Human Metapneumovirus Infection in Hospitalized Children

Discussion

To our knowledge the risk of AKI in our cohort of subjects with HMPV infection appeared to increase with age. Although older children were also more likely to have at least one comorbidity, the presence of specific comorbidities, aside from cystic fibrosis, was not associated with AKI. A recent large study using a national hospitalization database of > 2.5 million children in the United States reported the incidence of AKI to be 0.39%; the risk of AKI increased with age, with the median age of subjects with AKI being 10.8 y versus 2.0 y in subjects without AKI.¹³ Given that the incidence of AKI was 12.2% in our study, it appears that children with HMPV may be at higher risk of AKI than the general hospitalized pediatric population. A separate study of a general pediatric ICU population demonstrated an incidence of AKI as high as 26.9%,¹⁴ whereas the rate among children admitted to the pediatric ICU was 13.7% in our study. The development of AKI was not associated with the need for respiratory support, mechanical ventilation, or pediatric ICU admission in our study. This suggests that AKI during HMPV infection may not be strictly related to hypoxemia or hemodynamic instability, and that HMPV may have a more direct effect on the kidneys.

Other respiratory viruses have been associated with AKI via various mechanisms. Rat models have demonstrated direct viral infection of kidney tissues as well as cytokine-mediated kidney injury via infection with RSV.^15,16 Rhabdomyolysis-induced kidney injury is most commonly associated with influenza as well as other causes, including coxsackie virus, human immunodeficiency virus, and parainfluenza.^17,18

The presence of AKI in our cohort was based on ICD-9 coding; thus, we were unable to evaluate for the possible cause or classification of AKI. The mean degree of change in serum creatinine and estimated glomerular filtration rate among our subjects with AKI is consistent with AKI defined by either the KIDGO group (creatinine increase of 0.3 mg/dL or 150–200% of baseline, or urine output < 0.5 mL/kg/h for 6 h) or the modified RIFLE criteria (decrease in estimated glomerular filtration rate of 25% or urine output < 0.5 mL/kg/h for 8 h).^19,20 Urine output could not be determined accurately in our retrospective study. No data were collected regarding urinalysis results or creatinine phosphokinase levels in our cohort, so we are unable to comment further on the role of rhabdomyolysis. Future studies that evaluate a larger subset of HMPV-infected children with AKI would be useful to confirm our findings and delineate this risk further.

Although the incidence of HMPV decreases with age,²¹ it appears that the presence of comorbid conditions may be a risk factor for HMPV-related illness requiring hospitalization in older children. Comorbidities occurred at a higher rate among older children in our cohort. These findings are similar to a previous report in a general hospitalized population with HMPV showing 83% prevalence of comorbidities in subjects 5–17 y old versus 36% in those < 6 months old.¹⁰ Similar findings have been reported by others.²² Our cohort's baseline comorbidity rate (62%) was also comparable to those reported in other studies, which ranged from 56% to 68%.^6,10,23 Children hospitalized with HMPV are also more likely to have underlying comorbidities than children with RSV in a combined out-patient and in-patient cohort.⁸ Our findings imply that, although HMPV typically affects younger children, with increasing age it may be those with comorbidities who are at risk of requiring hospitalization.

Previous studies have reported mixed results on specific risk factors that contribute to worse hospital outcomes due to HMPV. In a pediatric ICU cohort, increased mortality was observed in females, those with underlying medical conditions, and those with hospital-acquired HMPV. However, a difference in hospital or pediatric ICU length of stay or need for mechanical ventilation based on demographics or comorbidities was not found.²² Comparatively, children with comorbidities with HMPV have been reported to have a significant increase in pediatric ICU admission (OR 4.3) and intubation (OR 7.1).¹⁰ These children also had an overall longer stay and more expensive hospital care. In that cohort, length of hospital stay was specifically greater in subjects with non-asthma pulmonary disease, neuromuscular diseases, and metabolic diseases compared to subjects with asthma or cardiac issues.¹⁰ Increased length of stay, need for ICU-level care, and longer use of noninvasive ventilation and supplemental oxygen in pediatric subjects with HMPV and underlying comorbidities have also been reported.²³

In comparison, our study identified the general presence of one or more of the specified underlying comorbidities to be a risk factor for needing respiratory support compared to previously healthy subjects. A history of asthma or prematurity conferred increased risk of respiratory support need. Subjects who were admitted for neutropenia in the setting of HMPV infection were at a lower risk of needing respiratory support, likely due to the indication for admission being a febrile neutropenia workup. This was also noted in previous studies.²³

Outside of these associations and the discussed AKI data, our results did not indicate any other underlying risk factors with regard to increased length of stay, ICU admission, or mechanical ventilation. It is important to note that some of the previous studies included central nervous system diseases, genetic syndromes, seizures, neuromuscular diseases, and chronic renal disease in their comorbidities, which may have an important impact on HMPV outcomes.

In our study cohort, 4 subjects with severe respiratory failure required ECMO. All 4 of these subjects had underlying lung disease and developed ARDS. All of their hospital courses were complicated by either bacterial or fungal coinfection. Although the children in our cohort who required ECMO survived to discharge, they had prolonged hospital stays. A previous report described 2 of 111 pediatric subjects admitted to 1 of 3 pediatric ICUs with laboratory-confirmed HMPV who required ECMO.²² Several case reports have described the need for ECMO in pediatric subjects with HMPV, including a 32-month-old patient with asthma who initially failed high-frequency oscillatory ventilation,²⁴ a 3-month-old infant who had been born at 27 weeks gestational age,²⁵ and a 9-month-old subject with biliary atresia who developed HMPV-related respiratory failure in the days following liver transplant.²⁶ All of these patients survived. Two pediatric patients a 7-month-old infant with acute lymphoblastic leukemia²⁷ and a previously healthy 2-year-old child who developed bilateral hemorrhagic bronchopneumonia, died despite the use of ECMO.²⁸

Our study has several limitations, including its retrospective nature, the use of ICD-9 codes to identify comorbidities, and the lack of inclusion of some comorbidities, such as neurologic and genetic conditions, that potentially are associated with hospital outcomes in other studies. The use of ICD-9 codes to identify AKI is also a limitation because this makes it difficult to draw definitive conclusions about the incidence of AKI during HMPV infection. The lack of urine output data as a supporting criterion for AKI diagnosis is another limitation. Additionally, as a 2-center study, we lacked sufficient numbers of pediatric subjects to provide a well-powered analysis of the multiple potential factors that could be associated with rare outcomes, including the development of AKI, need for ECMO, and death. Finally, not all subjects were tested for HMPV upon admission during the study period, and some cases were likely missed. Testing for HMPV was done at the discretion of the clinical care teams because no protocol or algorithm was in place for testing for HMPV.