Is insomnia a risk factor for new-onset asthma? A population-based study in Taiwan

Objectives To determine whether insomnia at baseline is a risk factor for new-onset asthma. Methods We recruited 48 871 patients with insomnia (insomnia group) newly diagnosed between 2002 and 2007, and 97 742 matched controls without insomnia (control group) from Taiwan’s Longitudinal Health Insurance Database 2000. All of the patients were followed up for 4 years to see whether new-onset asthma developed. Patients with previous asthma or insomnia were excluded. The Poisson regression was used to estimate the incidence rate ratios (IRRs) and 95% CIs of asthma. Cox proportional hazard regression was used to calculate the risk of asthma between the two groups. Results After a 4-year follow-up, 424 patients in the insomnia group and 409 in the control group developed asthma. The incidence rate of asthma was significantly higher in the insomnia group (22.01vs10.57 per 10 000 person-years). Patients with insomnia have a higher risk of developing new-onset asthma during the 4-year follow-up (HR: 2.08, 95% CI 1.82 to 2.39). The difference remained significant after adjustment (adjusted HR: 1.89, 95% CI 1.64 to 2.17). Conclusions This large population-based study suggests that insomnia at baseline is a risk factor for developing asthma.

The lack of data on BMI is a major problem, since obesity is overrepresented with insomnia and considered as a key factor in asthma development among adults. If adjustments could have been made for BMI (and, simirlarly, for smoking history) it would have added strength to the results. This very important limitation should at least be more clearly addressed. The results were adjusted for sleep apnea, but the prevalence was surprisingly low in both groups. Adjustments were also made for anxiety and depression, but also these conditions tend to be under-reported using only DSM-9 CM codes.
The authors refer to Sivertsen's paper from the Norwegian HUNT study published in 2014. However, I think that another, more recent and highly relevant publication from the HUNT study by Brumpton et al: Prospective study of insomnia and incident asthma in adults: the HUNT study (ERJ, 2017) also should be referred to.

GENERAL COMMENTS
In this prospective cohort study the authors have investigated insomnia as a risk factors for new onset asthma in Taiwanese population and found that patients with insomnia had almost twice the risk of developing asthma.
Approximately 50,000 patients with newly diagnosed insomnia and twice the number of matched control were followed up by four years for new onset asthma.
The diagnostic tool used for diagnosis of insomnia is not stated except that it was diagnosed during one hospital admission and three outpatient visits within a year. Insomnia is a syndrome ranging from acute transient sleep disturbance to chronic sleep disruption.
Similarly the basis for new onset asthma during follow up is not explained except that it was diagnosed during one hospital admission or three clinic visits. Prevalence of atopy related disorders (allergic rhinitis, atopic dermatitis and urticaria) was significantly higher in insomnia population than control group. This could influence diagnosis of new onset asthma in insomnia population. Authors need to discuss possible effect of above differences in diagnosing new onset asthma-despite Authors need to provide more details for inclusion and exclusion criteria.
Authors have not clarified if patients with asthma were actively excluded from case and control group and basis of exclusion.
Insomnia is dynamic disease and it might have changed /resolved in insomnia group and might have developed in control group during follow up-were above changes monitored during follow up.

Malvika Sagar
Mc Lane Children's Hospital/Baylor Scott and White REVIEW RETURNED 23-Sep-2017

GENERAL COMMENTS
It is a very large , well written study where authors have addressed a very important question-is insomnia a risk factor for asthma. I have few minor questions related to the paper. 1. Under the material and method section lines 41-44, Patients diagnosed with asthma before insomnia were excluded. It is possible that the patients with symptoms of asthma and not diagnosed yet may have been included. You may add that in limitation of the study.
2. Figure 1: Please consider rephrasing the legend to 'The probability of developing asthma in insomnia and control group'.
3. Sometimes in adult population COPD can be misdiagnosed as asthma and vice versa. The same patient may have been diagnosed with COPD and asthma. Could you look into that?
4. Do you have data on how many patients with insomnia were treated? It would be interesting to note if there were any differences in outcome in those insomnia patients who were treated versus those who were not. This may add weight to your conclusion that proper treatment of insomnia patients might help prevent the progress of airway inflammation.
5. Your sample size is very large. However, for the sake of completeness, did you do the power calculation?

