Clinical manifestations and outcome of viral acute lower respiratory infection in hospitalised children in Myanmar

This prospective study investigated the clinical characteristics, clinical outcomes, and viral aetiologies in children in Myanmar who were admitted to the Yankin Children Hospital (YKCH) in Yangon, Myanmar for ALRI from April 2017 to March 2019. YKCH, affiliated with the University of Yangon, Medicine II, is one of the largest children's hospitals in Yangon with 550 beds. The hospital plays a central role in the primary to tertiary care of children in the Yangon metropolitan area, which has a childhood population of more than 1.7 million [29]. We enrolled patients aged 1 month to 12 years who were admitted to the general wards or the intensive care unit at YKCH under the diagnosis of ALRI. The maximum age at admission at YKCH was 12 years. We asked each paediatrician to recruit certain numbers of patients each month to investigate the viral cause of ALRI throughout the study period.

Based on the World Health Organization (WHO) definition [30, 31], ALRI was defined as (1) history or measured fever of ≥ 38 °C, (2) cough, (3) onset within the last 10 days, and (4) hospitalisation requirement. Particularly for patients aged < 5 years, ALRI was defined as (1) fast breathing: less than 2 months: ≥ 60 bpm, 2 months to less than 12 months: ≥ 50 bpm, 12 months to less than 5 years: ≥ 40 bpm, or (2) difficulty of breathing with chest indrawing and cough.

With regards to the patient characteristics, stunting was categorised using the height-for-age z-score: normal, ≥ − 2; moderate, ≥ − 3 to < − 2; severe, < − 3 [26]. Tachycardia was defined as a heart rate ≥ 160 bpm in children aged < 1 year and ≥ 120 bpm in children aged ≥ 1 year [19]. Hypoxemia was defined as peripheral oxygen saturation (SpO₂) ≤ 90% by pulse oximetry [19, 31]. Tachypnea was defined as a respiratory rate of ≥ 60 bpm in children aged < 2 months, ≥ 50 bpm in children aged 2–11 months, and ≥ 40 bpm in children aged ≥ 1 year [19, 30, 31]. Fever was defined as a body temperature ≥ 38.0 °C [19].

The laboratory examinations, including blood tests, bacterial culture, and chest radiographs, were ordered according to the physicians’ discretion. Radiological findings were evaluated by senior paediatricians using an endpoint of consolidation, over-inflation, and whether or not the findings were unilateral or bilateral. Admission criteria were based on each physician’s decision. However, the majority of the reasons for admission were (1) requirement of supplemental oxygen due to hypoxia, (2) requirement of hydration due to intolerable oral feeding, and/or (3) poor general condition.

Exclusion criteria for this study included those who (1) were recently hospitalised with acute respiratory illness within 7 days, (2) had already been enrolled in this study < 28 days, (3) diagnosed with bacterial ALRI by physical examination and microbiological tests (blood, nasopharyngeal and oropharyngeal swab, or sputum cultures); (4) had an alternative diagnosis of other respiratory disorders; (5) had a tracheostomy tube; or (6) had baseline diseases causing immunocompromised or altered respiratory status, such as chromosomal abnormalities, cancer, known respiratory diseases (bronchiectasis, cystic fibrosis, etc.), cerebral palsy, and human immunodeficiency virus infection.

All patients were conclusively diagnosed with ALRI by senior paediatricians. They also interviewed enrolled children and their caregivers using standardised questionnaires to collect information regarding the baseline characteristics and medical history. Data from medical records were abstracted systematically after discharge.

All methods used in the current study were performed in accordance with the relevant guidelines and regulations. Written informed consent was obtained from all the study participants, including their parents and guardians, before participation in the study. This study was approved by the ethics committees of the Department of Medical Research, Ministry of Health and Sports, Myanmar (016616) and Niigata University, Japan (2547).

Severe ALRI was defined when children with ALRI had (1) central cyanosis or SpO₂ ≤ 90%, (2) severe respiratory distress, or 3) “general danger signs,” such as inability to drink, persistent vomiting, convulsions, or unconsciousness [30, 31]. The WHO definition was designed for children aged 2 months to 5 years, but we adapted the definition to 1 month to 12 years to simplify the study criteria at the study site.

