From 1993 to 2018, hantavirus infections were reported in 39 states, with hantavirus pulmonary syndrome (HPS) as the most common and fatal manifestation. To identify differences in the presentation of HPS between children and adults, we hypothesized that children with HPS would be diagnosed later in their illness course given the nonspecific clinical features of HPS.
This was an evaluation of the clinical and demographic characteristics of national HPS cases from 1993 to 2018. Data were from the Centers for Disease Control and Prevention database and 1 state department of health, comprising 97% of US cases. We compared children (0 to 12 years), adolescents (13 to 18 years), and adults using nonparametric and parametric analyses, with additional exploratory analyses to identify clinical variables associated with mortality.
Among 719 HPS patients, 22 (3.0%) were aged ≤12 years, 47 (6.5%) were 13 to 18 years old, and the remaining 650 (90.4%) were adults. Overall mortality was 35.4% and did not differ between age groups (P = .8). The time between symptom onset and death differed by age group, with children living a median of 2 days (interquartile range [IQR] 2 to 3), adolescents 4 days (IQR 3 to 5), and adults 5 days (IQR 4 to 8; P = .001). The mean highest hematocrit and median highest creatinine level were significantly associated with mortality in those 0 to 18 years old but not adults.
In our dataset representing the largest study of HPS in the United States, we found that children with HPS died more quickly than adults and that highest hematocrit and creatinine levels were associated with death only among those <19 years old.
HPS is a deadly disease and is difficult to recognize. To our knowledge, there have never been any large studies comparing clinical features and outcomes for children and adults with HPS in the United States in the past decade.
Our national dataset represents the largest US study of HPS. We found that children with HPS die more quickly than adults and that highest levels of hematocrit and creatinine are specifically associated with death in children.
In May of 1993, multiple cases of a rapidly progressing illness causing pulmonary edema, respiratory failure, and death were reported in New Mexico and were subsequently found to be secondary to a previously unknown hantavirus. This new hantavirus was called Sin Nombre virus (SNV), meaning “virus without a name” in Spanish.1 Hantaviruses are RNA viruses in the Hantaviridae family2 known to cause 2 types of illnesses in humans: (1) hantavirus pulmonary syndrome (HPS), also referred to as hantavirus cardiopulmonary syndrome, which occurs in North and South America and primarily affects lung tissue,1 and (2) hemorrhagic fever with renal syndrome, which occurs predominantly in Asia and Northern Europe and has a renal predilection.1–3
In the United States, HPS is most often caused by SNV and has a mortality rate of up to 35%.2,4 The deer mouse Peromyscus maniculatus is the reservoir species of SNV, although other rodents may harbor the virus.5,6 Infection typically occurs through the inhalation of particles from infected mouse droppings, urine, or saliva or through direct contact with an infected mouse.3,7 HPS cases have also been reported to occur after a mouse bite.8 No cases of human-to-human transmission have been reported in the United States.2
HPS became a nationally notifiable disease in the United States in 1995.9 From 1993 to 2020, the Centers for Disease Control and Prevention (CDC) and Nationally Notifiable Disease Surveillance System reported 833 cases of HPS and nonpulmonary hantavirus infections in the United States, with the highest caseload in New Mexico (118 cases).4 Ninety-four percent of hantavirus cases occurred in states west of the Mississippi River.4
Diagnosis of HPS, especially in the pediatric population, is difficult because of its rare occurrence and its similar presentation to other viral illnesses.9 Given the frequency with which common respiratory viruses affect children, providers may not consider a diagnosis of HPS in a child until late in the course of the illness. It is important that providers recognize clinical and laboratory findings of HPS because early diagnosis and critical care management at a hospital with extracorporeal membrane oxygenation (ECMO) capabilities improve survival.10–12 The purpose of this study was to (1) describe laboratory-confirmed HPS cases in the United States over a 26-year period (from 1993 to 2018), and (2) compare the presentation and outcomes of HPS between pediatric and adult patients. We hypothesized that the duration of time between the onset of symptoms and definitive diagnosis would be longer in children than in adults given the overlap in presenting symptoms with more common viral infections of childhood.
