STUDY OVERSIGHT

This investigation was reviewed in accordance with CDC procedures for protection of human research participants and was considered nonresearch public health surveillance activity intended for use in disease control or to inform policy during an emergency response. Details regarding the author contributions to the study are provided in the Supplementary Appendix, available with the full text of this article at NEJM.org.

CASE DEFINITIONS

State health departments, a Council of State and Territorial Epidemiologists Task Force, and the CDC developed and disseminated EVALI surveillance case definitions and data-collection tools beginning in August 2019.^7,19,20 For this analysis, a fatal case of EVALI was defined as a confirmed or probable¹² case of EVALI (as determined with the use of the primary or out-of-hospital case definitions shown in Tables S1 and S2 in the Supplementary Appendix) in a patient who died as of January 4, 2020, and whose death was reported to the CDC as of January 7, 2020. A nonfatal case of EVALI was defined as a confirmed or probable case of EVALI in a hospitalized patient who had not died and whose case was reported to the CDC as of January 7, 2020.

DATA COLLECTION

Health departments from 50 states, the District of Columbia, and two U.S. territories (Puerto Rico and the Virgin Islands) voluntarily reported cases of EVALI in hospitalized patients to the CDC and, when available, included data from medical-record abstractions and patient interviews. Proxies were interviewed if patients were too ill or had died. The CDC received data on patients hospitalized for EVALI through a standardized national case-report form. Medical records were the primary data source for the variables included in the standardized form for both fatal and nonfatal cases; patient interviews were used for nonfatal cases and proxy interviews for fatal cases when data from the medical-record abstraction were missing. Individual state health authorities determined the methods and procedures used for data collection (such as those to be used for the incorporation of data from proxy interviews), which resulted in variations in data completeness. Data on fatal cases were supplemented with data abstracted from medical-record review, which included additional variables such as body-mass index and initial vital signs and laboratory results that were not available for nonfatal cases. Imaging reports were reviewed by two investigators.

ANALYSES

The results of the descriptive analyses are presented as proportions for categorical variables and medians and ranges for continuous variables. Differences are not reported with P values or confidence intervals because this study describes the entire population of patients with fatal or nonfatal cases of EVALI reported as of January 7, 2020. Analyses were completed with the use of SAS software, version 9.4 (SAS Institute).

RESULTS OF THE NATIONAL STUDY

Table 1.Case Status, Demographic Characteristics, and Substance Use History of Patients with Fatal or Nonfatal Cases of EVALI.

As of January 7, 2020, a total of 2558 nonfatal cases of EVALI in hospitalized patients from 50 states, the District of Columbia, Puerto Rico, and the U.S. Virgin Islands and 60 fatal cases in patients from 27 states and the District of Columbia were reported to the CDC. The demographic characteristics and substance use histories of these patients are provided in Table 1. Among the patients with available data on sex, 32 of 60 (53%) of those with fatal cases and 1666 of 2498 (67%) of those with nonfatal cases were male. The median age was 51 years (range, 15 to 75) among patients with fatal cases and 24 years (range, 13 to 85) among those with nonfatal cases. Among the patients with available data on race and ethnic group, 39 of 49 (80%) of those with fatal cases and 1104 of 1818 (61%) of those with nonfatal cases were non-Hispanic white.

Among the patients with available data on e-cigarette, or vaping, product use (45 patients with fatal cases and 2021 with nonfatal cases), 30 (67%) and 1593 (79%), respectively, reported nonexclusive use of THC-containing products; 25 (56%) and 1109 (55%), respectively, reported nonexclusive use of nicotine-containing products; 12 (27%) and 803 (40%), respectively, reported using both THC-containing and nicotine-containing products; and 2 (4%) and 122 (6%), respectively, reported using neither (Table 1). Among the patients with available data on combustible tobacco use, 44 of 56 (79%) of those with fatal cases and 439 of 1327 (33%) of those with nonfatal cases were current or former smokers.

Table 2.Medical and Mental Health Conditions in Fatal or Nonfatal Cases of EVALI.

The medical histories of the patients with fatal or nonfatal cases of EVALI who had available data are provided in Table 2. The patients with fatal cases were more likely than those with nonfatal cases to have a history of any respiratory disease (25 of 57 [44%] vs. 371 of 1429 [26%]), cardiac disease (26 of 55 [47%] vs. 115 of 1169 [10%]), and any mental health condition (32 of 49 [65%] vs. 575 of 1398 [41%]). A total of 26 of 50 patients (52%) with fatal cases had obesity.

Table 3.Clinical Course, Initial Vital Signs, and Initial Laboratory Results in Hospitalized Patients with Fatal or Nonfatal Cases of EVALI.

