- Email: [email protected]
Improved immunization practices reduce childhood hepatitis B infection in Tonga
Vaccine 27 (2009) 4462–4467
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Improved immunization practices reduce childhood hepatitis B infection in Tonga Niklas Danielsson ∗ , Toakase Fakakovikaetau, Edit Szegedi World Health Organization, WHO Office Cambodia, P.O. Box 1217, Phnom Penh, Cambodia
a r t i c l e
i n f o
Article history: Received 4 April 2009 Received in revised form 3 May 2009 Accepted 10 May 2009 Available online 7 June 2009 Keywords: Hepatitis B vaccine HBsAg Elimination target Convenience testing Children Birth dose
a b s t r a c t Background: Hepatitis B infection is hyper-endemic in Tonga and 19% of pregnant women test positive for hepatitis B surface antigen (HBsAg). Routine childhood immunization against hepatitis B was introduced in 1989 and the target for elimination was set at <1% HBsAg prevalence in children. A study conducted in 1998, a decade after the introduction of hepatitis B immunization, found the HBsAg prevalence to be 3.8% in pre-school children. The finding resulted in the strengthening of the delivery of hepatitis B vaccine with emphasis on providing the first dose within 24 h after birth. The aim of this study was to measure the impact of improved immunization practices on the prevalence of hepatitis B infection in pre-school children, and to assess the progress towards hepatitis B elimination in Tonga. Measured outcome: Prevalence of HBsAg antigen. Type of study: Cross-sectional study. Methods: Children aged 6–59 months who were admitted to Vaiola Hospital, Nuku’alofa, Tonga, and had blood collected for clinical investigation, were tested for HBsAg with a rapid serological test (Abbott Determine® ). A total of 449 children were recruited and interviewed and 375 (84%) were tested for HBsAg. Convenience testing was chosen, as it was likely to be a relatively unbiased method in this situation where all children on the island have good and equitable access to hospital services. Immunization status was checked against the children’s immunization cards and cross-checked against the records kept by the public health nurses. Information about socio-economic status, parent education, blood transfusion, breast-feeding, mode of delivery, and place of birth was collected through interviews with mothers using a standardized questionnaire. Results: Three children tested positive for HBsAg resulting in a prevalence of 0.8% (CI 0.2–2.5%). Hepatitis B 1 (Hep B 1) immunization coverage was found to be high, 99.1% (CI97.7–99.7) and 91.9% (CI 88.9–94.2) of children received the first dose of hepatitis B vaccine within 24 h after birth. Coverage for the third dose of hepatitis B vaccine was 97.6% (CI 95.5–98.7) and of the children with complete immunization 84.7% (CI 80.9–87.9) had received all three doses by 6 months of age. Conclusions: The results show that the instituted changes in the delivery of hepatitis B vaccine have been effective in reducing the transmission of hepatitis B to children and indicate that Tonga appears to have achieved the elimination target of less than 1% HBsAg prevalence in children. This convenience survey used a simple, cost effective and reproducible study design. Convenience testing can be recommended for surveillance of the effectiveness of immunization programmes in settings where the population has good access to health services. © 2009 Elsevier Ltd. All rights reserved.
1. Introduction Approximately two billion people, or one third of the world’s population, have serologic evidence of past or present hepatitis B virus (HBV) infection [1]. WHO estimates that 367 million people worldwide are chronic carriers of HBV and that 500 000–700 000 people die each year as a result of HBV infection [2,3]. Mortality
∗ Corresponding author. Tel.: +855 012443 993; fax: +855 023216211. E-mail address: [email protected] (N. Danielsson). 0264-410X/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2009.05.051
from HBV is pre-dominantly from liver cirrhosis and hepato-cellular cancer (HCC) as a consequence of chronic HBV infection [3–5]. The risk of developing chronic hepatitis B infection is inversely related to the age at time of infection and the chronic carrier rate can be above 90% when transmission occurs in the neonatal period [6,7]. Death rates from HCC or liver cirrhosis are reported to be as high as 25% among HBV carriers infected in infancy. HBV infection is the commonest cause of liver cancer worldwide and hepatitis B vaccine is the first vaccine to be used for cancer prevention [8,9]. Immunization of infants can prevent >80% of hepatitis B related deaths [3,10] and in 1991, the World Health Assembly
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467 Table 1 Tongan immunization schedule.
recommended that HBV vaccine should be included in national immunization programmes worldwide. Today, 116 countries routinely immunize children against HBV and an elimination target has been set at less than 1% infection prevalence in preschool populations in endemic areas [11].
Hep B 1 BCG Hep B 2, DPT1/Hib* 1, OPV** 1 DPT/Hib2, OPV2 Hep B 3, DPT3/Hib3,OPV3 MR*** 1 MR 2
2. Background
*
2.1. Hepatitis B in Tonga
**
Hepatitis B infection is hyper-endemic in the Pacific islands and earlier studies in Tonga have found that 10–20% of the adult population are chronic carriers and that 94.6% in the 10–19 years age group have serological markers for previous or current infection [8,12]. Routine screening of blood donors at the central Vaiola Hospital consistently find hepatitis B surface antigen (HBsAg) prevalence rates of more than 10%. A study among pregnant women in 2005 showed that 19.6% were HBsAg positive [13]. In countries with high prevalence rates most HBV infection occurs as a result of mother to child transmission during delivery and child-to-child transmission during early childhood. Immunization against HBV in Tonga started in 1989 using a three-dose schedule for infants with the first dose delivered as early as possible after birth. A study of vaccine efficacy in 1998 found 3.8% of pre-school children on Tongatapu to be HBsAg positive [14] despite reported hepatitis B immunization coverage of 95%. 2.2. Improved immunization delivery The finding of the unexpectedly high HBsAg prevalence in the 1998 study meant that Tonga was falling short of the elimination target for hyper-endemic populations of less than 1% and resulted in efforts by the immunization programme to improve vaccine delivery. Most children (>95%) in Tonga are born in hospitals, and arguably the most important improvement made to the immunization practices after the 1998 study, was the decision to transfer the responsibility for delivering the first dose of hepatitis B vaccine from EPI staff to the midwives and nurses who attend deliveries. The delivery attendant is required to give the HBV vaccine as soon as the baby is born and to record both the time and date of immunization in the delivery chart as well as on the immunization card, which is given to the mother before she leaves the hospital. Steps have been taken in order to deliver HBV vaccine in advance to the small number of women who deliver on isolated islands without vaccine storage facilities. HBV vaccine is stable up to a month without refrigeration and by sending the vaccine out to the islands in advance of an anticipated delivery instead of waiting for notification that a child has been born, has ensured that also geographically disadvantaged children are immunized immediately after birth (Tables 1–3). 2.3. Aim of the study The main aim of the study was to estimate the prevalence of HBV infection in children measured as the presence of hepatitis B surface antigen in blood (calculated with 95% confidence limits), and to estimate the immunization coverage of selected EPI antigens.
4463
***
Within 24 h after birth At birth At 6 weeks of age At 10 weeks of age At 14 weeks of age At 1 year of age At 18 months of age
Hib: haemophilus influenzae type b vaccine. OPV: oral poliovaccine. Measles-rubella vaccine.
Table 2 Timeliness of Hep B 1 immunization for 433 children with known place of birth. Place of birth
<24 h
1–7 days
>7 days
Total
Abroad At home, no SBA* SBA, (HC or home) Hospital
15 (53.6%) 2 (50.0%) 6 (85.7%) 359 (91.1%)
4 (14.3%) 0 (0.0%) 0 (0.0%) 14 (3.6%)
9 (32.1%) 2 (50.0%) 1 (14.3%) 21 (5.3%)
28 (100%) 4 (100%) 7 (100%) 394 (100%)
*
SBA skilled birth attendant.
