MATERIALS AND METHODS

Study population

Okayama is located in western Japan, with a population of approximately 2 000 000. Kawasaki Medical School Hospital is the main hospital for the prefecture and includes all clinical specialities. Between April 1991 and December 2000, 4756 serum samples were collected from healthy individuals aged 6 months to 88 years (2488 males and 2268 females) who had lived in the same area of Okayama for more than 10 years (children under 10 years old had also lived in the same area since their birth). Serum samples were collected and stored at −70°C until measurement.

Antigen preparation

The C pneumoniae KK-pn15 strain, which was isolated in our laboratory from a patient with acute pharyngitis,³ was used as the antigen and was propagated in HEp-2 cells. The elementary bodies of the KK-pn15 strain were purified by a method of continuous Urografin (Schering AG, Berlin/Bergkamen, Germany) gradient centrifugation (40–52%), as described previously.⁴ The elementary bodies were resuspended in a solution of 2% yolk sac in phosphate buffered saline (pH 7.2) containing 0.02% formalin.

Serological testing

A MIF test was used to measure antibodies to C pneumoniae.³ The presence of IgG and IgM antibodies against C pneumoniae was detected using commercial fluorescein isothiocyanate conjugated goat antihuman IgG and IgM (Medical and Biological Laboratories, Nagoya, Japan). Rheumatoid factors were absorbed with Gullsorb (Gull Laboratories, Salt Lake City, Utah, USA) before IgM titrations. The criterion for a positive serological test was a titre of ≥ 1/16 for IgG.¹ An anti-C pneumoniae antibody of titre ≥ 1/16 for IgM was considered to indicate a recent infection. Recent reports concerning serological criteria indicate that the definition of acute infection using a titre of ≥ 1/512 IgG is controversial. Therefore, in our study, a titre of ≥ 1/512 for IgG was defined as a high titre. Statistical analysis was done by means of the χ² test.

RESULTS

Figure 1 shows the age related incidence of anti-C pneumoniae antibodies during the entire period. The antibody detection rate was 57%, broken down by age as follows: 6 months to 1 year old, 4%; 2–4 years old, 5%; 5–9 years old, 40%; 10–14 years old, 56%; 15–19 years old, 59%; 20–29 years old, 57%; 30–39 years old, 63%; 40–49 years old, 67%; 50–59 years old, 67%; 60–69 years old, 70%; and ≥ 70 years old, 71%. In subjects over 15 years old, anti-C pneumoniae antibodies were present in 1331 (67.5%) of 1971 males and 1078 (60.6%) of 1777 females. There were 14 cases (0.3%) of recent C pneumoniae infection among these healthy individuals, as determined by IgM.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Population prevalence of anti-Chlamydia pneumoniae IgG antibody by age and sex in 4756 individuals (closed circles, male; open circles, female individuals).

As shown in fig 1, the prevalence of antibody to C pneumoniae varies in the population in relation to age and sex. Therefore, to estimate year to year differences we randomly selected sera that were matched for stratified age (15–29, 30–49, 50–69, and ≥ 70 years old) and sex in each year and compared them. Table 1 shows the prevalence of anti-C pneumoniae antibody in the different year groups between 1991 and 2000. The antibody positivity rate was between 59.0% and 73.3%. In 1993 and 1999, there were distinct peaks in the positivity rate: 73.3% and 73.0%, respectively. In contrast, in 1996, the prevalence was 59.0%, which was the lowest during the observation period. There were significant differences in the prevalence between the highest years (1993 and 1999) and other years (1993, p = 0.00058; 1999, p = 0.00092). Table 1 also shows the prevalence of high titres of IgG (≥ 1/512) and IgM ≥ 1/16. The prevalence of high titre sera was between 2.6% and 7.6%. In 1999, there was also a peak in the prevalence rate of high titres and this was significantly higher than in other years (p = 0.00075).

We also analysed the season and seropositivity. No evidence of seasonality was noted, in that parallel seasons in different years did not show similar patterns of seroprevalence.

