ILDs are a heterogeneous group of disorders characterized by the development of lung fibrosis. The pathogenesis implicates a series of inflammation and fibrosis disrupting the interstitial bed and changing the lung parenchyma. Causing factors can be exogenous or endogenous. ILDs without identifiable causes are considered idiopathic.

ILD can be classified on the basis of histological and radiological parameters.

The main alterations concerning the interstitium and air space of the lung are ground-glass opacities, reticular markings, thickened airways, diffuse centrilobular nodules, pulmonary fibrosis, desquamative interstitial pneumonia (DIP), and others [13].

Among the exogenous causes of ILD, illicit drugs have been reported [1,2,3,13,14,15,16,17,18].

The substance of abuse most used by our patients was cannabis.

In Group 1, ILD emerged at chest CT in 75% of the subjects with cannabinoid-positive BAL and in 20% of the subjects with cannabinoid-negative BAL. This statistically significant difference (p = 0.0299) suggests the possible role of cannabinoids in causing ILD. Furthermore, the respiratory damage of second-hand tobacco smoke is known, but not that of cannabis smoke. Since cannabis is the most widespread psychotropic substance in the world, the observation of a patient passive smoker, who had a cannabinoid-positive BAL from second-hand smoke, suggests further studies, even in subjects who are not active users.

Moreover, in Group 1, among the nine patients with chest CT signs of ILD and cannabinoid-positive BAL, the main cannabis-induced organ lesions were ground-glass opacities in 44.4%, tree-in-bud sign in 22.2%, bilateral micronodules in 11.1%, and consolidations in 11.1%. Thus, this evidence would suggest, in cases where these radiological patterns are found, to better investigate the possible use of illicit drugs.

In Group 2, ILD was evident at chest CT in 4/10 (40%) patients with cannabinoid-positive surgical specimens, and in 0/5 (0%) with cannabinoid-negative surgical specimens. In this case, the association between cannabinoids in lung tissue and ILD that emerged from chest CT was not statistically significant (p = 0.2308); however, the complete absence of ILD radiological signs in the patients with cannabinoid-negative surgical specimens further suggests a role of cannabinoids in favoring ILD. Nevertheless, at histological examination, the lung tissue of the four cannabinoid-positive patients with ILD on chest CT scan was characterized by interstitial fibrosis and DIP-like reaction, thus suggesting a possible correlation among cannabinoid-positive lung tissue, pathological findings of ILD on the same tissue, and chest CT evidence of ILD.

In addition, in Group 2, among the four cannabinoid-positive patients with chest CT signs of ILD, we observed ground-glass opacities, consolidations, and micronodules, again suggesting, in case of these radiological findings, to consider a possible history of illicit drug use.

Moreover, the difference in chest CT findings of ILD, in 75% of the patients of Group 1 subjected to BAL, compared to 40% of those of Group 2 operated for pneumothorax, could be explained by the longer time of exposure to cannabis smoke. In fact, in the patients of Group 1, who were older, with a mean age of 38.42 years, the average duration of exposure to cannabis smoke was longer, 15 years, thus justifying the greater number of subjects with radiological evidence of ILD. In Group 2, composed of younger patients, mean age 26.86 years, the shorter average time of exposure to cannabis smoke, 9.5 years, may explain the smaller number of subjects with chest CT detection of ILD. This suggests the harmful action of cannabis in relation to the time of exposure.

Even the cellularity that emerged in the histology of surgical specimens in patients of Group 2 supports the hypothesis of the damaging action of cannabis in causing ILD. In fact, the respiratory system is subjected to the chronic irritative insult of tobacco and cannabis smoke, associated with the presence of pollutants responsible for the cellularity present in histological samplings. In the damaged tissue, neutrophils and monocytes are activated, which release toxic factors: proteases, elastase, and collagenase. The chronic irritative stimulus would favor the formation of emphysematous bullae earlier in cannabinoid-positive lung tissue [11]. The presence of macrophages or lymphoplasma cells or foreign body giant cells (Table 3) supports the action of cannabinoids in promoting the nonspecific chronic inflammatory process, with secondary lysis of the interalveolar septa, site of the chronic flogistic process, with interstitial fibrosis/DIP.