3. Material and methods 3.1. Study population This was a cross-sectional hospital-based sero-survey utilizing convenience testing for hepatitis B surface antigen in admitted children after informed consent had been given by the child’s guardian. Subjects were recruited from among all children aged 6–59 months, who were admitted to Vaiola Hospital and from children who had blood samples taken in the hospital’s outpatient clinic, during the period December 2004–April 2007. Vaiola Hospital is the only hospital on Tonga’s main island of Tongatapu where 70% of the country’s population resides. The majority of the recruited children were admitted to the paediatric ward of the hospital. Children were included once in the study. Exclusion criteria included admission for acute or chronic hepatitis and previous recruitment. Ethical approval for the study was obtained from the Medical Research Ethics Committee, Ministry of Health Tonga. 3.2. Material Sample size was calculated based on an expected HBsAg prevalence of 2%, approximately half the prevalence of 3.8% found in a 1998 study of 211 pre-school children on Tongatapu. The population of children between 6–59 months of age on Tongatapu is approximately 10,000, which required the recruitment of 335 subjects to provide a prevalence estimate with a two sided 95% CI of 1.5%. The precision was set at 1.5% because of the low expected prevalence [15]. Allowing for failed testing a total of 449 eligible children were recruited in the study, and of those, 375 (84%) were successfully tested for hepatitis B surface antigen using the Abbott Determine® rapid HBsAg test (specificity 99.95%, sensitivity 95.16%). Positive tests were re-tested with the same test for confirmation. Seventy-four children (16%) were interviewed and had their immunization data checked but did not have a test result for HBsAg. There were three reasons for failing to get a HBsAg result: (1) after
Table 3 Timeliness of second and third doses of hepatitis B vaccine. Interval between Hep B 1 and 2
Frequency
≥1 month (according to schedule) <1 month
440 6
Total
446
Percent (95% CI) 98.7 (96.9–99.5) 1.3 (0.5–3.1) 100.0
Interval between Hep B 2 and 3
Freqency
≥2 months (according to schedule) <2 months
355 83 438
Percent (95% CI) 81.1 (77.0–84.5) 18.9 (15.5–23.0) 100.0
4464
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
the child had been recruited there was no clinical indication for drawing blood and the child was discharged without having blood collected, (2) there was insufficient blood in the sample to conduct HBsAg serology after the clinically indicated tests had been performed, (3) the hospital was temporarily out of rapid hepatitis B tests and sera were not stored for later testing. The most common cause for no test was insufficient volume of blood in the first sample and the fact that the child did not have any subsequent blood samples taken before discharge. 3.3. Collected information Data was gathered using a questionnaire that captured information on the child’s socio-economic circumstances, immunization, medical and delivery history. Immunization status was obtained from the child’s immunization card and then checked against the records kept at the health centre where the child received immunizations, and against the central registry kept by the supervising public health nurse at the Ministry of Health. Paediatric nurses conducted the interviews in Tongan. When blood was collected for clinical investigation a special request was made for a rapid HBsAg test to be performed by technicians in the blood transfusion unit who routinely check donated blood for HBV infection using the same rapid test. 3.4. Data analysis All epidemiological data were made anonymous using survey ID numbers. The information from the questionnaires, the immunization data and the result of the HBsAg tests were consecutively entered into a Microsoft Access database. Data analysis was carried out using Epi InfoTM , Version 3.4.1 software [16]. Statistical analysis was performed using frequencies, means and proportions.
The proportion of households with a TV was higher than that with a refrigerator. The majority (79.4%) of the 422 mothers for whom information about education was available had attended secondary school reflecting the high literacy in Tonga and the favorable school attendance rate for girls. 4.3. DPT 3 and measles immunization Coverage for DPT3 was 97.6% for the 436 children who had immunization data available. Of the 270 children who were 1 year or older at recruitment 88.1% had been immunized with a measles containing vaccine (MCV). 4.4. Hepatitis B immunization data Immunization data on hepatitis B was unavailable for four children in the study despite considerable efforts to locate them. Coverage for hepatitis B 1 (Hep B 1) was 99.1% (CI 97.7–99.7). Of the 445 children for whom immunization data was available, 409 (91.9%; CI 88.9–94.2) received the first dose of Hep B vaccine within 24 h after birth and 424 (95.3%; CI 92.8–97.0) received the first dose within the first week of life. The remaining 21 children (4.7%; CI 3.0–7.2), received Hep B 1 one week or later after birth. All 445 children received at least one dose of Hep B. Coverage for BCG was 98% (CI 96.1–99.0). Only 14.1% (CI 11.0–17.8) of children, however, received their first dose of BCG within 24 h after birth. Complete hepatitis B immunization requires three doses of vaccine. Of the children in the study, 438 (97.6%; CI 95.5–98.7) received all three doses of hepatitis B vaccine and out of those, 371 (84.7%; CI 80.9–87.9) completed their hepatitis B immunization before 6 months of age. Eleven children (2.4%) out of a total of 449 either did not receive Hep B 3 (eight children) or did not have immunization records (three children).
4. Findings/results 4.5. HBsAg prevalence 4.1. Age and birth place distribution The 449 children in the study had a mean age of 20.6 months at the time of recruitment. The median age was 14.7 months (6.0–59.7 months) and there were 244 (54.3%) boys and 205 (45.7%) girls. All but eight (1.8%) children resided on Tongatapu. All the 51 villages of Tongatapu were represented in the sample. The resident address was missing for three (0.7%) of the children. Of the 433 children with known birthplace 405 (93.5%) were born in Tonga. Of these 394 (91%) were born in hospital, seven (1.6%) were born outside of hospital with the help of a skilled birth attendant (SBA) and four (0.9%) were born at home without SBA. The remaining 28 (6.5%) of the children were delivered abroad either in the USA, New Zealand or Australia. 4.2. Birth weight, breast-feeding, mode of delivery, socio-economic indicators and mother’s education The mean birth weight was 3467 g (SD ± 671 g) and the median weight was 3515 g (range 1200–5720 g). Prematurely born children were included in the group of low birth weight children (<2500 g). More than 22% of the children had an above normal birth weight (>4000 g). Information on breastfeeding was recorded for 408 children of whom 40 (9.8%; CI 7.2–13.2%) were never breastfed. The mean duration of breastfeeding among the 368-breastfed children was 9.1 months (SD ± 5.9). Data on the mode of delivery was available for 411 children. Of those 13% was born by caesarean section. Most of the mothers reported that the household had access to electricity (91.3%), owned a TV (85.6%) and a mobile phone (85.4%).
