The effect of smoking on COVID‐19 severity: A systematic review and meta‐analysis

Abstract Various comorbidities represent risk factors for severe coronavirus disease 2019 (COVID‐19). The impact of smoking on COVID‐19 severity has been previously reported in several meta‐analyses limited by small sample sizes and poor methodology. We aimed to rigorously and definitively quantify the effects of smoking on COVID‐19 severity. MEDLINE, Embase, CENTRAL, and Web of Science were searched between 1 December 2019 and 2 June 2020. Studies reporting smoking status of hospitalized patients with different severities of disease and/or at least one clinical endpoint of interest (disease progression, intensive care unit admission, need for mechanical ventilation, and mortality) were included. Data were pooled using a random‐effects model. This study was registered on PROSPERO: CRD42020180920. We analyzed 47 eligible studies reporting on 32 849 hospitalized COVID‐19 patients, with 8417 (25.6%) reporting a smoking history, comprising 1501 current smokers, 5676 former smokers, and 1240 unspecified smokers. Current smokers had an increased risk of severe COVID‐19 (risk ratios [RR]: 1.80; 95% confidence interval [CI]: 1.14‐2.85; P = .012), and severe or critical COVID‐19 (RR: 1.98; CI: 1.16‐3.38; P = .012). Patients with a smoking history had a significantly increased risk of severe COVID‐19 (RR: 1.31; CI: 1.12‐1.54; P = .001), severe or critical COVID‐19 (RR: 1.35; CI: 1.19‐1.53; P < .0001), in‐hospital mortality (RR: 1.26; CI: 1.20‐1.32; P < .0001), disease progression (RR: 2.18; CI: 1.06‐4.49; P = .035), and need for mechanical ventilation (RR: 1.20; CI: 1.01‐1.42; P = .043). Patients with any smoking history are vulnerable to severe COVID‐19 and worse in‐hospital outcomes. In the absence of current targeted therapies, preventative, and supportive strategies to reduce morbidity and mortality in current and former smokers are crucial.


SUPPLEMENTARY APPENDIX
The effect of smoking on COVID-19 severity: a systematic review and meta-analysis Rohin K. Reddy*, Walton N. Charles*, Alexandros Sklavounos, Atul Dutt, Paul T. Seed and Ankur Khajuria *Rohin K. Reddy and Walton N. Charles are co-first authors and contributed equally to this work.

Table of Contents
Page 1. Summary of previous systematic reviews i. Table S1: Key findings of previous reviews 2 ii. Conversion of Newcastle-Ottawa Scale scores to overall grade 16

Supplementary results
i. Table S8: Breakdown of Newcastle-Ottawa Scale grading for included studies 17 ii. Figure S1: Funnel plots for the effect of current smoking on A. Severe COVID- 19 18 B. Severe or critical COVID- 19 18 C. Mortality 18 iii. Figure S2: Forest and funnel plots for the effect of current smoking on A. Disease progression 19 B. ICU admission 20 C. Mechanical ventilation requirement 21 iv. Table S9: Sensitivity analyses of the effect of current smoking 22 v. Figure S3: Forest plot showing the effect of current smoking on mortality by country 23 vi. Figure S4: Funnel plots for the effect of a smoking history on A. Severe COVID- 19 24 B. Severe or critical COVID- 19 24 C. Mortality 24 vii. Figure Table S10: Sensitivity analyses of the effect of a smoking history 29 ix. Figure S6: Forest plot showing the effect of a smoking history on mortality by country 29

2.
Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? Criteria for grading overall quality: § High -No or one non-critical weakness: The systematic review provides an accurate and comprehensive summary of the results of the available studies that address the question of interest. § Moderate -More than one non-critical weakness*: The systematic review has more than one weakness, but no critical flaws. It may provide an accurate summary of the results of the available studies that were included in the review. § Low -One critical flaw with or without non-critical weaknesses: The review has a critical flaw and may not provide an accurate and comprehensive summary of the available studies that address the question of interest. § Critically low -More than one critical flaw with or without non-critical weaknesses: The review has more than one critical flaw and should not be relied on to provide an accurate and comprehensive summary of the available studies. § *Note: Multiple non-critical weaknesses may diminish confidence in the review and it may be appropriate to move the overall appraisal down from moderate to low confidence. Other well-established criteria include: BTS CURB65 score, IDSA/ATS criteria and Pneumonia Severity Index.
In the absence of well-established criteria, clinical endpoints of ICU admission, requirements for mechanical ventilation and/or death will be used as surrogate markers for severity.

