American Journal of Respiratory and Critical Care Medicine

Rationale: Asthma is heterogeneous and knowledge on exacerbation patterns is lacking. Previous studies have had a relatively short follow-up or focused on severe disease.

Objectives: To describe exacerbation patterns over a prolonged follow-up in a population that includes patients of all disease severity.

Methods: We used electronic health care records to identify patients with asthma aged 18–55 years and their exacerbations from 2007 to 2015. A cohort with greater than or equal to 7 years of data was used to describe exacerbation patterns by asthma severity defined by medication use. Effect estimates for risk factors were calculated for sporadic (single year of exacerbations) and recurrent (>1 yr) exacerbation patterns, using logistic regression. In a nested case–control design, the association between a history of exacerbations, spanning 5 years, and a future exacerbation was examined.

Measurements and Main Results: A total of 51,462 patients were eligible for the 7-year cohort; 64% had no exacerbations. Of those who exacerbated, 51% did so only once; exacerbation frequency increased with disease severity. Only 370 patients (0.7%) were characterized by a frequent-exacerbator phenotype (yearly exacerbations), of whom 58% had mild/moderate asthma. Exacerbation risk factors were not uniquely associated with a particular exacerbation pattern. A past exacerbation increased the risk of a future exacerbation more than all other factors, although this effect dissipated over 5 years.

Conclusions: During 7 years of follow-up, exacerbations occur in around one-third of patients. Of those who exacerbate, half do not do so again; the timing of future exacerbations is largely unpredictable. Just 2% exhibit a frequent-exacerbator phenotype. Past exacerbation patterns are the most informative risk factor for predicting future exacerbations.

Scientific Knowledge on the Subject

Available studies on exacerbation patterns are small (<800 patients), followed up with patients for 3 years or fewer, and usually included only those with more severe disease; the exception followed patients for just 12 months. Although studies have found a recent history of a past exacerbation increases future exacerbation risk, only one study looked at a prolonged past history (5 yr); however, it included only patients on high-dose inhaled corticosteroids.

What This Study Adds to the Field

To our knowledge this is the largest study of the pattern of asthma exacerbations; our cohort also included patients of all disease severity. Our findings suggest most patients with asthma seldom exacerbate and that when patients do exacerbate, half of the time it is a single sporadic event. However, in the other half, having an exacerbation is one of the biggest risk factors for having a subsequent one, even up to 5 years earlier. Our findings also suggest that from routinely collected information it is not possible to predict with accuracy which patients are likely to exacerbate, although certain factors increase the risk. These findings aid in understanding the natural history of asthma.

Asthma is a highly prevalent and inherently variable disease. Its severity ranges from mild, with barely noticeable symptoms, to very severe, causing significant impact on the patient’s life. Exacerbations, characterized as acute or subacute progressive worsening of asthma symptoms, diminish quality of life and impose a considerable health care burden on society (1, 2). Exacerbations can exhibit various patterns. Although some patients experience only sporadic events, others may suffer exacerbations in clusters, such as in response to seasonal changes; others, with more severe asthma, are described as having a “frequent exacerbator” phenotype (35). Multiple factors have been epidemiologically linked to exacerbations; they include asthma severity, sex, comorbidities, and a history of a recent exacerbation (69).

Although it is recognized that there are several patterns of exacerbations, most studies have followed patients for only short periods of time, usually 1 year and at most 3 years, and included only specific patient groups (4, 5, 1013). In particular, most studies have addressed only the small proportion of patients with asthma who have more severe or difficult-to-treat disease, because they are significant users of expensive health care. However, the greatest societal and health care burden may lie with most individuals with asthma who have much milder disease, yet are collectively responsible for substantial prescription costs and loss in work productivity (1416).

To better comprehend all adult exacerbation patterns and to determine the effect of both recent and distant exacerbations on the risk of future events, we have used one of the world’s largest databases of longitudinal medical records (17). Some of the results of this study have been previously reported in the form of an abstract (18).

Data Sources

Our data source was pseudonymized, primary care information from the Clinical Practice Research Datalink (CPRD), which has coverage of approximately 7% of the U.K. population, and has been shown to be nationally representative, including with respect to age and sex (17). To include hospital-treated exacerbations and asthma deaths, only CPRD patients whose records were linked to Hospital Episodes Statistics (HES; data collected during all hospital visits) and Office of National Statistics (mortality data) were included. HES is only collected from hospitals in England, therefore only approximately 60% of CPRD patients have HES-linked data. This linkage also provides data on socioeconomic status through the Index of Multiple Deprivation (IMD) (19).