Of 375 children who were tested for hepatitis B surface antigen (HBsAg), three tested positive, resulting in a prevalence of 0.8% (95% CI 0.2–2.5%). All three children had received a complete course of HBV vaccine at correct intervals and with the first dose given immediately after birth. All three infants were breastfed and none of the mothers had received blood transfusion. 5. Discussion 5.1. The results The results show that the immunization programme in Tonga has been successful in further reducing the number of children who become infected with HBV. The estimated prevalence of 0.8% indicate that Tonga may well have achieved the HBV elimination target of less than 1% infection in pre-school children although the sample is not large enough to make a firm conclusion. The HBsAg prevalence in pre-school children has declined from 3.8% in 1998 to 0.8% in the present study. The reduction has taken place in the absence of any major change in the proportion of hepatitis B infected mothers. Although it is possible that a reduction in horizontal transmission among children could have contributed to the reduced HBsAG prevalence, there are no indications that risk factors, such as hygiene practices, skin lesions and crowding have changed significantly between the two studies. Delivery practices and infection control in the hospital that could explain diminished vertical transmission from mother to child have not improved to a great extent since the previous study. The first birth cohorts to be immunized against hepatitis B are just about to enter reproductive
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
life. The risk of mother-to-child transmission of hepatitis B remains very high, which is confirmed by a 2005 study that found an HBsAg prevalence of 19.6% in pregnant women. The immunization programme has been using the same hepatitis B vaccine during this period. Overall, hepatitis B immunization coverage has remained in the same range and higher coverage is unlikely to explain why the proportion of HBsAg positive children has fallen to less than a quarter of what it was in 1998. Pregnant women are not screened for hepatitis B infection and hepatitis B immune globulin (HBIG) therapy is not available in Tonga to treat children of HBV infected mothers. The most prominent change in immunization practices has been the increased attention given to the timing of the first dose. Strong emphasis has been placed on the importance of administering the first dose as early as possible within 24 h after birth. We believe that the most plausible explanation for the decline is the strengthening of vaccine delivery that was instituted after the findings of the 1998 study. The importance of timeliness of the first dose of the hepatitis B vaccine was demonstrated in a study in British Columbia where a delay in the initial dose of vaccine was found to be associated with increased risk of infection [17]. Vaccination at birth has been shown to protect infants of HBeAg positive mothers in Thailand [18] where 0.7% of children were found to be HBV carriers in a follow-up study after a three-dose regime, similar to the Tongan schedule, was introduced [19]. Comparable conclusions have been made in Vietnam [20]. In Alaska, another hyper-endemic area, a 10 years assessment after the introduction of the neonatal immunization programme found that the HBsAg prevalence has been eliminated among children below the age of 10 [21]. Immunization programmes tend to be very efficient in immunizing children through a dedicated delivery system. Such a “vertical” approach is important for the successful immunization of children in countries with weak and poorly functioning health systems, but the lack of integration can become an obstacle to improving coverage in better functioning health systems. As long as the administration of the first dose of hepatitis B vaccine remained the exclusive responsibility of Tonga’s public health nurses, children continued to be missed or received a delayed first dose of hepatitis vaccine despite of being born in a health facility. Newborn children could be discharged from the delivery ward on weekends or at night when the EPI staff were not on duty to provide immunizations, thus depriving the child of the important birth dose of hepatitis B vaccine. Since the responsibility for delivering the birth dose of HBV vaccine has been delegated to the delivery staff and the vaccine is stored in the delivery ward, a much larger proportion of children now receive the first dose of HBV vaccine on the day they are born compared to BCG vaccine. Fig. 1 shows that few children are administered BCG vaccine within 24 h after birth according to the recommendations by the national immunization programme. Before delivery staff started giving HBV vaccine, most children would receive BCG and hepatitis B on the same day. To improve the timeliness of BCG vaccination, delivery staff could be authorized to administer BCG vaccine together with Hep B 1. It should be noted, however, that BCG coverage is close to 100% in
Fig. 1.
4465
Tonga and, in contrast to Hep B 1, the efficacy of BCG vaccination does not improve if it is given within 24 h after birth. The study showed that only half of the children born abroad received their birth dose of hepatitis B vaccine within 24 h. Although it is possible that Tongan women who deliver in Australia, New Zealand or USA are tested for hepatitis B during pregnancy and that the reason why the children were not immunized at birth was that their mothers were not infected, it is of concern that some of them might return unimmunized to Tonga. If infants of Tongan mothers and other Pacific island mothers are not recognized as being at risk of hepatitis B infection when born abroad, they are likely to return to Tonga unprotected. The results confirm the high immunization coverage reported by the national EPI programme in Tonga. One concern raised by the study is the comparatively lower coverage for the first dose of measles-rubella (MR1) vaccine. If the total number of unprotected individuals is allowed to increase over time, the susceptible population may become large enough to sustain an epidemic if either measles or rubella is introduced to Tonga from abroad. 5.2. The methodology Convenience testing of admitted children was chosen for this study because the study population was believed to be representative of all children on Tongatapu with regards to the risk of hepatitis B infection, and because we wanted to evaluate the feasibility of using the methodology for future surveillance of HBV transmission. Testing for HBsAg at a time when blood is collected for other reasons is a comparatively cheap method of monitoring HBV immunization efficacy. Convenience sampling [22,23] utilises existing infrastructures and is less staff intensive and has a lower opportunity cost compared to household surveys and other community based methodologies. Convenience testing allows for repeated surveillance surveys at regular intervals or for cost effective continuous surveillance over time. The method has been used for monitoring measles anti-body levels in order to predict the risk of epidemics in a population and as quality assurance for measles immunizations. 5.3. Limitations of the design There are limitations with hospital-based surveys of which the risk of selection bias is the most obvious. For bias to occur the risk of being HBsAg positive should be positively (or negatively) associated with the risk of being hospitalized. Hepatitis B transmission around birth is asymptomatic and HBV infection in early childhood rarely causes symptomatic disease. The possibility of children with symptomatic hepatitis B disease being over-represented in the group of hospitalised children can be addressed by excluding patients admitted for symptoms suggestive of liver disease. We chose to exclude children with liver cirrhosis, acute hepatitis and HCC from the study. However, no eligible patients with these conditions were admitted to the ward during the study. Selection bias would also occur if HBsAg positive children were more likely to be admitted to hospital for conditions other than cirrhosis and hepatitis, that is if they belonged to a sub-population with generally higher morbidity. This would happen if HBsAg positive children were more likely to fall ill with pneumonia, diarrhoea, meningitis, or other common paediatric conditions requiring admission to hospital. There is an association in general between chronic conditions, such as malnutrition, immune deficiencies and cystic fibrosis, and the risk of hospitalisation but there is no evidence that children who are infected with hepatitis B early in life are at increased risk of being admitted to hospital. The possibility that hepatitis B infection would be a protective factor
4466
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
against diseases requiring hospitalizing is remote and it can be assumed that a hospital based study is more likely to over-estimate the prevalence of HBsAg than the other way around. The argument that the probability of hepatitis B infection could be higher among children, who for socio-economic reasons are less likely to receive hospital care, is valid for many settings. Nevertheless, we found no reason to believe that an association between low socio-economic status and access to care would influence our findings. Tonga is a homogenous society and access to health care is very similar across the island of Tongatapu. Because health care is truly free, the financial obstacles to seeking hospital care are limited to the cost of transport. A good network of roads connects all villages with the hospital and the time to get there by bus or car does not exceed 35 min from anywhere on the island. The fact that the study recruited children from all villages on Tongatapu without attempting to have geographical representation supports that there is limited selection bias due to geographical inaccessibility. Vaiola Hospital, where the study was conducted, serves as a primary level hospital for the entire population of Tongatapu, the main island where 70% of Tonga’s population lives. Although Vaiola is also the referral hospital for the three district hospitals located on outer islands in other parts of the country, the number of referrals is in reality very low because of the limitations in transporting patients by air. Vaiola Hospital houses the only clinical laboratory on the island as well as the only paediatric facility. Except for seeking care abroad (New Zealand, the nearest developed country, is 3 h flight time from Tonga) there are no alternatives to Vaiola Hospital for acute child health care on Tongatapu. Education, and in particular the mother’s education, influences the adoption of immunization and the likelihood that a child is taken to hospital for care. Mothers with no or low education are less likely to access immunization for their children and less likely to seek early health care thereby increasing the risk of serious illness. Both factors could cause bias by increasing the risk of a child requiring hospital care. Nonetheless, primary education is compulsory in Tonga and the literacy rate is 98.8%. Ratios for secondary and tertiary enrolment are in favour of girls. The mothers of the three HBsAg positive children had all completed secondary education. In summary, we had no reason to expect that geographical, economical or educational barriers would result in selection bias towards a lower risk of hepatitis B infection among the children seen at the hospital and we believe that the prevalence of HBsAg found in this study provides an accurate estimate of the true HBsAg prevalence in children on Tongatapu. An important limitation in the design is the low expected prevalence of HBsAg which increases the required sample size. We found that it was difficult to maintain the motivation among the nurses in the ward to recruit patients. The design meant that the patient should first consent and be interviewed before HBsAg test could be performed. A better study design might have been to initiate anonymous testing for HBsAg in the laboratory of blood samples from children.
