You are here

  • Home
  • Archive
  • Volume 11, Issue 1
  • Association between COVID-19 severity and tobacco smoking status: a retrospective cohort study using propensity score matching weights analysis

Article Text

Article menu
Smoking
Association between COVID-19 severity and tobacco smoking status: a retrospective cohort study using propensity score matching weights analysis
  1. Musaad A Alshammari1,
  2. Ahmad Alamer2,
  3. Lina Al Lehaibi3,
  4. Mashael Alghamdi4,
  5. Haneen Alotaibi5,
  6. Mukhtar Alomar3,
  7. Fawaz Alasmari1,
  8. Faleh Alqahtani1,
  9. Abdualziz Alhossan6 and
  10. Tahani K Alshammari1
  1. 1Department of Pharmacology and Toxicology, King Saud University, Riyadh, Saudi Arabia
  2. 2Department of Clinical Pharmacy, Prince Sattam bin Abdulaziz University, Al Kharj, Saudi Arabia
  3. 3Department of Pharmaceutical Affairs, Dammam Medical Complex, Eastern Health Cluster, Dammam, Saudi Arabia
  4. 4Research Office, Saudi National Institute of Health, Riyadh, Saudi Arabia
  5. 5College of Pharmacy, King Saud University, Riyadh, Saudi Arabia
  6. 6Department of Clinical Pharmacy, King Saud University, Riyadh, Saudi Arabia
  1. Correspondence to Dr Musaad A Alshammari; malshammari{at}ksu.edu.sa

Abstract

Introduction The COVID-19 pandemic continues to be a global threat to public health, with over 766 million confirmed cases and more than 6 million reported deaths. Patients with a smoking history are at a greater risk of severe respiratory complications and death due to COVID-19. This study investigated the association between smoking history and adverse clinical outcomes among COVID-19 patients admitted to a designated medical centre in Saudi Arabia.

Methods A retrospective observational cohort study was conducted using patient chart review data from a large tertiary medical centre in the eastern region of the country. Patients admitted between January and December 2020 were screened. The inclusion criteria were ≥18 years of age and confirmed COVID-19 infection via reverse-transcription-PCR. The exclusion criteria were unconfirmed COVID-19 infection, non-COVID-19 admissions, unconfirmed smoking status, vaccinated individuals, essential chart information missing or refusal to consent. Statistical analyses comprised crude estimates, matching weights (as the main analysis) and directed acyclic graphs (DAGs) causal pathway analysis using an ordinal regression model.

Results The sample comprised 447 patients (never-smoker=321; ever-smoker=126). The median age (IQR) was 50 years (39–58), and 73.4% of the sample were males. A matching weights procedure was employed to ensure covariate balance. The analysis revealed that the odds of developing severe COVID-19 were higher in the ever-smoker group with an OR of 1.44 (95% CI 0.90 to 2.32, p=0.130). This was primarily due to an increase in non-invasive oxygen therapy with an OR of 1.05 (95% CI 0.99 to 1.10, p=0.101). The findings were consistent across the different analytical methods employed, including crude estimates and DAGs causal pathway analysis.

Conclusion Our findings suggest that smoking may increase the risk of adverse COVID-19 outcomes. However, the study was limited by its retrospective design and small sample size. Further research is therefore needed to confirm the findings.

  • COVID-19
  • Tobacco and the lung
  • Respiratory Infection
  • Non invasive ventilation
  • Clinical Epidemiology

Data availability statement

Data available upon request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjresp-2023-001976

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What is already known on this topic

    • Findings from the early waves of the pandemic indicated a paradoxical association between smoking and a decreased risk of adverse COVID-19 outcomes. This was attributable to a number of methodological gaps, such as a reliance on record linkage.

    What this study adds

    • To address the methodological limitations in the literature surrounding the topic of smoking status and COVID-19 outcomes, we used various methodological approaches to analyse our results and compare the outcomes.

    How this study might affect research, practice or policy

    • This study will inform researchers investigating the association between smoking and the risk of adverse COVID-19 outcomes about existing methodological gaps in the literature. It also offers insights into different methodological techniques that can reduce biased or paradoxical results.

    Introduction

    As of July 2023, the COVID-19 pandemic has continued to be a global public health threat, with over 766 million confirmed cases worldwide and more than 6 million deaths reported to the WHO.1 Since its emergence and rapid global spread in early 2020, COVID-19 has exhibited an extensive range of clinical manifestations, from asymptomatic infection to severe and even fatal pneumonia.2 The predominant symptoms observed in patients with the COVID-19 infection are fever, cough, chest tightness and dyspnoea.3 The severity of symptoms appears to be dependent on age and health status. Data from early studies on COVID-19 patients reveal that advanced age and the presence of pre-existing conditions or comorbidities such as chronic obstructive pulmonary disease, hypertension and diabetes are associated with increased severity of symptoms and an elevated need for intensive medical care.4

