Differences in clinical severity of respiratory viral infections in hospitalized children

Abstract

It is uncertain whether clinical severity of an infection varies by pathogen or by multiple infections. Using hospital-based surveillance in children, we investigate the range of clinical severity for patients singly, multiply, and not infected with a group of commonly circulating viruses in Nha Trang, Vietnam. RT-PCR was performed to detect 13 respiratory viruses in nasopharyngeal samples from enrolled patients. We apply a novel clinical severity score and examine associations with the odds of being severe and differences in raw severity scores. We find no difference in severity between 0-, 1-, and 2-concurrent infections and little differences in severity between specific viruses. We find RSV and HMPV infections to be associated with 2- and 1.5-fold increase in odds of being severe, respectively, and that infection with ADV is consistently associated with lower risk of severity. Clinically, based on the results here, if RSV or HMPV virus is suspected, PCR testing for confirmatory diagnosis and for detection of multiple coinfecting viruses would be fruitful to assess whether a patient's disease course is going to be severe.

Figure 1

Demographics of the study cohort. Figure shows the demographics of the cohort. Underlying conditions of chronic diarrhea, cardiac, neurological, renal, malnourished, and thalassemia had less than 5% prevalences each. Similarly with discharges of rhinitis, pharyngitis, tonsillitis, otitis media, croup, trach, bronchitis, meningitis, and sepsis had prevalences less than 5%. We note that the colors are arbitrary. SES, socioeconomic status.

Figure 2

Distribution of CSS by virus. Figure shows the distributions (counts) of CSS by virus. Multinomial confidence intervals are calculated using the method of Sison and Graz, as implemented in the R package MultinomialCI. ADV Adenovirus, HBoV Human bocavirus, HMPV Human metapneumoviruses, HPIV-1, -2, -3, -4 Human parainfluenza viruses 1–4, HRV Human rhinovirus, Flu A influenza A, Flu B influenza B, RSV Respiratory syncytial virus.

Figure 3

Percent of patients with severe disease by virus for no-, mono-, and multiple-infections and percent of severe and non-severe cases by virus and number of viruses detected in the NP sample. Panel (A) shows the percent of patients suffering severe disease (CSS $>3$) by virus and whether the patient was PCR negative (dashed black line), PCR positive for 1 virus (light blue), or PCR positive for more than 1 virus (dark blue). Line segments indicate 95% binomial confidence intervals. Panel (B) show the percent of severe and non-severe patients by virus and PCR positive for 1, 2, or $\ge 3$ viruses, respectively. ADV Adenovirus, HBoV Human bocavirus, HMPVHuman metapneumoviruses, HPIV-1, -2, -3, -4 Human parainfluenza viruses 1–4, HRV Human rhinovirus, Flu A influenza A, Flu B influenza B, RSV respiratory syncytial virus.

Figure 4

Difference in proportion severe cases between all viruses. Figure shows the difference in percent of patients suffering severe disease (CSS $>3$) by virus with each other virus. Dark squares outlined in black indicate significant differences between percents as tested by a one-sided proportion test. For example, the bottom row indicates the proportion severe ADV cases is significantly lower than proportions of severe cases from HBoV, HMPV, HPIV3, HRV, Flu A, and RSV. This is comparing mono- and multiply-infected individuals. ADVAdenovirus, HBoV Human bocavirus, HMPVHuman metapneumoviruses, HPIV-1, -2, -3, -4 Human parainfluenza viruses 1–4, HRV Human rhinovirus, Flu A influenza A, Flu B influenza B, RSV respiratory syncytial virus.