Danovaro-Holliday MC, LeBaron CW, Allensworth C, Raymond R, Borden TG, Murray AB, Icenogle JP, Reef SE. A Large Rubella Outbreak With Spread From the Workplace to the Community. JAMA. 2000;284(21):2733-2739. doi:10.1001/jama.284.21.2733
Author Affiliations: National Immunization Program, Centers for Disease Control and Prevention, Atlanta, Ga (Drs Danovaro-Holliday, LeBaron, and Reef); Douglas County Health Department, Omaha, Neb (Ms Allensworth); Nebraska Health and Human Services System (Dr Raymond and Mr Borden), Lincoln, Neb; National Center for Infectious Diseases, Centers for Disease Control and Prevention (Dr Icenogle and Ms Murray).
Context Childhood vaccination has reduced rubella disease to low levels in the
United States, but outbreaks continue to occur. The largest outbreak in the
past 5 years occurred in Nebraska in 1999.
Objectives To examine risk factors for disease, susceptibility of the risk population,
role of vaccine failure, and the need for new vaccination strategies in response
to the Nebraska rubella outbreak.
Design, Setting, and Patients Investigation of 83 confirmed rubella cases occurring in Douglas County,
Nebraska, between March 23 and August 24, 1999; serosurvey of 413 pregnant
women in the outbreak locale between October 1998 and March 1999 (prior to
outbreak) and April and November 1999 (during and after outbreak).
Main Outcome Measures Case characteristics, compared with that of the general county population;
area childhood rubella vaccination rates; and susceptibility among pregnant
women before vs during and after the outbreak.
Results All 83 rubella cases were unvaccinated or had unknown vaccination status
and fell into 3 groups: (1) 52 (63%) were young adults (median age, 26 years),
83% of whom were born in Latin American countries where rubella vaccination
was not routine. They were either employed in meatpacking plants or were their
household contacts. Attack rates in the plants were high (14.4 per 1000 vs
0.19 per 1000 for general county population); (2) 16 (19%), including 14 children
(9 of whom were aged <12 months) and 2 parents, were US-born and non-Hispanic,
who acquired the disease through contacts at 2 day care facilities (attack
rate, 88.1 per 1000); and (3) 15 (18%) were young adults (median age, 22 years)
whose major disease risk was residence in population-dense census tracts where
meatpacking–related cases resided (R2 = 0.343; P<.001); 87% of these persons
were born in Latin America. Among pregnant women, susceptibility rates were
13% before the outbreak and 11% during and after the outbreak. Six (25%) of
24 susceptible women tested were seropositive for rubella IgM. Rubella vaccination
rates were 90.2% for preschool children and 99.8% for school-aged children.
Conclusions A large rubella outbreak occurred among unvaccinated persons in a community
with high immunity levels. Crowded working and living conditions facilitated
transmission, but vaccine failure did not. Workplace vaccination could be
considered to prevent similar outbreaks.
Between March 23 and 26, 1999, a 29-year-old man residing in Douglas
County, Nebraska, sought treatment in 3 health care facilities for what he
thought was a sexually transmitted disease. At his third visit, he was noted
to have a maculopapular rash, low-grade fever, and lymphadenopathy. Though
no case of rubella had been reported in Douglas County in the previous 9 years,
the patient had a positive rubella IgM test result. The patient worked in
a meatpacking plant where a second rubella case was identified. Surveillance
for rash illness was enhanced, resulting in the detection of more individuals
in other meatpacking plants and in the community, almost all of whom were
young adults born in Latin American countries, where rubella vaccination only
recently has been implemented for children.
Vaccination campaigns in 7 Douglas County meatpacking plants targeted
3000 workers. Physicians and clinics enhanced immunization efforts, and the
health department provided other activities that included press, radio, and
television coverage, and distribution of information through churches and
community organizations. Despite these efforts, the outbreak continued for
4 months. Of the 125 cases reported from this outbreak, 83 occurred in Douglas
County (Omaha), 12 in other Nebraska counties, and 30 in neighboring Iowa.
Among the 83 cases in Douglas County, 14 were US-born children who attended
2 day care centers, and at least 7 pregnant women were infected. This outbreak
ultimately became the largest US rubella outbreak in the past 5 years.
