CLINICAL INVESTIGATION
Relapse after Emergency Department Discharge for Acute Asthma Brian H. Rowe, MD, MSc, CCFP(EM), Cristina Villa-Roel, MD, MSc, Marco L.A. Sivilotti, MD, MSc, FRCPC, Eddy Lang, MD, CCFP(EM), CSPQ, Bjug Borgundvaag, MD, PhD, CCFP(EM), Andrew Worster, MD, MSc, CCFP(EM), Allan Walker, MD, Scott Ross, MD, CCFP(EM)
Abstract Objectives: The objectives were to determine patient and treatment-response factors associated with relapse after emergency department (ED) treatment for acute asthma. Methods: Subjects aged 18–55 years who were treated for acute asthma in 20 Canadian EDs prospectively underwent a structured ED interview and telephone contact 2 weeks later. Results: Of 695 enrolled patients, 604 (86.9%) were discharged from the ED; follow-up was available in 529 (87.5%); 63% were female and the median age was 29 years. Most patients were discharged on oral (70.8%) and inhaled (60.1%) corticosteroids (CS); 2-week treatment adherences were 93.3 and 80.9%, respectively. Relapse occurred in 9.2% at 1 week (95% confidence interval [CI] = 7.1% to 12.0%) and 13.9% (95% CI = 11% to 17%) at 2 weeks. In multivariable modeling, factors associated with relapse were ethnicity (risk ratio [RR] white = 0.66; 95% CI = 0.52 to 0.83); female gender (RR = 1.57; 95% CI = 1.14 to 2.09); any ED visits in the past 2 years (RR = 1.47; 95% CI = 1.18 to 1.80); ever admitted for asthma treatment (RR = 1.83; 95% CI = 1.09 to 2.84); use of combined inhaled CS plus long-acting b2agonists (RR = 1.39; 95% CI = 1.07 to 1.78) and of oral CS (RR = 1.35; 95% CI = 1.12 to 1.59) at the time of ED presentation. Conclusions: Ethnicity (white), female gender, prior ED visits and admissions for asthma, and recent treatments (especially oral CS) were associated with asthma relapse, which remains relatively common. Future research is required to target this high-risk group. ACADEMIC EMERGENCY MEDICINE 2008; 15:709–717 ª 2008 by the Society for Academic Emergency Medicine Keywords: asthma, emergency department, corticosteroids, relapse
E
mergency departments (EDs) are key settings for the delivery of acute asthma care. It is estimated that nearly 2 million ED asthma visits occur
annually in the United States alone.1 Despite improved understanding of rates of presentation among different centers and risk factors associated with asthma
From the Department of Emergency Medicine (BHR, CVR, AW, SR) and the School of Public Health (BHR), University of Alberta, Edmonton, Alberta; the Department of Emergency Medicine, Queen’s University (MLAS), Kingston, Ontario; Sir Mortimer B. Davis-Jewish General Hospital (EL), Montreal, Quebec; Mt. Sinai Hospital (BB), Toronto, Ontario; Northeast Community Health Centre (AW), Edmonton, Alberta; McMaster University (AW), Hamilton, Ontario; and Sturgeon General Hospital, St. Alberta (SR), Alberta, Canada. All the Emergency Departments are in Canada and affiliated with the Canadian Association of Emergency Physicians Research Consortium (CAEP RC) and located in Canada. Received December 22, 2007; revision received March 17, 2008; accepted April 23, 2008. Presented in part at the Annual Meeting of the Society for Academic Emergency Medicine (SAEM) in San Francisco, CA (May 18– 21, 2006), and the International Conference on Emergency Medicine (ICEM) in Halifax, Nova Scotia (June 3–7, 2006). Conflicts: Within the past 2 years, Dr. Rowe has received research funding and speaker’s fees from the following respiratory companies: GSK, AstraZeneca (AZ), and Abbott. Dr. Lang has served on advisory boards for Abbott and Wyeth. Dr. Worster has received speaker’s fees from AZ. Drs. Rowe, Sivilotti, Lang, Borgundvaag, and Worster received funding from AZ for an investigatorinitiated, multicentered asthma trial 2005–2006. Drs. Villa-Roel, Walker, and Ross declare no known conflicts of interest. The authors thank GlaxoSmithKline (GSK) who provided an investigator-initiated, unrestricted, research grant to the CAEP RC to fund this study, CAEP for their support of the Research Consortium and this project, and the Department of Emergency Medicine Research Group (EMeRG) at the University of Alberta for their in-kind support of this project. Dr. Rowe is funded by the Government of Canada (Ottawa, ON) as a 21st Century Canada Research Chair. Address for correspondence and reprints: Dr. Brian H. Rowe; e-mail:
[email protected].