6. Conclusion and recommendations The first cohorts of hepatitis B immunized girls in Tonga will start reproducing within a few years. It is expected that the proportion of HBV infected pregnant women in Tonga will drop from the present level of 19% to below 5% within a decade and then continue to decrease to below 1%. This will dramatically reduce the number of children who are exposed to the risk of perinatal transmission of HBV. Although mother-to-child transmission is expected to fall quickly due to the fact that fewer pregnant women will be infected, the risk of horizontal transmission to newborns is likely to remain high for several generations in Tonga. Most Tongans live in extended
families and the number of adults who care for a particular child can be high [24]. The majority of the children in the study were born on the main island of Tongatapu, which prevented an assessment of the strategy to deliver vaccine in advance to children born on isolated islands. It remains to be done in further studies. The cause as to why some infants get infected in spite of correct immunizations requires further study as well. This prevalence study used a simple and reproducible study design that can be implemented in similar settings to monitor end-points for the immunization programme. The testing can be done with or without interviewing caretakers. The most cost effective model would be anonymous convenience testing of blood samples collected from admitted children. This method could be recommended for continued surveillance of the impact of HBV immunization in children in Tonga. The hospital laboratory could be asked to randomly test children of certain age groups. The Ministry of Health should also consider monitoring maternal HBsAg prevalence in the young adult population in order to validate the benefits of Hep B immunization. Cohorts of young pregnant women could be tested at regular intervals during antenatal care (ANC). This would be convenient as Tonga has recently introduced provider initiated HIV testing of all pregnant women and as the ANC attendance is nearly 100%. Anonymous convenience testing could include monitoring of vaccine efficacy of other vaccine preventable diseases, such as measles.
Acknowledgement The authors wish to thank Sela Paasi, the National EPI Coordinator, the nurses who conducted the interviews at the paediatric ward at Vaiola Hospital, all the caregivers who participated in the study and WHO Western Pacific Region for supporting the HBsAg tests.
References Reference List [1] Ruff TA. Temporary advisor WHO. Importance of hepatitis B birth dose; 2007. Microsoft Publisher Presentation. [2] WHO position paper on hepatitis B vaccines. Weekly Epidemiological Record 2004;79:253–64; 2007;255–62. (http://www.who.int/wer/2004/en/ wer7928.pdf). [3] Goldstein ST, Zhou F, Hadler SC, Bell BP, Mast EE, Margolis HS. A mathematical model to estimate global hepatitis B disease burden and vaccination impact. Int J Epidemiol 2005 Dec;34(6):1329–39. [4] Amin J, Law MG, Bartlett M, Kaldor JM, Dore GJ. Causes of death after diagnosis of hepatitis B or hepatitis C infection: a large community-based linkage study. Lancet 2006;368(9539):938–45. [5] Iloeje UH, Yang HI, Jen CL, et al. Risk and predictors of mortality associated with chronic hepatitis B infection. Clin Gastroenterol Hepatol 2007;5(8):921–31. [6] Broderick AL, Jonas MM. Hepatitis B in children. Semin Liver Dis 2003;23(1):59–68. [7] Boxall EH, Sira J, Standish RA, et al. Natural history of hepatitis B in perinatally infected carriers. Arch Dis Child Fetal Neonatal Ed 2004;89(5):F456–60. [8] Wainwright RB, McMahon BJ, Bender TR, et al. Prevalence of hepatitis B virus infection in Tonga: identifying high risk groups for immunization with hepatitis B vaccine. Int J Epidemiol 1986;15(4):567–71. [9] Zuckerman AJ. Prevention of primary liver cancer by immunization. N Engl J Med 1997;336(26):1906–7. [10] Mansoor OD, Salama P. Should hepatitis B vaccine be used for infants? Expert Rev Vaccines 2007;6(1):29–33. [11] Clements CJ, Baoping Y, Crouch A, et al. Progress in the control of hepatitis B infection in the Western Pacific Region. Vaccine 2006;24(12):1975–82. [12] Ushijima H, Dairaku M, Honma H, et al. Incidence of antibodies against human immunodeficiency virus, human T-cell lymphotropic virus type 1, hepatitis B virus, hemorrhagic fever with renal syndrome virus and Chlamydia in Tonga and western Samoa. Jpn J Med Sci Biol 1990;43(4):101–9. [13] World Health Organization. Second generation surveillance surveys of HIV, other STIs and risk behaviours in 6 Pacific Island countries 2004–2005. 2007. [14] Wilson N, Ruff TA, Rana BJ, Leydon J, Locarnini S. The effectiveness of the infant hepatitis B immunisation program in Fiji, Kiribati Tonga and Vanuatu. Vaccine 2000;18(26):3059–66.
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467 [15] Naing LWTRBN. Practical issues in calculating the sample size for prevalence studies. Arch Orofac Sci 2006;1:1–14. [16] Dean AG, Dean JA, Burton AH, Dicker RC. Epi Info: a general-purpose microcomputer program for public health information systems. Am J Prev Med 1991 May;7(3):178–82. [17] Marion SA, Tomm PM, Pi DW, Mathias RG. Long-term follow-up of hepatitis B vaccine in infants of carrier mothers. Am J Epidemiol 1994 Oct 15;140(8):734–46. [18] Lolekha S, Warachit B, Hirunyachote A, Bowonkiratikachorn P, West DJ, Poerschke G. Protective efficacy of hepatitis B vaccine without HBIG in infants of HBeAg-positive carrier mothers in Thailand. Vaccine 2002;20(31–32):3739–43. [19] Poovorawan Y, Theamboonlers A, Vimolket T, et al. Impact of hepatitis B immunisation as part of the EPI. Vaccine 2000 Nov;19(78):943–9.
4467
[20] Milne A, West DJ, Chinh DV, Moyes CD, Poerschke G. Field evaluation of the efficacy and immunogenicity of recombinant hepatitis B vaccine without HBIG in newborn Vietnamese infants. J Med Virol 2002 Jul;67(3):327–33. [21] Harpaz R, McMahon BJ, Margolis HS, et al. Elimination of new chronic hepatitis B virus infections: results of the Alaska immunization program. J Infect Dis 2000 Feb;181(2):413–8. [22] Rubin L, Hefer E, Dubnov Y, Warman S, Rishpon S. An evaluation of the efficacy of the national immunization programme for hepatitis B. Public Health 2007 Jul;121(7):529–33. [23] WHO Regional Office for the Western Pacific. Guidelines for the implementation of STI prevalence surveys. 2007. [24] Masters N. Hepatitis B transmission within families. Br J Gen Pract 1991 Jul;41(348):301.