Rubella is one of the most common causes of birth defects in the world,
resulting in spontaneous abortions, stillbirths, and congenital rubella syndrome
(CRS). The manifestations of CRS include hearing impairment, blindness, heart
defects, and mental retardation.1,2
According to the World Health Organization, in 1996, two thirds of the world's
population lived in countries where rubella vaccination was not practiced
routinely,3 and the number of infants with
CRS born each year worldwide was estimated to be 110,000 in 1999.4
In the United States, childhood vaccination strategies have reduced
rubella and CRS to record low levels,5,6
and in 1990, a goal was established to eliminate indigenous rubella and CRS
from the United States by the end of 2000.7
However, outbreaks have continued to occur.8
The pattern seen in Douglas County is typical of recent outbreaks: of the
14 rubella outbreaks reported to the Centers for Disease Control and Prevention
(CDC) in 1996-1998, 7 were workplace-associated and all disproportionately
affected Hispanics (median percent Hispanics, 92.5%).8
It has been assumed that high susceptibility levels to rubella in unvaccinated
populations play an important role in such outbreaks,9
but data to support this hypothesis are lacking. Furthermore, the role of
vaccine failure in the US-born population has been unclear: serosurveys of
school children have suggested that 33% of adolescents were seronegative and
might be susceptible to rubella, raising the possibility that they provide
a pool for potential outbreak spread.10,11
Mathematical models have suggested that poorly implemented childhood vaccination
strategies can actually increase CRS rates by increasing the number of susceptible
women of childbearing age12 beyond the 87%
thought to provide community-level herd immunity.13
We investigated the Douglas County rubella outbreak to help shed light
on the following issues: risk factors for disease, susceptibility levels in
a population at risk, whether failure to vaccinate children or vaccine failure
contributed to the outbreak, and whether additional strategies beyond childhood
vaccination are necessary to prevent future US rubella outbreaks.
We restricted our analysis to cases confirmed by laboratory testing
or epidemiological linkage.14 Laboratory confirmation
included a positive IgM result, isolation of rubella virus, or significant
increase in serum rubella IgG antibodies (4-fold increase or seroconversion)
from paired acute and convalescent samples. An epidemiologically linked case
was defined as a person who was exposed to a laboratory confirmed case during
the infectious period and who met the standard clinical case definition: (1)
acute onset of generalized maculopapular rash; (2) temperature greater than
99°F (37.2°C) if measured; and (3) arthritis/arthralgia or conjunctivitis
or lymphadenopathy. We analyzed only cases residing in Douglas County (where
83 of the 95 Nebraska cases occurred) between March 23, 1999 (rash onset of
the first case), and August 24, 1999 (2 incubation periods after rash onset
of the last reported case).
Case-finding efforts during the outbreak included active surveillance
for rash illness in all meatpacking and related facilities, weekly calls to
physicians and infection control practitioners in high-risk areas, weekly
calls to clinical laboratories to report any rubella serological tests requested,
and weekly contact with school nurses to investigate reports of rash illness
among the county's 64,463 schoolchildren.
Characteristics of the Cases. The Douglas County Health Department investigated each case using a
form developed in collaboration with the CDC. We obtained the following information:
birth date, country of birth, sex, race/ethnicity, vaccination status, place
of exposure, and census tract of residence. For comparison, we abstracted
census data for the general population of Douglas County.15
Childhood Vaccination Rates. We used data from the 1998 National Immunization Survey16
for Nebraska and the 1997 Douglas County school survey.17
Susceptibility Among Selected Pregnant Women. We identified 2 clinics serving a predominantly immigrant Hispanic population
in Douglas County and conducted a chart review for rubella IgG screening results
obtained from local laboratories for pregnant women between October 1998 and
March 1999 to estimate susceptibility rates before the outbreak. To estimate
the susceptibility rates when the outbreak started, serum samples obtained
for routine prenatal rubella IgG screening in the larger clinic between April
1 and November 30, 1999, were tested at the CDC. To assess risk factors for
susceptibility, women seen between June and November 1999 were given a self-administered
questionnaire in English and Spanish that sought information concerning age,
race/ethnicity, birth country, length of stay in the United States, US schooling,
and previous US deliveries of infants.
The CDC used the Trinity Biotech (Dublin, Ireland) IgM capture enzyme
immunoassay and Wampole (Cranbury, NJ) IgG enzyme-linked immunosorbent assay.
Following the manufacturer's standards, for both assays an antibody index
of 1.10 and greater (8.2 International Units [IU] for IgG) was considered
a positive result, 0.91 to 1.09 (6.6-8.1 IU) equivocal, and 0.90 and less
(6.5 IU) a negative result. Rubella susceptibility
was defined as either a negative IgG or a positive IgM result.