VERSION 1 -AUTHOR RESPONSE
Reviewer: 1 Reviewer Name: Fredrik Sundbom Institution and Country: Uppsala University, Sweden Comment: It was showed that having an insomnia diagnosis at baseline was a risk factor for development of new-onset asthma, and that the difference remained significant after adjustment for comorbidity. The paper is structured and follows the STROBE guidelines. The study benefits from a large population and a long observation time. However, I have some concerns, mainly about the methods and the definitions used.
Both asthma and insomnia were defined by DSM-9 CM codes. This may work well for asthma, but it may result in a false low prevalence of insomnia, since only individuals with severe insomnia actually have a insomnia diagnosis. Hence, the present study is difficult to compare to previous studies in the same field, of which most use questionnaires on insomnia symptoms. It may also lead to sampling bias, since a higher number of individuals likely to seek medical care may be found in the insomnia group.
Reply: We agree with you that ICD-9 CM code methods can only help us identify severe insomnia patient who needs medical help. However, this method has been applied for several previous studies (Ref. ). Therefore, this kind of analysis should be acceptable. In addition, due to the method differences, the findings in this study may not be used to compare to previous studies using questionnaires on insomnia symptom. Finally, the issues regarding sampling bias, such as a higher number of individuals likely to seek medical care may be found in the insomnia group were added as one of limitations. Comment: The lack of data on BMI is a major problem, since obesity is over-represented with insomnia and considered as a key factor in asthma development among adults. If adjustments could have been made for BMI (and, similarly, for smoking history) it would have added strength to the results. This very important limitation should at least be more clearly addressed. The results were adjusted for sleep apnea, but the prevalence was surprisingly low in both groups. Adjustments were also made for anxiety and depression, but also these conditions tend to be under-reported using only DSM-9 CM codes.
Reply: Lacking of some important confounding factors, such as BMI and smoking history, is the major limitation of our study. The under-reported prevalence of some related disease is also a limitation. We have listed them in our limitation section.
Comment: The authors refer to Sivertsen's paper from the Norwegian HUNT study published in 2014. However, I think that another, more recent and highly relevant publication from the HUNT study by Brumpton et al: Prospective study of insomnia and incident asthma in adults: the HUNT study (ERJ, 2017) also should be referred to.
Reply: Thank you for providing the new reference for our study. We have reviewed the article and added it to our reference.

Reviewer: 2
Reviewer Name: Dr Himender Makker Institution and Country: University College London Hospital, London UK Comment: In this prospective cohort study the authors have investigated insomnia as a risk factors for new onset asthma in Taiwanese population and found that patients with insomnia had almost twice the risk of developing asthma.
Approximately 50,000 patients with newly diagnosed insomnia and twice the number of matched control were followed up by four years for new onset asthma.
The diagnostic tool used for diagnosis of insomnia is not stated except that it was diagnosed during one hospital admission and three outpatient visits within a year. Insomnia is a syndrome ranging from acute transient sleep disturbance to chronic sleep disruption. Similarly the basis for new onset asthma during follow up is not explained except that it was diagnosed during one hospital admission or three clinic visits. Prevalence of atopy related disorders (allergic rhinitis, atopic dermatitis and urticaria) was significantly higher in insomnia population than control group. This could influence diagnosis of new onset asthma in insomnia population. Authors need to discuss possible effect of above differences in diagnosing new onset asthma-despite Reply: Thanks for the reviewer's comment. The definition of insomnia diagnosis was used the ICD-9-CM codes according to the previous published papers. Therefore, we think the definition of insomnia diagnosis may be acceptable. Moreover, we understand the prevalence of atopic diseases was higher in the insomnia group. However, the incidence asthma was still significant higher in the insomnia group after full adjustment.
Comment: Authors need to provide more details for inclusion and exclusion criteria.
Reply: Thanks for reviewer's comment. We add a flowchart for more clearly description.
Comment: Authors have not clarified if patients with asthma were actively excluded from case and control group and basis of exclusion.
Reply: Thanks for reviewer's comment. We add a flowchart for more clearly description.
Comment: Insomnia is dynamic disease and it might have changed /resolved in insomnia group and might have developed in control group during follow up-were above changes monitored during follow up.
Reply: Thank you for your comments. We understand that the Insomnia is dynamic disease, so the control group was selected as patients without any insomnia records. According to your advice, we revised the relevant part in the manuscript. Your questions are answered as follows.
We excluded all patients with any diagnosis of insomnia during year 2002 to 2011 in the control group.

Reviewer: 3
Reviewer Name: Malvika Sagar Institution and Country: Mc Lane Children's Hospital/Baylor Scott and White