Efficacy of High-Titer Live Attenuated Varicella Vaccine in Healthy Young Children

Transcription

1 8330 Efficacy of High-Titer Live Attenuated Varicella Vaccine in Healthy Young Children Tiina Varis and Timo Vesikari Department of Virology, University of Tampere Medical School, Tampere, Finland The efficacy of a high-titer, reformulated varicella vaccine was studied in to 30-monthold children. Vaccinees were randomly allocated to 5 groups to receive one of two lots of an original high-titer vaccine, one of two lots of a partially heat-inactivated vaccine, or placebo. Both vaccines were well tolerated. Seroconversion was detected in 100% and 99% of children immunized with the high- and low-titer vaccines, respectively. Sixty-five cases of serologically confirmed varicella-like disease were discovered during follow-up (mean, 29.3 months): 5 in the high-titer vaccine group, 19 in the low-titer vaccine group, and 41 in the placebo group (P ::::;.005 for each difference). Thus, the protective efficacy of live attenuated varicella vaccine is dependent on vaccine titer. High-titer varicella vaccine induces excellent protection in healthy young children. In 1974, Takahashi et a1. [1] introduced a live attenuated varicella vaccine for the prevention of chickenpox in high-risk children [1]. Subsequently, vaccines derived from the Japanese Oka strain were developed in Europe and the United States and licensed in many European countries for vaccination of immunocompromised children. However, use ofvaricella vaccine for the control of chickenpox in high-risk groups has remained limited. Even though varicella vaccine is well tolerated, the reactogenicity of the Oka strain vaccine is greater in immunocompromised than in healthy children, and efficacy is slightly lower [2-6]. Moreover, vaccination requires interruption of chemotherapy, which many pediatric oncologists find problematic. In healthy children, live varicella vaccine is well tolerated and more efficacious than in immunocompromised subjects [7-10]. Extensive immunization ofhealthy children at an early age would offer an alternative strategy for the protection of high-risk groups against varicella, provided that the vaccineinduced immunity was solid and long-lasting. In a series of US studies, modified cases of chickenpox were observed in 2.1 % and 2.4% of vaccinees in the first and second year after vaccination, respectively [11], and long-term follow-up suggests that the incidence of breakthrough cases may remain at the same level in subsequent years [11-13]. However, these studies have included few children younger than 2 years old. We previously found that 7% of successfully immunized toddlers, who were 13- to 17-months-old at vaccination, had clinically suspected breakthrough mild varicella during an 18 month follow-up period [10]. This age group may be regarded Parents gave informed consent. The study was approved by the Ethics Committee, Tampere University Hospital. Current Good Clinical Practice guidelines were followed. Reprints or correspondence: Dr. Timo Vesikari, University oftampere Medical School, P.O. Box 607, FIN Tarnpere, Finland. The Journal of Infectious Diseases 1996; 174(Suppl 3):S by The University of Chicago. All rights reserved /96/ $01.00 as the primary target if varicella were to be controlled by a general vaccination program. We reasoned that immunization of young children presents a particular challenge. Vaccine-induced immunity might wane more rapidly in younger than in older children, and a higher dose of varicella vaccine might be required to induce durable immunity. After a reformulated varicella vaccine with a high titer was developed, we undertook a randomized, placebo-controlled trial to determine the efficacy ofthis vaccine in children 10- to 30-months-old over a mean follow-up period of 29.3 months. Methods Vaccine. Live attenuated varicella vaccine was derived from Oka strain (obtained from the University of Osaka, Japan) and propagated in MRC-5 human fibroblast cells at SmithKline Beecham Biologicals (Rixensart, Belgium). To produce the vaccine, infected cells were harvested, concentrated, diluted in stabilizing reagents, and disrupted to release the intracellular varicellazoster virus (VZV). The final product was freeze-dried. We used two production lots of vaccine (VA103A42 and VA107A42) in this study. The titers of the respective lots were (10,000) and (15,850) pfujdose (0.5 ml). To obtain a low-titer vaccine, vials ofthe original vaccine lots were heated at 37 C for 10 days. This procedure was thought to mimic an aging process at prolonged storage. The resulting low-titer vaccines had titers of (1260) pfu and (630) pfu/dose for lots VA103A42 and VA107A42, respectively. We will refer to the two original vaccine lots as "high-titer" vaccine and the two heated lots as "low-titer" (or "heat-exposed") vaccine. The placebo was prepared in a manner similar to the original active vaccine except that no VZV was present. The freeze-dried vaccine and placebo preparations were similar in appearance: Both were beige-colored powders. When reconstituted, the placebo and the varicella vaccine both were clear liquids. The vaccines were stored at - 20 C for the duration of the study. Shortly before administration, the freeze-dried pellets were reconstituted with 0.6 ml of sterile water; 0.5 ml was injected subcutaneously in the left upper arm.

