Effectiveness of the Targeted Hepatitis B Vaccination Program in Greenland

Abstract

Objectives. To evaluate the effectiveness of the hepatitis B virus (HBV) vaccination program in Greenland, which targets children born to mothers who are positive for HBV surface antigen (HBsAg), we determined vaccination coverage, levels of postvaccination antibodies, and frequency of breakthrough infections in at-risk children.

Methods. We conducted a population-based retrospective cohort study with data from nationwide registries. We identified all children born to HBsAg-positive mothers from 1992 to 2007 and collected data on their HBV vaccination status. In 2008 to 2010, we tested the children for HBV core antibody, HBsAg, and anti-HBsAg antibody (HBsAb).

Results. Of 4050 pregnant women, 3.2% were HBsAg positive. Of 207 children born to these women, 20% received no vaccinations, and only 58% received at least 3 vaccinations. At follow-up, HBsAb levels in vaccinated children were much lower than expected, and 8 (6%) of 140 at-risk children had breakthrough infections, with 4 chronically infected (persistently HBsAg positive).

Conclusions. The prevention program targeting children at risk for HBV in Greenland is ineffective. HBV vaccination should be included in the universal childhood vaccination program, and postvaccination HBsAb levels should be monitored.

Hepatitis B virus (HBV) infection may cause chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma.¹ Perinatal mother-to-child transmission is a major cause of chronic HBV infection in endemic areas. However, up to 90% of transmissions can be prevented with immunoglobulin given within 48 hours postpartum in conjunction with 3 or 4 HBV vaccine doses, beginning at birth and completed within 12 months.^2,3 In 1992, the World Health Organization recommended that all countries include HBV vaccination in the universal childhood vaccination program by 1997⁴; by December 2007, 171 of the world's 193 countries had complied.⁵ In the remaining 22 countries, some rely on identifying high-risk groups (e.g., the low-endemic Scandinavian countries); others, mainly sub-Saharan countries where infection is endemic, have no HBV vaccination strategy.⁶

Like other Arctic populations, the Inuit in Greenland have a high prevalence of HBV infection. Overall, 40% to 45% of the population have been infected (i.e., they test positive for HBV core antibody [HBcAb]), and 5% to 10% are chronically infected (i.e., they also test positive for HBV surface antigen [HBsAg]). The prevalence of HBsAg has not changed in the past 30 years.^7–9 Reports have indicated that the incidence of cirrhosis and liver cancer is lower in Greenland than in other highly endemic countries and lower even than in Denmark, where HBV infection is not endemic.^7–11 Hence, policymakers have not considered HBV infection to be a major health problem at the population level, and HBV prevention has relied on vaccination of at-risk infants.

Since 1992, Greenlandic policy has been to screen all pregnant women for HBsAg and to vaccinate infants of HBsAg-positive mothers. The program recommends that children receive 200 international units intramuscular HBV-specific immunoglobulin (HBIG; Aunativ, Biovitrum AB, Stockholm, Sweden) and 4 doses of 10 micrograms intramuscular recombinant HBV vaccine (EngerixB, SmithKline, Rixensart, Belgium), with HBIG and the first vaccination given within 48 hours after birth and additional vaccinations given at ages 1, 2, and 12 months.

Recently, a study reported that 3 siblings of a known chronic carrier of HBV were found to be chronically infected.¹² The same study described horizontal transmission of HBV and hepatitis D among children in Greenland.¹² These observations raised concern that the targeted HBV vaccination program was not fully effective and was insufficient to reduce the burden of HBV-related disease in Greenland. We carried out a retrospective population-based cohort study with data from national registries, with 3 objectives: (1) to determine HBV vaccination coverage in children of HBsAg-positive mothers; (2) to estimate the effectiveness of HBV vaccination, as measured by HBsAg antibody (HBsAb) levels; and (3) to determine the frequency of breakthrough infections in at-risk children.

METHODS

Greenland is an island with 16 towns and approximately 60 settlements, all located along the coast (Figure 1). Of the 56 000 inhabitants, 85% are Inuit; the majority of non-Inuit are Whites from Denmark. Health care is provided free of charge through public hospitals located in each town and by nursing stations in the settlements. All births should take place in hospitals, and pregnant women are transferred from settlements to towns with hospitals 4 weeks before their due date.

