The Diagnosis of Potentially Fatal Asthma in Hospitalized Adults

The Diagnosis of Potentially Fatal Asthma in Hospitalized Adults* Patient Characteristics and Increased Severity of Asthma Th011UlS R Miller, M.D.; Paul A. Greenberget; M.D., F.C.C.R; and Roy Patterson, M.D., F.C.C.R

We evaluated various patient characteristics in patients hospitalized for asthma during 1987 to 1990. Potentially fatal asthma was identified in 26 of 87 adult patients (29.9 percent) hospitalized. Patients with PFA had increased frequency of prednisone use prior to hospitalization (p
diagnosis of PFA identifies a higher risk patient with asthma. The data suggest that at the time of hospitalization the PFA patient has had a shorter recognized prodrome of increased respiratory symptoms, reduced peak expiratory Row rates and greater likelihood of major psychiatric disease or noncompliance. Effective ambulatory control of PFA and non-PFA is advisable with earlier use and higher dosages of oral corticosteroids. (Chest 1992; 102:515-18)

A sthma has not always been recognized as a fatal 1'1. disease. Sir William Osler taught that asthmatic

revealed a cumulative mortality rate of 7.1 percent.23 These criteria identify patients at high risk ofincreased mortality compared with literature mortality rates of at least 1.7 per 100,000 population}2,33-36 The goals of this study were twofold: (1) to evaluate various patient characteristics in a hospitalized asthma patient population and estimate the frequency of PFA, and (2) to determine whether the clinical diagnosis of PFA was associated with increased disease severity and morbidity in surviving patients.

patients don't die, they just pant into old age} It has only been during this century that it has been widely accepted that asthma can be fatal. 2-6 Over the last decade there has been an increase in the prevalence and mortality of asthma in children and adults. 7- 1O This increased mortality has been accompanied by increased hospitalization rates. ll ,12 These trends have not only been observed in the United States but also worldwide. 13-15 Various patient characteristics and comorbid factors have been postulated to account for asthma mortality.I6-24 These include inadequate assessment with resultant underappreciation of the severity of the disease by patient, family or physicians, and thus undertreatment,22-26 the overuse of ~ agonists,27,28 noncompliance,22,23,26 and inconsistency of care.29-31 Certain psychosocial problems 22 ,23,26 or even geographic location of primary residenceS or other demographic factors,32 have been associated with increased mortality. In 1988, criteria to identify high risk patients were proposed to establish the diagnosis of PFA.22 In the original series, the mortality rate was 5.5 percent over 4.8 ± 4.3 years of follow-up. A more recent follow-up ofthese patients with some additional patients for an additional two years of follow-up *From the Division of Allergy-Immunology, Department of Medicine, Northwestern University Medical School, Chicago. Supported by USPHS grant AI11403 and the Ernest S. Bazley Grant to Northwestern University and Northwestern Memorial Hospital. Manuscript received August 19; revision accepted November 27. Reprint requests: Dr. Miller; Allergy-Immunology, 303 East Chicago Avenue, Chicago 60611

PFA = potentially fatal asthma; pH = negative logarithm of hydrogen ion activity

METHODS

Patient Identification A eomputer-generated list of patients ages 45 years and less who were discharged with the diagnosis of asthma from an adult tertiary care hospital (Northwestern Memorial Hospital, Chicago, Ill) was obtained. The inclusive dates were October, November and December for the years 1987, 1988, 1989 and 1990 because of the increased number of respiratory infections and asthma exacerbations during those months in the Chieagoland region. 37 Patients included needed to have the principal diagnosis of asthma without status asthmaticus (493.90) or asthma "'ith status asthmatieus (493.91). No other diagnoses were included (acute or chn>nic unspecified bronchitis, pneumonia, ete). There were 104 patient cases which fulfilled these criteria. Of these, eight records were miscoded (such that the admission was obviously not for an exacerbation of asthma) and nine reeords could not he located. This left 87 cases suitable for evaluation to be reviewed.

