Secondary Logo

Journal Logo

Original Studies

Change Over Time in Seasonality and Severity of Children Hospitalized With Respiratory Syncytial Virus Infection in Japan

Ozeki, Shoko MD*,†; Oshiro, Makoto MD; Fukumi, Daichi MD; Takeuchi, Tomoya MD, PhD; Mii, Sayaka MD; Nishikado, Yuichi MD

Author Information
The Pediatric Infectious Disease Journal: May 06, 2022 - Volume - Issue - 10.1097/INF.0000000000003568
doi: 10.1097/INF.0000000000003568

Abstract

Respiratory syncytial virus (RSV) causes acute respiratory tract illness in individuals of all ages, although its severity depends on a variety of factors, including age, underlying disease and history of infection. Risks of serious symptoms, such as lower respiratory tract infection include children under 6 months,1 children with chronic lung disease (CLD),1 preterm infants,1 infants with congenital heart disease (CHD),1 children with Down syndrome2 and children who are immunocompromised.3 In particular, RSV infections often cause respiratory failure in infants younger than 3 months, even without an underlying disease.

RSV causes seasonal outbreaks worldwide. In tropical and semi-tropical climates, seasonal outbreaks are usually associated with the rainy season.4 In temperate regions, many bronchiolitis epidemics from late autumn to winter are caused by RSV.4 In Japan, these epidemics usually occur from November to April, with a peak in January or February. Recently, the epidemic started in early August, a month earlier, based on the Infectious Disease Weekly Report for the 6 seasons from 2012 to 2017.5

In Japan, RSV rapid diagnosis kits are widely used; however, insurance is only applied to inpatients, outpatient infants less than 1 year of age and children who are eligible for palivizumab, a humanized monoclonal antibody. For this reason, RSV rapid antigen detection tests (RADTs) tend to be performed during the period when bronchiolitis is prevalent; thus, it is difficult to confirm whether other studies refer to actual seasonal variations. In our hospital, RSV RADTs are performed for almost all hospitalized children with suspected infectious diseases by means of control measures to prevent nosocomial infections during all seasons. Therefore, it has been possible to confirm RSV morbidity in children requiring hospitalization, even when RSV does not occur. The purpose of this study was to evaluate the seasonal variation in hospitalized patients with RSV infection over the past decade. Furthermore, we elucidated how different seasons influenced severity in children with RSV.

METHODS

Study Setting

From March 1, 2011 to February 28, 2019, children who were admitted to the pediatric ward for treatment of infection and underwent RSV RADTs were retrospectively enrolled at the Pediatric Department at Japanese Red Cross Aichi Medical Center Nagoya Daiichi Hospital, Nagoya. Our hospital has a relatively large pediatric ward in western Nagoya, and our department has 60 beds, 25 of which are for children with infectious diseases. We have performed RSV RADTs and influenza in almost all hospitalized children for the treatment of acute infection for patients cohorting to prevent nosocomial infection. The study was approved by the institutional review board of Japanese Red Cross Aichi Medical Center Nagoya Daiichi Hospital (institutional review board No. 2021-046).

Data Collection

We identified those who underwent RSV RADTs based on medical chart review. Here, we considered positive results of RSV RADTs as confirmation of laboratory-confirmed RSV infection. Both viral culture and polymerase chain reaction (PCR)-based assays have not been added because of the lack of rapidity for patient cohorting. As an exception, cases of infection that were diagnosed to be caused by an obvious pathogen were not considered for testing. The decisions in such cases were made by the physician in charge of admission. In some cases, the RSV RADT had been tested by the practitioner beforehand, and the physician decided not to perform the test again. In such cases, we adopted the test results for the referral letter. Additionally, some cases that underwent RSV RADTs before the date of hospitalization were considered included in patients who were positive if the RADT results were obtained within 7 days before hospitalization. Patients whose RADT was tested on and after the second day of the hospital admission were excluded because they could not be distinguished from nosocomial infections. A total of 5 cases were excluded: 2 cases admitted for gastroenteritis and dehydration tested positive on the third day of admission (no test on admission), 1 case admitted for cyclic vomiting required a test on the third day of admission, 1 case admitted for intestinal calculus tested positive on the 7th day of admission and 1 case admitted for acute disseminated encephalomyelitis tested positive on the 18th day of admission. In all the cases, upper respiratory tract symptoms were mild and there was no deterioration in respiratory status after the tests were performed. Patients who obtained negative RADT results upon admission were classified into the RSV-negative group.