Data were analyzed using SAS 6.12 (SAS Institute, Cary, NC). To determine
risk factors for susceptibility among pregnant women, we used the χ2 and Fisher exact tests (bivariate) and, where more than 1 factor was
found significantly associated with susceptibility, we performed multivariate
stratified analysis and logistic regression. To determine risk factors for
rubella attack rate by census tract, we used bivariate and multivariate linear
During the first 2 months of the outbreak (March 23-May 31, 1999), 75%
(46/61) of cases were identified as workers from a meatpacking plant or their
household contacts (Figure 1). During
the last 2 months (June 1-July 27, 1999), 73% (16/22) of cases identified
were related to 2 day care centers.
The 83 confirmed cases (69 laboratory confirmed and 14 epidemiologically
linked) fell into 3 groups (Table 1).
Meatpacking Plant–Related Cases. Of the 83 cases, the meatpacking plant group included 52 cases (63%):
39 meatpacking plant workers and 13 household contacts. Unlike the general
population of Douglas County, these persons were predominantly foreign-born,
Hispanic adult males (median age, 26 years). High attack rates (14.4 per 1000;
range, 3.3-33.3 per 1000 employees) were found in affected meatpacking plants
(Table 2), compared with a countywide
attack rate of 0.19 per 1000 persons.
Day Care Center–Related Cases. Of the 83 cases, 16 were from 2 day care centers (center A and center
B). All individuals were US-born non-Hispanic whites, of whom 14 were children
(aged 5-17 months), and 2 were the parents (aged 34 and 35 years) of 2 case-children.
Prior to the first case, certain children from center A were thought to have
been exposed to an asymptomatic individual with ties to the Hispanic community
who subsequently tested positive for rubella IgM. The source-case for center
B was a child who had transferred from center A. The highest attack rates
of the outbreak were found in these centers (88.1 per 1000; range, 53.2-138.5/1000
children) (Table 2).
Community-Related Cases. These 15 (18%) of the 83 cases could not be linked to either of the
above groups, but their demographic characteristics were statistically indistinguishable
from the meatpacking group (median age, 22 years; 87% non–US-born Hispanic).
Available information regarding the 12 cases that occurred in Nebraska
outside Douglas County suggests that they had similar characteristics to those
within the county (eg, 73% Hispanic, 64% non–US-born, and 67% meatpacking
Of the 83 cases, none had a documented history of rubella vaccination:
57 (69%) were known not to have received the rubella vaccine, and 26 (31%)
had unknown vaccination status. Of the 26 persons with unknown vaccination
status, 19 (73%) were born in countries where rubella immunization was not
routine, 5 (19%) began school in the United States before the rubella vaccine
was licensed, and 2 (8%) began school in the United States before school laws
for rubella vaccination were enforced. A state law required rubella vaccination
of children in day care centers, but in each center, all cases occurred in
1 classroom where children at or younger than the minimum age of vaccination
were grouped. Of the 14 children in the day care group, 9 (64%) were younger
than the minimum age of vaccination (<12 months), 3 (21%) were age-eligible
but not overdue (12-15 months) for vaccination, and 2 (14%) were overdue (16
and 17 months). Of the 159 children in the 2 day care centers, no vaccinees
acquired symptomatic disease.
Vaccination rates (95% confidence interval) for 1 dose of rubella vaccine
were 90.2% (3.5%) for children aged 19 to 35 months in Nebraska, according
to the 1998 National Immunization Survey, and 99.8% for the Douglas County
school population according to the 1997/1998 Douglas County school survey.
Prior to the outbreak, a second dose of rubella vaccine was recommended for
all Nebraska seventh graders, but the proportion who received it is unknown
(after the outbreak, a second dose was required, although no cases were documented
among school children).
Rates. In the retrospective chart review (October 1998-March 1999), rubella
IgG results were available for 197 (97.5%) of 202 medical charts. Of these,
26 patients (13%) were reported to be susceptible to rubella. In the prospective
study (April-November 1999), 216 (96.8%) of the 223 blood samples tested at
the CDC provided nonequivocal IgG and IgM results. Of these, 24 (11.1%) showed
susceptibility based on a 6.5-IU threshold. When the laboratory threshold
was varied, these susceptibility rates changed but were not high: 11% at 5
IU; 13% at 10 IU; and 17% at 15 IU. Susceptibility rates before (13%) and
after (11%) the outbreak onset did not differ significantly (P = .52).