    Given the robust scientific evidence demonstrating the harmful effect of smoking on the immune and respiratory systems, it has been hypothesised that smoking could be a risk factor for more adverse COVID-19 clinical outcomes.5 6 However, several early studies assessing the association between cigarette smoking and increased adverse clinical outcomes of COVID-19 among hospitalised patients have yielded contradictory findings.4 In a review of the literature on tobacco product use and the association between COVID-19 infection and unfavourable outcomes, Benowitz et al identified that out of 51 studies, 20 suggested a positive effect of cigarette smoking on COVID-19 infection, 26 suggested a negative effect and 13 suggested no significant association.4 Notably, more recent studies indicate an association between smoking status and unfavourable outcomes of COVID-19 infection. Indeed, the results of multiple systematic reviews and meta-analyses conducted since the outbreak of the pandemic support the hypothesis that smoking is a risk factor for more adverse clinical outcomes.7 For instance, a meta-analysis by Patanavanich et al of 34 articles covering a total of 35 193 COVID-19 patients identified a significant association between smoking and an increased risk of COVID-19 mortality.7

    In Saudi Arabia, smoking poses a substantial threat to the health of the population. According to data drawn from two nationwide cross-sectional surveys conducted with 7317 Arabic-speaking adults across all 13 regions of the country in 2018, the prevalence of cigarette smoking was 21.4%.8 Given that a sizeable segment of the population are cigarette smokers, there is an urgent need for studies that identify the impact of smoking on COVID-19 clinical outcomes. To address this gap, this study investigated the association between smoking history and adverse clinical outcomes among COVID-19 patients admitted to a tertiary hospital in Saudi Arabia.

    Methods

    Study design and setting

    This retrospective observational cohort study was conducted using patient chart review data obtained from one of the largest settings in the eastern region, the Dammam Medical Complex (DMC). As a tertiary care hospital, DMC holds accreditation from the Saudi Central Board of Accreditation of Healthcare Institutions and has a capacity of 423 beds. It was designated specifically to manage COVID-19 patients during the pandemic. It is also important to emphasise that healthcare in Saudi Arabia is primarily funded by the public sector through the Saudi Ministry of Health (MOH). The study was approved by the DMC Institutional Review Board (IRB Protocol No: PH-20). In this paper, we adhere to Strengthening the Reporting of Observational Studies in Epidemiology reporting criteria.9 Due to the retrospective nature of the study, the IRB granted a waiver for consent to access patients’ chart data. Verbal consent was obtained from individual patients when contacting them to confirm their exposure.

    Selection of participants

    We screened cases of COVID-19 hospitalised between January and December 2020. Our inclusion criteria were patients ≥18 years old with cases confirmed via reverse-transcription-PCR positive results obtained for SARS-CoV-2 tests conducted in accordance with the WHO and local interim guidance of the Saudi MOH. The exclusion criteria were non-confirmed COVID-19, non-COVID-19 admissions, admission outside the study recruitment period, unconfirmed exposure status (smoking), vaccination recipient, essential information missing in charts or refusing to consent when contacted. To minimise selection bias, we randomly selected patients’ charts to be screened and coded for analysis.10 All eligible patients were contacted to confirm their smoking status prior to COVID-19 admission. In the case of death, extended efforts were made to contact the patient’s relatives to confirm their smoking status.

    Data collection/data source

    Three researchers independently obtained and reviewed demographic information, clinical data, laboratory findings and clinical severity based on the WHO ordinal scale and other outcomes and coded them using the Research Electronic Data Capture cloud server system in a non-identifiable form.11 Patient data were acquired from DMC’s electronic health record (EHR) system. All patients were contacted over the phone to collect information missing in the records to confirm their exposure status (ie, smoking status) prior to COVID-19 admission. No data were missing regarding exposure, the covariates of interest or outcome variables.

    Definitions

    Patients were evaluated using the diagnostic and statistical manual of the American Psychiatric Association criteria for diagnosing tobacco use disorder.12 Smoking status was defined according to the National Center for Health Statistics. Patients defined as current smokers were adults who have smoked 100 cigarettes in their lifetime and who currently smoke cigarettes. Former smokers were defined as those patients who smoked at least 100 cigarettes in their lifetime but who had quit smoking at the time of the interview. Patients in the non-exposed (never-smoker) group were defined as patients who were not smokers or smoked fewer than 100 cigarettes in their lifetime prior to COVID-19 admission.13 Electronic cigarette (e-cigarette) smoking status was determined according to the US National Health Interview Survey.14 Current e-cigarette use was defined as use every day or some days. Former users were those who reported the use of an e-cigarette or other electronic vaping product (even if just once in a lifetime).

    The eight-point WHO ordinal scale was used to assess the severity of COVID-19 cases during hospital admission. Because the study was limited to the inpatient setting, three points or higher were used: 3=hospitalised with no oxygen use, 4=hospitalised with the use of oxygen therapy by mask or nasal prongs, 5=hospitalised with the use of non-invasive oxygenation or high-flow oxygen, 6=hospitalised with intubation or mechanical ventilation use, 7=hospitalised with ventilation and the provision of additional organ support (pressors) or extracorporeal membrane oxygenation and 8=death. The highest level of COVID-19 severity during hospital admission was coded.

    Outcomes

    The primary outcome was the association between smoking status (never-smoker vs ever-smoker) and the WHO ordinal COVID-19 severity scale. Secondary outcomes were the association of smoking status (non-smoker vs ever-smoker) with intensive care unit (ICU) admission, ICU length of stay (ICU-LOS) and hospital LOS. In addition, we evaluated the same outcomes for smoking status (never vs current or former smoker).

