Exhaled carbon monoxide has been linked to a lower risk of Parkinson’s disease

This is a very interesting observational study due to its scale – involving more than 500,000 adults aged between 30 and 79 who were followed for around 12 years – and, above all, because it offers a possible explanation for an association that we have been aware of for decades: smokers have a lower risk of developing Parkinson’s disease.

What is new is that the authors have directly measured exhaled carbon monoxide (CO) and found that, even amongst people who have never smoked, higher levels of CO are associated with a lower risk of Parkinson’s. This raises the possibility that the supposed protective effect of tobacco may be linked to CO and not necessarily to nicotine. It is an interesting hypothesis, moreover, because the potentially neuroprotective effects of CO had already been observed in experimental models.

The study is robust due to its large sample size, its prospective design and the objective measurement of CO, but caution is needed when interpreting the results. As this is an observational study, it demonstrates an association, not a causal relationship. Therefore, these results do not in any way imply that smoking is beneficial or that one should smoke to prevent Parkinson’s disease. Smoking remains a major cause of cancer, cardiovascular disease and mortality. Furthermore, the study participants are from the Chinese population, whose genetic and environmental characteristics, consumption habits, and patterns of exposure to smoke and pollution may differ from those of other populations. It is therefore prudent not to extrapolate these results directly to other populations until studies are available that replicate the association in different contexts.

From a clinical perspective, the most interesting aspect is that these results support research into the potential therapeutic use of very low, controlled doses of CO. This is a promising avenue, but it is still experimental and has no immediate clinical application.

The study by Bueno-López and colleagues provides a new insight into why smokers have a lower risk of developing Parkinson’s disease. After following more than half a million participants from the China Kadoorie Biobank for around 12 years, the authors confirm this association and present a very interesting finding: amongst people who had never smoked, those with higher levels of exhaled carbon monoxide (CO) also had a lower risk of Parkinson’s.

This result suggests that the reduced risk of Parkinson’s associated with smoking may not be due to tobacco itself, but that CO could contribute, at least in part, to explaining this relationship. This possibility is particularly interesting because for years it has been suggested that nicotine might be responsible for this effect, but clinical trials conducted to date have not demonstrated that it slows the progression of the disease.

The study stands out for its large number of participants, its long follow-up period and the control for numerous factors that could influence the results (second-hand smoke exposure, use of solid fuels in the home, physical activity and socio-economic status). Furthermore, the results are consistent with previous studies carried out on animal models, which show that, although CO is highly toxic at high concentrations, at low and carefully controlled doses it can activate antioxidant, anti-inflammatory and neuroprotective mechanisms.

In any case, and very importantly, this study does not prove that smoking protects against Parkinson’s disease and offers no reason to take up smoking or to continue smoking. In fact, the study itself confirms that smoking increases the risk of lung cancer, cardiovascular disease, stroke and premature death.

The research has some limitations: as it is observational, it shows an association but does not prove a cause-and-effect relationship; furthermore, exhaled CO was assessed only once, at the start of the study (using the average of two consecutive measurements), and its origin cannot be determined with precision. Although these results do not, for the time being, change clinical practice, they open up the possibility of exploring low-dose CO as a therapeutic strategy—an option that is already beginning to be evaluated in patients with Parkinson’s disease.

The study is epidemiologically sound: a prospective design, over half a million participants and 12 years of follow-up. But for me, its significance lies not where it might seem at first glance. Confirming that smoking is associated with a lower incidence of Parkinson’s does not constitute a decisive contribution, because that observation has been on the table for decades. What is striking is the subgroup of people who had never smoked: even amongst them, higher levels of exhaled carbon monoxide are associated with fewer subsequent cases of the disease. And here, the measurement is objective, with adjustments made for foreseeable external sources, ranging from passive smoking to the domestic use of solid fuels.

However, an association does not imply a protective effect. Carbon monoxide could be involved in some protective mechanism or simply indicate other biological processes or exposures that do influence the risk. The study does not distinguish between the two.

The mechanism is plausible. We produce carbon monoxide endogenously via the heme oxygenase pathway, and experimental findings link it to the response to oxidative stress and to neuroprotective effects, reducing the aggregation of α-synuclein in animal models.