Risk Factors. Of the 167 pregnant women from whom serum samples were obtained between
June and November 1999, 96 (58%) provided complete information on the questionnaire,
had nonequivocal laboratory results, and were included in the analysis (Table 3). Of the 96 women analyzed, 95
(99%) were Hispanic, 89 (93%) non–US-born, and their median age was
25 years (range, 15-39 years). They did not differ significantly from the
71 nonincluded (P≥.38) women in terms of susceptibility,
age, ethnicity, US-born status, US school attendance, or delivery of an infant
in the United States. Mother's young age and short stay in the United States
were risk factors in this population.
Infection. Of the 24 susceptible women tested, 6 (25%) had positive rubella IgM
results. None reported symptoms. All women shared the characteristics of the
rest of the pregnant women studied (Hispanic 100%, non–US born 100%,
age range, 18-23 years), and none had documentation of prior vaccination.
In March 2000, an infant with CRS (deafness, thrombocytopenia, and heart defect)
was born to one of the surveyed women who had tested IgM equivocal and had
had a rash during her first trimester. Ongoing surveillance for CRS in the
area should ultimately document the final number of CRS infants born as a
result of this outbreak.
Of the 107 census tracts in Douglas County, cases occurred in 35 (33%):
meatpacking plant–related cases in 26, community-related cases in 9,
and day care center–related in 8 (Figure
2). Tract-specific attack rates ranged from 0 to 6 per 1000 inhabitants.
In bivariate analysis, attack rate by census tract was significantly
associated with percent Hispanic (R2 =
0.279; P<.001), percent non–US born (R2 = 0.187; P<.001),
and population density (R2 = 0.118; P<.001) but not with percentage below poverty level
(R2 = 0.032; P
= .06). In multivariate analysis, percentage non–US born lost significance
when paired with percent Hispanic. The fit of a model, including percent Hispanic
and population density (R2 = 0.356; P<.001), improved when the interaction between the 2
terms was included (R2 = 0.505; P<.001). The pattern for meatpacking plant–related
cases followed that of the overall outbreak. Two thirds (10/15) of the community-related
cases resided in tracts where meatpacking plant–related cases resided
(attack rate, 0.11 per 1000), and a model combining the presence of a meatpacking
case and high population density best fit the community-related case pattern
(R2 = 0.343; P<.001).
Day care center–related cases showed none of the above patterns, and
only a weak correlation was found with residence in tracts with high poverty
rates (R2 = 0.043; P<.001).
In summary, we found that the largest rubella outbreak in the United
States in 5 years began in crowded working environments among adults born
in Latin American countries where rubella vaccination only recently has been
implemented. Disease spread to other members of the Hispanic community facilitated
by high population density. The force of transmission was such that a high
proportion of susceptible pregnant women attending prenatal clinics in the
outbreak area were infected, despite preoutbreak immunity levels among these
women (87%) which had been thought to protect against community-based rubella
transmission.13 Ultimately, the disease found
its way into day care centers distant from the outbreak, where it infected
US–born, non-Hispanic children, most of whom were younger than the minimum
age of vaccination. Fortunately, high county-wide vaccination levels among
children and adults vaccinated as children limited outbreak spread beyond
unvaccinated persons at risk, and no cases were documented among vaccinees.
These data suggest that neither vaccine failure nor failure to implement
current child-based vaccination strategies was responsible for the outbreak.
Rather, the findings imply that to prevent such outbreaks new approaches must
be found to achieve high vaccination levels among adults missed by the US
school-based strategy, particularly non–US-born adults working and living
in crowded conditions where the efficiency of rubella transmission appears
to be high.
Workplace-based rubella outbreaks, such as this one in Douglas County,
involving persons born in countries where rubella vaccination is not routine,
have become increasingly common.8,9
Concentration of large numbers of unvaccinated adults indoors may provide
a far more favorable environment for the airborne transmission of a disease
like rubella18 than farm fields or other outdoor
working conditions. Among military recruits, barracks conditions have been
shown to overcome high levels of immunity and produce outbreaks when disease
is introduced.19 Since meatpacking plant employees
frequently report that they travel to and from countries, such as Mexico (country
of birth of 62% of Douglas County cases) where approximately 40,000 rubella
cases annually have been reported between 1994 and 1998,20
the possibility of introducing disease is increased.
High exposure rates among Hispanics may explain why Hispanic ethnicity,
rather than birth outside the United States, was a primary risk factor for
disease incidence in our geographical analysis. It also may explain why the
proportion of US rubella cases of Hispanic ethnicity has risen to 83% (1998).