ª 2008 by the Society for Academic Emergency Medicine doi: 10.1111/j.1553-2712.2008.00176.x
ISSN 1069-6563 PII ISSN 1069-6563583
709
710
exacerbations,2,3 this disease continues to represent one of the principal causes of ED attendance, contributes to ED overcrowding, and results in impressive health care costs.4,5 In North America, approximately 80%–90% of the patients with acute asthma who present to the ED will be discharged.2,6,7 Of these patients, between 5 and 25% will relapse within the first week, and 21 to 35% will relapse within 3 weeks.7,8 Relapses represent important adverse health outcomes for the patient and the health care system, and efforts to reduce them have been ongoing. Although relapse events vary with both disease severity at presentation and treatment prescribed at discharge, other potential predictors have been postulated by some groups.2,9,10 Overall, a variety of demographic, historical, and treatment factors have been proposed; however, many of these studies have been methodologically weak, small, inconclusive and/or their conclusions are outdated. There is an urgent need to identify predictors of relapse in adult asthma patients after treatment and discharge from the ED. These patients represent an important target in clinical care mainly because of the possibility of stratifying their relapse risk. In addition, selected modifiable factors could be the focus of interventions for the prevention of subsequent relapses.11 The aim of this multicenter study was to determine patient and treatment-response factors associated with relapse within the first 2 weeks of ED treatment for acute asthma. METHODS Study Design This was a prospective cohort study that enrolled patients treated for acute asthma in 20 EDs across Canada between June 2004 and December 2005. Recruitment strategies varied at each site based on research staff availability; 24-hour-per-day sampling was not possible in any site. The institutional research ethics board of each site approved the study protocol, and written informed consent was obtained from all participants. Study Setting and Population The study sites were all members of the Canadian Association of Emergency Physicians Research Consortium (CAEP RC; http://www.caep.ca/rc). All but four sites were urban teaching sites assessing a median of 50,000 patients per year. Patients, aged 18–55 years with a clinical diagnosis of asthma (e.g., previous history of asthma, response to b2-agonist therapy, and worsening symptoms such as wheezing, dyspnea, cough, or chest tightness) and no evidence of chronic obstructive pulmonary disease (COPD) exacerbation were invited to participate in the study. Subjects with mild obstruction (>80% predicted peak expiratory flow [PEF] at presentation), >30 packyears of smoking, cognitive impairment, inability to provide informed consent, pneumonia, extremely severe asthma presentation that could not be stabilized in the ED (status asthmaticus or requirement of intubation), and those who were hospitalized were excluded from this analysis. A missed, refused, and other-
Rowe et al.
