This study was an investigator-initiated, randomized, multi-center, double-blind, parallel-group, placebo-controlled, phase 4 clinical trial. It focused on the potential of Swedish snus to reduce smoking and increase quit rates among cigarette smokers motivated to reduce their smoking or quit completely. The study was conducted in compliance with the ethical principles of the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice Guidelines (ICH-GCP) at two occupational health care centers in Belgrade, Serbia during January 2008 through March 2010 . The study was approved by the centers' institutional review board, and all participants provided written informed consent prior to entering the study. Before initiation, the study was registered on http://www.clinicaltrials.gov (identifier: NCT00601042).
Participants were recruited through posters and other printed material distributed at or in the vicinity of the study sites, and by word-of-mouth. The two sites were occupational health centers located at the head office of a large Serbian corporation (NIS-Jugopetrol) and at a major research institution in Belgrade (Vinča Institute of Nuclear Sciences). The inclusion criteria were: age between 20 through 65 years, history of daily smoking for more than one year, an average daily consumption of more than 10 cigarettes during the past month, motivation to substantially reduce or quit smoking, good general health, and acceptance not to take pharmaceutical nicotine products or any other non-protocol treatment to facilitate smoking cessation during the study period. Exclusion criteria were: uncontrolled hypertension (systolic > 140 mg Hg, diastolic > 90 mg Hg), history of coronary heart disease, other significant heart condition, or any medical condition that may interfere with study procedures, pregnancy or nursing, current abuse of alcohol or illicit drugs, current active oral disease that may interfere with use of study product, significant current psychiatric disease or psychosocial problems that may interfere with study procedures, and use of pharmaceutical or other products for smoking reduction or cessation within the past 3 months.
The products were manufactured by Swedish Match AB according to the GothiaTek® standard  and were supplied in identical, food-grade, plastic containers. The products came in sachets (pouches) that were placed in the anterior part of mouth between the upper gingiva and cheek for 30-60 minutes. The participants could choose from two different sachet sizes (0.5 g and 1 g) and two different flavors (liquorice and eucalyptus). Swedish snus according to the GothiaTek® standard is a low nitrosamine, moist, oral tobacco product with a water content of c. 45-55%, a nicotine content of c. 1%, and a pH of c. 8.5. The nicotine uptake from snus sachets in comparison with a 2 mg nicotine polacrilex gum was described previously . Snus was found to provide a more rapid uptake than the gum. However, the nicotine uptake from smoked products such as cigarettes is much more rapid than with oral tobacco due to the pulmonary mode of delivery .
The placebo snus products were almost identical to the snus products in physical appearance, mouth feel, pH, flavoring, and other sensory characteristics but they did not contain tobacco or nicotine.
With stratification by center, and using a block size of six, a predefined, central, computer-generated randomization sequence assigned participants in a 1:1 ratio to receive snus or matching placebo. Randomization was done by consecutively associating each included participant's identifiers with a unique, computer-generated sequential number. Lists at the study sites linked these numbers to specific study products, that is, snus or placebo. At the sites all study products were identified solely by identification numbers which ensured that both participants and investigators were blinded to treatment assignments. The protocol did not include procedures to assess the success of the blinding.
Each participant was scheduled to be followed for a total of 48 weeks. Study products were distributed to the participants during the entire study period. Participants were instructed to cut down on smoking as much as possible or quit smoking completely. Whenever they felt an urge to smoke, they were instructed to take a sachet of their allocated product. The number of sachets consumed per day was determined by the participants themselves. There was no prescribed tapering of product usage. Unblinding of treatment assignments was not done until after a participant had concluded the entire 48 week study period. Those who wanted to continue with snus after 48 weeks were obliged to buy commercially available products.
None of the study centers had previous experience with smoking cessation interventions (as quit smoking programs were non-existent in Serbia at the time of the trial) or smokeless tobacco (as there is no traditional use of such products in Serbia). However, the trialists attended training sessions prior to the initiation of the study which covered alternative approaches to smoking cessation, the chemical composition of snus, the epidemiology of snus use in Sweden, health effects of nicotine and snus versus cigarette smoking, how to use the study products, the need for adequate nicotine dosing to suppress cravings, methods for counseling of smokers who want to quit, and proper use of study equipment.
Potential participants were invited to seminars during which information was provided about health risks associated with smoking and available smoking cessation strategies. The physiological effects of nicotine were outlined, and an account given of the Swedish experience with snus including potential health risks associated with smokeless tobacco products. A few days-weeks after a seminar, those interested to participate were invited to a baseline visit during which their eligibility was determined and written informed consent to participate was obtained. All included participants were provided with their allocated study product at the baseline visit.
During the first 24 weeks the main study objective was to substantially reduce smoking so the participants were instructed to replace as many cigarettes as possible with their allocated study product or quit completely. They were informed that complete cessation should be the ultimate goal but that smoking reduction could be an important first step toward that aim. Information was given that one 1.0 g sachet used according to the instructions roughly should be able to replace one cigarette. All participants were encouraged to remain in the trial, attend all visits, and complete all assessments irrespective of study product usage or intensity of smoking. The participants were instructed to document on a weekly basis in a diary how many cigarettes they smoked on average per day, and how many study products they had used. Those who managed to achieve the protocol definition of a substantial smoking reduction at the week 24 visit or who had quit completely (see "Study end-points"), continued in the trial up to 48 weeks. During week 25-48 they were actively instructed to quit smoking completely. Participants who did not meet the protocol criteria for smoking reduction at the week 24 visit were counted as treatment failures in all efficacy analyses and were not actively followed after week 24.
