
36
As noted above, the onset issue was the subject of a prior evidentiary hearing conducted by the special master previously
assigned to the case, and I have incorporated her ruling in my decision. This fact determination does not serve as “law of the
case” (a doctrine that applies more to prior legal determinations in a proceeding) that I must necessarily follow. See
generally Banks v. United States, 741 F.3d 1268, 1276 (Fed. Cir. 2014) (indicating that law-of-the-case doctrine “posits that
when a court decides upon a rule of law, that decision should continue to govern the same issues in subsequent stages in the
same case”) (quoting Christianson v. Colt Indus. Operating Corp., 486 U.S. 800, 815-16 (1988)). Indeed, I am no more
formally bound by the onset proceeding than I would be bound by my determinations in a different case, the determinations
of a special master in another case, or even a Court of Federal Claims decision. Hanlon, 40 Fed. Cl. at 630.
However, neither party asked me to reconsider the Ruling Regarding Finding of Fact prior to the entitlement
hearing in this case, nor did Petitioner attempt to argue that newly-discovered or additional proof relevant to onset favored a
different outcome. Moreover, Dr. Roseff submitted a supplemental opinion in which he applied the findings from the Ruling
Regarding Finding of Fact but still determined that his original opinion was valid. Supplemental Expert Report by Dr. Roseff
dated Sept. 16, 2013 (ECF No. 48-1) (Pet’r’s Ex. 28) at 1. Given the above (plus the fact that the Ruling Regarding Finding
of Fact is comprehensive and was the result of a proceeding in which both sides had full opportunities to present evidence),
in the exercise of my discretion I find that it is appropriate to adopt the Ruling Regarding Finding of Fact herein. See Pacific
Gas & Elec. Co. v. United States, 114 Fed. Cl. 146, 149 (2013) (when a successor judge is transferred a case in which a prior
order has been rendered, the successor judge “should not overrule the earlier judge’s order or judgment merely because the
later judge might have decided matters differently,” but should exercise his discretion in determining if circumstances
warrant reopening the previously-determined issue) (quoting United States v. O’Keefe, 128 F.3d 885, 891 (5th Cir. 1997)).
26
During the hearing, both experts acknowledged this lack of record support linking Ms.
Sullivan’s vaccination to her RA. Aside from a single rheumatologist (who appears to have been merely
recounting Mr. Sullivan’s view that her RA was connected to the Gardasil vaccine), none of Petitioner’s
treating physicians ever identified Gardasil as a possible cause of her injury. Tr. at 74-75, 138. Dr. Rosé
indicated that (based on his own ample experience diagnosing RA) he did not himself see anything in
her treatment record that would support such a connection. Id. at 138. Dr. Roseff’s expertise in
rheumatology would have made him well-qualified to point out contrary evidence from the record (such
as a test result that suggested a connection between the Gardasil vaccine and the development of Ms.
Sullivan’s symptoms, even if overlooked by a treating physician), but he admitted he could identify no
such evidence either. Id. at 75. Petitioner has thus failed to offer preponderant evidence in support of the
second Althen prong.
C. Althen Prong Three
As I explained above, Ms. Sullivan did not provide a reliable theory for how the Gardasil
vaccine could cause RA. Accordingly, she cannot satisfy the third Althen prong either, since the
adequacy of the proposed timeframe in which the Gardasil vaccine could have caused the Petitioner’s
injury must relate to the medical theory for how this would occur in the first place. Bazan, 539 F.3d at
1352; Shapiro, 101 Fed. Cl. at 542. But even if I had found that Petitioner had provided preponderant
evidence in satisfaction of Althen prong one, I would still find that this third prong has not been
similarly satisfied.
Petitioner inconsistently argued what was a medically acceptable timeframe for onset in this
case – sometimes appearing to measure onset from the last in the series of three Gardasil vaccinations
she received, while other times suggesting that the series in total had a “cumulative effect,”37
compounding over time from the date of the first or second vaccine in the series. Neither argument was
persuasive.
Measuring from the last Gardasil vaccination Ms. Sullivan received (January 4, 2008), the onset
of her RA symptoms (beginning “somewhere between March and April 2008,” as determined at the
onset fact hearing) occurred ten to twelve weeks thereafter. Dr. Rosé proposed, however, that a
reasonable timeframe in which any vaccine might conceivably cause some kind of arthritic condition
37
During the hearing, Petitioner at times displayed confusion as to what timeframe she intended to try to prove in which the
vaccination resulted in the onset of her RA. Her pre-hearing filings characterized it as a “three month period of time between
the onset of symptoms and the vaccine” (Pet’r’s Pre-Hr’g Filing (ECF No. 54)), which is consistent with the Ruling
Regarding Finding of Fact. At certain points throughout his testimony, however, Dr. Roseff referred to a two-month latency
period as an appropriate temporal relationship between vaccination and onset of Petitioner’s symptoms. See, e.g., Tr. at 28,
33. Dr. Roseff further muddied Petitioner’s argument by indicating that it was not clear to him whether it was the first,
second, or third vaccination (or even a “cumulative effect” of all three vaccinations) that ultimately caused Petitioner’s RA.
Tr. at 54, 73.
27
would be no more than two to four weeks.38 Rosé Report at 6. Given the above, three months appears
too long a timeframe in which onset could have occurred.