Methods
This case series study used the CDC Viral Special Pathogens Branch database of confirmed US HPS cases. Individual state departments of health report cases of HPS by submitting Hantavirus Disease Case Report Forms to the CDC. From 1993 to 2018, 34 states reported confirmed domestically acquired HPS cases to the CDC. Of these, 31 states consented to have the CDC share their data with us. One additional state provided its data directly to us. HPS cases with assumed exposure outside of the United States were excluded.
The variables we requested from the CDC dataset included age and sex, date of symptom onset, specimen acquisition date, history of rodent exposure, fever (>101°F or 38.3°C), and highest temperature recorded, thrombocytopenia (<150 000 platelets/mm3) and lowest platelet count measured, highest hematocrit percentage measured, highest creatinine measured (mg/dL), chest radiograph suggestive of acute respiratory distress syndrome (ARDS; in the opinion of the clinician completing the case report form), intubation history, outcome (alive/dead/unknown), date of death (if deceased), and ECMO history. We omitted the case location to ensure anonymity.
We defined the duration of time between the onset of symptoms and definitive diagnosis (when the clinical team caring for the patient likely first considered the diagnosis) as the number of days between the variables “date of symptom onset” and “specimen acquisition date.” We defined the time between symptom onset and death as the number of days between the variables “date of symptom onset” and “death date.” Many records were missing entries for 1 or more variables, particularly in early years when data for variables such as “specimen acquisition date” and “ECMO” were not routinely collected. These fields were left blank, and analyses were conducted on the sample size that was available for each combination of variables.
Patient age was grouped as child (≤12 years), adolescent (13 to 18 years), or adult (≥19 years) and was the independent variable for most analyses. For exploratory analyses of mortality, the sample of children and adolescents was too small, and these 2 groups were combined to create a “pediatric” group. For analyses involving continuous outcome variables and 3 age groups, we used ANOVA or Kruskal-Wallis rank tests, and for those with binary outcomes, we used χ2 and Fisher’s exact tests. Two group predictor variable analyses with numeric outcomes used t tests or the Wilcoxon rank test. For analyses with small sample sizes and nonnormal distributions, we used nonparametric analyses and metrics such as logistic regression, median, and interquartile range (IQR). A threshold α value of 0.05 was used to indicate significance. The study was approved by the Institutional Review Board of the UNM Health Sciences Center (UNM HRRC # 20-401). The CDC’s Human Research Protection Office reviewed and approved the protocol (protocol #7416) in accordance with 45 C.F.R. § 46.104.
Results
The subset of hantavirus data we used from 1993 to 2018 included 720 out of a total of 742 domestically acquired HPS cases (97%) in the United States and 32 of 34 states with reported cases (excluding HPS cases with assumed exposure outside of the United States). We excluded 1 person for whom age was not available for a final sample size of 719. Overall mortality was 35.4% and did not differ by age group (P = .82; Table 1). Case counts per year ranged from 10 to 41. The highest number of cases occurred in May, June, and July.