Study Population and Design

All patients included were defined as having asthma using validated asthma-specific Read codes (the clinical terminology system used by GP practices and CPRD) (20). Patients were eligible once they had satisfied all of the following: received their first asthma Read code, their data were deemed research acceptable by CPRD quality control, they were continuously registered in a general practice with CPRD linkage, and were aged greater than or equal to 18 years. Patients were also only eligible on or after April 1, 2007 (before this date HES Accident and Emergency, A and E data were not available). A patient’s data were then censored on the first date of any of the following criteria: death, transfer out of a CPRD practice, last CPRD-data collection, last HES-data collection, aged 55 years, or on December 31, 2015. Two cohorts were then derived from this study population: a 7-year cohort and a nested case–control population (see Figure E1 in the online supplement).

Seven-year cohort

All patients included in the 7-year cohort had to have 7 or more years of eligible data; data were censored 7 years after the patient’s eligibility date.

Nested case–control

Cases were identified as patients that had experienced an exacerbation between September 30, 2014, and September 30, 2015; their date of the exacerbation in this time period was their index date. Three control subjects were matched to each case by age, sex, and general practice on the index date. All cases and control subjects had to have 5 or more years of eligible data before the index date.

Exacerbations and Other Variables

We defined a moderate exacerbation as one with 1) a prescription of oral corticosteroids, total amount less than or equal to 300 mg, not prescribed during an annual asthma review or on the day a patient was seen for another steroid-treated disease (list available from authors on request); 2) an A and E visit (identified using HES A and E data); or 3) a hospital admission of less than 1 day duration (identified using HES admissions data and International Classification of Diseases-10 codes for asthma). A severe exacerbation was defined by hospital admission 1 or more night for asthma. We used the highest level of care to define an exacerbation and considered those recorded within 14 days of another to be part of the same event.

In the 7-year cohort, to aid clarity in the different patterns of exacerbations, we applied the following definitions: a “sporadic” pattern was defined as one or more exacerbation in only 1 of the 7 years; a “recurrent” pattern was defined as more than 1 year of having an exacerbation; and a “frequent-exacerbator” phenotype was defined as one or more exacerbations every year.

In the nested case–control analysis, a recent exacerbation was defined as one in the year before the index date, whereas a distant exacerbation was defined as one occurring in any of the 4 years before that (Figure 1). The term “past year” refers to exacerbations occurring in that year (e.g., “Past Year 5” includes only exacerbations in the year that began 5 years before the index date). A past year was defined as a severe year if there was at least one severe exacerbation in that year.

All covariates in the 7-year cohort were ascertained at entry, whereas covariates in the nested case–control analysis were ascertained on the index date. Body mass index (BMI) was measured using kg/m2. Mean blood eosinophil counts (cells/μl) were derived using the mean of each patient’s counts during the 7 years. A history of atopy, rhinitis (allergic or nonallergic), gastroesophageal reflux (reflux), anxiety, or depression was recorded using appropriate Read codes (available on request). Chronic obstructive pulmonary disease (COPD) was identified by validated Read codes, a smoking history, and age greater than 35 years (21).

To measure asthma severity, we used the 2016 British Thoracic Society (BTS) stepped approach (incorporating medication class and dose), a recommended evidence-based method of measuring asthma severity in the United Kingdom (22). Patients were classified using all prescriptions in the year before their entry into the cohort, or their index date in the case–control analysis. The steps in our analyses correlate with the following pharmacologic management: step 1 = no maintenance medications (can be prescribed an inhaled reliever); step 2 = low-dose inhaled corticosteroid (ICS); step 3 = low-dose ICS + long-acting β2-agonist; step 4 = medium-dose ICS ± add-on therapy; step 5 = high-dose ICS ± add-on therapy; step 6 = high-dose ICS + continuous/frequent oral corticosteroids. The dosage was according to the BTS guidelines, beclomethasone equivalent: low-dose = 400 μg (maximum, 799 μg), medium-dose = 800 μg (maximum, 1,599 μg), and high-dose = 1,600 μg (no maximum). Further details on the methods for our approach are available in our earlier paper (7).

To assess for any selection bias of the 7-year cohort, we compared the characteristics of eligible and ineligible patients.

Statistical Analysis
Seven-year cohort

We used descriptive statistics to describe the exacerbation patterns and multivariable multinomial regression to estimate the association between risk factors and sporadic or recurrent exacerbations. The models were adjusted for sex, age (categorized), BMI, IMD, medication step, smoking history, COPD, depression, atopy, reflux, rhinitis, and anxiety.

