Background. Studies in Japan and the United States have shown that varicella vaccine is both safe and efficacious. In 1984, we undertook a 10-year prospective study using a research lot of Oka/Merck varicella vaccine to assess antibody persistence and breakthrough chickenpox rates. In 1987, we began a similar prospective study with lots made in production facilities that ended after 6 years because many children were given a second dose. The purpose of this study is to report humoral antibody persistence and breakthrough chickenpox rates after 6 to 10 years of prospective follow-up.
Methods. One hundred forty-three seronegative children received a research lot (950 plaque-forming units/dose) with 97.9% seroconversion by an assay for fluorescent antibody to membrane antigen (FAMA). One hundred thirty-eight children received production lots (1145 to 3265 plaque-forming units/dose) with 93.5% seroconversion. Yearly chickenpox exposure surveys were completed by phone, and children were seen by a study nurse whenever chickenpox was suspected. A subset in each group had serum collected every 2 years and tested for FAMA antibody.
Results. In the research group there have been 25 cases of chickenpox in 137 seroconverters in a period of more than 10 years (yearly rate of 1.7%). In the production lot group there have been 22 cases of chickenpox in 129 seroconverters in a 6-year period (yearly rate of 2.8%). In the research group the median titer rose from 1:16 to 1:64 between 1 and 10 years. In the production group, the median titer did not change between 1, 2, and 4 years. Median antibody titers were compared between the research and production groups at 1, 2, and 4 years and did not differ. The rate of development of modified chickenpox has not increased with time since vaccination, and neither has the case severity. Children with FAMA titers ≤8 at 6-weeks' postvaccination were four times more likely to develop chickenpox than those with titers ≥64.
Conclusions. 1) Modified chickenpox has occurred in approximately 2% to 3% of vaccinees per year, regardless of the vaccine lot given. 2) FAMA titers have risen between 1 and 10 years in research lot recipients and remained the same in production lot recipients. 3) The likelihood of modified chickenpox developing is inversely related to the 6-week postvaccination FAMA titer.
In March 1995, the twenty-one-year development and testing phase for varicella vaccine (VARIVAX, West Point, PA) culminated with licensure by the United States Food and Drug Administration. Before licensure, approximately 3 million doses of Oka strain vaccine had been given in Japan and Europe,1and in the first year after licensure, 2.3 million doses were distributed in the United States.2 Yet, concern remains that the vaccine will postpone rather than prevent chickenpox, and some physicians are not recommending vaccination.3 Mathematical models have predicted, however, that the universal use of varicella vaccine in infancy will decrease the overall incidence of disease, even though it may shift the mean age of disease upward.4 The worst outcome would occur if vaccine were not used universally, leaving a substantial unprotected cohort after childhood.3Therefore, it is important for physicians to have data demonstrating the ability of varicella vaccine to prevent chickenpox and its complications throughout long periods, so that administration of the vaccine at 12 months and to older susceptible children becomes universal. The recent increase in cases of invasive group A streptococcal infections complicating chickenpox is of concern, with a study in Los Angeles reporting that up to 12% of all varicella complications are attributable to group A streptococci.5Furthermore, there is a rate of 1 in 550 cases of chickenpox being hospitalized for a complication of varicella.5 Because the average age of a child with chickenpox complications is 3.8 years, this argues against delaying vaccination until the teen years as some doctors have proposed. A study in Boston done in 1995 in adults and immunocompromised patients6 showed that previous studies had underestimated the rate of hospitalization for all varicella complications by fivefold.
We report our results of a 10-year prospective study of varicella vaccine recipients aged 12 months to 13 years, who were followed by yearly phone calls for chickenpox exposures and seen by a physician whenever modified chickenpox was suspected by the parents. We also report varicella antibody titers, measured by fluorescent antibody to membrane antigen (FAMA) assay, in a subset of children returning biannually for serum collection, and compare the titers of children given two lots of Oka/Merck vaccine, differing in the site of manufacture and in viral titer.
MATERIALS AND METHODS
Two vaccine preparations were used. Research lot 977 was made in small-scale research lots and used in 1984 at a dose of 950 plaque-forming units (pfu). Four production lots prepared in a Merck Sharp & Dohme (West Point, PA) manufacturing facility were used in the second study from December, 1987 to May, 1989. The titers were 1145, 1770, 3010, and 3265 pfu/dose. These lots also varied in the total antigen content,7 with the production lots having a greater antigen content relative to the number of infectious particles.
