Medications and Adherence to Treatment Guidelines Among Children Hospitalized With Acute COVID-19

Abstract

Objective

COVID-19 treatment guidelines rapidly evolved during the pandemic. The December 2020 Infectious Diseases Society of America (IDSA) guideline, endorsed by the Pediatric Infectious Diseases Society, recommended steroids for critical disease, and suggested steroids and remdesivir for severe disease. We evaluated how medications for children hospitalized with COVID-19 changed after guideline publication.

Methods

We performed a multicenter retrospective cohort study of children ages 30 days to < 18 years hospitalized with acute COVID-19 at 42 tertiary care US children’s hospitals April 2020 - December 2021. We compared medication use before and after the December 2020 IDSA guideline (pre- and post-guideline) stratified by COVID-19 disease severity (mild-moderate, severe, critical) with interrupted time series.

Results

Among 18,364 patients who met selection criteria, 80.3% were discharged in the post-guideline period. Remdesivir and steroid use increased post-guideline relative to the pre-guideline period, although the trend slowed. Post-guideline, among patients with severe disease, 75.4% received steroids and 55.2% remdesivir, and in those with critical disease, 82.4% received steroids and 41.4% remdesivir. Compared to pre-guideline, enoxaparin use increased overall but decreased among patients with critical disease. Post-guideline, tocilizumab use increased and hydroxychloroquine, azithromycin, anakinra, and antibiotic use decreased. Antibiotic use remained high in severe (51.7%) and critical disease (81%).

Conclusions

Although utilization of COVID-19 medications changed following December 2020 IDSA guidelines, there was a decline in uptake and incomplete adherence for children with severe and critical disease. Efforts should enhance reliable delivery of guideline-directed therapies to children hospitalized with COVID-19 and assess their effectiveness.

Article Summary

We evaluated medication utilization and adherence to treatment guidelines in children hospitalized for COVID-19 across the United States.

Introduction

Coronavirus disease 2019 (COVID-19) forced clinicians, researchers, and public health experts to confront a novel pathogen with limited information. As the pandemic evolved, health systems rapidly developed policies and treatment guidelines to align with emerging evidence. The impact of the pandemic on United States (US) children has been substantial; as of April 11, 2022, children accounted for 19% of all US COVID-19 cases, over 41,000 hospitalizations, and almost 1,500 deaths.^1,2 Furthermore, as variants emerged, such as Delta and Omicron, prevalence and hospitalizations increased among children with COVID-19.^1,2

Management of acute COVID-19 in children was particularly challenging as the pediatric population was largely excluded from clinical trials and management recommendations were often based on expert consensus or extrapolated from adult data.³ Therapies studied during the pandemic included investigational agents such as remdesivir and repurposed medications (e.g., hydroxychloroquine, ivermectin).⁴ Furthermore, prophylactic anticoagulation to prevent thrombotic complications and antibiotics due to concern for secondary bacterial infections were commonly used.^5,6

National best practice guidelines evolved rapidly as new data emerged. By the end of March 2020, the Food and Drug Administration (FDA) issued an Emergency Use Authorization (EUA) for chloroquine and hydroxychloroquine, and in May 2020, remdesivir received an EUA for treatment of hospitalized patients with severe COVID-19.^7,8 However, by June 2020 the EUA for chloroquine and hydroxychloroquine was revoked.⁹ In August 2020, the Infectious Diseases Society of America (IDSA) and Pediatric Infectious Diseases Society (PIDS), recommendations focused on glucocorticoid and remdesivir use with some variation in recommendation strength by severity of illness.^10,11

By December 2020, coinciding with a winter surge in COVID-19, revised IDSA guidelines (endorsed by PIDS) were published online.¹² The guidelines focused on acute COVID-19 treatment recommended glucocorticoids for hospitalized critically ill patients, while noting that remdesivir appeared less beneficial in these patients (e.g., mechanical ventilation or extracorporeal membrane oxygenation (ECMO)).¹² Steroids and remdesivir were suggested for patients with severe disease (e.g., SpO2 ≤ 94% on room air).¹²

In the context of shifting recommendations, the extent to which therapies recommended and not recommended by the December 2020 national guidelines were utilized for hospitalized children with COVID-19 remains largely unknown. In this study, we describe how treatment of children with acute COVID-19 at US children’s hospitals evolved and the extent to which clinicians followed published national COVID-19 treatment guidelines from December 2020 onward.