FIGURE 1—

Map of Greenland showing 5 areas: North, West, Nuuk (the capital), South, and East.

Note. Greenland population = 56 000.

Since June 1972, all Greenlandic residents have been assigned a unique identification number. The Civil Registration System records information on date and place of birth, gender, current and past addresses, parents, and date of death or emigration. Inuit ethnicity is defined as having 2 parents born in Greenland, mixed ethnicity as having 1 parent born in Greenland, and non-Inuit as having both parents born outside Greenland; when the father is not identified, ethnicity is determined by the mother's place of birth only. Previous studies have demonstrated a good correlation between parental birthplace and genetic ethnicity in Greenland.^13,14

All HBV testing in Greenland is performed at Queen Ingrid's Hospital in Nuuk, the capital city. At the time of our study, Greenland's HBV Database contained HBV results from 1992 through September 2009 except for files for 2000 through 2004, which were lost during construction of the database. We used all available test results in our study; i.e., because of the lost files for 2000 through 2004, we evaluated HBsAg prevalence among pregnant women only for 1992 through 1999, but we used information from the Civil Registration System to identify all children born from 1992 through 2007 to all known HBsAg-positive mothers.

Pregnant Women and At-Risk Children Cohorts

We identified all babies born in Greenland from 1992 through June 2007 in the Civil Registration System. Through parental linkage, we identified the mothers and retrieved the results of antenatal HBV screening up to 40 weeks before a live birth from the HBV Database. To ensure that all mothers had had equal access to antenatal HBV testing, we included in the cohort only mothers who had been living in Greenland from 40 weeks before birth until delivery.

We defined at-risk children as all children born to a mother with an HBsAg-positive test result during any pregnancy from 1992 through June 2007. Babies born to a previously HBsAg-positive woman who tested HBsAg negative during the most recent pregnancy were defined as not at risk. For women who were HBsAg positive in an earlier pregnancy but without an HBV result in the current pregnancy, we reviewed birth records and classified the child as at risk if there was an indication that the mother was known to be HBsAg positive (e.g., the records noted that the mother was chronically HBV infected or that the newborn had received HBV vaccination).

We reviewed records to determine HBV vaccination coverage. Administration of the first HBV vaccine and HBIG is recorded in the mother's perinatal chart. The second, third, and fourth vaccinations are recorded in the child's medical chart, which should follow the child upon relocation, and on the child's immunization certificate, which is kept by the mother. We excluded children who had died or moved from Greenland before they were aged 18 months because vaccination history was incomplete. We accepted the number of vaccinations given if the last record of any childhood vaccination in the immunization chart was noted after 2 years of age, indicating that the immunization chart had been in use during the relevant time. We defined proper HBV protection as receipt of HBIG and at least 3 HBV vaccinations, with both HBIG and the first vaccination given no more than 48 hours after birth and the remaining 3 vaccinations within the first 18 months of life, as documented by any immunization record.³ We collected data on vaccination coverage from July 2008 through February 2010.

Follow-Up

To determine current HBV status of children born to HBsAg-positive mothers, we conducted a nationwide follow-up study between July 2008 and April 2010. With parental consent, children had blood samples drawn for HBV testing. We omitted the municipalities Upernavik and Qeqertarssuaq and 3 settlements in Aasiaat Council because access was difficult and few children were in need of follow-up. We excluded children who had died or moved from Greenland at the time of follow-up.

The registry-based studies fulfilled the Helsinki Declaration II requirements. Before enrollment in the follow-up study, we obtained informed parental consent and, for children aged 8 years or older, assent, but we immediately excused unwilling children of all ages. We referred incompletely vaccinated children to the consultant at the local hospital for further evaluation, including free vaccination.

We shipped serum samples, frozen at −20°C, to Statens Serum Institut, Copenhagen, Denmark, where we tested them with radioimmunoassay kits for total-HBcAb, HBsAg, and HBsAb (Abbott, Wiesbaden, Germany). According to international guidelines, the cut-off value of HBsAb is 10 international units per liter or higher.¹⁵ A randomly chosen subset of 20 HBcAb-negative samples was retested in an independent laboratory with Vidas Anti-HBs Total Quick test (Biomérieux, Marcy l'Etoile, France), yielding results that differed from the initial test by no more than 5 international units per liter.

