Can smoking reduce the incidence and prevalence of non-tuberculous mycobacterial pulmonary disease?

INTRODUCTION

The genus Mycobacterium comprises more than 200 members, which are differentiated based on pathogenicity, virulence, response to drugs, growth characteristics, and in vivo adaptation.^[1,2] Mycobacteria other than Mycobacterium tuberculosis complex and Mycobacterium leprosy are known as “Non-tuberculous mycobacteria (NTM)” and in the past, these have been known by various acronyms such as “Mycobacteria other than tuberculosis (MOTT)” or saprophytic mycobacteria. They attracted medical attention only after the acquired immunodeficiency syndrome (AIDS) epidemic, but predominantly from tuberculosis (TB) industrialized world and rarely from developing countries, which are also TB-endemic countries. Most probably because the chances of missing the NTM species are higher in TB-endemic countries, due to overworked manpower with TB diagnostic and management of TB cases. The information regarding their true incidence and prevalence in these countries is scarce.^[1] However, NTM can cause various types of clinical conditions, but pulmonary disease (PD), often called NTM-PD, is the most common and severe.

The NTM-PD is a growing health concern for pulmonologists and microbiologists. It is estimated that the prevalence of NTM-PD is around ~6.5/100,000 in European countries, which is rising in many geographical regions globally, but has not yet been systematically estimated.^[2] Co-existing lung diseases, including “bronchiectasis,” “chronic obstructive pulmonary disease (COPD),” and “cystic fibrosis (CF),” are common conditions associated with NTM-PD. There are some well-established risk factors, including smoking, COPD, emphysema, malnutrition, and bronchiectasis. Several studies across the globe observed the same risk factors described, along with a previous history of TB, steroids (given in any form), age, and interstitial lung disease.^[3]

In a study from Japan, Yeh et al.^[4] reported that the incidence of COPD was more than three times higher in patients with NTM infection than in the non-NTM cohort. From China, Hu et al.^[5] also found that 64% of patients with NTM had PTB in the past and 39% had bronchiectasis. Abbew et al.^[6] from sub-Saharan Africa in a scoping review of 785 potential articles identified that previous history of TB, smoking, and mining were the strongest risk factors.

Even though Lim et al.^[7] from Singapore reported that the majority of their patients diagnosed with NTM-PD were non-smokers (61%), but they did not overemphasized this fact, considering that the denominator for non-smokers will always be higher than smokers in most of the populations across the globe and the disease prevalence will also follow the same pattern in that population. The scientific validation of risk association can only be made in disease outcomes and considering the diseased population as the denominator. Lee and Jhun^[8] from South Korea also found that current smoking, bronchiectasis, and acid–fast bacilli smear positivity demonstrated a statistically significant negative impact on microbiological cure in NTM-PD.

However, we were surprised to read a very recent article from South Korea by Chung et al. ^[9] in the Scientific Reports and their conclusions that current smokers and current heavy smokers were at lower risk for NTM-PD and these conclusions invited our attention and here we present our observations.

OUR OPINION

It is well established that NTM infections are not reportable in any country. This leads to a lack of awareness even among the microbiologists who are responsible for identifying the organism, as well as the treating physicians. Furthermore, in most of the countries where TB is endemic, the laboratory infrastructure is lacking for culture and identification of NTM up to the species level, because of overwhelmed human resources, suboptimal infrastructure, financial constraints, and low priority given to these infections over the TB and other prevalent infectious diseases. It is also well documented that there are no globally standardized or accepted criteria to define NTM isolation from respiratory samples vis-a-vis NTM respiratory disease.^[1]

Chung et al. ^[9] conclude that current smokers and current heavy smokers were at lower risk for NTM-PD development than never smokers. The title and conclusions of the article promote smoking, and that could be a disastrous consequence of this publication.

The data are skewed, and the conclusions are questionable. First, the authors used NTM-PD and pulmonary TB (PTB) as denominators and not the number of smokers. In their study, taken together, there were 60.35% non-smokers and 39.65% smokers (former or current). The total events of NTM-PD are 2288 and 25436 that of PTB, and 676 individuals had both NTM and TB. The authors do not give details of NTMPD, how the disease was diagnosed, and if the patients had multiple episodes, a common condition in NTM disease, how these were filtered or not segregated at all? It is very likely that the repeated episodes would have been counted as a new episode in the records of the insurance claim, which might have given inflated numbers.^[10]

The basis of their conclusion is derived from an earlier study from Korea that only estimated the prevalence of NTM episodes based on medical insurance claims. In that study, the authors claimed that in South Korea, the prevalence of NTM disease has increased from 11.4/100,000 in 2010 to 56.7/100,000 in 2021.^[11] Therefore, the basis of their premise itself is on wrong footings, as no laboratory confirmation of NTM was done. Making such far-reaching conclusions could have been considered very cautiously.

