Epidemiology of parainfluenza virus types 1, 2 and 3 infections based on virus isolation between 2002 and 2011 in Yamagata, Japan


Katsumi Mizuta, Department of Microbiology, Yamagata Prefectural Institute of Public Health, Tokamachi 1-6-6, Yamagata, Yamagata 990-0031 Japan. Tel: +81 23 627 1373; Fax:+81 23 641 7486; e-mail: mizutak@pref.yamagata.jp


To clarify the epidemiology of viral acute respiratory infections (ARIs), 305 human parainfluenza virus types 1 (HPIV1), 154 HPIV2 and 574 HPIV3 strains were isolated from 16,962 nasopharyngeal swabs obtained between 2002 and 2011 at pediatric clinics in Yamagata, Japan. The total isolation frequency for HPIV1–3 was 6.1%. Unlike HPIV1 infections, HPIV3 showed clear seasonality with yearly outbreaks in the spring–summer season. HPIV2 tended to appear biannually in autumn–winter. Although no reliable techniques for the laboratory diagnosis of these infections have been established, the present results suggest that HPIV1–3 are an important causative agent of ARIs in children.

List of Abbreviations: 

acute respiratory infection


cytopathic effect


National Epidemiological Surveillance of Infectious Diseases


human malignant melanoma


human parainfluenza virus


respiratory syncytial virus


reverse transcription

Human parainfluenza viruses are enveloped, negative-sense RNA viruses that belong to the family Paramyxoviridae (1, 2). There are four genetically different types: HPIV1 to HPIV4; HPIV1 and HPIV3 belong to the genus Respirovirus and HPIV2 and HPIV4 to the genus Rubulavirus (1, 2). Although HPIV4 is rarely reported, HPIV1–3 are important causes of various ARIs in children, including the common cold, croup, bronchitis, bronchiolitis, and pneumonia. They also commonly reinfect older children and adults. Although such infections are generally mild in healthy persons, they can cause serious disease in immunocompromised hosts (3). In Japan, fewer HPIV strains have been detected than have strains of other respiratory viruses, such as RSV (4). There have been few epidemiological studies and negligible data collected on HPIVs in Japan (5–8). Herein, we describe the results of virus isolation from patients with ARIs in Yamagata, Japan between 2002 and 2011, with particular focus on HPIVs.

In collaboration with the Yamagata prefectural health authorities for the national surveillance of viral diseases in Japan, between January 2002 and December 2011 we collected 16,962 nasopharyngeal swab specimens from patients with ARI attending two pediatric clinics (Yamanobe and Katsushima Pediatric Clinics). Among these specimens, 12,189 (71.9%) were from patients ≤ 5 years old, 2763 (16.3%) from patients between 6 and 9, 1466 (8.6%) from patients between 10 and 14, and 469 (2.8%) from patients ≥ 14. We placed the nasopharyngeal specimens in tubes containing 3 mL of transport medium and transported them to the Department of Microbiology, Yamagata Prefectural Institute of Public Health for virus isolation (9).

We have used HHVMRG plates, the acronym being derived from the human embryonic lung fibroblast (HEF), human laryngeal carcinoma (HEp-2), Vero, Madin Darby canine kidney (MDCK), rhabdomyosarcoma (RD-18S), and green monkey kidney (GMK) cell lines that they include, for virus isolation since June 2001 (10) and the HHVe6MRG plate, in which we substituted the VeroE6 cell line for the Vero cell line, since January, 2004 (9, 11). We used 96-well tissue culture plates (Greiner Bio-one, Frickenhausen, Germany) vertically and prepared two rows of each cell line as described previously (9, 11). Beginning in 2008, we also prepared HMV-II cell lines as separate 96-well tissue culture plates and inoculated the specimens onto them, mainly to isolate HPIVs (12, 13). After centrifugation of the specimens at 1500 g for 20 min, we inoculated 75 μL of supernatant directly into two wells of each cell line. We stored the remainder of each specimen at −80 C. We centrifuged the inoculated plates at 450 g for 20 min, incubated them at 33 C in a 5% CO2 incubator and assessed them for CPE for 14 days, except for the Vero E6 cell lines, which we observed for approximately one month without changing the medium to isolate human metapneumovirus (11).

When we observed a CPE or hemagglutination test and/or found a hemadsorption test to be positive using guinea pig erythrocytes (0.8%), we performed virus identification using a hemadsorption inhibition test, RT-PCR and sequence analysis as described previously (9, 12).

