Article Text
PDF
Abstract
Objective To identify risk factors for asthma attacks and poor asthma control in children aged 5–16 years.
Methods Prospective observational cohort study of 460 children with asthma or suspected asthma from 10 UK general practices.
Gender, age, ethnicity, body mass index, practice deprivation decile, spirometry and fraction of exhaled nitric oxide (FeNO) were recorded at baseline. Asthma control scores, asthma medication ratio (AMR) and the number of asthma attacks were recorded at baseline and at 6 months.
The above independent variables were included in binary multiple logistic regression analyses for the dependent variables of: (1) poor symptom control and (2) asthma attacks during follow-up.
Results Poor symptom control at baseline predicted poor symptom control at 6 months (OR 4.4, p=0.001), while an increase in deprivation decile (less deprived) was negatively associated with poor symptom control at 6 months (OR 0.79, p=0.003). Higher FeNO levels (OR 1.02, p<0.001) and a recent history of asthma attacks (OR 2.03, p=0.02) predicted asthma attacks during follow-up. Asian ethnicity was associated with a lower OR for a future attack (OR 0.32, p=0.02).
A decrease in AMR was also associated with an increased OR for future asthma attacks (OR 2.99, p=0.003) when included as an independent variable.
Conclusions We identified risk factors for poor symptom control and asthma attacks in children. Routine assessment of these factors should form part of the asthma review to identify children at an increased risk of adverse asthma-related events.
- physiology
- data collection
Data availability statement
Data are available on reasonable request. The datasets used and analysed for this study are available from the corresponding author on reasonable request.
Statistics from Altmetric.com
What is already known on this topic?
Asthma is the most common chronic disease of childhood and is associated with significant morbidity.
There is broad consensus on the need to assess both current asthma symptom control and the risk of future asthma attacks at each asthma review.
The predictive value of spirometry and fraction of exhaled nitric oxide (FeNO) for future asthma attacks in children is unclear.
What this study adds?
This is the first prospective study in UK primary care to explore risk factors for future asthma attacks and poor symptom control in children.
Abnormal spirometry and FeNO are common in children who attend for asthma reviews in primary care.
Raised FeNO is a risk factor for future asthma attacks. FeNO testing should be funded for use in all care settings.
Introduction
Asthma is the most common chronic illness in children and is associated with significant morbidity and mortality.1 In the UK, a child is admitted to hospital with an asthma attack every 20 min.2
Current guidelines3–6 and the 2017 Lancet Commission on asthma7 highlight the need to incorporate assessment of current asthma control and attack risk at every review.
A recent systematic review identified several risk factors for asthma attacks in children aged 5–12 years.8 A previous asthma attack was the most strongly predictive. Other factors included: persistent symptoms, poor access to healthcare, suboptimal drug regimen (defined as a low preventer-to-reliever inhaler ratio), comorbid atopic/allergic disease, African-American ethnicity, poverty, tobacco smoke exposure, obesity and younger age.
Both the UK National Institute for Health and Care Excellence (NICE) and Global Initiative for Asthma recommend the use of spirometry for monitoring. However, Buelo et al 8 found inconclusive evidence to support the use of either spirometry or fraction of exhaled nitric oxide (FeNO) for asthma monitoring. Notably, only one study included in this review was from the UK and did not include spirometry or FeNO, limiting the generalisability of the evidence to the UK paediatric population.
Using data collected as part of the ‘Childhood Asthma Management in Primary Care - Implementation of Exhaled Nitric Oxide and Spirometry Testing (CHAMPIONS)’ study (NCT02913872), we aimed to identify risk factors for poor asthma symptom control and asthma attacks in children aged 5–16 years managed in UK primary care.
Methods
Study design
CHAMPIONS was a prospective observational study conducted in 10 general practices in the East Midlands, UK, designed to evaluate the implementation and clinical outcomes9 related to the delivery of spirometry and FeNO testing for children in primary care.
Children aged 5–16 years on the practice asthma register and children with suspected asthma, not on the asthma register but who, in the previous 12 months, were prescribed: inhaled corticosteroids (ICS) OR ≥2 short acting beta-2 agonist (SABA) inhalers OR oral corticosteroids for acute wheeze/cough/breathlessness were recruited.
Practice nurses performed clinical reviews independently. Lung function testing was performed by practice staff on the same day as the clinical review, supported by the research team.10 Patient records were interrogated by the research team for the included study variables.
Written consent was sought from all participants by the research team.
