PEDIATRICS Vol. 105 No. 1 January 2000, pp. 84-88

Postexposure Effectiveness of Varicella Vaccine

Barbara Watson, MBChB^*, , Jane Seward, MBBS, MPH^§, Alice Yang, RN^*, Patricia Witte, RN^*, James Lutz, MPA^*, Christopher Chan^*, Sandra Orlin, MSN, CRNP, and Robert Levenson, MBA^*

From the ^* Department of Public Health Philadelphia, Philadelphia, Pennsylvania; the  Albert Einstein Medical Center, Philadelphia; the ^§ Philadelphia Health Management Corporation; and the  Centers for Disease Control and Prevention, National Immunization Program, Atlanta, Georgia.

	ABSTRACT

Top Abstract Methods Results Discussion References

Objective.  1) To describe the postexposure effectiveness of varicella vaccine in a homeless shelter; and 2) to demonstrate an effective public health intervention and its implications.

Design.  A prospective observational study.

Setting.  A women and children's shelter in Philadelphia with 2 cases of varicella before intervention.

Outcome Measures.  Varicella in vaccinated and unvaccinated shelter residents; vaccine effectiveness for prevention of varicella when administered after exposure among children <13 years of age.

Results.  Sixty-seven shelter residents received varicella vaccine after exposure, including 42 children <13 years of age. One child who was unvaccinated developed varicella, but no vaccinated child developed typical disease. Vaccine effectiveness was 95.2% (95% CI, 81.6%-98.8%) for prevention of any disease and 100% for prevention of moderate or severe disease among the children <13 years of age.

Conclusion.  When used within 36 hours after exposure to varicella in a setting where close contact occurred, varicella vaccine was highly effective in preventing further disease. This study provides support for the recent recommendation by the Advisory Committee on Immunization Practices to administer varicella vaccine after exposure: this practice should minimize the number of moderate or severe cases of disease and prevent prolonged outbreaks.  Key words:  varicella, varicella vaccine, outbreak, postexposure, vaccine effectiveness.

Varicella-zoster virus (VZV) vaccine (Varivax, Merck, West Point, PA) was licensed in March 1995 for use in the United States among susceptible healthy persons aged 12 months or older. In May 1999, the Advisory Committee on Immunization Practices (ACIP)¹ updated its varicella vaccine recommendations to include postexposure vaccination and outbreak control. This recommendation was based on >20 years of research. In 1977, Asano et al² hypothesized that whereas natural varicella infection may require 5 to 7 days for the wild virus to propagate in the respiratory tract before primary viremia and dissemination occur, vaccine virus may elicit humoral and cellular immunity in less time, thus providing protection even if administered after exposure.

From 1975 to 1986, clinical trials using different formulations of the Oka/Merck vaccine in the United States³ and the Oka/Biken vaccine in Japan^24-11 demonstrated that the vaccines were 67% to 100% effective in preventing disease when administered after exposure. These studies examined dose and time as variables and, in general, found that doses of at least 1000 plaque-forming units (pfus) were effective if administered within 3 days. Four studies suggested the window of opportunity for postexposure vaccination may be as long as 5 days.^39-11

The most convincing data are from controlled (placebo group or unvaccinated control group) and uncontrolled (using historical controls) clinical trials using children exposed within households. Arbeter et al,³ Asano et al,^2,6 and Naganuma et al¹¹ demonstrated that the vaccine administered within 3 days prevented 75% to 100% cases after household exposure where, in the absence of an intervention, ~87% of exposed susceptible siblings are expected to develop disease.¹² Although Arbeter et al³ used an Oka/Merck vaccine formulation with a higher minimum dose (4350 pfus) than the United States licensed Oka/Merck vaccine (>1350 pfus), Asano⁶ and Naganuma¹¹ using Oka/Biken vaccines demonstrated effectiveness using doses lower than the currently licensed Oka/Merck United States vaccine. Most recently, in 1998, a small household study of postexposure administration of the currently licensed United States vaccine suggested that the vaccine is highly effective in preventing moderate and severe disease when administered at the standard dose within 3 days of exposure.¹³