Methods

Study design

We performed a multicenter retrospective cohort study of children hospitalized with COVID-19 at 42 tertiary children’s hospitals in the US who are part of the Children’s Hospital Association (Lenexa, KS) and contribute clinical and resource utilization data to the Pediatric Health Information System (PHIS). This billing and utilization database includes International Classification of Diseases 10^th Revision (ICD-10) codes for diagnoses and Current Procedural Terminology (CPT) codes for procedures, as well as demographics, and is subject to quality and reliability checks.^13,14 The Institutional Review Board (IRB) at the Children’s Hospital of Philadelphia determined the study did not meet criteria for human subjects research.

Study population

We identified the index hospitalization for children ages 30 days to less than 18 years discharged from April 1, 2020 to December 30, 2021. We included children with a primary COVID-19 diagnosis (ICD-10 code U.071) or a secondary COVID-19 diagnosis when their primary diagnosis was a complication known to occur concurrently with acute COVID-19 (Appendix Table 1).^14,15 The intent of these inclusion criteria was to capture patients with symptomatic COVID-19 as the primary reason for admission, as opposed to patients who screened positive on admission but were hospitalized for unrelated conditions. COVID-19 ICD-10 code U.071 was established April 1, 2020, rapidly adopted, and has high predictive value in identifying hospitalized patients with a positive polymerase chain reaction result.^14,16,17 We excluded neonates who never left the hospital and children less than 30 days to be consistent with previous studies that suggested this population is frequently admitted for other non-COVID conditions.¹⁴ We excluded hospitalizations for Multisystem Inflammatory Syndrome in Children (MIS-C) or Kawasaki Disease as our objective was to describe treatment for acute COVID-19 (Appendix Table 2). MIS-C is a post-COVID-19 complication requiring different treatment.

Medications

Medications for COVID-19 were included in the analysis if they had an FDA EUA and/or were listed in IDSA COVID-19 treatment guidelines.^4,7,8,12 Systemic corticosteroids included dexamethasone, methylprednisolone, prednisolone, and prednisone. Hydroxychloroquine was included as use in this population was more likely intended for COVID-19 therapy than antimalarial treatment. We also examined medications used to prevent thrombotic complications (enoxaparin) and treat potential secondary bacterial infections (antibiotics) as these were commonly used in this cohort. Antibiotics were identified using the National Healthcare Safety Network list of all pediatric antibacterial agents for systemic administration.¹⁸ Azithromycin was evaluated separately from other antibiotics as it was a proposed COVID-19 treatment. Famotidine was not included due to difficulty in determining whether it was intended to treat COVID-19.

Disease Severity Categories

COVID-19 disease severity was classified in mutually exclusive categories including mild-moderate, severe (non-critical), and critical disease (e.g., ECMO, shock, mechanical ventilation) to mirror IDSA guidelines and previous COVID-19 and respiratory disease-related research (Table 1).^4,12,14,19

Table 1.

COVID-19 Disease Severity Classification and Recommended Medications

Critical Disease	Severe (Non-critical) Disease	Mild – Moderate Disease
Classification of Disease Severity for Study Population
ICU and at least one of the following: Invasive mechanical ventilation Sepsis Shock (use of vasopressors) ECMO	ICU without any of the following: Invasive mechanical ventilation Sepsis Shock (use of vasopressors) ECMO	Admission (inpatient/observation) not meeting criteria for critical or severe disease
OR	OR
Death	Respiratory failure*
IDSA Guideline Recommendations for Select COVID-19 Medications as of December 2020 12
Recommend steroids Remdesivir less beneficial in patients on mechanical ventilation or ECMO Hydroxychloroquine not recommended Hydroxychloroquine + azithromycin not recommended Tocilizumab not recommended Anakinra (no guidelines) Ivermectin (no guidelines)	Suggests steroids Suggests remdesivir Hydroxychloroquine not recommended Hydroxychloroquine + azithromycin not recommended Tocilizumab not recommended Anakinra (no guidelines) Ivermectin (no guidelines)	No therapies recommended

^*Given lack of patient-level data on oxygen saturation or receipt of oxygen therapy, ICD-10 codes for respiratory failure (Appendix Table 1) were used as a surrogate measure of hypoxia requiring supplemental oxygen (SpO2 ≤ 94%) to categorize non-ICU patients as severe disease per IDSA guidelines (n = 48 patients).¹²

Abbreviations: COVID-19: coronavirus disease 2019; ECMO: Extracorporeal Membrane Oxygenation; ICD-10: International Classification of Diseases 10^th Revision; ICU: Intensive Care Unit; IDSA: Infectious Diseases Society of America

Therapy Guidelines by Time Period

Pre- and post-guideline periods were defined as April 2020 – November 2020 and December 2020 – December 2021, respectively, coinciding with the online publication of the IDSA guidelines endorsed by PIDS on December 2, 2020.¹² Although various guidelines were published and modified throughout the pandemic, by December 2020 there was general consensus around recommendations for the treatment of pediatric patients. As our goal was to assess guideline adherence, we purposefully chose the date that IDSA guidelines were published.