Ms. Sullivan attempted to substantiate three months as medically acceptable, pointing to
literature showing purportedly analogous diseases that can be latent for that same period or longer. For
instance, Dr. Roseff analogized RA to an inflammatory arthritic condition caused by Lyme disease (an
insect-borne bacterial infection). Before being first recognized as a separate entity, Lyme arthritis was
initially confused with early RA, so facially this comparison is not inapt. See Allen C. Steere & Lisa,
Glickstein, Elucidation of Lyme Arthritis, 4 Nature Reviews Immunology 143 (2004) (Pet’r’s Ex. 31 at
2). But Dr. Rosé persuasively argued that the timeframe for the development of Lyme arthritis was
more logically attributed to persistence of inflammation (despite the otherwise successful treatment of
the initial infection) rather than latent onset.39 Tr. at 23-24, 131-34. For his part, Dr. Roseff admitted
that he could not support the Lyme arthritis analogy with any comparable research involving RA. Id. at
24. Indeed, it appears to some extent that Dr. Roseff chose the Lyme arthritis analogy mainly because it
provided an instance of “a long period between infection and development of inflammatory arthritis,”
rather than for its demonstrable scientific similarity. Id. at 58. Further, Petitioner cited no evidence
suggesting that any component of the vaccine had ever been found in an RA patient’s synovial capsules
(in the same manner that the Lyme disease bacteria are found there).
The same is true for the other purportedly analogous cases (as referenced in the Toplak or
Agmon-Levin articles) involving latency periods cited by Dr. Roseff. The main common ground those
articles generally have with present circumstances is the fact that they display a delayed autoimmune
response following receipt of a vaccine. Otherwise, they all involve different vaccines and different
diseases, and thus cannot be assumed to be scientifically comparable to a sufficient degree to support
the theory that Ms. Sullivan could have experienced delayed onset of RA-related symptoms months
after any of the series of Gardasil vaccines that she received. Indeed, these pieces of literature
themselves are far more limited in the scope of their conclusions (as Dr. Roseff himself specifically
admitted with respect to the Toplak article (see, e.g., Tr. at 47-48)). The Toplak article concludes with
the specific observation that the flu vaccine did not “increase the percentage of positive autoantibodies
in the general healthy adult population,” even if some autoantibody increases were seen at certain
temporal points after vaccine administration – increases having “no clear clinical significance.” Toplak
at 138; Tr. at 30.
38
Other Vaccine Program cases involving RA or similar arthritic illnesses have found this same general timeframe to be
reasonable (although I do not adopt it as the “correct” one, or the only one a petitioner could ever prove). Capizzano v. Sec'y
of the Dep't of Health & Human Servs., No. 00-759V, 2006 WL 3419789, at *11 (Fed. Cl. Nov. 8, 2006); Doe/28 v. Sec'y of
Dep't of Health & Human Servs., 2009 WL 229665, at *20 (Fed. Cl. Jan. 21, 2009) (finding that petitioner has shown a
medically appropriate time interval between receipt of the hepatitis B vaccine and her development of nonspecific arthritis
where approximately one month elapsed between petitioner's third hepatitis B vaccination and the onset of her symptoms,
and “[t]he interval was so strikingly appropriate that four of her treaters ascribed her condition to the vaccination”).
39
See Allen C. Steere & Lisa Glickstein, Elucidation of Lyme Arthritis, 4 Nature Reviews Immunology 143 (2004) (Pet’r’s
Ex. 31) (indicating that “10% of patients with Lyme arthritis develop persistent synovitis, which lasts for months or even
several years after the apparent eradication of the spirochete from the joint with antibiotic therapy.”).
28
Petitioner was also unable to establish preponderant evidence that the series of three Gardasil
vaccinations she received could produce in concert the onset of her RA symptoms in the early spring of
2008. In an attempt to do so, she proposed (through Dr. Roseff’s testimony) something along the lines
of a “cumulative effect” or challenge-rechallenge theory – that Ms. Sullivan’s successive vaccinations
(whether due to build-up of aluminum in Ms. Sullivan’s body, or simply from her system being primed
immunologically) culminated in her initial knee symptoms and ultimate development of RA. But this
theory was thinly substantiated and inadequately explained. The record shows absolutely no evidence
that Ms. Sullivan experienced any reaction after receipt of the prior two Gardasil vaccines, as Dr. Rosé
noted would be expected to have occurred if her immune system was being challenged after each
successive vaccination. Tr. at 109; see also Hall v. Sec’y of Health & Human Servs., No. 02-1052V,
2007 WL 3120284, at *7 (Fed. Cl. Sept. 12, 2007) (a petitioner who suffered from a shoulder injury
was entitled to compensation based on a challenge-rechallenge theory when petitioner received two
doses of the hepatitis B vaccine and after each dose, she experienced problems in her shoulder within
two weeks of the vaccination; the amount of time between the first vaccination and the start of her
problems was more than the amount of time between the second vaccination and the start of problems;
and the reaction to the second vaccination was also more severe). And as discussed above, Dr. Roseff
himself did little more than speculate that it was possible that the Gardasil series Ms. Sullivan received
could have built up in her body in some manner to result in her initial symptoms, supporting this
speculation with unscientific literature like the Judicial Watch article rather than reliable scientific or
medical proof. Petitioner thus did not establish that it was more likely than not that the total series of
Gardasil vaccinations could cumulatively produce her initial symptoms two to three months after
receipt of the final vaccination.
Conclusion
I have great sympathy for the suffering Ms. Sullivan experienced after receiving the Gardasil
vaccinations. However, based on the records filed and testimony at hearing, I cannot conclude that she
is entitled to an award of compensation in this case. The Vaccine Act permits me to award
compensation only if a Petitioner alleging a “non-Table Injury” can show by medical records or
competent medical opinion that the injury was more likely than not vaccine-caused. Here, there is
insufficient evidence to support an award of compensation, leaving me no choice but to hereby DENY
this claim.
In the absence of a timely-filed motion for review (see Appendix B to the Rules of the Court),
the Clerk shall enter judgment in accord with this decision.
IT IS SO ORDERED.
/s/Brian H. Corcoran
Brian H. Corcoran
Special Master
29