If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author. Department of Allergeology and Pulmonology, Astrid Lindgrens Childrens Hospital, Karolinska University Hospital, S-171 86 Stockholm, Sweden. Tel.: +468 517 775 65; fax: +468 517 777 39.
Previous reports indicate that morbidity and mortality from asthma have increased during the past decades. Here, the mortality rate associated with asthma and possible risk factors in children and young adults in Sweden during the period 1994–2003 were evaluated.
Methods
The medical profession was asked to report suspected cases of death from asthma in individuals 1–34 years of age. All death certificates containing relevant ICD codes were reviewed. Medical records and autopsy reports were assessed and telephone interviews with next-of-kin performed.
Results
During the 10-year period 37 deaths due to asthma were identified. The median age at the time of death was 27 years and 6 of the deceased were younger than 15. The overall incidence of death from asthma decreased from 1.54 deaths per million in 1994 to 0.53 per million in 2003. Common risk factors were under-treatment (23/37), poor adherence to prescribed treatment (17/37) and adverse psychosocial situation (19/37). An alarming finding was that 11 of the 37 deaths were probably caused by food allergy and for 8 subjects death was associated with exposure to pet dander. The death certificates were found to contain inaccuracies with 30% of those for whom asthma was reported as the underlying cause having died from other causes.
Conclusion
Asthma mortality in children and young adults in Sweden decreased between 1994 and 2003. Food allergy and inadequate treatment were the major risk factors for such a death. Recognition and special care of patients with asthma who have shown signs of non-compliance, denial or severe food allergy must be encouraged.
At the same time the frequency and duration of hospitalization of school children and young adults due to asthma have in general decreased continuously during the past two decades, the exception being the 1–4 years old with wheezing illness, for whom the rate of hospitalization has remained unchanged.
The few epidemiological and case studies of mortality due to asthma in children and young adults have revealed decreasing rates in several European countries as well as in the United States.
Asthma mortality in Swedish children (1–14 years of age) and young adults (15–24 years old) remained low (1–6 deaths per million individuals each year) between 1952 and 1972,
found that this rate for young adults almost doubled during the period of 1981–1988.
An alarming observation in this latter study was that the severity of asthma in some of the young people, who died of acute attacks, was considered to be moderate or mild. Most of these patients had previously suffered from severe asthma and allergy, but their condition appeared to have improved since childhood and they were not receiving anti-inflammatory treatment. In addition, poor compliance to prescribed use of medication, due to inadequate education of the patients, denial and psychosocial factors, is also a risk factor for mortality from asthma.
The aim of the present investigation was to characterize changes in the incidence of mortality from asthma among inhabitants of Sweden 1–34 years of age from 1994 to 2003 and to identify risk factors for such mortality in children and young adults. In this context we also attempted to determine the accuracy with which deaths from asthma in these age groups were recorded.
Methods
In 1993 a national Swedish Task Force was established to monitor prospectively all deaths of young people from asthma. The expert panel included specialists in both adult and pediatric allergology and respiratory medicine and in forensic medicine as well as an asthma nurse.
Starting in 1994 copies of all death certificates for 1–34 years old containing the ICD-9 codes 490–496 and 995 or the ICD-10 codes J40–47 and T28 recorded in any part of the death certificate were received from Statistics Sweden and, later on, from the Centre for Epidemiology (EPC) at The National Board of Health and Welfare. Furthermore, the medical profession was asked to report all suspected deaths from asthma in this age group. Finally, the police reports, medical records and autopsy reports of all such individuals suspected to have died from asthma were collected. Whenever possible an asthma nurse conducted telephone interviews with the next-of-kin employing a modified standardized questionnaire developed by the British Thoracic Association and previously used by Mohan et al.
The information obtained concerning each suspected death from asthma was reviewed according to a standardized protocol by one member of the Task Force and subsequently presented to the rest of the panel, which met regularly to discuss about the patients and complete a final panel work-sheet. In each case the panel had to determine whether the patient had indeed died of asthma and, when possible, identify any predisposing factors for this death.