2 JID 1996; 174 (Suppl 3) High-Titer Varicella Vaccine in Healthy Children S331 Subjects. We studied 513 healthy children who were 10- to 30-months old at vaccination. We recruited participants by contacting by letter parents of children born in Tampere University Hospital and explaining the nature of the study. About 25% ofthe parents approached chose to participate. Inclusion criteria required the children to be free of obvious health problems, as established by clinical examination. Children with a history of chickenpox or zoster were excluded. Administration ofthe varicella vaccine was separated by 28 days from other routine childhood immunizations, notably measles-mumps-rubella vaccine. Study design. The study was double-blind, randomized, and placebo-controlled. The vaccine vials were individually numbercoded according to a randomization list. Each vaccinee was assigned the same running number as the number on the respective vaccine vial. Randomization was done in blocks of 18. Two-thirds ofthe subjects received varicella vaccine; one-third received placebo. The vaccine recipients were allocated to 4 subgroups to receive the two high-titer and two low-titer (heat-exposed) vaccine lots. Vaccinations were done between September 1991 and August Before vaccination, a venous blood specimen was collected. Parents were asked to monitor symptoms, take rectal temperatures each day for 28 days after vaccination (postvaccination period), and record findings on a check list. They were asked to contact the investigators immediately if any papular or vesicular skin reactions occurred. A blood specimen was collected from 35 to 63 days (target, 42 days) after vaccination. In addition, parents were told to contact the investigators immediately ifany papular or vesicular rash occurred after the postvaccination period or if the child was in contact with a suspected case of chickenpox or shingles. A separate clinic was established for these visits so that all children, even those with the mildest varicella-like disease, could be seen promptly. A full-time nurse was available for these consultations. Cases ofsuspected varicella were clinically examined by the investigator. A symptom diary card was started. Venous blood specimens were collected at the acute stage and 4-6 weeks later for determination ofvaricella antibodies. All children were seen by the investigators ~ 1 year after vaccination. The follow-up lasted until 30 June 1994, when the mean duration offollow-up was 29.3 months. Laboratory studies. Sera were stored at -20 cc until shipment on dry ice to SmithKline Beecham Biologicals for titration under code. Varicella antibodies were measured by an indirect immunofluorescence (IIF) method [14]. Samples that showed no fluorescence or barely visible fluorescence at the 1:4 starting dilution were considered to be seronegative. Seroconversion was defined as the appearance of antibodies in the serum of subjects who were initially seronegative (i.e., IIF titer < 1:4 before vaccination to ~ 1:4 after vaccination). A 4-fold rise in IIF antibody titer was regarded as significant in those initially seropositive. Geometric mean titers (GMT) of varicella antibodies were calculated using log transformation of positive titers and taking the antilog of the mean ofthe transformed values. Analysis ofresults. Sample size was calculated from assumptions on disease attack rates. The analyses only considered cases of varicella that were serologically confirmed. For these cases, an analysis of clinical severity was done by SmithKline Beecham Biologicals using the case report forms supplied by the clinical investigators. Table 1. Vaccine titer High (release) Low (heat-exposed) Placebo NOTE. CI, confidence interval; GMT, geometric mean titer. * p =.004, t test to compare GMTs between high- and low-titer vaccines. In addition, an analysis of postvaccination reactions was done for cases in which papular or vesicular skin manifestations were observed within the 28-day postvaccination period. Postvaccination serologic responses were also analyzed by SmithKline Beecham without revealing the code to the investigators. Protective efficacy was calculated as follows: Efficacy = 100 X [1 - (attack rate in vaccine recipients/attack rate in unvaccinated group)]. Fisher's exact