Hajek 2019.

Study characteristics
Methods	Design: Multicentre pragmatic randomized controlled trial to examine the efficacy of e‐cigarettes compared with nicotine replacement therapy Recruitment: participants attending UK stop‐smoking service and via social media Setting: U.K. National Health Service stop‐smoking services Study start date: 1 April 2015; Study end date: 31 March 2018
Participants	Total N: 886 N per arm: EC: 439; NRT: 447 Inclusion criteria: Adults who smoke (aged 18 or over) with no strong preference to use or not to use nicotine replacement or e‐cigarettes, and were currently not using either type of product Able to read/write/understand English Exclusion criteria: Pregnant or breastfeeding Strong preference to use or not use NRT or EC, currently not using either type of product 48% women; median age 41; median cpd 15 ; mean FTND 4.6; 41.5% reported past use of ECs Motivated to quit: Not reported
Interventions	EC: Refillable NRT: Informed of range of NRT products and selected preferred product, encouraged to use combination. Participants free to switch products. Supplies provided for up to 3 months EC: Starter pack (1 Kit, Aspire UK) provided along with 30 ml bottle of Tobacco Royale flavor e‐liquid, concentration 18 mg/ml. Participants showed how to use and asked to purchase future e‐liquid online or from local vape shops and to buy different EC device if the 1 provided did not meet their needs. Enouraged to experiment with e‐liquids of different strengths and flavors. If unable to obtain own supply, provided with further 10‐ml bottle (not proactively offered). Oral and written info on how to operate EC Both arms received multi‐session behavioral support as per UK stop‐smoking service practice (one‐to‐one sessions weekly with local clinicians, exhaled CO monitored for at least 4 weeks post‐TQD); signed behavioral contract not to use other therapy for at least 4 weeks
Outcomes	Weeks 4, 26 and 52 Cessation: Sustained and biochemically‐validated CO < 8 ppm Adverse events and biomarkers: “adverse reactions”: presence or absence of nausea, sleep disturbance and throat and mouth irritation, and respiratory symptoms (presence or absence of shortness of breath, wheezing, coughing and phlegm), death Other outcomes measured: Use and ratings of trial products Rating of withdrawal symptoms (weeks 1 ‐ 6) Reduction of cigarette consumption Cost effectiveness
Study funding	“Supported by the National Institute for Health Research (NIHR) Health Technology Assessment Programme (project number, 12/167/135) and by a grant (A16893) from the Cancer Research UK Prevention Trials Unit.”
Author declarations	From ICJME disclosure forms: “Miss Natalie Bisal has nothing to disclose. Dr. Dawkins reports personal fees from Johnson & Johnson, outside the submitted work; Dr. Goniewicz reports personal fees from Johnson and Johnson, outside the submitted work; Dr. Hajek reports grants and personal fees from Pfizer, outside the submitted work; Ms. Li reports grants from NCCHTA, during the conduct of the study; Dr. McRobbie reports grants from NIHR HTA program, during the conduct of the study; personal fees from Pfizer, personal fees from Johnson & Johnson, outside the submitted work; Dr. Myers Smith has nothing to disclose. Dr. Parrott has nothing to disclose. Dr. Pesola has nothing to disclose. Mrs Anna Phillips‐Waller has nothing to disclose. Dr. Przulj reports grants from Pfizer, outside the submitted work; Dr. Ross has nothing to disclose. Dr. Sasieni has nothing to disclose. Ms. Wu has nothing to disclose."
Notes	New for 2020 update, listed as ongoing study ISRCTN60477608 in 2016 review update Note higher use of allocated product at 12 m in intervention group compared to control group: “Among participants with 1‐year abstinence, 80% (63 of 79) were using e‐cigarettes at 52 weeks in the e‐cigarette group and 9% (4 of 44) were using nicotine replacement in the nicotine‐replacement group.”
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “Randomization took place on the quit date to limit differential dropout. Randomization sequences (1:1 ratio in permuted blocks of 20, stratified according to trial site) were generated with the use of a pseudorandom number generator in Stata software and were embedded into an application that only revealed the next treatment assignment once a participant had been entered into the database.”
Allocation concealment (selection bias)	Low risk	Refer to 'Random sequence generation'.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Not blinded, but as both arms contained active interventions performance bias judged unlikely
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Biochemical validation used
Incomplete outcome data (attrition bias) All outcomes	Low risk	At 12 months: EC Arm: 356/439 NRT Arm: 342/447
Selective reporting (reporting bias)	Low risk	All prespecified outcomes reported