Death was considered to have been due to asthma whenever typical symptoms, including wheezing and breathlessness had preceded death according to the medical records, police reports or next-of-kin and/or characteristic findings were observed upon autopsy.
A majority vote was needed to confirm or reject that death was caused by asthma, but a vote was never needed as the panel was congruent in all cases. Additionally, all cases were re-evaluated at the end of the study, which confirmed the original evaluation.
Definitions
For classification of the severity of asthma, cases with a history of highly sporadic and mild symptoms were defined as having had mild asthma; those with recurrent, but mild symptoms and/or in need of daily treatment with inhaled steroids as having had moderate asthma; and individuals with symptoms that restricted activities of daily life as having had severe asthma. The presence of allergy was defined as demonstration of a positive skin-prick and/or serological test, together with a history of development of typical symptoms in response to exposure to allergen.
Under-treatment was considered to have occurred whenever a recurrent failure to show up for scheduled outpatient appointments, non-compliance with prescribed medication or management or unwillingness to accept the diagnosis of asthma was documented or reported. Moreover lack of inhaled steroids in those with moderate or severe asthma, irrespective of whether the patient or the medical service was responsible, was considered to be under-treatment. Likewise, subjects, who despite suffering from a known food allergy that had previously caused anaphylactic reaction, did not carry an auto-injector with adrenaline were also recorded as being under-treated. An adverse psychosocial situation was considered to have existed for subjects who abused alcohol or narcotics; were psychologically depressed to an extent that required antidepressant medication or resulted in unemployment; or had severe language difficulties.
For analysis, the subjects were divided into the following age groups: 1–19 years old (n=12, including 6 male), 20–29 years old (n=12, 8 males) and 30–34 years old (n=13, 7 males).
Ethics
Ethical approval for the study was granted by Ethics committee at the Medical Faculty of Uppsala University.
Results
The incidence of death due to asthma and accuracy of the death certificates
During the 10-year period from 1994 to 2003, 75 deaths suspected to be due to asthma were reported and subsequently analyzed by the expert panel (Fig 1). In 26 of these cases the Task Force concluded that the death was obviously unrelated to asthma while for 12 other individuals the information available was insufficient for confident determination of the cause of death. Of the 46 cases for whom asthma was recorded as the underlying cause of death in the death certificate, the expert panel disagreed about 14 indicating an over-estimation (false positive rate) of 30%. Five of these 14 patients were severely mentally retarded with cerebral palsy syndromes. The time span between death and start of follow up was 0–25 months with median and mean of 7 months.
Figure 1Comparison of the cause of death according to the death certificate* and as evaluated by the Task Force**.
Of the 37 deaths classified by the Task Force as being due to asthma, autopsy reports were received for 28 and telephone interviews conducted with the next-of-kin for 24, with the same questions being answered in written form for an additional three. In five of these 37 cases, asthma was not regarded as the underlying cause of death on the death certificates, giving an under-estimation (false-negative rate) of 14%. In three of these five cases, asthma was recorded as a contributing cause of death; whereas in the other two, reported by the medical profession, asthma was not mentioned at all on the death certificate.
As illustrated in Fig. 2 the incidence of deaths due to asthma, as determined by the Task Force, decreased from 1.54 per million in 1994 to 0.53 per million in 2003. The median (mean) age at the time of these deaths was 27 (23.5) years, with four individuals dying at 10 years of age or younger (Fig. 3). Four of the 37 deaths occurred following admission to the hospital and of the 33 that occurred outside the hospital, 8 took place on the way to the hospital, 12 while waiting for ambulance and the remaining 13 were found dead, without having called anyone about their situation.
Figure 2The incidence of deaths due to asthma among 1–34 years olds in Sweden from 1994 to 2003 as determined by the Task Force.