Five-year nested case–control analysis

We used incidence density sampling to mimic time-to-event analysis; the design allows controls to later become cases, and the effect estimates, obtained here by conditional logistic regression (odds ratio [OR]), approximate to a rate ratio (23). Each model was adjusted for BMI, IMD, medication step, smoking history, COPD, depression, atopy, reflux, rhinitis, and anxiety.

Sensitivity Analyses

Because of the percentage of missing values the models were adjusted for blood eosinophil counts in the multivariate analyses of the 7-year cohort (missing 25%) but not in the nested case–control (missing in more than 50% of 3:1 matched cases and control subjects).

All statistical analyses were conducted using Stata version 14 (Stata Corp.).

Ethics

The protocol for this research was approved by the Independent Scientific Advisory Committee for MHRA Database Research (protocol 16_067); the approved protocol was made available during peer review. Generic ethical approval for observational research using the CPRD with approval from Independent Scientific Advisory Committee has been granted by a Health Research Authority Research Ethics Committee (East Midlands, Derby, REC reference number 05/MRE04/87). Linked pseudonymized data were provided by CPRD. Data are linked by NHS Digital, a statutory trusted third party, using identifiable data held only by NHS Digital. Select general practices consent to this process at a practice level with individual patients having the right to opt out.

Characteristics of the 7-Year Cohort

There were 51,462 patients with asthma with 7 years of continuous follow-up; 43.9% were men, and at the start of the cohort, they had a median age of 36.8 years (interquartile range [IQR], 28–42.7 yr); 51.4% were never smokers, 9.5% had a record of gastroesophageal reflux, 43.5% of rhinitis, 46.2% of atopy, 18% of anxiety, and 1.6% of COPD (see Table E1). Three-quarters of the cohort had an eosinophil count, of whom 8.0% had a count greater than 450 cells/μl.

The eligible population had a smaller proportion of patients in the oldest age category and larger proportion in the first medication step (see Table E2). The exacerbation rate was slightly higher in the ineligible population (eligible: incidence rate ratio = 2.02; 95% confidence interval [CI], 1.97–2.00) (ineligible: incidence rate ratio = 2.10; 95% CI, 2.09–2.11) (see Table E3).

Exacerbation Patterns

During the 7-year study period, 64% of the population had no record of an exacerbation, 24% only exacerbated once, and fewer than 5% had an exacerbation seven times or more.

Of the 18,555 patients who had an exacerbation during the 7-year follow-up, 51% (9,466 patients) experienced a sporadic exacerbation pattern, whereas the remainder (9,089 patients) experienced a recurrent exacerbation pattern. The number of patients who did not exacerbate, or displayed only a sporadic pattern, decreased with increasing asthma severity; the opposite was found for those who displayed a recurrent pattern, or had a frequent-exacerbator phenotype (Table 1 and Figure 2). Only 1.8% of patients with a single, sporadic year of exacerbations had more than one exacerbation during that year.

Table 1. Percentage of Patients with Asthma in Each Exacerbation Pattern by Asthma Severity

Exacerbation PatternStep 1Step 2Step 3Step 4Step 5Step 6All
No exacerbations73.164.556.649.836.34.363.9
Sporadic exacerbation16.419.620.621.320.44.318.4
Recurrent exacerbations10.415.622.227.439.887.817.0
Frequent exacerbations0.10.30.61.53.53.60.7

A sporadic exacerbation pattern was defined as ≥1 exacerbation in only one of the 7 years, a recurrent exacerbation pattern was defined as >1 year of having an exacerbation, and a frequent-exacerbator pattern was defined as ≥1 exacerbation every year.

For each consecutive year of the 7-year study, the number of patients that had exacerbated sporadically decreased (but remained over or around 10%, until the final year); in the seventh year of follow-up, 7% had exacerbated for the first time (Figure 3). During the consecutive years, of those who had their first exacerbation (as indicated in the pie charts), only 26% had a further exacerbation later during the 7-year follow-up.

The distribution of time between the first and second exacerbation was wide and skewed to the right, with a median of 1 year and an IQR of 0.3–2.1 years (see Figure E2); median time was marginally longer for patients in medication step 1 than for those in medication step 6 (step 1: median = 1.2 yr; IQR, 0.4–2.6) (step 6: median = 1.0 yr; IQR, 0.4–2.1).