Demographic and Clinical Characteristics of 719 Pediatric and Adult HPS Cases Reported to the CDC
Characteristic . | n . | All ages . | Children, n . | Children 0–12 Y . | Adolescents, n . | Adolescents 13–18 Y . | Adults, n . | Adults 19+ Y . | P . |
---|---|---|---|---|---|---|---|---|---|
Age, median (IQR) | 719 | 38 (26–51) | 22 | 10 (7–12) | 47 | 16 (14–18) | 650 | 40 (29–52) | |
Sex | 717 | 648 | .53 | ||||||
Male (n %) | 441 | 61.5 | 11 | 50.0 | 29 | 61.7 | 401 | 61.9 | |
Female (n %) | 276 | 38.5 | 11 | 50.0 | 18 | 38.3 | 247 | 38.1 | |
Number with fever >38.3°C | 589 | 499 (84.7) | 16 | 13 (81.2) | 37 | 30 (81.1) | 536 | 456 (85.1) | .75 |
Highest temp °C, mean, SD | 202 | 39.1, 0.90 | 7 | 39.5, 0.77 | 12 | 39.2, 1.3 | 183 | 39.1, 0.87 | .52 |
Percent with thrombocytopenia <150 000 platelets/mm3 | 597 | 576 (96.3) | 14 | 14 (100.0) | 41 | 41 (100.0) | 542 | 520 (95.9) | 1.0 |
Lowest platelets/mm3, median (IQR) | 329 | 51 000 (32 000–74 000) | 9 | 52 000 (46 000– 80 000) | 25 | 42 000 (33 000–68 000) | 295 | 51 000 (31 000–74 000) | .61a |
Highest hematocrit %, mean (SD) | 296 | 50.3 (10.3) | 7 | 46.2 (2.9) | 21 | 53.1 (9.7) | 268 | 50.2 (10.4) | .04a |
Highest creatinine mg/dL,c median (IQR) | 261 | 1.3 (1–2) | 7 | 0.7 (0.6–0.7) | 18 | 1.1 (0.9–2.6) | 236 | 1.3 (1.0–2.0) | .006a |
Highest creatinine, µmol/L | 261 | 115.0 (88.4–176.8) | 7 | 61.9 (53.1–61.9) | 18 | 97.3 (79.6–229.9) | 236 | 115.0 (88.4–176.8) | .006a |
Chest radiograph findings of ARDS | 590 | 564 (95.6) | 14 | 13 (92.9) | 39 | 36 (92.3) | 537 | 515 (95.9) | .25b |
Intubated | 628 | 398 (63.4) | 17 | 6 (35.3) | 43 | 28 (65.1) | 568 | 364 (64.1) | .05 |
ECMO | 71 | 14 (19.7) | 3 | 1 (33.3) | 0 | 0 | 68 | 13 (19.1) | 1.0 |
Outcome | 714 | ||||||||
Fatal | 253 (35.4) | 22 | 7 (31.8) | 46 | 18 (39.1) | 646 | 228 (35.3) | .82 | |
Time between symptom onset and specimen collection in days | 290 | Mean (SD): 8.0 (9.8) | 9 | Mean (SD): 7.6 (10.8) | 18 | Mean (SD): 12.8 (24.2) | 263 | Mean (SD): 7.7 (7.9) | .22a |
Median (IQR): 6 (4–8) | Median (IQR): 4 (3–6) | Median (IQR): 4.5 (2–9) | Median (IQR): 6 (4–8) | ||||||
Range: 0–92 d | Range: 2–36 d | Range: 0–92 d | Range: 0–79 d | ||||||
Time between symptom onset and death in days (median, IQR) | 221 | Mean (SD): 7.0 (6.4) | 6 | Mean (SD): 2.5 (0.8) | 14 | Mean (SD): 4.4 (2.5) | 201 | Mean (SD): 7.3 (6.6) | .001a |
Median (IQR): 5 (4–8) | Median (IQR): 2 (2–3) | Median (IQR): 4 (3–5) | Median (IQR): 5 (4–8) | ||||||
Range: 0–42 d | Range: 2–4 d | Range: 1–10 d | Range: 0–42 d |
Characteristic . | n . | All ages . | Children, n . | Children 0–12 Y . | Adolescents, n . | Adolescents 13–18 Y . | Adults, n . | Adults 19+ Y . | P . |
---|---|---|---|---|---|---|---|---|---|
Age, median (IQR) | 719 | 38 (26–51) | 22 | 10 (7–12) | 47 | 16 (14–18) | 650 | 40 (29–52) | |
Sex | 717 | 648 | .53 | ||||||
Male (n %) | 441 | 61.5 | 11 | 50.0 | 29 | 61.7 | 401 | 61.9 | |
Female (n %) | 276 | 38.5 | 11 | 50.0 | 18 | 38.3 | 247 | 38.1 | |
Number with fever >38.3°C | 589 | 499 (84.7) | 16 | 13 (81.2) | 37 | 30 (81.1) | 536 | 456 (85.1) | .75 |
Highest temp °C, mean, SD | 202 | 39.1, 0.90 | 7 | 39.5, 0.77 | 12 | 39.2, 1.3 | 183 | 39.1, 0.87 | .52 |
Percent with thrombocytopenia <150 000 platelets/mm3 | 597 | 576 (96.3) | 14 | 14 (100.0) | 41 | 41 (100.0) | 542 | 520 (95.9) | 1.0 |
Lowest platelets/mm3, median (IQR) | 329 | 51 000 (32 000–74 000) | 9 | 52 000 (46 000– 80 000) | 25 | 42 000 (33 000–68 000) | 295 | 51 000 (31 000–74 000) | .61a |
