Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 39
2
Approach
Charged with assessing the epidemiologic, clinical, and biological ev-
idence regarding the causal relationship between specific vaccines and
specific adverse events, the committee drew upon previous reports by com-
mittees of the Institute of Medicine (IOM, 1991, 1994, 2001a,b, 2002a,b,
2003a,b, 2004a,b), other vaccine safety researchers (Halsey, 2002; Loke
et al., 2008; WHO, 2001), general epidemiologic principles (Hill, 1965),
and other systematic reviews in clinical medicine and public health (Liberati
et al., 2009; Owens et al., 2010; Schunemann et al., 2011; Stroup et al.,
2000; USPSTF, 2008). The committee adopted, with one exception,1 the
wording for the categories of causal conclusions used by the Institute of
Medicine (IOM) committees in the past. The categories used previously
were considered appropriate and the benefits of consistency were deemed
compelling enough to extend the categories to this report.
Two streams of evidence from the peer-reviewed literature support the
committee’s causality conclusions: (1) epidemiologic evidence derived from
studies of populations (most often based on observational designs but ran-
domized trials when available), and (2) clinical and biological (mechanistic)
evidence derived primarily from studies in animals and individual humans
or small groups. Some studies provide evidence relevant to both epidemio-
logic and mechanistic questions. Drawing from both lines of evidence to
support causal inference is well established in the literature. When con-
fronted with epidemiologic and mechanistic evidence suggesting—however
1 Asdescribed in a subsequent section, previous IOM committees described the strongest evi-
dence as establishing a causal relationship; this committee uses the term convincingly supports.
39
OCR for page 40
40 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
strongly or however weakly—that a vaccine is associated with an adverse
event, one asks, “Does this make sense given what is known and generally
accepted about the biological response to the natural infection, to the vac-
cine, and what is known about the pathophysiology of the adverse health
outcome?”
LITERATURE SEARCHING
As described in Chapter 1, the committee was tasked to assess the rela-
tionship between a specific adverse health outcome and a specific vaccine.
A professional medical librarian conducted three waves of comprehensive
literature searches of the published, peer-reviewed biomedical literature
using MEDLINE (1950–present); EMBASE (1980–present); BIOSIS (1969–
2005); Web of Science, consisting of the Science Citation Index (1900–pres-
ent) and the Social Science Citation Index (1956–present); and search terms
specific to each vaccine–adverse event relationship under study. Appendix
C contains the search strategies used. The first wave of searches included
the earliest date of the database to the date of the first search. Follow-up
searches were conducted in August 2010 and late December 2010 to ensure
that articles published after the initial search were not missed. On occasion,
specialized searches were conducted to supplement the general searches.
Also, review of the reference list of an article sometimes revealed studies
not captured by the general search. These studies were retrieved.
Titles and abstracts, where available, were reviewed to screen out
articles that did not address one of the potential vaccine adverse events
to be reviewed or that were not primary research articles. See Figure 2-1.
For example, the committee did not assess review articles. The committee
restricted its review to those vaccines used in the United States, even if the
study was conducted outside of the United States, with a few exceptions
that will be discussed in the vaccine-specific chapters that follow. Articles
were retrieved and reviewed again for relevance to the committee charge.
Articles written in languages other than English were translated using
Google Translate or a professional translation service. The committee did
not include in its reviews data presented only in abstract form or in oth-
erwise unpublished formats, with one exception described in Chapter 9,
“Human Papillomavirus Vaccine.” An individual report from the Vaccine
Adverse Event Reporting System was reviewed only if it had been described
in a peer-reviewed research study and the committee wanted additional in-
formation. Decisions from the Vaccine Injury Compensation Program were
not reviewed, because they are not published in the peer-reviewed medical
literature. The committee did not review the conclusions contained in ear-
lier IOM reports. The committee reviewed the data and made conclusions
independently.
OCR for page 41
ORIGINAL RESEARCH IN HUMANS OTHER LITERATURE
Controlled Uncontrolled Reviews of
Animal and
Observational Studies and Clinical Studies and Case Reports Consequences of the
In Vitro Studies
Randomized Trials Surveillance Studies Natural Infection
Included studies Excluded reports Included reports
Excluded studies Excluded studies with case descriptions that did not meet with case descriptions
with very serious with very serious that met minimum minimum attribution that met minimum
limitations from limitations from attribution elements elements for attribution elements
the Epidemiologic the Epidemiologic in the Mechanistic Mechanistic in the Mechanistic
Evidence Evidence Evidence Evidence Evidence
Studies Included in the Studies Included in the
Weight of Epidemiologic Evidence Weight of Mechanistic Evidence
High Moderate
Strong Intermediate Weak Lacking
Limited
(increased or (increased or
null/decreased) null/decreased)
CAUSALITY CONCLUSION
41
FIGURE 2-1 Epidemiologic and mechanistic evidence reviewed by the committee.