test was used to compare attack rates for both vaccine titer levels. The O'Neill method was used to calculate vaccine efficacy confidence limits. Life table curves showing the incidence ofcases in each group were compared by Wilcoxon test. Results Serologic responses after varicella vaccination. No. with response/ no. studied (%) 164/164 (l00) 160/161 (99.4) 1/155 (0.6) GMT (95% CI) 69.6 ( )* 53.6 (47.0~61.2) 64 Titer range We studied 267 boys and 246 girls. The mean age at vaccination was 17.6 months (range, ). The vaccine and placebo groups did not differ significantly for age or sex. Twenty-one children were seen by the investigator because of suspected vaccine reactions (varicella-like papules or vesicles within 28 days after vaccination): 9 received high-titer vaccine, 6 were given low-titer (heat-exposed) vaccine, and 6 received placebo. None had fever or constitutional symptoms. Serologic analysis. Pre- and postvaccination sera were collected from 501 children. Of these, 13 samples were not obtained in the proper time (35-63 days) for a second blood sample and were excluded from analysis of serologic response. Eight children (1.6%) who were initially seropositive were excluded from seroconversion analysis. Of the remaining 480 seronegative children, 325 received vaccine and 155 were given placebo (table 1). All children who received the high-titer vaccines and all but 1 (99.4%) given the low-titer vaccines seroconverted. The GMT of postvaccination varicella antibodies was 69.6 in the high-titer vaccine recipients and 53.6 in the lowtiter vaccinees (table 1). Two (25%) ofthe initially seropositive children had a significant rise in varicella antibodies after vaccination. There was some age-related difference in the antibody response after vaccination: Younger children developed higher IIF antibody titers (P =.035). When high- and low-titer vaccine recipients were combined, the postvaccination GMT (95% confidence interval [CI]) was 66 ( ) for children ages months, 54 ( ) for those months, and 46 ( ) for those months old.

3 S332 Varis and Vesikari JID 1996; 174 (Suppl 3) Table 2. Cases of serologically confirmed varicella-like disease in the study population. No. of No. of Attack Study group children cases rate High-titer (release) vaccine % Low-titer (heat-exposed) vaccine % Placebo % NOTE. Fisher's exact test on attack rates. P <.001, high vs. placebo; P =.002, low vs. placebo; P =.005, high vs. low. Clinical efficacy. Four children in the placebo group contracted clinically apparent and serologically confirmed chickenpox between vaccination and the collection ofthe postvaccination blood specimen and were excluded from the efficacy analysis. One vaccine recipient developed a very mild case of serologically confirmed varicella (four papules on the face) before collection of the postvaccination blood sample and was likewise excluded. We also excluded 8 initially seropositive children and 7 children whose parents withdrew them from the study. The efficacy analysis was initiated from the postvaccination blood sample; the mean duration of follow-up at the time ofthis analysis was 29.3 months. A total of 493 children were included in the follow-up. Eighty-five cases of suspected varicella-like disease (case definition, ~ 1 papule or vesicle) were reported: 5 occurred before day 42, 1 was in an initially seropositive subject, and 14 were not confirmed serologically. As determined from paired sera, 65 cases were associated with a significant rise in varicella antibodies. These were regarded as serologically confirmed varicella cases. Table 2 shows the distribution of cases by study group. Forty-one placebo recipients developed varicella (attack rate, 25.5%) as did 24 children in the 2 vaccine groups. Varicella cases among vaccinees were unevenly distributed between the high- and low-titer vaccine groups (table 2): 5 and 19 cases, respectively (P =.005). Thus, the crude efficacy of the hightiter vaccine was 88% (CI, 72%-96%) versus 55% for the low-titer vaccine (CI, 31%-72%). The development of varicella in the 3 groups is shown as a function of time in the life table analysis in figure 1. The differences between the 3 groups analyzed were statistically significant. A severity analysis ofthe varicella cases showed that all but one episode