In the research group healthy toddlers 12 to 24 months of age and the older siblings as old as 12 years of age were enrolled as previously described.8 In the production group, children 12 months to 17 years were eligible, but only 2 of 138 children were >13 years of age. All parents/guardians of the children gave informed consent for three sera to be obtained: prevaccination, 6-weeks' postvaccination (PV), and at 1 year. Two sites were used for enrollment: a private suburban office and an urban teaching hospital.
Sera were assayed in Cleveland by our previously described modification of the FAMA assay.8 Titers are expressed as the reciprocal of the highest dilution with a positive result. A titer of 1:2 or higher was considered positive. Enzyme-linked immunosorbent assays (ELISAs) for varicella zoster virus (VZV) glycoprotein were performed by Merck Research Laboratories (West Point, PA) as previously described;9 these titers are reported only when insufficient sera were available to perform FAMA assays, which occurred in the production lot group only.
Patient Follow-up Methods
At 2-years' PV, all parents and those children >8 years of age were asked to participate in a 10-year serum collection study. A separate informed consent was signed to have serum drawn every other year for 10 years and to contact the researchers for examination if chickenpox was suspected. Families were contacted by phone yearly for 6- to 6½-years' PV in the production lot and 10 to 10½ years in the research lot to complete an exposure questionnaire, regardless of participation in the serum collection study. Exposures were classified as household or nonhousehold contacts. When unable to reach the participants, letters were sent. We also used contact phone numbers for friends or relatives obtained at study enrollment. For the subset of 20 children in the biannual serum survey at 10 years, phone calls were made at 12 years to see if anyone had developed chickenpox.
In the production lot group, sera were drawn at 1, 2, and 4 years, and then a second dose of vaccine was offered to all remaining participants as part of a study on revaccination. All revaccinated children were removed from this study. However, phone calls for exposure histories were done at 6 to 6½ years for those who remained in the study.
Families were asked to contact the investigators if a case of chickenpox or shingles was suspected, even if only a few lesions were present. Children with suspected varicella were seen by a physician, and an acute serum was obtained. If a vesicle was present, permission to culture it for varicella virus was requested. A 6-week postinfection serum sample was obtained from those children in whom the acute titer was either not obtained or did not document a fourfold or more increase from the most recently obtained sample. A proven case was defined as a disseminated papulovesicular rash with a fourfold or more increase in the FAMA titer or a positive vesicular culture; a probable case was a papulovesicular rash after varicella exposure, but without serologic confirmation.
Revaccination of Children Who Failed to Seroconvert
Fifteen children failing to seroconvert were offered a free second dose of varicella vaccine. Such children were not included in the biannual serum survey or in exposure studies, except for one phone call at the end of the study to determine if breakthrough disease had occurred. These 15 children are discussed because of interesting clustering of failures within families.
Because the data were not normally distributed, nonparametric analyses were used. A Friedman 2-way analysis of variance was conducted on median titers for the production lot and research lot to examine the stability of titers throughout a time period. Post hoc comparisons within groups were done using the Wilcoxon matched-pairs signed ranks test. Research and production lot titers were compared at 1-, 2-, and 4-years' PV by the Mann-Whitney U test. We compared 6 week and 1 year median titers in children in each lot who went on to develop chickenpox with those who did not get the disease using the Mann-Whitney U test. Kaplan-Meier survival curve analysis was used to compare the breakthrough rates of chickenpox in the two groups throughout an approximate 6-year period because that was the study period of the production group. The time of breakthrough was included, regardless of whether the child completed the study. Data were truncated at 74 months to minimize artifact in the tail of the distribution. A log-rank χ2 test was used to compare the breakthrough rates. The Mann-Whitney U test was used to compare the age of vaccination in children with or without modified chickenpox.
Study Population and Length of Follow-up
Two types of Oka/Merck varicella vaccine were studied: a research lot with 950 pfu/dose and production lots produced at manufacturing scale, with 1100 to 3265 pfu/dose. For comparison, the licensed Merck varicella vaccine now has a minimum of 1350 pfu/dose on the expiration date, but the titer is more than 2700 pfu per dose when used promptly. The age range of children at vaccination was 12 months to 13 years in both groups, but the research lot group was younger; median age 22 months for the research group and 27 months for the production group (P < .01). All agreed to be followed prospectively by yearly phone calls and to report any cases of suspected chickenpox to the study nurse. Contact was maintained successfully for 4 years with 93% of children who received the research lot and 81% of children who received the production lot.
