Hepatitis A antibody prevalence among people with an intellectual disability in Ireland

This manuscript aims to determine the prevalence of antibody to and risk factors for hepatitis A virus (HAV) in individuals attending three intellectual disability services through a cross-sectional survey held in Dublin, Ireland. Participants were 636 individuals aged four to 78 years attending three intellectual disability services. The main outcome measure was the measurement of anti-HAV (IgG antibody) in oral fluid swabs using an antibody capture enzyme immunoassay (EIA) technique. Risk factor information was obtained by questionnaire from the individual’s medical record. 
Participants were 362 males and 274 females. The median age was 36 years. The median age of the individuals differed significantly from one institution to another (P<0.001). The prevalence of antibody to HAV was 43% overall but the individual levels for the three institutions were 65%, 30% and 68% respectively, the difference being statistically significant (P<0.05). Although a number of factors were statistically significantly associated with prevalence of antibody on univariate analysis, only age was associated with the prevalence of antibody on multivariate analysis. Among clients living at home, both age and use of respite care were associated with having antibodies to HAV.
In conclusion, the prevalence of antibody increased with age and 14% had evidence of infection in the first ten years of life. We recommend that consideration should be given to immunising new entrants to the service with the combined hepatitis A and B vaccine.

Introduction HA V causes an acute self-limiting disease.It is transmitted via the faecal-oral route, and is most common in the areas of the world with poor sanitation.The clinical severity of HA V infection increases with age.Childhood infection is usually quite mild.the majority of children under 5 years showing no symptoms, but people infected as adults can suffer severe and prolonged illness.Hospitalisation and mortality rates increase with age I, The seroprevalence of HA V has been declining in most parts of the world in recent decades, probably due to improvements in living standards 2. HAVis a notifiable disease in Ireland and the number of HA V cases notified in this country decreased by 92% between 2000 aod 2003, from 309 cases in 2000 (8.5 casesiloo,ooo) to 25 in both 2002 and 2003 (0.61100,000) '.
There was an increase again in 2004 to 47 (1.1 / 100,000) 4.However, reporting practices changed in 2004, when, for the first time, case definitions were introduced and laboratories were obliged to notify infectious diseases s.Provisional figures for 2005 indicate that 48 cases of HA V were notified (1.21100,000) (Personal communication).
A population based prevalence survey in 1991 in an Irish town found a population immunity level to HAV of 43%, with immunity increasing with age 6.A study of Irish international travellers over a six year period (1986-1991) showed that over half (54%) those surveyed were immune 7 while a study of general practice attendees in Ireland over a three year period (1993-1996) found an immunity level of 67% A 1982 study reported that the prevalence of HA V antibody in residents in institutions for intellectual disability was higher than in the general population 9.This was attributed to suboptimal hygiene in institutionalised settings 10.An Irish seroprevalence surver in a nonresidentialleaming disability setting found a 70% prevalence in the early 19905 I .However, screening in three Irish intellectual disability schools found an overall prevalence level of HAV antibody of 10010 in 1999 among three to eighteen year olds 12. Individuals with Down syndrome have been reported to have a higher prevalence than other forms of intellectual disability 9, I) Factors reported previously to be associated with anti-HA V positivity include age 10, 1)•17 and duration of institutionalisation 10,1), IS,I7, However, Gil

Questionnaires
We used a questionnaire to obtain basic demographic data and other infonnation relevant to the spread of HA V from each client's medical records (Appendix 2).The questionnaires did not record name or address but had a personal identifier and date of birth.Each institution also had to complete a brief questionnaire on total numbers of people in the institution and on whether any outbreaks of HA V had occurred there (Appendix 3).The questionnaires were piloted for comprehension.

Laboratory
Oral fluid was collected according to the manufacturer's instructions for analysis for hepatitis A antibody.These samples recorded the client 's name, personal identifier and date of birth and were sent to the National Virus Reference Laboratory (NVRL) for analysis for hepatitis A IgG antibody.Measurement of anti-HAY (lgG antibody) using an antibody capture enzyme immunoassay (EIA) technique from the oral fluid samples was undertaken using a modification of a method used previously in the same laboratory which included specimen validation lJ .Comparison of anti-HA V results for 50 matched serum and oral fluid specimens showed sensitivity and specificity greater than 95% for this modification (Personal communication).Where a result was equivocal, it was repeated.

