Routine childhood varicella vaccination, introduced in the United States in 1995, has reduced varicella morbidity and mortality by 66% to 88%.1–3 Despite decreased disease incidence, outbreaks continue to occur, and increasingly involve high-vaccinated populations.4–7 In vaccinated persons, breakthrough varicella is often mild—characterized by a mostly maculopapular rash with few or no vesicles and fewer than 50 lesions, which resolve in less time than typical varicella vesicular lesions.8 Varicella in persons vaccinated with 1 dose of vaccine has been shown to be contagious, although a vaccinated person with fewer than 50 lesions is only one-third as likely to transmit the disease as an unvaccinated person.9
In June 2005, the Advisory Committee on Immunization Practices (ACIP) issued a recommendation for a second dose of varicella vaccine during outbreaks. In June of 2006, the ACIP expanded this recommendation to include a routine second dose of varicella vaccine for children at age of school-entry and a catch-up second dose for all ages.10 A key rationale for the second dose recommendation was the need to address continued outbreaks of varicella and the burden these outbreaks place on children, schools, families, public health departments, and the medical system. As of the 2008–2009 school season, 12 states require a second dose of varicella vaccine, and 9 additional states are actively considering adding the requirement in the near future.
The first opportunity to investigate the effectiveness of a second dose of varicella vaccine administered before a varicella outbreak came in September 2006, when the Arkansas Department of Health was notified of a varicella outbreak occurring in a school complex where approximately 40% of the students had received a second dose in February 2006. This article summarizes findings from this investigation and discusses their implications for varicella outbreak prevention and control.
On September 1, 2006, the Arkansas Department of Health was informed of an outbreak in a rural community in southeast Arkansas, where 8 of 31 cases were 2-dose varicella vaccine recipients. Most of the 31 case patients attended the elementary school complex that had experienced an earlier outbreak from January through March of 2006. As part of control measures10 in the previous outbreak, a vaccination clinic was held on February 2–3, 2006, at the elementary school complex, and nearly 45% of the 880 students who attended the school were vaccinated.
The elementary school complex consists of 3 schools: a prekindergarten (School A) with 100 students; a lower elementary from kindergarten through third grade (School B) with 429 students; and an upper elementary from fourth to sixth grade (School C) with 351 students. The school complex admits students from throughout the county (2000 census population: 20,000) as it is only 1 of 2 elementary school locations in the county. School A was housed in its own building with a separate playground. Schools B and C occupied different wings of the same building and students mixed a great deal because of shared facilities and activities on and off campus.
A case of varicella was defined as an acute generalized maculopapulovesicular rash with at least 3 lesions and without other apparent cause in a student from the elementary school complex with rash onset between September 1 and December 18, 2006. The 3 lesion minimum was used because biting insects are still present in southeast Arkansas in September and October. A person was considered vaccinated if his or her first or second dose was received at least 42 days before rash onset. Cases who received their first dose <42 days before rash onset were considered unvaccinated, as were noncases who received a first dose after September 1, 2006. Vaccine recipients who received their second dose <42 days before rash onset or after September 1 if a noncase were classified as having an indeterminate vaccination status and were excluded from vaccine effectiveness calculations. Varicella disease was categorized as either mild (3–49 lesions), moderate (50–249 lesions), or severe (>250 lesions, or having complications or requiring hospitalization).
Data about demographics, vaccination status, including dates of vaccination, underlying medical conditions, and history of prior varicella disease were collected using questionnaires distributed to all students in the school complex for their parents to complete. Varicella vaccination history was verified through the Arkansas vaccine registry, as well as through records maintained at the local county health unit. Cases were identified through the local county health unit, including referrals from the school nurse responsible for the school complex. Additional case finding was conducted using the parental questionnaire, which included a question regarding the presence of rashes, bumps, blisters, or spots on their child since the start of the school year. Brief telephone interviews were made to parents answering yes to the case-finding question. If the rash identified was consistent with the case definition, the child was considered a case. All suspected case patients were contacted by telephone and information was obtained regarding clinical and epidemiologic features of the illness, including potential sources of exposure.
