International Journal of Obesity (1998) 28, 278±281 ß 1998 Stockton Press All rights reserved 0307±0565/98 $12.00 Short Communication Controlled trial of hypnotherapy for weight loss in patients with obstructive sleep apnoea J Stradling, D Roberts, A Wilson and F Lovelock Chest Unit, Churchill Hospital, Oxford, OX3 7LJ, UK OBJECTIVE: To assess if hypnotherapy assists attempts at weight loss. DESIGN: Randomised, controlled, parallel study of two forms of hypnotherapy (directed at stress reduction or energy intake reduction), vs dietary advice alone in 60 obese patients with obstructive sleep apnoea on nasal continuous positive airway pressure treatment. SETTING: National Health Service hospital in the UK. MEASURES: Weight lost at 1, 3, 6, 9, 12, 15 and 18 months after dietary advice and hypnotherapy, as a percentage of original body weight. RESULTS: All three groups lost 2±3% of their body weight at three months. At 18 months only the hypnotherapy group (with stress reduction) still showed a signi cant (P < 0.02), but small (3.8 kg), mean weight loss compared to baseline. Analysed over the whole time period the hypnotherapy group with stress reduction achieved signi cantly more weight loss than the other two treatment arms (P < 0.003), which were not signi cantly different from each other. CONCLUSIONS: This controlled trial on the use of hypnotherapy, as an adjunct to dietary advice in producing weight loss, has produced a statistically signi cant result in favour of hypnotherapy. However, the bene ts were small and clinically insigni cant. More intensive hypnotherapy might of course have been more successful, and perhaps the results of the trial are suf ciently encouraging to pursue this approach further. Keywords: hypnosis; hypnotherapy; obesity; weight loss; sleep; sleep apnoea; apnea Introduction Obstructive sleep apnoea (OSA) is a condition characterised by repeated collapse of the pharyngeal airway during sleep. In severe cases the upper airway will collapse >300 times a night, producing a similar number of brief awakenings. These awakenings grossly fragment sleep leading often to severe daytime hypersomnolence. 1 Two of the dominant risk factors for developing this condition are upper body (particularly neck) obesity, 2 and subtle changes in lower facial shape, 3 both of which narrow the upper airway and allow the normal withdrawal of pharyngeal dilator muscle tone with sleep to produce critical narrowing of the lumen. It has long been appreciated that weight loss does improve OSA 4 but, as with many other obesity related conditions, it has proved dif cult to achieve and even Correspondence: Dr J Stradling, Chest Unit, Churchill Hospital, Oxford, OX3 7LJ, UK. Received 27 August 1997; revised 20 October 1997; accepted 10 November 1997 harder to maintain. 5 The usual treatment is nasal continuous positive airway pressure treatment (NCPAP), which requires the patient to wear an uncomfortable mask over the nose every night during sleep which is supplied with air from a pump set at a pressure slightly above atmospheric (about 10 cm H 2 O). 6 Despite the unpleasantness of this therapy, the vast majority of patients do not lose weight, even though in many cases this would allow them to dispense with the equipment. Our experience has also been that out of > 900 patients on NCPAP, only a handful have lost suf cient weight to be able to return their NCPAP machines as a consequence. Whether these patients are particularly resistant to weight loss techniques, compared to other obese groups (such as type 2 diabetics) is not known. Hypnotherapy can be used as an adjunct to weight loss strategies 7,8 and we wondered whether it might help patients with OSA to reduce their body weight. We have therefore carried out a controlled study in a National Health Service (NHS) setting to establish whether hypnotherapy could produce greater short (one month) and long term weight loss (18 months) than conventional dietary advice alone.