Criteria for Potentially Fatal Asthma The patients were categorized as having PFA22 or not having PFA from data prior to the index admission. A history of anyone of the follo\\;ng four criteria fulfilled the diagnosis of PFA: (1) intubation for respiratory failure or respiratory arrest, (2) respiratory acidosis without intubation (pH <7.35), (3) two or more hospitalizations despite administration of oral eorticosteroids or (4) two or more episodes of pneumothorax or pneumomediastinum associated with status asthmaticus (Table 1). CHEST I 102 I 2 I AUGUS'T, 1992

515

Table I-Criteria of PotentiaUy Fatal Asthma in the Study Population (n = 26)

Criteria 1 Intubation 2 Respiratory acidosis without intubation 3 ~2 hospitalizations despite administration of oral steroids 4 ~2 pneumothorax/pneumomediastinum with status asthmaticus 2and3 1 and 3 1,2 and 3

No. of Patients with Criterion 7 1 7

o 2

7

2

lbrameters Evaluated 1h determine disease severity and morbidity, various parameters were evaluated including duration of symptoms prior to hospitalization and the length of hospitalization. The duration of symptoms prior to hospitalization was defined as increased coughing or wheezin~ noted by the patient, above the patient's baseline level of symptoms. This information frequently was obtained from the admission history as noted by the medical student, medicine intern, resident and attending physician or specialty consultant (AllergyImmunolo~ or Pulmonary). If there were differences in terms of duration of symptoms, the most detailed evaluation from the most senior physician was used. Medications being used at the time of admission, including corticosteroids, were recorded. Peak 80\\' values and arterial blood gas levels were noted \\,hen obtained and recorded in the medical records. Some patients had two values recorded: one initial value and the next after varying amounts of therapy. One patient had three values recorded. A mean value for each patient while in the emergency room was calculated. Because of the varying amounts of therapy the patients had received at the time of the peak Row measurement, this value is a rough estimate. Many of the PFA patients who did not have a peak Row checked had an arterial blood gas value determination and the decision to hospitalize the patient would not have heen impacted by the peak Row measurement. Other variables used to reveal patient characteristics were evaluated. This information, such as past medical history, included comorbid medical diseases as well as major psychiatric illness, tobacco, drug or alcohol abuse and documentation of noncompliance in the past. In order to be included as having major psychiatric illness, the patient needed to have been given the diagnosis after a psychiatric evaluation at some time in the past. Patients who were taking an antidepressant or anti-anxiety medication without having had psychiatric evaluations were not counted in Table 2. Current documentation of noncompliance also was noted, such as a patient who was receiving a theophylline preparation and had a serum concentration level of <0.5 mgIL in the emergency room. To be defined as noncompliant, patients needed to have current evidence of noncompliance with medication or to have specific documentation of noncompliance noted in the chart. Standard demographic information including age, race and gender was noted as well as the location of the patient's residence (city or suburb). Statistical Analysis Various statistical computations were employed, depending on the type of data and sample size being analyzed. For the PFA patient characteristics including the demographic data and medical profile (except age), a chi square calculation was obtained unless one of the subsets contained five or fewer patients in which the Fisher's exact test was used. For age, a mean, standard deviation and standard error of the mean were calculated and an unpaired Student's t test used.

516

For the variables used as a predictor of disease severity a chi square or Fisher's test was used as described previously such as presence of prednisone at presentation, and previous hospitalizations (whether multiple, fe\\' or none). For peak Row, a mean, standard deviation and standard error of the mean were calculated and an unpaired Student's t test used. The distribution of values for duration of symptoms prior to hospitalization and length of hospitalization were nonparametric so a Mann-Whitney test was employed. RESULTS