The authors have done only univariate analysis, instead of multivariate with other conditions, such as nutritional status, age, and co-morbidities such as chronic kidney disease (CKD), Bronchiectasis, and COPD. There were several confounding factors in the data, but the authors emphasized only associating smoking as a protective factor for NTM disease. Low body mass index (BMI) was significantly associated with a high prevalence of NTM diseases as compared to normal and high BMI, but was ignored by the authors [Table 1].

The data provided in the flow chart and Table 1 of the original article do not match. According to the flow chart, a total of 29074 individuals developed either NTM-PD or PTB, even if there is no mention of a combination. While in Table 1, the data show that 2288 had NTM-PD, 25436 had PTB, and 676 had a combination of PTB and NTM-PD, still making a total of 28400 patients.

Their supplementary Tables 2 and 3 given in the original article indicate that comorbidities such as CKD, bronchiectasis, and COPD had a strong association with NTM disease, which is a well-known secondary infection.^[9] In their Table 2 6.86% (259,031 of 3,774,308) of insured persons had CKD, and 272 (9.18%) had NTM irrespective of smoking behavior.^[9] This analysis is incorrect and intended to inflate the data. Taking denominator of total NTM events (2288) instead of the total number (259,031) of CKD cases is not right. Ideally, the number of NTM cases out of all CKD patients should have been taken, and if the number is calculated in that manner, the actual prevalence comes to only 0.105%. As expected, the incidence of TB was much higher than NTM disease in persons with bronchiectasis (1.17% vs. 2.48%) and COPD (0.42% vs. 2.26%). NTM and Aspergillus superinfections in treated patients of cavitary PTB lesions are also well-known sequels, but the authors have not mentioned anything about the NTM disease in these patients.

Since the study has involved data from an insurance database, the reliability of declaring one’s true disease status in a national insurance database is questionable. That too sensitive information, such as smoking status and years of smoking, cannot be precisely estimated from this database and the two consecutive studies from Korea without having laboratory confirmation of NTM disease is misleading.

The age cut-off taken is 20 years. Most of the pediatric/young adult TB cases, as well as NTM infections, would have been missed. Because the highest smoking age group is teenagers between 14 and 20 years, with greater addiction potential, these data are not captured. The article mentions about “age of initiation of smoking” but analyses its impact only in the age group of above 20-year.

There is a plethora of evidence that prior lung diseases, including cavitation, fibrosis, chronic obstructive disease, and use of steroids and broad-spectrum antibiotics, significantly increase the risk of NTM lung disease.^[12] Smoking is a well-established risk factor for lung infections, including mycobacterial diseases, because of compromised lung function^[13] and especially the pathogen-associated molecular patterns (PAMP).^[14] Several diseases associated with anatomical and physiological changes in the lung tissues, such as “Cystic Fibrosis (CF),” non-CF “bronchiectasis,” “primary ciliary dyskinesia,” “COPD,” TB in the past, pneumoconiosis, silicosis, and probably COVID-19 have been recognized as predisposing conditions to NTM lung disease. Conditions that make the person immunosuppressed, including AIDS, organ transplantation, or use of a tumor necrosis factor-alpha inhibitor, are also associated with NTM disease. On the genetic level, uncommon defects in the interleukin-12 and interferon (IFN-gamma) axis, and anti-IFN gamma autoantibodies may also lead to disseminated NTM infection.^[15] Even though there is a paucity of knowledge about differences in the cell wall and their biochemical structures between the M. tuberculosis and considerably pathogenic NTM species, it is well known that cell walls of both have chains of mycolic acid linked to middle layer of arabinogalactan covalently attached to the inner layer of peptidoglycan (PPG). Together, these form the mycolyl-arabinogalactanpeptidogycan (mAGP) complex, which provides structure and contributes to the impermeability of the mycobacterial cell wall. The major lipids, phosphatidylinositol (PIM), lipomannan (LM), and lipoarabinomannan (LAM) are also similar with minor changes in Mycobacterial LAM.^[16] However, the authors have provided nothing in support of their hypothesis, why smoking should behave differently for NTM and M. tuberculosis.

Besides several other pitfalls in the study, it was expected from the authors that the observations would have been interpreted in the right context. Bay et al.^[17] have concluded in a similar study that even if the smoking alone was not detrimental to the successful outcome of the TB treatment, several confounders, such as better socioeconomic conditions leading to better ventilation, good nutrition, and high BMI, could have contributed more significantly. Therefore, we conclude that this study not only lacks scientific evidence but also that the conclusions are misleading and should have not been published by the journal. The reviewers are mainly responsible for not raising such questions and for recommending such a misleading study for publication.

CONCLUSIONS

This opinion piece is written as a journal club to discuss the strength and weakness of a recently published paper in the reputed journal Scientific. It was unfortunate that when we contacted the journal and the authors, our concerns were neither addressed by the author nor the journal agreed to publish our comments. We believe that such studies with skewed data interpretation with larger public health related misgivings should have not been published in the first place, and even if these were published by oversight of the editor, the comments should have been allowed by the journal. In our conclusion, we strongly opine that smoking cannot be promoted as a beneficial habit to any population, that too with incorrect analysis of the data.