Seroprevalence of Anti-H5 Antibody among Thai Health Care Workers after Exposure to Avian Influenza (H5N1) in a Tertiary Care Center

In recent years, infection with avian influenza (H5N1) virus in humans has been associated with high morbidity and mortality. The first human outbreak of H5N1 infection occurred in Hong Kong in 1997, followed by 2 cases of H5N1 infection in Hong Kong in February 2003 [1–3]. Subsequent outbreaks of H5N1 infection occurred in Vietnam and Thailand in January 2004 [4, 5]. Twenty-two patients in Vietnam and 12 patients in Thailand had confirmed cases of H5N1 infection; 23 (67%) of 34 infected patients died.

To date, there is a paucity of data regarding the effect of occupational exposure to H5N1 virus among health care workers (HCWs). One prior study suggested that, although rare, human-to-human transmission to HCWs occurred during the 1997 Hong Kong outbreak of H5N1 infection [6]. To further characterize occupational exposure to H5N1 infection, we conducted an epidemiologic investigation and seroprevalence study comparing HCWs in our institution who were exposed to an index case patient with H5N1 infection with those HCWs who were not exposed.

Case report. In March 2004, a 39-year-old woman was referred to Thammasat University Hospital Medical Intensive Care Unit (MICU; Pratumthani, Thailand) because of rapidly progressive pneumonia. At initial presentation to the referring hospital, she reported a 1-week history of fever, diarrhea, nausea, and vomiting, but no respiratory symptoms [7]. On day 5 of hospitalization, she developed cough and shortness of breath. Therapy with ceftazidime and amikacin was initiated, and the patient was transferred to our hospital. At our hospital, therapy with imipenem, azithromycin, and doxycycline was initiated, because adult respiratory distress syndrome (ARDS) progressively developed. Consultation with the infectious diseases department was requested 48 h after admission to our hospital, and additional history from the patient's family revealed that she had recently been exposed to several dead chickens in her neighborhood. Nasopharyngeal aspirates obtained from the patient were positive for the H5 strain of influenza A according to 2 RT-PCR primers and real-time RT-PCR. Viral sequencing subsequently yielded H5N1. The patient died of severe ARDS and multiorgan failure 1 day later. All family members were well and did not have any influenza-like illness.

During routine patient care, 30 HCWs at our hospital were exposed to the index case patient with unrecognized H5N1 infection; appropriate personal protective equipment (PPE) was not used for the initial 48 h of exposure. After case recognition (48 h after admission to our hospital), appropriate PPE (i.e., surgical mask, gown, and gloves) was used by all HCWs caring for this patient at MICU, regardless of patient contact. HCWs who were assigned to perform any aerosol-producing procedures, such as suction of respiratory secretions or administration of breathing treatments, wore N95 masks, cap and/or hood, and goggles, in addition to gown and gloves. Standard precautions were applied to all patients in MICU. All HCWs performed hand washing before and after caring for all MICU patients.

Setting. Thammasat University Hospital is a 350-bed tertiary care center. It has a level III, 8-bed MICU equipped with central air conditioning and 2 isolation rooms, with an estimated 400 admissions per year. There were no negative-pressure rooms and no special ventilation system. The MICU staff includes 1 MICU attending physician, 3 internal medicine residents, 20 nurses, and 3 ancillary and general health care assistants. An infectious diseases consultant, a pulmonologist, and a cardiologist were also involved in the care of this patient. The estimated patient-to-nurse ratio for the MICU was 2 : 1. There are 2 infection-control nurses (ICNs) and 1 hospital epidemiologist who are responsible for all aspects of infection control in the hospital, including the occupational health program.

Patients and methods. Paired acute-phase and convalescent-phase serum samples for anti-H5 antibody testing were obtained from HCWs who had been exposed to the index case patient and from HCWs who had not been exposed; the HCWs without exposure to the index case patient worked at the same hospital as those HCWs with exposure. Acute-phase serological testing for anti-H5 antibody was defined as serological testing involving samples obtained within 1 week after exposure to the index case patient, and convalescent-phase serological testing for anti-H5 antibody was defined as serological testing involving samples obtained 14 days after acute-phase serum samples were collected. HCWs with exposure were defined as those HCWs who were exposed to the H5N1 index case patient without use of appropriate PPE and who worked on the ward where the index case patient was hospitalized. HCWs without exposure worked at the same hospital but had not worked on the ward where the index case patient was hospitalized. Temporally-related influenza-like illness among HCWs was defined as a respiratory illness that began 1–14 days after exposure to the index case patient. All HCWs in this study were followed up prospectively for 2 weeks for temporally related influenza-like illness, and their temperature were measured twice weekly. This required all HCWs to report to the MICU head nurses before shifts (or to report to the ICN, for HCWs not routinely working in the MICU) for follow-up. Any HCWs who developed a temporally-related influenza-like illness or fever underwent rapid influenza A testing by enzyme immunomembrane filter assay (Directigen Flu A; Becton Dickinson). If the rapid influenza A test result was positive, antiviral medication was offered. All HCWs consented to participate in this study.