There were 87 cases suitable for evaluation consisting of 61 non-PFA and 26 PFA cases. No asthma fatalities occurred during the study period. The number of criteria for the diagnosis of PFA in 26 cases is presented in Table 1. A comparison of PFA and nonPFA cases is listed in Table 2. Most patients were receiving 12-h theophylline preparations and inhaled ~-adrenergicagonists. In the 26 PFA patients, 22 (84.6 percent) were using theophylline and 20 (76.9 percent), ~-adrenergic agonists prior to hospitalization. Of the 61 non-PFA patients, theophylline was being utilized in 40 (65.6 percent) and ~-adrenergicagonists in 43 (70.5 percent). As shown in Table 2, few patients were using inhaled corticosteroids. Patients with PFA were characterized by a statistically increased number of patients receiving prednisone (p
The morbidity and mortality from asthma have been increasing over the last decade. 7-10 Also ofinterest is the observation that the asthma mortality rate is disproportionately higher in specific urban areas. 8 The reasons for this are not clear and are most likely multifactoria1. 8 In 1988, four diagnostic criteria were proposed identifying patients with PFA.22 Based on observations from subsequent clinical experience, these high-risk patients required intensive anti-asthma therapy that usually consisted of high-dose long-term Diagnosis of Potentially Fatal Asthma (Miller, Greenberger, Patterson)

Table 2-Potentially Fatal Asthma: Characteristics and Morbidity Parameters evaluated for: Patient characteristics Gender (M/F) Race (Caucasian/ noncaucasian) Age (mean) Place of residence (city/ other) Tobacco use Drug or aleohol abuse Major psyehiatrie disease Noneomplianee Receiving inhaled cortieosteroids Disease severity/morbidity Prednisone administration on admission Duration of symptoms before hospitalization (days) Length of hospitalization (days) Multiple (~3) previous hospitalizations for asthma No previous hospitalizations Peak How value recorded in emergeney room (Umin)

PFA (n=26)

non-PFA (n=61)

p Value

4/22

14/47

NS

9/17 28.2

12/32 31.1

NS NS

2213 9 3 5 11

49/10 25 7 3 11

NS NS NS <0.05 <0.05

8

9

NS

14

6


3.2

8.3


3.7

3.0


22 0

6 37

<0.0001 <0.0001

184

222

<0.05

inhaled corticosteroids and daily or alternate-day doses of prednisone. 22.23 Other factors such as reasons for noncompliance with medical advice have been discussed. 22 The major factors were patient noncompliance and associated psychiatric disease. 22 It \\'as also believed that the underutilization of corticosteroids, whether inhaled or oral, may be contributing to increased patient morbidity and mortality. 22 In this series of patients, noncompliance and the presence of major psychiatric disease were associated with the diagnosis of PFA. The question of whether these factors were independent risk factors for increased severity of disease was analyzed. There was no significant association between psychiatric disease or patient noncompliance with any measure of disease severity with fe\\' exceptions. Compliant patients were more likely never to have been previously hospitalized (p = 0.(018), and noncompliant patients showed a higher rate of having one to two previous hospitalizations (p = 0.(08). Nevertheless, there was no difference for compliance or noncompliance in regard to multiple (~3) previous hospitalizations (p = 0.825). This observation suggests that noncompliance is a contributor to severity of disease but not an independent risk factor, especially for severe asthma. This finding also could be interpreted as evidence that the outpatient regimens for severe asthmatic patients were inadequate in that whether the patient was compliant with their

regimen or not anti-asthma medications did not prevent multiple hospitalizations. The inadequacy of the outpatient regimens these patients were receiving also is supported by the fact that only eight PFA patients and nine non-PFA patients were receiving inhaled corticosteroids. The use of inhaled corticosteroids has been recolnmended as an integral part for all moderate to severe asthmatic patients and in the National Heart Lung Blood Institute guidelines. 3H •39 There is also recent evidence supporting the use of inhaled corticosteroids as first-line treatment. 40 The ability to predict which patients are at increased risk for future asthma morbidity and mortality is needed and a laudable goal. The data in this series reveal that the diagnostic criteria of PFA does identify patients with increased asthma severity and morbidity in surviving patients. Of particular interest is the fact that patients with PFA deteriorated to the point of requiring hospitalization much more quickly than the non-PFA group (3.2 days vs 8.3 days, p<.OOI). Although prehospital daily examinations and serial spirometric values were not obtained, and the number of prehospitalization days of increased asthma symptoms is an approximation, even prior oral corticosteroid use did not prevent all hospitalizations. The use of short courses of prednisone in addition to regular inhaled corticosteroids has been shown to reduce readmissions for asthma. 41 These data suggest that earlier use of effective dosages of oral corticosteroids most likely would have reduced the number of hospitalizations in compliant patients in this series. Emergency room assessment using peak flow rates or arterial blood gas value determinations were obtained in 15 of 25 patients (60 percent) with PFA, whereas one additional patient required immediate intubation upon presentation. Similarly 41 of 61 (67.2 percent) non-PFA patients had peak flow values recorded. It is unclear why more patients did not have either measurement performed; however, in these patients we could not document any untoward clinical outcomes resulting from their absence. Management of patients with PFA requires its diagnosis, subsequent intensive pharmacotherapy, consideration of allergic factors as triggers of asthma, and recognition of psychologic, social and economic factors that influence patient outcomes. 22 .23 Deaths from asthma in patients fulfilling the criteria for PFA are not invariable but require physician expertise and patient compliance with medical advice. The findings of this study are consistent with the notion that PFA patients have a more aggressive or serious type of asthma resulting in longer hospitalizations, shorter recognized prodrome of respiratory symptoms before requiring admission, lower peak expiratory flow rates and far greater likelihood of prior asthma hospitalizations. The current hospitalizations CHEST / 102 / 2 / AUGUS'T, 1992

