Swine Influenza Virus Antibodies in Humans, Western Europe, 2009

Serologic studies for swine influenza viruses (SIVs) in humans with occupational exposure to swine have been reported from the Americas but not from Europe. We compared levels of neutralizing antibodies against 3 influenza viruses—pandemic (H1N1) 2009, an avian-like enzootic subtype H1N1 SIV, and a 2007–08 seasonal subtype H1N1—in 211 persons with swine contact and 224 matched controls in Luxembourg. Persons whose profession involved contact with swine had more neutralizing antibodies against SIV and pandemic (H1N1) 2009 virus than did the controls. Controls also had antibodies against these viruses although exposure to them was unlikely. Antibodies against SIV and pandemic (H1N1) 2009 virus correlated with each other but not with seasonal subtype H1N1 virus. Sequential exposure to variants of seasonal influenza (H1N1) viruses may have increased chances for serologic cross-reactivity with antigenically distinct viruses. Further studies are needed to determine the extent to which serologic responses correlate with infection.

Serologic studies for swine infl uenza viruses (SIVs) in humans with occupational exposure to swine have been reported from the Americas but not from Europe. We compared levels of neutralizing antibodies against 3 infl uenza viruses-pandemic (H1N1) 2009, an avian-like enzootic subtype H1N1 SIV, and a 2007-08 seasonal subtype H1N1-in 211 persons with swine contact and 224 matched controls in Luxembourg. Persons whose profession involved contact with swine had more neutralizing antibodies against SIV and pandemic (H1N1) 2009 virus than did the controls. Controls also had antibodies against these viruses although exposure to them was unlikely. Antibodies against SIV and pandemic (H1N1) 2009 virus correlated with each other but not with seasonal subtype H1N1 virus. Sequential exposure to variants of seasonal infl uenza (H1N1) viruses may have increased chances for serologic cross-reactivity with antigenically distinct viruses. Further studies are needed to determine the extent to which serologic responses correlate with infection.

P andemic (H1N1) 2009 infl uenza virus resulted from
genetic reassortment between at least 2 swine infl uenza viruses (SIVs) (1). Hemagglutinin (HA) of this novel subtype H1N1 virus is similar to that of classical swine infl uenza virus and the triple reassortant subtype H1N1 viruses that are endemic in swine populations in North America. At the time of its detection in humans, pandemic (H1N1) 2009 virus had never been detected in swine populations anywhere, but it is believed to have circulated undetected in regions with little or no surveillance for infl uenza viruses in swine. Because this virus has not been reported by the European Surveillance Network for Infl uenza in Pigs (www.esnip.ugent.be) since the network's inception in 2001, it was most likely absent in swine in western Europe. By the end of 2009, pandemic (H1N1) 2009 virus infection of swine had been reported in Norway (2); sporadic cases have been reported in a few other European countries (e.g., Germany, Italy, Denmark) (3). The swine were probably infected by contact with infected humans, whereas transmission from swine to humans has not yet been documented. Pandemic (H1N1) 2009 virus is the fi rst swine-origin virus that is readily transmitted between humans (4).
Human infections with SIVs are rare. During 1958During -2005, only 50 cases of zoonotic infections were reported; most were in persons who had contact with swine (5). Limited secondary transmission to close contacts has been reported but appears to be rare, and to our knowledge, sustained human-to-human transmission of enzootic SIVs has never been noted (6). Some serologic studies suggest that persons who work with swine are at increased risk for zoonotic infection with SIVs (7)(8)(9)(10)(11)(12).
The predominant subtype H1N1 SIVs in Europe were introduced from wild ducks to swine in 1979 and have an entirely avian-derived genome (13)(14)(15). These viruses are designated as avian-like viruses and are antigenically distinct from subtype H1N1 SIVs in North America and from pandemic (H1N1) 2009 virus. Few cases of human infection with these avian-like swine subtype H1N1 viruses have been reported; chains of transmission have not been found (5,9,15), and no serologic studies have provided indirect evidence of transmission of SIVs to humans in Europe (15).

