Lung involvement in children with COVID‐19 multisystem inflammatory syndrome

Abstract

Since the beginning of the COVID‐19 pandemic, multisystem inflammatory syndrome in children (MIS‐C) has been reported in increasing numbers, mostly focusing on cardiac dysfunction. Very few studies have evaluated lung involvement in terms of imaging findings, while data regarding pulmonary function in children with MIS‐C are not available. The purpose of our study was to evaluate lung involvement in MIS‐C by imaging and lung function by structured light plethysmography (SLP) at hospital admission and 6 months afterwards. Spirometry is the gold standard technique to evaluate lung function in children. However, SLP has the advantage of not requiring contact with the patient, offering an effective solution for the evaluation of lung function during the pandemic. To our knowledge this is the first study that aims to investigate pulmonary function by SLP in children with MIS‐C.

To the Editor,

Since the beginning of the COVID‐19 pandemic, multisystem inflammatory syndrome in children (MIS‐C) has been reported in increasing numbers, mostly focusing on cardiac dysfunction. Very few studies have evaluated lung involvement in terms of imaging findings, ¹ , ² while data regarding pulmonary function in children with MIS‐C are not available.

The purpose of our study was to evaluate lung involvement in MIS‐C by imaging and lung function by structured light plethysmography (SLP) at hospital admission and 6 months afterwards. Spirometry is the gold standard technique to evaluate lung function in children. However, SLP has the advantage of not requiring contact with the patient, offering an effective solution for the evaluation of lung function during the pandemic. To our knowledge this is the first study that aims to investigate pulmonary function by SLP in children with MIS‐C.

In this single‐center, retrospective study, we included 21 patients admitted to Buzzi Children's Hospital, in Milan, Italy from October 2020 to February 2021 with a diagnosis of MIS‐C, according to the Center for Disease Control and Prevention case definition. ³ Institutional ethics approval was obtained (N2021/ST/13).

Chest X‐ray and lung ultrasound (US) were performed within 24–48 h of admission. A pediatric radiologist reviewed the chest X‐ray and US examination of patients in two different sessions. A severity score of pulmonary X‐ray involvement was calculated adopting a simplified version of the Radiographic Assessment of Lung Edema Score. ⁴ An US severity score of pulmonary involvement was calculated using the global lung US aeration score. ⁵ To evaluate lung function, every patient underwent SLP evaluation. Tidal breathing was recorded for at least 3 min with an SLP device (Thora‐3DiTM, PneumaCare, Ltd.). The SLP recording allowed calculation of tidal breathing parameters and the ratio of the inspiratory flow at 50% of the tidal breathing to the expiratory flow at 50% of the tidal breathing (IE50). ⁶ Values of IE50 less than 1.3 were considered in the normal range, according to previous studies. ⁷

Statistical analysis was performed using R software, version 4.0.0. Continuous variables were presented as mean ± standard deviation or median (interquartile range) where their distributions were considered not symmetrical. Categorical variables were presented as counts and percentages. To show the direction of the Spearman correlation coefficients between the lung morphology (i.e., chest X‐ray and lung US) and function (IE50) and their general trend, we computed the 95% confidence interval (CI) for the correlation and we presented them graphically by means of forest plots.

Patient demographics and clinical characteristics are summarized in Table 1.

Table 1.

Demographics, clinical characteristics, and clinical course of patients with multisystemic inflammatory syndrome in children (MIS‐C)

