An Incorrect Technique Guarantees Failure

0145-6008/04/2808-1273$03.00/0 ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH

Vol. 28, No. 8 August 2004

LETTERS TO THE EDITORS Long-Term Effects of and Physiological Responses to Nitrous Oxide Gas Treatment During Alcohol Withdrawal: A Double-Blind, Placebo-Controlled Trial Dear Readers: In December 2002, a paper entitled “Long-Term Effects of and Physiologic Responses to Nitrous Oxide Gas Treatment During Alcohol Withdrawal: A Double-Blind, Placebo-Controlled Trial” was published in our journal (Alho et al., 2002). What follows is a series of correspondence relating to the original article including a Letter to the Editor by Drs. Mark Gillman and Fred Lichtigfeld and two letters from experts in the field who were asked to give their opinions on this subject. Alho et al.(2002) were asked to submit a rebuttal letter in response to the Gillman/ Lichtigfield Letter to the Editor. However, the rebuttal letter by Alho et al. (2002) is not reprinted here because they already published the same letter elsewhere (Gillman and Lichtigfeld, 2004). I hope that publication of this correspondence will enlighten our readers regarding this ongoing controversy about the use of nitrous oxide to manage alcohol withdrawal. Ivan Diamond, M.D., Ph.D. Editor REFERENCES Alho H, Methuen T, Paloheimo M, Strid N, Seppa K, Tiainen J, Salaspuro M, Roine R (2002) Long-term effects of and physiological responses to nitrous oxide gas treatment during alcohol withdrawal: a double-blind, placebo-controlled trial. Alcohol Clin Exp Res 26:1816 –1822. Gillman MA, Lichtigfeld FJ (2004) Correct use of analgesic nitrous oxide for the alcohol withdrawal state is essential. J Clin Psychopharmacol 24:238-239; author reply 239-240.

An Incorrect Technique Guarantees Failure Dear Editor: We read the paper by Alho et al. (2002) and conclude that the only relationship between their work and ours is that we both used subanesthetic concentrations of nitrous oxide (N2O). In our hands, both single (Gillman and Lichtigfeld, 1990; Gillman and Lichtigfeld 1991) and doubleblind (Gillman and Lichtigfeld, 2002; Gillman and Lichtigfeld, 2004a) and those who have been trained (Ojukangas and Gillman, 1994; Ojutkangas, 1991; Daynes, 1989; Daynes and Gillman, 1994) in our specific technique, all have been able to show that analgesic N2O and oxygen mixtures are useful in alleviating acute alcoholic withdrawal states. We have highlighted the major differences between our method and that used by Alho et al. (2002) more recently, including the fact that they did not even follow our Alcohol Clin Exp Res, Vol 28, No 8, 2004: pp 1273–1276

time frame of 20 min of N2O administration but used instead a 30 min period (Gillman and Lichtigfeld, 2004b). We contend that it is the complete difference in method used (Alho et al., 2002) which could explain the discrepant findings (Gillman and Lichtigfeld, 2004b). Indeed, their method emphasized the research aspect of the study at the expense of therapeutic efficacy. Clearly, the experimental subjects could have been distracted by the research atmosphere, by the questioning and research interventions (Alho et al., 2002). This could have led to a negative therapeutic milieu, which would have further inhibited the benefits of the N2O (Gillman and Lichtigfeld, 2004b). Their method of administering N2O, instead of producing anxiolysis (as used by ourselves) would have created a stress situation, aggravating instead of alleviating the stress associated with alcohol withdrawal states. Furthermore, their cohort CIWA scores on admission were extremely low, particularly in the N2O group (Alho et al., 2002). Indeed, the scores were so low that according to those who had designed the CIWA rating scale (Sullivan et al., 1989) it was their experience that no pharmacological treatment was indicated in patients showing such low scores. Sullivan et al. (1989) proposed that supportive care alone would be sufficient. Even in those cases where scores were between 10 and 20 the necessity for drug treatment was equivocal (Sullivan et al., 1989). The addition of diazepam during the treatment period can be criticized because the use of concomitant medications in a drug study “constitutes a major artifact because it can markedly weaken the drug/placebo difference.” (Janicak et al., 1993) In the present study, the use of diazepam as a concomitant medication occurred in all groups studied (Alho et al., 2002). Indeed, once diazepam was added as a daytime sedative, the study (Alho et al., 2002) became not analgesic nitrous oxide against a placebo, but rather nitrous oxide combined with diazepam versus placebo plus diazepam. Alho et al. (2002) justified their use of an end tidal (Et) volume of 30% N2O by referring to a Cochrane Protocol entitled “Conscious sedation for dental anxiety,” in which this concentration was allegedly recommended for safe conscious sedation. After perusing the above Cochrane Review (McGoldrick et al., 2001), we could find no such reference therein. We contacted Dr. McGoldrick who referred us to Dr. A. de Jongh, one of her coauthors of this Protocol (McGoldrick et al., 2001). De Jongh agreed with us that Alho et al. (2002) were in error and had misquoted their Cochrane Review by stating: “The target of N2O treatment was a 30% Et concentration (monitored as described above) that was considered sedating and safe (McGoldrick et al., 2001)....” 1273

