RSS Feeds
Observations on Asthma Mortality
- E.R. McFadden Jr., MD; and
- Edward L. Warren, MD
- From University Hospitals of Cleveland and Case Western Reserve University School of Medicine, Cleveland, Ohio. Grant Support: In part by Specialized Center of Research grant HL 37117 from the National Heart, Lung, and Blood Institute; the General Clinical Research Center grant MO-RR00808 from the National Center for Research Resources; and Pathophysiology of Exercise Induced Asthma RO 1 grant HL 33791. Requests for Reprints: E.R. McFadden Jr., MD, Division of Pulmonary and Critical Care Medicine, University Hospitals of Cleveland, 11100 Euclid Avenue, Cleveland, OH 44106-5067. Current Author Addresses: Drs. McFadden and Warren: Division of Pulmonary and Critical Care Medicine, University Hospital of Cleveland, 11100 Euclid Avenue, Cleveland, OH 44106-5067.
Abstract
The rate of death from asthma in the United States has been increasing over the last decade, but such deaths still remain uncommon.Mortality rates differ markedly by region and ethnicity, and case-fatality rates are highest in black men living in inner cities. In general, patients in greatest jeopardy are those with severe, unstable disease who are not being objectively monitored; however, death can occur in anyone if the attack is intense enough. Sudden catastrophic episodes of asthma occur but are very rare. Failure to recognize the seriousness of the terminal episode or to treat the episode appropriately remains the chief contributing cause of poor outcome. Little compelling evidence shows that adverse effects of medications play much of a role in asthma-related death. The disease characteristics that place patients at risk remain inadequately defined. Of those suggested, only a history of recurrent hospitalization and the need for ventilatory assistance are specific enough to be helpful. These characteristics, however, are found in only 36% and 6% of cases, respectively; thus, their absence is of no value in assessing risk. The prognosis after a near-fatal episode of asthma is poor, and approximately 10% of patients die in the year after the event. Given the current state of knowledge, it is wise to view all exacerbations of asthma that last longer than a few days as potentially fatal and to treat them accordingly. This is especially true in patients who have previously had a severe episode of asthma.
The subject of asthma-related death seems straightforward. Death is a readily recognized end point, and asthma is an easily diagnosed disease. However, because no prospective studies on the subject have been done, deductions about prevalence and causality must be drawn from retrospective analysis of death certificates, medical charts, and post hoc interviews with relatives or care givers [1-19]. Although such information leads to sobering considerations, it also contains inherent limitations. For example, disease-specific deaths must be analyzed in the context of the overall mortality rate in the region of interest, but this type of analysis has never been done. If the general longevity of a population is low, then an increase in asthma-related mortality rates may be only a marker of inadequate health care rather than a unique problem. Furthermore, changes in diagnostic criteria and inaccuracies in death certificates may also confuse the statistics. Such problems clearly exist in the United States and may have artifactually increased mortality rates by 35% [20-23].
In the past 30 years, asthma-related case-fatality rates in industrialized countries have varied from 1.5 to 8.5 per 100 000 persons [2, 24]; rates in the United States and Canada have consistently been at the lower end of this range. The rate of deaths increased in New Zealand, Australia, and the United Kingdom in the 1960s [25] and increased again in New Zealand a decade later [1]. In the United States, mortality statistics have cyclically changed, but the overall pattern for the 20th century has been flat [26]. In the past decade, the cycle was upward and the rate of asthma-related deaths increased from 1.3 to 1.9 per 100 000 [27]. Although this represents a large percentage increase, the actual number of patients dying is still small; most patients with asthma have nothing to fear. In fact, asthma has little impact on expected longevity [28]. To put the matter into perspective, in 1991 10.3 million persons had the disease and 5106 died of it [27]. Thus, 1 of every 2000 persons at risk and fewer than 20 of every 1 million citizens would have been affected. Such numbers clearly are not staggering, especially when contrasted with the 85 000 to 100 000 persons who die each year of chronic bronchitis and lung carcinomas.
Demographic Characteristics
Asthma-related mortality rates differ markedly by region and ethnicity. Persons of all races in all locations die of asthma, but four areas- New York, New York; Cook County, Illinois; Maricopa County, Arizona; and Fresno County, California-dominated the statistics in the 1980s [20, 29]. New York and Cook County contained only 6.5% of the 5- to 30-year-old persons in the United States, but 21.1% of all asthma-related deaths in this age group occurred in these areas. Other urban areas have similar difficulties [30]. The reasons for this situation are unclear. Many U.S. cities have large minority populations living in poverty, in whom the prevalence and severity of asthma is high [20, 27, 31, 32]. During the past decade, the rate of asthma-related death was consistently greater in black persons than in white persons nationwide; male persons were affected most [27]. Neither penury nor ethnicity, however, is the sole explanation. Elevated death rates are found in large cities [20, 30, 31, 33] but not in states with very low per capita income (such as Mississippi, Louisiana, and West Virginia), and the mortality rate in educated African-American persons is generally lower than that in less educated African-American persons [20, 34-36]. The social conditions in inner cities may increase risks for exposure and produce more aggressive disease. Because impoverished persons are less likely to seek or have access to high-level continuous medical care [35], a vicious cycle of neglected illness with partial recovery develops until the patient's reserves are severely compromised.