Statistical analysis

The distribution of patients overall and for each period was summarized by frequencies and percentages for categorical variables and median and interquartile range for continuous variables. Categories included age, sex, race and ethnicity, payer, COVID-19 disease severity, and medication received. Patient characteristics and medications used were compared between pre- and post-guidelines periods using chi-square tests and Wilcoxon Rank-Sum tests as appropriate. The proportion of patients receiving steroids, remdesivir, hydroxychloroquine, azithromycin, tocilizumab, anakinra, enoxaparin, ivermectin, and antibiotics were compared between pre- and post-guideline periods including stratification by COVID-19 illness severity. Monthly discharges by disease severity category were assessed over the study period, as well as monthly use of hydroxychloroquine and ivermectin. Interrupted time series were used to analyze monthly pre- and post-guideline use of guideline-recommended COVID-19 therapies (steroids and remdesivir), and enoxaparin, stratified by COVID-19 disease severity. SAS version 9.4 (SAS Institute, Inc) was utilized for all analyses. A p-value threshold of < 0.05 was considered statistically significant.

Results

Study Population

There were 18,364 patients included in the study, of whom 80.3% were discharged from their index hospitalization in the post-guideline period (Table 2). There was a higher percentage of Hispanic patients in the pre-guideline period versus the post-guideline period (42.6% vs. 26.2%, p < 0.001), and the opposite was observed for Non-Hispanic White patients (24.6% vs. 39.5%, p < 0.001). COVID-19 hospitalizations for each disease severity category fluctuated with COVID-19 incidence over time (Figure 1). The majority (64.2%) of the 134 patients with an ICD-10 code for respiratory failure were cared for in the intensive care unit (ICU), and this proportion was similar pre- and post-guideline. Only 48 non-ICU patients had an ICD-10 code for respiratory failure and were classified as severe disease.

Table 2.

Comparison of Demographics, COVID-19 Disease Severity, and Medications across Pre/Post-Guideline Periods of COVID-19 Pandemic

		Pre-guideline	Post-guideline
	All Inpatients	April 2020 - Nov 2020	Dec 2020 - Dec 2021	p-value
N	18364	3618 (19.7)	14746 (80.3)
Age in years
Median [IQR]	6 [0, 14]	6 [0, 14]	6 [0, 14]	0.056
0 – 4 years	8414 (45.8%)	1698 (46.9%)	6716 (45.5%)	0.004
5 – 11 years	3744 (20.4%)	666 (18.4%)	3078 (20.9%)
12 – 17 years	6206 (33.8%)	1254 (34.7%)	4952 (33.6%)
Sex
Female	8517 (46.4%)	1682 (46.5%)	6835 (46.4%)	0.881
Male	9847 (53.6%)	1936 (53.5%)	7911 (53.6%)
Race and Ethnicity
Non-Hispanic White	6716 (36.6%)	889 (24.6%)	5827 (39.5%)	< 0.001
Non-Hispanic Black	4551 (24.8%)	856 (23.7%)	3695 (25.1%)
Hispanic	5406 (29.4%)	1543 (42.6%)	3863 (26.2%)
Asian	451 (2.5%)	88 (2.4%)	363 (2.5%)
Other	1240 (6.8%)	242 (6.7%)	998 (6.8%)
Payor
Private	5343 (29.1%)	807 (22.3%)	4536 (30.8%)	< 0.001
Government	12101 (65.9%)	2673 (73.9%)	9428 (63.9%)
Other	920 (5%)	138 (3.8%)	782 (5.3%)
Disease Severity
Mild-Moderate	13496 (73.5%)	2606 (72.0%)	10890 (73.9%)	< 0.001
Severe (non-critical)	2269 (12.4%)	386 (10.7%)	1883 (12.8%)
Critical	2599 (14.2%)	626 (17.3%)	1973 (13.4%)
Medications
Steroids	7769 (42.3%)	1125 (31.1%)	6644 (45.1%)	< 0.001
Remdesivir	3750 (20.4%)	226 (6.2%)	3524 (23.9%)	< 0.001
Hydroxychloroquine	61 (0.3%)	37 (1.0%)	24 (0.2%)	< 0.001
Azithromycin	1131 (6.2%)	257 (7.1%)	874 (5.9%)	0.008
Tocilizumab	204 (1.1%)	19 (0.5%)	185 (1.3%)	< 0.001
Anakinra	391 (2.1%)	123 (3.4%)	268 (1.8%)	< 0.001
Enoxaparin	4620 (25.2%)	845 (23.4%)	3775 (25.6%)	0.005
Ivermectin	4 (0%)	0 (0%)	4 (0%)	0.322
Antibiotics*	8492 (46.2%)	1930 (53.3%)	6562 (44.5%)	< 0.001

^*Excluding azithromycin. Disease Severity: mild-moderate, did not require ICU or have respiratory failure ICD-10 code; severe (non-critical), respiratory failure ICD-10 code or in ICU but without invasive mechanical ventilation, ECMO, shock, or sepsis; critical, in ICU with invasive mechanical ventilation, ECMO, shock, sepsis, or died during admission.