Two of the patients who died had not been previously diagnosed as asthmatic. As illustrated in Table 1, the severity of the asthma which the deceased suffered had increased with age. Five had been hospitalized at least once for this condition during the 12 months preceding their death. Just before the final attack, the degree of asthma was characterized as severe in 10 of the 34 patients for whom this could be assessed. The condition of 17 of these 34 patients had deteriorated notably several hours or even days prior to their death, indicating that they had delayed in seeking medical help.
Table 1Characteristics of the children and young adults who died from asthma
Age group (year)
Number with the following characteristics:
1–19
20–29
30–34
Combined
n=12
n=12
n=13
n=37
Asthma recorded as the primary cause of death in the death certificate
10
11
11
32 (86)
Asthma mentioned as contributing cause of death in the death certificate
2
1
2
5 (14)
Autopsy performed
9
9
10
28 (76)
Interview conducted (orally or in written form)
8
11
8
27 (73)
Suffering from any allergy
10 [1]
8 [2]
7 [4]
25 (68)
Pollen allergy
5
7
6
18 (49)
Pet allergy
7
6
3
16 (43)
Food allergy
9
2
0
11 (30)
Mild asthma
5
3
2
10 (27)
Moderate asthma
5
6
6
17 (46)
Severe asthma
2
3
5
10 (27)
(x)=Percent of all asthma deaths.
[x]=Number not evaluated due to insufficient information.
Of the 37 deaths, 25 were known to suffer from allergy and 5 were classified as non-allergic with insufficient information concerning the possible presence of atopy for the other 7. Allergy to airborne allergens, i.e., pet dander (n=16) and pollen (n=18), dominated (Table 1).
Allergy to food (peanuts, soy and almonds) was considered to be responsible for 11 of the deaths (Table 2). All these individuals either exhibited typical symptoms of a sudden, severe and irreversible asthma attack immediately after eating the item of food to which they were allergic and/or typical characteristics of acute asthma in connection with autopsy. Among these patients who suffered a fatal asthma attack as a consequence of food allergy, five had mild, four moderate and two severe asthma.
Table 2Major factors which contributed to the deaths of the children and young adults due to asthma
Among the 1–19 years old (n=12), food allergy was the major cause of the fatal attack in 9 cases, two of whom suffered from severe asthma. In all but one patient, their allergy to food was previously known. For 8 patients with known allergy to pet dander, death due to asthma was associated with exposure to this specific allergen immediately prior to the fatal attack. Allergy was a less important risk factor for the individuals who were 30–34 years of age (Table 2).
For the entire group no seasonal variation in the frequency of mortality due to asthma was detected. Nor was any increase in mortality observed during the Swedish pollen season (April–August), irrespective of the presence or absence of pollen allergy. No gender difference was found regarding accuracy of death certificates and risk factors for asthma death.
Under-treatment
Altogether, of the 36 deaths for whom this could be assessed, 23 were considered to have been under-treated (Table 2). Poor adherence to asthma-treatment was responsible for this under-treatment in 16 cases, the medical profession in another six and both for one individual. Fifteen of the patients were not taking inhaled steroids at the time of death, even though two of these demonstrated clear evidence of severe asthma and another five of moderate asthma. Only 1 out of the 8 patients who suffered a fatal reaction to an item of food had an auto-injector with adrenaline available and used it. Three patients, all in the 1–19 years old group, were classified as under-treated on the basis of the fact that despite having severe food allergy that had previously caused anaphylactic reactions they were not equipped with an auto-injector containing adrenaline.
Inadequate treatment of asthma was most evident in the groups of 20–29 and 30–34 years old and was usually due to poor compliance, 7 patients in these groups being drug addicts. One patient was advised to discontinue treatment with inhaled steroids since she was pregnant, while in another two cases it was concluded that the medical service had failed as these patients were sent home without adequate treatment following an acute attack of asthma and subsequently died of asthma a few days later.
