On 2017 Jan 11, Thomas E Sussan commented:

Extrapolation of the levels of exposure in this model to those of human users is difficult due to a number of factors, including differences in respiratory mechanics, altered metabolic rates of nicotine and other constituents, and the fact that mice are passively breathing rather than actively puffing. Development of an appropriate animal model that is both representative of a human condition and also amenable to investigation can be challenging. As a result, animal models often push the limits of a typical human exposure in order to obviate the need for a protracted disease course. Additionally, due to interspecies differences, toxicity studies generally employ a 10-fold uncertainty factor when extrapolating from laboratory animals to humans, and further uncertainty factors are also typically used to account for vulnerable/susceptible populations. Findings in animal toxicity studies are generally applicable to humans, and animal studies have been used for decades to provide the basis for human studies.

Mice are frequently more resistant than humans to toxicant exposures, as is the case for cigarette smoke. The most widely accepted mouse model of cigarette smoke-induced lung disease utilizes daily exposure to high levels of cigarette smoke for 4-7 hours per day for 6 mo, resulting in a phenotype that is morphologically similar to that of a chronic smoker with mild COPD. Additionally, unlike humans, mice rarely progress beyond mild stages of disease. Due to the relative short history of e-cigarettes, we don’t have the ability to compare the pulmonary responses in mice to those of chronic e-cig users. Thus, while animal models have been indispensible for providing the basis of human studies, they all have limitations, and our model is no different. Based on duration of exposure, we believe that our published model is reasonable, although we do not dispute that the level of exposure is elevated compared to that of a typical e-cig user.

In our exposure model, mice breathed freely in a chamber containing a constant mix of 20% e-cig vapor/80% air for 1.5h twice per day. In 3 hours, a human would take approximately 2160-3600 breaths (12-20 breaths/minute), and thus a 20% e-cig vapor would be equivalent to 432-720 breaths. However, it is notable that a typical human breath is substantially larger than a typical e-cig puff. Thus, breaths and puffs are not equivalent, and our exposure is higher than that of a typical human e-cig user.

Drs. Mukhin and Rose used a different approach to estimate exposure in our model based on our measured cotinine levels and previously published values of metabolic rates for nicotine and cotinine in mice and humans, a published equation to calculate steady state cotinine, published steady state cotinine levels in mice after nicotine infusion, and a published volume of e-cig liquid consumption per puff. Based on these values, they determined that our model was equivalent to 3600-4600 puffs by a human user (although our e-cigs contained 1.8% nicotine, not 1.6%, which reduces these estimates slightly to 3333-4111). This puff estimate is approximately 10x higher than the average number of daily puffs taken by a typical e-cig user. These published estimates for nicotine and cotinine metabolism can vary considerably due to genetic variation within and between species, age, sex, diet, etc, making it difficult to accurately compare exposures based on nicotine. Regardless, I am not disputing the validity of their estimate nor questioning their intentions. In fact, our exposure estimate (based on number of breaths) is consistent with their estimate (based on nicotine metabolism). The reported serum cotinine concentrations demonstrated that these levels were consistent with those of human e-cig users, but it is correct that this is actually indicative of an elevated exposure due to differences in nicotine metabolism between mice and humans.

Drs. Mukhin and Rose also stated that due to lack of clarity in the methods, it is possible that the puff estimate could be as high as 13000 puffs per day. To clarify this area of confusion that was not clear in our methods, blood was collected from 5-10 mice almost immediately after conclusion of exposure, but the process of collecting blood from each mouse required a certain amount of time. Thus, we wrote in the methods that blood collection was made within 60 min of exposure, but we strived to collect it as soon as possible.

Regarding the claim by PHE that we simply made the mice sick and thus reduced their recovery, the mice displayed no overt signs of sickness (ie vomiting, diarrhea, dizziness, lethargy, etc). They exhibited only mild effects in the absence of infection. However, in response to infection they demonstrated immunosuppression, which was evident even when the airway macrophages were infected ex vivo. This demonstrates a clear impairment at the cellular level.

The relevance of this study’s findings is unclear. The phenotypes demonstrated in this model certainly warrant further exploration of potential immunosuppressive effects in human e-cig users, especially considering the well-published impaired immune responses observed in both humans and mice exposed to cigarette smoke, which are similar to the findings in the current study. Our current study should not serve as the definitive body evidence, but should instead guide future studies. There is some more recent evidence to suggest that human e-cig users exhibit gene expression signatures indicative of immunosuppression (PMID: 27288488), although it’s noteworthy that this commenter has also criticized this human study and several other e-cig studies. Another group has shown in a retrospective study that switching from cigarettes to e-cigs may reduce the number of COPD exacerbations (PMID: 27986085), suggesting that the immune response is improved as a result of switching. Additional studies are still greatly needed to determine the relevance of our current animal study to e-cig users.