Figure 1.

Volume of Inpatient Discharges Over Time for Children Hospitalized with COVID-19 Stratified by Disease Severity

The dotted line represents the number of children hospitalized with mild-moderate COVID-19 disease; dashed line, children with severe (non-critical) disease; solid line, children with critical disease.

Medication Use

The most common medications in the pre-guideline period included antibiotics (53.3%), steroids (31.1%), enoxaparin (23.4%), azithromycin (7.1%), and remdesivir (6.2%). Remdesivir and steroids significantly increased after the guidelines were published but remained used in a minority of hospitalized children. Hydroxychloroquine, azithromycin, anakinra, and antibiotics significantly decreased post-guideline. Ivermectin use was low post-guideline. There was a small but statistically significant increase in enoxaparin and tocilizumab post-guideline (Table 2).

Adherence to Treatment Guidelines

As the guideline recommendations are based on COVID-19 illness severity (e.g., steroids and remdesivir), Table 3 and Figure 2 can be used to illustrate adherence to treatment guidelines. Although the percentage of patients receiving steroids and remdesivir significantly increased within each disease severity category from the pre- to the post-guideline period (Table 3), the trend analysis suggests the slope was significantly lower post-guideline for mild-moderate and critical disease (Figure 2, Appendix Table 3). Overall, 75.4% of severe disease and 82.4% of critical disease patients received steroids in the post-guideline period, when steroid treatment was suggested or recommended, respectively (Tables 1 and 3).¹² In the post-guideline period, patients with severe disease received remdesivir more frequently (55.2%) than those with critical disease (41.4%), consistent with the recommendation for use in patients with severe disease who are not on mechanical ventilation or ECMO.¹² Remdesivir use both pre- and post-guideline was higher in children ≥ 12 years.

Table 3.

Medications by COVID-19 Disease Severity and Pre/Post-Guideline Periods

	Mild-Moderate			Severe (non-critical)			Critical
	Pre-Guideline	Post-Guideline	p-value	Pre-Guideline	Post-Guideline	p-value	Pre-Guideline	Post-Guideline	p-value
N	2606	10890		386	1883		626	1973
Steroids	474 (18.2%)	3599 (33.0%)	< 0.001	182 (47.2%)	1419 (75.4%)	< 0.001	469 (74.9%)	1626 (82.4%)	< 0.001
Remdesivir	62 (2.4%)	1669 (15.3%)	< 0.001	68 (17.6%)	1039 (55.2%)	< 0.001	96 (15.3%)	816 (41.4%)	< 0.001
Age < 12	12 (0.7%)	663 (8.7%)	< 0.001	23 (11.5%)	380 (39.5%)	< 0.001	38 (10.9%)	420 (35.1%)	< 0.001
Age ≥ 12	50 (6.3%)	1006 (30.9%)	< 0.001	45 (24.2%)	659 (71.6%)	< 0.001	58 (20.9%)	396 (51.1%)	< 0.001
Hydroxychloroquine	10 (0.4%)	19 (0.2%)	0.038	10 (2.6%)	1 (0.1%)	< 0.001	17 (2.7%)	4 (0.2%)	< 0.001
Azithromycin	146 (5.6%)	509 (4.7%)	0.048	57 (14.8%)	182 (9.7%)	0.003	54 (8.6%)	183 (9.3%)	0.623
Tocilizumab	0 (0)	32 (0.3%)	0.006	3 (0.8%)	67 (3.6%)	0.004	16 (2.6%)	86 (4.4%)	0.043
Anakinra	19 (0.7%)	30 (0.3%)	0.001	12 (3.1%)	33 (1.8%)	0.082	92 (14.7%)	205 (10.4%)	0.003
Enoxaparin	307 (11.8%)	1700 (15.6%)	< 0.001	167 (43.3%)	1020 (54.2%)	< 0.001	371 (59.3%)	1055 (53.5%)	0.011
Ivermectin	0 (0%)	0 (0%)	-	0 (0%)	0 (0%)	-	0 (0%)	4 (0.2%)	0.260
Antibiotics*	1149 (44.1%)	3989 (36.6%)	< 0.001	237 (61.4%)	974 (51.7%)	0.001	544 (86.9%)	1599 (81.0%)	0.001

^*Excluding azithromycin. Disease Severity: mild-moderate, did not require ICU or have respiratory failure ICD-10 code; severe (non-critical), respiratory failure ICD-10 code or in ICU but without invasive mechanical ventilation, ECMO, shock, or sepsis; critical, in ICU with invasive mechanical ventilation, ECMO, shock, sepsis, or died during admission.