•
ASTHMA RELAPSE
exclusions (no contact telephone number and patients who left the ED against medical advice) database was maintained at each site. Patients’ ED therapy and disposition were left to the discretion of the treating ED physician. PEF was recorded during triage or after the first ED treatment, and spirometry was performed before discharge. At discharge, emergency physicians (EPs) prescribed medications as deemed appropriate and provided patients with a discharge plan. Most patients were discharged with a short course of oral corticosteroids (CS; 50-mg prednisone tablets daily for 5 to 7 days). Outcome assessment was performed 2 weeks after discharge by follow-up telephone interviews. Study Protocol Once the patient’s condition was stabilized, eligibility was assessed by a trained research nurse. Those who met enrollment criteria underwent a structured interview in the ED to capture sociodemographic, disease, and treatment-related data. The subsequent ED course and disposition were verified from medical record abstraction. Data collection forms were reviewed before their submission to the coordinating center, where they were entered by trained data management personnel. Sociodemographic factors such as age, gender, ethnicity, educational level, marital status, and employment were collected. The use of preventive actions such as an action plan, asthma diary, spacer device, PEF meter for asthma monitoring, and influenza vaccination were recorded. Smoking status was coded as never smoked, previous smoker, or current smoker. Insurance status was recorded as the percentage of medication costs usually paid by patients and as the self-report of reduced medication use in the previous 12 months. Availability of primary care provider (PCP) was evaluated by asking ‘‘Do you have a family physician?’’ PEF was standardized as the percentage of the patient’s predicted value according to age, gender, ethnicity, and height.12 Chronic asthma factors, such as the previous use of CS for asthma, number of ED visits and hospitalizations for the treatment of asthma (ever and during the past 2 years), and history of intubation were ascertained by patient recall. The Canadian Triage and Acuity Score (CTAS) as recorded by the ED triage nurse was used to determine the priority with which patients required assessment. This five-level scale, which ranges from resuscitation (1) to less urgent (5), is the national standard for all ED patients and has been shown to be valid and reliable.13 Adherence to post-ED treatment was addressed at the follow-up phone call by asking whether the patient could obtain the ED prescription and whether the medication was taken as prescribed or instructed. Measures Relapse was defined as an urgent visit to any ED, clinic, or physician office for worsening asthma symptoms (such as cough or dyspnea); this definition is an accepted approach used in other similar asthma studies.2 Side effects (e.g., insomnia, hoarseness, swelling, sore throat, and nausea) were also ascertained at each
ACAD EMERG MED • August 2008, Vol. 15, No. 8
•
www.aemj.org
follow-up. Scheduled follow-up visits to a PCP or clinic for asthma checkups were documented; however, these visits were not considered relapses. Suspected patient relapses were identified at telephone follow-up and by medical record review. Objective evidence of worsening asthma was sought from the medical records (increase in severity of symptoms, need of urgent medical treatment, change in asthma medications or treatment plan, and ⁄ or hospitalization). Relapses were adjudicated by an external committee of two independent reviewers. Disagreements were resolved by consensus after including a third member. Data Analysis We considered a baseline relapse rate of 20% (based on the results of similar studies of adults with asthma discharged from EDs after treatment for acute asthma).2,9,10 A sample size of 506 adults was adequate for a univariable logistic regression analysis of asthma relapse and urgent visits to a clinic in the last year. From the sample obtained, a multivariable logistic regression (MLR) analysis of relapse and predictors that have a multiple correlation coefficient of 0.5, a sample size of 529 provided a power of 80% with a 5% level of significance.14 Descriptive data include proportions with 95% confidence intervals (CIs), means with standard deviations (SDs), or medians with interquartile range (IQR) as appropriate. Bivariable analyses for dichotomous variables were performed by chi-square test; bivariable analyses for continuous variables were performed by t-tests or Mann-Whitney tests, as appropriate. The database for refusals, misses, and other exclusions was analyzed. All multivariable analyses were performed using SAS (Version 9.1, SAS Institute Inc., Cary, NC). First, known factors associated with relapse and potential relapse predictor variables identified at p < 0.15 significance testing were included in the model. Next, a multivariable analysis was performed using a stepwise method to reduce the list of potential predictors to include only those with significant testing at p < 0.10. An a priori decision was made to control for the following important variables in the final model: age, gender, and ethnicity. Based on the distribution of the variable ‘‘Ethnicity,’’ it was dichotomized as white and nonwhite for the multivariable analyses. All variables remaining in the reduced model were further analyzed using the generalized estimating equation (GEE) method to adjust for the clustering of patients within the sites involved in this study. All first-order interaction terms were examined between factors in the GEE MLR model and if there were significant interactions terms, they were included in the multivariable model. The goodness of fit was assessed using deviance statistics. An MLR model was developed to estimate the risk of relapse following ED treatment for acute asthma, controlling for demographic characteristics, asthma history, severity of current exacerbation, ED course, and discharge medication. The odds ratio (OR) and 95% CI were corrected to derive a more appropriate estimation of the magnitude of the association: risk ratio (RR).15 A p-value of < 0.05 was considered statistically significant in the final MLR model.