In our patients, cellularity involved in a nonspecific chronic inflammatory process was present in 9/10 (90%) patients with cannabinoid-positive lung tissue and in 0/5 (0%) patients with cannabinoid-negative lung tissue. In cannabinoid-positive patients and in cannabinoid-negative ones, respectively, we observed macrophages in 6/10 (60%) and in 0/5 (0%), lymphoplasma cells in 2/10 (20%) and in 0/5 (0%), and foreign body giant cells in 3/10 (30%) and in 0/5 (0%), thus defining, as far as we know, for the first time in the literature, this phenotype associated with cannabinoids in surgical specimen.

Other studies suggest that marijuana causes pulmonary fibrosis [19,20]. In an autopsy study, the intensity of macrophage infiltration appears to be related to the dose of marijuana taken [19].

Desquamative interstitial pneumonia (DIP) is reported in users of large quantities of marijuana [20,21]. In our study, DIP-like reaction (Table 3) was found in 5/10 (50%) patients with cannabinoid-positive lung tissue and in 1/5 (20%) with cannabinoid-negative lung tissue. We observed the accumulation of macrophages in the five cannabinoid-positive patients with DIP-like reaction; absence in the only cannabinoid-negative patient with DIP-like reaction (patient n. 14, Table 3). Our finding suggests the responsibility of cannabis in causing DIP, especially taking into consideration the young age of the patients (24–27 years) (Table 3). Indeed, in DIP from other causes, the average age of onset of symptoms is between 40 and 60 years [22,23]. Not only can drugs be responsible, but also additives, adulterants, and cutting substances, such as cellulose, talc, and starch [2,6,24], as in the patient in Figure 2: the research in the granuloma was crucial for a correct etiological diagnosis.

Among the manifestations caused by substances of abuse, ground-glass opacities have been described. In a study of cocaine users, in a subset of patients with chest CT evidence of ILD, ground-glass opacities were present in 100% of the cases, consolidations in 50%, and nodular formations and tree-in-bud pattern in 12.5% [2]. This study, based on statistical–radiological data, suggests the responsibility of cocaine, and ours, based on statistical–radiological–toxicological data, further supports it. Indeed, in the four patients of Group 1 with cocaine-positive BAL (Table 1), chest CT showed ground-glass opacities in 25%, consolidations in 50%, and tree-in-bud pattern in 25% of cases.

To formulate an etiological diagnosis, the substance of abuse is usually searched for in the blood, but its diagnostic reliability has not yet been proven. Thus, to verify the usefulness of searching for the substance of abuse in the blood or in the organ where the disease is located, we studied a subgroup of patients of Group 1. In the subgroup, eliminating the confounding elements, represented by the four patients who simultaneously presented the substance in BAL and in blood (thus suggesting a transfer of the substance from the respiratory tract to the blood), 55.6% of the remaining patients resulted positive in BAL and none of them in blood, with a statistically significant difference (p = 0.0294). Although it was expected to find the substance in the blood of all the patients, this was not the case, underlining the usefulness of searching for the drug in the organ where the disease is located, thus in BAL and not in the blood, to formulate a reliable etiological diagnosis.

Lastly, our enrolled patients did not have any chronic diseases, infectious factors, nor were they users of other medications such as fibrotizing drugs that might have been responsible for the documented ILD.

This study has some limitations: the relatively limited number of patients and consequent power limitation due to the small sample size, especially the cannabinoid-negative cases, which could have been a confounding factor in the comparative analysis between positive and negative cases, causing potential biases; the lack of a true control group; the retrospective design of the study; and difficulties in recruiting patients who were reluctant to provide informed consent for the toxicological examination of their BAL or lung tissue or blood. Most of our patients were active tobacco smokers, which could have been a confounding or contributing factor in the pathologic findings. However, to the best of our knowledge, there are no studies demonstrating that tobacco smoking is associated with the direct development of ILD. Furthermore, co-exposure to other drugs might have some importance, but the major role should be attributed to the drug identified in the examined biological samples. The simple anamnestic data of substance abuse is relative. To ascribe a major role to the drugs used and not identified in biological samples is only a hypothesis of responsibility, as reported in other previous studies where only the anamnesis acted as an etiological diagnostic guide. Thus, not having found other responsible etiological factors, we should ascribe responsibility to cannabis in the cases in which it has been found in biological samples.