At the time of admission, paediatricians collected two nasopharyngeal swabs from each patient who satisfied the case definition and agreed to participate in the study. One swab was used for testing with a rapid test kit (Quick Navi-Flu + RSV, Denka Seiken Co. Ltd., Tokyo, Japan) at the medical ward at YKCH. The other swab was immediately placed in viral transport media, and stored at − 80 °C in a freezer at YKCH and subsequently transported to the laboratory of Niigata University, Japan for real-time reverse transcription PCR (RT-PCR). Both tests were performed for all the patients regardless of the rapid test results. The RT-PCR test results were used to estimate the sensitivity and specificity of the rapid tests for influenza and RSV.

After thawing, the samples were centrifuged at 1500 × g for 30 min at 4 °C, and the supernatants were used in the analysis. Viral RNA and DNA were extracted using a QIAamp MinElute Virus Spin Kit (QIAGEN, Valencia, CA, USA). Real-time reverse transcription PCR (RT-PCR) was performed using the One Step PrimeScript RT-PCR Kit (TaKaRa, Tokyo, Japan) with virus-specific primers and a TaqMan probe. For the RNA virus, the thermocycling settings were as follows: 42 °C for 5 min for cDNA synthesis, 95 °C for 3 min followed by 45 cycles at 95 °C for 5 s for denaturation, and 60 °C for 40 s for annealing and extension, except for human coronavirus (hCoV), which was under different conditions (46 °C for 40 s for the annealing and extension process). Regarding the DNA virus, the thermocycling settings were as follows: 95 °C for 3 min followed by 45 cycles at 95 °C for 5 s for denaturation, and 60 °C for 40 s for annealing and extension. Using this PCR assay with recorded cycle threshold (Ct) values to estimate relative RNA or DNA levels, 16 viruses were examined: RSV (RSV A and RSV B), hRV, influenza virus (influenza A and influenza B), hMPV, enteroviruses, PIV (PIV 1–3), human coronavirus (hCoV; NL63, OC43, 229E, and HKU), adenovirus, and hBoV [32‐41]. The primers and probes with concentrations used for each virus in the study are summarized in Additional file 1: Table S1. If the Ct value was above 35, we confirmed the existence of target bands in the PCR products by gel electrophoresis. The presence of target bands was confirmed as positive results.

In addition to the 16 viruses, a real-time PCR assay for Streptococcus pneumoniae [42‐44] was performed to estimate colonisation rates. Real-time PCR was performed using the specific primers designed to target the following genes: cps-A and Takara SYBR Premix Ex Taq II (Product code: RR820A; TaKaRa, Tokyo, Japan). The thermocycling conditions were as follows: 95 °C for 30 s, followed by 40 cycles at 95 °C for 5 s for denaturation, and 60 °C for 30 s for annealing and extension.

Bacterial culture, including blood, nasopharyngeal and oropharyngeal swab, or sputum, was obtained based on their availability when each physician suspected bacterial ALRI. Bacterial culture was repeatedly obtained when bacterial ALRI was suspected in patients who enrolled the study.

The primary outcome of this study was to clinically characterise the ALRI patients with in-hospital mortality. The secondary outcome was the identification of viral pathogens that caused ALRI.

Based on the YKCH statistics from 2015 to 2016, the annual inpatient admission of ALRI was reported to be approximately 3000 cases. The sample size was set to 300 patients each year, and we tried to recruit the patients each month to observe the pathogens causing ALRI at different time points in 2 years.

Paper-based survey data were entered independently and in duplicate by the study team. Data are reported as the median and interquartile range (IQR) or as the mean and standard deviation (SD) or percentages, where appropriate. Categorical variables were analysed using the χ² test, and continuous variables were assessed using the Mann–Whitney U-test. Statistical significance was set at p < 0.05. Statistical testing was performed using the Statistical Package for the Social Sciences (ver. 24) (IBM, Armonk, NY, USA).

In total, 5463 children diagnosed with ALRI were admitted to the YKCH between 1 April 2017 and 30 March 2019. The average number of admissions was 228 patients/month, and the number of patients admitted in 2017 and 2018 was 2,467 and 2,996, respectively. The number of ALRI cases peaked during the rainy season in both years, with the highest peak observed in September (solid line, Fig. 1). Of these, 570 (10.4%) eligible children were enrolled in the study. The clinical characteristics of the patients classified by age, severity, and detected number of viruses are summarised in Table 1 and Additional file 1: Table S2. The median age of the patients was 8 months (IQR: 4–15 months), and 56.5% were male. The patients were divided into the following 3 age groups; (1) 1–11 months, (2) 1–5 years, and (3) 6–12 years. Moderate to severe stunting (weight for age Z score < − 2) was observed in 138 (25%) patients. Among the eligible patients, 49.6% (n = 262) and 70.5% (n = 372) had received the 10-valent PCV10 and Hib vaccines (n = 528), respectively. Sick contact was observed in only 11% (65/570) of the patients. With regards to the time from the onset of disease, 70% (401/507) visited the hospital and received health care equal to or greater than 3 days after the onset of symptoms. The most common symptoms were cough (94%) and difficulty of breathing (73%).