Figure 2.

Trends in Medication Use Over Time for Children Hospitalized with COVID-19 Stratified by Disease Severity

Vertical line indicates IDSA Guideline Version 3.5.1 online publication December 2, 2020. Solid orange line represents the trend of medication use pre-guideline; dashed orange line, the post-guideline trend; blue line, the projected trend post-guideline if no change occurred with guideline publication; p-value, the change in slope pre- to post-guideline.

Across pre- to post-guideline periods, hydroxychloroquine significantly decreased across all disease severity categories, with a noticeably abrupt decline correlating with EUA revocation June 2020 (Table 3, Appendix Figure 1).⁹ Azithromycin use significantly decreased in patients with mild-moderate and severe disease (Table 3). In the post-guideline period, percent enoxaparin use significantly increased among mild-moderate and severe disease and decreased among critical disease (Table 3). The interrupted time series analysis suggests these trends began in the pre-guideline period for mild-moderate and critical disease but use in severe disease was declining pre-guideline and then significantly increased post-guideline (Figure 2, Appendix Table 3). Antibiotic use significantly declined across all disease severity categories post-guideline, although remained high in critical disease (81%) and severe disease (51.7%) (Table 3).

Discussion

In our retrospective cohort study of children with COVID-19 discharged from children’s hospitals, we found increased use of steroids and remdesivir and decreased use of hydroxychloroquine and azithromycin during the post-guideline period. Although changes in percent utilization pre- versus post-guideline appear consistent with IDSA guidelines, uptake in steroid and remdesivir use generally declined post-guideline and adherence to recommendations was incomplete. While guideline adherence for no use of hydroxychloroquine was close to 100%, steroids were used only in 82.4% of critically ill patients despite a strong recommendation.

Throughout the pandemic, clinicians and researchers rapidly developed and revised guidelines leading to changes in therapies for COVID-19 within children’s hospitals. Publications and webinars by various organizations led to early adaptation of therapies from adult experience. As experience grew several guidelines, some pediatric-specific, were published. Although we found compliance with COVID-19 guidelines was not 100%, the speed of uptake compared to other non-COVID-19 guidelines was extremely high.²⁰ It is frequently stated that it takes about 17 years for research evidence to reach clinical practice.²¹ Media attention, clinician webinars, use of social media, and gravity of COVID-19 illness might have contributed to rapid implementation and higher level of guideline compliance.

Recommended therapies such as remdesivir and steroids increased throughout the pandemic, although use was not universal in severe and critical illness. Steroids represent a commonly available and familiar medication and were consistently suggested or recommended for children with severe and critical disease, respectively. Post-guideline, only steroid use in severe disease continued to increase following the pre-guideline trend. The trend in steroid use for mild disease decreased consistent with guideline recommendations. However, the post-guideline trend in steroid use for patients with critical disease became static despite recommendations, leading to incomplete adoption. Despite their availability, adherence to this guideline-recommended therapy remains around 80% for children with critical disease and less for children with severe non-critical disease. However, both rates exceed the use of remdesivir in severe non-critical patients (55.2%), a population in which remdesivir is suggested throughout guideline revisions.

Trends in remdesivir use decreased post-guideline for all disease categories. This was consistent with the recommendation for no therapy in patients with mild-moderate disease and decreased benefit for patients with critical disease on mechanical ventilation or ECMO, but less consistent with suggested use in severe illness. Barriers to remdesivir use in the severely ill population may include perceived lack of efficacy on meaningful outcomes (e.g., recovery versus mortality) or FDA approval in October 2020 only for hospitalized patients ≥ 12 years and weighing ≥ 40 kg, while use for younger children weighing ≥ 3.5 kg remained under an EUA during the study period.⁷ In our study, children ≥ 12 years had higher percentage of remdesivir use compared to younger children, although use significantly increased for both age groups pre- to post-guideline.

Finally, the post-guideline use of steroids (33%) or remdesivir (15.3%) for mild-moderate illness warrants further exploration, given the suggestion against use in this population. The use of tocilizumab and anakinra was not common, reflecting the lack of clear recommendations for anakinra and recommendation against routine use of tocilizumab in the December 2020 IDSA Guidelines (Table 1). The recommendation against tocilizumab was revised in February 2021 to a suggestion for use in patients with progressive severe or critical disease with elevated markers of systemic inflammation in addition to standard of care (i.e., steroids), and further mention of anakinra was dropped.^12,22 While it is difficult to assess guideline adherence, we note that use of both these medications was higher in critically ill patients.