The message to the public must be unequivocal. These findings do not justify taking up smoking or delaying giving up smoking; the study itself finds, as might be expected, higher rates of lung cancer, cardiovascular disease and mortality amongst smokers. Nor do they justify deliberately exposing oneself to a gas that causes severe poisoning.

Furthermore, there remains a limitation that is easily overlooked. The fact that something is associated with a lower incidence of the disease says nothing about whether administering it would slow its progression in someone who has already been diagnosed. These are two distinct questions, and this research addresses only the first. Before discussing clinical applications, the finding would need to be replicated in other populations and its safety and efficacy demonstrated in controlled trials. This is a hypothesis worth exploring, but so far there is no reason to change clinical practice or tobacco prevention policies.

Is the research of good quality?

“It is a well-designed and well-executed study by a group of experts in epidemiology. The analysis is meticulous. The only point to bear in mind is that the follow-up period is relatively short for diseases such as Parkinson’s disease (PD), which have a late age of onset; as a result, they may be detecting fewer cases than actually occur in that sample.”

How does it fit in with the existing evidence?

“It is consistent with epidemiological studies showing a reduced risk of developing PD among smokers, even to a greater extent than in this study.”

Could findings of this kind encourage people to take up smoking or prevent them from giving up?

“No, the health risks associated with smoking are well known and severe. In fact, the study suggests that the protective effect can be dissociated between the effect of carbon monoxide (CO) and that caused by other compounds present in tobacco, such as nicotine, which was thought to be causally linked, albeit without clear evidence. It is also true that this study cannot prove that this is not the case (regarding nicotine or other tobacco compounds), but it does provide evidence of an inverse correlation between CO levels and the risk of developing PE in non-smokers, which is the most significant finding from my point of view.”

What are the implications of these findings for clinical practice?

“At present, they have no direct clinical implications. Strategies are being explored to stimulate the production of endogenous CO as an indirect antioxidant, due to its ability to activate heme oxygenase-1 and other antioxidant mechanisms.”

Does the study have any significant limitations that need to be taken into account?

“I believe the main limitation is the relatively short follow-up period for diseases such as PD, which have a late age of onset, and perhaps the fact that only a single measurement of exhaled CO was taken, which may not be representative of actual exposure over time.”

Is the research of good quality?

"This is a very high-quality piece of research, based on an epidemiological study with a coherent and robust methodological design, a very large sample size and long-term follow-up.

One of the inherent limitations of epidemiological studies is the difficulty in controlling for all the factors that may come into play throughout a person’s life and that could potentially influence the observed associations. Although a large sample size does not completely eliminate this limitation, it does help to minimise it and lends greater robustness to the results.

Overall, I consider this to be a very well-conducted study that provides relevant information for a better understanding of the relationship between smoking and Parkinson’s disease.”

How does it fit in with the existing evidence?

“The study fits in well with previous evidence, as it reproduces an epidemiological association that has been known for decades: smokers have, statistically speaking, a lower risk of developing Parkinson’s disease.

However, for me, one of the main contributions of this work is that it allows us to pinpoint this association to a specific factor: exhaled carbon monoxide (CO). It is particularly interesting that this association is observed even in people who have never smoked. This raises the possibility that the historically described association between smoking and a lower risk of Parkinson’s disease does not necessarily depend on tobacco itself or all its components, but may be mediated, at least partially, by specific factors related to CO.

Historically, another explanation has been put forward for this inverse association. Dopamine plays a key role in reward and addiction mechanisms and, given that Parkinson’s disease is characterised by a progressive loss of dopamine, it has been suggested that people in the very early or prodromal stages of the disease might have a lower predisposition to developing addictive behaviours, including smoking. According to this hypothesis, it is not smoking that reduces the risk of Parkinson’s, but rather that undiagnosed Parkinson’s disease might reduce the likelihood of becoming a smoker. This study provides new insights into how to interpret this relationship.”

Could findings of this kind encourage people to take up smoking or prevent them from giving up?

“They shouldn’t. In fact, I believe the study’s message points precisely in the opposite direction.