Factors Associated with Sporadic, Recurrent, and Frequent-Exacerbator Exacerbations

The factors significantly associated with exacerbations in order of the strength of their association were having asthma treated in a higher medication step; of older age; a codiagnosis of COPD; from a more deprived socioeconomic group; elevated eosinophil count; being female; being a current smoker; being obese; and diagnoses of rhinitis, reflux, depression, or atopy (Table 2; see Table E4). All risk factors were more strongly associated with a frequent-exacerbator than a recurrent pattern, and more strongly associated with a recurrent than a sporadic pattern. The only variable associated with a recurrent/frequent-exacerbator pattern, but not a sporadic one, was anxiety. The factors most strongly associated with an exacerbation in patients with a frequent-exacerbator pattern were asthma severity, medication step 4 (medium-dose ICS ± add-on therapy) or higher, and having a codiagnosis of COPD.

Table 2. Factors Associated with Each Exacerbation Pattern

 SporadicRecurrentFrequent
 Adjusted Odds Ratio95% CIP ValueAdjusted Odds Ratio95% CIP ValueAdjusted Odds Ratio95% CIP Value
Sex         
 MRefRef  Ref    
 F1.341.27–1.41<0.00011.771.67–1.89<0.00012.551.92–3.38<0.0001
Age category         
 18–25 yrRefRef  Ref    
 25–34 yr1.231.13–1.33<0.00011.191.09–1.30<0.00011.240.74–2.080.09
 35–44 yr1.251.16–1.35<0.00011.331.24–1.43<0.00011.621.00–2.61<0.01
 45–54 yr1.191.08–1.31<0.00011.441.32–1.62<0.00012.001.19–3.38<0.01
BMI         
 NormalRefRef  Ref    
 Overweight1.171.10–1.24<0.00011.211.13–1.29<0.00011.311.03–1.83<0.01
 Obese1.361.28–1.44<0.00011.701.59–1.81<0.00011.741.31–2.33<0.0001
 Underweight1.060.89–1.270.401.080.89–1.320.350.990.41–2.491.0
Socioeconomic status         
 1 (least deprived)RefRef  Ref    
 21.101.02–1.18<0.051.281.18–1.39<0.00011.951.3–2.96<0.01
 31.161.08–1.25<0.00011.411.30–1.59<0.00012.511.67–3.88<0.0001
 41.211.12–1.30<0.00011.481.37–1.71<0.00011.901.35–3.15<0.0001
 51.291.19–1.39<0.00011.511.37–1.66<0.00012.651.74–4.06<0.0001
BTS step         
 Step 1RefRef  Ref    
 Step 21.271.20–1.35<0.00011.631.52–1.74<0.00012.091.29–3.40<0.01
 Step 31.561.43–1.71<0.00012.632.43–2.92<0.00015.143.01–8.64<0.0001
 Step 41.811.68–1.95<0.00013.543.31–3.84<0.000113.048.68–19.0<0.0001
 Step 52.292.07–2.55<0.00016.826.2–7.5<0.000137.9926.0–57.2<0.0001
 Step 63.981.15–13.770.29  Insufficient sample size  
 Not BTS1.951.53–2.49<0.054.053.23–5.04<0.000121.9711.00–42.99<0.0001
Smoking history         
 NeverRefRef  Ref    
 Current1.221.15–1.30<0.00011.411.32–1.50<0.00011.781.30–2.38<0.0001
 Ex-smoker1.061.00–1.130.051.111.03–1.14<0.011.200.91–1.570.20
Rhinitis         
 NoRefRef  Ref    
 Yes1.121.06–1.18<0.00011.271.20–1.35<0.00011.371.07–1.75<0.05
Reflux         
 NoRefRef  Ref    
 Yes1.161.06–1.26<0.00011.271.16–1.38<0.00011.481.10–1.98<0.01
COPD         
 NoRefRef  Ref    
 Yes1.861.58–2.33<0.00014.663.87–5.68<0.000114.009.60–20.03<0.0001
Depression         
 NoRefRef  Ref    
 Yes1.181.06–1.20<0.00011.371.24–1.51<0.00011.461.13–1.89<0.01
Atopy*         
 NoRefRef  Ref    
 Yes1.141.08–1.20<0.00011.121.05–1.19<0.011.200.76–1.600.26
Anxiety         
 NoRefRef  Ref    
 Yes1.020.95–1.090.31.070.98–1.150.091.110.86–1.540.39

Definition of abbreviations: BMI = body mass index; BTS = British Thoracic Society; CI = confidence interval; COPD = chronic obstructive pulmonary disease; Ref = reference group.

Each variable is adjusted for all other variables shown. There were 46,216 patients included in the model (those with missing variables were excluded).

* Atopy variable here excluded those with allergic rhinitis because of collinearity.