Psychosocial risk factors
Psychosocial problems were considered to be a risk factor in 19 of the cases of mortality, including 14 (10 of whom were males) who had been under-treated. Alcohol and/or drug abuse was the major psychosocial factor (Table 2) and there was clear evidence of intravenous intake of narcotic drugs immediately prior to the fatal attack in six individuals. Analysis revealed the presence of amphetamine in four of these subjects and morphine in one, while no analysis was performed in one case despite suspected intake of amphetamine in association with the fatal attack. Another 6 patients had been diagnosed with psychiatric disorders, while two others were immigrants with limited proficiency in Swedish. Of the 18 individuals who delayed seeking medical help 12 exhibited adverse psychosocial factors.
Discussion
During the 10-year period from 1994 to 2003 the incidence of death from asthma among children and young adults was found here to decrease from 1.54 to 0.53 per million individuals. Of the deceased 1–34 years old evaluated, only a minority had been known previously to suffer from severe asthma. An anaphylactic and asthmatic reaction to food was the major cause of death in the group of 1–19 years old the majority of whom had mild to moderate asthma. Between 20 and 34 years of age, poor compliance was the predominant underlying cause of death. An adverse psychosocial situation, especially involving drug abuse, was the major reason for this poor compliance. In a minority of our cases death was a consequence of the failure of the healthcare system. A substantial number of asthma deaths could have been prevented if the patients had sought medical care at an earlier stage. In additional, we found an over-estimation of asthma as the underlying cause of death on death certificates.
In contrast to the increase in overall prevalence of asthma, the rate of mortality from this disease in Sweden decreased during the present study period, in agreement with what has been reported from several other countries.
Most likely, improvement in the management of asthma, and especially the introduction of treatment with inhaled steroids account for much of this decrease. In the case of Sweden, the major changes in treatment of asthma that emerged around 1985 were accompanied by regular information campaigns directed towards the medical profession, as well as the general public. Accordingly, despite an increase in the prevalence of asthma in Sweden from 1985 to 2000 Wennergren and co-workers have reported a continuous reduction in the frequency and length of hospitalization of 5–18 years old for this disease during this same period.
The most dramatic decrease in hospitalization was observed during the period when use of inhaled corticosteroids became more common which support that the decrease in asthma mortality represents an advance in the treatment of asthma. In attempts to identify risk factors for death due to asthma seasonal variations could provide important clues. Here, we detected no such variations, including no correlation between the presence of pollen allergy and the death rate during the pollen season, in agreement with previous findings on 1–19 years old Danish children.
However, in this previous study by Jorgensen and colleagues there was a positive correlation between the number of positive skin-prick tests and the risk of dying from asthma during the summer, with a peak in August. In our case we did not have all the information necessary for performing such detailed analyses. We did detect an association to exposure to pet dander in 8 cases, uniformly distributed between the three age groups.
Moreover we found that food allergy, especially allergy to soy and peanuts, was a significant cause of death by asthma, especially among 1–19 years old. The majority of these individuals suffered from mild or moderate asthma confirming the previously reported combined risk of a strong allergy to peanuts and/or soy protein and mild/moderate asthma in young persons.
These results also emphasize the role of asthma in fatal food anaphylaxis. However, the high proportion of our deceased patients who had mild (n=10) or moderate (n=17) asthma was unexpected and differs from several previous reports.
These relatively high numbers can only be explained in part by an allergic reaction to food or pets, which occurred in 19 of these subjects, while other factors must have played a role in the remaining 8 cases.
Death by asthma is more common among patients with psychosocial problems, especially where alcohol and/or drug abuse is involved.
Moreover, we observed that the frequency of both psychosocial problems and under-treatment were higher among the 30–34 years old than the younger subjects. The relatively large number of deaths that occurred in immediate association with intravenous administration of narcotics, especially amphetamine, is a remarkable finding here, which in contrast to the association inhalation of heroin has, to our knowledge, only been reported a few times previously.