In the RSV-positive group, medical records were reviewed for the following: (1) epidemiologic characteristics including age, sex and gestational age at birth; (2) whether they were eligible for RSV prophylaxis with palivizumab, wherein the vaccination had actually been done before admission; (3) the presence of underlying medical conditions, including prematurity, pulmonary disease, CHD, trisomy 21 or 18, congenital or acquired immunodeficiencies, neuromuscular disorders and presence of other congenital abnormalities; (4) whether they suffered from wheezing and (5) outcomes of care or disease-severity parameters, including length of stay, requirement of supplemental oxygen, continuous positive airway pressure (CPAP) support or high-flow nasal cannula (HFNC), the need and length of mechanical ventilation, requirement and length of stay in the ICU. Oxygen inhalation was initiated at peripheral capillary oxygen saturation (SpO2) ≤94%. The RSV-negative group only investigated the epidemiologic characteristics and the presence of wheezing.

The attending physician decided whether the patient had wheezing. Some patients are wheezing at the first consultation and developed wheezing during the course and are judged as wheezing positive.

Demographic and baseline characteristics of patients who were RSV-positive and RSV-negative were compared. Associations between categorical and continuous variables were analyzed using the χ2, Fisher’s exact and Mann–Whitney U tests, as appropriate. Descriptive analyses were performed using the frequency distributions or rates. Median values (25th–75th percentiles) were used to summarize the demographic data and baseline characteristics of patients.

To assess seasonality, we divided the year into 4 periods of 3 months each: spring (March-May), summer (June-August), fall (September-November) and winter (December-February) and created variables to represent the seasons. Since we knew from past trends that RSV has a cyclical pattern, we defined March through February of the following year as the RSV season. To examine the proportions of the 4 seasons between the different RSV seasons, we used the Cochran–Armitage trend test.

In view of the changing prevalence of RSV, we compared the before transition (2011–2012 to 2015–2016) and post-transition (2016–2017 to 2018–2019) periods. We compared the characteristics of patients with RSV who required oxygenation before and after the transition, as well as the proportion of patients who required oxygenation between seasons before and after the transition. Probability value (P value) of less than 5% (P < 0.05) was considered statistically significant. All statistical analyses were performed using the JMP ver. 15 (SAS Institute, Cary, NC).

RESULTS

From March 1, 2011 to February 28, 2019, 3750 hospitalized children with suspected infectious diseases who underwent RSV RADTs were retrospectively enrolled. A total of 945 cases (25.2%) were RSV-positive, whereas 2805 (74.8%) were RSV-negative. Of these, 2403 cases (64.1%) were children aged 24 months and younger, 740 cases (30.8%) were children with RSV and 1663 cases (69.2%) were children with non-RSV.

We compared the baseline demographic characteristics of children with and without RSV infections (Table 1). The median age of children hospitalized for RSV was significantly younger than that of children with non-RSV infections [11 (interquartile range {IQR} 4–21) vs. 19 (IQR, 8–42) months, respectively; P < 0.001]. The length of hospital stay of children who were RSV-positive was longer than that of children who were RSV-negative [5 (IQR, 4–6) vs. 4 (IQR, 3–6) days, respectively; P < 0.001]. The rate of wheezing was significantly higher in the RSV-positive group than in the RSV-negative group (53.5% vs. 16.2%; P < 0.001). The percentages of RSV tests we examined throughout the period are shown in Figure, Supplemental Digital Content (https://links.lww.com/INF/E728). The average testing rate over the study period was 48.6%; the RSV seasonal testing rate had declined over the years, from 51.9% in 2011–2012 to 43.3% in 2018–2019.

TABLE 1. - Baseline Characteristics
All Cases RSV Positive RSV Negative P value
Total 3750 945 (25.2%) 2805 (74.8%)
Age, months
3 702 (18.7%) 232 (24.6%) 470 (16.8%)
4–12 793 (21.2%) 265 (28.0%) 528 (18.8%)
13–24 908 (24.2%) 243 (25.7%) 665 (23.7%)
25 1347 (35.9%) 205 (21.7%) 1142 (40.7%)
Median(IOR) 11 (4–21) 19 (8–42) <0.001
Length of stay, days
1–-5 597 (63.2%) 1983 (70.7%)
6–29 342 (36.2%) 770 (27.5%)
30 6 (0.6%) 52 (1.9%)
Transfer 2
Median (IOR) 5 (4–6) 4 (3–6) <0.001
Wheeze on auscultation 506 (53.5%) 453 (16.2%) <0.001

F1
FIGURE 1.:
Changes of the seasonal distribution of children hospitalized for RSV-positive cases. The figure shows the number of people hospitalized in each of the 4 seasons and the percentage of the total RSV season.