711
RESULTS Enrollment Over the study period, 1,591 ED patients were screened for enrollment, and 1,105 (69%) potentially met eligibility criteria after excluding misses and refusals (Figure 1). An additional 410 patients were ineligible; the most common exclusions were for suspected pneumonia (31.1%), extremely severe asthma presentation that could not be stabilized in the ED (10.3%), cognitive impairment (8.2%), and COPD exacerbation or more than 30 pack-years of smoking (6.7%). Overall, 43.5% were ineligible for other reasons, such as chronic pathologic conditions, history of drug abuse, inability to communicate in English, and patients who left the ED against medical advice. The median age of the patients who were not enrolled in the study was 36 (IQR = 26 to 48) years and 60.1% were female. Of the 695 patients who were enrolled, 604 (87%) were discharged from the ED, and 529 were available for follow-up. Demographics Of the 604 study patients, 63% were female and the median age was 29 (IQR = 23 to 39) years. Ethnicity was visually assessed by the dedicated research nurse as white (85.4%), Aboriginal (2.8%), Oriental (2.1%), East Indian (2.8%), African-Canadian (2.9%), or other (3.6%). Outcome Among the 529 study patients with follow-up data, 9.2% (95% CI = 7.1% to 12.0%) relapsed at 1 week and
Figure 1. Patient flow.
712
Rowe et al.
13.9% (95% CI = 11% to 17%) within 2 weeks of the ED visit. Females were more likely to relapse, as were patients already receiving oral CS and any inhaled CS agent. No statistically significant differences were found among sociodemographic characteristics, educational level, and the availability of PCP between patients who relapsed and those who did not (Table 1).
•
ASTHMA RELAPSE
Preventive Actions Preventive actions such as having an ‘‘action plan’’ and owning a ‘‘spacer device’’ were more frequent in patients who relapsed than in those who did not. Otherwise, other preventive interventions (e.g., using an asthma diary, using a PEF meter for monitoring, and previous influenza vaccination) were similar
Table 1 Characteristics of Patients with Acute Asthma Discharged from the ED Categorized by Relapse Status
Characteristics Demographic factors Age, yr (±SD) Female Ethnicity White Aboriginal Oriental East Indian African-Canadian Other High school graduate Marital status (married ⁄ CL) Employment (full-time) Insurance status Patient pays Reported less medication use Preventive actions Have an ‘‘action plan’’ Use an asthma diary Own a ‘‘spacer device’’ Use peak flow meters Recent influenza vaccination Smoking status Never Previous Current Have PCP Chronic asthma factors Ever taken steroid medicine for asthma No. of ED visits in past 2 years Admitted for asthma in past 2 years Ever intubated for asthma ED usual site for problem asthma care ED usual source of asthma prescriptions Medications at presentation to the ED Inhaled SABA ICS Fluticasone* Budesonide* Other ICS* Any ICS agent (including combination agents) Inhaled LABA Combination agents (ICS + LABA) Leukotriene modifier ⁄ antagonist OCS Anticholinergics Short-acting anticholinergics* Long-acting anticholinergics* Combined SABA + anticholinergics Theophylline Mast cell stabilizers Antibiotics Nasal sprays
Relapse (n = 74) 33 ± 10.2 57 (77) 62 3 2 1 3 3 59 35 37 38 ⁄ 53 20 14 ⁄ 51
(83.7) (4.0) (2.7) (1.3) (4.0) (4.0) (79.7) (47.3) (50) (71.7) (0, 20) (27.4)
No Relapse (n = 455)
p-Value
31 ± 9.9 279 (61.3)