DISCUSSION

The seroepidemiology of C pneumoniae infection in Japan was studied by analysing the prevalence of anti-C pneumoniae antibody over a 10 year period. Our data and other epidemiological data support a major role for C pneumoniae infections in causing changes in the anti-C pneumoniae antibody positivity rate.^5,6 In our study, year to year variability was detected in the incidence of C pneumoniae infections. In our area, we encountered outbreaks of C pneumoniae infection in some schools and families in 1993 as confirmed by the isolation of cell cultures, the polymerase chain reaction, and serology. The high prevalence in 1993 was thus related to the occurrence of C pneumoniae infection in our area. Therefore, the higher antibody prevalence in 1999 may reflect the occurrence of an outbreak of C pneumoniae infection in the same year, as described by Karvonen et al.⁵ The high prevalence of extremely high IgG titres in 1999 may also point to the possibility that an epidemic of C pneumoniae occurred in our area.

Chlamydia pneumoniae is spread through person to person transmission by droplets, and outbreaks of infection have been reported in families, schools, military barracks, and nursing homes. Countrywide epidemics of C pneumoniae have also been recorded in Scandinavian countries.⁶ Epidemics, obviously caused by C pneumoniae infections, have occurred in the Scandinavian countries in 1976/7, 1981/2, and 1987/8 and in Finland in 1977 and 1987/8.^5–7 The epidemic cycle, from the peak of high incidence to low incidence, has been estimated to be approximately four to six years in Scandinavia and approximately 10 years in eastern Finland.⁵ In our area, we detected two waves of largely subclinical infections that occurred at approximately six year intervals. Epidemiological surveillance of C pneumoniae infection must be continued to determine whether this interval reflects the epidemic cycle.

“It is intriguing to speculate that Chlamydia pneumoniae infection might contribute to the sex specific difference in the incidence of cardiovascular disease”

Surveys in various parts of the world have demonstrated that antibodies to C pneumoniae occur in over 50% of adult populations. This finding supports the assumption that antigenic boosting occurs throughout life.^1,2 There have been only two seroepidemiological studies of C pneumoniae infection among the Japanese population, using the MIF test.^8,9 Our seroepidemiological studies demonstrated that a very small proportion of children under 5 years of age show serological evidence of past infection with C pneumoniae. The prevalence then increases dramatically from ages 5 through to 14 years, and by age 20 years approximately 55% of those studied have detectable amounts of antibody to the organism. The seroprevalence continues to increase among older age groups, but at a slower rate, and reaches approximately 70% in the elderly. These observations are consistent with former reports.^8,9 However, the antibody prevalence appears to be higher in the lower age groups in Japan than in Western countries.^1,2

Pronounced sex differences in the prevalence of antibodies to C pneumoniae, with an excess of seropositivity among men, have been observed in many countries.^1,2 However, we found very little sex differences in our study; seropositivity was 59.9% in male and 54.3% in female individuals. However, the prevalence of anti-C pneumoniae antibodies in elderly individuals (≥ 60 years old) was significantly higher in men than in women (75.5% v 64.8%; p = 0.00029). It is intriguing to speculate that C pneumoniae infection might contribute to the sex specific difference in the incidence of cardiovascular disease, but further investigation in representative population samples is required to confirm a higher prevalence of infection in men. Recently, O'Neill et al reported that subjects who had blood taken during the winter months were more likely to be seropositive than subjects who had blood taken during other seasons.¹⁰ They suggested that infection with C pneumoniae may exhibit seasonality, with a higher risk of acquisition of infection during winter. In our study, however, we detected no evidence of seasonality in the past decade. This may be because only one winter period was described in O'Neill's study.¹⁰

In conclusion, we have shown that C pneumoniae infection is as highly endemic in Japan as it is in Western countries, and have detected year to year variability. Long term studies in Japan are needed to clarify the epidemic occurrence or periodicity of C pneumoniae infection.