Emergency Medical Services Administration of Systemic Corticosteroids for Pediatric Asthma: A Statewide Study of Emergency Department Outcomes
For children with an asthma exacerbation, systemic corticosteroids (CS) administered in the emergency department (ED) decrease hospital admission rates and ED length of stay (LOS).1-4 Time-dependent effects favor earlier CS administration, ideally within the first hour of ED arrival.2, 3 Only one pediatric study has examined if earlier emergency medical services (EMS) administration of CS improves patient outcomes.5 That study found decreased hospital admission rates and ED LOS after adding dexamethasone to intravenous (IV) methylprednisolone as options for pediatric asthma. However, the study was of a single EMS agency and overall EMS CS use was low (<20%).5 To date, no large study has examined pediatric asthma outcomes after EMS treatment. This study's objective was to examine a statewide population of pediatric asthma patients to determine the effects of EMS administration of CS on ED outcomes.
This was a retrospective observational cohort study of pediatric asthma patients treated by EMS from 2011 to 2016. Patients were identified from Florida's EMS Tracking and Reporting System (EMSTARS) database. EMSTARS contains information on ED outcomes for patients deterministically linked to the Agency for Healthcare Administration (AHCA) database of hospital and ED information. Both the study institution and Florida DOH Institutional Review Boards approved the study (IRB201702645 and Protocol 180000U11, respectively). This study was funded by the National Center for Advancing Translational Sciences of the National Institutes of Health under awards KL2TR001429 and UL1TR001427.
Inclusion criteria were patients 2 to 18 years of age, transported by EMS to an ED, with an EMS provider primary impression of respiratory distress, and who were administered albuterol at least once to indicate an acute asthma exacerbation. Patients were excluded if they were classified as a trauma/injury, seizure, pregnancy-related complication, or interfacility transport. Patients less than 2 years of age were excluded to avoid confounding with wheezing from bronchiolitis. Patients who were not successfully linked with the AHCA database were excluded.
Emergency department discharges to home, law enforcement custody, or a psychiatric facility were classified as discharges. Admissions to the same facility and transfers to another facility (ED or inpatient) were classified as admissions. We chose to classify transfers to another ED (even if subsequently discharged) as admissions because most pediatric interfacility transports are for inpatient admission.6 ED LOS was documented by hour, not minute, which precluded further LOS analysis.
Demographic, clinical, and EMS variables were analyzed using descriptive statistics. Continuous variables were compared using Student's t-test and Kruskal-Wallis tests, as appropriate. Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate. Missing data were excluded.
To determine EMS administration of CS's association with ED discharge, we used candidate variables identified by univariate analysis to create a parsimonious logistic regression model using forward selection. Propensity scoring with a nearest-neighbor 1:1 matching algorithm was also used to compare more homogenous case (received CS) and control (did not receive CS) pairs. Statistical analysis was performed using SAS version 9.4.
Of the 388,187 patients ages 2 to 18 years in the EMSTARS database from 2011 to 2016, 11,667 met the study's definition of an asthma exacerbation. Of those patients, 3,812 had known ED outcomes. During the EMS encounter, in addition to albuterol, 37.7% of patients received nebulized ipratropium bromide and 9.7% of patients received CS in the form of IV methylprednisolone (only nine patients received dexamethasone; therefore, dexamethasone was excluded from further analyses). Most patients (75.3%) were discharged from the ED. Significant differences between admitted and discharged patients included age, race, initial prehospital heart rate, respiratory rate, pulse oximetry, respiratory effort, level of alertness, and EMS administration of magnesium sulfate or epinephrine (all p < 0.001).
In the final adjusted model, receiving CS from EMS was associated with decreased odds of ED discharge (odds ratio [OR] = 0.7, 95% confidence interval [CI] = 0.5–0.9). Other significant variables associated with ED discharge are detailed in Table 1.
| Variable | OR | 95% CI |
|---|---|---|
| Heart rate | 0.98 | 0.97–0.98 |
| Respiratory rate | 0.98 | 0.96–0.97 |
| Pulse oximetry | 1.04 | 1.02–1.06 |
| Glasgow Coma Scale score | 1.27 | 1.17–1.37 |
| EMS scene location | ||
| Healthcare facility | 0.29 | 0.21–0.39 |
| Public building | 1.84 | 1.20–2.82 |
| Business | 1.69 | 1.12–2.54 |
| Lights and sirens | 0.71 | 0.59–0.85 |
| EMS destination decision: physician choice | 0.27 | 0.13–0.59 |
| Labored respiratory effort | 0.80 | 0.66–0.98 |
| Fatigued respiratory effort | 0.51 | 0.30–0.85 |
| Systemic corticosteroid administration | 0.71 | 0.54–0.95 |
| Magnesium sulfate administration | 0.30 | 0.17–0.53 |
| EMS procedure: suction | 0.24 | 0.07–0.89 |
- Multivariable logistic regression model: Akaike Information Criterion = 3131.0, area under the curve = 0.77, Hosmer and Lemeshow goodness-of-fit test = 0.82.
