South Africa’s ban on the sale of tobacco products as part of Covid-19 lockdown regulations is the subject of a court battle, with current medical research coming under the spotlight. The available data settles little, with several surprising twists and turns. What is clear though, is that more research is needed.
South Africa went into lockdown on 26 March 2020, a drastic measure intended to slow the spread of Covid-19 and give the country time to prepare itself to fight the disease. Among the more controversial regulations during lockdown has been a ban on the sale of tobacco products.
This was challenged in court in May by the Fair-trade Independent Tobacco Association, a grouping of cigarette makers. In defence of the government's ban, minister of cooperative governance and traditional affairs Nkosazana Dlamini-Zuma said there was "nothing sinister" behind it and that it was based on "relevant medical literature”.
Statistics cited by the minister can be traced to a small review of available evidence published early in the outbreak. This analysed five studies from China, with a total sample size of 1,549 Covid-19 positive patients. Fifty-two of them were current smokers. (Note: The more recent reviews cited below include these studies, and others.)
The conclusion that the review reaches, and which has been shared by the minister, is that the severity of Covid-19 is greater in smokers.
Are smokers at greater risk? And what does more recently published research reveal? We took a look.
Negative effects of smoking
From a medical perspective, smoking is a destructive habit. A review of 45 studies on smoking shows it has a negative effect on virtually every organ system. It increases the risk of certain cancers, decreases immune function, and impairs reproductive, heart, lung, blood vessel and bone health.
The same review shows that smoking increases the duration of symptoms of viral lung infections by up to a day on average. Depending on the illness studied, current smokers are between 34% and 200% more likely to develop influenza-like illnesses than non-smokers.
South Africa also has to contend with a heavy burden of tuberculosis (TB), an infectious bacterial disease that mainly targets the lungs. Studies have shown that smoking more than doubles the risk of TB infection and increases the risk of dying from it by a similar margin.
In light of this, governments could be tempted to crack down on smokers during a viral public health crisis. Covid-19 has, however, turned established convention on its head.
Data suggests smokers are underrepresented in Covid-19 studies
At present there are no randomised controlled studies – the type of studies which are among the most rigorous – which have directly investigated the effect of smoking on Covid-19. This means we have to gather our data via other methods.
One way is to compare the characteristics of people infected with Covid-19 with the larger population. By looking at what happens to Covid-19 patients who are smokers, we can get an idea of the nature of the relationship between the disease and smoking.
One of the largest studies on this comes from a collaboration between the International Severe Acute Respiratory and emerging Infections Consortium and the World Health Organization. It is an ongoing project involving 208 acute care hospitals in England, Scotland and Wales. A total of 20,133 Covid-19 positive patients are included in the study. Of these, just over 14,000 were asked about their smoking status. Of those asked, 6% are current smokers, compared to 15% of the British population.
Another frequently cited report published by the US Centers for Disease Control and Prevention looked at 7,162 Covid-19 patients for whom they had complete data for underlying conditions. Of these, 1.3% were current smokers, compared to 15.6% of the American population. Twenty-two of the smokers were hospitalised and five admitted to intensive care units.
‘Pre-print’ studies not yet reviewed
Due to the short timeline of current Covid-19 research, a number of “pre-print” studies have been released. These are studies which have been submitted to a journal and are currently undergoing peer review, the process where studies are evaluated and validated.
Pre-print studies are generally best avoided when trying to provide the highest quality commentary or evidence on a topic, but much of the available Covid-19 literature exists in this form.
A pre-print study was conducted on 3,789 US veterans, of which 42.3% were smokers. It found that smokers were underrepresented among the 585 patients who tested positive for Covid-19. Among them 159 (or 27%) of smokers tested positive.