Methods Subjects Subjects were recruited from amongst patients on NCPAP for OSA at the Oxford Sleep Unit. The patients' association newsletter carried an advertisement asking for volunteers. Sixty patients were recruited (almost all of those volunteering) if their body mass index (BMI) was more than 30 kg/m 2 and if they agreed in advance to attend for all the dietary, hypnotherapy and subsequent weighing appointments. Techniques Dietary input. All patients received dietary advice from a fully quali ed and experienced State Registered Dietitian (FL) on two occasions, supplemented by appropriate literature (for example, Enjoy Healthy Eating, Health Education Authority, 1993). Prior to the rst interview a simple questionnaire was completed by the patients about their eating habits. This rst interview (>30 min) used the results of the questionnaire to focus on how to change to `healthy eating'. This focused on developing a positive approach to food and avoiding the concept of dieting and the prohibited foods. The patients were encouraged to identify changes they thought they could make over the succeeding four weeks prior to the next interview and these were documented. The second interview a month later consisted of groups of four patients (all of whom were trial participants) sharing their aims for change and whether they had been achieved or not. The dietitian reinforced the healthy eating messages and the patients were encouraged to consider future changes they could make to their eating patterns. In addition, the bene ts of regular exercise (and its de nition) were discussed along with any speci c problems an individual patient might have. Hypnotherapy. Those participants receiving hypnotherapy did so in two sessions, a month apart, from a medical hypnotherapist (a member of the British Society of Medical and Dental Hypnosis) and also a local general practitioner (AW). Each session lasted 30 min. The rst session was alone, and the second (a month later) in a group of four patients (all of whom were trial participants). The standard induction techniques of eye xation and progressive relaxation were used. 9 Each patient received an audiotape of part of their rst hypnotherapy session, with instructions on how to use it as a self-hypnosis exercise each day at home thereafter. Letters were sent out at one and eight months to encourage continuing use of the hypnotherapy tape. Two versions of hypnotherapy were used. The rst included ego strengthening under hypnosis and Hypnotherapy and weight loss in patients with obstructive sleep apnoea J Stradling et al centred on stress reduction, the second included ego strengthening but also tried to alter attitudes to food using the Spiegel and Spiegel approach (``overeating is harmful to your health and poisons your body''). 9 Protocol The study was provisionally discussed with all participants by telephone. Those eligible, attended for their rst appointment, having been randomised into one of three parallel groups: 1) dietary advice only, given on that occasion and one month later, 2) dietary advice and hypnotherapy (with stress reduction ) given on that occasion and one month later, 3) dietary advice and hypnotherapy (with speci c suggestions about food) given on that occasion and one month later. All patients' rst appointments with the hypnotherapist and dietitian were on their own. The one month appointments were in groups of four such patients, all of whom were trial participants. Weight and height were measured at the rst appointment, and weight measured subsequently at 1, 3, 6, 9, 12, 15 and 18 months by the unit secretary. The same scales (SECA, Hamburg, Germany) were used on every occasion. The protocol was approved by the Central Oxford Research Ethics Committee. Analysis The changes in weight were analysed using percent change from baseline for each individual patient. The percent change in weight data were then averaged at each time point to produce an overall statement of percent change in weight. This normalisation to baseline approach, controls for two potential biasing effects. If there are dropouts with weights different from the mean, then an apparent drop in absolute weight at any one time point might be due for example to the loss of a particularly heavy subject. In addition, initially heavier patients may nd it easier to lose an absolute amount of weight. Thus normalisation to starting weight goes some way to control for any chance differences in weight across the three groups. Percent changes in weight at each time point for each group, compared to baseline, were assessed by paired Student t-tests. Initial differences in baseline values, and subsequent differences in percent change in weight at the seven time points, between the three groups were assessed by analysis of variance (ANOVA) with Duncan's test of signi cance (SAS suite of programs, Cary, USA). An area under the curve' analysis using ANOVA was also performed using the percent change in weight data at all time points. 279
280 Results Hypnotherapy and weight loss in patients with obstructive sleep apnoea J Stradling et al Twenty subjects entered each of the three arms of the study and had their rst treatments. Table 1 shows the baseline characteristics and gender distribution of the three groups. Random chance produced differences in the mean starting weight of the three groups, but the baseline BMI levels were not signi cantly different. This was mainly due to the lower height of the dietary advice only group, which contained a higher proportion of women. Expressing change in body weight as a percentage of original weight should reduce the likelihood that the initially heaviest group, perhaps able to achieve more weight loss, could bias the results. Figure 1 shows the percentage fall in body weight in each of the three groups up to 18 months. Although 20 patients entered each of the study arms, by the end of the study there were 15 in two arms, and 16 in the hypnotherapy with stress reduction arm, 11 men and 3 women dropped out: a drop out rate of 25%. The majority of drop outs had occurred by six months, and the three female drop outs were all in the hypnother- Table 1 Change in weight for the three experimental groups Dietary advice only Hypnotherapy with stress reduction Hypnotherapy with food advice Experimental group mean and (s.d.) number weighed number weighed number weighed gender M/F 15/5 18/2 16/4 Initial BMI (kg/m 2 ) 37.6 (5.9) 40.7 (7.5) 39.4 (6.8) Initial weight (kg) 108.3 (17.8) 20 121.9 (21.9) 20 119.1 (18.1) 20 % change in weightð1 month 70.75 (2.01) 19 71.22 (2.03) 19 70.89 (1.40) 19 % change in weightð3 months 71.92 (3.57) 20 73.07 (2.66) 17 72.30 (3.06) 18 % change in weightð6 months 71.23 (4.80) 15 72.99 (3.58) 17 71.99 (3.71) 18 % change in weightð9 months 0.41 (5.53) 15 73.15 (4.32) 15 1.89 (4.14) 16 % change in weightð12 months 70.67 (6.02) 16 72.73 (4.42) 16 71.90 (3.64) 16 % change in weightð15 months 70.76 (5.75) 15 73.97 (3.14) 13 7 0.83 (2.31) 15 % change in weightð18 months 72.00 (7.17) 15 73.33 (4.40) 16 71.23 (3.91) 15 kg lost in 18 months 3.1 (8.6) 3.8 (5.8) 1.6 (4.6) BMI ˆ body mass index. Figure 1 Percentage fall in body weight over 18 months (mean and SEM).