    Sample size calculation

    Assuming the proportion of COVID-19 severe cases in the non-exposed group to be 20%—an increase in the risk of severe COVID-19 disease with an OR of 2—and after applying continuity correction, a sample size of 137 patients in the exposed group and 274 patients in the non-exposed group was required for a 1:2 matching procedure. The sample calculation was performed to achieve a power of 80% and a two-tailed significance level (alpha) of 0.05.

    Statistical analysis

    Descriptive statistics were generated using appropriate tests to delineate patients’ characteristics and outcomes. Categorical variables were reported as frequency rates and percentages. A χ2 square test was performed to test for differences between the two groups. Continuous variables were presented as medians with their 25th and 75th percentiles. A Mann-Whitney U test was conducted to test for differences between the two groups.

    Matching weights procedure

    Owing to the observational nature of the study and the large number of covariates, we applied matching weights procedures as our primary method of analysis.15 Matching weights is an extension of the inverse propensity score weighting methods. Compared with propensity score matching and inverse propensity treatment weighting (IPTW) methods, matching weights is more efficient as it uses all the data available. It confers numerical stability compared with IPTW as weights are increased to balance the distribution of the propensity score. For further details, we refer readers to the relevant papers by Li and Greene and Yoshida et al.15 16

    Using this method, covariates of interest associated with exposure status (never-smoker vs ever-smoker) or outcome were selected to create patient weights using a logistic regression model. Because exposure status could not be assessed in all the eligible patients, the estimate of interest was the average treatment effect among the treated (ATT). The covariates of interest were age, gender, weight, comorbidities and in-patient COVID-19 regimens. To assess the balance of covariates after matching, we calculated the absolute standardised mean difference (SMD) between the two groups for the precovariate and postcovariate adjustments.17 Conventionally, SMD values of <0.2 are desirable for an appropriate balance.18 We illustrated the balance of the absolute SMD in a love’s plot using the ggplot2 package in R.19 Next, we fitted an ordinal regression model (proportional odds model) for the WHO ordinal scale and exposure status using the matching weights. We then employed a sandwich-type robust variance estimator to handle the disparity arising from the pseudo-population size—which occurred due to the utilisation of matching weights—to ensure an accurate estimation of treatment effects. This was implemented using the svyVGAM package in R.20 We conducted logistic regression models for the individual components of the WHO ordinal scale and ICU admission. A proportional odds model was then used to analyse ICU-LOS and hospital LOS outcomes, which involved discrete, non-normally distributed data. Smoking status was included as a covariate in this model. The medians and 95% CIs for these variables were estimated using a bootstrap method in the rcompanion package.21 In the matching weight case, the svyby function from the survey package was used to obtain medians with a 95% CI.

    All analyses were conducted twice: once with the crude data and then with the matched weight data for our main analysis. For the proportional odds and logistic regression models, the results are presented with an OR alongside its 95% CI.

    Directed acyclic graphs (DAGs)

    We complemented our analyses with DAGs in which we identified the potential confounders for statistical adjustments. DAGs are non-parametric diagrammatic representations of assumptions and hypotheses about the data-generating process and how variables influence each other.22 To investigate the total causal effect of smoking exposure (non-smoker and ever-smoker) on COVID-19 severity outcomes, we carefully selected and conditioned on a sufficient adjustment set that included age, body mass index and gender. The DAG causal pathway was then implemented once again to test the association between exposure (never, current or former) and COVID-19 severity. Like the matching weights procedure, a similar choice of models was selected to perform the analyses.

    All analyses were conducted using R Core Team (2022) software (R Foundation for Statistical Computing, V.4.2.2, Vienna, Austria) using the following packages: ggplot2,19 svyVGAM20 and rcompanion.21

    Patient and public involvement

    None.

    Results

    Overall, we screened the records of 1190 patients and included 447 in the study (ever-smoker=126; never-smoker=321). Reasons for exclusions were refused consent (n=3), non-confirmed COVID-19 (n=13), unable to confirm smoking status (n=720), received vaccine (n=1), non-COVID-19 admission (n=5) and missing information (n=1) (see patient recruitment flow chart in figure 1). Within the crude data, the prevalence of different smoking statuses was as follows: conventional current smokers (n=33, 7.4%), conventional former smokers (n=92, 20.6%) and current electronic vape smokers (n=1, 0.2%). No individuals were identified as former electronic vape smokers.

    Figure 1
    Figure 1

    Patient recruitment flow chart.

    The median with IQR for age for the never-smoker group was 50 (IQR 40–58.0) versus 48 (IQR 39–57.8) in the ever-smoker group (p=0.380). There were significantly more male patients in the ever-smoker group (96.7%) than in the never-smoker group (63.9%) (p<0.001). Comorbidities were similar between the two groups except for diabetes, which affected 40.2% of patients in the non-smoker group versus 29.4% in the ever-smoker group (p=0.043). Home medications were well-balanced between the two groups except for non-aspirin antiplatelet therapy (2.4% in the ever-smoker group vs 1.9% in the never-smoker group, p=0.035). Inpatient treatment regimens were similar between the two groups. The full results are presented in table 1. After matching weights, an evident balance was observed in all baseline characteristics, as indicated by the absolute SMD <0.2. The love’s plot (figure 2) illustrates the prematching and postmatching weights covariate balance.