Study variables
We recorded baseline patient characteristics on the day of the asthma review.
The asthma control test (ACT)11 (≥12 years) and the childhood asthma control test (CACT)12 (4–11 years) were used to assess current symptom control. A score of <20 denotes current uncontrolled asthma.6 Follow-up questionnaires were posted 6 months following the review.
Asthma attacks
Electronic patient records were interrogated at baseline and at 6 months to identify the number of asthma attacks 6 months preasthama and postasthma review. An asthma attack was defined as any unplanned healthcare attendance with acute respiratory symptoms managed with asthma medications.
Socioeconomic status
We used the social deprivation decile of each patient’s registered practice as a surrogate marker for an individual participant’s socioeconomic status.
This is based on the Index of Multiple Deprivation 2015,13 an overall measure of deprivation experienced by people living in an area, which ranges from 1 (most deprived) to 10 (least deprived).
Asthma medication ratio (AMR)
The AMR is calculated from the number of prescriptions for ICS divided by the number of prescriptions for both ICS and SABA over a given time period.
AMRs range from 0 (only SABAs prescribed) to 1 (only ICS prescribed).14 Lower AMR scores suggest suboptimal treatment.
Spirometry and FeNO testing
All children attempted both tests. Spirometry was performed using a portable spirometer (CareFusion UK Ltd) according to American Thoracic Society/European Respiratory Society standards.15 Spirometry results were expressed as z-scores, and lower limits of normal (LLN) were calculated using Global Lung Initiative reference equations. Abnormal spirometry was defined as forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) or FEV1/FVC <LLN.16 FeNO measurements were attempted following spirometry, in accordance with the NICE diagnostic algorithm and standardised guidelines, using a hand held FeNO analyser (Circassia).17
Data analysis
Statistical analyses were carried out using IBM SPSS software V.24.0 (Armonk, New York, USA). Descriptive statistics are expressed using the mean (95% CIs) or median (IQR) for continuous variables and percentages for categorical variables.
Binary multiple logistic regression models were fitted for the dependent variables of: (1) ≥1 asthma attack during follow-up (yes/no) and (2) poor symptom control (ACT/CACT ≤20) at follow-up (yes/no). Independent variables included were: age, gender, history of asthma attacks at baseline (yes/no), poor symptom control at baseline (yes/no), FEV1 z-score, FEV1/FVC z-score, FeNO as a continuous variable, AMR at baseline, deprivation decile, body mass index (BMI) (overweight/obese or not)18 and ethnicity.
To control for changes in asthma medication prescriptions following the review, we repeated the regression analyses using the change in AMR from baseline (decrease, no change and increase) as an independent categorical variable instead of AMR at baseline.
Collinearity between independent variables was tested for by calculating the variance inflation factor (VIF).
A receiver operator characteristic (ROC) analysis was performed to determine the FeNO level that best predicted an asthma attack during follow-up.
All statistical tests were performed at the alpha=5% level.
Results
Six hundred and fourteen (40%) children attended reviews between June 2016 and August 2017, and 612 families consented to participate. Four hundred and fifty-six children attending clinic were on their GP asthma register and 158 (26%) were not.
Four hundred and sixty-five (76%) children were able to perform spirometry and FeNO. Children able to perform both were older, 11.0 (10.5–11.5) years versus 7.0 (6.6–7.4), p<0.001.
Five children moved practices during the study; their medical records were unavailable at follow-up.
Patient characteristics
Baseline characteristics are shown in table 1 and follow-up characteristics in table 2.
Baseline characteristics of children with spirometry and FeNO data (n=465)
Characteristics of children with follow-up data available (n=460 unless otherwise stated)
Twenty-five per cent of our cohort had an abnormal spirometry result. The median FeNO was 23 ppb.
Poor symptom control was reported by 41% of children at baseline, reducing to 25% at follow-up. Asthma attacks were documented in 22% of children in the 6 months preceding their review, reducing to 16% at follow-up.
The proportion of children prescribed ≥1 SABA inhaler remained the same at around 70% pre and post follow-up, but the proportion prescribed ≥1 ICS inhaler increased from 54% to 68%. Mean AMR increased from 0.53 to 0.60.
Risk of future poor symptom control
One hundred and seventy-five (38%) families returned the follow-up ACT/CACT questionnaires and were included in the binary multiple logistic regression analysis.