Among the vital signs, tachypnoea was the most common finding (78%), with a median of 60 bpm (IQR: 48–64 bpm), followed by fever (≥ 38.0 °C) (41%), and tachycardia (36%) (Table 2). On physical examination, chest indrawing, rhonchi, and coarse crackles were commonly observed in 67%, 63%, and 51% of the patients, respectively. Symptoms and signs divided by z-score did not differ significantly (P ≥ 0.05).

Patients with ALRI were admitted to the ward for a median of 4 days (IQR, 3–7 days). Intensive care unit management was required in 51 patients (9%). The in-hospital mortality rate was 5% (28/570).

Although a viral infection was suspected, antibiotic therapy was initiated in more than half of the patients (325/570, 57%) based on each peadiatrician’s decision, not by a protocol or guideline for the use of antibiotics in the hospital. Among them, 107 (33%) patients received two or more antibiotics during hospitalisation (Table 2), and the simultaneous use of two or more antibiotics was documented in 91 (28%) patients. Third-generation cephalosporins (55%) and amoxicillin/clavulanic acid (34%) were commonly prescribed, followed by amoxicillin/flucloxacillin (13%), cefoperazone/sulbactam (9%), ampicillin or penicillin G (7%), piperacillin/tazobactam (6%), azithromycin (4%), carbapenems (3%), and quinolones (2%).

A complete blood count was obtained in 63% of the patients. The median white blood cell count was 13.4 × 10³/μL (IQR: 9.3 × 103/μL) (Table 3). C-reactive protein was measured in 41% of the patients and had a median of 1.6 mg/dL (IQR: 0.3–4.6 mg/dL). A chest radiograph was obtained in 59% of the patients, and consolidation and over-inflation were observed in 44% and 43% of the patients, respectively. Rapid diagnostic tests for both RSV and influenza were performed in 99% of the patients, with 128 (23%) and 23 (4%) patients testing positive for RSV and influenza, respectively (Table 4).

Viruses were detected in 88% of the patients. The number of detected and detected viruses is summarised in Table 5. RSV B (36%) and hRV (28%) were the most commonly detected viruses, followed by hMPV (13%), adenovirus (13%), enteroviruses (12%), RSV A (10%), and influenza A (9%) (Table 5 and Fig. 2). Comparing the infants (less than 12 months) and the group aged 6–12 years, RSV was more likely to be detected in infants (198 [52%] vs 54 [29%], P < 0.001). In contrast, influenza was less likely to be detected in infants (26 [7%] vs 27 [14%], P = 0.004).

In total, 12% (68/570) of the patients had no virus detected on RT-PCR, while 18% (12/68) were positive for RSV on the rapid test but were negative on PCR. The median age of the undetected cases was 8 months (IQR: 4–15 months), which was not significantly different from the detected cases (P = 0.91).

The epidemiological curves of representative viruses, including RSV, hRV, and influenza viruses, are shown in Fig. 1. RSV predominated in both 2017 and 2018, with a peak in September. The detection of influenza started in June 2017, and the detection rate was lower in 2018. hRV has been detected throughout the years, with peaks in August 2017 and September 2018.

In total, 45% (259/570) of the patients were PCR-positive for pneumococcus (Table 5). Among those who had received the 10-valent PCV, 49% (130/264) were positive, which did not significantly differ from those who had not received the vaccine (42%, 129/306) (P = 0.09).

No significant differences were observed between the two groups (P = 0.15, P = 0.15, respectively) when we compared the positive rates for pneumococcus between the severe (39%) vs non-severe cases (47%) and the fatal (32%) vs non-fatal cases (46%).