OCR for page 42
42 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
The committee’s bibliographic retrieval was posted on the project web-
site with a request for public comment regarding missing articles.2 The
committee received one submission, which was reviewed. The bibliography
was separated into two sections. Section I contained those articles on which
the committee focused its initial review. Section II contained those citations
for articles that did not meet the committee’s criteria (i.e., original research,
vaccine used in the United States, adverse event within the committee’s
scope, animal or in vitro studies of relevance).
WEIGHT OF EVIDENCE
The committee made three assessments for each relationship reviewed.
The first assessment applies to the weight of evidence from the epide-
miologic literature; the second applies to the weight of evidence from the
biological and clinical (mechanistic) literature. The third assessment is
the committee’s conclusion about causality. In assessing the weights of
evidence, each individual article (or findings within an article if more than
one outcome or vaccine was studied) was evaluated for its strengths and
weaknesses. The committee then synthesized the body of evidence of each
type (epidemiologic or mechanistic) and assigned a “weight of evidence”
for each. These weights of evidence are meant to summarize the assessment
of the quality and quantity of evidence. The committee then reviewed the
two weight-of-evidence assessments in order to make a conclusion about
the causal relationship. The committee’s approach to each of these three
assessments will be discussed in the following sections.
Epidemiologic Evidence
Experimental studies (trials) are generally considered more rigorous
than observational studies; controlled studies are generally considered more
rigorous than uncontrolled studies. A brief description of major study
designs and methodological considerations can be found in Appendix A.
Surveillance studies were reviewed, but the absence of a control group lim-
ited their contribution to the weight of epidemiologic evidence; studies that
included individual case descriptions were reviewed for their contribution
to the evaluation of mechanistic evidence (discussed in subsequent sections).
Small clinical studies that were not controlled for vaccine administration
were generally reviewed for contributions to the mechanistic weight of
evidence.
2 See http://www.iom.edu/~/media/Files/Bibliography.pdf.
OCR for page 43
43
APPROACH
Evaluation of Individual Studies
Each epidemiologic study was evaluated for its methodological limita-
tions (e.g., flawed measurement of either vaccine administration or adverse
event, failure to adequately control confounding variables, incomplete or
inadequate follow-up, failure to develop and apply appropriate eligibility
criteria) and for the precision of the reported results (e.g., the width of the
95% confidence interval around an effect estimate, which also reflects the
statistical power to detect a significantly increased risk of an adverse event).
Studies that were deemed to be very seriously flawed did not contribute to
the weight of evidence; they are identified in the text for completeness but
are not discussed in depth.
It is important to note that a specific study could be well designed and
well conducted but also have very serious limitations for the purposes of
this committee’s analysis. A specific study could have fewer limitations for
some vaccines or some outcomes than for others. Small clinical studies can
be well conducted but the number of subjects may be too small to detect
most adverse events. Although most efficacy studies include a safety com-
ponent, the results are often nonspecific (e.g., “no serious adverse events
were detected”). Even some larger safety studies failed to detect an adverse
event. Studies in which no cases of a specific adverse event were identified
are uninformative for this review, because if the vaccinated cohort does not
include enough cases to approximate background rates, the study is under-
powered to inform an assessment. The upper limit of the 95% confidence
interval will always overlap with the background rate unless the vaccine is
protective. Some might use that information as means to approximate an
upper limit on risk, but the committee did not see that as its charge (see
Chapter 13). Studies such as these were considered to have very serious
limitations for the purpose of the committee’s assessment.