The study does not identify smoking as a protective factor, but rather points to a specific element that could help explain the observed epidemiological association. This is important because it would eventually allow us to study that mechanism without having to expose ourselves to the harmful effects of tobacco.

Furthermore, the study itself shows that smoking is associated with a higher incidence of heart attack, stroke and all-cause mortality. Therefore, these results in no way justify taking up smoking or continuing to smoke.

Even in the hypothetical scenario that smoking had some protective effect against Parkinson’s disease, the risk-benefit balance would still be clearly unfavourable. Parkinson’s disease is a treatable condition and, in many patients, has a relatively benign course over many years. It would make no sense to try to reduce the risk of Parkinson’s whilst simultaneously increasing the risk of cardiovascular and cerebrovascular diseases, cancer and mortality. Conclusion: smoking is harmful.”

What are the implications of these findings for clinical practice?

“At present, they are limited, and recommendations should remain unchanged: smoking remains harmful to health and must be discouraged.

The significance of the study is primarily scientific, as it helps to better understand a well-known epidemiological association. In science, it is essential to distinguish between association and causation. The fact that two phenomena are associated does not necessarily mean that one is the cause of the other. We have known for some time that smoking is associated with a lower incidence of Parkinson’s disease, but that does not prove that smoking is the cause of this reduced risk. Identifying the specific biological factors that explain this association and determining whether any of them actually have a causal relationship is the next step.

If future experimental studies were to provide a solid biological basis for the possible role of carbon monoxide (which this study does not demonstrate), a clinical trial might eventually be considered to investigate its effects via a controlled and safe method of administration, without any need to smoke. Only then could we speak of a possible therapeutic application. At present, we are a long way from being able to make such a recommendation.”

Does the study have any significant limitations that need to be taken into account?

“The main limitation stems from its very nature as an epidemiological study. The study allows us to identify associations, but not to establish a causal relationship or, on its own, to demonstrate the biological mechanism that explains them. Epidemiological studies are, moreover, subject to multiple potential confounding factors related to lifestyle, the environment, individual characteristics and cumulative exposures over the years. In this study, the large sample size and the robustness of the design allow these limitations to be considerably minimised, although not completely eliminated.

Therefore, the results should be interpreted as strong evidence of an association and as an interesting hypothesis for further investigation, but not as proof of a cause-and-effect relationship.

Overall, I consider this to be a high-quality study that helps to put the well-known link between smoking and Parkinson’s disease into context and, above all, paves the way for understanding the mechanisms that may actually underlie this association.”

"This is a vast study involving a very large population, carried out in China, with excellent clinical and laboratory methodology. The study examines a specific mechanism – one which has not, until now, been assessed in such depth and with such conclusiveness – relating to the role of carbon monoxide (CO) as a factor causally linked to the lower incidence of Parkinson’s disease (PD) amongst regular smokers.”

Is the research of high quality?

“The research is of excellent quality due to the number of participants, the rigorous inclusion criteria and definition of the different groups (regular smokers, former or occasional smokers, and never-smokers), the measurement of CO, and the detailed analysis of many potential associated factors that could influence the findings.”

How does it fit in with existing evidence?

“It unequivocally confirms the association between smoking and a lower risk of developing PD. It also confirms other known associations in the Caucasian population, such as the slightly higher incidence of PD in men than in women, the direct relationship between age and the risk of developing PD, etc. However, the most original and relevant finding is that they have identified, with statistical certainty, an inverse relationship between CO levels and the risk of developing PD. Thus, in the population that has never smoked, CO levels are the lowest in the entire study and the incidence of PE is highest.”

Could findings of this kind encourage people to take up smoking or prevent them from giving up?

“They certainly should not. As we often reiterate, this is scientifically valuable data that can help us understand the relationship with tobacco and a possible neuroprotective effect of CO, which should be explored in other ways.”

What are the implications of these findings for clinical practice?

“Essentially none. A study published in 2023 revealed that daily nicotine use did not lead to any improvement in Parkinson’s-related symptoms and, on the contrary, there was some deterioration.

Important: for those already suffering from PD, it makes absolutely no sense to start smoking, as this would almost certainly result in a deterioration in general health and neurological status. This must be understood without any doubt.