Causes of Life-Threatening Episodes
The events most often associated in the literature with fatal or near-fatal episodes of asthma are profound emotional upsets, air pollution, therapeutic misadventures, and insufficient assessment of the seriousness of the final episode [8-19, 37-41]. The last is the most significant. Although data indicate that psychological problems are more prominent among selected patients who have died than in controls [37], the applicability of these observations to all patients with asthma is uncertain. The role played by human-made air pollutants has temporally decreased with the enforcement of rigid air standards. However, seasonal elevations in natural particulates, such as outdoor molds, may still be important and may contribute to respiratory arrests in some cases [42, 43]. The roles of indoor pollution and occupational exposures await elucidation.
Deaths from tranquilizers, sedatives, β-blockers, and nonsteroidal anti-inflammatory agents are well documented [39-41], but the effects of other drugs remain problematic. Since the epidemic of asthma-related deaths in the 1960s [25], the potential adverse consequences of sympathomimetic agents have been a recurring concern. Most recently, studies from New Zealand have resurrected the possibility that adverse effects occur with a specific agent, fenoterol [1, 44-46]. As in the original epidemic, however, the data are contentious [1, 46, 47]. No temporal coincidence was seen between alleged cause and effect, and other countries in which the drug was used did not experience the same events [47]. Thus, the association noted may simply have resulted from the marketing of fenoterol as the agent of choice for severely ill patients. Of note, no class effect of sympathomimetic agents and death in the New Zealand experience has been found [1, 44-48].
A Canadian investigation purported to show a relation between asthma-related death and the amount of β2-agonist used [49, 50]. Consumption of two or more canisters of fenoterol or albuterol per month was originally said to be associated with an increased risk for death [49]. The limit was subsequently lowered to 1.4 canisters [50]. These reports received much attention in both the medical and lay press, with little dissent. Our analysis, however, raised major concerns [51]. For example, the amount of medication claimed to be deadly actually represents little more than a few days' worth of the recommended oral doses [51]. Equally important, use of β2-agonists by means of hand-held nebulizers did not confer increased risk, but the standard dose of albuterol administered by this route equals 15 canisters per month. Few experienced physicians would accept these inconsistencies. Finally, the magnitude of the problem was overstated. Only 46 deaths occurred among the 12 301 patients evaluated; the number was expanded by incorporating 85 imprecisely defined “near deaths” (that is, cases in which the PaCO2 value exceeded 45 mm Hg or cases that required intubation). Ten percent of adults and children with asthma who seek care for acute episodes in emergency departments meet the first criterion and have no risk for death at all [52, 53]. Moreover, a single PaCO2 value does not predict the course of a patient with asthma [54, 55]. Elevated levels decrease with appropriate treatment and usually do not necessitate assisted ventilation [52-55]. Finally, intubation cannot be used as a surrogate for “near death” without objective documentation that the procedure was performed solely because of impending suffocation. Too often, this is not the case [56]. These points of contention undoubtedly account for the small effect that the Canadian observations had in a meta-analysis [57] and for the lack of concurrence between these observations and those from New Zealand [1], Sweden [58], or the United Kingdom [59].
Several studies have shown a pattern of failure on behalf of either the patient or health care professionals to recognize the severity of the final episode of asthma [8-19]. In fairness, such data need to be tempered by the realization that these investigations were uncontrolled and were conducted by small groups of physicians who retrospectively judged events against an arbitrary standard of care. Nonetheless, the findings are disquieting. The British Thoracic Society panel believed that some aspect of care or supervision was deficient in 98% of the deaths reviewed [11]. The problems ranged from failure to diagnose asthma (10%) to gross undertreatment (92%). Patients had not acted appropriately 77% of the time. Other investigations found similar patterns [8-10, 12-19].
It is commonly believed that most deaths from asthma occur suddenly at home or work, suggesting that even if recognized, the deaths may not have been preventable [11, 18, 24]. In reality, slightly fewer than half of all deaths occur in a hospital [3, 4, 8-1013, 15, 16, 18, 19, 60]; in nearly 85% of deaths, the final episode lasted at least 12 hours, which would have allowed ample time for treatment [5, 9-1316, 17, 19, 60, 61].
