Introduction

Hepatitis A virus (HAV) causes acute inflammation of the liver and is transmitted via the fecal-oral route. The virus is endemic in developing countries where sanitary conditions are suboptimal and ≥90% of infections occur in children, are asymptomatic, and lead to lifelong immunity [1]. Incidence of symptomatic infection increases with age and 80% of adults infected will complain of symptoms [2]. In industrialised, low-endemic regions such as the Netherlands, population seroprevalence of antibodies is low and much of the adult population is susceptible if exposed. Spread is typically limited to travelers to HAV endemic countries, men who have sex with men (MSM) and contacts of hepatitis A patients. The secondary attack rate can be up to 50%, and though the infection is often mild and self-limiting, HAV infection can in rare cases, lead to fulminant hepatic failure [3].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Algorithm of contacts identified, included, treatments assigned and outcomes.

1. Cases notified with clinical signs or symptoms of infection and elevated amino-transferase levels, with detectable hepatitis A-specific immunoglobulin M (IgM) antibodies in the serum (in the absence of hepatitis A vaccination in the last 12 months) or an epidemiological link to a confirmed case. 2. Baseline serological status was unknown for 128: 63 were not tested, 65 were tested ≥15 days post exposure. 3. Vaccinated twice (or once if within one year of exposure) with inactivated hepatitis A vaccine. 4. Asymptomatic and total anti-HAV positive within 14 days of exposure. If ≤10 years, also IgM negative. 5. All IgM positive ≤14 days post-exposure. 6. Total anti-HAV negative and without symptoms. 7 Immunoglobulin given as PEP according to guidelines. See Table 1. 8. Hepatitis A vaccine given as PEP according to guidelines. See Table 1. Total anti-HAV negative at baseline and anti-IgM positive with jaundice (+/- HAV RNA on PCR) at follow-up, or in the absence of jaundice, anti-IgM positive and HAV RNA detected by PCR in the same follow-up sample.

https://doi.org/10.1371/journal.pone.0078914.g001

In the Netherlands, symptomatic HAV infection is notifiable and the Public Health Service (PHS) routinely traces contacts of cases, offering post-exposure prophylaxis (PEP) according to national guidelines. Since 2004, hepatitis A vaccine has been used routinely in post-exposure prophylaxis [4]. Until 2004, hepatitis A immunoglobulin (IG) was the only form of PEP recommended, and evidence dating largely from the 1940s to 1960s [5-8] and clinical experience since [9], supports its effectiveness. Widespread use of a human blood-derived immunoprophylactic has become controversial however [10,11], and in some countries its use is no longer recommended [12]. In the interim, a safe, immunogenic vaccine which provides long term protection has been developed [13,14] and evidence of its efficacy and effectiveness in preventing secondary infection is accumulating [15-17]. In a randomised controlled trial conducted in 2007 [18], which directly compared the vaccine and immunoglobulin, the vaccine was shown to be similarly effective in preventing laboratory confirmed, symptomatic secondary HAV infection in healthy persons 2-40 years old. While national advisory bodies are increasingly recommending HA vaccine as an alternative to IG in PEP [19-21], IG remains the PEP of choice in those considered at increased risk of severe disease: people with certain underlying conditions or those who are immunosuppressed, for example. Internationally however, guidelines diverge concerning what PEP should be used in healthy people over 40 years [4,12,21-24], where evidence of vaccine efficacy is lacking. In the Netherlands, based on a review of the evidence and expert consensus [25], contacts aged <30 years, or, those aged 30-50 years and vaccinated within 8 days post-exposure, are generally considered to be at low risk of severe infection and vaccination is recommended [4]. Otherwise human immunoglobulin is prescribed (details in Table 1). In this study we evaluate the routine use of hepatitis A vaccine as an alternative to immunoglobulin in prevention of secondary HAV infection in contacts considered at low risk of severe infection.

Methodology

Results (Figure 1)

Notified Cases

From 1 July 2004 and 31 December 2012, 200 cases of acute HAV infection were notified, of whom 64% were male (n=129), and the median age was 29 (IQR:12-41). Signs of jaundice included conjunctival or dermal icterus (n=178/200, 89%), pale stools or dark urine (n=22/200, 11%). Overall, 48% (n=97) were imported infections, 17% (n=34) were homosexually acquired, 11% (n=23) were domestically acquired infections and 23% (n=47) were of unknown source. The majority (75%, n=170) were born in the Netherlands, a further 21 (10%) were of other western origin, 7 were from Turkey or North Africa (3%), and 23 (11%) were other non-western countries.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Timeline of exposure, vaccination, symptom onset and confirmation in secondary hepatitis A cases