1274

And that the method of Alho et al. (2002) in using a target concentration of N2O of 30% Et: 1. “Differs substantially from the method routinely used in clinical dentistry where the concentration of N2O is titrated to the patient’s individual clinical need with no target concentration used. 2. Would have resulted in higher concentrations of N2O being administered to their subjects than those used in routine clinical dentistry.” He agreed on behalf of the group that the above quotation could be used as a personal communication (De Jongh, 2003). Clearly, Alho et al. (2002) to justify their choice of a 30% Et concentration of N2O misquoted this reference (McGoldrick et al., 2002), which had stated no such thing. Indeed, by using this excessive concentration of N2O, Alho et al. (2002) made it even less likely for them to reproduce our positive results. In addition, the double-blind nature of the study of Alho et al. (2002) must be questioned. In the past (Hirsch, 1983), a strict double-blind procedure referred to “the fact that everyone treating the patient is also kept ignorant of the true nature of the treatment being given to any particular patient. This is done to avoid any difference in handling the patient, or administering the treatments which could occur as a result of bias toward the treatment or assessments of the treatment’s effect.” This definition would seem to hold especially true of the N2O therapy, because its administration, unlike the administration of tablets, occurred over a 30-min period. According to their description (Alho et al., 2002), the nurse responsible for giving the gases had to ensure that the patient was not anesthetized. A fact complicated by the high doses of N2O used by Alho et al. (2002). This is because neither Alho and his colleagues or the nurse had been trained or were even aware of the correct clinical endpoint, which only becomes obvious when on-hands training has been received. Had she not been aware of the gases given and to whom they were given, she would have obviously had to have acted rather differently to each patient than she actually did, thus introducing a bias to the procedure of Alho et al. (2002). Alho et al. (2002) have shown that the doses that they used often produced increased peripheral muscle tone (Alho et al., 2002), which they attribute to increased sympathetic tone often associated with some patient arousal (Alho et al., 2002). This work confirms that of others (Ebert and Kampine, 1989), which showed increased sympathetic tone resulting in increased muscle tone. “There were striking progressive augmentations in muscle sympathetic outflow associated with increasing concentrations of nitrous oxide.” This was also associated with small elevations of nor-epinephrine content (Ebert and Kampine, 1989). These latter authors relate the increased sympathoexcitation to the effect seen in the early periods of exposure to N2O anesthesia, which can also be seen at lower preanesthetic doses of N2O. All these observations confirm that the doses used by Alho et al. (2002) to achieve the Et of

LETTERS TO THE EDITOR

N2O of 30% was higher than is used when the dose is titrated in accordance with our method of administration (Gillman and Lichtigfeld, 2004b). The fact that there was no statistically significant difference between the N2O and diazepam group and the placebo and diazepam group indicates that in their situation the N2O was as good as placebo (Alho et al., 2002). We have previously suggested “that part of the response of patients during treatment of the alcohol withdrawal state with air or oxygen could have been due to an opioid mediated placebo response” (Lichtigfeld and Gillman, 1989a). Since analgesic N2O has been shown to be an agonist at opioid receptors (Lichtigfeld and Gillman, 1989a, Lichtigfeld and Gillman, 1989b; Gillman, 1986), it could well be that the pharmacological action of N2O as “superplacebo” combines with psychological factors that could further activate the endogenous opioid system (Lichtigfeld and Gillman, 1989a; Lichtigfeld and Gillman, 1989b). For these reasons we have suggested that the therapeutic effect of N2O is mediated by a specific potentiation of the endogenous opioid system (Lichtigfeld and Gillman, 1989a; Lichtigfeld and Gillman, 1989b). Indeed, other workers have shown that negative suggestion can in fact reverse the analgesic effects of N2O (Dworkin et al., 1986), presumably by an inhibition of the endogenous opioid system. Thus any interference with positive suggestion has to be avoided to obtain the therapeutic effect found with N2O by others (Ojutkangas and Gillman, 1994; Ojutkangas, 1991; Daynes, 1989; Daynes and Gillman, 1994) and ourselves (Ojutkangas and Gillman, 1994; Ojutkangas, 1991; Daynes, 1989; Daynes and Gillman, 1994). The combination of physiologic and the psychological factors that can jointly enhance the activity of the endogenous opioid system has been emphasized in relation to placebo analgesia (Price and Soerensen, 2002). It is therefore significant that audioanalgesia only acts positively if given together with very strong suggestion (Melzack et al., 1973) and audioanalgesia is effective when combined with N2O (Clark and Brunick, 1999). Therefore we can conclude that the correct use of N2O for the alcohol withdrawal state combined with a positive therapeutic milieu is essential for the optimal effect of this intervention. Mark Gillman and Fred Lichtigfeld REFERENCES Alho H, Methuen T, Paloheimo M, Strid N, Seppa K, Tiainen J, Salaspuro M, Roine R (2002) Long-term effects of and physiological responses to nitrous oxide gas treatment during alcohol withdrawal: a double-blind, placebo-controlled trial. Alcohol Clin Exp Res 26:1816 –1822. Clark M, Brunick A (1999) Handbook of Nitrous Oxide and Oxygen Sedation. Mosby, St Louis. Daynes G (1989) The initial management of alcoholism using oxygen and nitrous oxide: a transcultural study. Int J Neurosci 49:83– 86. Daynes G, Gillman MA (1994) Psychotropic analgesic nitrous oxide prevents craving after withdrawal from alcohol, cannabis and tobacco. Int J Neurosci 76:13–16. De Jongh A (2003) Clinical characteristics of somatization in dental practice. Br Dent J 195:151–154.