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Improved immunization practices reduce childhood hepatitis B infection in Tonga Niklas Danielsson ∗ , Toakase Fakakovikaetau, Edit Szegedi World Health Organization, WHO Office Cambodia, P.O. Box 1217, Phnom Penh, Cambodia
a r t i c l e
i n f o
Article history: Received 4 April 2009 Received in revised form 3 May 2009 Accepted 10 May 2009 Available online 7 June 2009 Keywords: Hepatitis B vaccine HBsAg Elimination target Convenience testing Children Birth dose
a b s t r a c t Background: Hepatitis B infection is hyper-endemic in Tonga and 19% of pregnant women test positive for hepatitis B surface antigen (HBsAg). Routine childhood immunization against hepatitis B was introduced in 1989 and the target for elimination was set at <1% HBsAg prevalence in children. A study conducted in 1998, a decade after the introduction of hepatitis B immunization, found the HBsAg prevalence to be 3.8% in pre-school children. The finding resulted in the strengthening of the delivery of hepatitis B vaccine with emphasis on providing the first dose within 24 h after birth. The aim of this study was to measure the impact of improved immunization practices on the prevalence of hepatitis B infection in pre-school children, and to assess the progress towards hepatitis B elimination in Tonga. Measured outcome: Prevalence of HBsAg antigen. Type of study: Cross-sectional study. Methods: Children aged 6–59 months who were admitted to Vaiola Hospital, Nuku’alofa, Tonga, and had blood collected for clinical investigation, were tested for HBsAg with a rapid serological test (Abbott Determine® ). A total of 449 children were recruited and interviewed and 375 (84%) were tested for HBsAg. Convenience testing was chosen, as it was likely to be a relatively unbiased method in this situation where all children on the island have good and equitable access to hospital services. Immunization status was checked against the children’s immunization cards and cross-checked against the records kept by the public health nurses. Information about socio-economic status, parent education, blood transfusion, breast-feeding, mode of delivery, and place of birth was collected through interviews with mothers using a standardized questionnaire. Results: Three children tested positive for HBsAg resulting in a prevalence of 0.8% (CI 0.2–2.5%). Hepatitis B 1 (Hep B 1) immunization coverage was found to be high, 99.1% (CI97.7–99.7) and 91.9% (CI 88.9–94.2) of children received the first dose of hepatitis B vaccine within 24 h after birth. Coverage for the third dose of hepatitis B vaccine was 97.6% (CI 95.5–98.7) and of the children with complete immunization 84.7% (CI 80.9–87.9) had received all three doses by 6 months of age. Conclusions: The results show that the instituted changes in the delivery of hepatitis B vaccine have been effective in reducing the transmission of hepatitis B to children and indicate that Tonga appears to have achieved the elimination target of less than 1% HBsAg prevalence in children. This convenience survey used a simple, cost effective and reproducible study design. Convenience testing can be recommended for surveillance of the effectiveness of immunization programmes in settings where the population has good access to health services. © 2009 Elsevier Ltd. All rights reserved.
1. Introduction Approximately two billion people, or one third of the world’s population, have serologic evidence of past or present hepatitis B virus (HBV) infection [1]. WHO estimates that 367 million people worldwide are chronic carriers of HBV and that 500 000–700 000 people die each year as a result of HBV infection [2,3]. Mortality
∗ Corresponding author. Tel.: +855 012443 993; fax: +855 023216211. E-mail address: [email protected] (N. Danielsson). 0264-410X/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2009.05.051
from HBV is pre-dominantly from liver cirrhosis and hepato-cellular cancer (HCC) as a consequence of chronic HBV infection [3–5]. The risk of developing chronic hepatitis B infection is inversely related to the age at time of infection and the chronic carrier rate can be above 90% when transmission occurs in the neonatal period [6,7]. Death rates from HCC or liver cirrhosis are reported to be as high as 25% among HBV carriers infected in infancy. HBV infection is the commonest cause of liver cancer worldwide and hepatitis B vaccine is the first vaccine to be used for cancer prevention [8,9]. Immunization of infants can prevent >80% of hepatitis B related deaths [3,10] and in 1991, the World Health Assembly
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467 Table 1 Tongan immunization schedule.
recommended that HBV vaccine should be included in national immunization programmes worldwide. Today, 116 countries routinely immunize children against HBV and an elimination target has been set at less than 1% infection prevalence in preschool populations in endemic areas [11].
Hep B 1 BCG Hep B 2, DPT1/Hib* 1, OPV** 1 DPT/Hib2, OPV2 Hep B 3, DPT3/Hib3,OPV3 MR*** 1 MR 2
2. Background
*
2.1. Hepatitis B in Tonga
**
Hepatitis B infection is hyper-endemic in the Pacific islands and earlier studies in Tonga have found that 10–20% of the adult population are chronic carriers and that 94.6% in the 10–19 years age group have serological markers for previous or current infection [8,12]. Routine screening of blood donors at the central Vaiola Hospital consistently find hepatitis B surface antigen (HBsAg) prevalence rates of more than 10%. A study among pregnant women in 2005 showed that 19.6% were HBsAg positive [13]. In countries with high prevalence rates most HBV infection occurs as a result of mother to child transmission during delivery and child-to-child transmission during early childhood. Immunization against HBV in Tonga started in 1989 using a three-dose schedule for infants with the first dose delivered as early as possible after birth. A study of vaccine efficacy in 1998 found 3.8% of pre-school children on Tongatapu to be HBsAg positive [14] despite reported hepatitis B immunization coverage of 95%. 2.2. Improved immunization delivery The finding of the unexpectedly high HBsAg prevalence in the 1998 study meant that Tonga was falling short of the elimination target for hyper-endemic populations of less than 1% and resulted in efforts by the immunization programme to improve vaccine delivery. Most children (>95%) in Tonga are born in hospitals, and arguably the most important improvement made to the immunization practices after the 1998 study, was the decision to transfer the responsibility for delivering the first dose of hepatitis B vaccine from EPI staff to the midwives and nurses who attend deliveries. The delivery attendant is required to give the HBV vaccine as soon as the baby is born and to record both the time and date of immunization in the delivery chart as well as on the immunization card, which is given to the mother before she leaves the hospital. Steps have been taken in order to deliver HBV vaccine in advance to the small number of women who deliver on isolated islands without vaccine storage facilities. HBV vaccine is stable up to a month without refrigeration and by sending the vaccine out to the islands in advance of an anticipated delivery instead of waiting for notification that a child has been born, has ensured that also geographically disadvantaged children are immunized immediately after birth (Tables 1–3). 2.3. Aim of the study The main aim of the study was to estimate the prevalence of HBV infection in children measured as the presence of hepatitis B surface antigen in blood (calculated with 95% confidence limits), and to estimate the immunization coverage of selected EPI antigens.
4463
***
Within 24 h after birth At birth At 6 weeks of age At 10 weeks of age At 14 weeks of age At 1 year of age At 18 months of age
Hib: haemophilus influenzae type b vaccine. OPV: oral poliovaccine. Measles-rubella vaccine.
Table 2 Timeliness of Hep B 1 immunization for 433 children with known place of birth. Place of birth
<24 h
1–7 days
>7 days
Total
Abroad At home, no SBA* SBA, (HC or home) Hospital
15 (53.6%) 2 (50.0%) 6 (85.7%) 359 (91.1%)
4 (14.3%) 0 (0.0%) 0 (0.0%) 14 (3.6%)
9 (32.1%) 2 (50.0%) 1 (14.3%) 21 (5.3%)
28 (100%) 4 (100%) 7 (100%) 394 (100%)
*
SBA skilled birth attendant.