Poor symptom control at baseline (CACT/ACT<20) was associated with an OR of 4.4 (95% CI 1.9 to 10.2; p=0.001) and each increase in deprivation decile (less deprivation) an OR of 0.79 (95% CI 0.68 to 0.93; p=0.003) of having poor symptom control at follow-up (table 3).
Binary multiple logistic regression analysis for the likelihood of poor symptom control (ACT/CACT <20) at follow-up (n=175)
The VIF for each independent factor was <2, demonstrating no evidence of collinearity.
Risk of future asthma attacks
Follow-up data on asthma attacks were available for 460 (99%) children, who were all included in this analysis (table 4).
Binary multiple logistic regression analysis for the likelihood of ≥1 asthma attack during follow-up (n=460)
Increasing levels of FeNO (OR 1.02, 95% CI 1.01 to 1.02; p<0.001) and ≥1 asthma attacks in the 6 months preceding baseline assessments (OR 2.03, 95% CI 1.15 to 3.60; p=0.02) were found to be independently associated with ≥1 attack during follow-up. Asian ethnicity was associated with a lower OR for an attack (OR 0.32, 95% CI 0.13 to 0.83; p=0.02).
A ROC analysis (figure 1) determined the best cut-off value of FeNO to predict an asthma attack was 58 ppb, providing a sensitivity of 36% and specificity of 82%.
ROC curve of FeNO as a predictor for future asthma attacks. Area under the curve (AUC=0.567). The optimum cut-off point for FeNO is 58 ppb (sensitivity 36%, specificity 82%). FeNo, fraction of exhaled nitric oxide; ROC, receiver operator characteristic.
We repeated the binary multiple logistic regression analysis with FeNO as a binary variable (online supplemental table S1). Using FeNO ≥58 ppb as a cut-off provided an OR of 3.35 (95% CI 1.83 to 6.15; p<0.001) of an attack during follow-up. In the 6 months following the review, 27 out of 97 (28%) children with FeNO ≥58 ppb had at least one asthma attack, compared with 48 out of 363 (13%) children with FeNO <58 ppb.
Supplemental material
Impact of prescription changes
Binary multiple logistic regression analyses were repeated using the change in AMR (rather than baseline AMR) as an independent variable to see whether changes to treatment had an impact on the prediction models (online supplemental tables S2 and S3). The same variables identified as significant predictors in the original analyses remained significant. No other variables were significantly associated with future loss of symptom control. However, a decrease in AMR during the follow-up period was associated with an increased OR of an asthma attack during the same period (OR 2.99, 95% CI 1.44 to 6.21; p=0.003).
Discussion
To our knowledge, this is the first prospective paediatric study in UK primary care to explore patient risk factors for future asthma attacks and poor symptom control.
Neither low lung function nor raised FeNO were predictors for future poor symptom control. This is consistent with previous studies, which have reported a poor relationship between asthma symptoms with both lung function and FeNO.9 19 Only poor symptom control at baseline and higher levels of deprivation were associated with poor symptom control at follow-up.
The relationship between socioeconomic deprivation and poor asthma control in children has been reported before.20 The reasons for this relationship are multifactorial.21 Nevertheless, there have been reports of successful strategies to improve asthma control in vulnerable children, including the provision of asthma education and frequent reviews.21
Alongside a history of an asthma attack, FeNO was found to independently predict for future attacks. A recent systematic review also identified previous asthma attacks as a risk factor for future attacks, with a similar OR (range 2.1–4.1) to ours (OR 2.0; 95% CI 1.2 to 3.6) but found insufficient evidence to reach a conclusion regarding FeNO.8
We found the odds of having an asthma attack were over three times larger for children with FeNO ≥58 ppb compared with <58 ppb. This is consistent with a recent Cochrane review, which concluded that FeNO-targeted asthma monitoring may reduce the frequency of future asthma attacks22 in children.
When ‘change in AMR’ was included as an independent variable, a decrease in AMR was associated with an OR of 3.0 for an asthma attack over the same period, suggesting that either a reduction in ICS or increase in SABA use are risk factors for attacks.
We observed a non-significant trend towards more attacks in females and younger children. Previous studies have also identified younger age as a risk factor, but the evidence for gender is less consistent.8
In our study, Asian children had a lower OR for a future asthma attack, contrasting with previous studies reporting higher rates of asthma-related primary care and hospital visits among children from South Asian families.23 24 As attendance for an asthma review was voluntary in our study, it is possible that our cohort was biased generally towards families more engaged with their child’s asthma management.