With regard to HPIVs, we isolated 1033 (6.1%) HPIV1–3 strains, comprising 305 HPIV1 (1.8%), 154 HPIV2 (0.9%) and 574 HPIV3 (3.4%) strains, from the 16,962 specimens we obtained during the study period. After we introduced the HMV-II cell line, the annual virus isolation frequencies of HPIV1–3 increased from 1.6 to 7.9% between 2002 and 2008 and from 9.4 to 10.8% between 2009 and 2011. Figure 1 shows monthly numbers of HPIV1–3 isolates. HPIV1 was uncommon in winter but quite commonly isolated between April and October. Further, although we isolated HPIV2 year-round, we recovered 55% of isolates between September and December. For HPIV3, we recovered 86% of isolates between May and July, but none between November and February, indicating that HPIV3 infections have clear seasonality. Figure 2 shows a breakdown of HPIV1–3 infections by age. For HPIV3, 53.5% of the children were younger than 2 years and the proportion decreased with age apart from the ≥ 10 years age group. In contrast, we found the highest percentage of HPIV1 and HPIV2 infections in the 2–4 years (2.4–2.7%) and 3–5 years (1.1–2.0%) age groups, after which the percentage of infections generally decreased with age. Regarding the clinical diagnosis of patients with HPIV1, HPIV2 and HPIV3 infections, 236 (77.4%), 123 (79.9%), and 458 patients (79.8%) were diagnosed with upper respiratory infections such as rhino-pharyngitis, respectively; 25 (8.2%), 11 (7.1%), and 13 (2.3%) with croup, respectively; 32 (10.5%), 18 (11.7%), and 63 (11.0%) with lower respiratory infections such as bronchitis, bronchiolitis, and pneumonia; and the rest with other diseases including viral exanthema.

Figure 1.

Monthly distribution of human parainfluenza virus types 1-3 strains isolated from patients with acute respiratory infections in Yamagata, Japan between 2002 and 2011.

Figure 2.

Proportion of human parainfluenza types 1–3 isolated by age in patients with acute respiratory infections in Yamagata, Japan between 2002 and 2011.

Human parainfluenza virus infections follow both endemic and epidemic patterns (1, 3); therefore, the seasonality of HPIVs varies depending on place, distinct variations being observed from year to year. The number of isolates of various viruses detected in public health laboratories all over Japan is available in the Infectious Agents Surveillance Report, Japan, for each year since 1981, the data between 1980 and 1991 being documented in published supplements (7, 8). All annual data are available from the NESID system (14). This NESID system database includes the data from Yamagata described in this study.

Several previous studies have reported that HPIV1 infections have clear outbreaks in autumn, mostly in September and November, either every two years (15–18) or at irregular intervals (19). In this study, we found no clear seasonality for HPIV1 infections, although HPIV1 infections did appear to be more common in odd-numbered years. In Japan, no source, including the NESID system, has indicated a seasonal pattern in HPIV1 infections (5–8, 14).

In comparison to the clear seasonality of HPIV1 and HPIV3 outbreaks, smaller yearly or irregular outbreaks of HPIV2 have reportedly occurred in autumn (15–19). In this study, we recovered many HPIV2 isolates in the autumn-winter season, observing a particular increase in even-numbered years since 2004 in Yamagata, Japan. The NESID system data support this trend: in the years prior to 1986, HPIV2 infections occurred more commonly in even-numbered years, apart from 1981 and 1983 (7, 8, 14). Thus, HPIV2 infections have commonly occurred in the autumn-winter season every two years in Japan, although this seasonality is less clearly observable than that of HPIV3.

In this study from 2002 to 2011 in Yamagata, Japan, we found HPIV3 infections to be grouped in clear yearly seasonal outbreaks, mainly between May and July. The data in the NESID system also show that HPIV3 infections have peaked in the spring-summer season since 1980 (7, 8, 14). Many previous studies have reported that HPIV3 causes yearly outbreaks, mainly in the spring-summer season, around the globe (15–20); the clear seasonality of HPIV3 in Yamagata appears similar to that observed in other areas.

It is generally accepted that HPIV3 as well as RSV infections are common in infants and young children, whereas HPIV1 and HPIV2 infections tend to be commoner in older persons (1–3, 15, 19). Knott et al. reported that the age distribution of HPIV3 infections peaks at 6 months–2 years of age, whereas HPIV1 and HPIV2 peak at 2–5 years (15): findings that are similar to our observations in this study. Clinically, fewer of our patients were diagnosed with croup (2.3–8.2%) than was reported by Knott et al. (9–45%) (15). However, both studies supported the contention that HPIV1 and HPIV2 are more strongly associated with croup than is HPIV3, which is in agreement with the trends described in various textbooks (1, 3).

This study indicates that the annual isolation frequencies of HPIV1–3 are 1.6–10.8%, and that those for HPIV1–3 infections are equal to or higher than those previously reported for RSV (2.2–3.2%) (9). However, in Japan fewer cases of HPIV1–3 than of RSV are detected: 1870 and 3462 cases were reported, respectively, between 2001 and 2010 (4). This could be because there is no established reliable technique for the laboratory diagnosis of HPIVs. Identification of the suspected causative agents and development of a system for their laboratory diagnosis are the first steps needed for the proper management and treatment of patients with infectious diseases. We hope this study will lead to a better understanding of the epidemiology and etiology of HPIVs and hopefully aid in the development of a rapid antigen test, such as immunochromatography, similar to those currently available for use in clinical settings for influenza virus, adenovirus, RSV and human metapneumovirus.


We thank the doctors, nurses and people of Yamagata Prefecture for their assistance and collaboration in the surveillance of viral infectious diseases. This work was partially supported by grants-in-aid from the Japan Society for Promotion of Science and for Research on Emerging and Re-emerging Infectious Diseases from the Ministry of Health, Labor and Welfare.


All authors declare they have no conflicts of interests.