Highest hematocrit %, mean (SD) | 296 | 50.3 (10.3) | 7 | 46.2 (2.9) | 21 | 53.1 (9.7) | 268 | 50.2 (10.4) | .04a |
Highest creatinine mg/dL,c median (IQR) | 261 | 1.3 (1–2) | 7 | 0.7 (0.6–0.7) | 18 | 1.1 (0.9–2.6) | 236 | 1.3 (1.0–2.0) | .006a |
Highest creatinine, µmol/L | 261 | 115.0 (88.4–176.8) | 7 | 61.9 (53.1–61.9) | 18 | 97.3 (79.6–229.9) | 236 | 115.0 (88.4–176.8) | .006a |
Chest radiograph findings of ARDS | 590 | 564 (95.6) | 14 | 13 (92.9) | 39 | 36 (92.3) | 537 | 515 (95.9) | .25b |
Intubated | 628 | 398 (63.4) | 17 | 6 (35.3) | 43 | 28 (65.1) | 568 | 364 (64.1) | .05 |
ECMO | 71 | 14 (19.7) | 3 | 1 (33.3) | 0 | 0 | 68 | 13 (19.1) | 1.0 |
Outcome | 714 | ||||||||
Fatal | 253 (35.4) | 22 | 7 (31.8) | 46 | 18 (39.1) | 646 | 228 (35.3) | .82 | |
Time between symptom onset and specimen collection in days | 290 | Mean (SD): 8.0 (9.8) | 9 | Mean (SD): 7.6 (10.8) | 18 | Mean (SD): 12.8 (24.2) | 263 | Mean (SD): 7.7 (7.9) | .22a |
Median (IQR): 6 (4–8) | Median (IQR): 4 (3–6) | Median (IQR): 4.5 (2–9) | Median (IQR): 6 (4–8) | ||||||
Range: 0–92 d | Range: 2–36 d | Range: 0–92 d | Range: 0–79 d | ||||||
Time between symptom onset and death in days (median, IQR) | 221 | Mean (SD): 7.0 (6.4) | 6 | Mean (SD): 2.5 (0.8) | 14 | Mean (SD): 4.4 (2.5) | 201 | Mean (SD): 7.3 (6.6) | .001a |
Median (IQR): 5 (4–8) | Median (IQR): 2 (2–3) | Median (IQR): 4 (3–5) | Median (IQR): 5 (4–8) | ||||||
Range: 0–42 d | Range: 2–4 d | Range: 1–10 d | Range: 0–42 d |
Percentages are based on the number of cases for whom the characteristic was known and will not always sum to the total sample size of 719.
Used Kruskal-Wallis equality of populations rank test because of nonnormal distribution.
Used Fisher’s exact test because of cell sizes <5.
Values for creatinine were not provided on the case report forms, creatinine units were assumed to be in mg/dL.
There were 69 pediatric hantavirus cases and 650 adult cases; of the pediatric patients, 22 (31.9%) were ≤12 years old and 47 (68.1%) were 13 to 18 years old. The median age was 14 years, the IQR was 12 to 17 years, and the youngest patient with a known age was 5 years old. Table 1 contains a summary of demographic and clinical data for each group. There were many similarities between the child, adolescent, and adult groups. Adolescent and adult HPS patients were similar with respect to being disproportionately male, all 3 age groups had a high proportion of patients who presented with fever (>101°F), and almost all patients had thrombocytopenia. Nearly all patients had chest radiograph findings consistent with ARDS. Differences between age groups were evident in only 2 of the clinical variables. Highest hematocrit was lowest among the youngest age group (mean: 46.2; standard deviation [SD]: 2.9) and highest among adolescents (mean: 53.1; SD: 9.7; P = .04). Highest creatinine levels were lowest among children (median: 0.7; IQR: 0.6–0.7), and higher among adolescents (median: 1.1; IQR: 0.9–2.6) and adults (median: 1.3; IQR: 1.0–2.0).
Time between symptom onset and diagnosis did not differ significantly between age groups; the median time for children was 4 days, adolescents 4.5 days, and adults 6 days (Table 1). However, the median time between symptom onset and death was shortest for the youngest children (median: 2; IQR: 2–3), longer for adolescents (median: 4; IQR: 3–5), and longest for adults (median: 5; IQR: 4–8; P <.001). The longest survival time was 4 days for children, 10 days for adolescents, and 42 days for adults.