◊ One dose hepatitis A vaccine administered. Total anti-HAV negative and asymptomatic. ● Jaundice (conjunctival icterus +/- dark urine, pale stools) ○ Non-specific symptoms: nausea, fatigue, loss of appetite, malaise ■ Date of confirmed infection: IgM positive & PCR positive. i Case was aged 57 and was not immunised according to protocol. ii This case was PCR negative, but IgM positive and symptoms and signs confirmed by general practitioner. iii Child aged 1 year who was asymptomatic throughout, but HAV RNA confirmed on PCR

https://doi.org/10.1371/journal.pone.0078914.g002

Discussion

We evaluated the routine use of hepatitis A vaccine as an alternative to immunoglobulin in the prevention of secondary hepatitis A infection in contacts considered at low risk of severe infection over an 8-year period (2004-2012). In the Netherlands, all healthy contacts aged 1-30 years, and those aged 30-50 who are within 8 days of exposure to the index case, are given hepatitis A vaccine in preference to immunoglobulin. Of contacts vaccinated according to protocol, 6.0% seroconverted, developing a symptomatic laboratory confirmed, secondary infection. In the randomised controlled trial by Victor et al. [18], 4.4% of vaccinated contacts developed a symptomatic secondary infection. The trial included only contacts aged 2-40 years, and when the same age-group in our evaluation is compared, the proportion was similar (3.6%). Direct evidence of the efficacy of the vaccine in PEP in those over 40 is not available, so in the Netherlands, this recommendation was made based on a combination of epidemiological data, hospitalisation and case-fatality rates in the Netherlands, and vaccine immunogenicity data (seropositivity of 77% in persons aged 40-62 years post-vaccination in one study [28]). In our evaluation, 30% of contacts over 40 years vaccinated per-protocol developed symptomatic infection, and half of the 8 secondary infections were over 40 years, all of whom had been vaccinated. In a subgroup analysis of all healthy contacts vaccinated within 8 days of exposure, the risk of secondary infection in those over 40 years was 12 times that of contacts aged ≤15 years (95%CI:1.3-106.7). Internationally, guidelines vary regarding the upper age limit at which vaccination is offered in preference to immunoglobulin [21-24], and in some countries, vaccination is the only form of PEP offered [12]. Though in absolute terms, the number of secondary infections in this study is low, our findings suggest that a more conservative upper age limit for contact vaccination may be appropriate.

This real-world evaluation illustrates the complexities of post-exposure prophylaxis in a low-endemic country [29] where there are high levels of pre-existing immunity in some minority groups together with susceptible older adults at risk of severe infection. In the Netherlands, the incidence of acute hepatitis A infection has declined dramatically in recent years [30]. Although outcomes of this evaluation and that of 2004 are not directly comparable, the same definitions and timeframes were used throughout and the general picture in recent years is one of smaller households (median of 1 versus 3 family/ household contacts per index case), less pre-existing natural immunity among contacts (41% versus 54%), and a lower proportion of co-infections (11% versus 28%) among susceptible contacts. This picture, together with the overall decline in incidence, probably reflects reduced household exposure and transmission through improved household hygiene and reduced virus importation from countries where endemicity of HAV is decreasing as socioeconomic, sanitary and water supply conditions improve (Turkey and Morocco in particular).

The proportion of IgM positive seroconversions was also reduced, 16%, versus 34% in the previous evaluation. In the latter (where only immunoglobulin was offered), only one-fifth of those who became anti-IgM positive were symptomatic, despite the 34% IgM-positive seroconversion rate. This was attributed to the attenuation of symptomatic infection by immunoglobulin. As vaccination can induce a transient IgM seropositivity [31]. we considered only those who were jaundiced and/or had detectable HAV RNA as true secondary infections. Our rate of secondary infection may therefore be an underestimate (though we think it unlikely that viral RNA would no longer be detectable given the short follow-up time). Irrespective, this would imply a relatively greater proportion of symptomatic infections among those who became IgM positive in our vaccinated population. Whether this reflects a real difference, a vaccine effect or a hygiene effect is uncertain. Ultimately, the rate of symptomatic, laboratory confirmed secondary HAV infection among susceptible contacts was similar in both evaluations: 5.3% (6/112 from 2004-2012) versus 6.4%(12/186 from 1996-2000).

There were a number of study limitations. Firstly, as in the previous evaluation, those who did not undergo baseline serological testing (and were thus excluded) were younger than those who participated. As children are more likely to be susceptible than adults, the proportion with pre-existing immunity may therefore be overestimated. Secondly, contacts ≥10 years old who were asymptomatic and total anti-HAV positive at baseline were assumed to be immune and were not screened for IgM antibody. This could have resulted in the misclassification of contacts who in fact were asymptomatically co-infected, though the majority of older children and young adults develop symptoms. Thirdly, symptomatic contacts may have been more inclined to return for follow-up testing, leading to an overestimate of the secondary infection rate, but in any case, the proportion of seroconversions was small. Finally, we did not find any association between gender, ethnic background or household size and secondary infection, but as the absolute number of secondary infections in this study was small, the study may have been underpowered. It is possible that factors other than age (level of education for example) contributed to secondary HAV infection in those over 40 years who were vaccinated within 8 days of exposure.

In conclusion, timely administration (≤14 days post-exposure) of HAV vaccine in routine post-exposure prophylaxis was feasible and the incidence of laboratory confirmed, symptomatic secondary infection in contacts of hepatitis A cases aged under 40 years was low. Although larger studies are required, it may be prudent to limit routine use of vaccination in those over 40 years of age, irrespective of the duration post-exposure. Retaining the use of immunoglobulin as PEP of choice in older people and those at risk of severe illness seems appropriate.

Acknowledgments

The authors would like to thank the nursing staff in the Department of Infectious Diseases, Public Health Service Amsterdam who interviewed the patients and collected the data, and Prof. Martien Borgdorff for his critical review of the manuscript.