Statistical Analyses

To determine demographic factors associated with HBV status of the mother and vaccination coverage and HBsAb level in at-risk children, we estimated risk ratios by log-linear binomial regression analyses. We first performed univariate analyses. We used variables with P values lower than .05 in the multivariate models and estimated 95% confidence intervals.

We used χ² testing to evaluate the associations of demographic factors with vaccination coverage and breakthrough infections. We considered a 2-sided P value lower than .05 to be statistically significant. We performed all analyses with SAS version 9.1.3 (SAS Institute Inc, Cary, NC).

RESULTS

In total, 5725 women gave birth to 8515 children from 1992 through 1999 and were living in Greenland during their first pregnancy. Of these, we identified 4212 women with HBV test results in the HBV Database; we excluded 162 (3.9%) of these because they did not live in Greenland during the entire pregnancy. This left a sample of 4050 (70.7%) women who were screened in 5283 pregnancies (Figure 2a). Of these, 131 (3.23%) tested HBsAg positive (Table 1). Women in West and East Greenland had 4.2 and 3.4 times as high a risk of being HBsAg positive, respectively, than did women in South Greenland. Inuit mothers had a risk of being HBsAg positive that was 5.9 times as high as that of non-Inuit mothers. Overall, risk among mothers who were aged 30 years or older during their first pregnancy in the study period was 3 times as high as risk among mothers younger than 20 years.

FIGURE 2—

Participant flow chart for (a) the population-based study of HBV prevalence among pregnant women, (b) the retrospective vaccination coverage study on children born to HBsAg-positive mothers, and (c) the follow-up study of children born to HBsAg-positive mothers.

Note. HBV = hepatitis B virus; HBsAg = HBV surface antigen. According to the Danish Cause of Death Register, none of the 9 deceased children died of HBV-related disease.

TABLE 1—

Prevalence of Hepatitis B Virus Surface Antigen Among Pregnant Women: Greenland, 1992–1999

Characteristic	HBsAg Positive, No. (%)	HBsAg Negative, No. (%)	Crude RRa (95% CI)	Adjusted RRab (95% CI)	Adjusted Pb
Total (n = 4050)	131 (3.2)	3919 (96.8)
Place of birth					< .001
Westc (Ref)	82 (5.0)	1554 (95.0)	1.00	1.00
Eastd	12 (4.0)	300 (96.0)	0.77 (0.42, 1.39)	0.81 (0.45, 1.46)
Northe	11 (2.1)	512 (97.9)	0.42 (0.23, 0.78)	0.43 (0.23, 0.79)
Nuukf	17 (2.0)	833 (98.0)	0.40 (0.24, 0.67)	0.40 (0.23, 0.67)
Southg	9 (1.2)	720 (98.8)	0.25 (0.12, 0.49)	0.24 (0.12, 0.47)
Ethnicityh					.052
Inuit (Ref)	114 (3.3)	3306 (96.7)	1.00	1.00
Mixed	16 (3.5)	437 (96.5)	1.06 (0.63, 1.77)	1.01 (0.60, 1.70)
Non-Inuit	1 (0.6)	175 (99.4)	0.17 (0.02, 1.21)	0.17 (0.02, 1.24)
Mother's age, y					.02
≤ 19	10 (1.5)	670 (98.5)	0.30 (0.16, 0.59)	0.38 (0.17, 0.84)
20–24	24 (2.5)	932 (97.5)	0.52 (0.32, 0.82)	0.53 (0.32, 0.89)
25–29	38 (3.2)	1164 (96.8)	0.65 (0.44, 0.97)	0.62 (0.41, 0.94)
≥ 30 (Ref)	59 (4.9)	1153 (95.1)	1.00	1.00
RR/y			1.06 (1.03, 1.09)	1.06 (1.02, 1.09)	.002
Birth order					.1
First (Ref)	29 (1.8)	1623 (98.2)	1.00	1.00
Second or third	75 (4.1)	1759 (95.9)	2.33 (1.53, 3.56)	1.69 (1.03, 2.77)
Fourth or higher	27 (4.8)	537 (95.2)	2.73 (1.63, 4.57)	1.49 (0.80, 2.77)

Note. CI = confidence interval; HBsAg = hepatitis B virus surface antigen; RR = risk ratio. To avoid dependency between pregnancies, we only used the first test result during the first pregnancy with a result.

^aFor HBsAg-positive women.

^bAll categorized variables were adjusted for each other. Mother's age and number of children as continuous variables were also adjusted for each other as categorized variables.