Data extraction (selection and coding)
Two reviewers will independently screen the titles and abstracts of the papers retrieved in the searches for eligibility.
The full-texts of papers identified as being potentially eligible for inclusion will then be screened for inclusion. Any disagreements between the two authors will be resolved by discussion with a third reviewer.
Data will then be extracted from the studies selected for inclusion, as follows: first author, year of publication, study design, ethics approval, study setting, study population including multicentre studies, participant demographics (including smoking status), disease severity and clinical outcomes (ICU admission, mechanical ventilation requirement, and death).

Risk of bias (quality) assessment
The GRADE (Grading of Recommendations, Assessment, Development and Evaluations) framework will be used to evaluate the quality of each study based on: risk of bias, imprecision, inconsistency, indirectness and publication bias. Studies will be given a level of evidence of very low, low, moderate or high.
Strategy for data synthesis COVID-19 outcomes will be compared between smokers and non-smokers.
Outcome data will be pooled using Review Manager software (The Cochrane Collaboration, Denmark) and evaluated using odds ratios with 95% confidence intervals.
Heterogeneity between studies will be assessed in Review Manager, with a meta-analysis performed if the studies are relatively homogenous regarding methodology and outcomes.

PROSPERO International prospective register of systematic reviews
The results of the meta-analyses will be shown as forest plots.
If, however, a meta-analysis should prove to be inappropriate, a narrative synthesis will be performed instead. NB: Since the protocol was registered, a significant number of papers were found to report patient outcomes by smoking history. Hence, in addition to analysing the effect of current smoking, a separate analysis was conducted comparing those with a smoking history (current and former smokers) with never-smokers.

Stage
Started Completed

Preliminary searches Yes No
Piloting of the study selection process No No Formal screening of search results against eligibility criteria No No

Data extraction No No
Risk of bias (quality) assessment No No

Data analysis No No
The record owner confirms that the information they have supplied for this submission is accurate and complete and they understand that deliberate provision of inaccurate information or omission of data may be construed as scientific misconduct.
The record owner confirms that they will update the status of the review when it is completed and will add publication details in due course.

Versions
27 April 2020 PROSPERO This information has been provided by the named contact for this review. CRD has accepted this information in good faith and registered the review in PROSPERO. The registrant confirms that the information supplied for this submission is accurate and complete. CRD bears no responsibility or liability for the content of this registration record, any associated files or external websites.  Structured summary Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number.

Rationale
Describe the rationale for the review in the context of what is already known. 6 4.
Objectives Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS). 7

Protocol and registration
Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number.
Eligibility criteria Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.
Information sources Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched. 8 8.
Search Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

9.
Study selection State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis). 9

Data collection process
Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators. 9 11. Data items List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made. 9

Risk of bias in individual studies
Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis. 10

Study selection
Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram. For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Table 1 19. Risk of bias within studies Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Comparator group Outcome 2. Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol?
For Partial Yes: The authors state that they had a written protocol or guide that included ALL the following: For Yes: As for partial yes, plus the protocol should be registered and should also have specified: review question(s) a meta-analysis/synthesis plan, if appropriate, and Yes Partial Yes No a search strategy a plan for investigating causes of heterogeneity inclusion/exclusion criteria a plan for investigating causes of heterogeneity a risk of bias assessment

Did the review authors explain their selection of the study designs for inclusion in the review?
For Yes, the review should satisfy ONE of the following:

Did the review authors perform study selection in duplicate?
For Yes, either ONE of the following: at least two reviewers independently agreed on selection of eligible studies and achieved consensus on which studies to include Yes No OR two reviewers selected a sample of eligible studies and achieved good agreement (at least 80 percent), with the remainder selected by one reviewer.

Did the review authors perform data extraction in duplicate?
For Yes, either ONE of the following: at least two reviewers achieved consensus on which data to extract from included studies Yes No OR two reviewers extracted data from a sample of eligible studies and achieved good agreement (at least 80 percent), with the remainder extracted by one reviewer.

Did the review authors provide a list of excluded studies and justify the exclusions?
For Partial Yes: For Yes, must also have: provided a list of all potentially relevant studies that were read in full-text form but excluded from the review Justified the exclusion from the review of each potentially relevant study

Did the review authors describe the included studies in adequate detail?
For Partial Yes (ALL the following): For Yes, should also have ALL the following:

For Yes
Must have reported on the sources of funding for individual studies included in the review. Note: Reporting that the reviewers looked for this information but it was not reported by study authors also qualifies Yes No 11. If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results?

RCTs
For Yes: The authors justified combining the data in a meta-analysis Yes No No meta-analysis conducted AND they used an appropriate weighted technique to combine study results and adjusted for heterogeneity if present.