Initial PCR testing of material from lesions was performed at the Arkansas State Laboratory using real-time PCR targeting VZV ORF22. Additional testing for VZV DNA was performed at CDC using VZV real-time FRET PCR. Genotyping of isolates was performed at the CDC National VZV Laboratory, using single nucleotide polymorphism analysis and ORF22r1 fragment DNA sequencing as described earlier.11,12 Blood spots were tested for anti-VZV IgM and IgG antibodies using a capture IgM ELISA and IgG ELISAs, using whole extracted VZV proteins or purified glycoproteins (gpELISA) as a source of antigen.13,14
Limited environmental samples of patient bedding and clothing were tested from 2 patients with active lesions in an effort to confirm varicella. One pair of pajamas from each of the 2 patients and a pillowcase from 1 patient's bedroom was collected. Clothing and bedding were processed at the CDC National VZV Laboratory by soaking overnight in 1 liter of deionized distilled water, manual wringing, and filtration through Whatman filter paper. Resultant liquid was tested for VZV by PCR using methods described previously.12
Data were analyzed using Epi Info version 3.3.2 (CDC, Atlanta, GA), SAS 9.1 (SAS Institute Inc, Cary NC), and Stata 8.2 (Stata Corp, College Station, TX). χ2 tests were used to compare proportions, Wilcoxon rank sum test was used to compare medians and numbers of lesions, and 2-sided P values were reported with a significance level of P < 0.05. The analyses were restricted to students attending the elementary school complex. Vaccination coverage for both 1 and 2 doses at the start of the outbreak was defined as the proportion of persons who had received varicella vaccine among those who had no history of varicella disease. The vaccine effectiveness of both 1 and 2 doses of vaccine were compared against an historic attack rate of 86.8% in unvaccinated individuals.15 Incremental vaccine effectiveness was calculated comparing those who had received 2 doses of vaccine against those who had received only 1 dose.
Of the 880 students in the elementary school complex, 696 (79%) returned a questionnaire. Response rates by school were 81% (81/100) in school A, 84% (362/429) in school B, and 72% (253/351) in School C (Table 1). Non-Hispanic black students were less likely to return a questionnaire than either non-Hispanic white or Hispanic students (65% vs. 86%, P < 0.001). Response rates did not differ by gender or vaccination status.
Vaccine Coverage and Disease History
Vaccination history was available for 871 (99%) of the students. Among those with no prior history of varicella, the overall vaccination coverage was 97% for at least 1 dose. Coverage for the second dose among students with no history of varicella was 39% overall and was comparable between Schools B (42%) and C (47%). Second dose coverage was 0% among students in School A, the prekindergarten, as they were not enrolled in school at the time of the February 2006 vaccination clinic (Table 1, P < 0.001 in comparison to Schools B and C).
Vaccination procedures at the vaccination clinic held in February, including storage and handling before and during the clinic, were consistent with applicable recommendations. There were no lot number differences among cases and noncases who received vaccine during the February clinic, and no evidence that vaccine administered during the clinic was compromised.
Disease history was available for 696 (79%) students who returned questionnaires and from an additional 2 vaccination records (as a reason for nonvaccination). Positive disease history for varicella was reported for 106 (15%) students. The proportion varied among schools (Table 1).
Between September 1 and December 18, 84 cases of varicella were identified in the school complex; occurrence was evenly distributed over the time interval. Overall attack rate was 15% (15/100) in School A, 10% (42/429) in School B, and 8% (27/351) in School C (Table 1, P = 0.052). The median age of cases was 8 years (range: 3–13). Thirteen (17%) of the 84 cases were identified from the case finding question on the questionnaire.
All of the cases were vaccinated with at least 1 dose or had a positive disease history. Twenty-five cases (30%) had been vaccinated with 2 doses, 53 (63%) with 1 dose, and 6 (7%) were unvaccinated with a history of disease. Two of the 2-dose vaccine recipients had a previous history of varicella; 1 had an unknown history. Seven of the 1-dose vaccine recipients had an earlier history of varicella.
Two of the cases had rash onset on September 1, 2006, the earliest date of rash onset. No known varicella or zoster exposure was reported for either case.
Overall disease severity was mild, with only 7 children (six 1-dose recipients and 1 unvaccinated child) reporting more than 50 lesions and none reporting complications or hospitalization. Of those reporting fewer than 50 lesions, one-third reported having 10 or fewer lesions. None of the cases had more than 250 skin lesions. Forty-six percent reported vesicles. The median duration of illness was 6 days for 1-dose recipients and 4 days for 2-dose recipients and those with a history of prior disease, however this difference is not statistically significant. Fewer than 30% in both groups reported a fever; 2 (33%) of the 6 laboratory confirmed cases had a documented fever. There were no differences in the clinical characteristics between 1 and 2 dose recipients. (Table 2).
For many of the cases, lesions had resolved by the time of the investigation and were no longer available for sampling and laboratory testing. Five (42%) of the 12 case patients whose lesion specimens were tested by the Arkansas State Laboratory were positive by PCR for wild type VZV DNA. Two (40%) of the 5 PCR positive specimens were from 2-dose recipients. Serum spots from 16 case patients were tested for anti-VZV IgM antibody; only one (6%) 1-dose recipient had anti-VZV IgM identified.
Environmental samples obtained from 2 patients with active lesions were both positive for wild-type VZV strain: one set of pajamas and the pillowcase. Both of the patients with positive environmental samples also had positive results by PCR from clinical specimens (1 was a 1-dose vaccine recipient case, the other was a 2-dose vaccine recipient who was vaccinated less than 42 days before rash onset).
The vaccine effectiveness calculations do not include the one 2-dose case patient who had been vaccinated 30 days before rash onset or the 9 healthy students who had been vaccinated with their second dose <42 days before September 1 (most were vaccinated after the start of the outbreak). Nine students who received their first dose <42 days before September 1 were considered unvaccinated (7 were vaccinated after September 1).
The attack rate in students with 1 dose and no prior history of varicella was 14.6% (46/316) and the attack rate in 2-dose recipients with no disease history was 10.4% (22/211); these were not significantly different (RR: 0.72, 95% CI: 0.44–1.15). By comparison with a historic attack rate of 87.8% among unvaccinated children enrolled in a clinical trial,15 vaccine effectiveness (VE) was estimated to be 85.4% (95% CI: 80.8–88.9) among 1-dose recipients and 89.1% (95% CI: 83.8 – 92.7) among 2-dose recipients (This vaccine effectiveness calculation could also have used the unvaccinated from an historic school outbreak. Using the unvaccinated rates from Izurieta et al,16 vaccine effectiveness (VE) was estimated to be 83.4% (95% CI: 78.1–87.5) among 1-dose recipients and 88.1% (95% CI: 82.2 – 92.1) among 2-dose recipients). The incremental vaccine effectiveness of the second dose over the first dose was 28.4% (95% CI: −15.4–55.6) of the estimated remaining 14.6% not covered by the first dose. As no severe disease was identified in the school, the vaccine was 100% effective against severe disease in both 1 and 2 dose recipients.
This report documents the largest outbreak of varicella reported in recent years in the United States and the first outbreak involving both 1- and 2-dose varicella vaccine recipients. In addition, this investigation provides information regarding the clinical manifestations of varicella in 2-dose vaccine recipients. The outbreak lasted nearly 4 months and involved 84 students from the elementary school complex. Vaccine coverage levels at the school complex were high for 1 dose (97%), but only moderate (39%) for 2 doses. Vaccine effectiveness at 85% for 1 dose of vaccine was comparable with previously published reports of postlicensure studies17 comparisons to historic unvaccinated rates, but significantly lower than the 94% reported from the clinical trial by Kuter et al18 The 89% effectiveness of the second dose was also significantly lower than the 98% reported by Kuter et al18 The 2-dose vaccine recipients had a similar clinical presentation to 1-dose recipients and those with a disease history. The higher estimates found by Kuter et al might in part be attributable to the more ideal circumstances of a clinical trial, including controlled vaccine administration, laboratory confirmation of all cases, and the exclusion of patients who may have received medications possibly limiting the vaccine's performance.
This study reinforces the challenges of identifying and confirming varicella in vaccinated individuals, given the difficulty in diagnosing mild cases of varicella in vaccinated persons and distinguishing them from other rash illnesses or insect bites. Many of the cases confirmed by PCR had few macules or papules, and many did not have any vesicles. One such PCR confirmed case was in a 2-dose recipient without fever who had only 3 small papules on her arm. In addition, there are no laboratory tools to confirm varicella once transient lesions resolve.8 Even while lesions are present, varicella testing of vaccinated cases is likely to be less sensitive because of the challenges of sampling macular lesions or their lower viral load (Leung J, Schmid DS, personal communication 2006). To confirm the circulation of VZV during this outbreak, we also used limited environmental sampling. While we were successful in confirming the presence of VZV DNA in clothing and bedding from suspected cases, this approach is impractical because routine cleaning of clothing and bedding will limit available environmental specimens for testing, and the testing of the clothing was highly labor intensive.
In our investigation, the incremental vaccine effectiveness of 2 compared with 1 dose of varicella vaccine was only 28%, with confidence limits that include zero. Kuter et al18 reported from their prospective 10-year multicenter clinical trial that vaccine efficacy for 2 doses was 98.3% compared with 94.4% with 1 dose, which translates to a statistically significant incremental vaccine efficacy of 69% (of the estimated 5.6% not covered by the first dose). Most of the cases reported in the study by Kuter et al were sporadic rather than outbreak related. Our attack rates of 14.6% in 1-dose recipients and 10.4% in 2-dose recipients are comparable with attack rates among 1-dose recipients of 8% to 26% found in previous school outbreaks with moderate to high one dose coverage.6,16,19,20
Our study has several limitations. First, some degree of misclassification likely occurred due to the challenges in recognizing and confirming mild cases and in distinguishing them from other rash illnesses or insect bites. This misclassification is likely nondifferential with respect to vaccination status, leading to similar estimates of vaccine effectiveness in both 1-dose and 2-dose recipients. It is possible, however, that disease in 2-dose recipients is even milder than that present in 1-dose recipients. If even milder 2-dose disease is not included, this may overestimate the VE of the second dose. Second, although this outbreak was sizeable, it may not have been large enough to detect small but meaningful differences in the incremental vaccine effectiveness.
Our investigation did not find a significant difference between 1-dose and 2-dose vaccine effectiveness; however, the overall vaccine coverage for 2 doses within the school was only 39%. This outbreak occurred only shortly after the June 2006 ACIP recommendation for routine 2-dose coverage, fortuitously in a population that had received a second dose earlier in the year for outbreak control. It is possible that higher 2-dose coverage at the school would have prevented the second outbreak from occurring, or that high coverage with 2 doses of varicella vaccine will reduce the number of outbreaks in the population. As our understanding of the effectiveness of 1 dose of varicella has developed over time with additional postlicensure studies,17 additional studies will be needed to assess the impact of the routine 2-dose recommendation by the ACIP. Additionally, it will be important to monitor the effect of these recommendations on the number and size of outbreaks reported from states.
The authors thank the following for their important contributions to this study. The staffs of the 3 schools that make up the elementary school complex were generous with their time even while preparing for their annual state education inspection. Dr. Haytham Safi assisted us with the Arkansas electronic vaccine registry. Donetta McGriff and Sue Hinton of the Arkansas Department of Health, and Tureka Watson at CDC, and the staff of the county health unit all supported this effort.
1. Davis MM, Patel MS, Gebremariam A. Decline in varicella-related hospitalizations and expenditures for children and adults after introduction of varicella vaccine in the United States. Pediatrics
2. Galil K, Brown C, Lin F, et al. Hospitalizations for varicella in the United States, 1988 to 1999. Pediatr Infect Dis J
3. Nguyen HQ, Jumaan AO, Seward JF. Decline in mortality due to varicella after implementation of varicella vaccination in the United States. N Engl J Med
4. Outbreak of varicella among vaccinated children—Michigan, 2003. Morb Mortal Wkly Rep
5. Lee BR, Feaver SL, Miller CA, et al. An elementary school outbreak of varicella attributed to vaccine failure: policy implications. J Infect Dis
6. Lopez AS, Guris D, Zimmerman L, et al. One dose of varicella vaccine does not prevent school outbreaks: is it time for a second dose? Pediatrics
7. Tugwell BD, Lee LE, Gillette H, et al. Chickenpox outbreak in a highly vaccinated school population. Pediatrics
8. Chaves SS, Haber P, Walton K, et al. Safety of varicella vaccine after licensure in the United States: experience from reports to the vaccine adverse event reporting system, 1995–2005. J Infect Dis
. 2008;197(suppl 2):S170–S177.
9. Seward JF, Zhang JX, Maupin TJ, et al. Contagiousness of varicella in vaccinated cases: a household contact study. JAMA
10. ACIP. Provisional Recommendations for Prevention of Varicella
. Atlanta, GA: Advisory Committee on Immunization Practices; 2006.
11. Loparev V, Martro E, Rubtcova E, et al. Toward universal varicella-zoster virus (VZV) genotyping: diversity of VZV strains from France and Spain. J Clin Microbiol
12. Loparev VN, McCaustland K, Holloway BP, et al. Rapid genotyping of varicella-zoster virus vaccine and wild-type strains with fluorophore-labeled hybridization probes. J Clin Microbiol
13. Watson B, Civen R, Reynolds M, et al. Validity of self-reported varicella disease history in pregnant women attending prenatal clinics. Public Health Rep
14. Weinmann S, Chun C, Mullooly JP, et al. Laboratory diagnosis and characteristics of breakthrough varicella in children. J Infect Dis
. 2008;197(suppl 2):S132–S138.
15. Ross AH. Modification of chicken pox in family contacts by administration of gamma globulin. N Engl J Med
16. Izurieta HS, Strebel PM, Blake PA. Postlicensure effectiveness of varicella vaccine during an outbreak in a child care center. JAMA
17. Seward JF, Marin M, Vazquez M. Varicella vaccine effectiveness in the US vaccination program: a review. J Infect Dis
. 2008;197(suppl 2):S82–S89.
18. Kuter B, Matthews H, Shinefield H, et al. Ten year follow-up of healthy children who received one or two injections of varicella vaccine. Pediatr Infect Dis J
19. Marin M, Nguyen HQ, Keen J, et al. Importance of catch-up vaccination: experience from a varicella outbreak, Maine, 2002–2003. Pediatrics
20. Parker AA, Reynolds MA, Leung J, et al. Challenges to implementing second-dose varicella vaccination during an outbreak in the absence of a routine 2-dose vaccination requirement—Maine, 2006. J Infect Dis
. 2008;197(suppl 2):S101–S107.