Vaccine 32 (2014) 924–930

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

A cohort event monitoring to determine the adverse events following administration of mouse brain derived, inactivated Japanese Encephalitis vaccine in an endemic district in Sri Lanka K.N.L.S.K. De Alwis a , M.R.N. Abeysinghe b,1 , A.R. Wickramesinghe c , P.R. Wijesinghe b,∗,1 a

Dengue Control Unit, No. 555/5, Public Health Building, Narahenpita, Sri Lanka Epidemiology Unit, 231 De Saram Place, Colombo, Sri Lanka c Faculty of Medicine, University of Kelaniya, Annasihena Road, Ragama, 11010, Sri Lanka b

a r t i c l e

i n f o

Article history: Received 6 September 2013 Received in revised form 12 December 2013 Accepted 18 December 2013 Available online 7 January 2014 Keywords: JE Incidence AEFI Sri Lanka Cohort-event monitoring

a b s t r a c t Introduction of human immunization reduced Japanese Encephalitis (JE) cases dramatically in Sri Lanka. However, the increased reporting of adverse events following immunization (AEFI) affected vaccine acceptance by the community. Against this background, we describe the incidence of overall AEFI and incidence and profile of AEFI, thought to be causally related to the mouse-brain derived JE vaccine. A follow-up of 9798 vaccine recipients was performed for a period of two weeks post-vaccination. Parents self-recorded observed signs and symptoms. The self-records were collected by trained supervisors. All monitored children who manifested symptom/s were investigated in details by medical officers experienced in AEFI investigations within two weeks after ending the follow-up period. Using the results of the investigation, the causality assessment was performed. The estimated cumulative incidence rate of overall AEFI was 8.6 children per 100 immunizations. The same for observed AEFI consistent with causal association to the inactivated JE vaccine was 4.3 children (95% CI—3.9–4.7%) per 100 immunizations. The most frequent AEFI was fever (81%). The frequency of high fever (>102 ◦ F) was 26%. Other major AEFI were body ache (22%) vomiting (21%), urticaria (19%), pruritus (5%), and headache (5%). Though 83% of children with AEFI thought to be causally related to the vaccine sought medical care, only 6.6% required hospitalizations. The incidence rate of AEFI in the cohort event monitoring was several-fold higher than that reported through the national AEFI surveillance system. The incidence rate of allergic manifestations among SriLankan children approached what was reported for non-endemic settings and was higher than in other JE endemic populations elsewhere. Contrary to the belief of medical practitioners and the general public, incidence of seizures was low and vaccine related other neurological manifestations were absent. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Japanese Encephalitis (JE) is the most frequent cause of viral encephalitis in Asia [1]. Since the first outbreak in 1971, there have been no JE outbreaks in Sri Lanka till 1985 [2]. Following three consecutive, major outbreaks in 1985–1987, an immunization campaign with the mouse-brain derived (MBD) JE vaccine was launched in 1988. The MBD JE immunization schedule consisted

∗ Corresponding author. WHO SEARO IVD, Indraprastha Estate, New Delhi, India. Tel.: +91 11 4340000/+91 9810402943 (mob). E-mail addresses: [email protected] (D.A. K.N.L.S.K.), [email protected] (A. M.R.N.), [email protected] (W. A.R.), [email protected], [email protected] (W. P.R.). 1 Present Address: Immunization and Vaccine Development, Regional Office for South, East Asia, World Health Organization, New Delhi, India. 0264-410X/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.vaccine.2013.12.047

of a primary series of two doses offered at an interval of 1–4 weeks followed by two booster doses offered one and four years after the primary series, respectively. As a result, JE cases declined from 812 (incidence rate of 4.7/100,000) in 1987 [3] to 26 sporadic cases (incidence rate of 0.1/100,000) in 2006 [4]. Sri Lanka conducted annual immunization campaigns with the MBDJE vaccine (Beijing-1-strain) in 18 of the 26 districts until transitioning to the island-wide, routine immunization with the live attenuated JE vaccine (LAJEV) SA 14-14-2 in 2009 [5]. In spite of the high immunization coverage and the steady decrease in the disease burden, the reported morbidity due to adverse events following immunization (AEFI) started to impact the JE immunization campaign. Through the routine, passive AEFI surveillance system, the programme managers initially observed a gradual increase in the JE vaccine specific AEFI rate from 5.1/100,000 (1998) to14.6/100,000 immunizations (2002) [6]. In subsequent 4 years, the JE vaccine specific AEFI rate was