In our study, only four patients died in the hospital, whereas a majority of the 1–19 years old died while waiting for the ambulance following a fatal food reaction. In the groups of 20–29 and 30–34 years old, the number who died outside the hospital demonstrated no association with psychosocial situation.
The importance of identifying the subgroup of asthmatics with psychosocial problems has been clearly demonstrated by Mayo and co-workers, as well as by Ruffin and Harrison, who were able to reduce asthma mortality by organizing special clinics for such patients.
When interpreting data concerning mortality caused by asthma, it is important to know how accurate such data provided by death certificates are. Previous studies have revealed high frequencies of both under- and over-estimation of the numbers of deaths due to asthma by death certificates.
Death certificates may report a disease as being either the underlying or a contributory cause of death, with official statistics reflecting primarily the reported underlying causes.
In the present investigation, of the 37 cases classified as death due to asthma by the Task Force, asthma was recorded as the underlying cause of death on the death certificates of 86%. Thus the proportion of such deaths which were inaccurately diagnosed as due to other causes (14%) was similar to that reported previously from Great Britain (18%) and Canada (16%).
Moreover, 14 (30%) of those certified as having died from asthma were judged by our Task Force either to have died from another cause or of a cause that could not be determined on the basis of the information available. In contrast, in a Finnish study the validity of death certificates on asthma proved to be good and only 7% were found to misclassify the cause of death.
A large proportion of the deaths misdiagnosed as being due to asthma had also been diagnosed as suffering from neuromuscular diseases, cerebral paresis or congenital heart diseases. Although in similar previous studies, the shift from the ICD-9 to the ICD-10 code has been discussed as a possible confounding factor, we found no indication that this shift influenced our present results.
Retrospective studies such as ours have always limitations with uncertain data. However, our method for obtaining information concerning deaths due to asthma and evaluation of each individual case by a group of specialists are well established and minimize the time span between occurrence of death and collection of data.
In spite of this, the maximum time between death and start of follow up was 25 months (median 7 months). Doctors delay before issuing death certificate and the subsequent handling at authorities was the main reason for this time span. However, the panel had no indications that the length of this time had any significant impact in obtaining medical records, police reports or on interviews with next-of-kin. All death certificates mentioning asthma or any associated diagnose were sent routinely to the Task Force. Except for two cases, all such deaths reported by the medical profession were also documented by the death certificates, indicating good reliability in obtaining the information required. Individuals 1–34 years of age were selected for evaluation in order to avoid confusion with other pulmonary diseases, such as chronic bronchitis and emphysema, in older patients, as well as to avoid the difficulty in diagnosing asthma and to exclude confusion with congenital heart and lung diseases in very young children. Moreover, data concerning asthma as the underlying cause of death have been shown to be the most reliable for individuals of 5–34 years of age.
In conclusion, the present retrospective findings demonstrate that, in Sweden, mortality due to asthma in children and young adults is decreasing. A large proportion of such deaths occur in patients with mild or moderate asthma and among those not using inhaled steroids. Adverse psychological and social factors are frequently associated with death by asthma, especially among older patients. In the case of the younger patients, an allergic reaction to food was the major cause of such death. Additionally, in those sensitized, close exposure to pet dander could be hazardous. Recognition and special care of patients with asthma who have shown signs of poor compliance, denial or severe food allergy must be encouraged.
Conflict of interest
The authors have no potential conflict of interest related to this article.
Acknowledgments
Swedish Heart Lung Foundation.
Swedish Asthma and Allergy Association.
National Institute of Public Health.
Consul Th C Bergh´s Foundation.
References
Backlund A.B.
Perzanowski M.S.
Platts-Mills T.
Sandstrom T.
Lundback B.
Ronmark E.
Asthma during the primary school ages – prevalence, remission and the impact of allergic sensitization.
00130-3&id=gr1.jpg){kind=link}
00130-3&id=gr2.jpg){kind=link}
00130-3&id=gr3.jpg){kind=link}