    Figure 2
    Figure 2

    Love’s plot for the covariates. ACEI, ACE inhibitor; ARB, angiotensin receptor blocker; BB, beta blocker; CCB, calcium channel blockers; CVD, cardiovascular disease; CKD, chronic kidney disease; DOAC, direct oral anticoagulants.

    View this table:
    Table 1

    Baseline characteristics

    In the analysis of crude estimates, an increase was observed in the odds of severe COVID-19 disease among the ever-smoker group, as measured by the WHO ordinal scale (OR=1.28, 95% CI 0.85 to 1.94, p=0.226), but this did not reach statistical significance. Individual components of the analysis revealed a non-statistically significant increase in the odds of using an oxygen mask or nasal cannula in the ever-smoker group (68.3% vs 59.5%) with an OR of 1.09 (95% CI 0.99 to 1.21, p=0.087). Additionally, there were higher odds of non-invasive mechanical ventilation or high-flow nasal cannula usage in the ever-smoker group (7.1% vs 3.1%) with an OR of 1.04 (95% CI 1.00 to 1.08, p=0.058), although this just failed to reach statistical significance. Conversely, no effects were observed in the use of invasive mechanical ventilation (1.6% vs 1.9%) with an OR of 0.99 (95% CI 0.97 to 1.03, p=0.840) or the odds of death (OR=0.96, 95% CI 0.90 to 1.02, p=0.196).

    Rates of ICU admissions in both groups were found to be similar (17.5% vs 16.5%). However, a non-statistically significant increase was observed in the median ICU-LOS in the ever-smoker group (10 days, 95% CI 6 to 21) compared with the never-smoker group (8.8 days, 95% CI 5 to 10) with an OR of 2.06 with CI (95% CI 0.76 to 5.69, p=0.142). Notably, the ever-smoker group had a statistically significant longer median duration of hospital stay (8.6 days, 95% CI: 7 to 10) than the never-smoker group (6.9 days, 95% CI 6 to 7). The full results are presented in table 2.

    View this table:
    Table 2

    Clinical outcomes

    The results of the matching weights analysis were largely in agreement with the crude analysis. The ever-smoker group exhibited a non-statistically significant increase in the odds of severe COVID-19 with an OR of 1.44 (95% CI 0.90 to 2.32, p=0.130), use of oxygen mask or nasal cannula with an OR of 1.04 (95% CI 0.93 to 1.15, p=0.532) and non-invasive mechanical ventilation or high-flow nasal cannula with an OR of 1.05 (95% CI 0.99 to 1.10, p=0.101). Admission to the ICU was higher in the ever-smoker group (16.9%), although not statistically significant, compared with the never-smoker group (13%) with an OR of 1.05 (95% CI 0.97 to 1.14, p=0.254). The median ICU-LOS increased in the ever-smoker group with an OR of 1.87 with CI (95% CI 0.50 to 7.37, p=0.358). Finally, the median hospital LOS in the ever-smoker group exhibited a non-statistically significant increase with an OR of 1.49 (95% CI 0.95 to 2.35, p=0.085). The full results are presented in table 2 and figure 3.

    Figure 3
    Figure 3

    WHO ordinal scale outcomes. In the WHO ordinal scale, 3=no oxygen, 4=oxygen mask or nasal canula, 5=non-invasive mechanical ventilation or high-flow nasal canula and 6=invasive mechanical ventilation. Death is an overriding incident.

    The DAG causal pathways analysis revealed largely similar associations, except for the odds of increased use of invasive mechanical ventilation among the ever-smoker group with a value of 1.04 (95% CI 1.00 to 1.09, p=0.080) (see table 3 and figure 4). In the analysis of current, former or never-smokers, the results revealed an increase in COVID-19 severity as measured by the WHO ordinal scale in the current smoker group with an OR of 1.28 (95% CI 0.60 to 2.74, p=0.523) and in the former smoker group with an OR of 1.36 (95% CI 0.82 to 2.27, p=0.234) compared with the never-smoker group. In addition, a non-statistically significant increase was observed in the odds of non-invasive mechanical ventilation or high-flow nasal cannula in the current-smoker group compared with the never-smoker group, with an OR of 1.07 (95% CI 0.99 to 1.15, p=0.062). Finally, a significant increase was observed in the odds of hospital LOS in the former smoker group with an OR of 1.84 (95% CI 1.19 to 2.85, p=0.006). Full details are presented in online supplemental table S1) in the supplement.

    Figure 4
    Figure 4

    Directed acyclic graph (‘causal diagram’). The diagram illustrates assumptions and hypotheses about the data-generating process. Circles in red represent the minimal adjustments set (conditioning on the age, BMI and gender variables provided sufficient adjustment to isolate the total direct effect of smoking on COVID-19 severity). The same directed acyclic graph was used to estimate the total effect of smoking status as a never-smoker, current smoker or former smoker. The outcome was ‘severity of COVID-19 based on the WHO ordinal scale’. A similar directed acyclic graph model for a second outcome, ‘ICU admission’ was envisioned. BMI, body mass index; CKD, chronic kidney disease; DM, diabetes mellitus; DVT, deep-vein thrombosis; HLD, hyperlipidaemia; HTN, hypertension.