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930

30.8/100,000, 57.6/100,000,192.6/100,000 and 92.5/100,000 immunizations, respectively [5]. The gradual increase in JE vaccine specific AEFI rates till 2002 may be partly explained by the improved reporting of AEFI as a result of continuous strengthening of the AEFI surveillance system. However, the same could not be attributed to the increase in AEFI rates in subsequent 3 years. The abrupt increase in JE vaccine specific AEFI rates during 2003–2005, in particular in 2005, coincided with high numbers of systemic allergic manifestations and to a lesser extent with seizures. The reporting rate of allergic manifestations increased from 23.8/100,000 immunizations in 2003 to 125.6/100,000, and 57.4/100,000 immunizations in 2005 and 2006, respectively. This is in contrast to the reporting rates of allergic reactions of 0.29 and 0.56/100,000 immunizations observed in 1998 and 2002. While the proportion of allergic manifestations among all JE specific AEFI remained 5.8% and 5.5%, respectively, in 1998 and 2002, the same was higher in the range of 62% (2006)–76% (2003) in subsequent years, indicating the contribution of allergic manifestations to the abrupt increase in MBD JE specific AEFI rates after 2002. To a lesser extent, seizures which accounted for a reporting rate of 0.7/100,000 immunizations in 2003 had increased to 13.2/100,000 and 6.9/100,000 immunizations in 2005 and 2006 [5]. Higher numbers of allergic manifestations and seizures among vaccinees than seen in previous years were initially observed by the paediatricians. This led to their expression of safety concerns regarding the MBD JE vaccine. Subsequently, it resulted in intensified surveillance on these AEFIs. The resultant publicity had a negative impact on acceptance of the vaccine by parents due to fear of ill-effects following immunization [7,8]. In the study area, Anuradhapura district, the immunization coverage of the first JE dose declined to 93.2% in 2005 as compared to the median coverage of 98.4% in 1997–2004. Even further declines were seen for the second dose to 87.4% as compared to the median coverage of 98.2% during the same 7 year period of reference [9]. The reluctance of health workers to provide the second dose to those who manifested with allergic reactions and seizures following the first dose and the reluctance of some parents to vaccinate their children with the second dose as a result of the adverse publicity regarding allergic reactions attributed to the dramatic decline in the second dose. In this context, quantification of the burden of AEFI attributable to the MBD JE vaccine was essential to conclude if there was really a safety issue. Second, describing the profile of individual AEFI was required to substantiate or alleviate perceived ill-effects of this vaccine by parents, medical practitioners and health workers. Third, irrespective of the safety issues of MBDJE vaccine, due to cost considerations, there was a need for evaluating a safe and cost-effective alternative JE vaccine. All these were deemed necessary for determining the strategic direction of the national JE immunization campaign. In view of above, (a) we designed a study to determine the incidence of overall AEFI, incidence and profile of AEFI thought to be causally related to the MBD JE vaccine in a cohort of vaccinees in 2006 (b) conducted a study on safety and efficacy of the alternative LAJEV SA-14-14-2 in 2007 and (c) conducted post-marketing surveillance of AEFI due to LAJEV from 2009 to 2012. After completing the strategic decision making process based on this series of studies, we describe the first study on safety of the MBDJE vaccine conducted in the JE endemic, Anuradhapura district in Sri Lanka during the annual JE immunization campaign conducted in July–August 2006.

2. Materials and methods In annual JE immunization campaigns, children aged 1–10 years were targeted for either the primary series or booster doses of the