^cManiitsoq, Sisimiut, Kangaatsiaq, Aasiaat, Qeqertarsuaq, Qasigiannguit, and Ilulissat.

^dTasiilaq.

^eQaanaaq, Upernavik, and Uummannaq.

^fCapital city.

^gPaamiut, Qaqortoq, Narssaq, and Nanortalik.

^hInuit was defined as both parents born in Greenland; where father was unknown, ethnicity was based on mother's place of birth only. For 1 person, ethnicity could not be determined.

The 1675 (29.3%) pregnant women without test results in the HBV Database differed in ethnicity, place of birth, and mother's age: a higher proportion of non-Inuit and women from nonendemic areas were not tested. However, the overall HBsAg prevalence differed only slightly from the prevalence among tested women (3.18% vs 3.23%) when we used weighted estimates for each ethnic group (Inuit, mixed, and non-Inuit) in the total population. We obtained a similar result when we weighted according to place of birth (3.14%) and mother's age (3.33%).

Vaccination Coverage

In total, 135 women with an HBsAg-positive result during any pregnancy from 1992 through June 2007 gave birth to 279 children during the study period (Figure 2b). Of 248 children in our sample, we successfully retrieved information on vaccination coverage for 207 (83%). In total, 20.3% of at-risk children received no HBV vaccination at all, 58% received at least 3 vaccine doses, and only 34% received the 4 doses recommended in the Greenlandic guidelines (Flemming Kleist Stenz, MD, chief medical officer, Greenland, oral communication, August 2009 and August 2010). If we assumed that all 41 children whose records lacked information about vaccination had received 3 or more vaccinations, the estimated 3-dose vaccination coverage would be 63%; the estimate would be 47% if we counted these children in the group that received fewer than 3 doses. Of the 165 children who received at least the first vaccination, 79.4% received HBIG, 92.4% of these within 48 hours after delivery. In total, 87% (143/165) of the initially vaccinated infants received the first vaccination within 48 hours after delivery, 78% (110/140) received the second vaccination within 2 months after birth, and 55.8% (67/120) received the third vaccination within 4 months.

Vaccination coverage increased from 46% in 1992 to 1995 to 69% in 1996 to 2000, but coverage decreased in 2001 to 2007, when only 58% of children received 3 or more vaccinations (P < .001). Vaccine coverage was not significantly associated with children's gender, ethnicity, or birth order; mothers' age at children's birth; being born in highly endemic areas (West and East Greenland) compared with nonendemic areas (South, North, and Nuuk); or being born in a town rather than a settlement.

Immunogenicity and Breakthrough Infections

We invited 213 of the 248 children in the study to participate in the follow-up study (Figure 2c). Of these, 140 (66%) agreed and gave blood samples. These 140 participants did not differ from the other 73 invited children in ethnicity, residence in a settlement rather than a town, gender, or birth period (1992–1995, 1996–2000, or 2001–2007), but children from the highly endemic areas participated in the follow-up testing at a higher rate (121/165, or 73%) than did children from the nonendemic areas (19/48, or 40%; P < .001).

Eight (6%) children were HBcAb positive, indicating breakthrough infection. Four of these were HBsAg positive, and all remained positive in later follow-up, indicating chronic infection. Seven of the 8 HBcAb-positive children had received at least 3 HBV vaccinations, with the first vaccination within 48 hours. However, only 4 (2 immune and 2 chronically infected) had also received HBIG at birth. All breakthrough infections occurred in children born in West Greenland, where HBV is highly endemic. Among children with breakthrough infections, 63% were born to women from settlements, compared with 16% of HBV-negative children being born to women from settlements (63% vs 16%, P < .001).

We had information on vaccination coverage for 126 (95%) of the remaining 132 HBcAb-negative children (Table 2). Of 99 children with at least 1 vaccination, 67 (68%) had HBsAb levels lower than 10 international units per liter. Number of vaccine doses was positively associated with higher HBsAb level. However, of 69 children who received 3 or more doses, only 41% had HBsAb levels of 10 international units per liter or higher, and 51% had levels lower than 5 international units per liter, a level our laboratory technicians considered consistent with no exposure to HBsAg. A significantly lower proportion of children from West Greenland had HBsAb levels lower than 10 IU/liter, compared with other regions. We detected no association with time since HBV vaccinations (i.e., age at testing) and HBsAb levels lower than 10 international units per liter (Table 2). In the restricted analyses for the 69 children who received at least 3 vaccinations, a higher proportion of children aged older than 10 years had levels lower than 10 international units per liter compared with younger children (P = .02), which is consistent with a waning effect with time from vaccination. Living in a settlement, ethnicity, gender, and birth order were not significantly associated with HBsAb level.