Swine Infl uenza Virus Antibodies in Humans, Western Europe, 2009
Studies in the United States, United Kingdom, and Finland found antibodies against pandemic (H1N1) 2009 virus in elderly persons (16)(17)(18). These antibodies can be explained by antigenic evolution of seasonal human infl uenza (H1N1) viruses that are derived from the 1918 pandemic virus (such as the classical swine infl uenza [H1N1] virus) but have undergone greater antigenic drift than the swine virus (19). Antigenically, the infl uenza (H1N1) viruses that circulated among humans before the 1950s are probably more closely related to the classical swine virus and thus to the pandemic (H1N1) 2009 virus than to contemporary human subtype H1N1 viruses. We investigated whether persons whose professions involve contact with swine (swine workers [SWs]) have neutralizing antibodies against 3 infl uenza viruses: pandemic (H1N1) 2009 virus, a European avian-like subtype H1N1 SIV, and a 2007-08 seasonal infl uenza subtype H1N1 (seasonal infl uenza) virus.

Study Population
During July 20-28, 2009, blood was collected from 211 healthy persons with past or present professional contact with swine. All participants gave informed consent and completed a questionnaire about the nature of their swine contacts (occupation, duration, frequency), infl uenza vaccination, and infl uenza infection history. No participant reported having been infected with pandemic (H1N1) 2009 virus. A total of 224 control serum samples were obtained from the serum bank of the Laboratoires Reunis, Junglinster, Luxembourg. The samples, from the general population of Luxembourg, had been submitted in December 2008 for routine serologic testing. Because of ethical constraints, no further information was gathered from controls. The study was approved by the National Ethical Committee for Research in Humans.

Virus Neutralization Assay
According to recommended World Health Organization protocols (20), serum samples were tested by virus neutralization assay against an infl uenza A (H1N1) virus strain isolated from a patient in Luxembourg in July  Positive control serum was collected from 5 patients >5 weeks after recovery from a laboratory-confi rmed infection with pandemic (H1N1) 2009 virus and from a previously unexposed pig 4 weeks after it had been experimentally infected with A/swine/Belgium/1/98 (H1N1) (21). Before the assay was conducted, all samples were heated to 56°C for 30 min to inactivate complement and unspecifi c inhibitors. Titers were reported as the reciprocal of the highest dilution of serum that completely neutralized virus growth. Samples were fi rst screened in duplicate with a 1:10 dilution. All samples that showed virus neutralization in ≥1 well were further titrated in quadruplicate up to a dilution of at least 1:320. Control samples positive for both viruses were included in all assays.

Statistical Analyses
Geometric mean titers (GMTs) were calculated for each person from quadruplicate serum samples. All negative samples were given an arbitrary GMT of 5. GMTs were compared by using the nonparametric Wilcoxon rank-sum test. To examine bivariate risk factors associated with antibody prevalence, we dichotomized GMTs of all positive samples for different cutoff points (>10 to >80) and analyzed them by χ 2 test and, for low proportions, by z-test. The distribution of antibody levels was checked for associations with multiple risk factors by using proportional odds modeling (23,24). Statistical analyses were performed by using SigmaStat version 3.1 (San Jose, CA, USA) and SPSS version 18 (Chicago, IL, USA).

Antibodies against Pandemic (H1N1) 2009 Virus
GMTs of antibodies against pandemic (H1N1) 2009 virus ( Table 2) were signifi cantly higher for SWs than for controls (p = 0.004). Table 3 shows that 2× more SWs than controls had neutralizing antibodies against pandemic (H1N1) 2009 virus for the lowest cutoff value (p = 0.001). This ratio slightly increased with rising cutoff values and remained signifi cant to a cutoff >160 (Table 3). In all age groups, ≈2× more SWs than controls had antibodies against pandemic (H1N1) 2009 virus (cutoff >10), except for persons >60 years of age (Table 4). For SWs and controls >60 years of age, GMTs for pandemic (H1N1) 2009 virus were similar (p = 0.897; Table 2). GMTs were signifi cantly higher for younger than for older (>60 years) SWs (but not controls) ( Table 2). Among SWs, antibodies against pandemic (H1N1) 2009 virus tended to decrease with age for all cutoff values; among controls, the same was observed for cutoffs >10 to >40. Thus, younger SWs more often had higher levels of antibodies against pandemic (H1N1) 2009 virus than did controls and older SWs. The difference between SWs and controls disappeared in older age groups and was weaker when older and younger controls were compared.