925

MBD JE vaccine (Beijing-1 strain). The primary series consisted of two doses offered at an interval of 1–4 weeks on completion of the first year of life. The first booster dose (3rd dose) was due one year after the primary series and the second booster (4th dose) was recommended four years after the primary series. Accordingly, our study population comprised (1) children aged one year who received their primary doses, (2) children aged two years who received their first booster dose, (3) children aged five years who received their second booster dose and additionally, (4) any child who did not belong to these age groups but was under 10 years and received any of the due MBDJE vaccine doses from the public sector immunization clinics in the Anuradhapura district in 2006. The total target population for the immunization campaign was 55,055. Children resident in other districts who received JE vaccines from Anuradhapura district and private sector immunizations were excluded. These children received no other concomitant vaccines. Since all children immunized in the entire district could not be followed up, we selected a study sample. The required sample size was determined based on the population proportion of AEFI due to MBDJE Vaccine with a specified relative precision [10] on the assumption that the observed prevalence was equal to incidence given that the duration of majority of AEFI was of a very short duration (prevalence = incidence × duration) [11]. Due to the limitations of estimates derived from the national AEFI surveillance system, we selected more valid, research-based estimate of systemic AEFI due to MBDJE vaccine (4.4%) reported in USA and Thailand [12] as the population proportion. The required sample size when compensated for the possible loss to follow-up was 10,200. The required study sample was selected from 4 of the 19 Medical Officer of Health (MOH) areas in the district that performed the highest number of JE immunizations in 2005. The number of participants to enrol from each of four selected MOH areas was determined proportionate to the number of JE immunizations performed in each MOH area in 2005. Subsequently, 3465 from Anuradhapura NPE, 2940 from the Anuradhapura NPC, 2415 from Medawachchiya and 1680 were from Thalawa MOH areas were enrolled for the study. All children who were administered MBD JE vaccine in public sector immunization clinics in selected MOH areas whose parent/s consented for participating in the study were prospectively and consecutively enrolled for follow-up until the required sample size was achieved. Given that the campaigns were held in two rounds of one-week duration, two-weeks apart, we had to enrol our study sample in both rounds. The short duration of a round enabled us to enrol early and late vaccinees while enrolment in both rounds facilitated including recipients of first and second doses in addition to the recipients of booster doses. Parents/guardians were explained about the study, voluntary nature of participation and non-influence of their decision on up-take of subsequent vaccines in the event of not giving consent for participation. Parents of participating children signed a consent form. The majority (99.8%) consented based on the explanation that their participation contributed to a national cause and due to nonavailability of any invasive procedures in the follow-up. In the first stage, enrolled participants were followed-up for a period of two weeks post-vaccination to collect information on AEFI due to the MBD JE vaccine. In line with the national and global standard case definition, an AEFI was defined as any untoward medical occurrence subsequent to JE immunization during the follow-up period that did not necessarily have a causal relationship with the JE vaccine [13]. In this context, any unfavourable or unintended sign, symptom or abnormal laboratory finding was considered as an AEFI. For soliciting AEFI, a pre-tested, self-administered questionnaire with a self-recorded diary to record basic information on occurrence of any symptom and sign during the follow-up period of two weeks was handed over to parents/guardians of participants. The

926

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930

diary contained a user-friendly list of symptoms that may have occurred after JE vaccination and the possible time of onset from vaccination for reference. Additionally, it provided space for recording any other sign or symptom external to the provided list and time of onset from the vaccination. Trained surveyors distributed the questionnaire and explained parents/guardians as to how they should record symptoms and fill the questionnaire. Additionally, parents were instructed how axillary temperature should be measured twice daily and recorded in Fahrenheit. At the completion of two weeks, surveyors visited houses of each participant and collected questionnaires irrespective of children developing any sign or symptom. They inquired parents/guardians as to if study participants developed any sign or symptom and cross-checked the same in the questionnaire to verify if parents have recorded the event. In the second stage, children who manifested any sign or symptom during the follow-up period were subject to a detailed investigation. Four medical officers (MO) experienced in AEFI investigation visited houses of children having signs and symptoms. They interviewed parents/guardians or caregivers of these children. While verifying the reported signs and symptoms, they also requested permission to verify medical records wherever required. In these cases, they visited any health institution where the child was brought for care for the reported sign or symptom, met with the physician/general practitioner (GP), sought their opinion and reviewed available records for any valid diagnosis or detecting any abnormal laboratory finding. The interviews occurred within two weeks after collecting the self-administered questionnaire. In the third stage, the causality assessment was performed by a panel comprising four MOs who conducted the field investigation and four epidemiologists from the National Programme of Immunization (NPI). They were supported by a pharmacologist and a paediatrician. In the causality assessment, they reviewed available clinico-epidemiological and laboratory information elicited in the second stage of data collection. MOs provided their opinion on each reviewed case. For selected cases, the panel sought the advice of relevant sub-specialists. By using Brighton Collaboration case definitions for fever, local reactions, injection site pain, diarrhoea, seizures, encephalitis and anaphylaxis, we standardized definitions of these AEFI before they underwent causality assessment. For other reported signs and symptoms, nationally available standard case definitions were used. Initially, we considered certain, probable/possible adverse events following JE immunization as causally associated based on the WHO causality assessment classification existing at the time of the analysis. However, with the introduction of the new WHO classification [13], we re-classified all AEFI as those with inconsistent causal association to, indeterminate, non-classifiable and those with consistent causal association to MBD JE vaccine using available level of information for the present report. We shared overall results of the causality assessment and presented details for all serious AEFI that accounted for 11% to the National AEFI committee. We defined all AEFIs that caused deaths, hospitalizations, prolongations of existing hospitalizations, persistent incapacity and events that were potentially life-threatening as serious AEFIs. The schematic representation of investigated JE immunizations is given in Fig. 1. Based on this investigation, we calculated the cumulative incidence of (1) overall adverse events following JE vaccination (reported events) and (2) AEFI consistent with causal association to the MBD JE vaccine. We used the number of children manifested with at least one AEFI as the numerator in both calculations. Though there were children with more than one AEFI or distinct periods of illness during the follow-up period, we considered neither the number of symptoms nor the number of distinct periods of illness in the numerator for these calculations. Considering that some children manifested with more than one AEFI that were

