Role for Maternal Asthma in Severe Human Metapneumovirus Lung Disease Susceptibility in Children

Abstract Background Severity of human metapneumovirus (hMPV) lower respiratory illness (LRTI) is considered similar to that observed for respiratory syncytial virus (RSV). However, differences in severity between these pathogens have been noted, suggesting the degree of illness may vary in different populations. Moreover, a potential association between hMPV and asthma also suggests that hMPV may preferentially affect asthmatic subjects. Methods In a population-based surveillance study in children aged <2 years admitted for severe LRTI in Argentina, nasopharyngeal aspirates were tested by RT-PCR for hMPV, RSV, influenza A, and human rhinovirus. Results Of 3947 children, 383 (10%) were infected with hMPV. The hospitalization rate for hMPV LRTI was 2.26 per 1000 children (95% confidence interval [CI], 2.04–2.49). Thirty-nine (10.2%) patients infected with hMPV experienced life-threatening disease (LTD; 0.23 per 1000 children; 95% CI, .16–.31/1000), and 2 died (mortality rate 0.024 per 1000; 95% CI, .003–.086). In hMPV-infected children birth to an asthmatic mother was an increased risk for LTD (odds ratio, 4.72; 95% CI, 1.39–16.01). We observed a specific interaction between maternal asthma and hMPV infection affecting risk for LTD. Conclusions Maternal asthma increases the risk for LTD in children <2 years old hospitalized for severe hMPV LRTI.

The clinical presentation and virulence of hMPV LRTI in infants and young children resembles that of other viruses, including respiratory syncytial virus (RSV), being also capable of eliciting life-threatening disease (LTD) [10,11]. That is why, in the context of the changing landscape of new interventions against RSV, it is important to understand how hMPV disease behaves in children and identify novel, preventable risk factors for severe disease.
Previously known risk factors for poor clinical evolution in children hospitalized for hMPV LRTI, such as prematurity, congenital heart disease, immunodeficiency, and neuromuscular disease, are not hMPV specific [3,[12][13][14][15]. Considering that the pattern of T-helper 2 (Th2) cytokines observed in respiratory secretions during hMPV LRTI may sometimes resemble that of individuals with asthma (including eosinophilic inflammation, higher interleukin-4 [IL-4] and IL-5 levels in mouse models [16], and induction of thymic stromal lymphopoietin in human airway cells [17,18]) we speculated that a history of asthmatic predisposition could specifically modify severity of hMPV disease in children.

Study Design
We conducted a prospective, multicenter active surveillance study to characterize the burden of severe LRTI in children from a low-income region in Buenos Aires, Argentina from 2011 to 2013. Details of the population are provided elsewhere [19]. The study was held in 12 public hospitals, from a geographically defined low-income region in the Buenos Aires suburbs, that provide care to an estimated population of 56 560 children younger than 2 years who lacked medical insurance [20]. Previous studies in this population examined RSV LRTI severity and the role of cytokines in human rhinovirus (hRV) infections [19,[21][22][23][24][25]. The study was approved by institutional review boards at each participating hospital, the state of Buenos Aires, and Vanderbilt University. All participating families signed an informed consent to join the study.
Eligibility criteria included infants and children younger than 2 years admitted for severe LRTI, defined as the sudden onset of cough, wheezing, retractions and/or crackles, with or without fever, and an oxygen saturation <93% when breathing room air or the need for O 2 support on arrival to emergency rooms. LTD was defined as O 2 saturation ≤87% on admission, requirement for mechanical ventilation, and/or admission to the intensive care unit.

Demographic and Clinical Information
Information on demographic, epidemiological, and clinical status was recorded from all participating patients from the time of admission until discharge using specifically designed questionnaires. In the study, we evaluated risk factors for hospitalization including prematurity (<37 weeks' gestation at birth), age, sex, lack of breastfeeding, malnutrition, house and floor materials, smoking at home, sources of heating, crowding (defined as more than 3 persons/room), parent's education, and availability of sewage system. Maternal asthma was defined as a physician diagnosis of asthma in the mother before infant enrollment.

Laboratory Tests
Nasopharyngeal aspirates were collected at the time of admission in patients from consenting families. Samples were tested in duplicate by real-time reverse transcriptase polymerase chain reaction (RT-PCR) for hMPV. Samples were also tested for influenza A viruses, RSV, and hRV, as previously described [19].

Statistical Analysis
Rates of severe and life-threatening hMPV disease were calculated by dividing the number of patients hospitalized with hMPV severe respiratory infection by the estimated census annual population in the selected area [20]. Student t test and χ 2 were used to compare children's clinical and epidemiological characteristics when appropriate. For the evaluation of risk factors associated with hMPV LTD, we used a multivariable logistic regression model. Covariates selected a priori and with a P value < .1 in univariate analysis were included in the model and analyzed in a stepwise forward logistic regression. A P value of less than .05 was considered statistically significant. Statistical analyses were performed using the Stata package for IBM-PC (Stata Corp).

Burden of Human Metapneumovirus LRTI
A total of 4045 hospitalized infants and young children met inclusion criteria; 3947 (98%) agreed to participate in the study. Clinical manifestations on admission were similar for participating and nonparticipating children (not shown). hMPV was detected in 383 (10%) children. Of these, 75 (20% of hMPV infections) were coinfected with RSV, 64 (17%) with hRV, and 1 with influenza A. Six patients were simultaneously infected with hMPV, RSV, and hRV. hMPV cases were more frequent during the first year of the study and always followed the same seasonal pattern as RSV with a distinct midwinter peak ( Figure 1). In each year, RSV was the virus most frequently detected in study participants.

