A total of 201,289 patients met the case definition of asthma. Of these, 4.5% of entered the cohort by satisfying the inpatient component of case definition, 76.1% by satisfying the outpatient component, and 19.4% by satisfying the medication component. Among the initial cohort, 21,144 (mean age 28.7, 55.4% female) had at least one follow-up period compatible with severe asthma, constituting the study sample. These individuals contributed 40,803 patient-years of data for this study. Table 1 provides the basic demographic characteristics of the cohort. The average number of follow-up periods was 8.75 and on average 1.93 periods for each patient (22.0% of total follow-up time) were labelled as severe asthma.
Table 1: baseline characteristics of the sample
Variable
|
Value
|
|
Total sample
|
21,144
|
|
Female; N (%)
|
11,722 (55.4%)
|
|
Age at index date; Mean (SD)
|
28.7 (9.4)
|
|
Socio-economic status; N (%)
|
|
|
Quintile 1
|
4,624 (21.9%)
|
|
Quintile 2
|
4,030 (19.1%)
|
|
Quintile 3
|
4,077 (19.3%)
|
|
Quintile 4
|
4,073 (19.2%)
|
|
Quintile 5
|
3,916 (18.5%)
|
|
Unknown
|
424 (2%)
|
|
Variables estimated during follow-up
|
Patient years with severe asthma
|
40,803
|
|
Patient years associated with maintenance OCS use
|
3,348 (8.2%)
|
|
Episodic OCS use; mean number of episodes (SD)
|
0.89 (0.79)
|
|
Patient-years with at least one episodic use of OCS; N (%)
|
30,684 (75%)
|
|
Patient-years with asthma-related hospital admission; N (%)
|
1,502 (3.7%)
|
|
Patient-years with asthma-related physician visits; N (%)
|
|
|
GP visits
|
25,425 (62.3%)
|
|
Respirologist
|
1,983 (4.9%)
|
|
General internist
|
1,012 (2.5%)
|
|
SABA canisters used per patient-year, Mean (SD)
|
6.2 (5.4)
|
|
Medication Possession Ratio for ICS per patient-year; mean (SD)
|
26.7% (29.4%)
|
|
Comorbidity: N(%)
|
|
|
Group 1 (Myocardial infarction)
|
90 (0.2%)
|
|
Group 2 (Congestive heart failure)
|
369 (0.9%)
|
|
Group 3 (Peripheral vascular disease)
|
169 (0.4%)
|
|
Group 4 (Cerebrovascular disease)
|
173 (0.4%)
|
|
Group 5 (Dementia)
|
84 (0.2%)
|
|
Group 6 (Chronic pulmonary disease)
|
7,818 (19.2%)
|
|
Group 7 (Connective tissue disease)
|
641 (1.6%)
|
|
Group 8 (Peptic ulcer disease)
|
308 (0.8%)
|
|
Group 9 (Mild liver disease)
|
581 (1.4%)
|
|
Group 10 (Diabetes without complication )
|
2,105 (5.2%)
|
|
Group 11 (Diabetes with complication)
|
130 (0.3%)
|
|
Group 12 (Paraplegia and Hemiplegia)
|
81 (0.2%)
|
|
Group 13 (Renal disease)
|
276 (0.7%)
|
|
Group 14 (Cancer)
|
435 (1.1%)
|
|
Group 15 (Moderate or Severe liver Disease)
|
67 (0.2%)
|
|
Group 16 (Metastatic Carcinoma)
|
78 (0.2%)
|
|
Group 17 (AIDS/HIV)
|
162 (0.4%)
|
|
OCS: oral corticosteroid; ICS: inhaled corticosteroids; SABA: short-acting beta agonists; SD: standard deviation
The average prednisone-equivalent dose of OCS in patient-years labelled as severe asthma was 382.5 mg per patient-year (1.05 mg of prednisone per day). Overall, OCS use of any type declined by a relative 1.8% per year during the study period (P<0.001).
Trends in maintenance OCS use
In total, 8.2% (3,348) of follow-up periods were associated with maintenance OCS use. The average prednisone-equivalent daily dose of OCS among chronic users was 6.23 mg. Figure 1 depicts the trends of maintenance OCS use over time, age, and duration of severe asthma
The average proportion of patient-years with severe asthma associated with maintenance OCS use during the first three calendar years (2000 – 2002) was 9.1% (Figure 1 – top panel). The corresponding value for the last three years (2011 – 2013) was 6.3%. On average, maintenance OCS use declined by a relative 3.8%/year (p<0.001).
