US Supreme Court Gives Total Liability Shield to Big Drug Companies.pdf by VegasStreetProphet



Varicella Mortality: Trends before Vaccine Licensure in the United States,
Pamela A. Meyer,1,a Jane F. Seward,2 Aisha O. Jumaan,2                               Council of State and Territorial Epidemiologists and 2National

and Melinda Wharton2                                                             Immunization Program, Centers for Disease Control and Prevention,
                                                                                                                                  Atlanta, Georgia

                         We examined varicella deaths in the United States during the 25 years before vaccine li-
                       censure and identified 2262 people who died with varicella as the underlying cause of death.
                       From 1970 to 1994, varicella mortality declined, followed by an increase. Mortality rates were
                       highest among children; however, adult varicella deaths more than doubled in number, pro-
                       portion, and rate per million population. Despite declining fatality rates, in 1990–1994, adults

                                                                                                                                                        Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
                       had a risk 25 times greater and infants had a risk 4 times greater of dying from varicella than
                       did children 1–4 years old, and most people who died of varicella were previously healthy.
                       Varicella deaths are now preventable by vaccine. Investigation and reporting of all varicella
                       deaths in the United States is needed to accurately document deaths due to varicella, to
                       improve prevention efforts, and to evaluate the vaccine’s impact on mortality.

   Varicella is a highly infectious disease that is preventable by             Methods
vaccine. Before vaccine licensure in 1995, ∼4 million cases per
year resulted in 9300 hospitalizations [1] and 100 deaths each                    Sources of mortality data. We focused on the 25 years before
year [2]. Children bore the brunt of the health burden, accounting             vaccine licensure because there is a 2-year delay in the availability
for 190% of cases, 66% of hospitalizations, and 45% of deaths                  of national mortality public-use data (e.g., 1997 is now available)
                                                                               and because it is too early to evaluate the impact of the varicella
(Centers for Disease Control and Prevention [CDC], unpublished
                                                                               vaccination program, which became operational in most state
data); however, the risk of severe complications and death was                 health departments by the end of 1996. We analyzed death records
highest among infants, adults [2], and immunocompromised in-                   for 1970–1994 in the national Multiple Cause Mortality Database,
dividuals [3, 4]. Moreover, complications and deaths were de-                  which is maintained by CDC’s National Center for Health Statis-
scribed commonly among previously healthy individuals [5–9].                   tics (NCHS). The Multiple Cause Mortality Database includes all
   Use of the varicella vaccine, which is recommended for rou-                 deaths in the United States, except for 3 years (1972, 1981, and
tine use among susceptible people 112 months old [10, 11], is                  1982). In 1972, only 50% of deaths were coded for underlying cause
anticipated to alter the epidemiology of varicella by shifting                 of death, and in 1981 and 1982, only 50% of deaths were coded
                                                                               for multiple conditions in 19 of 50 states. To estimate the actual
the largest proportion of cases from children to adults. High
                                                                               national deaths, NCHS duplicated records for 1981–1982 in the
vaccine coverage in childhood, especially if combined with a
                                                                               public-use files, and we duplicated 1972 records for this study.
catch-up immunization program at adolescence, is expected to                      The cause or causes of death obtained from death records were
lead to a dramatic overall reduction in varicella cases and com-               coded according to the Eighth Revision International Classification
plications among both children and adults [12].                                of Diseases, Adapted for Use in the United States (ICD-8A) [13]
   There has been no long-term, comprehensive analysis of vari-                for 1968–1978 and according to the 9th revision of the International
cella mortality for all age groups in the United States. We, there-            Classification of Diseases (ICD-9) [14] for 1979–1994. The under-
fore, analyzed national mortality data to characterize varicella               lying cause of death is determined by a computer algorithm that
deaths among United States residents during 1970–1994, the 25-                 evaluates all the codes for cause of death [15]. The algorithm assigns
                                                                               an underlying cause of death on the basis of conditions and their
year period before varicella vaccine licensure. These data will
                                                                               positions as listed on the death certificates. We analyzed deaths for
serve as baseline data for an evaluation of the impact of the                  which varicella (ICD-8A code 0.52 and ICD-9 code 0.52) was de-
vaccination program on varicella mortality in the United States.               termined to be the underlying cause of death for consistency with
                                                                               official death statistics reported from NCHS. The algorithm for
  Received 14 February 2000; revised 21 April 2000; electronically published   assigning varicella as the underlying cause of death and the in-
12 July 2000.                                                                  structions for using the varicella codes were the same throughout
    Present affiliation: National Center for Environmental Health, Centers      the study period.
for Disease Control and Prevention, Atlanta, Georgia.
                                                                                  Preexisting conditions and complications. We defined preexist-
  Reprints or correspondence: Dr. Jane F. Seward, National Immunization
Program, Centers for Disease Control and Prevention, 1600 Clifton Rd. NE,      ing conditions as conditions listed on the death record that were
Mailstop E-61, Atlanta, GA 30333 (                            likely to have existed before the varicella infection and classified
                                                                               them according to their potential to increase the risk for severe
The Journal of Infectious Diseases 2000; 182:383–90
  2000 by the Infectious Diseases Society of America. All rights reserved.     varicella infection. High-risk conditions were defined as those for
0022-1899/2000/18202-0001$02.00                                                which the Advisory Committee on Immunization Practices does
384                                                                         Meyer et al.                                                             JID 2000;182 (August)