    View this table:
    Table 3

    Total direct effect of smoking on outcomes,estimates using a directed acyclic graph (‘causal diagram’)

    Discussion

    It is crucial to remain impartial in our interpretation when faced with a lack of statistical significance, as relying solely on hypothesis testing can lead to the rejection of studies that may nevertheless provide valuable insights.23 In the current investigation, notwithstanding the absence of statistically significant associations between individuals with a history of smoking and adverse clinical outcomes in COVID-19 patients, notable indications of a potential inclination towards unfavourable clinical outcomes were observed. Specifically, these were evident in the context of the WHO ordinal COVID-19 scale, use of oxygen therapy (non-invasive or invasive), ICU admission and hospital LOS among patients with a smoking history. These associations require careful interpretation within the clinical context. They consistently emerged across various analytical approaches, including crude analysis, matching weights and cautious adjustments for confounding factors.

    By adjusting for the confounding factors, we were able to make sense of the counterintuitive initial results produced by the crude analysis. For instance, as depicted in figure 3, the crude analysis results indicate that the death rate among never-smokers was higher than that among ever-smokers by approximately 11%. However, when the two groups were matched by weight, the death rates, although still slightly higher among never-smokers, varied considerably less. This discrepancy is likely due to the imbalance in baseline characteristics among the groups, with diabetes and hypertension being more prevalent in the never-smoker group. Had we relied solely on the crude analysis, we might have concluded that smoking protects against COVID-19 mortality.

    At a mechanistic level, in the case of SARS-CoV-2 infection, smoking is thought to enhance the expression of ACE2 receptors. It is also associated with immune suppression, oxidative stress inflammation and vascular injury. Consequently, smoking increases susceptibility to more severe forms of COVID-19 disease.4 Although one single study may not fully address the question under discussion, it is important to note that earlier data from the first wave of the pandemic presented paradoxical findings, particularly regarding the population of current smokers.

    For instance, in a UK study conducted by Gao et al24 during the first wave, current smokers exhibited a reduced risk of severe COVID-19. However, mortality outcomes increased after careful adjustment of potential confounders. Several researchers have highlighted methodological issues with this research, including reliance on record linkage, usage of data not specifically designed to answer the question, the risk of information and classification errors during the pandemic and non-random outcome misclassification, as one-fifth of those who died were never hospitalised and were, therefore, less likely to be tested and diagnosed if they were current smokers. To their credit, Gao et al supplemented their paper with post-hoc analyses that addressed some of the issues raised by referees and other researchers.25 Furthermore, these contradictory findings appeared in the OpenSAFELY study, where authors analysed data from the UK national primary care EHR linked to COVID-19 death data.26 It is also important to note that this study was not specifically designed to examine the association between smoking and COVID-19 mortality; however, as indicated by other researchers, interpreting the HRs in their tables as risk factors may represent a classic table 2 Fallacy.26

    Notably, a recent meta-analysis conducted by Gallus et al found that the pooled OR estimate for the risk of severe COVID-19 diseases was 1.44 (95% CI 1.31 to 1.58) for an ever-smoker group versus a never-smoker group, which was close to the findings in the current study.27 These findings were confirmed by a robust Mendelian randomisation study conducted by Clift et al using UK Biobank data. This was undertaken in the general population, making it less susceptible to selection bias. The findings suggest that smoking increases the risk of acquiring SARS-CoV-2 infection, hospitalisation and death.28

    To the best of our knowledge, this study offered the first opportunity to directly study the association between smoking and COVID-19 severity in Saudi Arabia. Previous studies in our population have discussed the smoking behavioural impact of COVID-19 or quality of life measures.29 30 Furthermore, the current study attempts to address some of the methodological limitations in the literature. First, not unique to our setting, the reliance on EHRs to determine smoking status has challenged researchers.31 In a study by Polubriaginof et al conducted in New York-Presbyterian Hospital/Columbia University Medical Centre, retrospective data were collected to illustrate the challenges of obtaining accurate information on smoking status. Despite the use of EHR systems, reporting smoking status was often dispersed in clinical notes, and there was no standardisation of reporting by the clinicians. In addition, their analysis revealed that when reported smoking status was treated as a dynamic concept (ie, it changes with time), implausible changes in smoking status were observed in 54.5% of cases. For example, some patients were reported as current smokers, but in other visits, they were reported as being never smokers.31 Determining exposure status in our study was essential to avoid or minimise misclassification bias.32 In particular, we were aware that such information may not be accurate if taken directly from the EHR system without confirmation. In addition, given the urgency of the pandemic, clinicians may have committed omission errors when acquiring such important information. Therefore, we contacted all eligible patients to confirm their exposure status prior to COVID-19 admission, which could have introduced a degree of recall bias, although this was less likely given the weight of the event (hospital admission due to COVID-19). Moreover, this recall bias was always present, as the nature of the exposure variable requires subjective data from the patient. Furthermore, because the study was conducted in a COVID-19 centre, the availability of SARS-CoV testing was not a limitation, as access to healthcare was not an issue.

