To the best of our knowledge, our study is the first to describeboth seasonal variations of hospital admission and in-hospitalmortality for IPF and non-IPF ILD in the United States in the 11year- period from 2006 to 2016. Our primary findings are that from2006 to 2016, all-cause hospital admission rate of patients withinterstitial lung disease (all interstitial lung disease andIPF-only subgroup) declined but their overall mortality remainedunchanged (except IPF subgroup). Acute respiratory failure relatedadmission account for 23% of all causes and pneumonia 17.6%.Mortality of ILD in general and ILD with acute respiratory failureis highest in winter, up to 8.13% ± 0.60% and 26.3% ± 10.2%respectively. Admission rate for all cause admissions are highestin months from January to April. Subgroup analysis also showedseasonal variations with highest hospitalization rates for allsubgroups (IPF, ARF, pneumonia) in the months from December toApril (winter to early Spring).
Our finding of highest all-cause mortality for all causes ofadmissions and subgroup of acute respiratory failure in the winterwas similar to the findings by Olson et al which used a differentdatabase for analysis [12]. Seasonal variations were observed inhospitalization rates across all subgroups (acute respiratoryfailure, IPF, pneumonia) as well. The two most common explanationsfor winter and early Spring increase in admission rates arerespiratory infection and cold temperature. Cold air couldhypothetically induce hyperpnoea, subsequently cause drying of theairways [13] and inducing proinflammatory substances productionleading to epithelial injury [8]. Infectious etiology was suggestedbecause strong seasonal variations have been reported in COPD,pneumonia and recognized viral illness [14]. There is some evidencethat a colder environment could also prolong the life span ofviruses. Many viruses such as influenza A, RSV andmycoplasma pneumonia which cause infections in humansalmost exclusively in winter to early spring [15, 16]. Oneinteresting findings is that although winter has highest admissionrate for all subgroups (IPF, ARF, pneumonia and ILD in general),the mortality does not have strong seasonal variations inidiopathic pulmonary fibrosis and pneumonia only subgroup. Onehypothesis could be the severity of IPF related admissions andpneumonia has no weather association. We could not find literatureto explain this finding thoroughly and it could be a topic forfuture research.
Respiratory causes of death accounted for 64 %-89% in patientswith ILD [17-19]. We found that acute respiratory failure accountsfor 23% of admission of interstitial lung disease and this types ofadmission has high mortality rate of 26.3% ± 10.2%. This findingconcurs with the results of Moua et al that IPF and non-IPFinterstitial lung disease both have very high and similar mortalityrates after admission for respiratory distress [20]. Based on astudy in Finland, ischemic heart disease, heart failure and lungcancer were the other causes of death [21]. All of those conditionsalso have been reported to have higher mortality in winter time inthe general population [12, 22], which may explain the highermortality in IPF and non IPF ILD patients in winter time
Of note, the in-hospital mortality of interstitial lung diseasewas noted to be significantly higher than the similar study inchronic obstructive lung disease (COPD) and asthma patients usingthe same national database ,8% vs 2%, 8% vs 1% respectively[23,24]. interestingly enough, the mortality rate was 14% higher in thewinter compared to the summer, which was less pronounced than theseasonal variations of all cause of deaths of COPD patients (25% to50% higher in the winter) [12, 25]. Although both COPD andinterstitial lung disease are both progressive illnesses with thepathogenesis involving accelerated cellular senescence[26]. Thisfinding suggests that the impact of weather and viral illness onmortality might not be as pronounced in ILD, compared to COPD.
One of the utmost important roles of physicians is to preventhospital admission for ILD patients. ILD and especially IPF relatedadmissions are significant events after which the lung function ofpatients will significantly deteriorate with the mean survival onlyfrom 2.8 months to 27.7 months [27]. From our study, we found thatall cause admission rates in ILD patients, subgroup of only IPF,acute respiratory failure and only pneumonia in the last 11 yearswere highest in the months December to April (winter to earlySpring). Spring in general had highest admission rates compared tothe average of other seasons, even when infectious lung diseaseswere ruled out. Moineddin et al in their study in the primary caresettings found a higher office visits due to respiratory disease inthe months from December to April [28].
In the period of 11 years from 2006 to 2016, we observed adecrease in admissions rate for all cause hospital admission forILD (all types ILD and subgroup of IPF) with the rise in populationtaken into account. The sharp decrease in 2016 hospital admissionsmight be a result of incomplete report of administrative datapossibly due to the transition from ICD-10 system in the thirdquarter of 2015. In addition, many advances have been introduced indiagnosis and treatment of interstitial lung disease [29] as wellas in hospital management in reducing hospital admissions [30].
The all-cause mortality rate from interstitial lungdisease from 2006 to 2016 has been unchanged. However, theall-cause mortality rate of idiopathic pulmonary fibrosis subgroupencouragingly decreased in this 11-year period. Anti-fibrotictreatment availability could be a possible explanation. A recentlarge database study by Demsey et al reported a decreased mortalityrisk in IPF patients in the first two years of anti-fibrotictreatment [31]. However, it is challenging to pinpoint a singlefactor that lead to this encouraging result based on our studyespecially when antifibrotic therapies were only approved since2014 [32].
Our study has limitations. We did not include all types ofinterstitial lung disease We excluded the interstitial lung diseasegroup with identifiable external agents (organic dust, drug,asbestos, silicosis, pneumoconiosis) because of two reasons.Firstly, it is for the comparison with the results of the study byOlson et al for the interstitial lung disease group from 1992 to2003 [12] and secondly, including ILD group with identifiableexternal agents with different pathogenesis will create moreheterogeneity to our population. IPF cases were identified based ona broad definition algorithm which has been commonly used in theepidemiology studies. However, this algorithm was sensitive but notspecific [33], thus could overestimate the prevalence of IPF in ourILD population. Although we have included all ICD-9-CM andICD-10-CM codes for interstitial lung disease, the results areinevitably susceptible to errors from coding inaccuracies.Nevertheless, this study has provided with an important andobjective overview on the seasonal variations and trends inadmissions and mortality of this entity spectrum over a long periodof time.