0.103 0.009
399 7 9 13 11 16 392 198 214 271 ⁄ 373 20 97 ⁄ 360
0.695
(87.6) (1.5) (1.9) (2.8) (2.4) (3.5) (86.1) (43.5) (47.0) (72.6) (0, 20) (26.9)
0.148 0.543 0.635 0.917 — 0.604
40 7 35 27 35
(54.0) (9.4) (47.2) (36.4) (47.3)
171 27 150 121 176
(37.5) (5.9) (32.9) (26.5) (38.6)
0.001 0.251 0.017 0.188 0.329
21 23 30 64
(28.3) (31.0) (40.5) (86.4)
155 123 177 364
(34.0) (27.0) (38.9) (80)
0.594
51 2 22 8 51 6
(68.9) (0, 5) (29.7) (10.8) (68.9) (8.1)
262 1 82 36 265 45
(57.7) (0, 3) (18.0) (7.9) (58.2) (9.8)
0.004 0.000 0.019 0.701 0.082 0.630
65 23 18 9 0 48 4 33 9 12 11 10 1 1 0 0 2 0
(87.8) (31.0) (24.3) (12.1)
370 135 118 42 7 228 17 135 29 29 15 14 3 15 5 0 5 1
(81.3) (29.6) (25.9) (9.2) (1.5) (50.1) (3.7) (29.6) (6.3) (6.3) (3.3) (3.2) (0.6) (3.3) (1.1)
0.174 0.806
0.365 0.365
(1.0) (0.2)
0.264 0.686
(64.8) (5.4) (44.5) (12.1) (16.2) (14.8) (13.5) (1.3) (1.3) (2.7)
0.188
0.018 0.495 0.011 0.074 0.003 0.000
Data are reported as number (%), mean ± SD, or or median (IQR). CL = common-law; ED = emergency department; ICS = inhaled corticosteroid; IQR = interquartile range; LABA = long-acting b2-agonists; OCS = oral corticosteroids; PCP = primary care provider; SABA = short-acting b2-agonists; SD = standard deviation. *Not mutually exclusive.
ACAD EMERG MED • August 2008, Vol. 15, No. 8
•
www.aemj.org
between the groups. A status of previous or current smoking was more common in patients who relapsed than in those who did not (Table 1). Asthma History More patients who relapsed had at least one ED or urgent clinic visit for acute asthma during the previous 2 years (71.2% vs. 53.4%, p = 0.005). Relapse patients were more likely to report at least 2 days of activity limitations before the ED visit (70.2% vs. 55.1%, p = 0.01). Differences between patients’ pre-ED treatments according to the occurrence of relapse are shown in Figure 2. Severity of Asthma Presentations and ED Management There were no statistically significant differences in the CTAS, ED course, or lung function tests (earliest ⁄ final PEF and percentage of predicted PEF), between patients who relapsed and those who did not (Table 2). Post-ED Treatment Most patients were discharged on oral (70.8%) and inhaled (60.1%) CS with self-reported adherence rates of 93.3% (352 of 377) and 80.9% (251 of 310), respectively. Adherence rates by relapse status are shown in Table 3. Follow-up Of the 604 patients who were discharged, 75 (12.5%) were not able to be contacted 2 weeks after the ED visit or refused to continue in the study. Follow-up interviews were completed in 529 patients (87.5%). Of the 529, 1 patient in the group who relapsed (1.3%) and 127 patients in the group who did not relapse (27.9%) were seen by a physician for ‘‘routine care’’ after the ED visit (p = 0.0001). Ninety-seven percent of the patients who relapsed did so by seeking ED care during the 2-week follow-up period (ED relapse rate 13.6%), resulting in 12 hospitalizations (16% of all relapses). One death (unrelated to asthma) was documented in the group of patients who did not relapse. Factors Associated with Relapse From the multivariable analyses, relapse was significantly associated with being female (RR = 1.57, 95%
Figure 2. Patient before emergency department (ED) treatment. SABA = inhaled short-acting b2-agonists; ICS = inhaled corticosteroid; LABA = inhaled long-acting b2-agonists; OCS = oral corticosteroids; LEUK = leukotriene modifier ⁄ antagonist.