Mortality was associated with 3 of the clinical measures (Table 2). For this small subsample of cases for whom this information was available, all pediatric patients (aged 0 to 18 years) were considered together. Highest hematocrit was associated with mortality, but only in pediatric patients. The mean (± SD) highest hematocrit among children who died was 55.6 ± 7.2% whereas, among children who survived, the mean was 48.4 ± 8.4% (P = .04). Highest hematocrit was not associated with mortality in the adult sample (P = .08).
Clinical Measures Associated With Mortality Among Pediatric Cases of HPS, 1993–2018
. | Pediatric Age ≤18 y . | Adult . | ||||||
---|---|---|---|---|---|---|---|---|
Characteristic . | n . | Lived . | Died . | P . | n . | Lived . | Died . | P . |
Highest hematocrit %, mean (SD)a | 27 | 48.4 (8.4) | 55.6 (7.2) | .04 | 267 | 49.4 (8.6) | 51.8 (13.5) | .08 |
Highest creatinine mg/dL,b OR (95% CI) | 25 | 7.3 (1.4–38.0) | .02 | 236 | 1.1 (0.9–1.3) | .46 | ||
Intubated, OR (95% CI) | 59 | 4.0 (1.2–13.0) | .02 | 565 | 11.0 (6.4–19.2) | <.001 |
. | Pediatric Age ≤18 y . | Adult . | ||||||
---|---|---|---|---|---|---|---|---|
Characteristic . | n . | Lived . | Died . | P . | n . | Lived . | Died . | P . |
Highest hematocrit %, mean (SD)a | 27 | 48.4 (8.4) | 55.6 (7.2) | .04 | 267 | 49.4 (8.6) | 51.8 (13.5) | .08 |
Highest creatinine mg/dL,b OR (95% CI) | 25 | 7.3 (1.4–38.0) | .02 | 236 | 1.1 (0.9–1.3) | .46 | ||
Intubated, OR (95% CI) | 59 | 4.0 (1.2–13.0) | .02 | 565 | 11.0 (6.4–19.2) | <.001 |
% Analysis done using logistic regression.
Analysis done using t test.
Units for creatinine were not provided on the case report forms; creatinine units were assumed to be in mg/dL.
Highest creatinine level was also associated with mortality among pediatric cases (Table 2). The odds of dying among patients aged 18 years and younger increased by 630% (odds ratio [OR]: 7.3; 95% confidence interval [CI]: 1.4–38.1) for each 1-point increase in the highest creatinine measurement among children with HPS (P = .02). The median highest creatinine level among children who had HPS and died was 4 times that among those who survived (P <.001). Among adults with HPS, increases in highest creatinine levels were not associated with mortality (P = .46).
There was a strong relationship between experiencing intubation and mortality among both pediatric and adult HPS patients. Among 59 children for whom these data were available, the odds of mortality were 4 times as high among intubated patients as they were among those who were not intubated. Among adults, the odds ratio was 11.0 (Table 2). There was no relationship between receiving ECMO and mortality (OR: 0.99; 95% CI: 0.27–3.7), but only 68 patients had complete data for this analysis, only 2 of whom were children.
Discussion
In this study, we compare the clinical presentation of HPS cases among children, adolescents, and adults from a nearly complete US clinical population reported to the CDC from 1993 to 2018. The resulting analysis provides unique insight into how this disease affects children and adults differently. It is also, to our knowledge, the most complete sample of pediatric and adult HPS cases in the United States reported and compared in the literature to date.
Our study is the first to reveal that progression from symptom onset to death in children with HPS is significantly shorter than in adults. This has important clinical implications and argues for a high index of suspicion of the disease among health care providers working in hantavirus endemic regions. The median time to death in children from symptom onset in our study was only 2 days. This, in combination with a 32% mortality for HPS in children, provides evidence supporting immediate transfer to a tertiary medical center with ECMO capabilities for children with presentations concerning for the disease, with no role for outpatient observation, even in well-appearing children. Given that HPS is most common in rural areas,13 primary care providers may be the initial point of contact for such patients.