of varicella-like disease in vaccinees was very mild (table 3). The exception was a vaccinee who had fever and 30 vesicles. This child had not seroconverted after vaccination with low-titer vaccine and was regarded as a primary vaccine failure. In the placebo group, 61% of the cases were associated with fever. Ofthe breakthrough cases, 32% oflowtiter vaccinees had fever. No cases in the high-titer group were associated with fever. Most breakthrough varicella cases in vaccinees presented with a papularrash with few or no vesicles. (/)Ql -Ql 0-- Ql"i '~a:l _ 90 ~~ o ~ 85 Ql> 00) ~.5 80 c:c: ~ co --E 75 QlQl Non-heat-exposed varicella vaccine o Heat-exposed varicella vaccine o Placebo 70..t--..JL...--r--...,r--~--r--r--"T"'""-r----,--r-..., a Day 42 Days Figure 1. Life table analysis ofchildren remaining free ofvaricella in high-titer (non-heat-exposed) vaccine group (A), low-titer (heatexposed) vaccine group (D), and placebo group (0) during followup. No. of children analyzed (n = 493) by group: 166, high titer; 166, low titer; 161, placebo. Wilcoxon test for equality of survival curves: P <.001, high titer vs. placebo; P =.019, low titer vs. placebo; P =.012, high vs. low titer. In contrast, in the placebo group, 61% of patients had > 30 vesicles. By group, the median number of vesicles was as follows: placebo group, 30; low-titer vaccine group, 1; and high-titer vaccine group, 2 (table 3). All but 1 vaccinee with breakthrough varicella seroconverted after vaccination. The postvaccination GMT in the vaccinees who subsequently developed varicella was 51 (95% CI, ) compared with 62 (95% CI, ) in vaccinees who did not develop varicella: however, this difference was not statistically significant (P =.37). Table 4 shows the occurrence of varicella during follow-up in relation to age at vaccination. In the placebo group, the attack rate appeared higher in older children, while in the vaccine groups there was no apparent difference in the age-specific attack rate. Discussion During the 20-year history of live varicella vaccine, it has become apparent that the vaccine will not be used extensively in high-risk children. In the United States, varicella vaccine is Table 3. Clinical features ofvaricella occurring in the seroconverted vaccinees and in the placebo group. No. of vesicles No. with Study group fever > * Median (range) High-titer vaccine (0-10) Low-titer vaccine 6 t I (0-30) Placebo (0-300) * Papules only. t 1 subject who did not respond to vaccine had varicella with 30 vesicles and fever.

4 JID 1996; 174 (Suppl 3) High-Titer Varicella Vaccine in Healthy Children S333 Table 4. Age-specific attack rate ofvaricella in the vaccinees (highand low-titer combined) and placebo recipients. Age at No. of cases/no. of subjects (%) vaccination Total Vaccine (months) children Placebo recipients Vaccinees efficacy /110 (20.0) 15/211 (7.1) 64% /36 (36.1) 5/77 (6.5) 82% /15 (40.0) 4/44 (9.1) 77% not licensed for use in immunocompromised patients, and in Europe, despite licensure, only a small number of doses are given to high-risk children. The vaccine is more reactogenic in children with malignancies than in healthy children, and two doses are often recommended for induction of satisfactory immunity [15]. Furthermore, interruption of chemotherapy for vaccination may compromise the treatment of malignancies. At the same time, the problem of severe, even life-threatening chickenpox in immunocompromised children remains, despite advances in antiviral chemotherapy [16]. Theoretically, complications of varicella in high-risk and healthy children could be averted if all children were immunized with an effective varicella vaccine at an early age. Experience with live Oka strain varicella vaccine in healthy children has been largely encouraging. The vaccine appears safe; general symptoms after vaccination are rare, and mild skin reactions occur in < 10% of recipients [7-10]. The first efficacy trial outside Japan resulted in a 100% protection rate in schoolage children over one varicella epidemic season [8]. In Japan, long-term protective efficacy appears to be excellent [17, 18], and the Oka strain vaccine is licensed for use in healthy children. However, in a previous uncontrolled study, we found that 7% of children vaccinated at ages months experienced mild