    Using our methods, we also avoided overadjustments or adjusting potential colliders or mediators, which are known to attenuate the effect of the estimates or lead to reverse causality.33 34 It was imperative to navigate the causal pathway and carefully select variables to ascertain the total causal effect of smoking status on COVID-19 severity. In addition, our analytical approach used matching weights, in which we balanced all groups in terms of comorbid conditions. This was illustrated in the love’s plot (figure 2). One advantage of using the SMD to examine covariate balance between the two groups was that it was independent of the sample size, thereby allowing a comparison to be made of mean differences in units of the pooled SD. As argued by Imai et al, the use of significance testing to assess baseline balance in a matched sample is inappropriate due to reduced statistical power and the property of balance being specific to a particular sample.35 Moreover, given that there is value in examining outcome variables as ordinal rather than binary, using ordinal regression models was appropriate. There are also several notable advantages to using an ordinal regression model. For instance, it increases the power of the analysis because it respects the ordinality of the outcome. Furthermore, it can provide more nuanced information about the relationship between predictor and outcome variables when the covariate balance between two groups is examined.36

    The current study, however, involves several limitations that need to be addressed. First, the calculation of the sample size was based on the assumption of identifying an association that carries twice the risk of COVID-19 severity.37 Increasing the sample size could have produced statistically significant results for the severity outcome and given a more precise estimate of the smoking effect. However, we were limited by the number of patients to be screened who were admitted prior to the introduction of vaccines in our centre. Nonetheless, our results should be interpreted in the broader clinical context and the existing literature, as they provide clinical insights into the risk associated with smoking and COVID-19 severity. Second, because we relied on receiving confirmation of smoking status, we necessarily excluded many patients from the original sample pool due to unconfirmed smoking status, the inclusion of which could have increased the power of the study. Third, due to the low number of e-cigarette smokers, we could not perform an analysis to address this type of smoking. Fourth, only 33 patients were current smokers. No deaths occurred in this population, which could have resulted in a bias in the true estimates for COVID-19 severity in this cohort. We estimated COVID-19 severity to be similar in most of the WHO categories, with an emphasis on the non-significantly higher odds of the use of non-invasive mechanical ventilation or high-flow nasal cannula (adjusted OR=1.07 with a 95% CI 0.99 to 1.15, p=0.062, (online supplemental table S1). Using the DAGs method, we carefully selected the adjustments set for this analysis. The results disagree with some of the literature, which suggests that current smoking is protective against severe COVID-19, confirming the concerns raised by Westreich and colleagues on the possibility of a table 2 Fallacy in interpreting some of these data.26

    Another limitation of our study is that because we used multiple significance tests, our results are vulnerable to a type I error, which occurs when researchers conduct multiple significance tests and emphasise the test with the highest statistical significance. This increases the probability of a statistically significant ‘false-positive’ result.38 Although adjustments for the multiple testing problem exist, conducting these would have exposed our results to a type II error or led to a failure to detect significant effects, even when they are present.

    Finally, the generalisability of our study may be limited as new strains of the coronavirus continue to emerge and ignite new waves of infection around the globe. It is imperative that we reach a better understanding of the way in which risk factors such as cigarette smoking interact with disease progression and outcomes. Therefore, we recommend that larger, in-depth studies be conducted to investigate COVID-19 outcomes among ever-smokers and examine smoking behaviour, including current versus former status, type of smoking and frequency. Moreover, we recommend further investigation into the long-term impact of e-cigarette use, given the increasing prevalence of their usage on a local and global scale.39 40

    Conclusion

    We investigated the association between smoking history and adverse clinical outcomes among COVID-19 patients admitted to a tertiary hospital in Saudi Arabia. Although our analysis revealed no statistically significant association between individuals with a history of smoking and adverse COVID-19 clinical outcomes, we found notable indications of a potential inclination towards unfavourable clinical outcomes. Among the objectives that we sought to achieve was to address the methodological limitations in the literature surrounding the topic of smoking status and COVID-19 outcomes. To achieve this, we used various approaches to analyse our results and compare the outcomes.

    Data availability statement

    Data available upon request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants and was approved by the Dammam Medical Complex Institutional Review Board in Dammam, Saudi Arabia, IRB Protocol No: PH-20. Participants gave informed consent to participate in the study before taking part.