In contrast, the trend of maintenance OCS use over years of age showed an increase (Figure 1 – middle panel). Among the youngest five groups (14 – 18 years of age), an average of 4.0% of patients were maintenance OCS users, which increased to 12.1% among the oldest five groups (54 – 58 years of age). On average, maintenance OCS use increased by a relative 2.9% per one-year increase in age (p<0.001).
The trend over the time-course of severe asthma was also upward (Figure 1 – bottom panel). In the first three years of severe asthma, 8.0% of patient-years were associated with maintenance OCS use, which increased to 11.4% in the last three years (years 11 – 13 since the onset of severe asthma). On average, with each year since severe asthma onset, the proportion of maintenance OCS users increased by a relative 5.6% (P<0.001).
Trends in episodic OCS use
In total, 75.2% (n=30,684) of follow-up periods were associated with at least one episode of OCS use. The average number of OCS episodes per follow-up period was 0.89 (SD 0.79). Figure 2 depicts the trends of episodic OCS use over the three time axes of interest (calendar time, years of age, and years since the onset of severe asthma.
Episodic OCS use demonstrated a slightly upward trend over calendar time (Figure 2 – top panel). In the first three years (2000–2002), the average number of episodes of OCS use was 0.82/year, whereas in the last three years (2011–2013) it was 0.93/year (with a clear decline in the last year of observation). On average, episodic OCS use increased by a relative 1.1%/year (P<0.001).
The trend over years of age was also increasing (Figure 2 – middle panel). Among the youngest five groups (14 – 18 years of age), the average number of episodes of OCS use was 0.79/year, increasing to 0.95/year among the oldest five age groups (54–58 years of age). On average, there was a relative increase of 0.5% (P<0.001) in episodic OCS use for every one year increase in age.
The trend over time-course of severe asthma was also upward (Figure 2 – bottom panel). In the first three years of severe asthma, the average number of episodes of OCS use was 0.94, which increased to 1.01 in the last three years of available data (years 11–13). Within the time-course of severe asthma, the average number of episodes of OCS use increased by a relative 1.4%/year (P<0.001).
Factors associated with time trends of maintenance and episodic OCS use
Table 2 provides the results of the regression analyses for maintenance and episodic OCS use. Female sex and older age were associated with a higher likelihood of both maintenance and episodic OCS use, while higher socio-economic status was associated with a higher likelihood of maintenance but not episodic OCS use. Higher intensity of ICS use in the previous period was associated with higher maintenance and episodic use of OCS, most likely reflecting the severity of asthma. This was also the case for reliever medication use. Similarly, asthma-related inpatient and outpatient encounters with general practitioners (versus no encounters) were associated with higher likelihood of maintenance and episodic OCS use in the next year. Outpatient encounters with specialists (internists and respirologists, versus no encounters with any specialist) were associated with higher likelihood of maintenance use, but such associations for episodic use were not statistically significant. Several comorbid conditions were associated with chronic use, and fewer with episodic use. After controlling for covariates, the annual relative rate of decline in maintenance use was 2.8% and the annual relative increase in episodic use was 0.9%; both remained statistically significant.