Another meta-analysis, or “study of studies”, looked at 15 Chinese studies, two small studies from the US and one from Italy. The results showed that the number of smokers hospitalised for Covid-19 was lower than expected when compared to the prevalence of smoking in the countries.
|Smoking prevalence in Covid-19 patients compared to country populations
|Percentage of hospitalised Covid-19 patients that were current smokers
|Prevalence of smoking in the country
Source: What is Happening with Smokers and COVID-19? A Systematic Review and a Meta-Analysis
A pre-print study from a hospital in Paris, France, interviewed 340 Covid-19 positive in-patients and 139 out-patients on their smoking status. Given the level of smoking in the general population, they concluded the odds of being hospitalised as a smoker were about 20% of what would be expected.
This data yields a bizarre trend - smokers do seem to be less likely to end up hospitalised with Covid-19. It is contrary to everything we know about smoking and the expected disease outcomes for smokers.
What happens to smokers in hospitals?
What then of disease outcomes in smokers? It’s one thing to say that smokers may be less likely to be admitted to hospital with Covid-19 than non-smokers but what happens once they’re in hospital?
One of the largest studies looking at this was led by well-known evidence-based medicine advocate Ben Goldacre and a team from the University of Oxford and the London School of Hygiene and Tropical Medicine in the United Kingdom. They examined the records of 5,683 people who died from Covid-19 and the demographic and behavioural risk factors associated with death.
They found “weak evidence of a slightly lower risk” of dying from Covid-19 for current smokers. This means that when averaged out over a study population, there seems to be a tiny protective effect. In individual cases, however, this effect does not always appear.
The study noted that “even if smoking does have a small protective effect against Covid-19, this would still be massively outweighed by the well-established adverse health effects of smoking”.
A 2020 meta-analysis from the University of California in the US looked at 19 peer-reviewed papers with data from a total of 11,590 Covid-19 patients. In 29.8% of patients with a history of smoking, the disease became more severe or resulted in death. In comparison, this happened to 17.6% of non-smoking patients. The analysis noted that these findings were “not surprising” due to the negative effects of smoking on the overall function of the immune system within the lungs.
Taking a step back and looking at all of the data presented here we can tentatively agree with the hypothesis that smokers are less likely to be admitted to hospital with Covid-19, but once they are in hospital, their outcomes are similar at best and likely worse than nonsmokers.
Nicotine an avenue for research – but no proof yet
Exactly why this happens is unclear. The short answer is that we simply don’t know. But some research groups are starting to put together educated guesses.
Smoking tobacco produces a complex mixture of cancer promoting substances, toxins and other chemicals. None of these are known as yet to have any protective effects against viruses in general or Covid-19 specifically.
A team from the University of West Attica in Greece has proposed that nicotine may be responsible for the stark difference between the percentage of Covid-19 patients who are smokers and smoking prevalence in the general population.
The current theory is that nicotine might alter how our bodies respond to Covid-19 infection. A 2011 study from the University of Brighton in the UK exposed laboratory mice to fragments of bacteria to induce lung injury and inflammation. The mice showed lower levels of inflammation when treated with nicotine.
This nicotine connection is currently unsupported by real-world experimental data, but is a new avenue for research.
Findings should be interpreted with caution
As a team from the University of Sydney’s School of Public Health in Australia points out, we need to interpret all of these results with caution. Bias or a distortion due to some unaccounted-for error may play a role in why we’re seeing these unexpected results.
First among these is selection bias, meaning that the samples in these studies aren’t truly random. A sample may consist of people disproportionately less likely to be smokers. This includes health care workers, who are at higher risk of catching Covid-19 “but usually show lower prevalence of current smoking”. Many of these studies had incomplete datasets and sample sizes had to be drastically cut because smoking status was not recorded for many patients.
Social desirability bias may cause people to deny smoking. Patients may wish to be seen as making good decisions when it comes to their health. As a doctor in South Africa, I have experienced many patients reporting their smoking status incorrectly when being tested for Covid-19. In cases where someone has not smoked for a few days, they may report this as having “quit”. If not interrogated it may be recorded on their file.
Lastly, as mentioned above, many of these papers are still undergoing peer review, and await final approval before being published formally. While they provide new, interesting areas to explore they should be interpreted with caution – especially by those using them to justify health regulations.
Petrie Jansen van Vuuren is a medical doctor based in Gauteng, with a postgraduate background in human physiology research at the University of Pretoria.