Antibodies against SIV
Similar to fi ndings for pandemic (H1N1) 2009 virus, GMTs for SIV were higher among SWs than controls; however, the difference was not signifi cant ( Table 2; p = 0.168). More SWs than controls had positive SIV titers regardless of the cutoff (Table 3). These differences were signifi cant for cutoffs >20 to >160 and increased with higher cutoffs (Table 3). Comparable to fi ndings for pandemic (H1N1) 2009 virus, for age groups up to 60 years antibodies against SIV were found in 1.2-2× more SWs than controls (cutoff >10; Table 4); GMTs were signifi cantly higher among SWs than controls in this age group (Table 2; p = 0.028). Seroprevalences and GMTs were similar for persons >60 years of age from each group (Tables 2, 4).
Thus, antibody titers for SIV were found more often and were higher among SWs than controls. In contrast to fi ndings for pandemic (H1N1) 2009 virus, titers for SIV were found more often and were higher for older than younger controls; for SWs, titers were found more often among older persons but values were similar.

Antibodies against Pandemic (H1N1) 2009 Virus and SIV
Among SWs, for all cutoff values seroprevalence was higher for SIV than for pandemic (H1N1) 2009 virus. The same was found for controls but only for lower titers (≥10 and ≥20; Table 3). The differences between antibody positivity for each of the 2 viruses increased with age among SWs and controls (Table 4). Comparing seroprevalences for pandemic (H1N1) 2009 virus to those for SIV, differences were signifi cant only for SWs >60 years (p = 0.002). Also, signifi cantly more controls of the same age group (>60 years) had antibodies against SIV (62.2%) than against pandemic (H1N1) 2009 virus (6.7%, p<0.001; Table 4). The proportion of older (>60 years) SIV-seropositive controls (62.2%) differed signifi cantly from the proportion Thus, for both groups, more persons had antibodies against SIV than against pandemic (H1N1) 2009 virus, and differences in positivity decreased with increasing titers. Antibodies against SIV were more common among older persons, and antibodies against pandemic (H1N1) 2009 virus were more common among younger persons.

Seasonal Infl uenza Virus Compared with Pandemic (H1N1) 2009 Virus and SIV
GMTs for seasonal infl uenza virus were signifi cantly higher among SWs than controls (Table 2), and signifi cantly more SWs than controls had antibodies against seasonal infl uenza virus, at least for titers ≈10 to 40 (Table 3). Among all age groups, more SWs than controls had antibodies against seasonal infl uenza (Table 4). GMTs among controls >60 years of age were signifi cantly higher than those among younger controls ( Table 2). Signifi cantly more SWs and controls had antibodies against seasonal infl uenza virus than against pandemic (H1N1) 2009 virus and SIV (Table 3)