The reasons for this age-related rapid progression to death are not clear. Previous work has noted that survivors of HPS are more likely to develop immunoglobulin G antibody responses after infection and to ultimately attain higher titers of IgG.14 Higher levels of neutralizing antibody at admission are also associated with a milder disease course.15 It is possible that adults may mount more robust immunologic responses early in infection, which may account for increased survival times. If true, one may expect higher mortality in children than adults, as opposed to the equivalent mortality found in our study. However, given the better outcomes demonstrated by children with many viral infections compared with adults16–23 and the relative lack of medical comorbidities in children, mortality in children equal to adults for HPS may actually indicate relatively worse pediatric outcomes. For example, although different diseases than HPS, mortality rates may be up to 100-fold higher in older adults compared with young children with influenza and 1000-fold higher with severe acute respiratory syndrome coronavirus 2 infection.22,23 Given that reinfection with hantaviruses is not well described, protective preexisting immunity (due to previous subclinical infection) in adults is unlikely to play a role.
We also noted a novel association between increased creatinine and hematocrit in fatal versus nonfatal HPS in children, but not adults. Associations between mortality and elevated creatinine24–27 and hematocrit24,25,27,28 have been reported, although never specific to children. However, previous work has suffered from smaller numbers of pediatric cases25 or has analyzed pediatric and adult cases in aggregate.24,26–28 Our study was able to use a large, population-level dataset permitting the analysis and comparison of children and adults, which may have detected previously unappreciated associations.
Elevations in creatinine and hemoconcentration may simply serve as markers for extreme vascular permeability and end-organ damage, which places such patients at increased risk of death. However, it is unclear why this effect would be more pronounced in children. The underlying pathology of HPS is primarily driven by alterations in vascular permeability as a result of viral-induced inhibition of endothelial cell integrity due to hantavirus interactions with αγβ3 integrins and immune dysregulation.29–31 Age-related differences in factors affecting the integrity of vascular endothelium (either systemically or localized to the kidney) may differentially increase susceptibility to the destabilizing effects of hantaviruses on endothelial cells.32 The development of renal insufficiency is a risk factor for mortality among children with sepsis in general, however,33,34 and thus may represent a manifestation of the immature kidney to infection independent of any hantavirus-specific effects.
It should be noted that previous analyses of hantavirus datasets in the United States27 incorporated retrospective cases reported in the literature before the establishment of the CDC database in 1993. However, we chose not to include these cases because they may introduce a selection bias for mortality given their reliance on autopsy data,35,36 use of different methodology for confirmation of infection,35–37 and a detection bias, given incomplete reporting of laboratory data in many cases.35,37,38
This study was limited by the design, for which data were captured at one point in time on case report forms designed to fill needs not specific to research. Longitudinal information about the progression of the disease would have been invaluable. The use of specimen acquisition date was a proxy measure for the time that the diagnosis of HPS was first considered because providers may have ordered the test before seriously considering the diagnosis or may have ordered the test from scavenged blood collected previously. In addition, the use of days (as opposed to a more granular measurement such as hours) complicates the interpretation of the exact time course of the disease as it relates to time to hospitalization and death. We chose to use laboratory values that were consistent across the various case report forms over most study years and thus did not capture white blood cell count and notes on the clinical course of illness. To ensure anonymity given the rarity of the disease, we did not collect patient demographic information, including ethnicity, race, and location. Although nearly a comprehensive study, the population of pediatric cases available for study is thankfully small. Several cases were missing data for some variables. Nonetheless, this is a near-population-level study and adds substantially to what is known about pediatric HPS.
Conclusion
HPS is a deadly disease that kills more than one-third of the people it infects. Children merit particular attention because they have a brief window between symptom onset and death. For patients of any age, this disease must be recognized promptly by providers to initiate ICU level of care at an ECMO-capable facility. Children differ from adults in that highest hematocrit and highest creatinine levels were associated with mortality. Further research is needed to better characterize these differences that appear to represent prognostic markers for worse outcomes. Health care providers working in rural regions endemic to hantavirus must maintain a high index of suspicion for this disease and expedite transfer to an ECMO-capable facility for any patients with suspected HPS.
Acknowledgments
We want to thank the New Mexico Department of Health for their expertise and assistance with the study, as well as additional individual departments of health located throughout the United States for granting us permission to use data.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or any institutions with which the authors are affiliated.
Drs Dehority and Thorp conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Fullerton analyzed and interpreted the data, conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript; Ms Whitesell provided CDC data and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest relevant to this article to disclose.
Comments