Early studies suggested high incidence of thrombosis in children and adults with potential benefits to anticoagulation.^23,24,25 International Society of Thrombosis and Hemostasis (ISTH) interim guidelines recommended D-dimer screening and anticoagulation in adults with COVID-19 which correlated to decreased mortality.^5,26 Although there were no pediatric-specific guidelines, many children’s hospitals routinely used anticoagulation in hospitalized patients based on adult studies.²⁷ Subsequently, ISTH pediatric consensus guidelines and COVID-19 Pediatric Intensive Care Unit (PICU) guidelines advocated for anticoagulation in children with COVID-19 with an elevated D-dimer (five times above normal) or other coagulopathy risk factors.^27,28 Early in the pandemic, observational studies reported heparin use in 18.1% and enoxaparin use in 14.8% of hospitalized children with COVID-19.²⁹ Another single-center study found 65% of PICU patients and 16% of all hospitalized children with COVID-19 received anticoagulation during the early period of the pandemic.³⁰ However, a systematic review of the literature published through January of 2021 concluded that thrombotic and thromboembolic events are rare in children with COVID-19 or MIS-C.³¹ In December 2020, updated National Institutes of Health (NIH) guidelines stated that for hospitalized children with COVID-19, indications for venous thromboembolism prophylaxis should be the same as those for children without COVID-19.³² In our current study, 59.3% of children with critical illness received enoxaparin pre-guideline and its use significantly decreased to 53.5% post-guideline; however, the overall change in trend was not significant, possibly due to the rise in use during fall 2021. Enoxaparin use in mild-moderate and severely ill patients increased significantly pre- to post-guideline, with a significant positive upward trend post-guideline in severe patients. Overall, it appears that anticoagulation is still common in children with COVID-19 despite current recommendations, acknowledging the difficulty in deimplenting low-value care.³³

The interest in hydroxychloroquine for COVID-19 treatment emerged early in the pandemic with publication of an open-label nonrandomized study.³⁴ After this publication, the FDA issued a letter on March 28, 2020, granting EUA of chloroquine and hydroxychloroquine for patients with COVID-19.⁸ The use of hydroxychloroquine fell appropriately to < 1% after EUA revocation.⁹ Azithromycin use for COVID-19 therapy in combination with hydroxychloroquine also declined after IDSA and NIH COVID-19 Treatment Guideline publication.^35,36 Finally, ivermectin was not mentioned in December 2020 and use outside a clinical trial was not recommended in the 2021 IDSA guidelines.^22,37 The results of initial clinical trials were reported in late 2020, and meta-analyses soon afterwards suggesting benefit were later retracted.³⁸ The infrequent use of ivermectin in our study was heartening, suggesting that guidelines continued to be followed in pediatric patients despite media attention.

The decrease in antibiotic use over time across disease severity categories may indicate more confidence among clinicians as time progressed with the management of COVID-19 as a viral illness, both in the spectrum of illness and radiographic appearance of COVID-19 pneumonia. In our study, we could not determine whether there was presumed or confirmed secondary bacterial infection that warranted antibiotics or if antibiotics were used initially in critically ill patients because of a concern for sepsis. However, we note that antibiotics were commonly used in this cohort of children hospitalized for COVID-19.

There are several limitations that may influence study interpretation. Guideline implementation during a pandemic and may not be representative of other situations due to rapid guideline development with limited data over time. Furthermore, the PHIS database predominantly includes tertiary and quaternary care children’s hospitals, where guideline uptake may differ from community hospitals or smaller centers. Additionally, our primary analysis estimated guideline adherence as a percentage for each study period, but this pre- and post-design may not account for within-period trends, which are visually displayed for key medications in our figures. Patients may have been misclassified by utilizing ICD-10 diagnosis codes as a proxy for disease severity. For example, it is possible children who were asymptomatic or required supplemental oxygen in a non-critical setting without the ICD-10 code for respiratory failure were miscategorized as mild-moderate disease. Additionally, PHIS is a billing database and lacks patient-level clinical data such as duration of symptoms, oxygen saturation or receipt of oxygen therapy, and laboratory test results. Data on timing of disease onset was not available for our study but may have affected choice of therapy since remdesivir is thought to be more effective in the early phase of symptomatic disease.³⁹ Furthermore, we were not able to determine therapy-specific indications for non-COVID-19 treatments, such as steroids for asthma. Finally, the PHIS database does not reliably include experimental medications for some uncommonly used therapies (e.g., convalescent plasma) and these were not included in our study. Despite these limitations, our study reports important trends in adherence to national treatment guidelines in children hospitalized with COVID-19 and identifies potential targets for care improvement.