3. Material and methods 3.1. Study population This was a cross-sectional hospital-based sero-survey utilizing convenience testing for hepatitis B surface antigen in admitted children after informed consent had been given by the child’s guardian. Subjects were recruited from among all children aged 6–59 months, who were admitted to Vaiola Hospital and from children who had blood samples taken in the hospital’s outpatient clinic, during the period December 2004–April 2007. Vaiola Hospital is the only hospital on Tonga’s main island of Tongatapu where 70% of the country’s population resides. The majority of the recruited children were admitted to the paediatric ward of the hospital. Children were included once in the study. Exclusion criteria included admission for acute or chronic hepatitis and previous recruitment. Ethical approval for the study was obtained from the Medical Research Ethics Committee, Ministry of Health Tonga. 3.2. Material Sample size was calculated based on an expected HBsAg prevalence of 2%, approximately half the prevalence of 3.8% found in a 1998 study of 211 pre-school children on Tongatapu. The population of children between 6–59 months of age on Tongatapu is approximately 10,000, which required the recruitment of 335 subjects to provide a prevalence estimate with a two sided 95% CI of 1.5%. The precision was set at 1.5% because of the low expected prevalence [15]. Allowing for failed testing a total of 449 eligible children were recruited in the study, and of those, 375 (84%) were successfully tested for hepatitis B surface antigen using the Abbott Determine® rapid HBsAg test (specificity 99.95%, sensitivity 95.16%). Positive tests were re-tested with the same test for confirmation. Seventy-four children (16%) were interviewed and had their immunization data checked but did not have a test result for HBsAg. There were three reasons for failing to get a HBsAg result: (1) after
Table 3 Timeliness of second and third doses of hepatitis B vaccine. Interval between Hep B 1 and 2
Frequency
≥1 month (according to schedule) <1 month
440 6
Total
446
Percent (95% CI) 98.7 (96.9–99.5) 1.3 (0.5–3.1) 100.0
Interval between Hep B 2 and 3
Freqency
≥2 months (according to schedule) <2 months
355 83 438
Percent (95% CI) 81.1 (77.0–84.5) 18.9 (15.5–23.0) 100.0
4464
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
the child had been recruited there was no clinical indication for drawing blood and the child was discharged without having blood collected, (2) there was insufficient blood in the sample to conduct HBsAg serology after the clinically indicated tests had been performed, (3) the hospital was temporarily out of rapid hepatitis B tests and sera were not stored for later testing. The most common cause for no test was insufficient volume of blood in the first sample and the fact that the child did not have any subsequent blood samples taken before discharge. 3.3. Collected information Data was gathered using a questionnaire that captured information on the child’s socio-economic circumstances, immunization, medical and delivery history. Immunization status was obtained from the child’s immunization card and then checked against the records kept at the health centre where the child received immunizations, and against the central registry kept by the supervising public health nurse at the Ministry of Health. Paediatric nurses conducted the interviews in Tongan. When blood was collected for clinical investigation a special request was made for a rapid HBsAg test to be performed by technicians in the blood transfusion unit who routinely check donated blood for HBV infection using the same rapid test. 3.4. Data analysis All epidemiological data were made anonymous using survey ID numbers. The information from the questionnaires, the immunization data and the result of the HBsAg tests were consecutively entered into a Microsoft Access database. Data analysis was carried out using Epi InfoTM , Version 3.4.1 software [16]. Statistical analysis was performed using frequencies, means and proportions.
The proportion of households with a TV was higher than that with a refrigerator. The majority (79.4%) of the 422 mothers for whom information about education was available had attended secondary school reflecting the high literacy in Tonga and the favorable school attendance rate for girls. 4.3. DPT 3 and measles immunization Coverage for DPT3 was 97.6% for the 436 children who had immunization data available. Of the 270 children who were 1 year or older at recruitment 88.1% had been immunized with a measles containing vaccine (MCV). 4.4. Hepatitis B immunization data Immunization data on hepatitis B was unavailable for four children in the study despite considerable efforts to locate them. Coverage for hepatitis B 1 (Hep B 1) was 99.1% (CI 97.7–99.7). Of the 445 children for whom immunization data was available, 409 (91.9%; CI 88.9–94.2) received the first dose of Hep B vaccine within 24 h after birth and 424 (95.3%; CI 92.8–97.0) received the first dose within the first week of life. The remaining 21 children (4.7%; CI 3.0–7.2), received Hep B 1 one week or later after birth. All 445 children received at least one dose of Hep B. Coverage for BCG was 98% (CI 96.1–99.0). Only 14.1% (CI 11.0–17.8) of children, however, received their first dose of BCG within 24 h after birth. Complete hepatitis B immunization requires three doses of vaccine. Of the children in the study, 438 (97.6%; CI 95.5–98.7) received all three doses of hepatitis B vaccine and out of those, 371 (84.7%; CI 80.9–87.9) completed their hepatitis B immunization before 6 months of age. Eleven children (2.4%) out of a total of 449 either did not receive Hep B 3 (eight children) or did not have immunization records (three children).
4. Findings/results 4.5. HBsAg prevalence 4.1. Age and birth place distribution The 449 children in the study had a mean age of 20.6 months at the time of recruitment. The median age was 14.7 months (6.0–59.7 months) and there were 244 (54.3%) boys and 205 (45.7%) girls. All but eight (1.8%) children resided on Tongatapu. All the 51 villages of Tongatapu were represented in the sample. The resident address was missing for three (0.7%) of the children. Of the 433 children with known birthplace 405 (93.5%) were born in Tonga. Of these 394 (91%) were born in hospital, seven (1.6%) were born outside of hospital with the help of a skilled birth attendant (SBA) and four (0.9%) were born at home without SBA. The remaining 28 (6.5%) of the children were delivered abroad either in the USA, New Zealand or Australia. 4.2. Birth weight, breast-feeding, mode of delivery, socio-economic indicators and mother’s education The mean birth weight was 3467 g (SD ± 671 g) and the median weight was 3515 g (range 1200–5720 g). Prematurely born children were included in the group of low birth weight children (<2500 g). More than 22% of the children had an above normal birth weight (>4000 g). Information on breastfeeding was recorded for 408 children of whom 40 (9.8%; CI 7.2–13.2%) were never breastfed. The mean duration of breastfeeding among the 368-breastfed children was 9.1 months (SD ± 5.9). Data on the mode of delivery was available for 411 children. Of those 13% was born by caesarean section. Most of the mothers reported that the household had access to electricity (91.3%), owned a TV (85.6%) and a mobile phone (85.4%).