varicella-like illness during 18 months of follow-up [10]. After the advent of a reformulated and more stable varicella vaccine with a release titer of 10,000 pfu, we tested whether a high-titer vaccine might evoke better protection in young children than that previously observed. For comparison, we used partially heat-inactivated vials of the same vaccine lots with a 10-fold lower titer. This was obtained by exposing the vials at 37 C for 10 days, which was thought to mimic an aging process for 2 years at refrigerator temperature. However, it is now known that in reality this vaccine does not lose much potency when stored in the refrigerator as instructed by the manufacturer; recent information from the manufacturer shows that the two lots of varicella vaccine used in this study had titers of 7940 pfu and 2540 pfu after 2 years storage at 4 C (Peetermans J, personal communication). Both vaccines were safe and immunogenic. They were also efficacious compared with placebo, although the high-titer vaccine was the more efficacious ofthe two. The follow-up period in the present study was 29.3 months, and the incidence of breakthrough cases in the low-titer vaccine group was 4.7% per year (similar to our earlier experience). In contrast, the rate of breakthrough cases in the high-titer vaccine group was only 1.2% per year. The annual incidence ofnatural varicella in the placebo group was 10.4%. While this study provides the best indication of a correlation between vaccine titer and clinical protection to date, Clements et a1. [19] had similar findings. Our results support the original concept that it is more difficult to induce durable immunity in young vaccinees, particularly those below the age of 18 months. With the exception of vaccinee age, comparison of vaccine efficacy between our study and previously published studies in the United States and Japan is difficult for several reasons. First, vaccine titers cannot be directly compared because titration methods may differ. Second, the rate ofmild clinical reinfections is dependent on the intensity of follow-up. We are confident that our surveillance (in a single study center specialty clinic) was close and sensitive and therefore that our rate ofbreakthrough varicella probably represents a maximum rate. A summary report ofus studies with an observed breakthrough case rate of 2% per year included many sites with an increased possibility ofvarying surveillance intensity [11]. In Japan, very few breakthrough cases have been observed, but there is no specific information about the nature ofsurveillance [17, 18]. It is of interest that there was a rather poor correlation between level of postvaccination varicella antibodies and protection rate. The heat-exposed low-titer vaccine was almost as immunogenic as the high-titer vaccine as shown by IIF antibody determination. A possible explanation is that the lowtiter vaccine was prepared by gentle heat inactivation, which preserved the antigenicity of the product, and the inactivated viral material may have contributed to the induction ofantibody response. Watson et a1. [20] reported that a partially inactivated varicella vaccine preserved its immunogenicity and they suggested, without efficacy data, that the lower titer would not result in loss of potency [20]. However, we found a clear difference in protective efficacy between the two titer levels, which suggests that mechanisms other than varicella antibody must contribute to induction of protective immunity. Cell-mediated immunity induced by live multiplying virus is likely to play a critical role in this respect. In conclusion, we believe the high-titer varicella vaccine is well suited for use in healthy young children since it provides a high rate ofprotective immunity. The same vaccine, ifpartially inactivated, remains immunogenic but loses some of its efficacy. However, the clinical relevance of the difference in vaccine efficacy between the high- and low-titer preparations should not be overemphasized, as most of the cases of breakthrough varicella were very mild and of little clinical significance. Furthermore, treatment of the control vaccine for this trial was more severe than necessary to mimic cold storage; the heat treatment resulted in a greater loss in titer than would 2 years of storage at refrigerator temperature. Immunization of healthy young children against chickenpox is medically justified to prevent complications of varicella that