The point by Pruen is interesting. We previously did some preliminary testing to compare the disposable NJOY e-cigs versus the rechargeable NJOY e-cigs, and we noted considerable differences between the two with respect to the average life of a cartridge. We expected them to be similar, but were surprised to see that the rechargeable e-cigs lasted much longer than the disposables. Our animal exposure used the rechargeable e-cigs. Our exposure included a real-time light scattering monitor connected to the inlet of the exposure chamber to determine when the output of each e-cig began to wane. E-cig cartridges were replaced whenever the monitor detected a drop in vapor density. No one was puffing the e-cigs to ensure that the vapor remained pleasant to the taste, but we can definitively state that there was no dry puffing and thus overheating. We also puffed each e-cig for 2 seconds, once per minute to further prevent overheating of the coil and dry wicking.

Our study compared e-cig exposure to room air, but did not directly compare e-cig vapor to cigarette smoke. The questions of whether e-cigs are safe and whether they are safer than cigarettes are both valid questions, but our current study only addressed the first question. While our study demonstrated an effect on the immune response that was similar in nature to that previously seen with cigarette smoke, we did not directly compare the relative effects between cigarettes and e-cigs. No individual study can address all questions, and as stated above, further studies are needed to determine the effect of e-cigs on never-smokers, former-smokers, and dual users.

Dr Farsalino’s comment is simply a restatement of a paragraph in the Discussion of our publication, which stated the following: “Cigarette smoke contains 1014 free radicals per puff…We determined that E-cig vapor contains 7x1011 free radicals per puff…this concentration is several orders of magnitude lower than in cigarette smoke.”

On 2016 Dec 20, Zvi Herzig commented:

Important methodological issues have been raised in response to this study.

• Mukhin and Rose calculate, based on cotinine levels, that aerosol exposure levels in this study are the mouse equivalent of 11,000-13,000 puffs per day! link
• Public Health England's 2015 report on e-cigarettes notes that suddenly exposing nicotine-naive subjects to quantities of nicotine tolerated only by heavy smokers would be expected to result in stress, sickness and vomiting, which explains the reduced recovery from infection in exposed mice. link
• Mayer notes that it has been known for decades that nicotine has anti-inflammatory effects associated with immune suppression in mice and rats, but this is apparently not reflected in humans. Thus, this study's findings of reduced recovery after infection are not surprising, but lack relevance. link
• Pruen explains that this study's design protected poorly from inadvertent overheating, likely resulting in unrealistic toxicant exposures from pyrolysis. link
• Without cigarette smoke exposed controls it's quite difficult to estimate magnitude of effect in relation to real-world human consumption.

Additionally, Farsalinos notes:

"Concerning free radicals, the authors found 7x1011 spins/puff compared to 1014 spins/puff for smoking (other reports have measured up to 1017 spins/puff for smoking). That is about 150 times lower compared to tobacco cigarettes." link

On 2016 Dec 20, Zvi Herzig commented:

Important methodological issues have been raised in response to this study.

• Mukhin and Rose calculate, based on cotinine levels, that aerosol exposure levels in this study are the mouse equivalent of 11,000-13,000 puffs per day! link
• Public Health England's 2015 report on e-cigarettes notes that suddenly exposing nicotine-naive subjects to quantities of nicotine tolerated only by heavy smokers would be expected to result in stress, sickness and vomiting, which explains the reduced recovery from infection in exposed mice. link
• Mayer notes that it has been known for decades that nicotine has anti-inflammatory effects associated with immune suppression in mice and rats, but this is apparently not reflected in humans. Thus, this study's findings of reduced recovery after infection are not surprising, but lack relevance. link
• Pruen explains that this study's design protected poorly from inadvertent overheating, likely resulting in unrealistic toxicant exposures from pyrolysis. link
• Without cigarette smoke exposed controls it's quite difficult to estimate magnitude of effect in relation to real-world human consumption.

Additionally, Farsalinos notes:

"Concerning free radicals, the authors found 7x1011 spins/puff compared to 1014 spins/puff for smoking (other reports have measured up to 1017 spins/puff for smoking). That is about 150 times lower compared to tobacco cigarettes." link