causally consistent with MBD JE vaccine, additionally, we described the profile of these AEFI (sign/symptom) consistent with causal association to the MBD JE vaccine, determined their prevalence among vaccinees who had AEFI consistent with causal association to the MBD JE (n = 422) and calculated their cumulative incidence. In order to calculate the incidence, we used the frequency of each sign/symptom/abnormal laboratory finding thought to be causally associated with the MBD JE vaccine as the numerator while the number of immunizations was considered as the denominator. When there was (a) more than one sign/symptom in a child or (b) more than one distinct period of illness during the follow up period, we considered all manifested symptoms in the former and also the same in the distinct period of illness concluded as causally consistent with the MBDJEV for describing the profile and calculating the cumulative incidence of all individual AEFI thought to be causally associated with the vaccine. This study was approved by the Ethical Review Committee of the Faculty of Medicine of the University of Colombo (EC/05/017).

3. Results The target population for the immunization campaign in the district was 55,055 children. Of the 10,200 selected children, 10,181 (99.8%) children who received MBDJE vaccine were enrolled based on parental consent. However, only 9798 (96.2%) self-administered questionnaires and the diaries were completed at the end of the follow-up period. Three hundred and eighty three (3.8%) did not respond. The basic information of non-respondents [age (p = 0.53), sex (p = 0.29), MOH area (p = 0.57) and the dose of JE vaccine (p = 0.46)] collected at the enrolment were not systematically different from the respondents. The number of children who manifested any AEFI during the follow-up was 843. This amounted to a cumulative incidence of 8.6 (95% CI = 8.0–9.1) per 100 vaccinations. In the detailed investigation, 51 of these 843 children could not be contacted. Thus, among 792 children who were subject to detailed investigations, the number (proportion) that had AEFI with consistent causal associations to the MBDJE vaccine was 422 (53%). The corresponding cumulative incidence rate was 4.3 (95% CI = 3.9–4.7) per 100 vaccinations. The cumulative incidence of AEFI with consistent causal association to the MBD JE vaccine by the age, sex of vaccine recipients and scheduled dose is given in Table 1. Cumulative incidence was similar among males (4.3/100; 95% CI—3.7–4.9/100) and females (4.4/100; 95% CI—3.8–4.9/100). The highest cumulative incidence was found among children aged 1up to 3-years (5.1/100; 95% CI—4.6–5.6/100) and 5- up to 7-years (4.6/100; 95% CI—3.0–6.1/100). As regards the doses, the lowest cumulative incidence was reported for the fourth dose (2.7/100; 95% CI—1.8–3.6/100). The profile of observed AEFI with consistent causal association to the MBD JE vaccine and their cumulative incidence are given in Table 2. Fever was the most common symptom observed among children having AEFI thought to be causally associated with the MBD JE vaccine. A little over a half (55%) of these children experienced fever in the range of 100.4–102 ◦ F while slightly above one fourth (26%) reacted to vaccination with fever exceeding 102◦ F. The cumulative incidence of fever was 2.4 and 1.1/100 immunizations, respectively. Allergic conditions manifested as urticaria accounted for 19% (cumulative incidence—0.8/100 immunizations) while seizures comprised only 1.4% (cumulative incidence—0.06/100 immunizations) in children having AEFI causally associated with the JE vaccine. The median duration of onset of urticaria was two days. The proportion of children with adverse events causally associated to the MBD JE vaccine that sought medical care was 83%. The majority (98%) of these had sought out-patient care either from an

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930

927

The number of study parcipants idenfied to be enrolled 10200

The number enrolled for the study based on parental consent 10181

The number completed quesonnaires/diaries 9798

Number reported AEFI

Number did not report any AEFI

843

8955

Number invesgated in details

Number could not be traced 51

792

Inconsistent causal associaon to immunizaon (coincidental)

Consistent causal associaon to immunizaon

370

422

Fig. 1. Schematic representation of status of investigated JE immunizations.