The propensity score was comprised of variables significantly associated with receiving CS from EMS: scene and transport time; systolic and diastolic blood pressure; sex; EMS administration of ipratropium bromide, oxygen, magnesium sulfate, normal saline, and subcutaneous epinephrine; IV access; electrocardiogram; and “adult assessment” (treating the patient as an adult rather than a child). The propensity score (reflecting the probability of receiving CS) was associated with significantly decreased odds of ED discharge (OR = 0.3, 95% CI = 0.2–0.6). Propensity scoring produced 266 matched pairs. Among those pairs, EMS administration of CS did not have a significant effect on ED discharge (chi-square p = 0.25; logistic regression OR = 0.8, 95% CI = 0.6–1.2).
To date, this is the largest study linking EMS administration of CS with pediatric asthma patients’ ED outcomes. This study found that EMS administration of CS was associated with decreased odds of ED discharge. Several potential reasons for this result (which conflicts with ED studies) are revealed by the data. First, few patients received CS from EMS (<10%), in contrast to ED studies where the vast majority of patients were given CS.1-3 In this study CS were reserved for more severe patients, as evidenced by the propensity score's association with significantly decreased odds of ED discharge and its composite variables. Additionally, a prior study of ED administration of CS showed more pronounced benefits to patients with mild and moderate exacerbations.3 Our results may reflect a population of severe patients who were likely to be admitted to the hospital regardless of EMS interventions.
An important question raised is why were CS reserved for only severe patients? The vast majority of CS were administered in the form of methylprednisolone, with only nine administrations of dexamethasone. Methylprednisolone administration requires IV placement, which may not be otherwise necessary for a mild or moderate asthma patient. Additionally, studies of pediatric IV placement by EMS show low percentages of attempts.7 Yet based on publicly available EMS protocols, IV methylprednisolone is the most common form of CS authorized.8 Therefore there is a practical mismatch between ED research and guidelines incorporating oral CS and EMS protocols authorizing only IV CS.
The latest 2018 version of the National Association of State EMS Officials model clinical guidelines makes a first step in correcting that mismatch by listing oral dexamethasone after IV methylprednisolone as a CS option.9 That guideline also states that other CS formulations at equivalent dosing may be used.9 However, the NASEMSO guideline is not pediatric-specific and does not proscribe interventions by patient severity or EMS transport time.9 Only a prospective, pediatric-specific study can answer the those questions and ascertain if earlier, EMS-administered oral CS improve patient outcomes.
This is a retrospective study of one state and so its conclusions may not be generalizable to other areas of the United States. Only one-third of patients had known ED outcomes. We were unable to ascertain any medications patients received prior to EMS arrival or in the ED. It is possible that patients with asthma did not receive albuterol and were excluded; however, >99% of all pediatric respiratory distress encounters in EMSTARS were serviced by paramedics, who are licensed to give albuterol, suggesting respiratory distress from another source (e.g., pneumonia). Compared with previous EMS research, most variables in EMSTARS had few missing values.10
In the largest study to date linking emergency medical services interventions to ED outcomes for pediatric asthma, corticosteroids administered by emergency medical services did not significantly affect ED discharge rates. Corticosteroids were administered to the most severe patients, perhaps due to emergency medical services protocols authorizing only intravenous methylprednisolone. To extend evidence-based pediatric asthma care beyond the ED, a prospective trial of oral corticosteroids administered by emergency medical services is warranted.
The study investigators acknowledge Steve McCoy; Brenda Clotfelter; Karen Card, DrPH; and Joshua Sturms from the Florida Department of Health's Bureau of Emergency Medical Oversight for their assistance and data management. The study investigators also acknowledge Colleen Kalynych, Michelle Lott, and Justin Masud from the Department of Emergency Medicine, Division of Research, University of Florida–Jacksonville, and Alexis Thomas and Paul Zwick, from the University of Florida GeoPlan Center, for their assistance with this study.