Additionally, many non–US-born Hispanics may not be counted in census
surveys for immigration reasons, and the presence of these persons may further
facilitate disease transmission through increased population density. High
population density has been shown to facilitate community spread of measles21 but previously has not been identified as a risk
factor for community rubella outbreaks.
However, it is important to note that the attack rate in the community
was 131 times lower than in the workplace (0.11 per 1000 inhabitants vs 14.4
per 1000 workers), suggesting that population immunity levels were sufficient
to contain what could have become a far larger outbreak. Moreover, the force
of transmission was likely lower outside the Hispanic community. If we impute
to Douglas County the age-specific seropositivity patterns found in the US
general population,22 approximately 40,000
persons (13%) were rubella seronegative but fewer than 100 were identified
as rubella cases. Disease did spread beyond the Hispanic community but only
to 2 day care classrooms where children younger than the minimum age of vaccination
The data from Douglas County do not support the hypothesis that secondary
vaccine failure provides a large pool of susceptible individuals that can
contribute to outbreak transmission. While proving a negative is inherently
difficult, surveillance for rash illness was aggressive for the 64,463 schoolchildren
during the outbreak period, so that it seems unlikely that many cases, much
less a widespread outbreak, would have been missed. Failure to vaccinate susceptible
individuals rather than vaccine failure also was suggested from outbreak investigations
in the early 1990s by Lindegren et al.23 Most
Douglas County cases occurred among unvaccinated young adults, suggesting
that the school population was spared, not because of their age, but because
of protective levels of immunity.
The problem of rubella transmission in US workplaces is likely to grow
because certain industries increasingly depend on non–US-born workers,24 and 67% of the 1996 world's population lived in countries
where rubella vaccination was not routine.3
To prevent future rubella outbreaks, our data suggest that childhood
immunization strategies alone may not be enough, and that workplace vaccination
of high-risk adults needs to be considered. The Advisory Committee on Immunization
Practices only recommends rubella vaccination in health care settings.6,25 However, industries employing many
non–US-born individuals, such as meatpacking plants, could help prevent
rubella outbreaks—and the disruption of their own operations—by
requiring or encouraging rubella vaccination as a condition of employment.
Based on the Douglas County data, work-based strategies are most likely
to reach men, which will reduce, but not prevent, disease among women. The
Special Supplemental Food Program for Women, Infants and Children (WIC) comprises
44% of the pregnant women in the US, and vaccination initiatives through WIC
have been successful in vaccinating children,26,27
but greater efforts are needed to include women. The role of missed opportunities
in health care settings should not be ignored: it has been estimated that
62% of mothers of children with CRS had at least 1 missed opportunity for
rubella vaccination, 81% of which were postpartum.28
This is consistent with Douglas County data: 93% of the women in our survey
were born outside the US, but almost half (44%) had previous US deliveries.
Since containing 1 case of rubella costs about $5000 (P. Kramaz, MD, et al,
oral presentation, 33rd National Immunization Conference, Dallas, Tex, June
1999) and the lifetime cost of 1 child with severely disabling CRS may run
into millions, the savings to society can be great for routine rubella vaccination
of all susceptible women of childbearing age.
As a descriptive investigation, our findings lack the degree of certainty
associated with prospective or case-control studies. We were not able to document
any vaccine failures, but almost all our cases were adults and most adults
lack documentation of childhood vaccination even when it has been received.
Susceptibility rates were low among pregnant women and did not differ before
and after the outbreak. Antibody levels obtained by chart review may have
not been comparable to those determined at the CDC since different tests might
have been used. Additionally, these rates may not have been representative
of workers in meatpacking plants where most cases occurred. However, equally
low susceptibility rates were found among meatpacking workers in Kansas and
among pregnant women in a similar outbreak in Arkansas (trip report, P. Kramaz,
MD, 1998, and M. Hladik, MD, 1999, written communication). The pregnant women
in our serosurvey did acquire rubella disease, suggesting that they were indeed
part of the risk population whose susceptibility we wished to survey. Our
finding that population density facilitated transmission was based on census
tract correlations, and cases were compared ecologically to the general population;
thus, these results should be interpreted with caution.
Despite these limitations, we believe that the Douglas County data demonstrate
that rubella outbreaks can occur in communities where unvaccinated individuals
are concentrated in workplace or other environments where rubella is introduced.
New vaccination strategies targeting high-risk adults will be needed if such
outbreaks are to be prevented in the future.