713
CI = 1.14 to 2.09), having any ED visits in the past 2 years (RR = 1.47, 95% CI = 1.18 to 1.80), ever being hospitalized for asthma (RR = 1.83, 95% CI = 1.09 to 2.84), using combination inhaled CS plus long-acting b2-agonists (RR = 1.39, 95% CI = 1.07 to 1.78), and using oral CS (RR = 1.35, 95% CI = 1.12 to 1.59) at the time of ED presentation; white race protected against relapse (RR = 0.66, 95% CI = 0.52 to 0.83). No interactions were retained in the final MLR model as significant. Ethnicity and gender were correlated, but because they do not have a directional correlation, they both were included in to the model. There were nine missing data points in the variables, ‘‘number of ED visits in the past 2 years’’ and ‘‘ever admitted to a hospital for asthma in the past 2 years.’’ In these particular cases, reference values were imputed. The goodness of fit of the final GEE model was 0.761 (deviance value ⁄ degree of freedom) as assessed by deviance statistics (Table 4). Female Gender Because females were observed to relapse more than men, we examined this group in greater detail. At presentation, the proportion of females with CTAS 1–2 was similar to males (19.0% vs. 18.1%, p = 0.796); however, the proportions of women already taking oral CS (9.8% vs. 4.1%, p = 0.019) and inhaled CS (34.2% vs. 22.2%, p = 0.004) were higher than men. When discharged, 72.3% of the females were sent home on oral CS and 58.0% on inhaled CS. DISCUSSION Asthma relapses after ED discharge represent important negative outcomes for patients and the health care system, and attempts to reduce them are worthy of continued research efforts.2,9,10,16 While there are treatment approaches that are known to reduce relapses, such as the use of systemic17 and inhaled18 CS, there is a general lack of valid evidence for EP decision-making regarding nontreatment factors associated with relapse. This study attempted to identify the key factors associated with relapse and identified demographics (ethnicity, female gender), markers of poor control (previous ED visits and admissions for asthma), and current usage of preventer medication (combined inhaled CS and long-acting b2-agonists and oral CS) as being strongly associated with relapse. While these factors require validation, they do indicate that key markers at presentation may predict outcomes and suggest that interventions, such as more aggressive treatment and follow-up, may be targeted at such individuals. Even though the relapse rate in this study is lower than that reported in other North American centers (13.9% vs. 17%),2,9,10 it confirms that a considerable proportion of patients present for additional urgent care within days to weeks after ED discharge. It is likely that other factors specific to the Canadian health care system account for some of this difference. For example, Canada has universal health coverage and has demonstrated adoption of national asthma guidelines with evidence-based treatments, including systemic and inhaled CS at discharge. Other factors, such as the selection of patients enrolled in
714
Rowe et al.
•
ASTHMA RELAPSE
Table 2 Acute Asthma Presentation and ED Course Categorized by Relapse Status Characteristics CTAS 1, 2 3 4,5 Not documented Duration of symptoms Less than 24 hours 24+ hours No. of inhaled b2-agonist puffs within 24 hours of ED Vital signs Pulse Respiratory rate SaO2 Temperature (C) ED course Received inhaled b2-agonists in first hour No. of treatments Received inhaled b2-agonists over ED stay No. of treatments Given any corticosteroid treatment Received inhaled anticholinergics in first hour No. of treatments Received inhaled anticholinergics over ED stay No. of treatments MgSO4 medication in the ED ED length of stay (hours) ED length of stay > 6 hours Lung function Earliest PEF* Final PEF Change in PEFà Earliest %predicted PEF§ Final %predicted PEF– Change in %predicted PEF||
Relapse (n = 74)
No Relapse (n = 455)
p-Value
3 (3, 4) 14 (18.9) 36 (48.6) 18 (24.3) 6 (8.1)
3 (3, 4) 85 (18.6) 235 (51.6) 108 (23.7) 27 (5.9)
0.963
5 (6.7) 69 (93.2) 8 (2, 14)
41 (9.0) 414 (90.0) 8 (1, 14)
0.523
100.9 ± 17.5 22.6 ± 4.7 95.9 ± 2.8 36.7 ± 0.6 46 (62.1) 1 (0, 2) 68 (91.8) 3 (2, 5) 53 (71.6) 43 (58.1) 1 (0, 2) 65 (87.4) 3 (2, 4) 1 (1.3) 3.8 (2.0, 6.0) 19 (26.0) 299.5 ± 118.2 406.1 ± 165.5 136.4 ± 125.9 64.4 ± 22.0 85.5 ± 33.0 29.4 ± 26.2
99.9 22.9 95.9 36.6
± ± ± ±
18.3 6.2 3.4 0.6
0.424 0.638 0.849 0.619 0.138
300 (65.9) 1 (0, 2) 439 (96.4) 3 (2, 5) 360 (79.3) 266 (58.4) 1 (0, 2) 385 (84.6) 3 (1, 4) 10 (2.2) 3.4 (2.4, 4.0) 65 (14.0)
0.786 0.858 0.054 0.868 0.290 0.489 0.862 0.542 0.382 0.636 0.317 0.011
287.9 ± 110.9 425.8 ± 178.3 159.1 ± 163.8 59.7 ± 23.1 88.5 ± 38.0 33.3 ± 34.0
0.704 0.438 0.374 0.115 0.930 0.489
Data are reported as number (%), mean ± sd, or median (IQR). Each nebulizer treatment was counted as equivalent to six ‘‘puffs’’ from a metered-dose inhaler. CTAS = Canadian Triage and Acuity Score; ED = emergency department; IQR = interquartile range; MgSO4 = magnesium sulfate (IV); PEF = peak expiratory flow; SaO2 = oxygen saturation; SD = standard deviation. *Available for 425 patients. Available for 485 patients. àAvailable for 397 patients. §Available for 412 patients. –Available for 468 patients. ||Available for 385 patients.