In the research group 18% had developed modified chickenpox, 48% had been lost to follow-up, and 34% remained available for exposure history at the end of 10 years. At the end of the 6-year production lot study, 17% had developed modified chickenpox, 25% had been lost to follow-up or revaccinated, and 58% remained available for exposure history.
Seroconversion and Modified Chickenpox
FAMA seroconversion rates for research and production groups were 97.9% and 93.5%, respectively (Table1). Phone surveillance was 10 to 10½ years for the research lot group and 6 to 6½ years for the production lot group. There were 47 cases of modified chickenpox, of which 46 were considered proven cases by a fourfold or more FAMA titer rise within 4 weeks after disease and 1 by a positive vesicular culture. There were 7 probable cases, of which 6 had vesicular lesions seen by a study nurse or doctor, but the parent refused blood drawing; and 1 case was not seen by a doctor, but had 25 vesicles accompanied by fever. The median age at vaccination for children who developed chickenpox was higher (23 months) than for those who did not (17 mo; P = .006) in the research group. There was no difference in age at vaccination between those with and without chickenpox in the production group (28 months vs 26 months).
The average annual percentage of children who developed chickenpox was 1.7% in the research lot and 2.8% in the production lot (Table 1). Figure 1 shows the cumulative rate of development of modified chickenpox for the research and production groups during equal time periods by Kaplan-Meier survival curves. The cumulative risk of modified chickenpox by 74 months was 13% in the research group and 25% in the production group (P = .06). The failure rate between the two groups was similar up to 48 months' PV (0.07% in the research group versus 0.09% in the production group; P = .85). However, after 48 months, the curves indicate by divergence a trend (P = .09) towards a higher failure rate in the production group (0.16%) compared with the research group (0.06%).
In the research group throughout the 10-year period there were 57 household exposures and 643 nonhousehold exposures, resulting in 25 cases of chickenpox. In the production group, there were 37 household and 352 nonhousehold exposures throughout a 6-year period, resulting in 22 cases of modified chickenpox. In the research group, 8.6% of household exposures resulted in chickenpox; in the production group, the rate was 13.5%. In the research group 3.1%, of nonhousehold exposures led to chickenpox; in the production group, the rate was 4.8%.
Characteristics of Modified Chickenpox
As other investigators have reported, our cases of modified chickenpox were considerably milder than natural disease.10-14 Fifty-five percent of children were seen by a study nurse or doctor while rash was present, and the other 45% were by parental report of the rash. The median number of papulovesicular lesions was 30 for the research lot recipients (range, 5 to 140), and 28 for the production lots (range, 5 to 150). A temperature >100°F was present in 36% of the research lot and 27% of the production lot recipients, but generally lasted only 1 day. There was no increase in the severity of illness during the 6- to 10-year period as judged by either the number of vesicles or the frequency of fever. We also examined the characteristics of modified chickenpox in the 5 children who declined revaccination after failing to seroconvert. Four developed mild chickenpox (median 40 vesicles and all afebrile), and 1 reported no exposures. Of the 9 who were revaccinated after initial failure, all seroconverted, but 2 subsequently developed mild chickenpox.
FAMA Antibody Persistence After Vaccination
All children had sera drawn prevaccination, 6-weeks' PV, and 1-year PV; these results have been previously reported for the research group.8 Table 2 shows the median FAMA titers for a subset of children who agreed to participate in a biannual serum collection study for up to 10 years. Children who developed modified chickenpox are indicated in Figs2 and 3 and were excluded from the median titer calculation (titers after breakthrough were 128 to 4096). The median titer fell between 1 to 2 years in both vaccine groups, although this change was only statistically significant for the research lot group. In the production lot group there was no overall change in titers between 1 and 4 years (Table 2); however, there was a statistically significant rise from 1 to 10 years in the research lot group (Table 2). We also compared the median FAMA titers between the two groups at time periods of 1-, 2-, and 4-years' PV; there was no significant difference in titer at any of the three times analyzed.