The clinical characteristics of the hospitalized patients with fatal or nonfatal cases of EVALI who had available data are provided in Table 3 (full details are provided in Table S3). Approximately half the hospitalized patients with fatal cases (25 of 54 [46%]) and 479 of 2320 (21%) of those with nonfatal cases had presented initially in an outpatient setting (e.g., primary care physician location, urgent care center, or emergency department) without being admitted for EVALI-related symptoms. Among the hospitalized patients with fatal cases, 40 of 55 (73%) had hypoxia, 25 of 54 (46%) had tachycardia, and 26 of 52 (50%) had tachypnea at the time of admission to the hospital. The laboratory values at the initial admission showed that 37 of 52 patients (71%) with fatal cases had leukocytosis (white-cell counts >11,000 per cubic millimeter) and 29 of 45 (64%) had neutrophil predominance (white-cell count with >80% neutrophils). Most of the patients with fatal cases (93%) died in the hospital. Radiologic findings are provided in Table S4.

CASE REPORTS OF THREE PATIENTS WHO DIED FROM EVALI

Case 1: Multiple Preexisting Medical Conditions

A patient with coexisting asthma, chronic obstructive pulmonary disease, morbid obesity, preexisting heart failure, and reported use of THC-, cannabidiol-, and nicotine-containing e-cigarette, or vaping, products was admitted to the hospital 8 days after the onset of symptoms because of severe respiratory distress. Treatment with broad-spectrum antibiotics, diuretics, bronchodilators, systemic glucocorticoids, and bilevel positive airway pressure was initiated. Within 48 hours, acute respiratory distress syndrome developed, and the patient underwent intubation and mechanical ventilation. Evaluation and treatment decisions were complicated by respiratory failure and multiple chronic conditions. The provision of adequate oxygenation and ventilation remained difficult even though the patient was in a prone position. The patient died on hospital day 21. The full case report is provided in the Supplementary Appendix.

Case 2: Recrudescence of Symptoms

A patient visited two different health facilities 2 days after the onset of illness and was discharged home with a prescription for antibiotics and antiemetics for presumed community-acquired pneumonia. Three days later, the patient visited a different health facility and was admitted; treatments with bronchodilators and broad-spectrum antibiotics were initiated. The patient reported using nicotine-containing, but not THC-containing, e-cigarette, or vaping, products. The patient had increasing respiratory distress and fever (to a temperature of 38.9°C), which led to the initiation of systemic glucocorticoids on hospital day 3. On hospital day 5, the patient was discharged home with prescriptions for levofloxacin and a glucocorticoid on a tapering schedule.

Two days after discharge, the patient’s family reported that nausea, vomiting, and malaise recurred, and they found the patient to be unresponsive. Despite resuscitation efforts and readmission, the patient died a few hours later. According to the family report, the patient had used THC-containing products (Dank Vapes and Maui Wowi). Residue on the product cartridges tested positive for THC and vitamin E acetate. The full case report is provided in the Supplementary Appendix.

Case 3: No Glucocorticoids during Initial Hospitalization

A patient visited an outpatient facility the day after the onset of symptoms and was treated for viral gastroenteritis and sent home. This patient presented again with headache, nausea, vomiting, and diarrhea on day 6 of illness and was admitted to the hospital for dehydration. A chest radiograph on hospital day 4 showed prominent bilateral interstitial markings. The patient reported using nicotine-containing e-cigarette, or vaping, products, received treatment with azithromycin, and was transferred to the intensive care unit because of respiratory distress.

On hospital day 15, the patient was discharged home with instructions to abstain from using such products. During an interview with public health department staff, the patient reported having used the THC-containing product Dank Vapes. On day 24, because of continued respiratory compromise, the patient was readmitted to the hospital to undergo intubation and mechanical ventilation. Treatment with systemic glucocorticoids was initiated. On day 40, the patient underwent extracorporeal membrane oxygenation. On day 41, the patient had severe hypotension and cardiac arrest and died. The products submitted for testing after death included a THC-containing product, which was found to contain castor oil (no vitamin E acetate). The full case report is provided in the Supplementary Appendix.

This study assesses the clinical characteristics of hospitalized patients with fatal or nonfatal cases of EVALI. Consistent with the findings in previous reports,^12,13,15,17 most of the patients with fatal cases reported using THC-containing e-cigarette, or vaping, products. Nearly one third of the patients with fatal cases reported nicotine use only, and almost one quarter reported dual use, which suggests that other causal mechanisms may need to be considered. The median age was higher among the patients with fatal cases than among those with nonfatal cases, and the proportion of patients who had chronic medical conditions or were current or former combustible tobacco smokers was higher among those with fatal cases. Some patients with fatal cases had initially reported gastrointestinal symptoms (e.g., diarrhea and nausea), although such symptoms were reported more frequently among patients with nonfatal cases. Clinicians should consider EVALI in patients who present primarily with constitutional symptoms, gastrointestinal symptoms, or both.²¹