this study, could also be responsible for this variation and this should be taken into account when interpreting our results. Additionally, few patients withdrew from the study (12.5%) during the follow-up period, further strengthening our conclusions. We do not feel these dropouts biased the findings since the gender distribution of the study population was similar, and the ages were only marginally younger (28 ± 8.8 years). Given these observations, and the widespread availability of new asthma treatments, relapse events are still surprisingly common. Based on previous findings among Canadian EDs19 and the higher percentage of women presenting with asthma relapse after treatment and discharge from the ED, EPs should consider using these demographic data to stratify risk when formulating follow-up strategies. Our results are consistent with suggestions from other studies that women with asthma experience more severe symptoms that may result in more frequent
relapses.20 Moreover, it has been described that women have a phenotype of asthma that is more resistant to controller medications such as inhaled CS.21 While future mechanistic research will likely contribute to our understanding of these differences, in the mean time women discharged from the ED should be considered at risk for relapse. Historical severity was also strongly related to relapse. ED and urgent clinic visits within the prior months and 1 year previously have been found to be associated with a higher incidence of relapse.2,10,22 This suggests that patients with more unstable disease are likely to take longer to regain control and may relapse even though appropriate care is delivered. Diseaserelated factors, such as functional limitation and current inhaled CS, were strongly associated with relapse in univariable analyses.9,10,22 In our study, however, these two factors were not retained in the final multivariable model.
ACAD EMERG MED • August 2008, Vol. 15, No. 8
•
www.aemj.org
Table 3 Discharge Medication and Two-week Follow-up Categorized by Relapse Status Relapse (n = 74)
Characteristics Sent home Taken as Sent home Taken as
on OCS prescribed on ICS prescribed
Sent home with OCS alone Sent home with ICS alone Sent home with OCS and ICS combined Sent home with OCS and ICS and LABA or antileukotrienes Sent home with other combinations None of these
54 48 ⁄ 53 39 29
(72.9) (90.5) (52.7) (74.3)
No Relapse (n = 455) 325 301 ⁄ 322 272 219 ⁄ 272
p-Value
(71.4) (93.4) (59.7) (80.5)
0.785 0.633 0.256 0.254
21 (28.3)
112 (24.6)
0.224
8 (10.8)
67 (14.7)
28 (37.8)
197 (43.3)
3 (4.0)
6 (1.3)
3 (4.0)
15 (3.3)
11 (14.8)
58 (12.7)
Data are reported as number (%) ICS = inhaled corticosteroids; OCS = oral corticosteroids; LABA = long acting b2-agonists.
Table 4 Predictors of Relapse after ED Discharge from Multivariable Modeling Domains
Factors
Demographics
Ethnicity (white) Female gender Any ED visits in the past 2 years Ever admitted for asthma Combined ICS plus LABA at ED presentation OCS at ED presentation
Asthma history Treatment
RR (95% CI)
p-Value
0.66 (0.52, 0.83) 1.57 (1.14, 2.09) 1.47 (1.18, 1.80)