    References

    1. ↵
      1. World Health Organization
      . WHO COVID-19 dashboard. Geneva; 2023. Available: https://covid19.who.int/ [Accessed Jul 2023].
    2. ↵
      1. Chen T,
      2. Wu D,
      3. Chen H, et al
      . Clinical characteristics of 113 deceased patients with Coronavirus disease 2019: retrospective study. BMJ 2020;368:m1091. doi:10.1136/bmj.m1091
    3. ↵
      1. Lake MA
      . What we know so far: COVID-19 current clinical knowledge and research. Clin Med (Lond) 2020;20:124–7:124. doi:10.7861/clinmed.2019-coron
      OpenUrl
    4. ↵
      1. Benowitz NL,
      2. Goniewicz ML,
      3. Halpern-Felsher B, et al
      . Tobacco product use and the risks of SARS-Cov-2 infection and COVID-19: Current understanding and recommendations for future research. Lancet Respir Med 2022;10:900–15:S2213-2600(22)00182-5. doi:10.1016/S2213-2600(22)00182-5
      OpenUrl
    5. ↵
      1. World Health Organization
      . Smoking and COVID-19: scientific brief. 2020. Available: https://www.who.int/publications/i/item/WHO-2019-nCoV-Sci_Brief-Smoking-2020.2 [accessed Jul 2023]
    6. ↵
      1. van Zyl-Smit RN,
      2. Richards G,
      3. Leone FT
      . Tobacco smoking and COVID-19 infection. Lancet Respir Med 2020;8:664–5:S2213-2600(20)30239-3. doi:10.1016/S2213-2600(20)30239-3
      OpenUrl
    7. ↵
      1. Patanavanich R,
      2. Siripoon T,
      3. Amponnavarat S, et al
      . Active Smokers are at higher risk of COVID-19 death: A systematic review and meta-analysis. Nicotine & Tobacco Research 2023;25:177–84. doi:10.1093/ntr/ntac085
      OpenUrl
    8. ↵
      1. Algabbani AM,
      2. Almubark R,
      3. Althumiri N, et al
      . The prevalence of cigarette smoking in Saudi Arabia in 2018. FDRSJ 2018;1:1. doi:10.32868/rsj.v1i1.22
      OpenUrl
    9. ↵
      1. von Elm E,
      2. Altman DG,
      3. Egger M, et al
      . The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg 2014;12:1495–9:S1743-9191(14)00212-X. doi:10.1016/j.ijsu.2014.07.013
      OpenUrlCrossRefPubMed
    10. ↵
      1. Vassar M,
      2. Holzmann M
      . The retrospective chart review: important methodological considerations. J Educ Eval Health Prof 2013;10:12. doi:10.3352/jeehp.2013.10.12
    11. ↵
      1. Harris PA,
      2. Taylor R,
      3. Thielke R, et al
      . Research electronic data capture (Redcap)--A Metadata-driven methodology and Workflow process for providing Translational research Informatics support. J Biomed Inform 2009;42:377–81. doi:10.1016/j.jbi.2008.08.010
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      Diagnostic and statistical Manual of mental disorders [American Psychiatric Association]. In: Diagnostic and Statistical Manual of Mental Disorders Fifth Edition Text Revision—DSM-5-TR. Washington, DC, USA: American Psychiatric Association, Available: https://psychiatryonline.org/doi/book/10.1176/appi.books.9780890425787
    13. ↵
      Adult tobacco use information centers for disease control and prevention, special topics [National Center for Health Statistics]. 2017. Available: https://www.cdc.gov/nchs/nhis/tobacco/tobacco_glossary.htm#:~:text=Smoking%20status%3A%20A%20recoded%20variable,smoked%2C%20and%20smoking%20status%20unknown [Accessed Jul 2023].
    14. ↵
      1. Villarroel MA,
      2. Cha AE,
      3. Vahratian A
      . Electronic cigarette use among US adults. 2018.
    15. ↵
      1. Li L,
      2. Greene T
      . A weighting analogue to pair matching in propensity score analysis. Int J Biostat 2013;9:215–34. doi:10.1515/ijb-2012-0030
      OpenUrlPubMed
    16. ↵
      1. Yoshida K,
      2. Hernández-Díaz S,
      3. Solomon DH, et al
      . Matching weights to simultaneously compare three treatment groups: comparison to three-way matching. Epidemiology 2017;28:387–95. doi:10.1097/EDE.0000000000000627
      OpenUrlCrossRefPubMed
    17. ↵
      1. Desai RJ,
      2. Franklin JM
      . Alternative approaches for confounding adjustment in observational studies using weighting based on the propensity score: a primer for practitioners. BMJ 2019;367:l5657. doi:10.1136/bmj.l5657
    18. ↵
      1. Austin PC
      . Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011;10:150–61. doi:10.1002/pst.433
      OpenUrlCrossRefPubMed
    19. ↵
      1. Wickham H
      . Ggplot2. In: ggplot2: Elegant Graphics for Data Analysis. Cham: Springer-Verlag New York, 2016. Available: http://link.springer.com/10.1007/978-3-319-24277-4
    20. ↵
      1. Lumley T
      . svyVGAM: design-based inference in vector generalised linear models. 2021.
    21. ↵
      1. Mangiafico SS
      . Functions to support extension education program evaluation. 2023. Available: https://CRAN.R-project.org/package=rcompanion/ [Accessed Feb 2024].
    22. ↵
      1. Tennant PWG,
      2. Murray EJ,
      3. Arnold KF, et al
      . Use of directed Acyclic graphs (Dags) to identify confounders in applied health research: review and recommendations. Int J Epidemiol 2021;50:620–32. doi:10.1093/ije/dyaa213
      OpenUrlPubMed
    23. ↵
      1. Greenland S,
      2. Senn SJ,
      3. Rothman KJ, et al
      . Statistical tests, P values, confidence intervals, and power: a guide to Misinterpretations. Eur J Epidemiol 2016;31:337–50. doi:10.1007/s10654-016-0149-3