The proportions of exacerbations occurring in each season did not differ significantly between patients with sporadic or recurrent exacerbations, but sporadic ones occurred slightly more often in the winter months than recurrent, whereas recurrent occurred slightly more in the spring months than sporadic (see Figure E3).

Characteristics of the Nested Case–Control Population

The nested case–control population included 4,734 cases and 11,752 control subjects. Overall this population was younger, had fewer males (30%), and higher proportion with comorbidities than the 7-year cohort. Their median age was 45.2 years (IQR, 39.2–50.2 yr), 42.5% had never smoked, 4.2% had gastroesophageal reflux, 17.3% rhinitis, 33.6% atopy, 25.2% anxiety, and 5.8% COPD (see Table E5).

Association between Previous Exacerbations (Recent and Distant) and a Future Exacerbation

The odds of having a future exacerbation were significantly higher for patients who had experienced a past exacerbation but decreased over the 5-year period such that a recent exacerbation conferred a higher risk than a distant exacerbation (recent exacerbation: Past Year 1, OR = 6.7, 95% CI, 6.1–7.4; distant exacerbations, Past Year 2, OR = 3.8, 95% CI, 3.5–4.1; Past Year 3, OR = 3.0, 95% CI, 2.8–3.3; Past Year 4, OR = 2.9, 95% CI, 2.7–3.2; Past Year 5, OR = 2.6, 95% CI, 2.4–2.8). The adjusted odds related to a past history of exacerbations were higher than the odds for all other measured risk factors (see Table E6).

The odds of a future exacerbation were higher for patients with a past history of multiple exacerbations than for patients with a single exacerbation in that year; this effect was unchanged throughout the 5 years (Table 3). The odds of a future exacerbation were much higher for patients with a past history of a severe exacerbation, compared with a moderate one, if the severe exacerbation had occurred in the preceding 12 months; however, a severe exacerbation more than 1 year previously had little influence on the risk of a future exacerbation, compared with a moderate exacerbation (Table 3).

Table 3. Association between History of a Previous Exacerbation and a Future Exacerbation, by Number of Exacerbations in That Year and Severity of Exacerbations

Number of ExacerbationsSeverity of Exacerbations
Year of Previous ExacerbationAdjusted Odds Ratio95% CIYear of Previous ExacerbationAdjusted Odds Ratio95% CI
Past Year 1 (recent)  Past Year 1 (recent)  
 Single5.75.0–6.4 Moderate6.66.0–7.3
 Multiple8.87.6–10.1 Severe14.18.4–23.6
Past Year 2 (distant)  Past Year 2 (distant)  
 Single3.12.8–3.5 Moderate3.83.4–4.1
 Multiple5.14.5–5.8 Severe4.83.2–7.1
Past Year 3 (distant)  Past Year 3 (distant)  
 Single2.52.3–2.8 Moderate3.02.8–3.3
 Multiple4.33.7–4.9 Severe2.61.8–3.8
Past Year 4 (distant)  Past Year 4 (distant)  
 Single2.52.2–2.8 Moderate2.92.7–3.2
 Multiple4.03.5–4.5 Severe3.52.3–5.2
Past Year 5 (distant)  Past Year 5 (distant)  
 Single2.11.9–2.4 Moderate2.62.3–2.8
 Multiple3.83.3–4.4 Severe3.02.0–4.4

Definition of abbreviation: CI = confidence interval.

Each model (severity model and number of exacerbations model) is adjusted for body mass index, smoking history, medication step, chronic obstructive pulmonary disease, anxiety, depression, reflux, rhinitis, atopy, and socioeconomic status. Matched on sex, age, and general practice. There were 4,734 cases in each model and 11,752 control subjects. “Past year” refers to all exacerbations in that year (e.g., “Past Year 5” includes all exacerbations in the year that began 5 years before the index date).

In this U.K. cohort of more than 50,000 adults with asthma, nearly two-thirds did not experience an exacerbation during 7 years of observation. Of the 18,555 individuals that did, just half had a further exacerbation during follow-up. The proportion having recurrent (≥2) years of exacerbations increased with accumulative asthma severity but only those in the highest medication step had a higher proportion displaying a recurrent pattern, than either a sporadic pattern or no exacerbations. Presently, there is a paucity of published data describing the natural history of the general asthma population. Although it is important to understand the minority that have severe asthma and/or a frequent-exacerbator phenotype, because of their high morbidity and mortality, it is equally important to recognize the exacerbation patterns that occur in the greater majority of individuals with mild or moderate disease. Our analysis has shown that exacerbations occur even in those with apparently mild disease on low levels of treatment and can sometimes occur recurrently. There are around 4 million work-days lost annually to asthma in the United Kingdom (2), so understanding the risk and likely exacerbation pattern for working-age individuals will help to inform those with asthma, their families, and their employers.