The clinical symptoms and signs of ALRI classified by each single detected virus are summarised in Table 6. Comparing the symptoms and signs of the patients infected with each virus and those infected with a single virus, patients infected with RSV were more likely to develop rhonchi (99 [79%] vs 95 [57%], P < 0.001), and those infected with hMPV were less likely to develop prolonged capillary refill time (0 [0%] vs 23 [90%], P = 0.042). Furthermore, those infected with adenovirus had higher rates of coarse crackles (7 [88%] vs 143 [51%], P = 0.040). Other significant findings included higher rates of fever in those infected with influenza (16 [64%] vs 107 [40%], P = 0.021) and higher rates of cyanosis in those infected with hBoV (3 [75%] vs 34 [12%], P = 0.007) than in those infected with viruses other than the virus listed.

Two or more viruses were detected in 211 patients (37%) (Table 5 and Fig. 2), with the most common combinations being RSV and hRV (31%), followed by hRV and adenovirus (9%), and hRV and hMPV (7%) (Table 7). Among the cases with two or more viruses detected, 2.3% of the patients were fatal. There were no dominant viruses or specific combinations of the viruses specifically identified in fatal cases (Table 7).

Next, we evaluated the clinical impact of the number of viruses detected on the clinical outcomes. The in-hospital mortality did not differ among those detected with one virus (7%), two viruses (1%) (P = 0.07), or 3–5 viruses (6%) (P = 0.59). Similarly, the median length of hospitalisation did not differ significantly among those detected with one virus (4.0 days [IQR: 3.0–7.0 days]), two viruses (4.0 days [IQR: 3.0–6.0 days], P = 0.98), or 3–5 viruses (4.5 days [IQR: 3.0–7.0 days], P = 0.57) (Additional file 1: Table S3).

Next, we analysed the mortality and median length of hospitalisation in the patients detected with RSV only and those detected with RSV and hMPV and hBoV, which are often detected simultaneously with RSV. The mortality of those infected with RSV with or without hMPV (4% vs 0%, P = 0. 66) or hBoV (25% v. 0%, P = 0.22) did not differ significantly. Similarly, the median length of hospitalisation among the patients infected with RSV with or without hMPV (4.0 days [IQR: 3.0–5.0 days] vs 4.0 days [IQR: 3.0–6.0 days], P = 0.90) or hBoV (7.5 days [IQR: 3.0–14.2 days] vs 3.5 days [IQR: 3.0–5.8 days], P = 0.65) did not differ significantly.

In total, 107 (19%) severe cases were identified. The median age did not differ between the severe (8 months [IQR: 3–12 months]) and non-severe cases (8 months, [IQR: 4–16 months]) (P = 0.39). The rate of cases that did not receive any routine immunisation was higher in the severe (22%) than in the non-severe cases (10%) (P = 0.01). A single virus was detected in 58 (54%) patients with severe cases, with RSV B (34%) and hRV (33%) being the most common. Among the severe cases, 20/107 (19%) died during hospitalisation, while there were only 8/463 (2%) mortalities in the non-severe cases (P < 0.001).

In total, 28 patients (4.9%) died. The median age did not differ between the fatal (7 months, IQR: 3–21 months) and non-fatal cases (8 months, IQR: 4–15 months) (P = 0.88). Stunting (39% [11/28] vs 23% [127/542], P = 0.056) and lack of immunisation (32% [9/28] vs 12% [63/542], P = 0.050) were more frequently observed in the fatal cases than in non-fatal cases.

Significant differences were observed in all the vital signs on admission between the fatal and non-fatal cases. Compared to the non-fatal cases, the fatal cases had a higher heart rate (160 bpm [IQR: 140–168 bpm] vs 140 bpm [IQR: 120–152 bpm], P = 0.001), body temperature (38.7 °C [IQR: 37.8–39.0 °C] vs 37.7 °C [IQR: 37.0–38.3 °C], P < 0.001), respiratory rate (69 bpm [IQR: 62–78 bpm], vs 60 bpm [IQR: 48–64 bpm], P < 0.001), and lower SpO₂ (92% [IQR: 80–98%)] vs 96% [IQR: 94–98%], P = 0.024).

On physical examination, the following were more frequently observed in the fatal cases than in the non-fatal cases: coarse crackles (93% vs 49%, P < 0.001), wheezing (61% vs 36%, P = 0.008), grunting (75% vs 20%, P < 0.001), cyanosis (71% vs 8%, P < 0.001), and prolonged capillary refilling time ≥ 3 s (11 [39%] vs 39 [7%], P < 0.001).

In 25/28 (89%) fatal cases, at least one virus was detected: one virus (n = 20, 71%), two viruses (n = 2, 7%), and three viruses (n = 3, 11%). Influenza A (29%) and RSV B (21%) were the most commonly detected viruses.