517

in PFA patients occurred despite more frequent oral corticosteroid administration before admission. N either noncompliance nor major psychiatric illness were independent risk factors for asthma severity. The shorter prodrome of increased respiratory symptoms is very important in its clinical implications. It points to the fact that PFA patients require very aggressive and prompt treatment with oral steroids to prevent the rapid deterioration of respiratory status. In a 12-month prospective study of 225 patients with asthma who received either uusual care" or an intensive self-management, patient education and counseling program, the number of hospitalizations and emergency department visits were reduced by approximately 69 percent in both groups.42 These data could be interpreted as demonstrating the value of an initial evaluation for asthma, enrollment in a study and continuity of care. Prevention of asthma admissions and hopefully fatalities requires effective physician management and patient compliance. The diagnosis of PFA identifies a higher-risk group of subjects with asthma as compared with lower-risk patients with asthma. Effective control of asthma is advisable so that the patients have as normal a respiratory status as possible. Prevention of hospitalization in patients with and without PFA most likely also requires the use of inhaled corticosteroids on a regular basis and the earlier use and perhaps higher dosages of prednisone or other oral corticosteroids in addition to other anti-asthma modalities. 38-41 REFERENCES 1 Osler W The principles and practice of medicine. 4th ed. Edinbur~h: Pentland, 1901 2 Huber ML, Koessler KK. The pathology of bronchial asthma. Arch Intern Med 1922; 30:689-760 3 Bullen SS Sr. Correlation of clinical and autopsy findings in 176 cases of asthma. J Allergy 1952; 23:193-203 4 Williams DA. Deaths from asthma in England and Wales. Thorax 1953; 8: 137-40 5 Houston JC, de Navasquez S, Trounce JR. A clinical and pathological study of fatal cases of status asthmaticus. Thorax 1953; 8:207-13 6 Earle B\Z Fatal bronchial asthma: a series of fifteen cases with a review of the literature. Thorax 1953; 8:195-206 7 Gergen PJ, Mullally DI, Evans R. National survey of prevalence of asthma amon~ children in the United States, 1976 to 1980. Pediatrics 1988; 81: 1-7 8 \\eiss KB, Wagener DK. ChanJtin~ patterns of asthma mortality, identifyin~ target populations at hi~ risk. JAMA 1900; 264:1683-87 9 Buist AS, Vollmer WM. Reflections on the rise in asthma morbidity and mortality. JAMA 1990; 264: 1719-20 10 Sly RM. Increases in deaths from asthma. Ann Allergy 1984; 53:20-25 11 Gergen PJ, Weiss KB. Changing patterns of asthma hospitalization among children: 1979 to 1987. JAMA 1990; 264:1688-92 12 Robin E. Death from bronchial asthma. Chest 1988; 93:614-18 13 Woolcock AJ. Worldwide differen<..-es in asthma prevalence and mortality: why is asthma mortality so low in the USA? Chest 1986; 90:40-45 14 Sears MR, Beaglehole R. Asthma morbidity and mortality: New