Of 375 children who were tested for hepatitis B surface antigen (HBsAg), three tested positive, resulting in a prevalence of 0.8% (95% CI 0.2–2.5%). All three children had received a complete course of HBV vaccine at correct intervals and with the first dose given immediately after birth. All three infants were breastfed and none of the mothers had received blood transfusion. 5. Discussion 5.1. The results The results show that the immunization programme in Tonga has been successful in further reducing the number of children who become infected with HBV. The estimated prevalence of 0.8% indicate that Tonga may well have achieved the HBV elimination target of less than 1% infection in pre-school children although the sample is not large enough to make a firm conclusion. The HBsAg prevalence in pre-school children has declined from 3.8% in 1998 to 0.8% in the present study. The reduction has taken place in the absence of any major change in the proportion of hepatitis B infected mothers. Although it is possible that a reduction in horizontal transmission among children could have contributed to the reduced HBsAG prevalence, there are no indications that risk factors, such as hygiene practices, skin lesions and crowding have changed significantly between the two studies. Delivery practices and infection control in the hospital that could explain diminished vertical transmission from mother to child have not improved to a great extent since the previous study. The first birth cohorts to be immunized against hepatitis B are just about to enter reproductive
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
life. The risk of mother-to-child transmission of hepatitis B remains very high, which is confirmed by a 2005 study that found an HBsAg prevalence of 19.6% in pregnant women. The immunization programme has been using the same hepatitis B vaccine during this period. Overall, hepatitis B immunization coverage has remained in the same range and higher coverage is unlikely to explain why the proportion of HBsAg positive children has fallen to less than a quarter of what it was in 1998. Pregnant women are not screened for hepatitis B infection and hepatitis B immune globulin (HBIG) therapy is not available in Tonga to treat children of HBV infected mothers. The most prominent change in immunization practices has been the increased attention given to the timing of the first dose. Strong emphasis has been placed on the importance of administering the first dose as early as possible within 24 h after birth. We believe that the most plausible explanation for the decline is the strengthening of vaccine delivery that was instituted after the findings of the 1998 study. The importance of timeliness of the first dose of the hepatitis B vaccine was demonstrated in a study in British Columbia where a delay in the initial dose of vaccine was found to be associated with increased risk of infection [17]. Vaccination at birth has been shown to protect infants of HBeAg positive mothers in Thailand [18] where 0.7% of children were found to be HBV carriers in a follow-up study after a three-dose regime, similar to the Tongan schedule, was introduced [19]. Comparable conclusions have been made in Vietnam [20]. In Alaska, another hyper-endemic area, a 10 years assessment after the introduction of the neonatal immunization programme found that the HBsAg prevalence has been eliminated among children below the age of 10 [21]. Immunization programmes tend to be very efficient in immunizing children through a dedicated delivery system. Such a “vertical” approach is important for the successful immunization of children in countries with weak and poorly functioning health systems, but the lack of integration can become an obstacle to improving coverage in better functioning health systems. As long as the administration of the first dose of hepatitis B vaccine remained the exclusive responsibility of Tonga’s public health nurses, children continued to be missed or received a delayed first dose of hepatitis vaccine despite of being born in a health facility. Newborn children could be discharged from the delivery ward on weekends or at night when the EPI staff were not on duty to provide immunizations, thus depriving the child of the important birth dose of hepatitis B vaccine. Since the responsibility for delivering the birth dose of HBV vaccine has been delegated to the delivery staff and the vaccine is stored in the delivery ward, a much larger proportion of children now receive the first dose of HBV vaccine on the day they are born compared to BCG vaccine. Fig. 1 shows that few children are administered BCG vaccine within 24 h after birth according to the recommendations by the national immunization programme. Before delivery staff started giving HBV vaccine, most children would receive BCG and hepatitis B on the same day. To improve the timeliness of BCG vaccination, delivery staff could be authorized to administer BCG vaccine together with Hep B 1. It should be noted, however, that BCG coverage is close to 100% in
Fig. 1.
4465
Tonga and, in contrast to Hep B 1, the efficacy of BCG vaccination does not improve if it is given within 24 h after birth. The study showed that only half of the children born abroad received their birth dose of hepatitis B vaccine within 24 h. Although it is possible that Tongan women who deliver in Australia, New Zealand or USA are tested for hepatitis B during pregnancy and that the reason why the children were not immunized at birth was that their mothers were not infected, it is of concern that some of them might return unimmunized to Tonga. If infants of Tongan mothers and other Pacific island mothers are not recognized as being at risk of hepatitis B infection when born abroad, they are likely to return to Tonga unprotected. The results confirm the high immunization coverage reported by the national EPI programme in Tonga. One concern raised by the study is the comparatively lower coverage for the first dose of measles-rubella (MR1) vaccine. If the total number of unprotected individuals is allowed to increase over time, the susceptible population may become large enough to sustain an epidemic if either measles or rubella is introduced to Tonga from abroad. 5.2. The methodology Convenience testing of admitted children was chosen for this study because the study population was believed to be representative of all children on Tongatapu with regards to the risk of hepatitis B infection, and because we wanted to evaluate the feasibility of using the methodology for future surveillance of HBV transmission. Testing for HBsAg at a time when blood is collected for other reasons is a comparatively cheap method of monitoring HBV immunization efficacy. Convenience sampling [22,23] utilises existing infrastructures and is less staff intensive and has a lower opportunity cost compared to household surveys and other community based methodologies. Convenience testing allows for repeated surveillance surveys at regular intervals or for cost effective continuous surveillance over time. The method has been used for monitoring measles anti-body levels in order to predict the risk of epidemics in a population and as quality assurance for measles immunizations. 5.3. Limitations of the design There are limitations with hospital-based surveys of which the risk of selection bias is the most obvious. For bias to occur the risk of being HBsAg positive should be positively (or negatively) associated with the risk of being hospitalized. Hepatitis B transmission around birth is asymptomatic and HBV infection in early childhood rarely causes symptomatic disease. The possibility of children with symptomatic hepatitis B disease being over-represented in the group of hospitalised children can be addressed by excluding patients admitted for symptoms suggestive of liver disease. We chose to exclude children with liver cirrhosis, acute hepatitis and HCC from the study. However, no eligible patients with these conditions were admitted to the ward during the study. Selection bias would also occur if HBsAg positive children were more likely to be admitted to hospital for conditions other than cirrhosis and hepatitis, that is if they belonged to a sub-population with generally higher morbidity. This would happen if HBsAg positive children were more likely to fall ill with pneumonia, diarrhoea, meningitis, or other common paediatric conditions requiring admission to hospital. There is an association in general between chronic conditions, such as malnutrition, immune deficiencies and cystic fibrosis, and the risk of hospitalisation but there is no evidence that children who are infected with hepatitis B early in life are at increased risk of being admitted to hospital. The possibility that hepatitis B infection would be a protective factor
4466
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467
against diseases requiring hospitalizing is remote and it can be assumed that a hospital based study is more likely to over-estimate the prevalence of HBsAg than the other way around. The argument that the probability of hepatitis B infection could be higher among children, who for socio-economic reasons are less likely to receive hospital care, is valid for many settings. Nevertheless, we found no reason to believe that an association between low socio-economic status and access to care would influence our findings. Tonga is a homogenous society and access to health care is very similar across the island of Tongatapu. Because health care is truly free, the financial obstacles to seeking hospital care are limited to the cost of transport. A good network of roads connects all villages with the hospital and the time to get there by bus or car does not exceed 35 min from anywhere on the island. The fact that the study recruited children from all villages on Tongatapu without attempting to have geographical representation supports that there is limited selection bias due to geographical inaccessibility. Vaiola Hospital, where the study was conducted, serves as a primary level hospital for the entire population of Tongatapu, the main island where 70% of Tonga’s population lives. Although Vaiola is also the referral hospital for the three district hospitals located on outer islands in other parts of the country, the number of referrals is in reality very low because of the limitations in transporting patients by air. Vaiola Hospital houses the only clinical laboratory on the island as well as the only paediatric facility. Except for seeking care abroad (New Zealand, the nearest developed country, is 3 h flight time from Tonga) there are no alternatives to Vaiola Hospital for acute child health care on Tongatapu. Education, and in particular the mother’s education, influences the adoption of immunization and the likelihood that a child is taken to hospital for care. Mothers with no or low education are less likely to access immunization for their children and less likely to seek early health care thereby increasing the risk of serious illness. Both factors could cause bias by increasing the risk of a child requiring hospital care. Nonetheless, primary education is compulsory in Tonga and the literacy rate is 98.8%. Ratios for secondary and tertiary enrolment are in favour of girls. The mothers of the three HBsAg positive children had all completed secondary education. In summary, we had no reason to expect that geographical, economical or educational barriers would result in selection bias towards a lower risk of hepatitis B infection among the children seen at the hospital and we believe that the prevalence of HBsAg found in this study provides an accurate estimate of the true HBsAg prevalence in children on Tongatapu. An important limitation in the design is the low expected prevalence of HBsAg which increases the required sample size. We found that it was difficult to maintain the motivation among the nurses in the ward to recruit patients. The design meant that the patient should first consent and be interviewed before HBsAg test could be performed. A better study design might have been to initiate anonymous testing for HBsAg in the laboratory of blood samples from children.