Having established that only half of those who exacerbated had a recurrent pattern, we next sought to understand if certain factors were more or less likely to be associated with exhibiting a sporadic or a recurrent pattern, such that one could perhaps predict a patient’s exacerbation pattern. Smaller studies have compared infrequent and frequent exacerbators with each other, over a 1-year study period, and found significant differences between these phenotypes (12, 24). However, these were small studies and did not make comparisons to patients without exacerbations. Our analysis compared exacerbators with those who did not exacerbate but found no clear differences in specific patient demographics or clinical characteristics, except that all risk factors were more strongly associated with the recurrent phenotype than the sporadic one. The factors most strongly associated were requiring a higher medication step, having a codiagnosis of COPD, elevated eosinophil count, aged between 35 and 55 years old, being female, obese, or a current smoker. There was a considerable reduction in the odds of having an exacerbation between those that were ex-smokers compared with those that were currently smoking.

Interestingly, when looking at seasonal patterns, although sporadic and recurrent exacerbations both followed the winter peak and August dip that have previously been described (5, 25, 26), sporadic exacerbations did exhibit a slightly higher winter to springtime ratio. This may be because patients with recurrent exacerbations are more likely to have exacerbations triggered by multiple factors, including viral exposure, multiple allergens, pollution, and occupational exposures; in contrast, sporadic exacerbations may occur predominantly in patients with well-controlled asthma who are less susceptible to common triggers, but have had a single high exposure to a trigger, such as a viral infection. Further adding to the difficulty in predicting the natural history of an individual’s exacerbations, the time interval between exacerbations varied widely, and was only mildly influenced by disease severity. These findings confirm the unpredictable nature of asthma exacerbations, regardless of disease severity, that patients and physicians see every day. In addition, they support the recommendations of the U.K. Royal College of Physicians asthma review, that all patients with asthma should be educated in their disease, and compliant with suggested management (16).

A patient’s past exacerbation history was the strongest risk factor for a future exacerbation. Because most patients do not exacerbate every year, those that have an exacerbation history are likely to have a distant exacerbation history, which we chose to define here as an exacerbation occurring more than 1 year ago. It has been well documented that a recent exacerbation (≤12 mo) is a significant risk for a future exacerbation (69), but the risk a distant exacerbation imposes has been less well explored. A study of patients with asthma on high-dose ICS, identified from a claims database in the United States, also looked at exacerbations over a 5-year period; although they did not directly address the difference between recent and distant exacerbations, they did show that a baseline exacerbation risk persisted during the 5-year study follow-up (27). We also found, in our general asthma population, that any exacerbation history, even up to 5 years ago, still carried a significantly increased risk of a future exacerbation, albeit the risk dissipated over the 5-year time period. Both a recent and distant history of multiple exacerbations during 1 year conferred a higher risk than a single exacerbation in that year. In contrast, and possibly less intuitively, only the severity of a recent exacerbation had a significant effect on the risk of a future exacerbation; the severity of a distant exacerbation had no significant effect.

Strengths and Limitations

The biggest strength of the study was its large sample size and length of longitudinal follow-up. A possible limitation was the identification of patients with asthma; although we used validated Read codes with a high positive predictive value (28), it is still possible some patients with asthma, likely to have mild disease, were excluded. There is consistent evidence that asthma may be overdiagnosed in adults. A Canadian study reported that around a third of adults with a community diagnosis of asthma did not have current disease; one in three were using daily maintenance medication (29). A small U.K. study of primary care patients found that a third of patients with asthma had normal spirometry and bronchoprovocation results (30). This potentially high prevalence of overdiagnosis may partly explain the large proportion of individuals with an asthma diagnosis in our cohort who did not seem to exacerbate. To try to ameliorate the problem of overdiagnosis of asthma in patients with COPD we used a maximum age cutoff of 55 years; in this respect only a very small percentage of the patients had a COPD codiagnosis.

Another strength was that the study populations were drawn from a nationally representative population, but because of the eligibility criteria to have extensive longitudinal records, and only include patients with HES-linked data, it is possible there was some selection bias, reducing the generalizability of the results. To address this potential limitation, we compared the patients eligible to enter the 7-year cohort with the ineligible patients; although there were some differences in the populations the rates of exacerbations were generally similar. Therefore, we believe the study cohort was a reasonably representative general asthma population although there was possibly some selection bias caused by the longevity of the patient follow-up. This study only addressed moderate or severe exacerbations; those exacerbations that were treated in hospital out-patients only, treated with an increase in inhaled corticosteroids, or not treated at all, were not included; because of the nature of the documentation in CPRD, it is not possible to identify a short increase in inhaled corticosteroids. Only 62% of Accident and Emergency data were captured by CPRD between April 1, 2007, and 2015; because these data contributed only a small fraction of all exacerbations, this should not have significantly affected the results.