Changes of the Seasonal Distribution of Children Hospitalized for RSV-Positive Cases From 2011 to 2019

We analyzed the total number of hospitalized children per RSV season. Figure 1 displays the number of children hospitalized for RSV-positive cases from 2011–2012 to 2018–2019. Looking at the trends in notified cases per season per year, there was a tendency during summer to be troughs from 2011–2012 to 2015–2016, and then it changed to a tendency during summers to be the peaks, whereas autumns and winters to be the troughs after 2016–2017.

F2
FIGURE 2.:
Changes of the seasonal distribution of children hospitalized for RSV-positive cases required oxygen supply. The figure shows the number of children hospitalized for RSV-positive cases required oxygen supply in each of the 4 seasons and the percentage of the total RSV season.

For children who are RSV-positive, the percentage of hospitalizations during summer significantly increased from 6.8% (13/190 cases) in 2011–2012 to 46.3% (38/82 cases) in 2018–2019 (P = 0.02) (Fig. 1). On the other hand, the proportion of children admitted to the hospital was clearly decreasing during autumn [50.0% (95/190 cases) in 2011–2012 to 20.7% (17/82 cases) in 2018–2019 (P = 0.001)] and during winter [28.4% (54/190 cases) in 2011–2012 to 20.7% (17/82 cases) in 2018–2019 (P < 0.001)]. There was no significant difference in the percentage of hospitalizations during the spring (P = 0.5).

Changes of the Number of Children Hospitalized for RSV-Positive Cases Required Oxygen Supply From 2011 to 2019

In the positive group, 73.7% (696/945 cases) required an oxygen supply. Intubation was performed in 8 cases. Except for the intubated cases, eight patients required CPAP or HFNC. One of the intubated cases was transferred to the hospital for extracorporeal membrane oxygenation.

Figure 2 shows patients who required respiratory support during the RSV season. The total percentage of children requiring oxygen supply did not change over time. In addition to the number of hospital admissions, the number of children requiring oxygen supply tended during summers to be the peaks, whereas autumns and winters to be the troughs after 2016–2017. Considering the trend change with a turning point in 2016, we divided the data into before transition (from 2011–2012 to 2015–2016) and after transition (from 2016–2017 to 2018–2019). Table 2 shows a comparison of the characteristics of children hospitalized for RSV-positive cases that required oxygen supply by season. It was found that before transition more older children were hospitalized during spring than during autumn and winter. However, no significant difference was observed after the transition. Moreover, we compared of the characteristics of children hospitalized for RSV-positive cases that required oxygen before and after the transition, and the results showed that the age of children decreased only during spring, whereas the number of oxygen administration cases and length of stay increased (Table 3).

TABLE 2. - Comparison of the characteristics of children hospitalized for RSV-positive cases required oxygen supply by season
Before transition
Seasons Spring Summer Autumn Winter
Age, median (IQR), month 17 11 14 10
(9–6.75) (3–18.25) (4–25) (2–18)
With undelying disease, % 14.9 16.7 17.6 14.1
Wheeze on ausculation, % 46 48.2 56.7 56.4
Length of stay, median (IQR), days 4 (3–6) 5 (4–6) 5 (4–6) 5 (4–6)
After transition
Seasons Spring Summer Autumn Winter
Age, median (IQR), month 12.5 12 10 7
(7–18.5) (3–23) (3–18) (2–20)
With undelying disease, % 28.1 18.4 19.8 13
Wheeze on ausculation, % 59.4 44.8 51.4 54.4
Length of stay, median (IQR), days 5 5 5 5
(4.25–6) (4–6) (4–6) (4–6.25)
Statistically significant values are shown in bold.
(As for age, significant difference was observed during spring and summer, spring and autumn, spring and winter, autumn and winter.)

TABLE 3. - Comparison of the characteristics of children hospitalized for RSV-positive cases required oxygen supply before and after transition
Spring Summer
Before or after transition Before After P Value Before After P Value
Oxgen supply, % 58.1 87.5 0.01 61.1 72.4 0.19
Age, median (IQR), month 17 (936.8) 12.5 (718.5) 0.034 11 (3–18.25) 12 (3–23) 0.61
Wheeze on ausculation, % 46 59.4 0.29 48.2 44.8 0.73
Length of stay, 4 (3–6) 5 (4.3–6) 0.03 5 (4–6) 5 (4–6) 0.5
Median (IQR), days
Autumn Winter
Before or after transition Before After P Value Before After P Value
Oxgen supply, % 72 78.4 0.20 79.1 78.3 >.99
Age, median (IQR), month 14 (4–25) 10 (3–18) 0.094 10 (2–18) 7 (2–20) 0.76
Wheeze on ausculation, % 56.7 51.4 0.37 56.4 54.4 0.87
Length of stay, 5 (4–6) 5 (4–6) 0.34 5 (4–6) 5 (4–6.25) 0.71
Median (IQR), days
Statistically significant values are shown in bold.