Conclusion

COVID-19 medications for hospitalized children evolved throughout the pandemic as clinicians confronted a novel pathogen with limited therapeutic data in children. Despite these challenges, guidelines rapidly emerged, and clinicians increasingly adhered to them. However, approximately 1 in 5 critically ill patients did not receive steroid treatment in our post-guideline period despite a strong recommendation. As the COVID-19 pandemic continues to impact the world and increasingly our pediatric patients, consideration should be given to approaches aimed at more reliable delivery of guideline-directed therapy to children, to best use our collective knowledge to improve outcomes.

What’s Known on This Subject

The impact of the COVID-19 pandemic on children has been substantial. Guidelines for the management of COVID-19 in hospitalized children changed rapidly as new evidence emerged. The December 2020 Infectious Disease Society of America (IDSA) guidelines consolidated key treatment recommendations.

What This Study Adds

This study describes the patterns of medication utilization for children hospitalized with COVID-19 during 2020–2021, as well as adherence to the December 2020 IDSA treatment guidelines. We note guideline adoption successes and areas for improvement.

Abbreviations:

CI

confidence interval

COVID-19

coronavirus disease 2019

CPT

Current Procedural Terminology

ECMO

Extracorporeal Membrane Oxygenation

EUA

Emergency Use Authorization

FDA

Food and Drug Administration

ICD-10

International Classification of Diseases 10^th Revision

ICU

Intensive Care Unit

IDSA

Infectious Disease Society of America

IRB

Institutional Review Board

ISTH

International Society of Thrombosis and Hemostasis

MIS-C

Multisystem Inflammatory Syndrome in Children

PHIS

Pediatric Health Information System

PICU

Pediatric Intensive Care Unit

PIDS

Pediatric Infectious Diseases Society

US

United States

Appendix Table 1.

ICD-10 Codes for Complications Known to Occur Concurrently with Acute COVID-19¹⁴

Diagnosis	ICD-10 Code
Acute kidney injury	N17.9, N17.0, N17.1, N17.2, N17.8, S37.0, S37.091, S37.092A, S37.099A, S37.00
Acute liver injury	R74.01, R94.5, K72, K72.90, K72.91, K72.00, K72.01
Aseptic meningitis	A87, A87.8, A87.9, G03.8, G03.9, G03.0, G03.9, G03
Asthma	J45.21, J45.22, J45.31, J45.32, J45.41, J45.42, J45.51, J45.52, J45.901, J45.902, Exclude patients < 2 year of age
Bronchiolitis	J21.0, J21.1, J21.8, J21.9: exclude patient > 2 years
Cerebral infarction	I63, I630, I631, I632, I633, I634, I635, I636, I638, I639
Croup	J04.0, J04.1, J04.2, J04.30, J04.31, J05.0, J05.10, J05.11, J06.0
Encephalitis	A85, A85.8, A86, B941, G04, G04.0, G04.8, G04.9, G05, G05.3
Encephalopathy	G04.3, G04.30, G93.4, G93.40, G93.49, J10.81, J11.81
Febrile infant	Age < 90 days AND one of the following codes: R50, P81, R50.9, R50.81
Febrile seizure	R56.0, R56.00, R56.01, R56.9
Gastroenteritis/dehydration	A08.4, A08.8, A08.0, E86.0
Myocarditis/pericarditis/heart failure	I51.4, I40.1, I40, I41, I40.0, I40.8, I40.9, I51.4, B33.22, I30, I30.9, I30.8, I30.0, I30.1, B33.23
Nonfebrile seizure	G40.1, G40.20, G40.21, G40.30, G40.31, G40.4, G40.5, G40.6, G40.7, G40.8, G40.9, R56.8, R56.9
Pancreatitis	K85, K86.1
Pneumonia, lung effusion, or empyema	J12.0, J12.1, J12.2, J12.3, J12.81, J12.89, J12.9, J13, J14, J15.0, J15.1, J15.20, J15.211, J15.212, J15.29, J15.3, J15.4, J15.5, J15.6, J15.7, J15.8, J15.9, J16.0, J16.8, J17, J18.0, J18.1, J18.2, J18.8, J18.9, J69.0, J69.1, J69.8, B34.9, B34, J81.0
Respiratory failure*	J96, P28.5, J96.90, J96.91, J96.9, J96.00, J96.01, J96.02, J96.0, R09.2, J96.2, J80
Sepsis*	R65.20, R65.21, A41.9, A41.89, A41, I95.9, I95.8, I95
Sickle cell complications	D57.01, D57.211, D57.411, D57.811, D57.02, D57.212, D57.412, D57.812, D57.41, D57.81
Thrombotic complications	I82.49, I82.59, I26, I60, I61, I62
Unspecified Influenza-like Illness	B97.89, H66.90, H66.91, H66.92, H66.93, J00, J01.90, J06.9, J20.9, J40, R05, R50

Abbreviations: COVID-19: coronavirus disease 2019; ICD-10: International Classification of Diseases, 10^th Revision

^*These ICD-10 codes were also used for disease severity classification.