      OpenUrlCrossRefPubMed
    24. ↵
      1. Gao M,
      2. Aveyard P,
      3. Lindson N, et al
      . Association between smoking, E-cigarette use and severe COVID-19: a cohort study. Int J Epidemiol 2022;51:1062–72. doi:10.1093/ije/dyac028
      OpenUrlPubMed
    25. ↵
      1. Lam TH
      . Commentary: smoking, nicotine and COVID-19 outcomes: unprecedented challenges to Epidemiologists. Int J Epidemiol 2022;51:1075–7. doi:10.1093/ije/dyac138
      OpenUrl
    26. ↵
      1. Westreich D,
      2. Edwards JK,
      3. van Smeden M
      . Comment on Williamson et al. (Opensafely): the table 2 fallacy in a study of COVID-19 mortality risk factors. Epidemiology 2021;32:e1–2. doi:10.1097/EDE.0000000000001259
      OpenUrl
    27. ↵
      1. Gallus S,
      2. Scala M,
      3. Possenti I, et al
      . The role of smoking in COVID-19 progression: a comprehensive meta-analysis. Eur Respir Rev 2023;32:220191. doi:10.1183/16000617.0191-2022
    28. ↵
      1. Clift AK,
      2. von Ende A,
      3. Tan PS, et al
      . Smoking and COVID-19 outcomes: an observational and Mendelian Randomisation study using the UK Biobank cohort. Thorax 2022;77:65–73. doi:10.1136/thoraxjnl-2021-217080
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Al Kalif MSH,
      2. Alghamdi AA,
      3. Albagmi FM, et al
      . The negative impact of tobacco smoking on health-related quality of life among Saudi adolescents during the COVID-19 pandemic. Asian Pac J Cancer Prev 2021;22:4051–6:89916. doi:10.31557/APJCP.2021.22.12.4051
      OpenUrl
    30. ↵
      1. Tobaiqy M,
      2. MacLure A,
      3. Thomas D, et al
      . The impact of COVID-19 on smoking Behaviours and support for smoke-free zones in Saudi Arabia. Int J Environ Res Public Health 2021;18:6927. doi:10.3390/ijerph18136927
    31. ↵
      1. Polubriaginof F,
      2. Salmasian H,
      3. Albert DA, et al
      . Challenges with collecting smoking status in electronic health records. AMIA Annu Symp Proc 2017;2017:1392–400.
      OpenUrl
    32. ↵
      1. Warden G
      . Definitions of bias in clinical research. Methods Mol Biol 2021;2249:35–52. doi:10.1007/978-1-0716-1138-8_3
      OpenUrl
    33. ↵
      1. Schisterman EF,
      2. Cole SR,
      3. Platt RW
      . Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology 2009;20:488–95. doi:10.1097/EDE.0b013e3181a819a1
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Griffith GJ,
      2. Morris TT,
      3. Tudball MJ, et al
      . Collider bias undermines our understanding of COVID-19 disease risk and severity. Nat Commun 2020;11:5749. doi:10.1038/s41467-020-19478-2
    35. ↵
      1. Imai K,
      2. King G,
      3. Stuart EA
      . Misunderstandings between Experimentalists and Observationalists about causal inference. Journal of the Royal Statistical Society Series A 2008;171:481–502. doi:10.1111/j.1467-985X.2007.00527.x
      OpenUrlCrossRefWeb of Science
    36. ↵
      1. Scott SC,
      2. Goldberg MS,
      3. Mayo NE
      . Statistical assessment of Ordinal outcomes in comparative studies. J Clin Epidemiol 1997;50:45–55. doi:10.1016/s0895-4356(96)00312-5
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Reddy RK,
      2. Charles WN,
      3. Sklavounos A, et al
      . The effect of smoking on COVID-19 severity: A systematic review and meta-analysis. J Med Virol 2021;93:1045–56. doi:10.1002/jmv.26389
      OpenUrlPubMed
    38. ↵
      1. Molloy SF,
      2. White IR,
      3. Nunn AJ, et al
      . Multiplicity adjustments in parallel-group multi-arm trials sharing a control group: clear guidance is needed. Contemp Clin Trials 2022;113:106656. doi:10.1016/j.cct.2021.106656
    39. ↵
      1. Althobaiti NK,
      2. Mahfouz MEM
      . Prevalence of electronic cigarette use in Saudi Arabia. Cureus 2022;14:e25731. doi:10.7759/cureus.25731
    40. ↵
      1. Tehrani H,
      2. Rajabi A,
      3. Ghelichi-Ghojogh M, et al
      . The prevalence of electronic cigarettes Vaping globally: a systematic review and meta-analysis. Arch Public Health 2022;80:240. doi:10.1186/s13690-022-00998-w

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors MAA and AA contributed to the design of the work, data acquisition, supervised data collection, data analysis, interpretation of the data and writing the manuscript. MA contributed to data analysis, interpretation and writing the manuscript. TKA contributed to the study design and interpretation of the data and provided academic support. FA, FMA and AMA contributed to the study design and resources and provided academic support. LA-L and HA contributed to the sample collection and data management and provided logistic support. MA contributed to the study design and logistic support. MAA accepts full responsibility for the finished work and the conduct of the study, had access to the data, and controlled the decision to publish. All authors have read and agreed to the published version of the manuscript.

    • Funding Financial funding was provided by King Abdul-Aziz City for Science and Technology, General Directorate for Fund and Grants to King Saud University to implement this work through the Fast Track Program for COVID-19 Research Project No.5-21-01-001-0015.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.