Demographic and clinical data included age, sex, occupation, employment history, history of smoking and chronic diseases, history of recent travel, history of respiratory illnesses, exposure to the index case patient, and poultry exposure. Poultry exposure was defined as any of the following: contact with ill poultry, shopping at a market that sold live poultry, having freshly butchered or live poultry in the home within the previous 2 weeks, a history of butchering poultry, or a history of living on a poultry farm.

Blood samples were obtained from HCWs with exposure to the index case patient a median of 3 days (range, 2–3 days) after the last possible exposure without use of appropriate PPE. Blood samples were obtained from HCWs who were not exposed to the index case patient a median of 3 days (range, 2–3 days) after H5N1 infection was suspected in the index case patient. Serum samples were stored at -20°C until tested for antibody against H5N1. Twenty-five (83%) of 30 HCWs with exposure to the index case patient had paired acute-phase and convalescent-phase blood samples available for testing. Blood samples were obtained from both exposed and nonexposed HCWs by infection-control staff after H5N1 infection was suspected in a case patient. Exposed HCWs without a blood sample collected >10 days after the last possible exposure to an index case patient were excluded from the analysis.

All serum samples were sent on dry ice to the Influenza Branch of the Centers for Disease Control and Prevention (Atlanta, GA) for testing for H5-specific antibody using influenza A/Thailand/16/2004 (H5N1) as antigen. The microneutralization assay was based on standard methodology and was conducted in biosafety level 3 containment [8]. Acute-phase and convalescent-phase serum samples, serially diluted from 1 : 20 to 1 : 2560, were tested in at least duplicate assays. On the basis of previously established criteria, positivity was defined as a neutralizing antibody titer of >1 : 80 and confirmatory Western blot [8]. Human serum samples from a Thai subject with confirmed H5N1 infection were used as a positive control. For adults aged <60 years, the sensitivity and specificity of the combined tests were 80% and 96%, respectively. Adults aged ⩾60 years were excluded from the analysis, because the H5N1 microneutralization and H5 Western blot assays are less specific for this age group [8].

Results. Thirty HCWs were exposed to the index case patient with H5N1 infection at our hospital. For this epidemiological investigation, participation included 25 (83%) of 30 exposed HCWs and 24 nonexposed HCWs. There was no difference in age, sex, presence of chronic medical conditions, occupation, sex, travel history, and history of poultry exposure between the exposed and nonexposed HCWs (table 1). During the 2-week follow-up period, no HCWs developed any temporally-related influenza-like illness or fever, and none had evidence of anti-H5 seropositivity or seroconversion (table 2).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

The characteristics of and type of exposure reported by 30 health care workers exposed to a patient with H5N1 infection at Thammasart University Hospital, Thailand.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 2

Characteristics of health care workers (HCWs) exposed to an index case patient with H5N1 infection, compared with HCWs not exposed to the index case patient.

Discussion. The major finding from this cohort study was that none of the 25 HCWs who were exposed to H5N1 subsequently had clinical or laboratory evidence of seropositivity or seroconversion. Notably, all of the exposed HCWs were without appropriate PPE, and the majority (90%) worked within 3 feet of the index case patient; none of the HCWs developed fever or a temporally-related influenza-like illness. These summary findings compliment the conclusions from 2 previously reported cohort studies [6, 9] that identified low rates of anti-H5 antibody in household contacts, social contacts, and HCWs after exposure to individuals with cases of H5N1 infection. Together, these 3 studies suggest that human-to-human transmission of H5N1 infection is low [6, 9]. Because influenza A is constantly changing in antigens [1], it is plausible that the strain of H5N1 involved in the 1997 outbreak in Hong Kong was similar to yet distinct from the first H5N1 strain during the 2004 outbreak in Thailand.

From the perspective of study limitations, we were unable to quantify the true prevalence or risks associated with seroconversion among HCWs because of small sample size. Nonetheless, our study is the first to evaluate the seroprevalence of anti-H5 antibodies among HCWs using paired acute-phase and convalescent-phase serum samples after this initial outbreak of H5N1 infection in Thailand. Furthermore, our data suggest that HCWs remain at low occupational risk for acquisition of H5N1 infection when without appropriate PPE, but continued precautions and monitoring are essential in case the virus evolves to become more transmissible among humans.

• Received August 23, 2004.
• Accepted September 20, 2004.

• © 2005 by the Infectious Diseases Society of America