Table 1 The cumulative incidence of AEFI with consistent causal association to the MBDJE vaccine by age, sex and scheduled dose of JE vaccine. Variable

Sex Male Female Total Age (in years) 1- up to 3 3- up to 5 5- up to 7 7- up to 9 9 up to 10 Total Dose 1st dose 2nd dose 3rd dose 4th dose Total

Number of immunizations

Number of children with AEFI reported

Number of children with an AEFI causally associated with MBD JE vaccine

Cumulative incidence per 100 immunizations per 2 weeks (95% CI)

5076 4671 9747

406 386 792

217 205 422

4.3 (3.7–4.9) 4.4 (3.8–4.9) 4.3 (3.9–4.7)

7112 629 698 668 640 9747

677 40 58 11 6 792

365 16 32 7 2 422

5.1 (4.6–5.6) 2.5 (1.3–3.7) 4.6 (3.0–6.1) 1.0 (0.2–1.8) 0.3 (0–1.1) 4.3 (3.9–4.7)

4249 2412 1727 1359 9747

335 209 136 112 792

171 127 88 36 422

4.0 (3.4–4.6) 5.4 (4.5–6.3) 5.2 (4.2–6.3) 2.7 (1.8–3.6) 4.3 (3.9–4.7)

928

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930

Table 2 Profile and cumulative incidence of AEFI (sign/symptom) with consistent causal association to the MBD JE vaccine. AEFI (sign/symptom) with consistent causal association to MBD JE vaccine

Number (%) among all children having AEFI with consistent causal association to MBD JE vaccine (n = 422)

Cumulative incidence rate per 100 immunizations per 2 weeks (95% CI)a

Fever 100.4–102 ◦ F Fever >102 ◦ F Body achesb Vomiting Urticaria Pruritus Headacheb Diarrhoea Abdominal painb Difficulty in breathingb Swelling at the injection site Seizures Dizzinessb

232 (54.9) 110 (26.0) 94 (22.3) 90 (21.3) 79 (18.7) 21 (4.5) 21 (4.5) 13 (3.0) 11 (2.6) 8 (1.9) 8 (1.9)

2.4 (2.0–2.7) 1.1 (0.9–1.3) 0.9 (0.7–1.1) 0.9 (0.7–1.1) 0.8 (0.6–1.0) 0.2 (0.1–0.3) 0.2 (0.1–0.3) 0.1 (0.03–0.16) 0.1 (0.03–0.16) 0.08 (0.02–0.14) 0.08 (0.02–0.14)

6 (1.4) 4 (0.9)

0.06 (0.01–0.11) 0.04 (0–0.07)

a Denominator for calculating the incidence rates was number of immunizations which was 9747. b These subjective events are difficult to be determined in non-verbal vaccine recipients in lower age groups. Hence, the cumulative incidence of these events could be under-estimates.

independent GP (59%) or Out Patient Departments (OPD) of government hospitals (39%). The proportion of children who required hospital admissions for in-patient medical care was only 6.6%. 4. Discussion The evolving situation of low acceptance of JE immunization due to safety concerns demanded valid, empirical evidence on AEFI before strategizing to boost the confidence of general public on JE vaccination in endemic districts. For this purpose, an active, cohort event monitoring was required over the passive, routine AEFI surveillance given its inherent limitations. Our data collection method that entailed self-recording of signs and symptoms during the follow-up period by parents supplemented with a verification process by surveyors at the end of the follow-up was more sensitive than the passive AEFI reporting. However, the study had its own limitations. The absence of a control group was a major limitation of the study making it impossible to determine whether rates of AEFI were higher than the background rates or rates experienced following other routine childhood vaccines. However, given that nearly a half of the reported AEFI were causally inconsistent with MBDJE vaccine in the causality assessment, this may be considered a limited proxy to having a control group. Our data collection method too had some limitations in terms of completeness and reliability. Since case investigations were performed 2–4 weeks after the immunization, investigators were not able to observe and verify manifestations reported by parents. Some GPs who did not maintain patient records had difficulties in recalling events 4 weeks after the event. The incompleteness of self-reporting was well demonstrated in the case of 21 events of pruritus as opposed to the 79 urticarial manifestations given that usually urticaria is associated with pruritus. The other limitation was that we were not able to perform any auxiliary laboratory investigation on the advice of sub-specialists to assist the causality assessment as more than 4 weeks have passed since immunization at the time of causality assessment. Thus, our method of causality assessment using available levels of information and observations of MOs who investigated each AEFI case with above stated limitations was, to a certain extent, crude,