F3
FIGURE 3.:
Comparison of the percentage of children hospitalized for RSV-positive cases required oxygen supply before and after transition. The bar chart showing the percentage of children hospitalized for RSV-positive cases required oxygen supply in each of the 4 seasons. The numbers in the bars shows the actual number of them.

Furthermore, analyzing per season per period, there are more cases requiring oxygen during autumn and winter than during spring or summer before the transition (Figure 3). However, no significant difference was observed after the transition.

DISCUSSION

In this retrospective cohort study, the number of children hospitalized for RSV-positive cases changed to a tendency during summers to be the peaks, whereas autumns and winters to be the troughs after 2016–2017. So far, no conclusion has been reached regarding whether the epidemic period has shifted or prolonged with the earlier start of the RSV epidemic. However, at this time, the trough was also captured, suggesting that during summer to the peak. The strength of our study is that all infections were examined during the entire season, which was more accurate than other reports that only examined suspected RSV cases or the epidemic period.

There are several possible reasons for the RSV epidemic moving during summer. First is the climatic factor. In temperate regions, the factors, low temperature and high relative humidity were consistently shown to be associated with the number of RSV cases.6 Yusuf et al7 suggested that RSV infections were likely to be prevalent when the temperature range from 24–30 °C to 2–6 °C and humidity of 45%–65%. Shobugawa et al8 suggested that high temperature and relative humidity were positively associated with RSV outbreaks during summer. As summer is becoming hotter and hotter every year, it can explain the change in the RSV epidemic in Japan. The peak of the RSV epidemic in Okinawa, which has a subtropical climate, was seen during summer,8 but global warming may have accelerated the peak across the country. In addition, the increase in the number of visitors in Japan is also considered to be another cause. It is possible that the flow of people from Asia, especially from tropical and subtropical regions, affects the timing of the RSV epidemic. Actually, the diffusion of the severe acute respiratory syndrome coronavirus 2 and the implementation of restrictive measures led to a drastic reduction in RSV diffusion.9 In 2021, when RSV began to circulate again, the trend was different from previous years and warrants consideration.

In terms of seasonality and severity, the number of children requiring oxygen supply also had a tendency during the summers to be the peaks, whereas autumns and winters to be the troughs after 2016–2017. We assessed the data divided into before transition (from 2011–2012 to 2015–2016) and after transition (from 2016–2017 to 2018–2019), and we found that there were more cases of oxygen administration during autumn and winter before transition. Comparing the pretransition patient backgrounds, we found that the median age was higher during spring and the length of hospital stay was shorter than in any other season, except during summer. Therefore, during spring there were many older children hospitalized and few cases required oxygen, but during fall and winter, there were many younger children and many cases required oxygen. After the transition, there was no significant difference in oxygen supply or patient background. These things suggest that in the past, many younger children were hospitalized during the autumn and winter epidemics and required more oxygen, but recently, the effect of age has decreased and seasonal factors in the need for oxygen are no longer seen.

The risk of severe disease is known to be less than 6 months, which includes CLD, preterm birth, and CHD. Genetic predisposition is also known to be associated with severe diseases. In addition, it has been shown that epidemic strains change from subgroup A to B and B to A every 2–3 years.10 Although there was no evidence of seasonal variation in the risk of severe disease to date, our results suggest that the change in severe cases was due to the forward shift of the RSV epidemic.

It has been reported that the hospitalization rate of the patients with RSV is high in the younger age group,11 and the median age of hospitalization in this study was positive and predominantly younger. However, it cannot be denied that the median age of the negative group may be high because of the high admission rate of children with underlying diseases in patients who are febrile at our hospital. With regard to the length of hospitalization, there was no difference between the RSV-positive and RSV-negative groups in the group below 24 months of age, but when all ages were included, the length of hospitalization in the RSV-positive group was longer than those in the RSV-negative group. The wheezing rate of RSV has been reported to range from 74%–92% of hospitalized cases,11,12 but it was low in the present case. This may be due to the large number of cases admitted in the preliminary stages leading to bronchitis or bronchiolitis as a result of active screening for RSV infection using RADTs.