The committee was rigorous in assessing the strengths and weaknesses
of each epidemiologic study. For many of the conditions and adverse events
considered by the committee, the expected incidence and prevalence rates
in the general unvaccinated population as well as in unvaccinated but po-
tentially susceptible subgroups may be very low. Assembling a valid stan-
dard for comparison (e.g., an unvaccinated cohort of similar demographic
composition and followed over a similar time period of risk, or a control
group free of the adverse event but otherwise sufficiently similar to cases
diagnosed with the adverse event) and objectively verifying the timing and
type of vaccination and the details surrounding the onset and diagnosis
of the adverse event are complex if not prohibitively expensive research
endeavors. Although randomized clinical trials aiming to study vaccine
efficacy may provide the most valid, controlled circumstances in which to
also study vaccine safety, such trials inevitably enroll too few study par-
OCR for page 44
44 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
ticipants to be able to detect anything but extreme increases in the risks
of relatively rare adverse events of potential concern. Some studies, as will
be documented in chapters that follow, reviewed are likely the most meth-
odologically sound that can be done given the nature of the exposure and
the outcomes, even if the studies have some residual limitation due to the
challenges that often attend such research. The reader will see in the sum-
mary paragraphs for the epidemiologic studies and, in some circumstances,
the causality conclusion the committee’s interpretation of the evidence more
fully than can be captured with the formal and consistent wording of the
conclusions used in this report.
Evaluation of the Body of Studies
The committee reviewed methodological approaches of other system-
atic review efforts, but it was unable to identify one approach that incor-
porated all of the committee’s needs and could be adopted for immediate
use. Cochrane reviews, for example, focus on randomized controlled trials,
which is an uncommon design in vaccine safety studies. Other efforts
focused on evidence for or against a clinical practice or intervention (Guyatt
et al., 2008; USPSTF, 2008).
Consequently, the committee adopted key components of these other
approaches to develop a summary classification scheme that incorporates
both the quality and quantity of the individual studies and the consistency
of the group of studies in terms of direction of effect (i.e., is the effect of
the vaccine to increase risk, decrease risk, or have no effect on risk). A key
concept to these classifications is confidence, which refers to the confidence
the committee has that the true effect lies close to that of the estimate of the
average overall effect for the body of evidence (i.e., collection of reports)
reviewed (Schunemann et al., 2011), and integrates committee evaluation
of validity, precision, and consistency. Validity refers to the absence of
confounding, selection bias and information or measurement bias (i.e.,
internal validity), and the generalizability (external validity) of the findings
(Rothman et al., 2008b). Precision refers to the width of the confidence
interval (e.g., a 95% confidence interval) around an effect estimate, which
reflects the sample size of the study as well as the variability of the outcome
measurement (Rothman et al., 2008a). The wider the 95% confidence inter-
vals, the less statistical power to detect a difference as significant.
The four weight-of-evidence assessments for the epidemiologic litera-
ture are as follows:
• High: Two or more studies with negligible methodological limita-
tions that are consistent in terms of the direction of the effect and
taken together provide high confidence.
OCR for page 45
45
APPROACH
• Moderate: One study with negligible methodological limitations, or
a collection of studies generally consistent in terms of the direction
of the effect, provides moderate confidence.
• Limited: One study or a collection of studies lacking precision or
consistency provides limited, or low, confidence.
• Insufficient: No epidemiologic studies of sufficient quality found.
Assessments of high and moderate include a direction of effect. These
are to indicate increased risk of the adverse event, decreased risk of the
adverse event, or no change (“null”) in the risk of the adverse event. As-
sessments of limited or insufficient include no direction of effect.
The committee does not consider a single study—regardless of how
well it is designed, the size of the estimated effect, or the narrowness of
the confidence interval—sufficient to merit a weight of “high” or, in the
absence of strong or intermediate mechanistic evidence, sufficient to sup-
port a causality conclusion other than “inadequate to accept or reject a
causal relationship.” This requirement might seem overly rigorous to some
readers. However, the Agency for Healthcare Research and Quality advises
the Evidence-based Practice Centers that it has funded to produce evidence
reports on important issues in health care to view an evidence base of a
single study with caution (Owens et al., 2010). It does so due to the inabil-
ity to judge consistency of results, an important contributor to a strength
of evidence, because one cannot “be certain that a single trial, no matter
how large or well designed, presents the definitive picture of any particular
clinical benefit or harm for a given treatment” (Owens et al., 2010). It is
acknowledged by the committee and others (Owens et al., 2010) that policy
makers must often make decisions based on only one study. However, the
committee is not recommending policy, rather evaluating the evidence using
a transparent and justifiable framework.
Mechanistic Evidence
The committee assessed the mechanisms of vaccine adverse events by
identifying and evaluating clinical and biological evidence. First, the com-
mittee looked for evidence in the peer-reviewed literature that a vaccine was
or may be a cause of an adverse event in one or more persons (from case
reports or clinical studies) in a reasonable time period after the vaccination.
Then the committee looked for other information from the clinical and
biological (human, animal, or in vitro studies) literature that would provide
evidence of a pathophysiological process or mechanism that is reasonably
likely to cause the adverse event or to occur in response to specific im-
munization. Chapter 3 contains a discussion of the major mechanisms the
OCR for page 46
46 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
committee invokes as possible explanations of how a given adverse event
can occur after vaccination.