6. Conclusion and recommendations The first cohorts of hepatitis B immunized girls in Tonga will start reproducing within a few years. It is expected that the proportion of HBV infected pregnant women in Tonga will drop from the present level of 19% to below 5% within a decade and then continue to decrease to below 1%. This will dramatically reduce the number of children who are exposed to the risk of perinatal transmission of HBV. Although mother-to-child transmission is expected to fall quickly due to the fact that fewer pregnant women will be infected, the risk of horizontal transmission to newborns is likely to remain high for several generations in Tonga. Most Tongans live in extended
families and the number of adults who care for a particular child can be high [24]. The majority of the children in the study were born on the main island of Tongatapu, which prevented an assessment of the strategy to deliver vaccine in advance to children born on isolated islands. It remains to be done in further studies. The cause as to why some infants get infected in spite of correct immunizations requires further study as well. This prevalence study used a simple and reproducible study design that can be implemented in similar settings to monitor end-points for the immunization programme. The testing can be done with or without interviewing caretakers. The most cost effective model would be anonymous convenience testing of blood samples collected from admitted children. This method could be recommended for continued surveillance of the impact of HBV immunization in children in Tonga. The hospital laboratory could be asked to randomly test children of certain age groups. The Ministry of Health should also consider monitoring maternal HBsAg prevalence in the young adult population in order to validate the benefits of Hep B immunization. Cohorts of young pregnant women could be tested at regular intervals during antenatal care (ANC). This would be convenient as Tonga has recently introduced provider initiated HIV testing of all pregnant women and as the ANC attendance is nearly 100%. Anonymous convenience testing could include monitoring of vaccine efficacy of other vaccine preventable diseases, such as measles.
Acknowledgement The authors wish to thank Sela Paasi, the National EPI Coordinator, the nurses who conducted the interviews at the paediatric ward at Vaiola Hospital, all the caregivers who participated in the study and WHO Western Pacific Region for supporting the HBsAg tests.
References Reference List [1] Ruff TA. Temporary advisor WHO. Importance of hepatitis B birth dose; 2007. Microsoft Publisher Presentation. [2] WHO position paper on hepatitis B vaccines. Weekly Epidemiological Record 2004;79:253–64; 2007;255–62. (http://www.who.int/wer/2004/en/ wer7928.pdf). [3] Goldstein ST, Zhou F, Hadler SC, Bell BP, Mast EE, Margolis HS. A mathematical model to estimate global hepatitis B disease burden and vaccination impact. Int J Epidemiol 2005 Dec;34(6):1329–39. [4] Amin J, Law MG, Bartlett M, Kaldor JM, Dore GJ. Causes of death after diagnosis of hepatitis B or hepatitis C infection: a large community-based linkage study. Lancet 2006;368(9539):938–45. [5] Iloeje UH, Yang HI, Jen CL, et al. Risk and predictors of mortality associated with chronic hepatitis B infection. Clin Gastroenterol Hepatol 2007;5(8):921–31. [6] Broderick AL, Jonas MM. Hepatitis B in children. Semin Liver Dis 2003;23(1):59–68. [7] Boxall EH, Sira J, Standish RA, et al. Natural history of hepatitis B in perinatally infected carriers. Arch Dis Child Fetal Neonatal Ed 2004;89(5):F456–60. [8] Wainwright RB, McMahon BJ, Bender TR, et al. Prevalence of hepatitis B virus infection in Tonga: identifying high risk groups for immunization with hepatitis B vaccine. Int J Epidemiol 1986;15(4):567–71. [9] Zuckerman AJ. Prevention of primary liver cancer by immunization. N Engl J Med 1997;336(26):1906–7. [10] Mansoor OD, Salama P. Should hepatitis B vaccine be used for infants? Expert Rev Vaccines 2007;6(1):29–33. [11] Clements CJ, Baoping Y, Crouch A, et al. Progress in the control of hepatitis B infection in the Western Pacific Region. Vaccine 2006;24(12):1975–82. [12] Ushijima H, Dairaku M, Honma H, et al. Incidence of antibodies against human immunodeficiency virus, human T-cell lymphotropic virus type 1, hepatitis B virus, hemorrhagic fever with renal syndrome virus and Chlamydia in Tonga and western Samoa. Jpn J Med Sci Biol 1990;43(4):101–9. [13] World Health Organization. Second generation surveillance surveys of HIV, other STIs and risk behaviours in 6 Pacific Island countries 2004–2005. 2007. [14] Wilson N, Ruff TA, Rana BJ, Leydon J, Locarnini S. The effectiveness of the infant hepatitis B immunisation program in Fiji, Kiribati Tonga and Vanuatu. Vaccine 2000;18(26):3059–66.
N. Danielsson et al. / Vaccine 27 (2009) 4462–4467 [15] Naing LWTRBN. Practical issues in calculating the sample size for prevalence studies. Arch Orofac Sci 2006;1:1–14. [16] Dean AG, Dean JA, Burton AH, Dicker RC. Epi Info: a general-purpose microcomputer program for public health information systems. Am J Prev Med 1991 May;7(3):178–82. [17] Marion SA, Tomm PM, Pi DW, Mathias RG. Long-term follow-up of hepatitis B vaccine in infants of carrier mothers. Am J Epidemiol 1994 Oct 15;140(8):734–46. [18] Lolekha S, Warachit B, Hirunyachote A, Bowonkiratikachorn P, West DJ, Poerschke G. Protective efficacy of hepatitis B vaccine without HBIG in infants of HBeAg-positive carrier mothers in Thailand. Vaccine 2002;20(31–32):3739–43. [19] Poovorawan Y, Theamboonlers A, Vimolket T, et al. Impact of hepatitis B immunisation as part of the EPI. Vaccine 2000 Nov;19(78):943–9.
4467
[20] Milne A, West DJ, Chinh DV, Moyes CD, Poerschke G. Field evaluation of the efficacy and immunogenicity of recombinant hepatitis B vaccine without HBIG in newborn Vietnamese infants. J Med Virol 2002 Jul;67(3):327–33. [21] Harpaz R, McMahon BJ, Margolis HS, et al. Elimination of new chronic hepatitis B virus infections: results of the Alaska immunization program. J Infect Dis 2000 Feb;181(2):413–8. [22] Rubin L, Hefer E, Dubnov Y, Warman S, Rishpon S. An evaluation of the efficacy of the national immunization programme for hepatitis B. Public Health 2007 Jul;121(7):529–33. [23] WHO Regional Office for the Western Pacific. Guidelines for the implementation of STI prevalence surveys. 2007. [24] Masters N. Hepatitis B transmission within families. Br J Gen Pract 1991 Jul;41(348):301.