and relatively low in specificity, though it was consistent with the national practice and programmatically feasible. The observed cumulative incidence of adverse events following MBDJE vaccine (8.6/100 immunizations) could still be an underestimate given the possibility of under-reporting of subjective symptoms in lower age groups. Nevertheless, we could convincingly conclude that only half of the reported AEFI were consistent with a causal association to the incriminated vaccine (4.3/100 immunizations; 95% CI—3.9–4.7%). This underscores the fact that without strengthening the causality assessment of reported AEFI, many coincidental events have the potential to be incorrectly attributed to the MBD JE vaccine. Our estimate of the adverse events following MBD JE vaccine (8.6/100 immunizations) confirmed the existing under-reporting of the same (range: 0.005–0.19/100 immunizations) through the national AEFI surveillance system from 1998 to 2006 [5,6]. In general, the MBD JE vaccine has been considered relatively safe, although local reactions and mild systemic side effects tend to occur among some recipients [14]. To ensure this safety, the infected mouse brain suspension is purified in the manufacturing process extensively and the myelin based protein content is controlled below 2 ng/ml. However, gelatine which is used as a stabilizer is connected to occurrence of allergic reactions observed in recipients [15,16]. According to the WHO position paper [17], local reactions such as tenderness and swelling are experienced in about 20% of vaccinees and mild systemic symptoms appear in a similar proportion of vaccinees. Nonetheless, evidence on AEFI has been generated largely in non-endemic settings mainly among adults and travellers using varying products and batches. These make comparisons complex in endemic settings especially for child rates. Despite this complexity, the proportion of children with any AEFI in our cohort was lower than that is in the WHO position paper. Though hypersensitivity reactions have been a major concern especially among vaccinees from non-endemic settings [15,18–22], the cumulative incidence of potential allergic manifestations in the form of urticaria, pruritus and difficulty in breathing (0.8, 0.2 and 0.08/100 immunizations) causally consistent with JE vaccine observed in our cohort was slightly higher than the range of 1–64/10,000 immunizations reported in the literature[17,20,22,23]. However, it has to bear in mind that many of these global rates have the potential to be under-estimates given that they were generated in passive reporting in addition to the variations in study settings, populations, vaccine type, batches and number of doses. In our study, systemic allergic manifestations were mainly non-immediate type reactions. Though we did not measure anti gelatine antibodies in sera, as Sakaguchi et al. pointed out [15], some of these children would have been sensitised to gelatine during exposure to previous JE doses and other vaccines used in NPI containing gelatine as a heat stabilizer (measles, measles–rubella, and rubella). The reported high absolute numbers and rates of allergic reactions to the LAJEV which is also a freeze dried product with gelatine stabilizer observed in the postmarketing surveillance in 2009(68.8/100,000), 2010(65.4/100,000) and 2011(72/100,000) [5] indirectly suggest the possible biological basis of gelatine for the AEFI due to both JE vaccines in Sri Lankan children, Our estimate for urticaria could have been affected by the potential misclassification in the causality assessment using supplementary information collected based on parents’ and physicians’ recall for some patients. The self-reported pruritus was definitely an under-estimate reflecting the limits of data collection method. Though, our findings approach what is reported in the literature for non-endemic settings and indicate that the concerns of the medical community about the allergic reactions may indeed be unfounded, manifestations of allergic reactions in Sri-Lanka need special attention given that urticaria has not been reported as a common AEFI