518

Zealand. J Allergy Clin Immunol1987; 80:383-88 15 Jackson RT, Beaglehole R, Rea HH, Sutherland DC. Mortality from asthma: a new epidemic in New Zealand. Br Med J 1982; 285:771-74 16 Benatar SR. Fatal asthma. N Engl J Med 1986; 314:423-29 17 Buist AS. Asthma mortality: what have we learned? JAMA 1989; 84:257-83 18 Baniot Rioo B. Prevention of fatal asthma. Chest 1987; 92:460-66 19 Evans R III. Recent observations reflecting increases in mortality from asthma. J Allergy Clin Immunoll987; 80:377-79 20 Cushley MJ, Tattersfield AE. Sudden death in asthma: discussion paper. J Royal Soc Med 1983; 76:662-66 21 Rubinstein S, Hindi RD, Moss RB, Blessing-Moore J, Lewiston NJ. Sudden death in adolescent asthma. Ann Allergy 1984; 53:311-18 22 Greenberger PA, Patterson R. The diagnosis of potentially fatal asthma. NER Allergy Proc 1988; 9:147-52 23 Walker CL, Greenberger PA, Patterson R. Potentially fatal asthma. Ann Allergy 1990; 64:487-93 24 Strunk RC, Mrajek DA, Fuhrmann GS~ LaBrecque JF. Physiologic and psychological characteristics associated with deaths due to asthma in childhood: a case controlled study. JAMA 1985; 254:1193-98 25 Sears MR. Why are deaths from asthma increasing? Eur J Respir Dis 1986; 69(suppl 147): 175-81 26 Rea HH, Scrogg R, Jackson R, et aI. A case control study of deaths from asthma. Thorax 1986; 41 :833-39 27 Sears MR. Increasing asthma mortality fact or artifact? J Allergy Clin Immunoll988; 82:957-60 28 Boushey HA, Nichols J. Asthma mortality (medical staff conference). West J Med 1987; 147:314-20 29 Neuspiel DR, Kuller LH. Sudden and unexpected natural death in childhood and adolescence. JAMA 1985; 254: 1321-25 30 Mak H, Johnston Abbey H, Talamo RC. Prevalence of asthma and health service utilization of asthmatic children in the inner city. J Allergy Clin Immunol 1982; 70:367-72 31 Death due to chronic obstructive pulmonary disease and allied conditions. MMWR 1986; 35:507-10 32 Stewart CJ, Nunn AJ. Are asthma mortality rates changing? Br J Dis Chest 1985; 79:229-34 33 Rackemann FM, Edwards MC. Asthma in children: a follow-up study of 688 patients after an interval of twenty years. N Engl J Med 1952; 246:815-23,858-63 34 Blair H. Natural history of childhood asthma: 20 year follow-up. Arch Dis Child 1977; 52:613-19 35 Ogilvie AG. Asthma: a study in prognosis of 1,000 patients. Thorax 1962; 17:183-89 36 McCracken D. Prognosis in bronchial asthma. Br Med J 1950; 1:409-12 37 Cooper BJ, Patterson R. The corticosteroid dose graph. J Allergy Clin Immunol 1976; 58:635-46 38 National asthma education pr()~ram: expert panel report. Guidelines for the diagnosis and management of asthma. J Allergy Clin Immunoll991; 88(suppl):451-59 39 Hargreave FE, Dolovich J, Newhouse MT. The assessment and treatment of asthma: a conference report. J AlIerJO' Clin Immunoll990; 85:1098-1111 40 Maahtela T, Jarvinen M, Kava T, Kiviranta K, Koskinen S, Lehtonen K, et aI. Comparison of a ~2-agonist, terbutaline, with an inhaled corticosteroid, budesonide, in newly detected asthma. N Engl J Med 1991; 325:388-92 41 Mayo PH, Richman J, Harris MW Results of program to reduce admissions for adult asthma. Ann Intern Med 1990; 112:864-71 42 Bailey we, Richards JM, Brooks M, Soong S, Windsor RA, Manzella BA. A randomized trial to improve self-mana~ement practices of adults with asthma. Arch Intern Med 1990; 150: 1664-68

e

e

Diagnosis of Potentially Fatal Asthma (Miller. Greenberger. Patterson)