                      Table 1. Eighth Revision International Classification of Diseases, Adapted for Use in the United States
                      (ICD-8A) [13] and Ninth Revision International Classification of Diseases (ICD-9) [14] codes.
                      Condition or
                      complication                                          ICD-8A                                        ICD-9
                      High-risk preexisting conditions
                        Malignancies                        140.0–163.9; 170.0–174.0; 180.0–209.0        140.0–165.9; 170.0–175.0; 179.0–208.9
                        Immune deficiency                    275.0–275.1                                  279.0–279.9
                        Human immunodeficiency                                —                           042.0–044.9, 795.8
                        Central nervous system              347.9, 780.4, 065.0, 323.0, 781.7,           334.3, 334.9, 781.3, 049.9, 323.8,
                          complications                       320.8, 320.9, 045.9, 347.9                   323.9, 348.3, 047.8, 047.9,
                                                                                                           322.0–322.9, 321.7,
                                                                                                           348.1–348.5,348.9, 331.8
                        Hemorrhagic                         286.4–286.5, 287.0–287.2, 287.9              286.6, 286.9, 287.0–287.5, 287.9

                                                                                                                                                                             Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
                        Secondary infections                034.1, 035.0, 039.0–039.9, 732.0,            034.1, 035, 040.0–040.9, 041.0–041.9,
                                                              420.0–421.9, 320.0–320.9,                    729.4, 420.0–420.9, 421.0–421.9,
                                                              680.0–684.0, 686.0, 686.9,                   429.0, 320.1–320.3, 320.8–320.9,
                                                              038.0–038.9, 322.0, 324.0                    680.0–684.0, 686.0, 686.8–686.9,
                                                                                                           038.0–038.9, 790.7, 324.0, 324.1,
                                                                                                           324.9, 326,
                        Pneumonia                           481.0–482.3, 482.9, 480.0, 486.0,            481.0, 480.8, 480.9, 486.0, 483.0,
                                                              483.0–484.0, 485.0                           485.0
                           Cerebral ataxia, encephalitis, myelitis, encephalopathy, encephalomyelitis, meningitis, other diseases of brain
                      (ICD-8A includes Reye’s syndrome), cerebral degeneration, and other (ICD-9 includes Reye’s syndrome).
                           Purpura, thrombocytopenia, and other hemorrhagic conditions.
                          Other bacterial infections, infective myositis and other inflammatory diseases of tendon and fascia, endocarditis,
                      pericarditis, bacterial meningitis, skin infections, cellulitis, impetigo, septicemia, and intracranial/intraspinal abscess.
                           Bacterial, viral, organism unspecified, other, and bronchopneumonia.

not recommend vaccine administration because of the risk of severe                     death rates for the foreign-born population, we used the deaths
varicella disease in the vaccine recipient—for example, malignan-                      for 1980–1994 (the only complete 5-year intervals for which for-
cies and AIDS (comprising human immunodeficiency virus [HIV]/                           eign-born status was recorded on the death records) and applied
AIDS). We searched death records for additional conditions that                        the foreign-born population estimates obtained from the Census
may increase the risk for severe varicella infection either due to the                 Bureau’s current population reports for 1980 and 1990 [16]. To
underlying condition or to treatment for the condition (e.g.,                          account for changes in the population age distribution over the 25-
asthma, systemic lupus erythematosus, cystic fibrosis, rheumatoid                       year period, we calculated age-adjusted varicella mortality rates by
arthritis and allied conditions, diabetes, and scleroderma). Because                   making use of the estimated United States year 2000 population
of the low proportion of these potentially moderate risk conditions                    [17]. For calculation of case-fatality rates (CFR), we estimated the
(2.7%) over the 25-year period, we analyzed them with death re-                        number of varicella cases by using the 1970–1994 National Health
cords of the previously healthy—that is, those that did not list any                   Interview Surveys [18], which are maintained by NCHS. The Na-
of the high-risk preexisting conditions (table 1).                                     tional Health Interview Surveys collect information on health-re-
   We defined varicella-associated complications as conditions listed                   lated issues by conducting personal interview surveys annually and
on the death record that were likely to have occurred as a conse-                      uses a nationwide sample of the civilian, noninstitutionalized popu-
quence of varicella infection. We focused on 4 categories: pneumonia,                  lation in the United States. We calculated age-specific varicella
central nervous system (CNS) complications, hemorrhagic condi-
tions, and secondary infections. Pneumonia was defined as either
viral or bacterial pneumonia. CNS complications included conditions
that affected the CNS as a direct consequence of varicella infection,
such as encephalitis, cerebellar ataxia, and Reye’s syndrome. Hem-
orrhagic conditions included purpura and thrombocytopenia. Sec-
ondary infections were defined as bacterial infections, including those
that affect the CNS, such as brain abscess (table 1).
   Data analysis. Analyses were performed with SAS statistical
software, version 6.12 (SAS Institute, Cary, NC). We defined chil-
dren as people 19 years old and adults as people 20 years old.
Neonatal deaths included infants !28 days old. We assessed sea-
sonality by examining the number and percentage of varicella
deaths, by age group and by month of death. We examined sea-
sonality for all years and for the unduplicated years, to determine
if duplicating records distorted findings; however, the results were                    Figure 1. Percentage of varicella deaths, by month of death and age
similar, so we present seasonality for all years. To calculate varicella               group, United States, 1970–1994.
JID 2000;182 (August)                           Varicella Deaths in the United States, 1970–1994                                            385