apy and food advice group. When each arm was analysed separately and percent weight lost assessed at each time point, only the hypnotherapy and stress reduction group maintained a signi cant weight loss compared to baseline up to 18 months (P < 0.02), although in absolute terms the weight lost was clinically insigni cant (3.8 s.d. 5.8 kg). In the dietary advice group, only at three months was there a signi cant weight loss compared to initial. No patient lost enough weight at 18 months to get off NCPAP (maximum lost 23.5 kg from a starting weight of 147 kg). ANOVA and Duncan's test was used to compare the three groups together, to assess any statistically signi cant difference between them. At no individual time period, was there a signi cant difference between the weight lost in the three groups. However, using an `area under the curve' approach, which considers the whole time period, ANOVA found that the hypnotherapy with stress reduction produced signi cantly more weight loss than the other two treatment arms (P < 0.003), which were not signi cantly different from each other. Discussion This study shows that the addition of hypnotherapy to ordinary dietary advice did make a statistically signi cant difference to the weight lost, lasting up to 18 months, although in absolute terms the loss was clinically insigni cant (mean < 4 kg). It could be argued that we had not provided a proper control for hypnotherapy and should have used sham hypnotherapy. Our hypnotherapist felt this was not possible, hence the use of two hypnotherapy groups, one devised for weight loss and the other not. Contrary to expectations, the hypnotherapy centred on changing attitudes to food, was no better than dietary advice alone and not as successful as the `control' hypnotherapy group centred on stress reduction. This has several possible interpretations, all of which are speculative. For example, it could be that stress is one of the more important factors producing excessive food consumption in this group. If hypnotherapy was a non-speci c placebo, one would not expect there to have been a signi cant difference between the two hypnotherapy arms, which there was. A more general conclusion that can be drawn is that this group of patients is clearly resistant to weight loss. In a sense they have already demonstrated this resistance by choosing not to lose weight when they know that such an action might allow them to stop using their NCPAP every night, an arduous treatment with signi cant side effects. Hypnotherapy and weight loss in patients with obstructive sleep apnoea J Stradling et al Two recent reviews on the use of hypnotherapy for weight reduction 7,8 found con icting results and lamented the absence of controlled trials and longer term data, as well as great heterogeneity in the study population. In addition they stressed that hypnotherapy was unlikely to work in isolation and should be part of a wider programme. They suggested that attempts to improve a patient's belief in their ability to control events and reduce the level of stress in their lives should also help. Obese individuals tend to be more tense and often use excessive eating to help relieve this. 8 This is the rst controlled trial of hypnotherapy vs limited dietary advice for weight loss in patients with obstructive sleep apnoea on NCPAP treatment. In this study, the level of dietetic input was essentially a control for the hypnotherapy, rather than a trial of modern approaches to `healthy eating' advice. Unfortunately one is forced to conclude that this level of input in this patient group did not produce returns commensurate with the effort and costs. We accept we may not have tried hard enough and that it is entirely possible that a more intensive course of hypnotherapy (and/or dietary advice) could have brought about a better response, but this would of course have greater resource implications. Our unit has over 900 patients on NCPAP for OSA, two thirds of whom have a BMI > 30. The level of dietary advice we employed is the most that is available to the majority of obese patients in our hospital, and indeed in most NHS hospitals, largely due to the recognition that they are a poorly responsive group to such advice. References 1 Ferguson KA, Fleetham JA. Sleep-related breathing disorders: 4 ± Consequences of sleep disordered breathing. Thorax 1995; 50: 998±1004. 2 Davies RJO, Ali NJ, Stradling JR. Neck circumference and other clinical features in the diagnosis of the obstructive sleep apnoea syndrome. Thorax 1992; 47: 101±105. 3 Maltais F, Carrier G, Cormier Y, Series F. Cephalometric measurements in snorers, non-snorers, and patients with sleep apnoea. Thorax 1991; 46: 419±423. 4 Rubinstein I, Colapinto N, Rotstein LE, Brown IG, Hoffstein V. Improvement in upper airway function after weight loss in patients with obstructive sleep apnea. Am Rev Respir Dis 1988; 138 1192±1195. 5 Wing RR, Jeffery RW. Outpatient treatments of obesity: a comparison of methodology and clinical results. Int J Obes 1979; 3: 261±279. 6 Grunstein RR. Sleep-related breathing disorders: 5ÐNasal continuous positive airway pressure treatment for obstructive sleep apnoea. Thorax 1995; 50: 1106±1113. 7 Cochrane G. Hypnosis and weight reduction: which is the cart and which is the horse. Amer J Clin Hypnosis 1992; 35: 109± 118. 8 Vanderlinden J, Vandereycken W. The (limited) possibilities of hypnotherapy in the treatment of obesity. Amer J Clin Hypn 1994; 36: 248±257. 9 Spiegel H, Spiegel D. Trance and treatment: clinical uses of hypnosis. Basic Books: New York, 1978. 281