Appendix Table 2.

Study Inclusion and Exclusion Criteria

Inclusion Criteria - Status: Inpatient/Observation (included ICU admission) - Age: 30 days – less than 18 years old - Primary diagnosis of COVID-19 (ICD-10: U.071) with any secondary diagnosis - Secondary diagnosis of COVID-19 with primary diagnosis of a COVID-19 complication

Exclusion Criteria - Newborn who never left the Hospital - Diagnosis of Kawasaki Disease or MIS-C (ICD-10: M30.3, M35.8)

Abbreviations: COVID-19: coronavirus disease 2019; ICD-10: International Classification of Diseases, 10^th Revision; ICU: Intensive Care Unit; MIS-C: Multisystem Inflammatory Disease in Children

Appendix Table 3.

Trends in Medication Use Over Time for Children Hospitalized with COVID-19 Stratified by Disease Severity, Interrupted Time Series Parameter Estimates on Logit Scale (Corresponding to Figure 2)

		Mild-Moderate		Severe (non-critical)		Critical
		Estimate (95% CI)	p-value	Estimate (95% CI)	p-value	Estimate (95% CI)	p-value
Enoxaparin
	Pre-guideline slope	0.02 (−0.04, 0.08)	0.448	−0.08 (−0.18, 0.02)	0.124	−0.03 (−0.12, 0.06)	0.505
	Level change immediately after guideline	−0.2 (−0.44, 0.04)	0.108	−0.28 (−0.71, 0.15)	0.208	0.17 (−0.19, 0.54)	0.355
	Change in slope	−0.02 (−0.08, 0.04)	0.577	0.13 (0.02, 0.23)	0.017	0.04 (−0.06, 0.13)	0.440
	Post-guideline slope	0.01 (−0.01, 0.02)	0.466	0.05 (0.02, 0.08)	< 0.001	0.01 (−0.02, 0.03)	0.610
Remdesivir
	Pre-guideline slope	0.39 (0.23, 0.54)	< 0.001	0.22 (0.08, 0.36)	0.002	0.33 (0.2, 0.46)	< 0.001
	Level change immediately after guideline	−0.65 (−1.01, −0.3)	0.001	−0.47 (−0.95, 0)	0.058	0.45 (0.02, 0.87)	0.046
	Change in slope	−0.33 (−0.48, −0.17)	< 0.001	−0.13 (−0.27, 0.01)	0.063	−0.21 (−0.34, −0.08)	0.002
	Post-guideline slope	0.06 (0.04, 0.07)	< 0.001	0.09 (0.06, 0.11)	< 0.001	0.12 (0.1, 0.15)	< 0.001
Steroids
	Pre-guideline slope	0.15 (0.1, 0.2)	< 0.001	0.13 (0.03, 0.24)	0.009	0.23 (0.12, 0.33)	< 0.001
	Level change immediately after guideline	−0.09 (−0.28, 0.09)	0.340	−0.18 (−0.61, 0.26)	0.431	0.41 (−0.05, 0.88)	0.086
	Change in slope	−0.12 (−0.17, −0.06)	< 0.001	−0.05 (−0.15, 0.06)	0.394	−0.21 (−0.31, −0.1)	< 0.001
	Post-guideline slope	0.04 (0.02, 0.05)	< 0.001	0.09 (0.06, 0.12)	< 0.001	0.02 (−0.01, 0.05)	0.242

Abbreviations: CI: Confidence Interval

Event point is IDSA Guideline Version 3.5.1 online publication December 2, 2020. Slope is trend in medication use. Level change immediately after guideline represents change in series at event point (end of pre-guideline versus start of post-guideline period); change in slope is difference in slopes between pre- and post-guideline periods.

Appendix Figure 1.

Hydroxychloroquine and Ivermectin Use Over Time in Children Hospitalized with COVID-19 Stratified by Disease Severity

The dotted line represents the number of children hospitalized with mild-moderate COVID-19 disease; dashed line, children with severe (non-critical) disease; solid line, children with critical disease.