TABLE 2—

Anti–Hepatitis B Virus Surface Antigen Levels at Follow-Up Among At-Risk Children Born to Antigen-Positive Mothers in 1992–2007: Greenland, 2008–2010

Characteristic	HBsAb Level, IU/L, Median (IQR)	HBsAb < 10 IU/L, No. (%)	HBsAb ≥ 10 IU/L, No. (%)	HBsAb > 10 IU/L, Crude RR (95% CI)	HBsAb > 10 IU/L,a Adjusted RR (95% CI)	Adjusted Pa
Vaccinations, no.						.004
0	0.00 (0.00, 0.48)	25 (92.6)	2 (7.4)	0.16 (0.04, 0.61)	0.19 (0.05, 0.73)
1	0.04 (0.00, 1.54)	15 (88.2)	2 (11.8)	0.25 (0.06, 0.94)	0.29 (0.08, 1.10)
2	0.25 (0.07, 1.41)	11 (84.6)	2 (15.4)	0.32 (0.09, 1.20)	0.44 (0.12, 1.53)
3b	0.72 (0.02, 11.4)	18 (72.0)	7 (28.0)	0.59 (0.29, 1.18)	0.55 (0.28, 1.09)
4c (Ref)	8.07 (1.25, 35.1)	23 (52.3)	21 (47.7)	1.00	1.00
Age at follow-up, y						.2
< 5	0.00 (0.00, 1.41)	5 (100.0)	0 (0)	…	…
5–9	4.00 (0.17, 28.6)	17 (56.7)	13 (43.3)	1.61 (0.91, 2.85)	1.53 (0.97, 2.41)
10–14 (Ref)	1.29 (0.12, 11.4)	49 (73.1)	18 (26.9)	1.00	1.00
≥ 15	0.05 (0.00, 0.78)	21 (87.5)	3 (12.5)	0.47 (0.15, 1.44)	0.91 (0.28, 2.94)
Birthplace						.002
West Greenlandd (Ref)	1.44 (0.00, 21.8)	57 (64.8)	31 (35.2)	1.00	1.00
Rest of Greenlande	0.38 (0.07, 1.29)	35 (92.1)	3 (7.9)	0.22 (0.07, 0.69)	0.26 (0.09, 0.80)

Note. CI = confidence interval; HBsAb = anti-HBsAg antibody; IQR = interquartile range; RR = risk ratio. Ellipses indicate not estimated. Sample size was n = 126.

^aNumber of vaccinations (0–4), age at follow-up (< 5, 5–9, 10–14, or ≥ 15 years), and place of birth (West or rest of Greenland) adjusted for each other as categorized variables.

^bFor the 25 children who received 3 vaccinations, the age-specific median (IQR) for HBsAb for each age category at follow-up was 5–9 years, 3.6 IU/L (0.54, 36.2); 10–14 years, 1.9 IU/L (0.18, 11.4); ≥ 15 years, 0 IU/L (0, 14.1).

^cFor the 44 children who received 4 vaccinations, the age-specific median (IQR) for HBsAb for each age category at follow-up was 5–9 years, 23.8 IU/L (7.4, 81.7); 10–14 years, 3.7 IU/L (1.1, 29.9).

^dManiitsoq, Sisimiut, Kangaatsiaq, Aasiaat, Qeqertarsuaq, Qasigiannguit, and Ilulissat.

^eQaanaaq, Upernavik, Uummannaq, Nuuk, Paamiut, Qaqortoq, Narssaq, Nanortalik, and Tasiilaq.

DISCUSSION

Our results show that despite a well-intentioned program, Greenland is having difficulty providing broad coverage of HBV immunization by targeting at-risk children born to women who tested HBsAg positive during pregnancy. We found that 20% of at-risk children received no HBV vaccinations at all, only 58% received at least 3 HBV vaccine doses, and only 34% received the 4 doses recommended in the Greenlandic guidelines. At follow-up in 2008 to 2010, HBsAb levels in children who received at least 3 vaccinations were surprisingly low. We found breakthrough infections in 6% of at-risk children, most of them in children who received at least 3 vaccinations, and half of these infections resulted in the children testing HBsAg positive repeatedly, indicating chronic infection.