Table 2: Factors associated with maintenance and episodic oral corticosteroid use
Parameter
|
Maintenance use
|
Episodic use
|
OR
|
95%CI
|
P
|
RR
|
95%CI
|
P
|
Female sex
|
1.11
|
1.00 - 1.23
|
0.05*
|
1.08
|
1.05 - 1.11
|
<0.01*
|
SES quintile
|
1.04
|
1.01 - 1.07
|
0.02*
|
0.99
|
0.99 - 1.00
|
0.09
|
Calendar year
|
0.97
|
0.96 - 0.98
|
<0.01*
|
1.01
|
1.01 - 1.01
|
<0.01*
|
Age
|
1.31
|
1.25 - 1.38
|
<0.01*
|
1.02
|
1.01 - 1.03
|
<0.01*
|
ICS Medication Possession Ratio#
|
1.38
|
1.19 - 1.60
|
<0.01*
|
1.11
|
1.07 - 1.15
|
<0.01*
|
Number of dispensed SABA canisters#
|
1.03
|
1.02 - 1.03
|
<0.01*
|
1.01
|
1.00 - 1.01
|
<0.01*
|
Asthma-related hospital admission#
|
2.90
|
2.48 - 3.40
|
<0.01*
|
1.15
|
1.12 - 1.20
|
<0.01*
|
Asthma-related specialist visit#
|
2.86
|
2.47 - 3.29
|
<0.01*
|
1.05
|
0.99 - 1.12
|
0.09
|
Asthma-related internist visit#
|
1.58
|
1.31 - 1.92
|
<0.01*
|
1.06
|
0.97 - 1.15
|
0.17
|
Asthma-related general practitioner visit#
|
1.26
|
1.14 - 1.40
|
<0.01*
|
1.08
|
1.06 - 1.10
|
<0.01*
|
Comorbidity
|
|
|
|
|
|
|
Group 1 (Myocardial infarction)
|
0.96
|
0.51 - 1.81
|
0.91
|
1.07
|
0.87 - 1.32
|
0.53
|
Group 2 (Congestive heart failure)
|
1.34
|
0.92 - 1.97
|
0.13
|
0.96
|
0.86 - 1.08
|
0.49
|
Group 3 (Peripheral vascular disease)
|
1.57
|
0.98 - 2.51
|
0.06
|
0.91
|
0.80 - 1.03
|
0.13
|
Group 4 (Cerebrovascular disease)
|
1.04
|
0.63 - 1.70
|
0.88
|
1.01
|
0.88 - 1.17
|
0.85
|
Group 5 (Dementia)
|
0.91
|
0.40 - 2.09
|
0.83
|
1.08
|
0.85 - 1.38
|
0.53
|
Group 6 (Chronic pulmonary disease)
|
1.58
|
1.43 - 1.74
|
<0.01*
|
1.17
|
1.13 - 1.20
|
<0.01*
|
Group 7 (Connective tissue disease)
|
1.67
|
1.27 - 2.21
|
<0.01*
|
0.91
|
0.83 - 0.98
|
0.01*
|
Group 8 (Peptic ulcer disease)
|
0.84
|
0.54 - 1.31
|
0.44
|
1.02
|
0.91 - 1.14
|
0.74
|
Group 9 (Mild liver disease)
|
0.80
|
0.57 - 1.11
|
0.18
|
1.00
|
0.92 - 1.08
|
0.98
|
Group 10 (Diabetes without complication )
|
1.22
|
1.03 - 1.44
|
0.02*
|
1.03
|
0.97 - 1.08
|
0.36
|
Group 11 (Diabetes with complication)
|
0.99
|
0.74 - 1.32
|
0.94
|
1.04
|
0.94 - 1.14
|
0.43
|
Group 12 (Paraplegia and Hemiplegia)
|
1.06
|
0.73 - 1.54
|
0.77
|
1.05
|
0.88 - 1.26
|
0.58
|
Group 13 (Renal disease)
|
1.59
|
1.29 - 1.96
|
<0.01*
|
0.97
|
0.91 - 1.05
|
0.48
|
Group 14 (Cancer)
|
1.27
|
1.08 - 1.49
|
<0.01*
|
1.00
|
0.95 - 1.05
|
0.91
|
Group 15 (Moderate or Severe liver Disease)
|
0.92
|
0.68 - 1.26
|
0.61
|
1.01
|
0.95 - 1.07
|
0.74
|
Group 16 (Metastatic Carcinoma)
|
1.56
|
1.26 - 1.92
|
<0.01*
|
0.93
|
0.86 - 1.01
|
0.08
|
Group 17 (AIDS/HIV)
|
0.99
|
0.89 - 1.09
|
0.84
|
0.99
|
0.96 - 1.01
|
0.54
|
* Significant at 0.05 level
# These variables were measured during the preceding period to avoid reverse causality bias in estimates of associations
ICS: Inhaled corticosteroids; SABA: Short-acting beta-agonist; SES: Socio-economic status; OR: odds ratio; RR: relative rate
Sensitivity analysis
When we removed 19.2% of patients with other chronic respiratory conditions, results largely remained the same. The annual relative rate of change for maintenance OCS use was as follows: -4.7% per calendar year, 2.7% per year of age, and 4.3% per year since the inception of asthma. For episodic OCS use, the patterns were as follows: 1.0% per calendar year, 0.4% per year of age, and 0.1% per year since the inception of asthma.