There were some possible limitations in the measurement of medication step. First, the 2016 BTS asthma management guidelines were used, because of their clarity of inhaled corticosteroid doses in the stepwise approach, but our data predate these. Second, in the multinomial logistic regression model, the medication step used was that identified at entry to the cohort, but this may have changed over time in some patients. Last, the medications prescribed may not have been dispensed or adhered to, yet it would be expected that the prescription itself was an indication of the patient’s disease severity. There is no measure of adherence in CPRD, therefore it was not possible to estimate the effect of lack of adherence on the risk of a future exacerbation. The prevalence of atopy (just below 50% in the 7-yr cohort) may have been slightly lower than expected because atopy tests (including IgE and allergy skin tests) are infrequently performed in primary care in the United Kingdom.

Conclusions

Asthma exacerbation patterns are variable but the most likely course is no exacerbation, or a sporadic one except in the minority that have the most severe asthma, for whom a frequent-exacerbator phenotype is common. Exacerbations occurred across the spectrum of disease severity and treatment levels, including in those on the lowest treatment step. There were no clear patient risk factors that were associated with a sporadic rather than a recurrent exacerbation phenotype, although there was a small difference in seasonality. A past history of exacerbations was significantly associated with a future exacerbation, even up to 5 years ago; but the severity of a previous exacerbation was only influential in the year prior, not in subsequent past years.

