Intended for healthcare professionals


How much does relapse after one year erode effectiveness of smoking cessation treatments? Long term follow up of randomised trial of nicotine nasal spray

BMJ 1998; 316 doi: (Published 14 March 1998) Cite this as: BMJ 1998;316:830
  1. John A Stapleton, senior lecturer (j.stapleton{at},
  2. Gay Sutherland, lecturer,
  3. Michael A H Russell, professor
  1. Institute of Psychiatry, Tobacco Research Section, National Addiction Centre, London SE5 8AF
  1. Correspondence to: Dr Stapleton
  • Accepted 8 July 1997

Recent research on treatments to stop smoking has focused almost entirely on nicotine replacement, and several meta-analyses testify to the efficacy of four delivery systems.1 Although the ultimate goal of treatment is lifelong cessation, few trials have published results of abstinence beyond one year. Little consideration has therefore been given to whether the treatment is effective in reducing the major health risks of smoking. This effect would become evident only after many years of abstinence. Our randomised trial showed that the use of nicotine nasal spray compared with a placebo spray was associated with more than double the number of abstainers at one year.2 We report the results from a longer term follow up to estimate the impact of relapse after one year on effectiveness.

Subjects, methods, and results

A total of 227 heavy smokers entered the trial; 116 were given the nicotine spray and 111 the placebo. Of these, 47 sustained abstinence from smoking for 1 year. They constituted the long term follow up group; 33 were in the nicotine group, 14 in the placebo group. Criteria for long term sustained abstinence were the same as for the first year. Since the follow up was completed mainly over a 2 month period, the time interval from randomisation varied according to when the smoker entered the trial over 15 months. Standard survival methods were used to analyse the data. Survival times of those who were not contacted beyond 1 year (3 subjects in the nicotine group, 2 in the placebo group) and those who had successfully given up were censored at their last follow up. The Kaplan-Meier method was used to estimate cumulative abstinence up to 31/2 years.

Mean follow up period was 3 years 4 months (range 21/2 to 4 years) and was shorter by 21 days for the nicotine group. All observed relapses occurred within 31/2 years. The table shows that the nicotine spray maintained an advantage over placebo up to 31/2 years. Relapse after 1 year's abstinence was similar in the two groups and totalled 23% at 2 years, 38% at 3 years and 48% (95% confidence interval 32% to 64%) at 31/2 years. Although subjects had been recommended to use the spray for three months only, they were allowed to continue for 1 year. Of those remaining abstinent in the nicotine group, 19 used the spray for 1 year and 14 for <1 year (range 1-39 weeks). There was no difference in relapse after 1 year in the nicotine group between those who used the spray for 1 year and those who stopped earlier (difference 5%, 95% confidence interval −33% to 43%).

Results of long term follow up of randomised trial of nicotine nasal spray

View this table:


Our results show that the spray is an effective aid to long term smoking cessation and that those who used the spray for 1 year had a similar relapse profile to those who stopped using it earlier. They also indicate substantial relapse after the time that most studies have completed their final follow up to assess treatment efficacy. Although the success ratio of active to placebo treatment (about 2.5) was unchanged by relapse, the absolute difference was reduced considerably, and hence the estimated number needed to treat to achieve each success was increased (from 6.3 to 10.8).

Our relapse rate is similar to that in a trial of nicotine patches (37% between years 1 and 3)3 and in a study using supportive counselling and nicotine gum for 5 years (40% between years 1 and 5).4 High relapse rates after 1 year are also common in those not attending for treatment. A large general population survey estimated a relapse rate of 35% from non-validated self reports of the duration of abstinence.5

Success rates after 1 year or less of follow up substantially overestimate lifelong cessation after a single treatment episode.


Contributors: MAHR, GS, and JAS designed the original study, which GS conducted. JAS and GS designed the long term follow up phase, which GS conducted. JAS analysed the data and wrote the text of the paper. JAS and GS will act as guarantors for the paper.

Funding: Medical Research Council.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.