More than half the patients with fatal cases (63%) received a diagnosis of acute respiratory distress syndrome, which can lead to life-threatening acute hypoxemic respiratory failure.²²Patients with EVALI who have multiple chronic conditions may be at higher risk for acute respiratory distress syndrome and its complications. Because traditional mechanical ventilation can worsen lung injury in patients with acute lung injury or acute respiratory distress syndrome, which would then increase the risk of nonpulmonary organ or system failure,²³ clinicians should consider evidence-based principles regarding the management of acute respiratory distress syndrome when treating patients with severe cases of EVALI.²⁴

A higher proportion of hospitalized patients with fatal cases than those with nonfatal cases had cardiac disease, asthma, or chronic obstructive pulmonary disease. Cardiac or respiratory conditions in patients who use e-cigarette, or vaping, products may complicate assessment, radiologic interpretations, and management, particularly if the exposure history is incomplete. Patients may withhold medically relevant information because they fear being judged negatively by their clinical provider, they do not want to hear how harmful their behavior was, or they are embarrassed.²⁵ Cases 2 and 3 show how incomplete exposure history may delay treatment. Illegality of THC use can contribute to an incomplete exposure history, which reinforces the need for empathetic, nonjudgmental communication between the clinician and patient (or proxy) in a confidential setting.

Two thirds of patients with fatal cases (65%) had a known mental health condition — a higher percentage than that among patients with nonfatal cases (41%) and the U.S. general population (19%).²⁶ Patients, particularly former smokers, with mental health conditions are more likely to report lifetime and current e-cigarette, or vaping, product use.²⁷ The prevalence of mental health conditions is higher among persons reporting concurrent use of cannabis and conventional cigarettes.²⁸ Addressing underlying mental health conditions among patients with EVALI is an important component of hospital care, discharge planning, and postdischarge follow-up care.²⁹

In the current study, the percentage of patients who had obesity was higher among those with fatal cases (52%) than in the general population (40%). Data on body weight were not available for the patients with nonfatal cases. Obesity, a major risk factor for respiratory diseases,^30-33can lead to changes in pulmonary physiologic variables, complicate respiratory disease, and present challenges in achieving adequate mechanical ventilation.^30-32,34 More studies are warranted to determine if there is an association between obesity and development of EVALI and whether obesity contributes to the severity of EVALI outcomes.

In outpatient settings, when symptoms are attributable to infections, glucocorticoid use is cautioned⁹; limited research suggests that glucocorticoids may be beneficial in the treatment of EVALI.^8,15,29,35 Glucocorticoid administration may be warranted in severely ill patients when it is considered safe and feasible, after ruling out any life-threatening infections, such as influenza. Among the patients with fatal cases who had available data, the time from the onset of symptoms to glucocorticoid administration varied widely, and coadministration with antibiotics was frequently observed. Few patients with fatal cases had reported improvement with glucocorticoids; administering glucocorticoids too late in the clinical course may partially explain this finding. Glucocorticoid administration has been attributed to rapid clinical improvement, including improvement in oxygenation.^21,36 Glucocorticoid administration is recommended if there is no improvement with antibiotics, respiratory support, or both, and the dosing and duration should be determined on a case-by-case basis.³⁶ Current interim clinical guidance from the CDC provides more details on glucocorticoid administration.²⁹

The data in our study have limitations. We counted any patient with EVALI who died as a fatal case, although many patients had preexisting medical conditions that may have contributed to death. Preexisting medical conditions may have influenced hospitalization; however, a previous report noted that as of November 5, 2019, nonhospitalized patients with EVALI were similar to hospitalized patients with EVALI and composed only 5% of all cases.⁹ If death takes place outside a hospital, it may be difficult to attribute the cause of death to EVALI, so some fatal cases may have been missed. Many patients were either too ill to interview or died before they could be interviewed, so some data are subject to limitations of reporting by proxies. More in-depth data collection and follow-up were used for patients with fatal cases but not for those with nonfatal cases; this could account for some reported differences. We did not compare data on glucocorticoid use between patients with fatal cases and those with nonfatal cases because of limited information from those with nonfatal cases. Misclassification of substance use could have occurred, and ascertainment bias may have been present in the reporting of substance use between the patients with fatal cases and those with nonfatal cases because more proxy interviews were performed for the patients who died.

About half the patients in this series were seen at least once in an outpatient setting before their final hospital admission; earlier diagnosis and appropriate treatment may have improved outcomes.^15,35 Eight patients had been discharged from a hospital before they died, which suggests that there may be opportunities to review stability before discharge, to tailor messaging and interventions more toward the avoidance of e-cigarette, or vaping, products, to communicate the importance of adherence to a glucocorticoid-tapering schedule, to enhance coordinated postdischarge care, and to identify worsening symptoms with timely follow-up after discharge.²⁹

Chronic conditions, such as asthma, chronic obstructive pulmonary disease, cardiac disease, and any mental health condition, were common among hospitalized patients with EVALI. Clinicians and public health practitioners should ensure an accurate and timely assessment of exposures and cessation of those exposures.