The baseline visit was followed by 9 clinical visits over a total of 48 weeks. Participants were also contacted by telephone on two occasions (after 1 and 9 weeks). Each follow-up clinical visit comprised protocol assessments, a check of the participant's diary information, assessment of adverse events, and brief counseling (< 5-10 minutes).
The assessment at the baseline visit included medical history, history of smoking including previous quit attempts, measurements of height and weight, blood pressure, CO in exhaled air (Bedfont Micro Smokerlyzer, Sittingbourne, U.K.), assessment of nicotine dependence with the Fagerström Test for Nicotine Dependence (FTND) , pulmonary function tests (EasyOne Spirometer, ndd Medical Technologies, Zurich, Switzerland), and blood samples to assess the following biomarkers: total leukocytes (S-WBC), C-reactive protein (S-CRP), total S-cholesterol, high density lipoprotein (S-HDL), low density lipoprotein (S-LDL), S-fibrinogen, and S-cotinine.
Follow-up clinical visits were scheduled after week 2, 6, 12, 18, 24, 30, 36, 42, and 48. They included assessment of CO in exhaled air, self-reported smoking status and study product usage based on the participant's diary information, adverse events, and blood pressure. The pulmonary function tests, blood tests and measurement of weight were repeated at four of these visits (week 12, 24, 36, and 48). The FTND was administered at two follow up visits (week 24 and 48). The results were only considered relevant for those who reported continued smoking.
The telephone contacts scheduled after 1 and 9 weeks included assessment of self-reported smoking status, study product usage, and adverse events.
The primary end-point was smoking reduction at 24 weeks defined as a self-reported reduction of ≥ 50% in the average number of smoked cigarettes per day during week 21-24 compared to baseline, verified by a reduced concentration of carbon monoxide (CO) in exhaled air of at least 1 ppm. The protocol also included exploratory analyses of extent of smoking reduction (preceding 7 day period including abstinence days) compared to baseline according to predefined categories (100%, 75-99%, 50-74%, 25-49%, and < 25%).
Secondary end-points were evaluated at the week 12, 24, 36 and 48 clinical visits and included: CO-verified smoking reduction (≥ 50%) after 12 weeks (preceding 7 day period including abstinence days), point-prevalence estimate of smoking cessation (defined as self-reported total abstention from cigarettes during the preceding 7 day period verified by a concentration of CO in exhaled air of < 10 ppm at the clinical visit), estimates of continued smoking cessation (defined as self-reported total abstention from cigarettes during the preceding 4, 12, or 24-week period verified by a concentration of CO in exhaled air of < 10 ppm at all measurements during the specified period), and clinical tests and biomarkers including body weight, body mass index (BMI, defined as weight/height2), blood pressure, CO in exhaled air, pulmonary function tests including forced expiratory volume during one second (FEV1.0), forced vital capacity (FVC), the ratio between FEV1.0 and FVC (FEV%), and blood biomarkers (S-WBC, S-CRP, total S-cholesterol, S-HDL, S-LDL, S-fibrinogen, and S-cotinine).
An adverse event (AE) was defined as any symptom, physical sign or disease that either emerged during the study or, if present at the baseline visit, worsened during the study, regardless of the suspected cause of the event. Each AE was described by the responsible trialist in terms of duration, frequency, intensity, association with the study medication, assessment of possible causes, actions taken, and outcome. AEs for which an association with the allocated study product was considered "possible", "probable", or "definite" are reported in this paper as treatment-related. A serious AE (SAE) was defined as any AE that was fatal or life-threatening, permanently disabling, resulting in unplanned or prolonged hospitalization, or if medical interventions were required to prevent any of the mentioned outcomes.
Study monitoring and data handling
The study was coordinated and monitored by research personnel from an external contractor with a local office in Belgrade (i3 Research). They were also responsible for all data handling.
Efficacy data and intent-to-treat comparisons are reported for all randomized participants. All statistical methods were based on the International Conference on Harmonization (ICH) E9 document "Statistical Principles for Clinical Trials" . The statistical analyses were performed using SAS® v9.2 for Windows by an external contractor (i3 Statprobe).
In order to reliably detect (p < 0.05, statistical power > 80%) a more than two-fold increase in the odds of achieving smoking reduction at 24 weeks among the snus versus placebo groups, and assuming a smoking reduction rate of 15% in the placebo group versus 28% in the snus group (corresponding to an odds ratio of 2.2), the target sample size was estimated at 156 per treatment group for a total size of 312 study participants.
Demographics, vital statistics and other clinical data were summarized using summary statistics for continuous variables or by way of group frequencies and percentages for categorical variables.
In the efficacy analyses participants with missing or incomplete information, typically because of non-compliance with follow-up visits, were counted as not having achieved the end-point in question. Intent-to-treat comparisons of the proportion of participants achieving smoking reduction or cessation were done using logistic regression techniques allowing for allocated treatment, center, age at baseline, gender, and the interaction between treatment and center. Odds ratios were computed along with the 95% confidence intervals (95% C.I.) and two-sided p-values. The exploratory analyses of level of smoking reduction in the two treatment groups were done using Pearson's chi-squared test. Changes over time of average number of cigarettes smoked, vital signs, and biomarker data were analyzed using mixed effects repeated measures models. The models included allocated treatment, center, age at baseline, gender, and the interaction between treatment and center as fixed effects, and used unstructured residual covariance matrices for repeated records within subjects.