Table 2.
Evaluation on the causality between humidifier disinfectants and lung injury based on Hill’s criteria
| Criterion |
Evidence | Met or not | |
|---|---|---|---|
| Hill | US Surgeon General Expert Committee | ||
| Strength of association | Strength of association | In a previous case-control epidemiological study, the OR of HD exposure (95% CI) was 47.3 (6.1, 369.7) in adults (hospital control group) [12], 116.1 (6.5, 2,063.7) (community control group) [14], and 2.73 (1.41, 5.90) in children [13], showing strong association | Met |
| Consistency | Replication of the findings | Association was found both in adults [12,14] and children [13], and in a case-control study, consistent results were found in different control groups (hospital [12,13], community [14], and family [15]); In addition, a significant association was reported not only in a case-control study, but also in a retrospective cohort study [16] | Met |
| Specificity | Specificity of the association | In an epidemiological study, lung disease of unknown cause could not be explained by other causes than HDs [12,13]; It was not consistent with clinical, radiological, and pathological findings of lung disease of other well-known causes such as viral, bacterial, or immunological causes [7,9] | Met |
| Temporality1 | Temporal relationship | Lung injury of unknown cause had not been reported before HD was introduced to the market | Met |
| Biologic gradient | Dose-response relationship | As a result of a community-based case-control study, OR increased as the amount and period of HD use increased [14]; In a case-control study with a family control group, the increased exposure-OR association was shown according to sleep time, time to use humidifier per day, disinfectant concentration in the atmosphere, and the distance between a bed and humidifier in a room where humidifier containing disinfectant is turned on [15]; In a nationwide report of patients, lung injury or relevant mortality risk increased as the concentration was high in case of long and repetitive use of HD [16] | Met |
| Plausibility | Biological plausibility | As the size of aerosol containing HD sprayed through humidifier was <100 nm, it has been proven that only a small size can reach the peripheral bronchiole and get precipitated [1]; Lung injury was induced in an intra-tracheal drip animal study using a diluted concentration similar to the concentration of HDs that was on the market [1] | Met |
| Coherence to previous knowledge | Consistency of other knowledge | Previously known inhalation toxicity-induced lung lesions have characteristics such as lobular, diffuse infiltrative, and peri-bronchial infiltration, which were also shown in lung disease of unknown cause [7,9]; In toxicity evaluation reports published in other countries, there was no evaluation of inhalation toxicity as the major components of HDs had low volatility at room temperature. However, toxicity from oral exposure or dermatologic transmission was reported [1] | Met |
| Experiment | In a cellular toxicity experiment exposing normal pulmonary cells to the major components of HD, dose-dependent toxicity was expressed, and dose-dependently reactive oxygen was developed as a result of evaluating reactive oxygen production [1]; In an inhalation animal study using diluted concentration that was present on the market, histopathological findings similar to those found in patients with lung disease of unknown cause were observed [17] | Met | |
| Analogy | Ardystil syndrome, an interstitial lung disease developed in workers using spray paint in a western country in the 1990s, is similar to the case of lung disease due to HDs; A component of paint, acramin, has a very similar chemical structure to that of polyhexamethylene guanidine , which is found in HDs [18] | Met | |
| Consideration of alternative explanations | In an epidemiological study, the degree of the association between fungi and lung injury was substantially small compared to the association between HDs and lung injury; In other epidemiological studies [12,13], as there was no association between fungi and lung injury, presence of fungi can be interpreted as resulting from the use of the humidifiers; In case of hypersensitivity pneumonitis (humidifier fever) caused by toxins (endotoxin) from bacteria colonizing in humidifier [19] and white-dust related fever due to inhalation of metal deposits such as calcium and magnesium included in the water of humidifiers [20], there is distinctive difference in clinical, radiological and pathological findings compared to HD-induced lung injury | Met | |
| Cessation of exposure1 | There is no new incidence since HDs have been withdrawn from the market in November 2011; Despite its withdrawal, disease progress was irreversible in patients who had developed the disease before | Met | |
OR, odds ratio; CI, confidence interval; HD, humidifier disinfectant.
1
Hill’s criteria do not have exposure cessation items, the fact that there is no disease incidence after exposure cessation can be interpreted as temporal relationship.