Why should patients with a reversible disease die in a hospital? A disappointing reality is that here, too, the answer often lies in the care given. Many surveys have shown inadequate or incomplete assessments of patients with asthma in both outpatient [62, 63] and inpatient settings [8-1015, 16], especially in persons who died. By relying only on clinical acumen, practitioners tragically underestimated the morbid potential of the attacks. Specialists also make similar mistakes. An evaluation of 140 patients with asthma who were admitted to thoracic medicine or general medicine wards revealed that arterial blood gases were measured in just 58% of patients, even though they were hospitalized for acute respiratory impairment [64]. On the general medicine ward, this percentage was an appalling 27%. Only 57% and 8% of patients in the thoracic and general medicine wards, respectively, had objective assessments done of the severity of their obstruction. Thirty-six percent of the patients did not receive corticosteroids or were given a single bolus of corticosteroids. One could argue that such observations are no longer valid because of improved awareness of the deficiencies in asthma care. Yet, a recent study showed that underassessment and undertreatment remain common problems [65].
Sudden Unexpected Deaths
Most deaths occur in patients with severe, poorly controlled disease whose condition gradually deteriorates over days or in patients who have a history of unstable illness and are partially responsive to treatment, upon which a major attack is superimposed [9, 11, 13, 15, 18]. Rarely, death can be rapid and unexpected, so-called sudden-onset fatal asthma [11, 18, 66-70]. The usual history is the abrupt development of poorly responsive symptoms that quickly progress to extreme respiratory embarrassment. If ventilatory support is forthcoming, resolution is rapid and normal gas exchange and pulmonary mechanics are restored in hours rather than days. The clinical picture suggests airway anaphylaxis with either overwhelming smooth-muscle spasm and edema involving the entire tracheobronchial tree or the presence of strategically located obstruction in the major conducting airways. With sudden-onset asthma, there does not seem to be time to develop the sequelae associated with the cellular infiltration, mucociliary transport impairments, and mucosal denudation seen with more subacute inflammatory processes [71, 72].
Save for the speed of onset and recovery, the physiologic abnormalities of these patients mirror those of other patients with life-threatening illness [56, 73-75]. The distinguishing features are pathologic. At autopsy, the airways often lack inspissated secretions [71, 72] and contain more neutrophils than eosinophils in the submucosa [72]. Although these findings are provocative, they have been reported in only three patients. Consequently, the precise biological characterizations of this facet of the asthmatic diathesis may not yet be completely elucidated. True sudden asphyxic asthma is uncommon. In one investigation, nine cases of ventilatory arrest and sudden death occurring in the hospital were found among a total of 1169 consecutive admissions for asthma [68]. In another study, 10 patients were seen in a decade, only 5 of whom had respiratory arrest [70]. Clearly, additional investigation is needed before we will understand or be able to prevent this phenomena.
Patients experiencing near-fatal or fatal episodes of asthma have marked hypercapnic ventilatory failure with respiratory acidosis [56, 70, 73-75]. Many also have superimposed metabolic acidosis [69, 73, 75] that develops when the airflow limitation and its concomitant hyperinflation compromise cardiac output. Death typically results from respiratory arrest rather than arrhythmias. No data are available on the degree of obstruction necessary to produce death, but some sense can be gained from the arterial gas tensions in patients who present to emergency departments. Hypercarbia occurs about 10% of the time and is associated with an FEV1 ranging between 10% and 20% of predicted [52, 53, 76, 77]. The elevation in PaCO2 is usually modest, averaging 50 to 55 mm Hg. Arterial oxygen tends to be well preserved, and values of 40 mm Hg or less at sea level are uncommon [52, 53, 77, 78]. Consequently, the profound abnormalities described above are most atypical. One striking feature that requires exploration is marked hypercapnia. Quite often, the PaCO2 value is greater than 100 mm Hg, with PaO2 tensions averaging 150 mm Hg [56, 70, 73, 74]. Such values can only exist if the nitrogen in the alveoli is replaced by administering 100% oxygen. Thus, it is possible that the use of uncontrolled oxygen rather than simple overwhelming airway narrowing is the culprit. Until this issue is clarified, it may be wise to regulate the use of supplemental oxygen [78].
Identification of High-Risk Patients
The identification of high-risk patients is challenging. Patients with severe disease are thought to be in greatest peril, but those with mild illness may also die if therapy is inappropriate [79, 80]. Marked circadian variation in lung function, a large bronchodilation response, psychosocial instabilities, the use of three or more medications, frequent visits to the emergency department, recurrent hospitalizations, and previous life-threatening asthma attacks have been touted as portending poor outcomes [6, 7, 10, 14, 18, 56, 68, 81-83]. All but the last factor are too nonspecific to be of much predictive value, but even this last factor is problematic. Of the more than 900 patients in whom fatal or near-fatal asthma has been reported, only 36% were admitted in the year preceding their episode and just 6% had a history of ventilatory assistance or admission to an intensive care unit [10, 14, 18, 55, 56, 73-75]. The existence of such events is extremely important in a given patient, but the absence of these events is of no value in assessing risk. One noteworthy clinical point is that patients with asthma who experience respiratory failure with one episode of asthma tend to experience failure with subsequent episodes [70, 73-7582, 83]. Thus, prudence dictates that such patients be treated with special care [82, 83].