5 S334 Varis and Vesikari JIO 1996; 174 (Suppl 3) include pneumonia and encephalitis [21]. An immunization program is likely to be cost-effective if the vaccine price is comparable to that of other childhood vaccines [22, 23]. If general immunization against varicella is done at an early age, vaccine-induced immunity should protect against illness later in life and during immunosuppressive disease. The high-titer varicella vaccine that we describe appears to be a promising candidate for general use in young children. Acknowledgments We thank study nurses Tellervo Sonnunen and Birgit Blomqvist for dedicated work, Francois Meurice (Smith Kline Beecham) for collaboration in the analysis of clinical results, and Daniele Vandevoorde for serologic studies. References 1. Takahashi M, Otsuka T, Okuno Y, et al. Live attenuated varicella vaccine used to prevent the spread ofvaricella in hospital. Lancet 1974; 2: Tzawa T, Thara T, Hattori A, et al. Application of a live varicella vaccine in children with acute leukemia or other malignant disease. Pediatrics 1977;60: Brunell PA, Shehab Z, Geiser C, Waugh JE. Administration of live varicella vaccine to children with leukemia. Lancet 1982; 2: I Gershon A, Steinberg S, Gelb L, et al. Live attenuated varicella vaccine: efficacy for children with leukemia in remission. JAMA 1984;252: Gershon A, Steinberg S, Gelb L, et ai. Use of live attenuated varicella in immunocompromised children and adults. Pediatrics 1986; 78: Gershon A, Steinberg S, NTAID Varicella Vaccine Collaborative Study Group. Persistence of immunity to varicella vaccine with leukemia immunized with live attenuated varicella vaccine. N Engl J Med 1989; 320: Arbeter AM, Starr SE, Weibel RE, et ai. Live attenuated varicella vaccine: immunization of healthy children with the Oka strain. Pediatrics 1982; 100: Weibel RE, Neff BJ, Kuter BJ, et al. Live attenuated varicella virus vaccine: efficacy trial in healthy children. N Engl J Med 1984; 310: Johnson JE, Shurin PA, Flattlar D, et al. Live attenuated varicella vaccine in healthy 12- to 24-month-old children. Pediatrics 1988;81: Vesikari T, Ohrling A, Baer M, et ai. Evaluation oflive attenuated varicella vaccine (Oka-RIT strain) and combined varicella and MMR vaccination in month-old children. Acta Paediatr Scand 1991;80: White CJ, Kuter B, Ngai A, et al. Modified cases of chickenpox after varicella vaccination: correlation of protection with antibody response. Pediatr Infect Dis J 1992; I 1: Kuter BJ, Weibel RE, Guess HA, et al. Oka/Merck varicella vaccine in healthy children: final report of a 2-year efficacy study and 7-year follow-up studies. Vaccine 1991; 9: Watson BA, Piercy SA, Plotkin SA, Starr SE. Modified chickenpox in children immunized with the Oka/Merck varicella vaccine. Pediatrics 1993; 91: Williams V, Gershon A, Brunell P. Serological response to varicella-zoster membrane antigens measured by indirect immunofluorescence. J Infect Dis 1974; 134: Gershon A. Viral vaccines of the future. PediatrClin North Am 1990;37: Feldman S, Lott L. Varicella in children with cancer: impact of antiviral therapy. Pediatrics 1987;80: Asano Y, Nagai T, Miyota T, et ai. Long-term protective immunity of recipients of the Oka strain of live varicella virus vaccine. Pediatrics 1985;75: Kamiya H, Sakurai M, Ihara T, et al. Clinical use of Oka live varicella vaccine. Acta Paediatr Jpn 1988; 30: Clements DA, Maggio MV, Ursano AM, Wilfert CM. Safety, immunogenicity and 5-year efficacy of Ok a/merck varicella vaccine in normal infants and adolescents. Pediatr Res 1993; 33: 165A. 20. Watson B, Piercy S, Soppas D, et al. The effect of decreasing amounts of live virus, while antigen content remains constant, on immunogenicity of Oka/Merck varicella vaccine. J Infect Dis 1993; 168: Jackson MA, Burry F, Olson LC. Complications of varicella requiring hospitalization in previously healthy children. Pediatr Infect Dis J 1992; 11:441 ~ Preblud SR, Orenstein WA, Koplan JP, Bart KJ, Hinman AR. A benefitcost analysis of a childhood varicella vaccination program. Postgrad Med J 1985;61(suppl 4): Preblud SR. Varicella: complications and costs. Pediatrics 1986; 78: