We defined a suspected case of disease as community-acquired pneumonia (as primary clinical diagnosis), confirmed by chest radiograph, in a resident of the Blue Mountains 15 to 75 years of age who was admitted to a local hospital between March 1 and June 30, 2002. Patients with a history of congestive cardiac failure and chronic obstructive pulmonary disease were excluded. Active surveillance for suspected cases of psittacosis was initiated in the first week of June 2002. Surveillance was performed by review of patient medical records and daily contact with the emergency department and infection control staff at the Blue Mountains and other local hospitals. We attempted to contact all suspected case-patients by telephone to invite them to participate in the study and provide serum samples for laboratory testing.

We defined a probable case of disease as psittacosis in a suspected case-patient in which any of the following were demonstrated: seroconversion, a 4-fold rise in immunoglobulin (Ig) G titer by microimmunofluorescence (MIF), or a single or static high convalescent-phase MIF IgG titer to Chlamydophila psittaci.

We conducted a case-control study to identify independent risk factors for psittacosis in the outbreak. Only probable case-patients were included in the analysis. To identify controls, random digit dialing was used to select household telephone numbers from the randomly sorted Blue Mountains telephone directory. All randomly selected households were telephoned, and 1 person 15–75 years of age from every household was randomly selected and invited to participate in the study (18,19).

Interviews were conducted by trained interviewers in a computer-assisted telephone interview service, from June 18 to July 2, 2002, 7 days per week, during the day and the evening. Up to 10 attempts were made to contact each case-patient and control. Case-patients and controls completed a detailed telephone questionnaire, which included questions on demographics; contact with poultry, pet, and free-ranging birds; types of bird contact; other animal contact; and gardening and other outdoor activities undertaken in the 3 weeks before onset of illness (for case-patients) or the 3-week period April 1–21, 2002 (for controls).

Statistical analysis was conducted by using SAS Version 8.02 (SAS Institute Inc., Cary, NC, USA) (20). In univariate analysis, we compared characteristics and potential risk factors reported by probable case-patients and controls. Univariate analysis was performed by using chi-square tests and logistic regression analysis. When expected cell counts were <5, Fisher exact 2-sided p value was used. We performed multivariable logistic regression modeling using the backward stepwise elimination method to identify independent risk factors for psittacosis (21).

Serum samples were tested for Chlamydia and Chlamydophila species IgG (to endpoint titer) with a MIF assay (Chlamydia IgG SeroFIA, Savyon Diagnostics, Ashdod, Israel) using purified Chlamydophila pneumoniae, Chlamydia trachomatis, and Chlamydophila psittaci elementary bodies as antigen and a complement fixation (CF) assay using a Chlamydia genus-specific glycoprotein (Virion, Ruschlikon, Switzerland) (22,23). According to the manufacturer’s instructions, MIF titers >1:64 for Chlamydophila psittaci and Chlamydia trachomatis and >1:512 for Chlamydophila pneumoniae were regarded as indicative of current or recent infection. Serum samples were also tested for Chlamydia trachomatis IgG and IgA antibodies by using an enzyme immunoassay (EIA) (Medac Diagnostika, Wedel, Germany). All paired samples were tested in parallel in a single laboratory. Complement-fixing antibodies to Mycoplasma pneumoniae, influenza A and B viruses, adenovirus, and Coxiella burnetii were determined. Legionella pneumophila serogroups 1–6 and L. longbeachae total antibodies were tested by using an in-house immunofluorescence assay. In the later stage of the outbreak, upper respiratory tract specimens were collected from patients with pneumonia for Chlamydophila pneumoniae polymerase chain reaction (PCR) and for respiratory virus isolation. PCR for Chlamydophila pneumoniae was performed according to the published method (24).

To identify birds with avian chlamydiosis, we invited all local veterinarians and wildlife workers by letter to submit sick or dead birds for testing at a regional veterinary laboratory. The avian diagnostic test was based on direct immunofluorescence of a conjugated monoclonal antibody against a common, shared, genus-specific lipopolysaccharide antigen (Chlamydia-CEL VET IF TEST, Cellabs Diagnostics Pty Ltd, Brookvale, NSW, Australia). Household contacts of all probable case-patients were asked to submit any birds that the patient was known to have had contact with before onset of illness.

Figure

Figure. . Hospitalized patients with cases of probable psittacosis, Blue Mountains, by week of hospital admission.

Of 95 suspected cases identified, 59 (62%) had serologic evidence of psittacosis (probable cases), 30 (32%) were seronegative, and 6 (6%) were lost to follow up. The first patient with a probable case of psittacosis was hospitalized on March 11, 2002, marking the start of the outbreak, which peaked in late April to early May, with the last probable case-patient admitted to the hospital on June 29 (Figure). Of the 59 patients with probable cases of psittacosis, 36 (61%) were men, 50 (85%) resided in the upper Blue Mountains (altitude 770–1044 m), and 32 (54%) were 50–64 years of age. The average length of hospital stay was 7 days (range 2–29). No deaths occurred, although 2 patients required intensive care with intubation and mechanical ventilation.

We identified a large outbreak of probable psittacosis among residents of a forested district of Australia. Contact with live or dead free-ranging birds and lawn mowing without a grass catcher explained 70% of the cases. Possible reasons for the cessation of the outbreak include the onset of colder winter weather, which may reduce the prevalence or transmissibility of the infection among free-ranging birds and yard exposure among humans.

This study had several limitations. Case-patients and controls were exposed to media speculation over the role of parrots in the outbreak, so that people who believed they had psittacosis may have been more likely to report parrot exposure. Among patients, 28 completed hypothesis-generating questionnaires early in the investigation (between May 30 and June 7), which may have provided them with additional prompts to recall parrot exposure. However, diagnosis was unknown at the time patients completed the questionnaire, and few had laboratory evidence of psittacosis at the time they completed the questionnaire. Potential recall bias may also be countered because similar proportions of case-patients and controls reported no difficulty recalling contact with birds (case patients = 83%, controls = 85%, OR = 0.9, 95% CI 0.4–2.0, p = 0.7).

The serologic diagnosis of chlamydial infections is difficult because many assays lack specificity, and published data on the sensitivity and specificity of the MIF assay used in this study are limited (26). In this study, acute Chlamydophila psittaci infection, rather than Chlamydophila pneumoniae or Chlamydia trachomatis infection, was supported by 3 factors: 1) the seroconversion and 4-fold rises in Chlamydophila psittaci MIF IgG titers in samples from most case-patients (with most also having 4-fold rises in CF antibodies to Chlamydia), compared with lower MIF IgG titers to Chlamydophila pneumoniae and Chlamydia trachomatis, 2) the similar background rates of Chlamydia trachomatis–specific IgG on EIA in samples from the Chlamydophila psittaci seropositive probable and seronegative suspected case-patients, and 3) history of bird exposure among case-patients.

Analysis of hospital admission and discharge data indicated that, compared with the NSW state average and with the lower Blue Mountains average, increased rates of pneumonia have been seen among residents of the upper Blue Mountains during the autumn of previous years (unpub. data). This excess in pneumonia may be due to seasonal outbreaks of psittacosis in the area. Free-ranging birds are plentiful in the Blue Mountains, and chlamydiosis has been clinically diagnosed among juvenile birds, in particular crimson rosellas, in the area during the autumn of previous years (M. Cannon, pers. comm.).

Reasons for the positive association between probable psittacosis and residence in the upper Blue Mountains are speculative. The residents of the upper Blue Mountains are slightly older than that of the lower Blue Mountains (unpub. data) and may be more likely to remain at home and be active in their yards than the slightly younger residents of the lower Blue Mountains, who tend to commute to work in Sydney every day. This age difference may result in greater proximity between residents and free-ranging birds in the upper Blue Mountains.

Although the reason probable psittacosis case-patients were more likely to be 50–64 years of age is unclear, younger people may have had milder infection or been less likely to seek medical attention and require hospitalization (27,28). That persons with congestive cardiac failure or chronic obstructive pulmonary disease were excluded from the study may explain why persons 65–75 years of age appeared less at risk of disease.

Why a history of asthma seemed protective against psittacosis is uncertain. Persons with a history of asthma may be less likely to perform outdoor yard activities that generate dust and pollen. That persons who reported contact with caged or domestic birds appeared protected against disease may be related to previous exposure to Chlamydophila psittaci. While the development of protective immunity following infection with Chlamydophila psittaci has not been proven, laboratory-confirmed cases of reinfection have seldom been reported (2,29). That disease was less likely to develop in persons who reported potting or weeding may be because these activities generate little contaminated dust compared with activities such as lawn mowing.

Reported symptoms of the probable psittacosis case-patients in this outbreak were similar to those described elsewhere (1–3,5,10,15). One other community-wide outbreak of psittacosis has been reported to be potentially associated with contact with free-ranging birds (15). In 1995, 16 cases of psittacosis among residents of a rural town in a forested area in southern Australia were linked to trimming and mowing lawns and time spent in a yard, which are thought to be proxies for exposure to infectious particles shed by free-ranging birds with chlamydiosis (15). While this 1995 study did not find a significant relationship between illness and handling of dead birds (OR 2.4, 95% CI 0.25–21.05), the OR supports the positive association between illness and handling live or dead free-ranging birds (adjusted OR 7.4, 95% CI 2.5–22, p < 0.001) found in our study. The 1995 study reported the risk of illness associated with lawn mowing was not linked to use of a grass catcher, though whether the authors differentiated between those who never used a catcher and those who sometimes used a catcher is not clear. Previous studies have linked lawn mowing with psittacosis and with primary pneumonic tularemia (15,30). In our case-control study questionnaire, we differentiated between persons who mowed only with, with and without, and only without a grass catcher. We found this probable psittacosis outbreak to be linked to lawn mowing only without the use of a grass catcher. We could find no evidence in the literature suggesting that the use of a grass catcher provides lawn mower operators some protection from inhaling infectious particles, though grass catcher marketing materials of lawn mower manufacturers in Australia suggest that grass catchers reduce the generation of dust and hence the amount of dust expelled into the operator’s breathing zone (31).

This outbreak could have been missed altogether if not for alert clinicians. Laboratory evidence of psittacosis requires acute-phase and convalescent-phase (collected at least 6 weeks after onset) species-specific serologic testing for Chlamydophila or Chlamydophila psittaci PCR or culture. These tests require request by the treating doctor and referral to a specialized laboratory and enough concern to convince the patient to return for phlebotomy after recovery. In most clinical settings, underdiagnosis is therefore likely.

Clinicians should consider a diagnosis of psittacosis in persons with respiratory disease who reside in or have visited areas frequented by free-ranging birds. Health departments should also be alert for such outbreaks to facilitate diagnosis. Residents of areas frequented by free-ranging birds should avoid direct contact with dead or sick birds and bird droppings, use a particulate mask and gloves if contact is unavoidable, and avoid mowing lawns without a grass catcher. Further study is needed to identify effective measures, such as lawn mower design features, for reducing human exposure to Chlamydophila psittaci and other zoonoses.