The committee identified many case reports in the literature describ-
ing adverse events following vaccination. For the purposes of this report,
case report refers to a description of an individual patient; one publication
could describe multiple case reports. The cases considered by the commit-
tee in weighing evidence of mechanisms were not derived from the large
epidemiology studies considered above; there was no “double counting.”
The committee evaluated each case report using a well-established set of
criteria (“attribution elements”) for case evaluation (Miller et al., 2000).
At a minimum, for a case to factor into the weight-of-evidence assessment,
it had to include specific mention of the vaccine administered, evidence of
clinician-diagnosed health outcome,3 and a specified and reasonable time
interval (i.e., temporality or latency) between vaccination and symptoms.4
Case descriptions that did not have the three basic elements described above
were not considered in the mechanistic weight-of-evidence determinations.
As discussed in the next section, however, these three criteria were only
necessary but not sufficient to affect the weight of mechanistic evidence.
After identifying cases with the three basic elements, the committee looked
for evidence in the case descriptions and in other clinical or biological litera-
ture of a possible operative mechanism(s) that would support a judgment
that the vaccination was related to the adverse event. See Chapter 3 for a
description of possible mechanisms identified by the committee.
Rechallenge cases, in which an adverse event occurred after more than
one administration of a particular vaccine in the same individual, could
influence the weight of evidence. Each rechallenge, however, must meet
the same attributes of reasonable latency, documentation of vaccination
receipt, and clinician diagnosis of the health outcome. It is possible that
one or more of the “challenges” in an individual case patient reporting is
related to a coincidental exposure; thus, the committee looked for other
information, as described below, that would support a role for the vaccine
in each challenge. The value for the committee of rechallenge cases is much
greater for monophasic conditions (events that typically happen only once,
3 On occasion, the case report author describes clinical test results or observations but does
not proffer a diagnosis. In these cases, the committee assigned the case report to the health
outcome it felt appropriate. Some authors of older case reports use a diagnosis appropriate for
the time, but by today’s understanding of clinical disease and pathophysiology, the committee
offers a different diagnosis and the case report is described within that committee-directed
assessment.
4 What constitutes reasonable latency will vary across vaccines and across adverse events.
For example, most adverse reactions from live virus vaccines would not be expected to occur
within hours of vaccination; rather, time must elapse for viral replication.
OCR for page 47
47
APPROACH
e.g., vasculitis) than for relapsing-remitting conditions, such as multiple
sclerosis or rheumatoid arthritis.
Another factor that affected the weight of evidence was information in
the clinical workup that eliminated well-accepted alternative explanations
for the condition, thus increasing the possibility that the vaccine could be
associated with the adverse event. For example, Guillain-Barré syndrome
(GBS) is known to be associated with specific infections (e.g., Campylo-
bacter). Case reports of GBS following vaccination weighed more heavily in
the committee’s assessment if the authors reported that tests for those com-
mon infections were negative, thus eliminating some likely causes for the
GBS other than vaccination. Another particularly strong piece of evidence
in the case description that affected the weight of evidence is isolation of
vaccine strain virus from the patient.
The committee follows the convention of previous IOM committees in
considering the effects of the natural infection as one type, albeit minor, of
clinical or biological evidence in support of mechanisms.5 Other evidence,
described above, provided much stronger evidence in support of the mecha-
nistic assessment.
Evidence from animal studies is also informative if the model of the
disease is well established as applicable to humans or if the basic immunol-
ogy of the vaccine reaction is well understood. In vitro studies can also be
informative, but such data must be eyed with skepticism regarding their
relationship to the human experience. Specific examples of relevant clinical
or biological information are discussed in Chapter 3 generally and in the
vaccine-specific Chapters 4 through 11.
Evaluation of the Body of Clinical and Biological (Mechanistic) Evidence
The committee reviewed the approach of previous IOM committees
addressing vaccine safety (IOM, 1991, 1994, 2001a,b, 2002a,b, 2003a,b,
2004a,b) in evaluating the body of evidence of biological mechanisms. The
committee also searched for other appropriate frameworks for evaluating
biological evidence as support for causation analyses. The committee devel-
oped four categories for the weight-of-evidence assessment. Each category
includes consideration of the clinical information from case reports and
consideration of clinical and experimental evidence from other sources.
5 The committee relied on standard textbooks of infectious disease or internal medicine for
this evaluation; the committee did not review original research to come to this determination.
This is consistent with previous IOM committees tasked with reviewing evidence of causality
for vaccine safety. Evidence consisting only of parallels with the natural infections is never
sufficient to merit a conclusion other than the evidence is inadequate to accept or reject a
causal relationship.
OCR for page 48
48 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
The following are the categories for the mechanistic weight-of-evidence
assessments:
• Strong: One or more cases in the literature, for which the commit-
tee concludes the vaccine was a contributing cause of the adverse
event, based on an overall assessment of attribution in the available
cases and clinical, diagnostic, or experimental evidence consistent
with relevant biological response to vaccine.
• Intermediate: At least two cases, taken together, for which the
committee concludes the vaccine may be a contributing cause of
the adverse event, based on an overall assessment of attribution
in the available cases and clinical, diagnostic, or experimental evi-
dence consistent with relevant biological response to vaccine. On
occasion, the committee determined that at least two cases, taken
together, while suggestive, are nonetheless insufficient for the com-
mittee to conclude the vaccine may be a contributing cause of the
adverse event, based on an overall assessment of attribution in the
available cases and clinical, diagnostic, or experimental evidence
consistent with relevant biological response to vaccine. This evi-
dence has been identified in the text as “low-intermediate.”
• Weak: Insufficient evidence from cases in the literature for the com-
mittee to conclude the vaccine may be a contributing cause of the
adverse event, based on an overall assessment of attribution in the
available cases and clinical, diagnostic, or experimental evidence
consistent with relevant biological response to vaccine.
• Lacking evidence of a biologic mechanism: No clinical, diagnos-
tic, or experimental evidence consistent with relevant biological
response to vaccine,6 regardless of the presence of individual cases
in the literature.
CAUSALITY ASSESSMENT
The committee adopted the categories of causation developed by previ-
ous IOM committees. Implicit in these categories is that “the absence of
evidence is not evidence of absence.” That is, the committee began its as-
sessment from the position of neutrality; until all evidence was reviewed,
it presumed neither causation nor lack of causation. The committee then
6 The committee considered the clinical manifestations of the natural infection against which
the vaccine is directed to be sufficient for a weight of evidence of weak, rather than lacking.
As will be discussed in a subsequent section, a mechanism weight of evidence of weak alone
is never sufficient to support a causality conclusion other than the evidence is inadequate to
accept or reject a causal relationship.
OCR for page 49
49
APPROACH
moved from that position only when the combination of epidemiologic
evidence and mechanistic evidence suggested a more definitive assessment
regarding causation, either that vaccines might or might not pose an in-
creased risk for an adverse event.
The following are the categories of causation used by the committee:
Evidence convincingly supports7 a causal relationship—This ap-
•
plies to relationships in which the causal link is convincing, as with
the oral polio vaccine and vaccine-associated paralytic polio.
Evidence favors acceptance of a causal relationship—Evidence is
•
strong and generally suggestive, although not firm enough to be
described as convincing or established.
Evidence is inadequate to accept or reject a causal relationship—
•
The evidence is not reasonably convincing either in support of
or against causality; evidence that is sparse, conflicting, of weak
quality, or merely suggestive—whether toward or away from cau-
sality—falls into this category.8 Where there is no evidence meeting
the standards described above, the committee also uses this causal
conclusion.
Evidence favors rejection of a causal relationship—The evidence
•
is strong and generally convincing, and suggests there is no causal
relationship.
The category of “establishes or convincingly supports no causal rela-
tionship” is not used because it is virtually impossible to prove the absence
of a relationship with the same certainty that is possible in establishing the
presence of one. Even in the presence of a convincing protective effect of
vaccine in epidemiology, studies may not rule out the possibility that the re-
action is caused by vaccine in a subset of individuals. Thus, the framework
for this and previous IOM reports on vaccine safety is asymmetrical. The
committee began not by assuming the causal relationship does not exist,
but by requiring evidence to shift away from the neutral position that the
evidence is “inadequate to accept or reject” a causal relationship.
The committee then established a general framework by which the
two streams of evidence (epidemiologic and mechanistic) influence the final
causality conclusion. It is important to note that mechanistic evidence can
only support causation. Epidemiologic evidence, by contrast, can support
(“favors acceptance of”) a causal association or can support the absence
of (“favors rejection of”) a causal association in the general population
and in various subgroups that can be identified and investigated, unless or
7 Previous IOM committees used the term establishes instead of convincingly supports.
8 See Chapter 13 for further discussion.
OCR for page 50
50 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
until supportive mechanistic evidence is discovered or a rare, susceptible
subgroup can be identified and investigated. This framework needed to ac-
commodate the reality that for any given causality conclusion one or both
of the types of evidence could be lacking, the two types of evidence could
conflict, or neither type of evidence might definitively influence the causal-
ity conclusion.
The framework also had to accommodate known limitations of both
types of evidence. Epidemiologic analyses are usually unable to detect an
increased or decreased risk that is small, unless the study population is
very large or the difference between the groups (e.g., vaccinated vs. unvac-
cinated) at risk is very high (e.g., smoking increases the risk of lung cancer
by at least 10-fold). Epidemiologic analyses also cannot identify with cer-
tainty which individual in a population at risk will develop a given condi-
tion. These studies also can fail to detect risks that affect a small subset of
the population. Mechanistic evidence, particularly that emerging from case
reports, occasionally can provide compelling evidence of an association
between exposure to a vaccine and an adverse reaction in the individual
being studied, but it provides no meaningful information about the degree
of risk to the population or even to other individuals who have the same
predisposing characteristics. The occurrence rate of the adverse event or
condition in the general population cannot be estimated from case reports,9
nor can one be certain that the risk is homogeneous across potentially vul-
nerable subgroups within the general population (e.g., the developing fetus
and infants under 24 months, immunologically compromised individuals,
or individuals with a rare genetic predisposition).
The framework does not accommodate any information regarding the
benefit of the vaccine to either population or individual health. The focus
of this particular committee is only on the question of what particular vac-
cines can cause particular adverse effects.
In general, the framework shown in Figure 2-2 illustrates how causality
conclusions can be based primarily on epidemiologic evidence, primarily on
mechanistic evidence, or on a combination of the two, and that on occa-
sion expert judgment, such as that provided by the complement of expertise
represented on the committee, is needed to weigh uncertain or competing
evidence.
Evidence Convincingly Supports a Causal Relationship
The framework allows for a causality conclusion of “convincingly sup-
ports” based on an epidemiologic weight-of-evidence assessment of high in
9 See Chapter 13 for further discussion.
OCR for page 51
EPIDEMIOLOGIC ASSESSMENT MECHANISTIC ASSESSMENT CAUSALITY CONCLUSION
Inadequate
Low-
High Moderate Favors Favors Convincingly
Inter- to Accept
Inter-
(decreased risk (decreased risk
High Moderate
(increased risk) (increased risk) Limited Strong Rejection Acceptance Supports
mediate Weak Lacking or Reject
mediate
High
(increased risk)
Convincingly
Supports
Strong
Moderate
(increased risk)
Favors
Acceptance
Inter-
mediate
High Favors
(decreased risk
Rejection
*
Moderate ,
Inadequate
to Accept
or Reject
Low-Intermediate,
Weak, or Lacking***
* Causality conclusion is favors rejection only if mechanistic assessment is not strong or intermediate.
** Causality conclusion is inadequate to accept or reject only if mechanistic assessment is not strong or intermediate.
*** Causality conclusion is inadequate to accept or reject only if epidemiologic assessment is not
51
FIGURE 2-2 Strength of evidence that determined the causality conclusions.
OCR for page 52
52 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
the direction of increased risk (which requires at least two well-conducted
epidemiologic studies).
The framework also allows strong mechanistic evidence, which requires
at least one case report in which compelling evidence exists that the vac-
cine indeed did cause the adverse event, to carry sufficient weight for the
committee to conclude the evidence convincingly supports a causal relation-
ship. The committee considered laboratory-confirmed, vaccine-strain virus
isolation compelling evidence to attribute the disease to the vaccine-strain
virus and not other etiologies. The committee recognizes that vaccine-strain
virus can transiently appear in otherwise sterile spaces after vaccination;
however, the committee determined that the accurate detection of vaccine-
strain virus in symptomatic individuals to be strong evidence that the
vaccine caused the symptoms. This conclusion can be reached even if the
epidemiologic evidence is rated “high” in the direction of no increased risk
or even decreased risk. The simplest explanation in this circumstance is that
the adverse effect is real but also very rare. Another way of stating this is
that if the vaccine did cause the adverse effect in one person, then it can
cause the adverse effect in someone else (IOM, 1994). It might seem that
the committee “overvalued” case reports in allowing one case to provide
convincing evidence of causation; however, it is a rare case report that is so
convincing. For most of the specific causality conclusions in this category,
more than one compelling case report existed.
The isolated report of one convincing case provides no information
about the risk of the adverse effect in the total population of vaccinated
individuals compared with unvaccinated individuals. If the one convincing
case has an underlying condition that may increase susceptibility to the
adverse effect, it might have no relevance to the otherwise not-susceptible
population.
As will be described in subsequent chapters of the report, the committee
concluded the evidence convincingly supports 14 specific vaccine–adverse
event relationships. In all but one of these relationships, the conclusion was
based on strong mechanistic evidence with the epidemiologic evidence rated
as either limited confidence or insufficient. When moderate or strong epide-
miologic evidence is not available to support the committee’s conclusions
favoring causality, it is difficult, if not impossible, to quantify the risk of the
adverse event in either the entire population or the susceptible subgroup.
See Chapter 13 for a discussion of this issue.
Evidence Favors Acceptance of a Causal Relationship
A conclusion of “favors acceptance of a causal relationship” must be
supported by either epidemiologic evidence of “moderate” certainty of
an increased risk or by mechanistic evidence of intermediate weight. The
OCR for page 53
53
APPROACH
framework requires more than one epidemiologic study or more than one
case report (with supporting but not conclusive mechanistic information)
in support of this causality conclusion. A weight of mechanistic evidence of
low-intermediate was not sufficient, without concurring epidemiologic evi-
dence, to support a conclusion favoring acceptance of a causal relationship.
As will be described in subsequent chapters of the report, the committee
concluded the evidence favors acceptance of four specific vaccine–adverse
event relationships.
Evidence Favors Rejection of a Causal Relationship
The framework allows the committee to “favor rejection” of a causal
relationship only in the face of epidemiologic evidence rated as high or
moderate in the direction of no effect (the null) or of decreased risk and
the absence of strong or intermediate mechanistic evidence in support of a
causal relationship. As described above, the committee requires more than
one epidemiologic study to merit a conclusion that the evidence favors
rejection of a causal relationship.
As will be described in subsequent chapters of the report, the commit-
tee concluded the evidence favors rejection of five specific vaccine–adverse
event relationships.
Evidence Is Inadequate to Accept or Reject a Causal Relationship
The committee identified two main pathways by which it concludes
that the evidence is “inadequate to accept or reject” a causal relationship.
The most common pathway to this conclusion occurs when the epidemio-
logic evidence was of limited certainty or insufficient and the mechanistic
evidence was weak or lacking. Another pathway occurs when the epidemio-
logic evidence is of moderate certainty of no effect but the mechanistic evi-
dence is intermediate in support of an association. The committee analyzed
these sets of apparently contradictory evidence and ultimately depended
upon its expert judgment in deciding if a conclusion to favor acceptance
based on the intermediate mechanistic data was warranted or if the conclu-
sion remained as “inadequate to accept or reject” a causal relationship. The
committee required more than one epidemiologic study to conclude other
than that the evidence is inadequate to accept or reject a causal relationship.
As will be described in subsequent chapters of the report, the committee
concluded the evidence was inadequate to accept or reject the vast majority
of specific vaccine–adverse event relationships. See Chapter 13 for a discus-
sion of this conclusion.
OCR for page 54
54 ADVERSE EFFECTS OF VACCINES: EVIDENCE AND CAUSALITY
SPECIAL CONSIDERATIONS
As described in Chapter 3, the committee recognized that the risk
of an adverse effect of a vaccine can be influenced by host factors, some
known and others not yet understood. Where the committee thought the
evidence—whether from epidemiologic analyses or from the clinical stud-
ies—regarding risks to subpopulations was informative, evidence-based,
and biologically sound, it made separate conclusions. For example, the
risk of invasive disease following varicella vaccine, a live virus vaccine, is
likely much higher in immunocompromised persons than in persons who
are immunocompetent. Other subpopulation analyses in the report include
age and sex for some specific adverse events.
In their consideration of several adverse events, the committee con-
cluded that the mechanism of injury was likely unrelated to the specific
antigenic or other components of the vaccine. The committee concluded
that the exposure of concern is not the injected vaccine, rather the injection
of the vaccine. The adverse events include syncope, complex regional pain
syndrome, and deltoid bursitis. These are covered in Chapter 12.
REFERENCES
Guyatt, G. H., A. D. Oxman, G. E. Vist, R. Kunz, Y. Falck-Ytter, P. Alonso-Coello, H. J.
Schunemann, and G. W. Grp. 2008. GRADE: An emerging consensus on rating quality of
evidence and strength of recommendations. British Medical Journal 336(7650):924-926.
Halsey, N. A. 2002. The science of evaluation of adverse events associated with vaccination.
Seminars in Pediatric Infectious Diseases 13(3):205-214.
Hill, A. B. 1965. Environment and disease—association or causation? Proceedings of the
Royal Society of Medicine-London 58(5):295-300.
IOM (Institute of Medicine). 1991. Adverse effects of pertussis and rubella vaccines: A report
of the committee to review the adverse consequences of pertussis and rubella vaccines.
Washington, DC: National Academy Press.
IOM. 1994. Adverse events associated with childhood vaccines: Evidence bearing on causality.
Washington, DC: National Academy Press.
IOM. 2001a. Immunization safety review: Measles-mumps-rubella vaccine and autism.
Washington, DC: National Academy Press.
IOM. 2001b. Immunization safety review: Thimerosal-containing vaccines and neuro-
developmental disorders. Washington, DC: National Academy Press.
IOM. 2002a. Immunization safety review: Hepatitis B vaccine and demyelinating neurological
disorders. Washington, DC: The National Academies Press.
IOM. 2002b. Immunization safety review: Multiple immunizations and immune dysfunction.
Washington, DC: National Academy Press.
IOM. 2003a. Immunization safety review: SV40 contamination of polio vaccine and cancer.
Washington, DC: The National Academies Press.
IOM. 2003b. Immunization safety review: Vaccinations and sudden unexpected death in
infancy. Washington, DC: The National Academies Press.
IOM. 2004a. Immunization safety review: Influenza vaccines and neurological complications.
Washington, DC: The National Academies Press.
OCR for page 55
55
APPROACH
IOM. 2004b. Immunization safety review: Vaccines and autism. Washington, DC: The
National Academies Press.
Liberati, A., D. G. Altman, J. Tetzlaff, C. Mulrow, P. C. Gotzsche, J. P. A. Ioannidis, M.
Clarke, P. J. Devereaux, J. Kleijnen, and D. Moher. 2009. The prisma statement for
reporting systematic reviews and meta-analyses of studies that evaluate health care inter-
ventions: Explanation and elaboration. PLoS Medicine 6(7).
Loke, Y., D. Price, and A. Hermheimer. 2008. Chapter 14: Adverse effects. In Cochrane hand-
book for systematic reviews of interventions, edited by J. P. T. Higgins, and S. Green:
The Cochrane Collaboration.
Miller, F. W., E. V. Hess, D. J. Clauw, P. A. Hertzman, T. Pincus, R. M. Silver, M. D. Mayes, J.
Varga, T. A. Medsger, Jr., and L. A. Love. 2000. Approaches for identifying and defining
environmentally associated rheumatic disorders. Arthritis & Rheumatism 43(2):243-249.
Owens, D. K., K. N. Lohr, D. Atkins, J. R. Treadwell, J. T. Reston, E. B. Bass, S. Chang, and
M. Helfand. 2010. AHRQ series paper 5: Grading the strength of a body of evidence
when comparing medical interventions—Agency for Healthcare Research and Quality
and the effective health-care program. Journal of Clinical Epidemiology 63(5):513-523.
Rothman, K. J., S. Greenland, and T. L. Lash. 2008a. Precision and statistics in epidemiologic
studies. In Modern epidemiology, 3rd ed., Philadelphia: Lippincott Williams & Wilkins.
Pp. 148-167.
Rothman, K. J., S. Greenland, and T. L. Lash. 2008b. Validity in epidemiologic studies. In
Modern epidemiology, 3rd ed., Philadelphia: Lippincott Williams & Wilkins. Pp. 128-147.
Schunemann, H. J., S. Hill, G. H. Guyatt, E. A. Akl, and F. Ahmed. 2011. The GRADE
approach and Bradford Hill’s criteria for causation. Journal of Epidemiology and Com-
munity Health 65(5):392-395.
Stroup, D. F., J. A. Berlin, S. C. Morton, I. Olkin, G. D. Williamson, D. Rennie, D. Moher,
B. J. Becker, T. A. Sipe, S. B. Thacker, and M. Grp. 2000. Meta-analysis of observational
studies in epidemiology—a proposal for reporting. Journal of the American Medical
Association 283(15):2008-2012.
USPSTF (U.S. Preventive Services Task Force). 2008. U.S. Preventive Services Task Force
procedure manual. AHRQ Publication No. 08-05118-EF AHRQ.
WHO (World Health Organization). 2001. Causality assessment of adverse events following
immunization. Weekly Epidemiological Record 76(12):85-89.
OCR for page 56