                        Table 2. Number and percentage of people who died with varicella as the underlying cause
                        of death, by age group, sex, race, and foreign-born status for all years and for each 5-year
                        interval, United States, 1970–1994.
                                              All years       1970–1974      1975–1979      1980–1984      1985–1989      1990–1994
                        Variable             (n p 2262)       (n p 555)      (n p 471)      (n p 335)      (n p 376)      (n p 525)
                        Age group
                          !1 year              187   (8.2)     60   (10.8)    28   (6.0)     20   (6.0)     35   (9.3)     44   (8.4)
                          1–4 years            335   (14.8)   118   (21.3)    76   (16.1)    37   (11.0)    38   (10.1)    66   (12.6)
                          5–9 years            560   (24.8)   216   (38.9)   138   (29.3)    70   (20.9)    58   (15.4)    78   (14.8)
                          10–14 years          175   (7.7)     46   (8.3)     44   (9.3)     36   (10.7)    20   (5.3)     29   (5.5)
                          15–19 years           78   (3.5)      4   (0.7)     16   (3.4)     10   (3.0)     25   (6.7)     23   (4.4)
                            20 years           927   (41.0)   111   (20.0)   169   (35.9)   162   (48.4)   200   (53.2)   285   (54.3)
                          Female              1105 (48.9)     291 (52.4)     231 (49.0)     165 (49.2)     170 (45.2)     248 (47.2)
                          Male                1157 (51.1)     264 (47.6)     240 (51.0)     170 (50.8)     206 (54.8)     277 (52.8)

                                                                                                                                                   Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
                          African American     157 (6.9)       57 (10.3)      33 (7.0)       17 (5.1)       14 (3.7)       36 (6.9)
                          White               1895 (83.8)     488 (87.9)     425 (90.2)     285 (85.0)     304 (80.9)     393 (74.8)
                          Other                210 (9.3)       10 (1.8)       13 (2.8)       33 (9.9)       58 (15.4)      96 (18.3)
                        Birthplace, age
                             !20 years           NA              NA             NA            5 (1.5)       10 (2.7)       11 (2.1)
                               20 years          NA              NA             NA           35 (10.4)      43 (11.4)      62 (11.8)
                          United States
                             !20 years           NA              NA             NA          167 (49.9)     166 (44.1)     229 (43.6)
                               20 years          NA              NA             NA          126 (37.6)     155 (41.2)     222 (42.3)
                             !20 years           NA              NA             NA            1 (0.3)        0 (0.0)        0 (0.0)
                               20 years          NA              NA             NA            1 (0.3)        2 (0.6)        1 (0.2)
                          NOTE.    Data are no. (%). NA, not applicable.

incidence rates for each of the 5-year periods and multiplied those            adults 50 years old, compared with adults 20–49 years old
rates by the averaged United States population for the appropriate             and children !20 years old. There was a strong seasonal pattern
age group and 5-year period.                                                   for deaths regardless of the presence or absence on the death
                                                                               record of a high-risk preexisting medical condition.
                                                                                  Varicella mortality rates. Varicella mortality rates have been
                                                                               consistently higher among children !15 years old than among
   From 1970 to 1994, 2262 death records listed varicella as the               adults; the highest rates occurred among infants !12 months old.
underlying cause of death, an average of 90 deaths per year;                   The overall varicella mortality rate showed a pattern of decline
these ranged from 47 deaths in 1986 to 138 in 1973. Table 2                    followed by an increase (figure 2). The pattern of change in
shows the demographic characteristics of decedents whose un-                   mortality rates differed by age group (table 3). Mortality rates
derlying cause of death was varicella. For the 25-year period,                 among children !10 years old declined through the 1970s and
most deaths occurred among children (59.0%) and whites                         early 1980s, with the greatest decline among children 5–9 years
(83.8%). Adults 150 years old accounted for 19.1% of all vari-                 old, but they then increased in the late 1980s and early 1990s.
cella deaths, and infants !28 days old accounted for 1.2% of                      In contrast to the overall decline in mortality rates among
all varicella deaths. There was a shift in the age distribution of             children from 1970–1974 to 1990–1994, mortality rates have in-
varicella deaths from children (80.0%) in 1970–1974 to adults                  creased among adults, neonates, and foreign-born adult residents.
(54.3%) in 1990–1994. Whites accounted for most varicella                      Adult mortality rates increased 83% from 0.17 per million pop-
deaths during the study period; however, the proportion of                     ulation during 1970–1974 to 0.31 per million population during
deaths among decedents of races other than white or African                    1990–1994. Among neonates, the varicella mortality rate in-
American increased from 1.8% to 18.3% during the 25-year                       creased from 0.59 per million live births in 1970–1974 to 2.47 in
period. Foreign-born status was available only for the 5-year                  1990–1994. There also has been an increase in age-specific var-
intervals from 1980–1984 through 1990–1994. Of the 166 for-                    icella mortality rates among foreign-born residents. During
eign-born decedents in these years, 84.3% were 20 years old.                   1980–1984, when information on place of birth became available,
   Seasonality. Varicella deaths showed a strong seasonal pat-                 mortality rates were higher among foreign-born than among
tern consistent with patterns for varicella cases. Overall, 44.2%              United States–born residents who were 45 years old (3.48 vs.
of deaths occurred between March and May and 8.4% occurred                     1.06) and among those 20–44 years old (2.36 vs. 0.74), but not
between August and October. Figure 1 shows a seasonal pattern                  among those !20 years old (2.36 vs. 2.38). Mortality rates in-
in all age groups, but the pattern was less pronounced among                   creased among both foreign-born and United States–born resi-
386                                                             Meyer et al.                                                   JID 2000;182 (August)