The targeted approach for childhood HBV prevention does not appear to be effective. By contrast, coverage of other vaccines in the universal vaccination program in Greenland is comparable with coverage in Western countries.¹⁶ In 1991, the World Health Organization recommended that all countries with HBsAg prevalence higher than 8% implement HBV vaccination in their universal vaccination programs by 1995.⁴ The high prevalence in Greenland, 7% to 10%, clearly places the country in that category. Yet in 1989 and 1999, suggestions to include HBV vaccination were rejected because of (1) lack of data to challenge the effectiveness of the existing HBV-targeting program and the apparently low risk of cirrhosis and hepatocellular carcinoma in persons with chronic HBV infection in Greenland, and (2) competing financial priorities.

We found that 3.2% of pregnant women were HBsAg positive. This was slightly lower than rates for the adult population observed in other studies in Greenland.^7,9 However, we found large regional differences, with high rates of HBsAg prevalence in West and East Greenland and low rates in South Greenland and Nuuk. Our regional findings were compatible with previous observations of frequencies of up to 30% HBsAg positive in some West and East Greenlandic settlements and only 3% in Nuuk.^12,17 Furthermore, the HBV prevalence was 3 times as high in pregnant women aged 30 years or older as in women younger than 20 years during pregnancy, implying a high degree of HBV transmission among persons aged 20 to 30 years.

Vaccination coverage of targeted children increased during the 1990s but decreased thereafter. In 2000, Greenland introduced a new perinatal chart with a dedicated space to mark HBsAg result and report HBV prophylaxis at birth. However, this perinatal chart remains in the mother's hospital record and does not accompany the child's record, which means that the health care provider who should follow up the HBV vaccination series may not know the at-risk status of the child. Even in areas with the highest prevalence (West and East Greenland), we found that no special attention was directed toward follow-up of children of HBsAg-positive mothers. Furthermore, the country has a high turnover of doctors and nurses among both Greenlanders and the many Danish practitioners. Thus, communication and follow-through were difficult.

Following up in 2008 to 2010, we detected a remarkably high number of breakthrough infections in at-risk children despite a history of having 3 vaccinations (8/140). However, only half of the children with breakthrough infections received HBIG plus vaccination, a combination known to be superior to vaccination alone.³ Four of these breakthrough infections resulted in chronic infection.

Our data showed a rate of breakthrough infections higher than was found in previous studies, which reported that 1% to 10% of at-risk children became HBcAb positive but none became repeatedly HBsAg positive.^15,18–22 In accordance with these findings, we observed lower HBsAb levels among vaccinated children than among children of the same age who received equivalent vaccine doses in other areas of the world. Of children who received at least 3 doses between 4 and 16 years earlier, 58% had HBsAb levels lower than 10 international units per liter. Other studies have reported that 5 to 10 years after 3 or 4 HBV vaccinations during infancy, only 10% to 30% of children had a level lower than 10 international units per liter.^{15,19,23–26}

Antibody response after revaccination later in life is related to the peak level achieved after primary vaccination.¹⁵ Although antibody levels wane over time, it is thought that boosters are not needed for immunocompetent persons as long as a satisfactory antibody level has been achieved following the initial vaccination series.^15,27 We do not know the primary immune response of the vaccinated children. Even though a considerable immune memory may be retained after loss of detectable HBsAb,^15,28 the very low antibody levels we found could suggest an inadequate primary response.

Several factors could explain the low levels of HBsAb and the breakthrough infections we found in our follow-up. First, they may reflect a higher frequency of poor responders to HBV vaccine in Greenland than in other places. Studies on antibody response elsewhere have shown that 5% to 10% of healthy neonates are nonresponders.²⁹ The reasons are unclear but may relate to human leukocyte antigen genotypes.³⁰ However, no studies on genetic polymorphisms and HBV infection have been carried out in Greenland. Greenlanders are a relatively uniform and distinct ethnic group and therefore could have an unusual genetic distribution predisposing them to poor response.