This study has some limitations. First, we identified patients with RSV based on the results of the RSV rapid test. It is known that the accuracy is high,13 but it is less sensitive than viral culture or PCR-based assays. Although the sensitivity and specificity were not 100%, it was judged to have little impact on the results because the study was aimed at seasonal differences. Second, the testing rate at our hospital was around 50%, which was lower than expected. It was because we did not include cases in which testing was performed at general practitioner’s office and those in which infection was already determined to be caused by an obvious pathogen. Cases in which the previous physician’s RSV test result was negative and infection by a pathogen other than RSV was suspected were not retested. In addition, the average rate of testing among hospitalized infectious disease cases declined over the years. This could be attributed to the increased proportion of cases for which rapid tests were performed in outpatient clinics. Moreover, the test coverage varied depending on the season due to the prevalence of diseases caused by specific pathogens (rotavirus and mycoplasma virus, etc.) Even if the test rate is reduced, the validation of RSV diagnosis is ensured, so the results will not be affected. Third, there are few serious cases, and only the requirement for supplemental oxygen could be used as an index of severity. In this study, it was clear that a shift to the summer months occurred with respect to severe cases evaluated solely in terms of oxygen delivery. However, factors such as days in the hospital, days of oxygen supply, use of respiratory devices, and duration of intensive care unit admission also need to be assessed when evaluating severe cases. Only 1.9% of patients required a respiratory device, which is generally considered to be a serious condition, and further evaluation is needed to increase the number of cases.

Our findings that the number of children hospitalized for RSV-positive cases changed to a tendency during summers to be the peaks, whereas autumns and winters to be the troughs after 2016–2017. This suggests that the period shifted with the earlier start of the RSV epidemic. We found that until 2016, there was an association between seasonality and severity, such that many younger children were hospitalized during the autumn and winter epidemics and required more oxygen, but after 2016, this association was no longer observed. However, further research is needed with more cases of severe illnesses.

REFERENCES

1. Boyce TG, Mellen BG, Mitchel EF Jr, et al. Rates of hospitalization for respiratory syncytial virus infection among children in medicaid. J Pediatr. 2000;137:865–870.
2. Beckhaus AA, Castro-Rodriguez JA. Down syndrome and the risk of severe RSV infection: a meta-analysis. Pediatrics. 2018;142:e20180225.
3. Anderson NW, Binnicker MJ, Harris DM, et al. Morbidity and mortality among patients with respiratory syncytial virus infection: a 2-year retrospective review. Diagn Microbiol Infect Dis. 2016;85:367–371.
4. Obando-Pacheco P, Justicia-Grande AJ, Rivero-Calle I, et al. Respiratory syncytial virus seasonality: a global overview. J Infect Dis. 2018;217:1356–1364.
5. Yamagami H, Kimura H, Hashimoto T, et al. Detection of the onset of the epidemic period of respiratory syncytial virus infection in Japan. Front Public Health. 2019;7:39.
6. Julian WT, Tze PL. Correlations between climate factors and incidencea contributor to RSV seasonality. Rev Med Viral. 2014;24:15–34.
7. Yusuf S, Piedimonte G, Auais A, et al. The relationship of meteorological conditions to the epidemic activity of respiratory syncytial virus. Epidemiol Infect. 2007;135:1077–1090.
8. Shobugawa Y, Taleuchi T, Hibino A, et al. Occurrence of human respiratory syncytial virus in summer in Japan. Epidemiol Infect. 2015;143:1110–1118.
9. Greta DM, Raffaella N, Enrica M, et al. During the COVID-19 pandemic where has respiratory syncytial virus gone? Pediatr Pulmonol. 2021;56:3106–3109.
10. Tsutsumi H, Onuma M, Suga K, et al. Occurrence of respiratory syncytial virus subgroup A and B strains in Japan, 1980 to 1987. J Clin Microbiol. 1988;26:1171–1174.
11. Caroline BH, Geoffrey AW, Aaron KB, et al. Respiratory syncytial virus-associated hospitalizations among children less than 24 months of age. Pediatrics. 2013;132:e341–e348.
12. Mary TC, Xing Q, Brenda T, et al. Development of a global respiratory severity score for respiratory syncytial virus infection in infants. J Infect Dis. 2017;215:750–756.
13. Caroline C, Nicolas T, Christian R, et al. Diagnostic accuracy of rapid antigen detection tests for respiratory syncytial virus infection: systematic review and meta-analysis. J Clin Microbiol. 2015;53:3738–3749.
Keywords:

respiratory syncytial virus; seasonality; severity; Japan; children

Supplemental Digital Content

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.