Analysis
Analysis (chi square test (i) and logistic regression) was carried out using the statistical package R 24. •

Results
The response rate was 69 % (636/926) overall.There was no recor~ ~fany of,the participants being given immunoglobulin (in the previous six months) or hepatitIs A vaccme (ever) so all questionnaires were eligible for inclusion.No outbreaks of HA V were known ,10 ha~e occurred in any of the services.Not all questions were answered, and so the denommator IS below 636 for some results.I.The questions: length of time in this institution (years) and length of time in other institutions combined (years) were not reliably recorded and were therefore excluded from analysis.
HA V antibody level There were 635 oral fluid samples received.The overall prevalence of HA V antibody was 43% (271/635) with the individual levels for the three institutions being 65% (institution A), 30% (institution B) and 68% (institution C) respectively.The differences were statistically significant (p<0.05).The in iti al results were equivocal for 41 samples so these were repeated.One patient whose result was equivocal had died in the interim period.A comparison of antibody by age group in this study and in the population study 6 is shown in Figure I.A comparison test between the two studies found no significant difference in overall antibody prevalence but there was an obvious difference in the 0-9 year age group.

... ~
The prevalence of antibody in those living at home was 25% compared to 60% for those living in either an institution or a community dwelling (p<O.OOI).The distribution of HA V antibody status with respect to the variables in the questionnaire is shown in Table 2. Univariate logistic regression is outlined in Table 3. Multivariate logistic regression analysis for the entire study group and for those living at home is outlined in Tables 43 and 4b respectively.The only factor significant ly associated with prevalence of HA V antibody on multivariate analysis for the whole group was age.Among those living at home, in addition to age, the use of respite care was also associated with prevalence of antibody.

•
An explanation or some or the terms used to Tables 2-4b is outlined below Confidence inlerval: A confidence interval provides a range of plausible values for the unknown parameter.These are usually quoted as 95% intervals, which are constructed so that one can be 95% confident that the true value of the unknown parameter lies between the limits.
Odd.~ ral;o: The odds in favou r of being exposed in subjects with the target disorder divided by the odds in favour of being exposed in those without the target disorder.
P va/lie: This provides a standardised estimate of the likelihood of encountering values if there were no difference or effect.• Due to the small numbers of people included in this category (n=-4), the regression estimates were not calculated. ,.