Age, median (IQR)	10.1 (5.7–12.6) years
Sex, No. (%)
Male	17 (81)
Female	4 (19)
Ethnicity, No. (%)
Caucasian	16 (76.2)
Hispanic	4 (19)
Afro‐American	1 (4.8)
Asian	0
Comorbidities (at least one underlying condition), No. (%)	3 (14.3)
Symptoms and sign at presentation, No (%)
Fever	21 (100.0)
Gastrointestinal	20 (95.2)
systemic (asthenia, myalgia, headache)	7 (33.3)
Mucocutaneous	7 (33.3)
Cough	2 (9.5)
Dyspnea	1 (4.8)
Neurologic	1 (4.8)
Clinical presentation
Duration of symptoms prehospitalization, median (IQR)	5 (4–6) days
Organ systems involved, median (range)	3 (2–5)
Type of organ involvement, No. (%)
Cardiovascular	18 (85.7)
Respiratory	16 (76.1)
Gastrointestinal	14 (66.7)
Neurologic	4 (19.0)
Renal	3 (14.3)
Hematologic	4 (19.0)
Treatment, No. (%)
Intravenous immunoglobulin	21 (100.0)
Corticosteroid	20 (95.2)
Inotropes	7 (33.3)
LMWH	18 (85.7)
Clinical outcomes
Duration of hospitalization, median (IQR)	12 (10–16) days
Admission to ICU, No. (%)	16 (76.2)
Oxygen support, No. (%)	3 (14.3)
Noninvasive positive pressure ventilation, No. (%)	8 (50.0)
High flow nasal cannula (HFNC), No. (%)	1 (6.25)
Invasive mechanical ventilation, No. (%)	0
Duration of ventilatory support, median (IQR)	3 (2–4) days

Abbreviations: EF, ejection fraction; EFt0, ejection fraction during hospitalization; EFt6, ejection fraction at 6 months; FEV1, forced expiratory volume in the 1st second; FVC, forced vital capacity; ICU, intensive care unit; IE50, inspiratory to expiratory flow at 50% tidal volume; IQR, interquartile range; LMWH, low molecular weight heparin; No., number; US, ultrasound.

The therapy protocol adopted, including immunoglobulin and corticosteroids, is described in detail elsewhere. ⁸

Sixteen children (76.2%) required admission to the intensive care unit, all needed noninvasive ventilation; seven patients also required inotropic support (43.8%).

Most patients (10/21 [47.6%]) had three organ systems involved, 6/21 two systems, 2/21 four systems and 3/21 with five organs implicated. The most commonly were the cardiovascular (85.7%), respiratory (76.1%) and gastrointestinal (66.7%), followed by hematological (19%), neurological (19%), and renal (14.3%) (Table 1).

Pulmonary abnormalities were found in all patients at radiography, while lung US were unremarkable in 3 patients. Perihilar interstitial thickening was the most common finding observed on chest X‐rays. Consolidations were always found in the lower lobes and in only one case a bilateral pleural effusion was noted. Median chest radiograph severity score was 3 (interquartile range [IQR]: 2–4). At lung US, we found B lines in the lower lobes while pleural effusion was never found. The median lung US score was 3 (IQR: 2–6).

Pulmonary function was assessed by SLP. A mild obstructive disorder was observed in 15/21 patients (71.4%), with IE50 above normal values with a median of 1.4 (IQR: 1.29–1.6).

Pulmonary parenchymal alteration at hospital admission, identified by chest X‐ray score, showed a strong positive correlation with parenchymal anomalies identified by US score (r = 0.91, p = 0.001). Neither correlation between pulmonary morphological alterations at chest X‐ray at admission and heart dysfunction, identified by ejection fraction (ejection fraction [EF]%) (r = 0.30, p = 0.576), nor between US score and EF% (r = 0.23, p = 0.411) were found.

The presence of lung obstructive dysfunction, expressed by IE50, showed no correlation with heart dysfunction (r = 0.18, p = 0.238) nor with parenchymal alterations at admission expressed by a chest X‐ray of score (r = −0.19, p = 0.909).

The length of hospitalization was not correlated with heart involvement as systolic EF at admission (r = −0.05), otherwise it was positively associated with lung involvement in terms of both US score (r = 0.35, p = 0.035) and IE50 (r = 0.33, p = 0.06).