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930

in Asia where MBDJEV is widely used. The study also confirmed the under-reporting of allergic manifestations through the routine AEFI surveillance system (range: 0.02–0.12/100 immunizations) [5]. Given that hypersensitivity is a major, cited reason for low vaccine acceptance in Sri Lanka [7], in the light of our findings, we recommended as demonstrated in Brazil [24], the importance of strengthening the routine AEFI surveillance linked with mass vaccination campaigns to systematically capture these allergic manifestations, collecting serum for Ig E testing where feasible, performing detailed investigations for causality assessments, preparing country specific guidelines on vaccinating these children with remaining primary and booster JE doses and ensuring preparedness for field level management of anaphylactic reactions. No vaccine related encephalitis cases were reported in this study. This is consistent with study results in Japan from 1955 to 1966 although several cases of acute encephalitis temporarily linked to JE vaccination have been reported in the Republic of Korea [25]. However, our study would have been under-powered to detect encephalitis cases with a sample size calculated based on the 4% anticipated proportion of overall AEFI. Though seizures are often reported as a common occurrence, the cumulative incidence was as low as 0.06/100 immunizations. However, with 26% of AEFI being fever exceeding 102 ◦ F, some of these children could have been potentially vulnerable to febrile seizures. Similar to results in previous studies [26], we too observed no notable differences of AEFI after the first, second and third doses. AEFI were twice less than among recipients of the fourth dose than the first, second or third doses. This could possibly be explained by non-administration of MBD JE vaccines to those who manifested AEFI to previous doses or non-acceptance of the same by parents of these children. On the other hand, since the fourth dose is given when the child is at least 5 years old, lower rates of adverse events after the 4th dose may be linked to the age rather than the dose number. The proportion of children with adverse events causally associated to the MBD JE vaccine that sought medical care was very high. It may be possible due to the fact that only those events that had sufficient documentation were able to meet criteria for causal association and seeking medical care could have enabled achieving that level of documentation. Despite the high level of seeking medical care, our study revealed that only 6.6% of AEFI thought to be causally related to the JE vaccine required hospitalization which is a criterion for serious AEFI for prolonged and specialized medical care. This amounted only to a cumulative incidence rate of 0.29 (95% CI = 0.18–0.39) per 100 immunizations. The study concluded that the cumulative incidence of overall adverse events following MBD vaccine was many fold higher than the estimates derived through the national AEFI surveillance system. A half of these AEFI were coincidental events with inconsistent causal association to the incriminated vaccine. The MBD JE vaccine was relatively safe and only the incidence of observed allergic manifestations substantiated reported evidence of the medical community. This safety was also supported by the fact that the cumulative incidence rate of AEFI that required hospitalizations for in-ward medical care was as low as 0.3 per 100 immunizations. These study results and conclusions were reviewed by the national Advisory Committee on Communicable Disease (ACCD). Based on the recommendations of the ACCD, a series of activities was conducted to alleviate the safety concerns of health practitioners and the general public with a view to boosting their confidence on the MBDJEV. In spite of this, given that there was an abrupt increase in allergic manifestations reported in 2004–2005, due to the availability of excessive amounts of allergens in at least some batches, the NPI strengthened the post-JE campaign AEFI surveillance island-wide, revised overall guidelines for JE immunization and in particular, for immunizing those with an allergic history

929

in subsequent years. With transitioning to the live JE vaccine SA14-14-2 in 2009, MBD vaccine is no longer available in the public programme. However, MBD vaccine is still offered to children in the private sector. Hence, the findings of the present study remains valuable to date even after the transition as a reference for the baseline rates of MBDJE vaccine specific AEFI in Sri Lanka. Conflict of interest statement All four authors assure that they do not possess potential conflicts of interests with any individual or agency that could have inappropriately influenced the work related to this submission. Contributors NA in his official capacity as the national EPI manager conceptualized the need for this study for Sri Lanka. All four authors were involved in developing the methodology, data analysis, and interpretation of data. KNLSK De A implemented, monitored and supervised the field work. KNLSK De A, PRW and NA were involved in causality assessment. ARW acted as an advisor in statistical analysis in addition to his involvement in all other aspects of the study. PRW prepared the manuscript which was critically reviewed by other authors. The final draft was approved by all four authors for submission. Acknowledgement This study was financially supported by the Program for Appropriate Technology for Health (PATH), USA. However, PATH did not involve in the study design, data collection, data analysis, interpretation and report writing. Nor did they influence the decision to publish the work. Though NA and PRW are currently employed by the WHO, the opinions expressed in this article are those of the authors and do not represent the opinion of the World Health Organization. References [1] Tsai TF. Japanese Encephalitis vaccines. In: Plotkin SA, Mortier E, editors. Vaccines. Philadelphia, PA: WB Saunders; 1994. p. 671–713. [2] Abeysinghe MRN. Aspects of the outbreak of Japanese Encephalitis in Anuradhapura district. In: Doctor of medicine. Colombo, Sri Lanka: Post-graduate Institute of Medicine (PGIM); 1987, submitted. [3] Epidemiology Unit. Epidemiology and control of Japanese Encephalitis. Weekly Epidemiological Report 1993;93(17):1–2. [4] Epidemiology Unit. Surveillance report on Japanese Encephalitis—2007. Epidemiological Bulletin (Sri Lanka) 2008;49(3):17, http://www.epid.gov.lk/web/ (Accessed 13 attachments/article/167/Volume 49 3rd qtr 2008.pdf September, 2013). [5] Epidemiology Unit. Japanese Encephalitis: a manual for Medical Officers of Health. Colombo: Epidemiology Unit, http://www.epid.gov.lk/web/ attachments/article/141/JE%20book.pdf; [accessed 24 August, 2013]. [6] Epidemiology Unit. Adverse events following JE vaccine. Weekly Epidemiological Report 2005;32(10):1–2. [7] Epidemiology Unit. National Immunization Summit, Sri Lanka. In: Proceedings of the 2nd meeting of the immunization stakeholders. Epidemiology Unit, Ministry of Health; 2007. [8] Abeysinghe MRN, Wijesinghe PR. The landscape of new vaccines for Japanese Encephalitis: country-level strategies for introduction. International Journal of Infectious Diseases 2008;12:e34, http://www.sciencedirect.com/science/ article/pii/S1201971208002567 (Accessed 24 August, 2013). [9] DeAlwis KNLSK. To evaluate the effectiveness of the Japanese Encephalitis immunization programme. In: Doctor of medicine. Colombo, Sri Lanka: Postgraduate Institute of Medicine (PGIM); 2006, submitted. [10] Lwanga SK, Lemeshaw S. Sample size determination in health studies: a practical manual. Geneva: World Health Organization; 1991. [11] Hennekens CH, Burring JE. Epidemiology in medicine. 1st ed. Boston, MA: Lippincott Williams & Wilkins; 1987. [12] Halstead SB, Tsai TF. Japanese encephalitis vaccines. In: Plotkin SA, Mortier E, editors. Vaccines. Philadelphia, PA: WB Saunders; 2004. p. 919–58. [13] WHO. Causality assessment of adverse event following immunization (AEFI). In: User manual for the revised WHO classification. Geneva: WHO; 2013,