                                                                        among people who died from varicella was pneumonia (27.6%),
                                                                        followed by CNS complications (21.1%), secondary infections
                                                                        (8.6%), and hemorrhagic conditions (4.8%). Fatal pneumonia
                                                                        complications affected all age groups (range, 21.3% for 10–19-
                                                                        year-olds to 32.5% for people 20 years old). Over the 25-year
                                                                        period, pneumonia complications declined among people who
                                                                        died from varicella, especially among decedents with preexisting
                                                                        high-risk conditions, but also among decedents who were pre-
                                                                        viously healthy adults (60.4%). Yet among previously healthy
                                                                        children who died of varicella, the proportion with pneumonia
                                                                        complications was relatively stable at ∼20% (table 5).
Figure 2. Varicella mortality rates (age-adjusted to year 2000 pop-        CNS complications occurred in 44.1% of varicella deaths in

                                                                                                                                                       Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
ulation), United States, 1970–1994.                                     1970–1974, but only in 5.7% in 1990–1994. CNS complications
                                                                        were more common among children who died of varicella who
                                                                        were !15 years of age and among decedents of all ages who
dents during 1990–1994 and were consistently higher for foreign-
                                                                        were previously healthy. Nonetheless, fatal varicella compli-
born residents who were 45 years old (5.11 vs. 1.16), 20–44
years old (2.58 vs. 1.56), and !20 years old (4.16 vs. 3.31).           cations varied by age and risk status over time. Among children
   Mortality rates have been similar for males and females              who died of varicella who were previously healthy, the pro-
throughout the study period. Mortality rates among male pa-             portion with CNS complications declined from 64.0% in
tients and female patients declined through the 1970s and early         1970–1974 to 8.9% in 1990–1994. Although fatal CNS com-
1980s, then increased for males in the late 1980s and for females       plications were less common among previously healthy adult
in the 1990s. Mortality rates were higher among whites than             decedents than among child decedents, the proportions also
among African Americans, and mortality rates declined in the            declined, from 11.9% in 1970–1974 to 1.4% in 1990–1994.
early 1970s and 1980s; however, rates began to increase among           Reye’s syndrome, which is included in our definition of CNS
African Americans in 1990–1994. Mortality rates for people in           complications, occurred almost exclusively among children !15
all other racial categories combined have been higher than rates        years old who died of varicella and declined sharply in the early
for either whites or African Americans and have been steadily           1980s. During 1970–1979, there were 170 (16.6%) varicella
increasing.                                                             deaths records that listed possible Reye’s syndrome, compared
   CFR. Overall, the CFR were highest among adults, fol-                with 28 (2.3%) during 1980–1994.
lowed by infants, with the lowest rates among children 1–4                 In contrast to the decline in CNS complications and pneu-
years old, followed closely by children 5–9 years old. With the         monia reported in varicella deaths over the 25 years under
exception of children 15–19 years old, CFR declined between             study, hemorrhagic conditions and secondary infections in-
51.3%–72.2% in all age groups, with the greatest decline oc-            creased. Hemorrhagic conditions reported in varicella death
curring among adults. The fluctuation in CFR among the                   records increased from 2.5% in 1970–1974 to 8.0% in
15–19-year-old children reflects the small numbers of deaths in
this age group. In 1990–1994, adults had a risk 25 times greater
and infants had a risk 4 times greater of dying from varicella          Table 3. Five-year average varicella mortality rates, by age group,
                                                                        sex, and race, United States, 1970–1994.
than did children 1–4 years old (table 4).
                                                                                                 1970–      1975–       1980–        1985–    1990–
   Preexisting conditions. Overall, 622 (27.5%) varicella death
                                                                        Variable                  1974      1979         1984         1989     1994
records listed a preexisting high-risk condition; malignancies ac-
                                                                        Age group
counted for 88.4% of all high-risk conditions; however, the pro-          !1 year                 3.61       1.72       1.11         1.84     2.24
portion of preexisting conditions varied during the study period.         1–4 years               1.73       1.21       0.54         0.52     0.86
Malignancies were listed on 22.5% of varicella death records in           5–9 years               2.29       1.59       0.86         0.67     0.85
                                                                          10–14 years             0.44       0.45       0.40         0.24     0.32
1970–1974, 35.9% in 1975–1979, and 11.4% in 1990–1994. HIV/               15–19 years             0.40       0.15       0.10         0.27     0.26
AIDS codes first appeared on varicella death records during                  20 years              0.17       0.23       0.20         0.23     0.31
1985–1989. By 1990–1994, HIV/AIDS was listed on 2.1% of                 Sex
                                                                          Female                  0.47       0.40       0.28         0.27     0.37
varicella death records for children 19 years old, 13.5% for              Male                    0.46       0.43       0.31         0.35     0.45
those 20–49 years old, and 2.8% for people 150 years old. Most          Race

varicella deaths occurred in previously healthy individuals, rang-        African American        0.43       0.24       0.15         0.09     0.22
                                                                          White                   0.47       0.43       0.29         0.30     0.36
ing from 66.5% of child varicella deaths and 56.8% of adult               Other                   0.53       0.64       1.33         1.50     1.68
varicella deaths in 1980–1984 to 88.8% and 74.7% of child and                Total                0.46       0.41       0.29         0.31     0.40
adult varicella deaths, respectively, in 1990–1994.                       NOTE. Mortality rate is per 1,000,000 population.
   Complications. Overall, the most common complication                     Sex and race are age-adjusted to 2000 U.S. population.
JID 2000;182 (August)                                  Varicella Deaths in the United States, 1970–1994                                               387

                        Table 4. Varicella case-fatality rate, by age group and year of death (5-year interval) and 95%
                        confidence intervals (CIs), United States, 1970–1994.
                        Age, years        1970–1974             1975–1979            1980–1984            1985–1989            1990–1994
                        !1             7.6   (3.0–11.9)       7.2   (1.2–13.1)      5.1   (0.1–10.1)     6.8   (1.8–11.8)     3.7   (1.3–6.2)
                        1–4            2.3   (1.3–3.2)        1.7   (0.8–2.5)       0.7   (0.2–1.3)      0.6   (0.2–1.0)      0.8   (0.4–1.3)
                        5–9            2.8   (1.9–3.6)        2.0   (1.3–2.8)       1.0   (0.5–1.5)      0.7   (0.3–1.1)      1.0   (0.5–1.5)
                        10–14          4.2   (1.5–6.9)        2.0   (0.7–3.4)       1.6   (0.4–2.8)      0.8   (0–1.6)        1.6   (0.3–2.8)
                        15–19          1.0   (1.2–3.1)        4.9   (0.5–10.2)      1.4   (0.5–3.3)      4.1   (0.5–7.7)      5.9   (0.5–11.2)
                          20          76.6   (44.7–108.5)    77.4   (51.3–103.5)   17.6   (11.6–23.7)   19.5   (13.4–25.5)   21.3   (15.8–26.8)
                          Total        3.6   (2.9–4.2)        3.2   (2.6–3.9)       2.0   (1.5–2.5)      2.0   (1.5- 2.4)     2.6   (2.1–3.1)
                           NOTE. Data are deaths per 100,000 varicella cases (95% CIs). The numbers of varicella cases are estimated
                        from the 1970–1994 National Health Interview Surveys (NHIS). NHIS are conducted annually and use a nationwide
                        sample of the civilian, noninstitutionalized population in the United States.
                             Case-fatality rates for adults had very wide CIs during 1970–1974 and 1975–1979 because the estimates for

                                                                                                                                                             Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
                        the number of adult varicella cases were based on responses in the NHIS and because there were few adult varicella

1990–1994, whereas secondary infections increased from 3.1%                          offered appropriately to neonates whose mothers are infected
to 13.1%, most prominently among previously healthy children.                        with varicella around the time of delivery. Alternatively, it could
                                                                                     reflect an increase in deaths in the neonatal period from con-
                                                                                     genital varicella syndrome (not identifiable on death records
                                                                                     with a specific ICD code) secondary to maternal infection dur-
   This is the first long-term analysis of deaths due to varicella                    ing the first 2 trimesters of pregnancy.
in the United States to describe mortality rates, CFR, com-                             The decline in mortality rates among children in the 1970s
plications, and preexisting conditions among both child and                          and 1980s occurred concurrently with improved medical treat-
adult decedents. This study demonstrates important changes in                        ment and survival of people with malignancies [23], effective
varicella mortality in the United States before the availability                     antiviral therapies for varicella zoster virus infections, and im-
of varicella vaccine. Most striking was the shift in the prepon-                     proved treatment of life-threatening varicella complications. In
derance of varicella deaths from children in the 1970s to adults                     addition, reported cases of Reye’s syndrome declined dramat-
in the 1990s. During the 25-year time period under study, as                         ically after publication of the association between aspirin and
deaths declined among children, who made up the majority of                          Reye’s syndrome and common viral illnesses, often influenza
cases and deaths in the 1970s, the number of adult deaths                            or varicella [24, 25].
increased 3-fold, and, consequently, the relative proportion of                         The dramatic decline in varicella CNS complications among
adult deaths increased. The increase in adult deaths is consistent                   child decedents paralleled decreases in reports to the CDC of
with data from the United Kingdom that indicate increases in                         cases of postinfectious encephalitis during the late 1970s and early
both the adult varicella mortality rate and the proportion of                        1980s [26] and of varicella-associated Reye’s syndrome cases [27]
varicella cases among adults [19, 20].                                               during the 1980s. Studies conducted in the United States in the
   Increased mortality among adults may partly reflect in-                            1970s [28] reported higher rates of encephalitis as a complication
creased migration to the United States of people from tropical                       of varicella than those conducted in the early 1990s [29]. Similarly,
countries, such as Mexico, the Philippines, China, Vietnam, and                      a study in the United Kingdom reported a marked decline in
India, where adults are more likely to be susceptible to varicella                   deaths due to encephalitis among children between 1967 and 1985
[21]. Varicella death rates were higher among adults born in                         [19]. It is plausible that Reye’s syndrome was misdiagnosed as
other countries, compared with adults born in the United                             encephalitis. A postmortem study of 32 children who died of
States. During the time when place-of-birth information was                          varicella during 1952–1977 found that 12 decedents had acute
available on the death records (1985–1994), ∼20% of all adult                        encephalopathy compatible with Reye’s syndrome, but only 2
varicella deaths occurred among foreign-born adults.                                 had definite encephalitis, which suggests that true encephalitis is
   The increase in neonatal deaths, although relatively few in                       a rare event in fatal varicella [30]. After CNS complications de-
number, contrasts with a decline in neonatal mortality in the                        clined, pneumonia became the most common lethal varicella
United Kingdom over a similar time period (1967–1990) [22].                          complication among healthy children and adults, followed by
The increase in neonatal deaths parallels the increase in adult                      secondary infections. The increase in secondary infections among
deaths, which occurred while adult CFR declined, which sug-                          child decedents may reflect a true increase in severe secondary
gests that adult varicella cases have increased in the United                        infections, especially invasive group A streptococcal infections,
States over the 25-year period. This may have resulted in more                       which have been associated with varicella [31], or it may reflect
mothers acquiring infection during pregnancy or around the                           improved laboratory diagnosis.
time of delivery. The increase in neonatal deaths in the United                         In contrast to this decline, we are unable to explain the in-
States may be due to varicella-zoster immune globulin not being                      crease in mortality that occurred among adults and children
388                                                                                Meyer et al.                                                         JID 2000;182 (August)

Table 5.    Varicella complications among varicella decedents, stratified by age group, risk status, and 5-year interval, United States, 1970–1994.
                                                          !20 years old                                                                20 years old
                                         High-risk                            Healthy                                 High-risk                                Healthy
5-year interval          Complications          All deaths, n     Complications        All deaths, n   Complications         All deaths, n   Complications          All deaths, n
Central nervous system
  1970–1974                  5   (5.8)               86             229   (64.0)           358            3   (6.8)               44             8    (11.9)              67
  1975–1979                  6   (7.6)               79             133   (59.6)           223            5   (5.4)               93            10    (13.2)              76
  1980–1984                  2   (3.4)               58              24   (20.9)           115            0   (0.0)               70             2    (2.2)               92
  1985–1989                  2   (5.4)               37              12   (8.6)            139            2   (3.6)               56             4    (2.8)              144
  1990–1994                  4   (14.8)              27              19   (8.9)            213            4   (5.6)               72             3    (1.4)              213
  1970–1974                 42   (48.8)              86              82   (22.9)           358           23   (52.3)              44            43    (64.2)              67
  1975–1979                 40   (50.6)              79              48   (21.5)           223           35   (37.6)              93            36    (47.4)              76
  1980–1984                 15   (25.9)              58              23   (20.0)           115           18   (25.7)              70            36    (39.1)              92

                                                                                                                                                                                    Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
  1985–1989                  2   (5.4)               37              27   (19.4)           139            5   (8.9)               56            45    (31.3)             144
  1990–1994                  3   (11.1)              27              42   (19.7)           213            6   (8.3)               72            54    (25.4)             213
Secondary infection
  1970–1974                  7   (8.6)               86               7   (1.9)            358            3   (6.8)               44             0    (0.0)               67
  1975–1979                  8   (10.4)              79              21   (9.3)            223            5   (5.4)               93             7    (9.1)               76
  1980–1984                  3   (5.2)               58              12   (10.4)           115            1   (1.4)               70             5    (5.4)               92
  1985–1989                  4   (10.8)              37              20   (14.4)           139            8   (14.3)              56            15    (10.4)             144
  1990–1994                  5   (18.5)              27              35   (16.4)           213            7   (9.7)               72            22    (10.3)             213
  1970–1974                  3   (3.7)               86              10   (2.8)            358            1   (2.3)               44             0    (0.0)               67
  1975–1979                  3   (3.9)               79               4   (1.8)            223            2   (2.2)               93             0    (0.0)               76
  1980–1984                  1   (1.7)               58              12   (10.4)           115            1   (1.4)               70             9    (9.8)               92
  1985–1989                  3   (8.1)               37              10   (7.2)            139            3   (5.4)               56             4    (2.8)              144
  1990–1994                  1   (3.7)               27              19   (8.9)            213            4   (5.6)               72            18    (8.5)              213
  NOTE.     Data are no. (%), except where noted.