A BMI ≥91st centile was not a significant risk factor for attacks or poor symptom control in our cohort. Buelo et al reported a small, but significant, increased risk of attacks in overweight or obese children. However, due to the small effect size, only the largest studies (n>4000) in their review observed a significant association. Our cohort may have been underpowered to detect this association.
We found no association between low lung function and risk of future asthma attacks or poor symptom control. However, the high prevalence of abnormal lung function in our cohort of children managed within the community is concerning and highlights that subjective assessments of a child’s asthma and lung health are inadequate on their own.25
Longitudinal studies of lung function have demonstrated that lower lung function trajectories are associated with adverse health outcomes26 and low lung function in later adult life,27 28 including chronic obstructive pulmonary disease29 and early cardiopulmonary-related mortality.30 Therefore, identification of persistently low lung function in children warrants referral to a specialist service for assessment.
A quarter of children eligible to participate in this study were not on their GP practice’s asthma register. Without a formal diagnosis and addition to the register, these children are not reviewed and risk being overtreated or under-treated with asthma medications.
Limitations
Despite two written invitations, <50% eligible children were brought for an asthma review. It is possible that only more symptomatic children attended, biasing our cohort towards those with poorer control at baseline. Without spirometry or FeNO data in children who did not attend for comparison, the high prevalence of abnormal results we observed may be partially explained by this selection bias. Moreover, we did not obtain data on asthma attacks in non-attenders. The lack of data for non-attenders limits the generalisability of our findings.
Follow-up CACT/ACTs were only returned by 175 families, introducing further potential bias to our analyses.
We did not set out to control for any effects of seasonality and only followed patients for 6 months. However, we recruited patients over a 14-month period across all four seasons, mitigating any seasonal influence on our findings.
Healthcare professionals performing reviews were not blinded to spirometry or FeNO results. In conjunction with patient-reported symptom control, abnormal test results may have prompted a review and optimisation of asthma treatment plans. The relationship between objective tests and future risk of attacks might then have been underestimated. We performed further data analysis, including a ‘change in AMR’ as an independent variable, to attempt to control for this.
We included deprivation data at the practice level only. This may overestimate or underestimate individual deprivation depending on the socioeconomic diversity of a practice’s catchment population.
Finally, the CHAMPIONS study was not initially designed to investigate risk factors for asthma attacks or loss of symptom control. Therefore, the number of clinical variables available for regression modelling were limited to those collected in the original study. These did not include other potentially important factors identified in a recent systematic review, including atopy, poor access to care or exposure to environmental tobacco smoke.31
Conclusions
We have identified several risk factors for poor asthma-related outcomes in children managed in UK primary care. Any child with a history of a recent asthma attack, poor symptom control, raised FeNO or a decrease in their ICS to SABA use should prompt a structured review, with standardised assessments, to optimise asthma management.
We welcome the recent changes to the primary care quality outcomes framework (QOF) in England,32 which now incentivises the use of a validated ACT questionnaire, and documentation of previous asthma attacks, as part of the annual asthma review. Based on our findings, we feel that review of medication prescriptions, and the use of FeNO for monitoring, should be included as part of the asthma QOF and funding made available to support purchase of equipment.
However, incentivising these additional measures as part of a tick-box only exercise is not enough, and any initiative to improve asthma care must include the appropriate provision of paediatric-specific training to professionals who look after children with asthma.33
Data availability statement
Data are available on reasonable request. The datasets used and analysed for this study are available from the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
Ethics approval was obtained from the NHS research ethics committee (16/EM/0162).
Acknowledgments
We would like to thank the staff, children and families at all participating general practices for their support of this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @david__lo, @damian_roland
Contributors EAG, DL, CB, DR, MR, YY and AW participated in the initial design of the study. DL and LD were responsible for data collection and providing training to the general practices. DL was responsible for data analyses and initial manuscript preparation with input from EAG and CB. Statistical input was provided by MR. All coauthors contributed to the interpretation of results and provided revisions and approval of the final manuscript.
Funding This study was funded from grants provided by the Midlands Asthma and Allergy Research Association and Circassia Pharmaceuticals to EAG. The project fellow (DL) was funded by Health Education East Midlands.
Competing interests EAG: consultancy work for Boehringer Ingelheim in November 2016 and Anaxsys in July 2018 with money paid to the institution (University of Leicester), investigator led research grant from Circassia and Gilead. Research collaboration with Medimmune. Travel grants from Vertex.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.