AND investigated the causes of any heterogeneity
For NRSI For Yes: The authors justified combining the data in a meta-analysis Yes No No meta-analysis conducted AND they used an appropriate weighted technique to combine study results, adjusting for heterogeneity if present AND they statistically combined effect estimates from NRSI that were adjusted for confounding, rather than combining raw data, or justified combining raw data when adjusted effect estimates were not available AND they reported separate summary estimates for RCTs and NRSI separately when both were included in the review

If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis?
For Yes: included only low risk of bias RCTs Yes No No meta-analysis conducted OR, if the pooled estimate was based on RCTs and/or NRSI at variable RoB, the authors performed analyses to investigate possible impact of RoB on summary estimates of effect.

Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review?
For Yes: included only low risk of bias RCTs Yes No OR, if RCTs with moderate or high RoB, or NRSI were included the review provided a discussion of the likely impact of RoB on the results

Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review?
For Yes: There was no significant heterogeneity in the results

Yes No
OR if heterogeneity was present the authors performed an investigation of sources of any heterogeneity in the results and discussed the impact of this on the results of the review

If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review?
For Yes: performed graphical or statistical tests for publication bias and discussed the likelihood and magnitude of impact of publication bias

Yes No
No meta-analysis conducted 16. Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review?
For Yes: The authors reported no competing interests OR Yes No The authors described their funding sources and how they managed potential conflicts of interest  44 used a locally devised criteria based on chest X-ray findings § Each lung was divided into three zones (upper zone from superior hilar markings to the apices, middle zone from inferior to superior hilar markings and lower zone from the costophrenic sulcus to inferior hilar markings), with each zone graded on whether an opacity was absent (0) or present (1) § Total scores were calculated (maximum of 6) § Severe cases defined as a score of [3][4][5][6] Due to variations in the diagnostic criteria for grading severity (e.g. any patient meeting the major criteria of the IDSA/ATS criteria for 'severe' disease would actually be classified as having 'critical' disease according to the Chinese COVID-19-specific criteria), we performed further sensitivity analyses restricting to only studies using the Chinese criteria developed specifically to evaluate COVID-19 severity.  14 § Progression from non-severe to severe or critical disease § Severity graded according to the Chinese COVID-19-specific criteria Huang J et al. 27 § Required oxygen support (including high-flow oxygen supply, as well as invasive or non-invasive mechanical ventilation) Ji D et al. 31 § Progression from non-severe to severe disease § Severity graded according to the Chinese COVID-19-specific criteria, with additional criteria including mechanical ventilation requirement or worsening lung CT scan findings Liu W et al. 37 § Progression from non-severe to severe or critical disease or death § Progression from severe to critical disease or death § Progression from critical disease to death § Severity graded according to the Chinese COVID-19-specific criteria Yu T et al. 51 § Worsening lung CT scan findings 1-week post-treatment Yu X et al. 52 § Progression from non-severe (mild/moderate) to severe disease § Severity criteria not specified Zheng Y et al. 57 § Progression from non-severe to severe or critical disease § Severity graded according to the Chinese COVID-19-specific criteria Scoring components (any to score) Non-scoring components Selection (out of 4) 1.
Is the case definition adequate? § Independent validation (e.g. more than 1 person/record/time/process used to extract information) § Reference to primary record source § Record linkage (e.g. ICD codes in databases of unverifiable registry-based studies) § Self-report with no reference to primary record § No description 2.
Representativeness of the cases § Consecutive/obviously representative series § All eligible cases with outcome of interest over defined period of time/hospital § Only investigated a specific population of hospitalised COVID-19 patients (only those with e.g. cancer/IBD/admitted to ICU etc.) § Potential for selection bias or not stated 3.
Selection of controls § Community controls (i.e. same community as cases and would be cases if they had outcome) § Hospital controls § No description 4.
Definition of controls § No history of study endpoint stated § No description Comparability (out of 2) 5.
Age § Adjusted for age in multivariate model § Exposed and unexposed cases matched in design by age § Statements of no difference between groups/ differences were not statistically significant 6. Any other significant risk factor § As above, but for any other significant risk factor (such as gender or comorbidities) § Statements of no difference between groups/ differences were not statistically significant Exposure (out of 3) 7.
Ascertainment of exposure § Secure record § Structured interview conducted blind to case/ control status § Self-reported § Interview not blinded to case/control § No description 8.
Same method of ascertainment for cases and controls § Yes § Different method used 9. Non-response rate § Same rates for both groups § Different rates, even if difference explained Scoring components (any to score) Non-scoring components Selection (out of 4) 1.
Representativeness of the exposed cohort § Representative cohort of COVID-19 patients § Specific cohort of hospitalised COVID-19 patients (only those with e.g. cancer/IBD/ admitted to ICU etc.) § No description 2.
Selection of the nonexposed cohort § Drawn from the same community as the above cohort § Drawn from a different community (e.g. compared patients to healthcare staff) § No description 3.
Ascertainment of exposure § Secure record § Structured interview § Written self-report § No description 4.
Demonstration that outcome of interest was not present at start of study § Stated (e.g. in studies of progression to severe disease, must state that only mild/moderate disease patients were included or that severe disease patients were not included) § Not stated Comparability (out of 2) 5.
Age § Adjusted for age in multivariate model § Cohorts matched by age in design § Statements of no difference between groups/ differences were not statistically significant 6.
Any other significant risk factor § As above, but for any other significant risk factor (such as gender or comorbidities) § Statements of no difference between groups/ differences were not statistically significant Outcome (out of 3) 7.
Assessment of outcome § Secure record/database § Independent assessment (e.g. more than 1 person/record/time/process) § Self-reported § No reference to original medical records § No description of how outcomes extracted 8.
Was follow-up long enough for outcomes to occur? § Follow-up of at least 30 days stated/calculated from last date of enrolment NB: Short follow-up selected in recognition of current urgency of COVID-19 research § Follow-up less than 30 days § Patients still in hospital at the end of the study period with potential for outcomes of interest to occur 9.
Adequacy of follow up of cohorts § Complete follow-up (all patients accounted for) § Still score if small amount (15% or less) not accounted for if they're unlikely to introduce bias or if paper describes what happened to them (e.g. discharged/lost contact etc.) § Follow-up rate less than 85% § No description of patients unaccounted for Scoring components (any to score) Non-scoring components Selection (out of 4) 1.
Representativeness of the sample § Representative cohort of COVID-19 patients § Specific cohort of hospitalised COVID-19 patients (only those with e.g. cancer/IBD/ admitted to ICU etc.) § No description 2.
Sample size § Justified and satisfactory § Not justified 3.
Non-respondents (e.g. excluded patients) § Characteristics of respondents and nonrespondents compared (e.g. described basic characteristics of the excluded patients) § No description 4. Ascertainment of exposure § Secure record § Structured interview conducted blind to case/ control status § Self-reported § Interview not blinded to case/control § No description Comparability (out of 2) 5.
Age § Adjusted for age in multivariate model § Cohorts matched by age in design § Statements of no difference between groups/ differences were not statistically significant 6.
Any other significant risk factor § As above, but for any other significant risk factor (such as gender or comorbidities) § Statements of no difference between groups/ differences were not statistically significant Outcome (out of 2) 7.
Assessment of outcome § Secure record/database § Independent assessment (e.g. more than 1 person/record/time/process) § Self-reported § No reference to original medical records § No description of how outcomes extracted 8.
Statistical test § Statistical test(s) appropriate and clearly described, with OR/RR, confidence intervals and p-values presented § Statistical test(s) not appropriate or described