from the mid-1980s. Although the number of death records                                   sification could result in underestimating varicella deaths if
that mentioned HIV/AIDS increased during this time, most                                   complications of varicella, such as pneumonia or group A strep-
deaths occurred among otherwise healthy children and adults.                               tococcal infections, were assigned as the underlying cause of
This suggests that among healthy people, there may be less                                 death and varicella was not listed on the death records or was
awareness of the potential severity of varicella, which could                              listed as a contributing cause of death. On the other hand,
result in delays in seeking health care, in complications not                              misclassification could overestimate varicella deaths if dissem-
being recognized, or in aggressive therapy for complications                               inated herpes zoster, especially among people who were older
not being offered as early to this group. It is also possible that                         [28, 29] and/or immunocompromised was diagnosed and re-
the decrease in deaths among high-risk people reflects health                               corded as varicella. We found some evidence of misclassification
care providers’ vigilance in administering antiviral treatment to                          in the less-pronounced seasonality among deaths in older
people with malignancies and other preexisting high-risk con-                              adults, which suggests that varicella diagnosis codes in this age
ditions. Furthermore, high-risk conditions may have existed but                            group may have a lower specificity [29]. Furthermore, it can
were not reported. For example, information collected for death                            be difficult to clinically distinguish varicella from disseminated
records is limited to medical conditions, with no information                              herpes zoster. Although there are no published studies on the
provided on immune suppressive medical therapies, such as                                  validity of the varicella code as the underlying cause of death
systemic steroid use. Therefore, this study may underestimate                              on death certificates, quality assurance of death certificate data
the number of decedents with preexisting high-risk conditions.                             is maintained by trained nosologists who code conditions at
   There are several limitations associated with using death re-                           the state level and, in turn, by nosologists at NCHS who pe-
cords. These include coding errors, antemortem diagnostic er-                              riodically review data from a sample of the submitted death
rors, inadvertent omissions, underreporting of preexisting con-                            certificates. Furthermore, we compared decedents with varicella
ditions, such as HIV infection [32], unavailability of medical                             listed as the underlying cause of death with those with varicella
records to the certifying physician, death record completion                               listed as a contributing cause of death and found similar trends
before medical workup, difficulty in determining the underlying                             in mortality over time, which suggests that the changes ob-
cause of death when several disease processes are involved, and                            served are real and not attributable to a change in coding.
misunderstanding of the certification process [33].                                         Currently, the CDC is collaborating with one large state to
   Three other limitations should be considered when inter-                                assess the accuracy of the varicella code on death certificates
preting the data we report. First, misclassification may have                               by reviewing medical records.
occurred in assigning the underlying cause of death. Misclas-                                 Second, several ICD codes for medical conditions of interest
JID 2000;182 (August)                                   Varicella Deaths in the United States, 1970–1994                                                             389

in this study lack specificity. For example, Reye’s syndrome is                       6. Centers for Disease Control and Prevention. Varicella-related deaths among
                                                                                           adults—United States 1997. MMWR Morb Mortal Wkly Rep 1997;46:
coded under the ICD-8A in a broad category, “other disease
of the brain,” and there is no specific code for varicella en-                        7. Fleisher G, Henry W, McSorley M, Arbeter A, Plotkin S. Life-threatening
cephalitis. This limited our ability to monitor varicella deaths                           complications of varicella. Am J Dis Child 1981; 135:896–9.
with these 2 CNS complications. Third, information collected                         8. Peterson CL, Vugia DJ, Meyers HB, et al. Risk factors for invasive group
for death records is limited to medical conditions, with no in-                            A streptococcal infections in children with varicella: a case-control study.
formation on immunosuppressive medical therapies, such as                                  Pediatr Infect Dis J 1996; 15:151–6.
                                                                                     9. Jackson MA, Burry VF, Olson LC. Complications of varicella requiring hos-
systemic steroid use. Therefore, this study may underestimate
                                                                                           pitalization in previously healthy children. Pediatr Infect Dis J 1992;11:
the number of decedents with preexisting high-risk conditions.                             441–5.
   It is too early to evaluate the impact of the varicella vaccine                  10. Centers for Disease Control and Prevention. Prevention of varicella: rec-
on mortality. Nevertheless, with the availability of a safe and                            ommendations of the Advisory Committee for Immunization Practices
effective vaccine, preventing varicella deaths through vaccina-                            (ACIP). MMWR Morb Mortal Wkly Rep 1996; 45(RR-11):1–36.