Discussion
We examined the prevalence of antibody and risk factors for HAV antibody i~ those ,with an intellectual disability in three centres in lreland.The rate is much lower than In an Insh nonresidential learning disability setting in the early 19905 II, The overall level of 10% among those aged 3 to 18 years in special schools in 1999 is similar to the rate of 10.9% found in those aged 0-\9 years in this study 12 • OUT overall response rate was good, the prevalence of HA V may have been somewhat underestimated in institution C as the non-responders there were older and the overall response rate for that institution was poorer.The epidemiology of HA V has changed fundamentally with the advent of hepatitis A vaccine.Before that, it was primarily cyclical but this is now changing in the US 25, the UK and Scandinavia 26, We are not yet in a position to say whether this cyclical pattern has also disappeared in Ireland.We note that the rates have been at their lowest level over the past four to five years but we do not know what will happen to the cyclical pattern in the years to come.
As in the population study in the provincial town 6 and the prevalence study in the special schools 12 oral fluid (salivary) sampling was used to assess the prevalence of antibody in this study.Individuals, including patients can collect saliva in a non-invasive manner with modest training 27.This method was found to be acceptable to the clients in this study and avoids the need for blood letting which may be difficult in these circumstances.It is simple and rapid 23.
There is a paucity of studies available 00 this topic and many of those available are old.Comparisons may not be valid given the change in social circumstances.The overall prevalence of antibody of 43% is similar to that found in two US studies 14, 11 in 1970 and in 1994-95.Given the overall decline in the prevalence of HA V in recent decades, the timing of the various studies will have a very significant impact on the antibody level.The only European studies found in the intellectual disability setting were in France (1992-3: anti-HA V prevalence of 50%) and in Spain (1996-7:54%) 16, 18.These rates were marginally higher than we found but the Spanish study was carried out in an area of intennediate endem icity.
As found previously.the prevalence of antibody in this study sign ificantly increased with age 10, 1).\1 on both univariate and t analysis.While a number of factors were significantly assoc iated with anti-HA V on both univariate and t analysis, the only factor significantly associated with anti-HA Von multivariate logistic regression analysis for the whole group was age reflecting the fact that people do not lose their immunity and that older people are more likely to have been infected during their lifetime.Down syndrome was found to be negatively associated contrary to the findings of other studies 9.1).In countries with a low prevalence of anti-HAY, studies have found a close correlation with age and duration of institutionalisation and a high prevalence of HA V in institutionalised intellectually disabled patients 10, Il, u.
Among those living at home, both age and the use of respite care were significantly associated with HA V antibody.
Neither the World Health Organization nor the Centers for Disease Control Atlanta lists people in institutions for intellectual disabilities as a high•risk group needing v~ccinatio~ with h "" A . 25  acknowledging that the population prevalence study was done IS years ago, there was a much higher prevalence of antibody in the 0-9 year age group in our study.It would be useful to have ongoing seroprevalence studies for the Irish population.
The primary course of the combined vaccine (3 doses) for an adult currently costs €102.84 compared with €57.15 for hepatitis B vaccine (GlaxoSmithKline, personal communication, March 2(06).The equivalent paediatric costs are €57.15 and €41.88 respectively.Therefore, the differential is very small for the pediatric vaccine.As it is current policy that those attending intellectual disability services receive hepatitis B vaccine 19 , and given the higher level of immunity in this study in the 0-9 years age group in comparison to the previous population study 6 , consideration should be given to offering new entrants the combination vaccine.A high uptake of the combined vaccine among special schools' pupils was achieved with one school having an 86% uptake of first and second dose (n= 19) and the second having a 97% uptake (n=30).The third school was offered hepatitis A vaccine only in response to an outbreak 12.This vaccine is safe and easy to administer with hepatitis B 3 1.It must be remembered that a single lapse of appropriate hygiene during exposure to the virus is sufficient to cause infection 20.
These results will be available to Irish vaccination makers and thus can be used to fonnulate policy.In particular, we recommend that consideration should be given to immunising new entrants to the intellectual disability service with the combined hepatitis A and B vaccine.In the interim, strict hygiene and infection control polices should be maintained in the work and living areas of those with an intellectual disability.
Participation was voluntary.Study approval was received from the Research Ethics Committee of the Faculty of Public Health Medicine of the Royal College of Physicians of Ireland Conduct and management or the study A Steering Comm ittee (Appendix 4) guided the study.It comprised medical representatives from the intellectual disability services participating in the study, a nurse from Institution B, two Specialists in Public Health Medicine and an infection control nurse (until 2003) from the Department of Public Health, a Public Health Specialist in an academic Department, the laboratory manager in the NVRL.and a representative of parents of clients.One of the authors (S8) carried out the data collection in two institutions while a staff member did data collection in the third.A Specialist in Public Health Medicine in the Department of Public Health, reporting to the Steering Committee managed the study.The Eastern Regional Health Authority (now the Health Service Executive Eastern Region) provided the funding for the study.Individual results of the clients were placed in their medical record in the institutions and are available to the client's family/guardian and the client where possible.However, the level of disability of the client detennines infonning the client.

For
The response rate by institution was as follows: 73% (130/ 178) in institution A, 69% (415/600) in institution Band 61% (91 / 148) in institution C. Ninety five percent afthe non-responders did not return the fonn with the remaining 5% saying no.The median age of responders was 36 years and of non-responders, 38 years (p=O.06,Wilcoxonrank-sumtest).Further analysis showed a strong interaction between age and institution, ti.= 26.8,p = 2.8e•7).For this reason, analysis of non•response was done separately for each institution.The response rate by age and gender by institution is shown in Table

Table I :
Analysis of response r ate by institution, age and gender The 636 participants included 362 males and 274 fema les.The participants' median age differed significantly (P<O.OOI) from one institution to another with individuals attending service B substantially younger than those in institution A or C. The level of disability of the respondents was as follows: mild (86/634: 13%), moderate (258/634:41 %) and severe (290/634:46%).Twenty nine percent (1811628) had Down syndrome.The majority attended a day unit such as a school or a workshop (601 /635:95%).

Table 4b : Multivariate Logistic Regression for those living at home (n=308)
IS C lent group is vaccinated in a number of European countries a~cording.to the findings of the EUROHEP.NET project 30.A large US study conducted 10 the mid 1990s concluded that the current need for such an intervention is not clearly demonstrated 11.Changes in living arrangements, universal precautions and other changes may have substantially decreased the contemporary risk for HAV 11.While J. PrevlG.syacd.atlG.wllil btpadtll A YlCdaadoa Yes I No! Do.'t luIow ...... OIlob.n.wlthl.pUl6 mo.tll.(UNIG ) YtS/ Nol Don't k.ow II.Iflivlnl at bome, II rtlplte cart ayalled of Ytl l No I Don' t know (OJ lfyes.. Name of Ualt."C" _ _ _ _ _ _ _ _ _ __ __ _ _ _ _ 9. TypnorOayUnlt