On January 15, 1998, the director of a Philadelphia homeless shelter for women and children contacted the Division of Disease Control, Philadelphia Department of Public Health (PDPH), concerning 2 varicella cases among residents, a 27-year-old mother and her 11-month-old son. Both cases were prescribed acyclovir at the emergency room where they were diagnosed; however, they did not fill the prescription because of cost. Many young children who were likely to be susceptible to varicella resided at the shelter, and concern focused on the possibility of a prolonged outbreak that would require closing the shelter to new admissions for several months. To minimize the number of new cases, the Division of Disease Control staff offered varicella vaccine to all susceptible persons in the shelter. We report the use and effectiveness of varicella vaccine administered 36 hours following exposure to prevent an outbreak in a densely inhabited building.

	METHODS

Top Abstract Methods Results Discussion References

Philadelphia is 1 of 3 study sites participating in the Varicella Active Surveillance Project in collaboration with the Centers for Disease Control and Prevention. In October 1994, the Board of Health, PDPH, issued a regulation requiring reporting of varicella and herpes zoster in Philadelphia. As part of the Varicella Active Surveillance Project, active surveillance was established in an area of West Philadelphia; reporting sources included schools, day care centers, universities, health-care providers (including private physicians, public clinics, hospitals, and emergency rooms) and sites in the Special Supplemental Nutrition Program for Women, Infants, and Children. To identify other cases among residents of the West Philadelphia surveillance area, all homeless shelters in Philadelphia were notified of the project. Disease surveillance investigators perform surveillance in specific areas of the city for all reportable conditions with regularly scheduled phone calls or visits to sentinel sites and laboratories. As a sentinel-reporting site, this shelter usually promptly notifies PDPH about vaccine preventable and other reportable diseases that occur there.

Study Population and Setting

This shelter serves women and children and is 1 of 2 long-term facilities in Philadelphia with a drug and alcohol rehabilitation program. All residents are on medical assistance and the children are eligible for Special Supplemental Nutrition Program for Women, Infants, and Children. The shelter has rooms for the residents on 3 floors; each room is ~100 square feet and houses a family unit of up to 6 people. Residents share bathrooms and a common room ~700 square feet for training, eating, and recreation. During the day, the women attend drug and alcohol rehabilitation classes in the common room. Women participating in the "Back to Work" program go to their work site. School-aged children attend school. Preschool-aged children who are not enrolled in a preschool program remain at the shelter where they are attended by their mothers and the shelter staff; during this study, most preschoolers remained at the shelter, because the child care provider for these children was closed at the time. During the weekend, residents remain in the shelter. On January 15, 1998, the shelter was at full capacity with 154 residents (83 children, 35 mothers, and 36 women without children).

Index Cases

Two varicella co-index cases had rash onset on the evening of January 14, 1998. Case A was a 27-year-old mother and case B was her 11-month-old son. The cases were reported to the PDPH on January 15, 1998. The mother had >500 macular-papular-vesicular lesions, which lasted for 10 days. She had an associated severe headache and a cough, which persisted for 5 days. Her 11-month-son had cough, coryza, 300 to 500 macular-papular-vesicular lesions, and fever of 101°F for 3 days. His vesicular lesions were present for 14 days.

Study Variables

Exposure Because all residents shared the common eating and recreation room, and cases A and B were present in this area during the 2 days before rash onset and immediately after, we assumed all residents were exposed to varicella. We used appearance of rash in the index cases as the onset time of exposure, although persons with varicella may be contagious 2 days before rash onset.

Susceptibility to Varicella and Vaccine Coverage

We interviewed all adult residents of the shelter to identify persons susceptible to varicella based on history and varicella vaccination status. We confirmed vaccination status for all shelter residents on January 15, 1998, by reviewing immunization records held at the center, by checking the PDPH's Kids Immunization Database/Tracking System immunization registry, and by telephoning primary health-care providers named by the mothers. We obtained history of disease on January 16, 1998, from self-reports or, for children, from their mothers. One mother of a 15-month-old child reported that the child had varicella at 3 months of age. Because varicella is uncommon at that age, we reviewed hospital records to verify this case.

All unvaccinated persons with negative or unknown varicella history status were considered susceptible. Although a specific postexposure recommendation was not in force at the time, this practice was, nonetheless, consistent with varicella vaccination recommendations of both the ACIP¹⁴ and the American Academy of Pediatrics.¹⁵ Serologic testing for VZV immunity was not feasible in this setting. From previous experience, venous access in this population is difficult because of past drug abuse. Additionally, there is a high level of concern that blood will be tested for human immunodeficiency virus.

We calculated vaccine coverage before the intervention for children >12 months of age by dividing the number of susceptible children who had been vaccinated by the number of children eligible for vaccination.

Case Definition and Disease Severity

We defined a case of varicella as an illness with acute onset of diffuse (generalized) maculo-papulo-vesicular rash without other apparent cause. We classified severity of disease according to number of skin lesions. Disease was classified as mild if there were <50 skin lesions, moderate if there were between 50 and 500 skin lesions, and severe if there were >500 lesions. We considered varicella-like rash that occurred within 42 days after vaccination to be attributable to wild VZV; we did not attempt virus strain identification on the individuals who developed rash after vaccination.

Outbreak Control Measures

Cases A and B were confined in their single room along with the 3 susceptible children in the family who were aged 4, 5, and 6 years in an attempt to limit direct contact with other susceptible shelter residents. Complete isolation was not possible because the infected persons used the common bathroom facilities. Quarantine continued until lesions had crusted; for 10 and 14 days for the mother and the 11-month-old child, respectively. Residents who reported a history of varicella delivered meals.

Vaccination

On January 16, 1998, ~36 hours after rash onset in the index cases, we offered varicella vaccinations at no cost to all susceptible shelter residents. Vaccine information statements were provided.¹⁶ In addition to varicella, we offered other needed age-appropriate vaccinations to all shelter residents based on a review of immunization records.

Surveillance

In accordance with standard public health policy for control of outbreaks of contagious diseases, the PDPH closed the shelter to new admissions on January 15, 1998, and established surveillance for rash, febrile illnesses, and vaccine adverse events with daily visits by the disease surveillance investigators. Surveillance continued until 42 days (2 incubation periods) after onset of the last case. All individuals (vaccinated and those with history of disease) remained at the shelter for 42 days. No new susceptibles were admitted to the shelter during this time.

Vaccine Effectiveness (VE)

Because adults with a negative or uncertain history of disease are 70% to 90% likely to be immune when tested by serology,^17-18 we restricted our analysis of postexposure effectiveness to children 12 months through 12 years of age. VE was calculated as: [where VE = vaccine effectiveness; AR_U = attack rate in unvaccinated; AR_V = attack rate in vaccinated^19,20]. First we calculated VE for prevention of all disease and for prevention of moderate and severe disease using data from this study. Because only 1 child remained unvaccinated, we also calculated VE using a historical secondary attack rate of 87% among unvaccinated siblings in a household setting.¹² We calculated a 95% confidence interval around the VE point estimate using Katz's method (a Taylor series approximation to the variance of a relative risk).²⁰

	RESULTS

Top Abstract Methods Results Discussion References

Susceptibility and Vaccine Coverage

Of the 83 children in the center, 52 (62.7%) were susceptible to varicella based on lack of disease history and/or no record of varicella vaccine (Table 1). This included 1 child with a reported history of varicella at 3 months of age who was subsequently reclassified as susceptible based on review of hospital emergency department records (see below). Susceptibility rates declined from 100% among infants <1 year of age to 25.0% for children aged 10 to 12 years. Coverage rates for varicella vaccination were highest among children aged 1 to 4 years (25%). Only 1 of 23 children aged >5 years who were susceptible by history had been vaccinated (4.4%).

Postexposure Vaccination

A total of 67 individuals (25 women and 42 children) were vaccinated with varicella vaccine ~36 hours after exposure. Of the 52 susceptible children, 10 were not vaccinated; 9 of these were <12 months of age and thus not eligible for vaccination. The other was initially incorrectly classified as not susceptible based on the mother's report that the child had varicella at 3 months of age. Other children not vaccinated included 23 with a history of varicella and 8 who had been previously vaccinated. Adults not vaccinated included 34 women who reported a history of varicella and 13 women who were either not present or refused vaccination. In addition to varicella vaccine, a total of 79 other vaccinations were administered to persons identified as not up to date; these included 33 measles-mumps-rubella and 24 hepatitis B vaccinations.

Surveillance for Illness and Shelter Closure

Vaccinated Children We identified 2 children with varicella-like rash with onset on January 28, 1998. Both children were sons of case A. Both had rash consisting of 15 lesions (3 macules, 10 papules, and 2 vesicles and 4 macules, 9 papules, and 2 vesicles, respectively) with no fever or malaise; their lesions lasted 3 days.

Unvaccinated Children

We identified 1 case of varicella in a 15-month-old child who had not been vaccinated. The family consisted of a mother, a 1-month-old infant, and the 15-month-old son. The mother reported that the 15-month-old had varicella at age 3 months. Records from his emergency department visit were requested on January 16, 1998, and subsequent review of the records did not confirm the mother's reported history. By the time this information was available, the family had left the shelter because the 1-month-old infant was having apnea spells; he was admitted to a pediatric hospital where the family also stayed from January 25, 1998, to February 3, 1998. The hospital was notified that the 2 children had been exposed to varicella and that the 15-month-old child was susceptible. On returning to the shelter on February 4, 1998, the 15-month-old child developed a fever of 103°F, cough, and irritability that lasted 5 days, and 400 lesions that lasted for 14 days.

None of the 13 adults or the 23 children who gave a history of disease nor any of the 9 children <12 months of age who were ineligible for vaccination developed varicella. No additional cases of varicella were identified between February 5, 1998, and February 26, 1998. Although surveillance continued until March 18, 1998, the shelter reopened on February 27, 1998, because there were no susceptible residents in the shelter. No adverse reactions to vaccination were reported.

Vaccine Effectiveness

Using the attack rate of 100% in the 1 unvaccinated child, the vaccine was 95.2% (95% CI, 81.6%-98.8%) effective in preventing all disease and 100% effective in preventing moderate and severe disease. Using a historical attack rate of 87% among the unvaccinated, VE was 94.5% for any disease and 100% for moderate and severe disease.

	DISCUSSION

Top Abstract Methods Results Discussion References

In this population of shelter residents, varicella vaccine given 36 hours after exposure was highly effective in preventing disease; no cases of moderate or severe disease occurred among vaccinated children. Both of the cases in children who received vaccine were mild, and both occurred in children of case A, who certainly had the most intense exposure. The other child of case A did not contract varicella. Although none of the 9 susceptible infants <12 months of age who were not eligible for vaccination developed varicella, 6 of these infants were <3 months of age and were therefore likely to have had passive immunity from maternal antibody.²¹ Previous studies have shown a correlation between effectiveness and early administration of vaccine after exposure^2,3; in this study, administration of vaccine <2 days after exposure may have contributed to its effectiveness.

The recent study by Salzman and Garcia¹³ of 10 siblings who received the currently licensed vaccine within 3 days of exposure demonstrated that the vaccine prevented disease in 5 out of 10 of the vaccinated children; moreover, the vaccine prevented moderate or severe disease among 9 of the 10 children, assuming that all rash that occurred in siblings was attributable to wild virus. The disease described among the siblings who developed rash was very similar to the 2 cases seen in our study, with 4 children having 20 lesions or less and 1 child having 83 lesions.

Four limitations of our study should be considered when evaluating the VE rate we report. First, the VE calculation is based on only 1 child being unvaccinated; for this reason we provided a second method for calculating VE. This method used historical controls and we assumed that transmission in this densely inhabited shelter would be similar to transmission in a household setting. Second, the effectiveness rate reported here (as well as the rates reported by Salzman and Garcia¹³) may be overestimated, because all unvaccinated children with a negative history of clinical varicella were assumed to be susceptible and were vaccinated; the actual rate of seronegativity among the children was not assessed. The effect, however, is likely to be small. Previous studies have demonstrated that the majority of unvaccinated children with a negative or uncertain history of varicella are, in fact, susceptible. One study showed that serologic evidence of immunity among children aged 7 to 12 years with a negative or uncertain history of varicella ranged from 38% for 12-year-olds to 13% for 7-year-olds.²² Attack rates in a study of the administration of immune globulin to control varicella within a household showed that only ~4% of children susceptible by history did not contract disease.¹² In contrast, adults with negative or uncertain disease histories are 70% to 90% likely to be immune.^17,18 To examine the effect of this assumption, we calculated an adjusted effectiveness rate for this population by reducing the number of susceptible, vaccinated children. Estimating immunity levels among children with negative or uncertain history at 20% for children aged 1 to 9 years and 50% for those aged 10 to 12 years, the adjusted VE rate is 94.0%.

Third, viral strain identification was not performed in our study, and thus, we could not determine if rashes in the vaccinated children were because of the vaccine or the wild virus. However, because any rash after vaccination was assumed to be wild virus, this led to underestimating, not overestimating, VE. It is not possible to differentiate vaccine from wild type rash based on clinical examination. The rashes seen during follow-up were similar to the postvaccination rashes described among children within 2 to 4 weeks of vaccination^23,24 and to breakthrough disease from wild virus that may occur after vaccination.^25,26

A fourth limitation is common to VE studies in general; we assumed that all children in the population were exposed to varicella. Although this shelter was considered to be similar to a large household in terms of likelihood of exposure, it is possible that not all residents were exposed to the 2 index cases. This assumption would bias toward higher VE estimates.

Although varicella vaccine was highly effective in preventing varicella among susceptible persons, this intervention had another dramatic, positive consequenceminimizing the time during which varicella cases occurred in the homeless shelter. Because PDPH standard procedure is to close a shelter to new admissions from the time a varicella case is identified, and to remain closed for 3 weeks after the last case is identified, a prolonged outbreak of disease affects not only current residents, but also others who need assistance. Moreover, because this closure occurred in winter, with temperatures below freezing, the need to reopen the shelter as soon as possible was amplified. In this outbreak, the shelter was closed for only 6 weeks. In contrast, a varicella outbreak in another Philadelphia homeless shelter that used standard control measures without postexposure vaccination resulted in 63 varicella cases and shelter closure for 6 months.

Our finding supports that of Salzman and Garcia¹³; both studies demonstrate that postexposure use of the currently licensed varicella vaccine is highly effective in preventing or modifying varicella when administered after exposure. Further study of the use of this licensed vaccine to prevent disease in postexposure settings within households will provide additional information on effectiveness at intervals of >36 hours and will define the outer limits of efficacy. Data are limited on administration of vaccine 4 days after exposure. Based on our findings, we anticipate that the vaccine will be effective in prevention of varicella among exposed susceptible children after household exposures and other close exposures, especially if vaccine is offered soon after identification of varicella case(s). As recently recommended by the ACIP, use of varicella vaccine after exposure and for outbreak control should limit the spread of varicella, prevent its associated complications and dramatically shorten outbreaks.

	ACKNOWLEDGMENTS

We thank Mary McCauley for editorial assistance in the preparation of this manuscript and the Office for Emergency Shelter Services for rapid reporting of varicella cases and assistance and support for outbreak control activities.

	FOOTNOTES

Received for publication May 27, 1999; accepted Aug 3, 1999.

Reprint requests to (B.W.) Philadelphia Department of Public Health, 500 S Broad St, Philadelphia, PA 19146. E-mail: barbara.watson{at}phila.gov

	ABBREVIATIONS

VZV, varicella-zoster virus; ACIP, Advisory Committee on Immunization Practices; pfu, plaque-forming unit; PDPH, Philadelphia Department of Public Health; VE, vaccine effectiveness.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Centers for Disease Control and Prevention. Prevention of varicella. Update recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep. 1999;48(RR-6)
2. Asano Y, Nakayama H, Yazaki T, Protection against varicella in family contacts by immediate inoculation with varicella vaccine. Pediatrics. 1977; 59:3-7 [Medline]
3. Arbeter AM, Starr SE, Plotkin SA Varicella vaccine studies in healthy children and adults. Pediatrics. 1986; 78:748-756 [Abstract/Free Full Text]
4. Asano Y, Yazaki T, Miyata T, Application of a live attenuated varicella vaccine to hospitalized children and its protective effect on spread of varicella infection. Biken J. 1975; 18:35-40 [Medline]
5. Asano Y, Nakayama H, Yazaki T, Protective efficacy of vaccination in children in four episodes of natural varicella and zoster in the ward. Pediatrics. 1977; 59:8-12 [Medline]
6. Asano Y, Hirose S, Iwasa S, Protective effect of immediate inoculation of a live varicella vaccine in household contacts in relation to the viral dose and interval between exposure and vaccination. Biken J. 1982; 25:43-45 [Medline]
7. Takahashi M, Otsuka T, Okuno Y Live vaccine used to prevent the spread of varicella in children in hospital. Lancet. 1974; 2:1288-1290 [Medline]
8. Asano Y, Iwayama S, Miyata T, Spread of varicella in hospitalized children having no direct contact with an indicator zoster case and its prevention by a live vaccine. Biken J. 1980; 23:157-161 [Medline]
9. Katsushima N, Yazaki N, Sakamoto M Application of a live varicella vaccine to hospitalized children and its follow-up study. Biken J. 1982; 25:29-42 [Medline]
10. Katsushima N, Yazaki N, Sakamoto M Effect and follow-up on varicella vaccine. Biken J. 1984; 27:51-58 [Medline]
11. Naganuma Y, Osawa S, Takahashi R Clinical application of a live attenuated varicella vaccine (Oka strain) in a hospital. Biken J. 1984; 27:59-61 [Medline]
12. Ross AH Modification of chickenpox in family contacts by administration of gamma globulin. N Engl J Med. 1962; 267:369-376
13. Salzman MB, Garcia C Postexposure varicella vaccination in siblings of children with active varicella. Pediatr Infect Dis J. 1998; 17:256-257 [CrossRef][Medline]
14. Centers for Disease Control and Prevention. Prevention of varicella: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep. 1996;45(RR-11)
15. American Academy of Pediatrics, Committee on Infectious Diseases Recommendations for the use of live attenuated varicella vaccine. Pediatrics. 1995; 95:791-796 [Abstract/Free Full Text]
16. Centers for Disease Control and Prevention. Chickenpox Vaccine: What You Need To Know Before You or Your Child Gets the Vaccine. Vaccine Information Statement (interim). Atlanta, GA: Centers for Disease Control and Prevention; 1996
17. Alter SJ, Hammond JA, McVey CJ, Susceptibility to varicella-zoster virus among adults at high risk for exposure. Am J Infect Control. 1986; 7:448-451
18. McKinney WP, Horowitz MM, Battiola RJ Susceptibility of hospital-based health care personnel to varicella-zoster virus infections. Am J Infect Control. 1989; 17:26-30 [CrossRef][Medline]
19. Orenstein WA, Bernier RH, Dondero TJ, Field evaluation of vaccine efficacy. Bull World Health Organ. 1985; 63:1055-1068 [Medline]
20. Kahn H, Sempos C. Statistical Methods in Epidemiology. New York, NY: Oxford University Press; 1989
21. Gershon AA, Raker R, Steinberg, S, et al Antibodies to varicella-zoster virus in parturient women and their offspring during the first year of life. Pediatrics. 1976; 58:692-696 [Abstract/Free Full Text]
22. Lieu TA, Black SB, Takahashi H, Varicella serology among school age children with a negative or uncertain history of chickenpox. Pediatr Infect Dis J. 1998; 17:120-129 [CrossRef][Medline]
23. White CJ, Kuter BJ, Hildebrand CS, Varicella vaccine (VARIVAX) in healthy children and adolescents: results from 1987-1989 clinical trials. Pediatrics. 1991; 87:604-610 [Abstract/Free Full Text]
24. Watson B, Foster JA Appropriate use of varicella vaccine. Clin Immunother. 1995; 4:197-206
25. Bernstein HH, Rothstein EP, Pennridge Pediatrics Associates, et al Clinical survey of natural varicella compared with breakthrough varicella after immunization with live attenuated Oka/Merck varicella vaccine. Pediatrics. 1993; 92:833-837 [Abstract/Free Full Text]
26. Watson BM, Piercy SA, Plotkin SA, Starr SE Modified chickenpox in children immunized with the Oka/Merck varicella vaccine. Pediatrics. 1993; 91:12-22