930

[14] [15] [16]

[17]

[18] [19] [20]

D.A. K.N.L.S.K. et al. / Vaccine 32 (2014) 924–930 www.who.int/entity/vaccine safety/publications/aevi manual.pdf (Accessed 24 August, 2013). Gupte S. Recent advances in paediatrics. In: Special volume 18-pediatric neurology. 1st ed. New Delhi: Jaypee Publishers; 2008. Sakaguchi K, Nakashima H, Takahashi T, Nakayama H, Fujita S. Inoyue. Allergy 2001;56:804. Tauber E, Kollaritsch H, von Sonnenberg F, Lademann L, Jilma B, Firbas C, et al. Randomised, double blind, placebo controlled, phase 3 trial of the safety and tolerability of IC51, an inactivated Japanese Encephalitis vaccine. Journal of Infectious Diseases 2008;198:493–9. WHO. WHO position paper on Japanese Encephalitis vaccines. Weekly Epidemiological Record 2006;81:325–40, http://www.who.int/immunization/topics/ japanese encephalitis/en/index.html. Monath TP. Japanese encephalitis vaccines: current vaccines and future prospects. Current Topics in Microbiology and Immunology 2002;267:105–38. Plesner AM. Allergic reactions to Japanese Encephalitis vaccines. Immunology Allergy Clinics in North America 2003;23(4):665–97. Plesner AM, Ronne T. Allergic muco-cutaneous reactions to Japanese encephalitis vaccine. Vaccine 1997;15(11):1239–43.

[21] Zanin MP, Webster DE, Martin JL, Wesselingh SL. Japanese encephalitis vaccines: moving away from the mouse brain. Expert Review of Vaccines 2003;2(3):407–16. [22] Centers for Disease Control. Inactivated Japanese Encephalitis virus vaccine. Recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbidity and Mortality Weekly Return 1993;42(No. RR1):8–10. [23] Tsai TF, Chang GJ, Yu YX, Plotkin SA. Japanese Encephalitis vaccines. In: Plotkin SA, Orenstein WA, editors. Vaccines. Philadelphia, PA: WB Saunders; 1999. p. 672–703. [24] Freitas DR, Moura E, Araújo G, Cardoso A, Scheidt P, Ferraz E, Madalosso G, Chen RT, Hatch DL. Investigation of an outbreak of hypersensitivity-type reactions during the 2004 national measles–mumps–rubella vaccination campaign in Brazil. Vaccine 2013;31(6):950–4. [25] Sohn YM. Japanese encephalitis immunization in South Korea: past, present and future. Emerging Infectious Diseases 2000;6(1):17–24. [26] Northdurft HD, Jelinek T, Manchang A. Adverse reactions to Japanese Encephalitis vaccine in German travellers. Journal of Infection 1996;32: 119–22.