FAMA Protection From Disease
Figures 2 and 3 show that for the subset of children participating in biannual serum collection, no one with a titer greater than 1:16 had modified chickenpox within a 2-year period after testing. Of 27 cases of chickenpox, 18 had a titer of <2 or 2 within the 2 years before breakthrough. We also examined the relationship between the 6-week PV titer and the ultimate development of modified chickenpox. ELISA titers had to be used for this analysis in production lot children because of insufficient sera for FAMA titers. Table3 shows that there was a significantly lower 6-week PV median ELISA titer in the production lot children who developed chickenpox (4.2 vs 11.2; P = .02) and a lower 6-week FAMA titer in the research lot children who developed chickenpox (4 vs 16; P < .01). Lower titers in breakthrough cases were also seen when patients were analyzed by 1-year PV titers (Table3). Low FAMA titer at 6-weeks' PV in the research group was associated with a high rate of breakthrough chickenpox (Fig4). Seroconverting children with a 6-week PV titer ≤8 had an 83% likelihood of developing modified chickenpox within 10 years.
Family Clusters of Vaccine Failure
There were three families in which two siblings had poor responses to varicella vaccine. In the first case, the siblings failed to seroconvert, were revaccinated, and seroconverted, but one went on to become seronegative 2-years' PV and contracted modified chickenpox (48 vesicles). The second pair were fraternal twins who failed to seroconvert, declined revaccination, but have remained uninfected for 6 years; their sibling seroconverted. The third pair are brothers who failed to seroconvert, responded normally to the second dose, and have remained uninfected for 6 years.
Our data show that protection against chickenpox is correlated with persistence of antibody throughout a 10-year period after vaccination. They confirm the findings of two other studies that the 6-week PV antibody titer is predictive of protection from modified chickenpox.10,11 Furthermore, our data show close similarity between antibody persistence and rates of breakthrough chickenpox in children receiving either research lot vaccines produced in 1984 or production lots produced from 1987 to 1988, which more closely resemble the licensed vaccine in site of manufacture and number of viral particles. In 1989, we examined the 3-year PV experience with children in our research lot group. We found that 11% of children showed fourfold FAMA antibody boosts without any clinical symptoms when they entered nursery school or day care, presumably because of the frequent exposures to chickenpox that the parents reported.12 Our 4-year and 10-year PV serologic data now show that although antibody titers may decline between 1- and 2-years' PV, the median titer is unchanged after 4 years for the production lot and actually rises after 10 years for the research lot. In Japan, previous studies have shown that FAMA antibodies to VZV were present in 97% of children 7- to 10-years' PV,15 and at 20-years' PV the FAMA titers were even higher than at 10-years' PV.16 Such results are probably attributable to contact with circulating virus, because vaccine coverage is <20% in Japan.17 In the United States, boosting from wild virus exposure may fall with increasing rates of vaccination; however, a recent analysis of the influences of zoster on VZV transmission suggests that many decades may pass before boosting due to zoster ends.18 The magnitude of the effect depends on how often elderly adults interact with susceptible children.
Cases of modified chickenpox never developed in children with a FAMA titer >1:16 in the preceding 1 to 2 years, indicating that antibody persistence is related to protection. The 6-week FAMA response to vaccination was predictive of breakthrough chickenpox risk. This has previously been shown by White et al10 and Clements et al11 for glycoprotein ELISA antibody titers, but not for FAMA titers. The effectiveness of the varicella vaccine we studied can be crudely estimated from the natural attack rate in children 1 to 9 years of age of 8.3% to 9.1% per year.17 Our research lot rate of 1.7% of vaccinees with modified chickenpox per year is a reduction of 80% to 81%, and the production lot rate of 2.8% is a 66% to 69% reduction. However, we believe most cases of PV chickenpox will be too mild to be recognized as disease and few will seek medical attention, so the number of typical cases will be much more greatly reduced.
It is very likely that cell-mediated immunity also plays a role in preventing varicella.7 We observed, as have others,11,13 that children failing to seroconvert were nevertheless protected from natural chickenpox and developed very mild disease. Arvin7 demonstrated that T-cell proliferative responses to varicella glycoprotein I and IE62 proteins is stimulated by both the research lot and the production lots used in this study. This suggests that PV antibody testing may not be cost-effective in healthy children, because seronegative children may have protection from severe disease.