An alternative—but at present unsupported—explanation could be presence of mutations in the gene encoding for HBsAg^29,31 that enable the virus to escape monoclonal HBsAb, with an increased risk of breakthrough infection despite vaccination. Research is needed on whether this applies to Greenland, but the demonstration of a novel HBV subgenotype (B6) among Inuit in Greenland, Arctic Canada, and Alaska suggests that particular genetic HBV strains may exist in the Arctic.³²

Vaccine quality must also be taken into consideration. All vaccines are imported from Denmark from the same manufacturer. Studies have reported that the cold chain can be a problem during vaccine distribution because the immunogenicity of the HBV vaccine diminishes with freezing.^33,34 In Greenland, the average temperature is below 0°C for a great part of the year, and vaccine imported to Greenland may be stored in unheated warehouses until final delivery. Since 2006, imported vaccines have come with temperature monitors, but the effect of this cold-chain monitoring has not yet been evaluated.

Our study had several strengths. First, to avoid overestimation of at-risk children we conducted a population-based nationwide register study that only examined children born to women known to be HBsAg positive during pregnancy. Second, health care and vaccinations in Greenland are free of charge, so the effectiveness of this targeting program could be ascribed to the health care system, without confounding attributable to women failing to seek medical advice for economic reasons.

Limitations

The HBV Database was incomplete because of technical errors during its construction, which destroyed almost all HBV test results for 2000 through 2004. Thus, not all women who tested HBsAg positive in Greenland during a pregnancy in the study period were identified. However, the errors in the files were most likely random and not related to test results or other confounders and would therefore not create selection bias in estimations of the associations of HBV prevalence and covariates (e.g., mother's age, place of birth). We might not have identified every at-risk infant born to an HBsAg-positive mother, but those identified were most likely at risk.

We might not have assessed all vaccinations recorded. However, we contacted both the place of birth and the present local hospital to retrieve information on vaccination coverage. It is possible that not all vaccinations were recorded, which would bias the result toward underreporting vaccination. However, the very high vaccination coverage in Greenland's universal vaccination program indicated a high degree of awareness about registering vaccinations. Furthermore, we examined all available medical records from the hospital where each child was born as well as the local hospital serving the child at the time of follow-up.

We omitted 3 districts and 3 settlements from the follow-up study of at-risk children because of difficult access. Although we did not detect any differential misclassification, we cannot rule it out.

Conclusions

Our results have value for decision-making regarding disease prevention in Greenland and similar areas. The current targeted prevention program is not sufficiently effective, apparently because of both administrative failures and ineffective vaccinations.

The high degree of breakthrough infections was likely attributable in part to failure of the vaccination program, both in identifying children who required HBIG and at least 3 vaccinations and in maintaining proper temperature during vaccine distribution to maintain its efficacy.

In countries with endemic HBV infection and high rates of migration and health care staff turnover, such as Greenland, systematic and general programs are likely to work better than targeted programs because the steps for identifying and targeting a subgroup of children are complicated, expensive, and, as our results show, not completely effective. The majority of studies of cost–benefit analyses on universal HBV vaccination support this approach.⁵ Although we observed breakthrough infections in our sample, HBV vaccination remains the most effective public health means of preventing HBV infection.³⁵ In addition, universal childhood vaccination would also prevent horizontal transmission among children and adults through lifelong as well as herd immunity.^15,24 The risk of cold-chain failure will probably diminish as HBV vaccination becomes a common practice and procedures are universally formalized. Therefore, HBV control strategies in Greenland should include HBV vaccination as part of the universal childhood vaccination program.

Until HBV prevalence is greatly reduced in Greenland, antenatal HBV screening of pregnant women will still be needed so that children born to HBsAg-positive mothers will receive HBIG and HBV vaccine within 48 hours after delivery.

Our data indicate that complete antibody-based protection cannot be assumed in Greenland, even in persons with a history of vaccination. The vaccination response should therefore be measured in children aged 15 months who were vaccinated as newborns as to elucidate the primary response. In addition, a study of booster vaccinations in teenagers would clarify whether they have sufficient cellular memory to develop a protective HBV antibody response in the event of encountering the virus. Finally, the prevalence of genetic resistance to immunization and escaped HBV mutations in the population needs further study and response.

In September 2010, Greenland introduced routine HBV vaccination in the childhood vaccination program, a decision influenced by a hepatitis D outbreak among children in a settlement in the country in 2005 to 2009 and by the low efficiency of the targeted vaccine program, as shown in our study.