As the saying goes, ‘smoking kills’, and the adverse effects on general health are far more widespread and severe than any potential benefit in reducing the incidence of PD.”

Does the study have any significant limitations that need to be taken into account?

“Certainly. Although the sample is huge, there is a bias towards a mean age at which the incidence of PD is not yet very high. Regular, long-term smokers in China, in large numbers, are not yet that common, which may have an effect on the data. In any case, the inverse association described between CO and PE is very robust. A more significant limitation is that CO is a gas obviously associated with changes in the air caused by smoking, but chemically speaking, it is easy to understand that many other direct changes—involving other molecules—as well as indirect changes occur. Therefore, at this stage, it is not possible to confirm or establish a direct causal relationship between CO and a lower incidence of PE.”

The harmful effects of tobacco on health are well established. Tobacco use and exposure cause cancer and cardiovascular and respiratory diseases, which are responsible for the majority of the seven million deaths attributable to tobacco worldwide each year and the nearly 60,000 deaths that occur each year in Spain. Paradoxically, various observational studies have shown that, compared with non-smokers, smokers appear to have a lower incidence of Parkinson’s disease, although the cause of this association has not yet been clearly established.

Although some experimental studies suggest that nicotine may have protective effects on dopaminergic neurons, which might explain a certain degree of protection against Parkinson’s disease, clinical trials have not demonstrated nicotine’s efficacy in preventing this neurodegenerative disorder.

In the recently published study, the authors raised the possibility that it might be another component of tobacco smoke other than nicotine – namely carbon monoxide – that exerts a protective effect against Parkinson’s disease. To test this hypothesis, they measured carbon monoxide levels in over half a million people. After a 12-year follow-up, it was observed that people with higher-than-normal carbon monoxide levels had lower rates of Parkinson’s disease, whether they were smokers exposed to the substance via tobacco smoke or non-smokers who had been exposed via smoke emitted from traditional domestic cookstoves and fireplaces.

It is important to bear in mind that these findings do not imply any health benefit from tobacco use – given that half of regular smokers are known to die from diseases caused by such consumption – nor do they refute the serious health consequences of acute or chronic exposure to a gas as toxic as carbon monoxide. However, and despite the methodological limitations of observational studies, the results of this research provide data for further investigation into the aetiology of neurodegenerative diseases such as Parkinson’s, which may help to improve their prevention and treatment.

Strange as it may seem, we have known for years that tobacco use is associated with a lower risk of developing Parkinson’s disease. And, although we still do not know the reasons why, various studies point to nicotine as a possible cause. However, no improvements were observed in patients treated with nicotine patches, and some experienced a slight deterioration. This suggests that other factors may be responsible for the ‘beneficial effect’ of smoking on this condition.

This study analyses, for the first time, the possible link between tobacco use, exhaled carbon monoxide levels and the risk of developing neurodegenerative diseases, including Parkinson’s disease. To this end, a cohort of over 500,000 people, comprising both smokers and non-smokers from different regions of China, was followed up for more than 10 years. The results show that, among smokers with higher levels of exhaled carbon monoxide, the risk of developing Parkinson’s disease is reduced by up to 30 per cent, but not for other neurodegenerative diseases. However, the most interesting finding is that the risk of developing Parkinson’s disease is reduced to a similar extent in individuals—particularly women—who have never smoked but who had higher levels of exhaled carbon monoxide. This suggests that carbon monoxide may be responsible for the beneficial effects observed in the smoking population, acting as a defence mechanism. However, further research is needed to understand the mechanisms involved and to determine whether the results could be extrapolated to other populations, as the study focused on a population sample in China with specific socio-economic, environmental, cultural and genetic characteristics.

The results point to carbon monoxide produced by our own bodies as the factor responsible for the positive effects of tobacco in patients with Parkinson’s disease. Most importantly, this would diminish the role of tobacco components as protective factors against the disease. It is therefore important to emphasise that the evidence suggests the lower incidence of Parkinson’s among smokers is not directly related to the components of tobacco. Identifying factors of this kind would open the door to potential new, non-toxic and targeted treatments that could delay the progression of the disease, without the undeniable negative consequences that tobacco use has on human health.