Recent alerts demonstrate that the transmission of avian influenza to humans, which re-emerged at the beginning of 2004, is far from over in South East Asia [1]. As efforts to control the epidemic and prevent further human cases continue, the need to assess the effectiveness of current control measures grows. An executive summary of the final report of the outbreak of avian influenza A/H7N7 in the Netherlands has recently been published in English [2].

Between March and May 2003, an unprecedented outbreak of avian influenza occurred in humans in the Netherlands. During an extensive epizootic of influenza A virus H7N7 on commercial poultry farms, 86 cases in poultry workers and 3 cases in people with no poultry contact were initially confirmed by PCR. The predominant symptom was conjunctivitis [3]. One veterinarian developed fatal respiratory distress syndrome after close contact with infected poultry [4].

A questionnaire survey was carried out as a follow up to the outbreak. Approximately 400 poultry farmers and their families and almost 900 people who were involved in controlling the epidemic participated in this investigation. Blood samples were taken from 500 of these participants to determine possible infection with the avian flu virus. Additional studies were performed for 62 household contacts of 25 persons with avian flu virus infection.

Routine serological tests failed to detect any antibodies, even in the group of persons with confirmed avian influenza virus conjunctivitis. A modification of the haemagglutination assay was developed, based on observations that avian influenza viruses favour binding to red blood cells from horses rather than turkeys [5]. As at least 50% of the people exposed to infected poultry had H7 antibodies detectable with the modified assay, it was estimated that avian influenza A/H7N7 virus infection occurred in at least 1000, and perhaps as many as 2000 people. The seroprevalence of H7 antibodies in people without contact with infected poultry, but with close household contact to an infected poultry worker, was 59%. This suggests that the population at risk for avian influenza was not limited to those with direct contact to infected poultry, and that person to person transmission may have occurred on a large scale. Specificity of the unconventional assay was confirmed by the absence of reactivity in sera from 100 controls recently vaccinated with influenza vaccine (2002/2003) (specificity 100%). Assay specificity was further supported by the results of the cohort study: having measurable antibodies was associated with having conjunctivitis (RR 1.72; 95% CI 0.99-2.99), and a lower proportion of the exposed persons who took prophylactic antiviral medication developed antibodies (corrected OR 0.48; 95% CI 0.25-0.89).

Neither poultry farmers nor those engaged in controlling the epidemic complied satisfactorily with preventive measures. Only 6% of farmers reported consistent use of facial masks and 1% reported consistent use of goggles while working with infected poultry. In cullers, compliance was only slightly better: 25% consistently used facial masks and 13% used goggles. The results of the epidemiological study suggest that oseltamivir protected against conjunctivitis (corrected OR=0.14; 95% CI=0.08–0.27) as well as against infection without specific symptoms. No protective effect was demonstrable for safety goggles or mouth-nose masks [2].

After the outbreaks of group A subtype H5N1 (A/H5N1) avian influenza viruses in Hong Kong in 1997, in which 6 people died, the hypothesis was put forward that not only pigs but also humans themselves might serve as mixing vessels for the next pandemic influenza virus [6]. The outbreak of avian influenza A/H7N7 in the Netherlands and the recent unprecedented expansion of avian influenza A/H5N1 in Asia have reinforced this concern. A review of the outbreak and control efforts in the Netherlands highlights important lessons for preparedness: while separate systems are in place to signal and control animal diseases and human diseases, an outbreak of a zoonotic disease illustrates the importance of coordination between the two. In the Netherlands, the people infected came from a wide geographic region, and included foreign poultry workers. While the movement of animals was restricted, these people were out of the reach of the public health authorities while infectious and shedding the virus.

Although the disease in humans is more severe for A/H5N1, both avian influenza outbreaks illustrate that crossing the species barrier is less rare than previously recognised, that avian influenza virus adaptation occurs rapidly, and that if such jumps between species occur, human behaviour in the broad sense may accelerate dissemination [7].