                                                                                                                                                                             Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014
                                                                                    11. Committee of Infectious Diseases, American Academy of Pediatrics. Recom-
tion should be a public health priority. Physicians play a key
                                                                                           mendations for the use of live attenuated varicella vaccine. Pediatrics
role by ensuring that their patients are properly vaccinated. A                            1995;95:791–6.
high proportion of vaccinated people in the population is                           12. Halloran ME, Cochi SL, Lieu TA, Wharton M, Fehrs L. Theoretical epi-
needed to achieve herd immunity for the protection of people                               demiologic and morbidity effects of routine varicella immunization of pre-
who can not be vaccinated and who remain at high risk of                                   school children in the United States. Am J Epidemiol 1994; 140:81–104.
serious disease or death, such as infants and immunocom-                            13. National Center for Health Statistics. Eighth revision international classification
                                                                                           of diseases, adapted for use in the United States. Public Health Service pub.
promised people. Coverage levels must increase substantially
                                                                                           1693. Washington, DC: US Government Printing Office, 1967.
above the 1998 national vaccine coverage of 43% among chil-                         14. World Health Organization. Manual of the international statistical classifi-
dren 19–35 months old, to reduce disease burden, including                                 cation of diseases, injuries and causes of death, based on the recommen-
mortality, among all age groups [34]. Physicians should be                                 dations of the Ninth Revision Conference, 1975. Geneva: World Health
aware that most varicella deaths occur among healthy people,                               Organization, 1977.
                                                                                    15. National Center for Health Statistics (NCHS). Vital statistics, ICD-9 ACME
that adults as well as children die from this disease, and that
                                                                                           decision tables for classifying underlying causes of death. NCHS instruc-
varicella deaths continue to occur, despite the availability of an                         tion manual, pt 2c. Hyattsville, MD: Public Health Service, published
effective vaccine [5, 6]. Physicians are encouraged to report all                          annually.
varicella-related deaths to state health departments [35]. In-                      16. US Bureau of the Census. Table 7: age and sex of the foreign-born popula-
vestigating and reporting all varicella-related deaths will pro-                           tion: 1870 to 1990. Available at
                                                                                           documentation/twps0029/tab07.html. Released 9 March 1999.
vide more accurate and complete data on people who die from
                                                                                    17. Anderson RN, Rosenberg HM. Age standardization of death rates: imple-
varicella, such as immunocompromising conditions, therapies,
                                                                                           mentation of the year 2000 standard. National vital statistics reports 47(3).
vaccination status, and childhood residence, which can be used                             Hyattsville, MD: National Center for Health Statistics, 1998.
to improve prevention efforts.                                                      18. National Center for Health Statistics. Public use data tape documentation,
                                                                                           pt I, National Health Interview Survey, 1970–1994 [machine-readable data
Acknowledgments                                                                            file and documentation]. Hyattsville, MD: National Center for Health
                                                                                           Statistics (producer). Springfield, VA: National Technical Information Ser-
   We thank Ken Kochanek, of the National Center for Health Statistics,                    vice, US Department of Commerce (distributor).
Division of Vital Statistics, for technical assistance in interpreting Inter-       19. Joseph CA. Noah ND. Epidemiology of chickenpox in England and Wales,
                                                                                           1967–1985. BMJ 1988; 296:673–6.
national Classification of Diseases codes and mortality data, and Barry
                                                                                    20. Fairley CK, Miller E. Varicella-zoster virus epidemiology: a changing scene?
Sirotkin, of the National Immunization Program, Division of Data Man-
                                                                                           J Infect Dis 1996; 174(Suppl 3):S314–9.
agement, for compiling the 25-year varicella mortality file.
                                                                                    21. US Department of Commerce, Bureau of the Census. We the American…
                                                                                           foreign born, 1993. Available at
                                                                                           we-7.pdf. Accessed 27 March 2000.
                                                                                    22. Miller E, Marshall R, Vurdien J. Epidemiology, outcome and control of
 1. Wharton M, Fehrs L, Cochi SL, et al. Health impact of varicella in the 1980s           varicella-zoster infection. Rev Med Microbiol 1993; 4:222–30.
       [abstract 1138]. In: Program and Abstracts of the 30th Interscience Con-     23. Feldman S, Lott L. Varicella in children with cancer: impact of antiviral
       ference on Antimicrobial Agents and Chemotherapy (Atlanta, GA).                     therapy and prophylaxis. Pediatrics 1987; 80:465–72.
       Washington, DC: American Society for Microbiology, 1990:276.                 24. Arrowsmith JB, Kennedy DL, Kuritsky JN, Faich GA. National patterns
 2. Preblud SR. Age-specific risk of varicella complications. Pediatrics 1981; 68:          of aspirin use and Reye syndrome reporting, United States, 1980 to 1985.
       14–7.                                                                               Pediatrics 1987; 79:858–63.
 3. Feldman S, Hughes WT, Daniel CB. Varicella in children with cancer: sev-        25. Belay ED, Bresee JS, Holman RC, Khan AS, Shahriari A, Schonberger LB.
       enty-seven cases. Pediatrics 1975; 56:388–97.                                       Reye’s syndrome in the United States from 1981 through 1997. N Engl
 4. Perronne C, Lazanas M, Leport C, et al. Varicella in patients infected with            J Med 1999; 340:1377–82.
       the human immunodeficiency virus. Arch Dermatol 1990; 126:1033–6.             26. Centers for Disease Control and Prevention. Annual summary 1982: reported
 5. Centers for Disease Control and Prevention. Varicella-related deaths among             morbidity and mortality in the United States. MMWR Morb Mortal Wkly
       children—United States, 1997. MMWR Morb Mortal Wkly Rep 1998;47:                    Rep 1983; 32:1–149.
       365–8.                                                                       27. Centers for Disease Control and Prevention. Reye syndrome surveillance:
390                                                                   Meyer et al.                                                    JID 2000;182 (August)

       United States, 1986. MMWR Morb Mortal Wkly Rep 1987; 36:689–91.               immunodeficiency virus epidemic on mortality trends in young men,
28. Preblud SR, D’Angelo LJ. Chickenpox in the United States, 1972–1977. J           United States. Am J Public Health 1990; 80:1080–6.
       Infect Dis 1979; 140:257–60.                                           33. Flanders WD. Inaccuracies of death certificate information. Epidemiology
29. Choo PW, Donahue JG, Manson JE, Platt R. The epidemiology of varicella           1992; 3:3–5.
       and its complications. J Infect Dis 1995; 172:706–12.                  34. Centers for Disease Control and Prevention. Notice to readers: national vac-
30. Takashima S, Becker LE. Neuropathology of fatal varicella. Arch Pathol           cination coverage levels among children aged 19–35 months—United
       Lab Med 1979; 103:209–13.                                                     States, 1998. MMWR Morb Mortal Wkly Rep 1999; 48:829–30.
31. Davies HD, McGeer A, Schwartz B et al. Invasive group A streptococcal     35. Centers for Disease Control and Prevention. Evaluation of varicella reporting
       infections in Ontario, Canada. N Engl J Med 1996; 335:547–54.                 to the National Notifiable Disease Surveillance System—United States,
32. Buehler JW, Devine OJ, Berkelman RL, Chevarley FM. Impact of the human           1972–1997. MMWR Morb Mortal Wkly Rep 1999; 48:55–58.

                                                                                                                                                                  Downloaded from at UNIVERSITY OF ARIZONA on April 23, 2014

To top