The relationship between noninvasive ventilation and intensive care unit hospitalization with pulmonary morphological or functional alterations was explored by logistic univariate models. The need for noninvasive ventilation was correlated with US‐score (p = 0.0255), while it was not associated with IE50. The odds ratio from the coefficients of the logistic regression model resulted 1.81 (1.16, 3.47 95% CI). Figure 1A reports the correlation coefficient estimates and their 95% CIs.

Figure 1.

(A) Forest plot for Spearman rank correlation coefficients and 95% confidence interval for parameters at hospital admission. (B) Forest plot for Spearman rank correlation coefficients and 95% confidence interval for parameters at 6‐month follow‐up [Color figure can be viewed at wileyonlinelibrary.com]

Six months after hospital discharge, all patients underwent echocardiography, lung US, SLP and spirometry. EF% as well as lung US were both unremarkable. Pulmonary function, evaluated by spirometry, was normal in all patients. At SLP, 14/21 patients showed IE50 values in the normal range, while 7/21 (33.3%) had IE50 values persistently above upper limits (>1.3).

IE50 values at 6 months did not show correlations with US score and X‐ray score during acute phase (r = −0.20, p = 0.414; r = −0.38, p = 0.091).

IE50 showed a strong, negative correlation with FEV1 percentage and FVC percentage (r = −0.81, p = <0.001; r = −0.82, p = <0.001). In Figure 1B, the correlation coefficients along with their 95% CI are presented in a forest plot.

The chest X‐ray showed lung involvement in all patients with a median chest radiography severity score of 3, similar to that reported by Rostad et al. ⁴ According to literature, perihilar interstitial thickening was the most common finding, and consolidations showed lower zone predominance.

At lung US, we predominantly observed the presence of multiple B‐lines with lower lobe localization, according to literature data. ² , ⁹ Our data showed a strong correlation between chest X‐ray score and US score in identifying lung parenchymal abnormalities; hence, the employment of US in clinical practice may be considered a useful tool useful tool in children with MIS‐C.

Conversely, no correlation was observed between cardiac involvement (EF%) and lung involvement, in terms of both imaging and function, raising the hypothesis that lung involvement might be independent from cardiac involvement.

Regarding pulmonary function, we found a mild obstructive pulmonary function defect in children affected by MIS‐C during the acute phase, revealed by the increased value of IE50 at SLP. To our knowledge, this is the first study reporting data on pulmonary function in children affected by MIS‐C.

At the 6‐month follow‐up both lung US and echocardiogram showed complete recovery after the acute phase, in agreement with literature data. Spirometry parameters were unremarkable in all patients, showing a significant correlation with SLP parameter IE50.

Even if heart dysfunction has been considered the main feature in patients with MIS‐C, ¹⁰ our findings suggest that the lung are also involved. This could be due to the cytokine storm itself rather than secondary to cardiac involvement.

However, these results need confirmation by further studies.

In conclusion, our study suggests the importance of further investigating the lungs as the target organ of MIS‐C in prospective powered studies. In addition, the pandemic has encouraged clinicians to explore novel diagnostic techniques. In this context, SLP has proved valuable for investigating lung function. It could also be used in non pandemic conditions in the more vulnerable patients who need nonvolitional techniques.

AUTHOR CONTRIBUTIONS

Michele Ghezzi: conceptualization; methodology; data curation; writing ‐ original draft; writing ‐ review & editing; investigation. Emma Longoni: methodology; data curation; investigation; writing ‐ original draft. Alice Munari: methodology; data curation; investigation. Irene Raso: methodology; data curation; investigation. Giacomo Biganzoli: formal analysis. Gianvincenzo Zuccotti: conceptualization; writing ‐ review & editing; supervision. Enza D'Auria: conceptualization; writing ‐ review & editing; supervision; writing ‐ original draft.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

Ghezzi M, Longoni E, Munari A, et al. Lung involvement in children with COVID‐19 Multisystem Inflammatory Syndrome. Pediatric Pulmonology. 2023;58:615‐618. 10.1002/ppul.26224

DATA AVAILABILITY STATEMENT

Data are not available.