Discussion
At the time of blood collection from SWs (late July 2009), pandemic (H1N1) 2009 had spread to all continents, but intensity was still low in Europe, especially in Luxembourg and its neighboring countries. The only countries in which infection rates increased were the United Kingdom, Ireland, and Spain (where sporadic outbreaks occurred) (25). In 2009, Luxembourg had an intensive active surveillance system for infl uenza-like illnesses. The virus neutralization assay used measures neutralizing antibodies mainly against HA because antibodies were in the assay only during the virus entry phase (20). Nevertheless, we cannot exclude that residual antibodies against NA and M (93% and 98% aa identity between pandemic [H1N1] 2009 virus and SIV, respectively) may contribute to neutralization (27).
Because there is no correlate of protection for neutralizing antibodies or a defi nition of a positive titer measured by virus neutralization assay (28), we analyzed titers by using running cutoff values for positivity and compared GMTs. This analysis showed signifi cantly higher prevalence of neutralizing antibodies against pandemic (H1N1) 2009 virus in SWs than in controls, and seropositivity decreased with age. Younger (<60 years) SWs had higher titers, and 2× more SWs than age-matched controls had neutralizing antibodies against pandemic (H1N1) 2009.
No evidence indicates that pandemic (H1N1) 2009 virus was present in swine in Europe in or before July 2009. Reactivity with pandemic (H1N1) 2009 virus correlated best with antibodies against SIV. Although this correlation was highly signifi cant among SWs with relatively high titers for SIV, no such correlation was found among controls, in whom antibody levels against SIV were low. We speculate that the difference between the cohorts may refl ect cross-reactive antibodies to another infl uenza virus more similar to SIV (with or without a minor contribution of antibodies against seasonal infl uenza) in SWs, in contrast to low, mainly cross-reacting seasonal infl uenza virus antibodies in controls. Serologic cross-reaction between SIV and pandemic (H1N1) 2009 virus in pigs was recently reported (22). Our results also showed that reactivity with pandemic (H1N1) 2009 (or SIV) in either cohort cannot be explained by cross-reactivity with a recent seasonal infl uenza virus used in this study. Nevertheless, because more SWs than controls were exposed to seasonal infl uenza virus, we cannot exclude the possibility that antibodies to pandemic (H1N1) 2009 virus or to SIV in the SWs may be caused by a more complex history of exposure to seasonal infl uenza virus of subtype H1 or to subclinical infections with pandemic (H1N1) 2009 virus during the fi rst months of the pandemic.
Our fi nding of low levels of neutralizing antibodies against pandemic (H1N1) 2009 in controls (general population) is in agreement with fi ndings of previous studies (29). Our fi ndings that titers were less common but higher for older controls contrast with reports from the United Kingdom and Finland (16,17) but agree with fi ndings of 2 studies in China, where elderly persons (>60 years) had few or no neutralizing antibodies against this virus (30,31).
Our study also showed signifi cantly higher prevalence of neutralizing antibodies against SIV in SWs than in the controls at cutoffs >20 to >160, but differences in GMTs were not signifi cant. Similar serologic studies in humans in the United States showed markedly elevated antibody titers for North American SIVs of subtype H1N1 and H1N2 in SWs compared with controls (5,8,10,11,32,33). These studies used hemagglutination inhibition instead of virus neutralization assays and reported ORs for increased serologic responses instead of seroprevalence rates. The reported ORs, however, seem to be higher than those in our study (8,32,33) and could be partially explained by exclusion of persons with swine exposure in the US control groups.
Most persons undergo sequential infections with multiple antigenic variants of human infl uenza subtype H1N1 and H3N2 viruses throughout their lives. Such infections strongly increase the odds for serologic crossreactions with antigenically distinct H1 viruses, as documented in experimental studies with pigs (22), and may explain why older persons in the general population have higher antibody titers to SIV than their younger counterparts. Both older and younger controls are unlikely to have been infected with SIV, but older persons have been exposed to a wider variety of human seasonal infl uenza viruses. This exposure is also refl ected by a signifi cant difference in GMTs for recent seasonal infl uenza virus in older than younger controls. In Luxembourg, elderly persons may have had contact with swine because during 1920-1947 in Luxembourg, 50%-22% of all households kept >5 pigs, but before 1979, there was no apparently substantial swine infl uenza activity in this part of Europe (14). Apart from antibodies to SIV, a few controls also had antibodies to pandemic (H1N1) 2009 virus, but these did not correlate with each other, suggesting a different crossreactivity pattern than that for SWs. These fi ndings show that in the absence of paired serum samples, presence of neutralizing antibodies to a given infl uenza virus does not necessarily refl ect infection with that virus. Elevated antibody titers to SIV in part of the SWs may have resulted from exposure to the virus, but further studies are required to determine all possible causes.
In conclusion, titers of antibodies against pandemic (H1N1) 2009 virus and against an avian-like subtype H1N1 infl uenza virus were found more frequently and were higher for SWs than for controls. These titers cannot be explained by cross-reactivity with antibodies from recent seasonal infl uenza viruses. Neutralizing antibodies to both subtype H1N1 viruses showed some degree of correlation.
Further studies are needed to determine incidence of zoonotic SIV infections and the extent to which serologic responses correlate with infection. Neutralizing antibodies should confer at least partial protection against infection, reducing the risk that the avian-like subtype H1N1 SIV will cause major outbreaks of disease in humans.