1. Lloyd A, Price D, Brown R. The impact of asthma exacerbations on health-related quality of life in moderate to severe asthma patients in the UK. Prim Care Respir J 2007;16:2227.
2. Mukherjee M, Stoddart A, Gupta RP, Nwaru BI, Farr A, Heaven M, et al. The epidemiology, healthcare and societal burden and costs of asthma in the UK and its member nations: analyses of standalone and linked national databases. BMC Med 2016;14:113.
3. Dougherty RH, Fahy JV. Acute exacerbations of asthma: epidemiology, biology and the exacerbation-prone phenotype. Clin Exp Allergy 2009;39:193202.
4. Kupczyk M, ten Brinke A, Sterk PJ, Bel EH, Papi A, Chanez P, et al.; BIOAIR Investigators. Frequent exacerbators: a distinct phenotype of severe asthma. Clin Exp Allergy 2014;44:212221.
5. Sears MR. Epidemiology of asthma exacerbations. J Allergy Clin Immunol 2008;122:662668.
6. Blakey JD, Price DB, Pizzichini E, Popov TA, Dimitrov BD, Postma DS, et al. Identifying risk of future asthma attacks using UK medical record data: a respiratory effectiveness group initiative. J Allergy Clin Immunol Pract 2017;5:10151024.
7. Bloom CI, Nissen F, Douglas IJ, Smeeth L, Cullinan P, Quint JK. Exacerbation risk and characterisation of the UK’s asthma population from infants to old age. Thorax 2018;73:313320.
8. Chipps BE, Zeiger RS, Borish L, Wenzel SE, Yegin A, Hayden ML, et al.; TENOR Study Group. Key findings and clinical implications from The Epidemiology and Natural History of Asthma: Outcomes and Treatment Regimens (TENOR) study. J Allergy Clin Immunol 2012;130:332342.
9. Miller MK, Lee JH, Miller DP, Wenzel SE; TENOR Study Group. Recent asthma exacerbations: a key predictor of future exacerbations. Respir Med 2007;101:481489.
10. Kimura H, Konno S, Makita H, Taniguchi N, Shimizu K, Suzuki M, et al.; Hi-CARAT Investigators. Prospective predictors of exacerbation status in severe asthma over a 3-year follow-up. Clin Exp Allergy 2018;48:11371146.
11. Suruki RY, Daugherty JB, Boudiaf N, Albers FC. The frequency of asthma exacerbations and healthcare utilization in patients with asthma from the UK and USA. BMC Pulm Med 2017;17:74.
12. Boulay M-È, Pruneau-Pomerleau C, Villeneuve H, Deschesnes F, Ringuette L, Boulet L-P. Comparative features of asthma with frequent or infrequent exacerbations: a longitudinal study of retrospective and prospective events. J Asthma 2018;55:231243.
13. Loymans RJ, Sterk PJ. Exacerbation-prone asthma: a separate bioclinical phenotype? Am J Respir Crit Care Med 2017;195:275277.
14. Mukherjee M, Gupta R, Farr A, Heaven M, Stoddart A, Nwaru BI, et al.; Burden and True Cost of Asthma in the UK Research Team. Estimating the incidence, prevalence and true cost of asthma in the UK: secondary analysis of national stand-alone and linked databases in England, Northern Ireland, Scotland and Wales: a study protocol. BMJ Open 2014;4:e006647.
15. Cisternas MG, Blanc PD, Yen IH, Katz PP, Earnest G, Eisner MD, et al. A comprehensive study of the direct and indirect costs of adult asthma. J Allergy Clin Immunol 2003;111:12121218.
16. Levy M, Andrews R, Buckingham R, Evans H, Francis C, Houston R, et al. Why asthma still kills: the National Review of Asthma Deaths (NRAD) confidential enquiry report. London, UK: Royal College of Physicians (RCP); 2014.
17. Herrett E, Gallagher AM, Bhaskaran K, Forbes H, Mathur R, van Staa T, et al. Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015;44:827836.
18. Bloom CI, Palmer T, Quint J. Exacerbation patterns in adults with asthma in England: a population based study. Presented at the British Association for Lung Research 2018. September 6, 2018, Birmingham, UK.
19. Department for Communities and Local Government. The English indices of deprivation 2010: statistical release; 2010.
20. Nissen F, Quint J, Morales D, Mullerova H, Smeeth L, Douglas I. Validation of the recording of asthma diagnosis in UK electronic health records (Clinical Practice Research Datalink) [abstract]. Am J Respir Crit Care Med 2017;195:A2025.
21. Quint JK, Müllerova H, DiSantostefano RL, Forbes H, Eaton S, Hurst JR, et al. Validation of chronic obstructive pulmonary disease recording in the Clinical Practice Research Datalink (CPRD-GOLD). BMJ Open 2014;4:e005540.
22. British Thoracic Society. BTS/SIGN British guideline on the management of asthma; 2016.
23. Pearce N. What does the odds ratio estimate in a case-control study? Int J Epidemiol 1993;22:11891192.
24. Hoda U, Sousa A, Corfield J, Shaw D, Fowler S, Fleming L, et al. Characteristics of the frequent exacerbator in U-BIOPRED adult severe asthma cohort. Eur Respir J 2015;46:OA1461.
25. Jackson DJ, Sykes A, Mallia P, Johnston SL. Asthma exacerbations: origin, effect, and prevention. J Allergy Clin Immunol 2011;128:11651174.
26. Fleming DM, Cross KW, Sunderland R, Ross AM. Comparison of the seasonal patterns of asthma identified in general practitioner episodes, hospital admissions, and deaths. Thorax 2000;55:662665.
27. Schatz M, Meckley LM, Kim M, Stockwell BT, Castro M. Asthma exacerbation rates in adults are unchanged over a 5-year period despite high-intensity therapy. J Allergy Clin Immunol Pract 2014;2:570574.
28. Nissen F, Morales DR, Mullerova H, Smeeth L, Douglas IJ, Quint JK. Validation of asthma recording in the Clinical Practice Research Datalink (CPRD). BMJ Open 2017;7:e017474.
29. Aaron SD, Vandemheen KL, FitzGerald JM, Ainslie M, Gupta S, Lemière C, et al.; Canadian Respiratory Research Network. Reevaluation of diagnosis in adults with physician-diagnosed asthma. JAMA 2017;317:269279.
30. Shaw D, Green R, Berry M, Mellor S, Hargadon B, Shelley M, et al. A cross-sectional study of patterns of airway dysfunction, symptoms and morbidity in primary care asthma. Prim Care Respir J 2012;21:283287.
Correspondence and requests for reprints should be addressed to Chloe I. Bloom, M.B. Ch.B., B.Sc., Ph.D., M.Sc., National Heart and Lung Institute, Imperial College London, Emmanuel Kaye Building, 1b Manresa Road, London SW3 6LR, UK. E-mail: .

Author Contributions: C.I.B. and P.C. conceived and designed the study. C.I.B. obtained and managed the data. C.I.B. and T.P. analyzed the data. C.I.B., T.P., J.F., J.K.Q., and P.C. contributed to data interpretation and writing.

This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.

Originally Published in Press as DOI: 10.1164/rccm.201808-1516OC on December 3, 2018

Author disclosures are available with the text of this article at www.atsjournals.org.

Comments Post a Comment




New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine
199
4

Click to see any corrections or updates and to confirm this is the authentic version of record