Two additional groups of patients have abnormal responses to airway narrowing that put them at risk for recurrent respiratory failure. One group is unable to sense the presence of even marked airway obstruction and does not develop symptoms until the respiratory reserve is almost exhausted [84]. The other group has blunted hypoxic ventilatory drives and does not develop the hyperpnea characteristic of acute attacks [85, 86]. In this group, alveolar hypoventilation can rapidly develop with even moderate degrees of obstruction. Both groups are rare, but failure to recognize them can result in needless tragedy.
To detect potential danger, the manner in which decompensation develops must be appreciated. With the infrequent exception of so-called sudden asphyxic asthma, the risk for death in chronic asthma increases in direct proportion to the severity and frequency of the exacerbations and increases inversely with the extent of recovery. The typical course is an oscillatory one of recurrence and partial improvement that may extend from days to weeks. No single sign or symptom is prognostic [87]; rather, it is the overall pattern of increased frequency of nocturnal awakenings and worsening daytime dyspnea and wheezing that is important. When such a pattern develops, aggressive measures must be instituted and daily objective assessments of the severity of the airflow limitation become essential. To avoid disaster, treatment adjustments must be based on the patient's quantitatively determined needs rather than on the physician's or patient's subjective impressions.
The harbingers of impending disaster in emergent situations are a pretreatment arterial saturation less than 90%; normal or high CO2 tensions after receipt of appropriate therapy; persistent metabolic acidosis; and severe obstruction that does not improve by 30% to 40%, or worsens, with sympathomimetic agents [70, 73-7578, 88]. The presence of any of these conditions requires serial objective assessment in an environment capable of providing ventilatory assistance.
Preventing Death
The prevention of asthma-related death begins with the use of appropriate medications and is anchored to the ability of care givers and patients to recognize the danger signs listed in the preceding section. Anti-inflammatory agents are required in any patient with active disease [24], and the key to safety lies in an educated patient armed with a written plan for emergencies [65]. Objective assessment and close communication between physician and patient during periods of decompensation are essential. In emergency situations, there is a pressing need to develop and implement standardized protocols that incorporate effective drugs and serial measurement of the patient's progress [78, 88]. Clinical evaluation must always give way to the findings derived from objective monitoring of the severity of obstruction and the associated abnormalities in gas exchange. A useful maxim is to assume that every exacerbation is potentially fatal and to act accordingly. This is particularly true with patients who have previously had a life-threatening episode. In this circumstance, every effort should be made to prevent new attacks from developing. Here, long-term objective monitoring of lung function with preemptive changes in treatment is critical [75].
Prognosis after a Near-Fatal Episode of Asthma
The prognosis after a near-fatal asthma attack depends in part on how such an event is defined. To some, a near-fatal asthma attack is an episode of hypercapnia [75]; to others, it is the need for intubation and ventilatory assistance [70]. The former occurs 10 to 20 times more frequently than the latter and does not necessarily imply an ominous short-term or long-term prognosis [52-5476, 77]. Airway instrumentation, however, does have this implication. The published mortality rates in intensive care units for acute asthma range from 0% to 40%; most of these deaths were caused by complications of mechanical ventilation [54, 55, 73, 88-90]. In two studies that involved 339 patients with status asthmaticus [54, 55], the prevalence of hypercapnia varied from 27% to 62% but the need for ventilatory assistance ranged from 8% to 30%. It is impossible to ascertain how the decision to intubate was made in these studies; all too often, however, subjective criteria rather than objective observations of gas exchange are used [56]. Standardized approaches are desperately needed.
The long-term prognosis after a near-fatal asthma attack is poor. Approximately 17% of patients ultimately die of their illness; 10% do so in the year after the event [4, 56, 73, 91, 92]. The largest series is a 6-year study of 145 persons who had undergone mechanical ventilation for respiratory failure [75]. Of the 121 patients discharged, 60% were subsequently rehospitalized at least once, 19% were rehospitalized twice, 19 needed mechanical ventilation during a second episode, and 17 died. It is shocking that despite near-fatal episodes, lung function was objectively assessed after the initial hospitalization in only 59% of the survivors; only 21 of these survivors were regularly recording their peak flow. In asthma, familiarity does indeed breed contempt [93].
- Copyright ©2004 by the American College of Physicians
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