A similar rate of modified chickenpox has been observed in other studies, and rates of disease varied only slightly with the dose of virus administered. Watson et al19 found a lower rate of modified chickenpox in research lot trials (approximately 0.7% per year) compared with 1987 to 1989 production lots, with a rate of 2.2% per year. They speculated that this is because of higher 6-week antibody titers in the former group. In the 5-year to 10-year PV results published by Clements et al,11 the modified chickenpox rate was 18.6%. In the collaborative study by Merck Research Laboratories, the rate of modified disease was 2.1% and 2.4% in year 1 and 2, respectively.10 We did not find any difference in the 1, 2, or 4 year PV-FAMA titers between the two lots of vaccine, but we did not perform FAMA titers on 6-week sera of all children, so we cannot be certain there were no significant differences in the 6-week FAMA response to the two types of vaccine. Given the low rate of seroconversion failure (4%), our finding of three families with seroconversion failure in two siblings suggests that, as with hepatitis B, inherited immune response genes may play a role in vaccine failure.20
Now that varicella vaccine is licensed for use in the United States, it is important that all health care providers encourage its use, to prevent the accumulation of susceptible adults. Twenty-two years of United States studies, plus use in Japan and Europe, have shown it to be safe and efficacious.21 It is projected to be cost-effective,22 but the cost per life saved rises as the vaccine coverage rate falls below 70%. A mathematical model was developed by Halloran et al4 in 1994 to study the effects of varying levels of vaccine coverage and efficacy. At 97% vaccine coverage, regardless of whether efficacy was as low as 83% or as high as 95%, varicella transmission was predicted to effectively end in 30 years. However, if the level of coverage were only 70%, even with a second dose catch-up program at 12 years of age, over 760 000 cases per year of chickenpox would still occur, and at an older age than is presently seen. Thus, it is important to vaccinate 12-month-old children and all older children and adults who are still susceptible if we wish to eradicate varicella and its complications. The public sector campaign that began in 1996 is especially important, because the poor and uninsured are the ones most likely to be left susceptible into adulthood because of missed vaccination opportunities.
Now that varicella vaccine is part of the Vaccines for Children program, pediatricians and family practitioners have the opportunity to offer free vaccine to uninsured children. As Plotkin3 wrote in a recent editorial, “Physicians who withhold vaccination may create a self-fulfilling prophecy that more adults will get varicella.”
Partial funding for this study was provided by Merck Research Laboratories.
We wish to thank the following nurses and doctors for recruiting patients for our study: Micheal Michael, Sue Vladimiroff, Cheryl Henderson, Cynthia Strieter, Teresa Kammerman, Cynthia Black, Robert Mack, Paul Aronowitz, Larry Monger, Lynn Ridge, Alvin Novisky and Janise Parker. We also thank Doris Fulton and Sandy Prokay for laboratory assistance; Barbara Kuter, PhD, and Thomas R. Johnson, PhD, for review of the manuscript; and Mario Cleves, PhD, for statistical advice.
- Received February 18, 1997.
- Accepted May 7, 1997.
Reprint requests to (C.E.J.) Department of Pediatrics, MetroHealth Medical Center, 2500 MetroHealth Dr, Rm H-455, Cleveland, OH 44109.
↵† Leonard P. Rome is deceased.
Presented in part in abstract form at the Society for Pediatric Research, Seattle, WA, May 1994.
- FAMA =
- fluorescent antibody to membrane antigen •
- pfu =
- plaque-forming units •
- ELISA =
- enzyme-linked immunosorbent assay •
- VZV =
- varicella zoster virus •
- PV =
- Jones CP
- Plotkin SA
- Halloran ME,
- Cochi SL,
- Lieu TA,
- Wharton M,
- Fehrs L
- Choo PW,
- Donahue JG,
- Manson JE,
- Platt R
- Arvin AM
- Johnson CE,
- Shurin PA,
- Fattlar D,
- et al.
- Johnson CE,
- Rome LP,
- Stancin T,
- Kumar ML
- Watson BM,
- Piercy SA,
- Plotkin SA,
- Starr SE
- Bernstein HH,
- Rothstein EP,
- Pennridge Pediatric Associates, et al
- Asano Y,
- Nagai T,
- Miyata T,
- et al.
- Asano Y,
- Suga S,
- Yoshikawa T,
- et al.
- ↵Centers for Disease Control and Prevention. Prevention of varicella. MMWR. 1996;45(RR-11):1–36
- Ferfuson NM,
- Anderson RM,
- Garnet GP
- Watson B,
- Gupta R,
- Randall T,
- Starr S
- Copyright © 1997 American Academy of Pediatrics