Conversion of Newcastle-Ottawa Scale scores to overall grade of study quality
Score requirements in case series and cohort studies:

Overall
Bi X et al. 14 Zheng Y et al. 57 Huang J et al.   15 Azar K et al. 12 CDC COVID-19 Response Team 17 Kuderer N et al. 34 Kalligeros M et al. 32 Huang C et al.

C. Mechanical ventilation requirement
Overall Brenner E et al. 15 Kuderer N et al. 34 Hur K et al. 29 Bhargava A et al. 13 Goyal P et al.   Sensitivity analyses were not performed for outcomes with only one study following restrictions.

Overall
Wang K et al. 46 Subtotal Subtotal

USA
Yao Q et al. 49 Chen T et al. 20 Brenner E et al. 15 Zhou F et al.

International
Docherty A et al.

UK
Azar K et al.

Overall
Li X et al. 35 Feng Y et al. 22 Hu L et al.

B. Disease progression
Disease progression was the only outcome in which Harbord's test implied publication bias (p=0.049).

Overall
Liu W et al. 37 Yu X et al. 52 Zheng Y et al. 57 Yu T et al. 51 Ji D et al.

Overall
Hur K et al. 29 Goyal P et al. 23 Toussie D et al. 44 Kuderer N et al.  Sensitivity analyses were not performed for outcomes with only one study following restrictions. Inciardi R et al. 30 Azar K et al.

Subtotal
Yang X et al. 48 Li Y et al. 36 Wang K et al. 46 Subtotal

Italy
Klang E et al.

USA
Shi Q et al. 41 Docherty A et al.