During 2010–2011, a pertussis outbreak occurred in a suburban town (town A) in Nobeoka City in Miyazaki Prefecture, Japan. Town A has a population of 4,227 persons and an elementary school and junior high school. The first pertussis case (in a 17-year-old boy) was reported in September 2010. From that time, we conducted a laboratory-based active surveillance study in the area until April 2011, in cooperation with clinics, hospitals, and local health departments. Pertussis-suspected cases were defined as an illness with cough lasting for >2 weeks, and pertussis-confirmed cases were defined as the presence of 1 of the following laboratory findings: a culture-positive result for Bordetella species from NPS specimens, or a positive result for molecular testing for Bordetella species.

For molecular testing, we conducted conventional PCR specific for insertion sequence IS481, which is known to detect B. pertussis and B. holmesii, and B. pertussis–specific loop-mediated isothermal amplification assays (10–12). To further confirm B. holmesii infection, B. holmesii–specific real-time PCR specific for the recA gene was also performed as described (8). For confirmed cases of B. holmesii infection, we collected the general information for patients (clinical symptoms, treatment, contact information, and outcome) by face-to-face interview or questionnaire.

During the pertussis outbreak, we identified 76 suspected pertussis cases. Among these suspected cases, 35 cases were confirmed by laboratory testing and involved persons 2–63 years of age (median age 13 years); 1 case occurred in a 2-year-old child, 14 in 6- to 12-year-old children, 14 in 13–15-year-old children, and 6 in persons >16 years of age. Despite pertussis vaccination rates in childhood of 82.3%−92.6%, most (80%) patients were students 6–12 and 13–15 years of age. Among the 35 confirmed case-patients, 29 and 6 patients showed B. pertussis and B. holmesii infection, respectively. Confirmed cases of B. holmesii infection were observed within the last half of the epidemic curve, i.e., weeks 1–12 of 2011 (Figure 1). There were no cases of co-infection with B. pertussis and B. holmesii.

All NPS specimens from patients with B. holmesii infection showed a negative result for the B. pertussis loop-mediated isothermal amplification, but showed positive results in the IS481 PCR and B. holmesii recA real-time PCR (Table 1). B. holmesii-like organisms were obtained from 5 patients, and these were identified as B. holmesii by recA gene sequencing. Patient 6 had been treated with antimicrobial drugs before the culture test, which probably resulted in a culture-negative test result in this patient. Real-time PCR confirmed that patient 6 had a low B. holmesii DNA load (threshold cycle 36.6) in her NPS specimen. All patients experienced paroxysms of coughing without posttussive vomiting, especially at night, and 3 also experienced a whoop (Table 2). Five of 6 patients had a persistent cough lasting >2 weeks. No patients experienced any complications, and they were treated mainly with azithromycin, resulting in complete recovery. None of the patients were immunocompromised.

Our surveillance study showed epidemiologic linkage between 5 patients (patients 2–6), but not for patient 1 (Figure 2). Patient 1, a student in high school A, had the first case of B. holmesii infection (probable index case). He lived in a dormitory outside Nobeoka City. However, his family home was in town A in Nobeoka City, and he returned to his home at the end of 2010 and remained there in 2011. Patients 2, 3, and 4 were students at the same junior high school (B) in town A and were close friends. Patient 5 was a teacher at junior high school B and was in charge of patient 4. Patient 6 was a medical practitioner at clinic C, which is 1 of 2 clinics in town A. Patients 1–5 visited clinic C in early January, late February, late February, early March, and late March, 2011, respectively. The duration of illness in patient 1 did not overlap with that of the other patients, whereas that of patients 2–6 clearly overlapped.

In the past 16 years, ≈70 B. holmesii clinical strains have been isolated from human patients in several countries, mainly the United States. All reported cases of B. holmesii infection have been sporadic occurrences. Thus, the reservoir of B. holmesii is currently unknown. Moreover, whether B. holmesii is transmitted between humans is not known. In this report, we have demonstrated that 5 patients infected with B. holmesii showed epidemiologic linkage. In particular, the fact that 4 of these patients attended the same junior high school suggests that B. holmesii may be transmitted from person to person.

A previous report suggested that B. holmesii and B. pertussis may co-circulate in young adults (7). However, the relationship between pertussis epidemics and B. holmesii infection is not fully understood. Our active surveillance study showed that B. holmesii infection spread concurrently with the B. pertussis epidemic, but that there was no co-infection of B. holmesii and B. pertussis. Our observations demonstrate that accurate diagnosis is needed to discriminate between B. holmesii and B. pertussis infections during a pertussis outbreak because symptoms associated with these 2 diseases are similar.

In 2012, a patient with B. holmesii infectious pericarditis was reported in Japan (13). This is probably the first case report of B. holmesii infection in Asia. Previous surveillance studies conducted in the United States and Canada have shown low rates (0.1%–0.3%) of B. holmesii infection in patients with cough (7,8). However, in a recent study, B. holmesii DNA was detected in 20% of NPS specimens collected from patients in France who had been given a diagnosis of B. pertussis infection (14). These surveillance data indicate that B. holmesii infection is present in adolescents and adults, and that the organism is associated with pertussis-like symptoms. However, other causes of viral or bacterial respiratory infection cannot be excluded. Because of lack of specific diagnostic tools to detect bordetellae, B. holmesii infection may have been underestimated. Accurate diagnosis and further studies are required to fully elucidate the nature of B. holmesii infection.