A Subgroup of Infants at High Risk for Life-threatening hMPV LRTI
We subsequently explored whether children with hMPV (as a single pathogen or as a component of a coinfection) experienced disease of different severity than children infected with other viruses. Indeed, hMPV LRTI was milder in our population than disease elicited by RSV, hRV, and/or influenza A (OR, 0.53; 95% CI, .36-.79) ( Figure 3). But even though disease associated with hMPV was generally milder, a subgroup of infected children still experienced LTD. We therefore investigated whether a subpopulation with identifiable risk factors was disproportionality represented in this subgroup of subjects.
Given the Th2 polarization in respiratory secretions [16][17][18] and the reported association between hMPV hospitalizations and asthmatic children [12,26,27], we hypothesized that children of asthmatic mothers would be overrepresented among patients infected with the virus experiencing LTD.
Indeed, the rate of hMPV LTD in children of asthmatic mothers was 277/1000 compared to children of nonasthmatic mothers whose hMPV LTD rate was 3.4 times lower at 82.3/1000 (OR, 4.29; 95% CI, 1.42-12.93; P = .01 for LTD in children of asthmatic mothers vs children born to nonasthmatic mothers). The observed effect for maternal asthma remained significant after adjusting for important confounders (detailed in Table 2 Figure 4). Interestingly, when compared with patients with no virus detected as the control group, we observed that children admitted with hMPV infection and whose mothers were not asthmatic had a lowest risk of experiencing LTD (OR, 0.51; 95% CI, .33-.78; interaction P value = .024). However, when the mother was asthmatic, hMPV infection did not significantly increase the risk of having LTD compared to those without viral detections (Supplementary Table 1). Finally, given the rate of hMPV coinfections with RSV in our population (20% of hMPV infections), we explored whether hMPV played an instrumental role in the association between maternal asthma and LTD, or the effect was potentially attributable to effects of RSV. In the subgroup of children with hMPV infections in the absence of RSV (hMPV single infections and coinfections with hRV or influenza A), maternal asthma promoted LTD (OR, 3.8; 95% CI, 1.12-12.88; P = .03). The impact of maternal asthma in RSV infections alone was, conversely, not significant.

DISCUSSION
Our study reveals 2 interesting features of hMPV LRTI. While the virus appears to elicit milder disease among hospitalized children than other agents, among children admitted with hMPV LRTI, those born to asthmatic mothers are at increased risk for LTD. Twenty-eight percent of hMPV-infected patients hospitalized with LRTI and born to asthmatic mothers experienced LTD, in comparison to only 8% hMPV-infected patients hospitalized and born to nonasthmatic mothers. If confirmed in other studies and populations, these observations would be of considerable public health importance, given that asthma is the most frequent chronic disease during pregnancy with rising prevalence, and is now between 10% and 12% [28,29].
As reported by other groups, our study also showed that patients admitted for hMPV LRTI with previous comorbidities and those with severe complications during hospitalization were at higher risk for LTD. These risk factors have been described before for hMPV and other respiratory viruses [3,[12][13][14][15].
Several reports describe coinfections between hMPV and other respiratory viruses, mainly with hRV and RSV [41]. As others and our results suggest, hMPV may play an important role in these coinfections [42][43][44]. Perhaps this phenomenon is explained in part by "viral interference, " when one virus totally or partially blocks the replication of the other pathogen [45]. However, the role for hMPV as a coinfecting agent is still unclear and requires further study [46][47][48]. In a changing epidemiological landscape driven by preventive interventions against RSV, hMPV may soon alter its role and overall importance as a childhood pathogen should viral replacement occur [49]. Targeting children of asthmatic mothers for prophylaxis against hMPV may be a cost-effective strategy in the era of personalized medicine.
In addition, our study provides important disease burden information about hMPV. The hMPV infant hospitalization rate was surprisingly similar in our population to that in the United States, which is estimated at 2-4.9 per 1000 infants [5,26], and is considerably lower than that of RSV in other countries [23]. Importantly, the hMPV case fatality ratio in hospital was almost half that reported for RSV in the same population, highlighting that, to date, RSV remains the most important target for prevention [23].
Our study has limitations. First, because (as is the case for most studies) we did not analyze our samples for all other viruses, such as coronavirus or parainfluenza virus type 3, their effect on our observations is unknown. Second, given the age of our subjects and the prospective nature of this program, whether severity depended causally on maternal disease versus infant atopy/asthma cannot be answered. Third, given the observational nature of our study design focused on generating new hypotheses, no adjustment for multiple testing was performed. Additional research is needed to confirm the role of maternal asthma as a risk factor for severity in hMPV-infected children. However, our program also has significant strengths, including its prospective design and the evaluation of respiratory samples using state-of-the-art laboratory RT-PCR techniques. Furthermore, to our knowledge, this is the first population-based study seeking to define the burden of hMPV disease in a vulnerable population from a low-or middleincome country. In addition, this large study addressed numerous prenatal and perinatal risk factors, examining how they modify the severity of infection.
In summary, we report a novel observation that may alter our thinking about severe hMPV infections and, if confirmed in other studies, require us to reformulate its prevention strategies. hMPV represents a threat for life-threatening disease to young children born to asthmatic mothers.

Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.   Financial support. This work was supported by the Bill and Melinda Gates Foundation (grant numbers OPP1018614 and OPP1157162 to F. P. P.).
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.