1275

LETTERS TO THE EDITOR

Dworkin SF, Schubert M, Chen CAN, Clark DW (1986) Psychological preparation influences nitrous oxide analgesia: replication of laboratory findings in a clinical setting. Oral Surg Oral Pathol 6:108 –112. Ebert TJ, Kampine JP (1989) Nitrous oxide augments sympathetic outflow: direct evidence from human peroneal nerve recordings. Anesth Analg 69:444 – 449. Gillman MA (1986) Analgesic (sub anesthetic) nitrous oxide interacts with the endogenous opioid system: a review of the evidence. Life Sci 39:1209 –1222. Gillman MA, Lichtigfeld FJ (1990) Analgesic nitrous oxide for alcohol withdrawal: a critical appraisal after 10 years’ use. Postgrad Med J 66:543–546. Gillman MA, Lichtigfeld FJ (1991) Placebo and analgesic nitrous oxide for treatment of the alcohol withdrawal state. Br J Psychiatry 159:672– 675. Gillman MA, Lichtigfeld FJ (2002) Randomized double-blind trial of psychotropic analgesic nitrous oxide compared with diazepam for alcohol withdrawal state. J Subst Abuse Treat 22:129 –134. Gillman MA, Lichtigfeld FJ (2004a) Enlarged double-blind randomized trial of benzodiazepines against psychotropic analgesic nitrous oxide for alcohol withdrawal. Addict Behav 29:1183–1187. Gillman MA, Lichtigfeld FJ (2004b) Correct use of analgesic nitrous oxide for the alcohol withdrawal state is essential. J Clin Psychopharmacol 24:238 –239; author reply 239 –240. Hirsch SR (1983) The rationale of clinical trials: evaluation and planning. in The Scientific Basis of Psychiatry. Weller M. ed, pp 164 –173. Balliere Tindall, London. Janicak PG, Davis JM, Preskorn SH, Ayd FJ (Jr) (1993) Assessment of drug efficacy and relevant clinical issues, in Principles and Practice of Psychopharmacotherapy, pp 29 –35. Williams & Wilkins, Baltimore. Lichtigfeld FJ, Gillman MA (1989a) Analgesic nitrous oxide for alcohol withdrawal is better than placebo. Int J Neurosci 49:71–74. Lichtigfeld FJ, Gillman MA (1989b) The effect of placebo in the alcohol withdrawal state. Alcohol Alcohol 24:109 –112. Melzack R (1973) The puzzle of pain. Penguin Educational, Middlesex, UK. pp. 34 –39. McGoldrick P, Bannister J, de Jongh A, Durham R, Levitt J (2001) Conscious sedation for dental anxiety (Protocol). Cochrane Oral Health Group. Cochrane Database Syst Rev 2. Ojutkangas R (1991) Psychotropic analgesic nitrous oxide: rapid safe therapy for addictive withdrawal. Postgrad Med J 67:1027–1028. Ojutkangas R, Gillman MA (1994) Psychotropic analgesic nitrous oxide for treating alcohol withdrawal in an outpatient setting. Int J Neurosci 76:35–39. Price DD, Soerensen LV (2002) Endogenous opioid and non-opioid pathways as mediators of placebo analgesia, in The Science of the Placebo (Guess HA, Kleinman A, Kusik JW, Engel LW eds), pp 183– 206. BMJ Books, London. Sullivan JT, Sykora K, Schneiderman J, Naranjo CA, Sellers EM (1989) Assessment of alcohol withdrawal: the revised clinical institute withdrawal assessment for alcohol scale (CIWA-Ar). Br J Addiction 84: 1353–1357.

The Nitrous Oxide Study: Two Points (Expert Response to Gillman/Alho Letters to the Editor) Dear Editor: There are methodological and ethical difficulties in studying treatments for a potentially life-threatening condition such as alcohol withdrawal. Alho et al. point out the previous absence of true randomized blind controlled studies. It was clinically necessary that they offered diazepam when the CIWA-Ar score exceeded 10, and their ethics committee may have demanded that. Thus it is true, as Gillman and Lichtigfeld (2004) state, that the study of Alho et al. was, for a sub group of the patients, a comparison of

nitrous oxide plus diazepam versus placebo gas plus diazepam. But there is an advantage in this: the authors sensibly utilized quantity of diazepam prescribed as an outcome measure (and found that the nitrous oxide group did not require any less than the placebo group). Gillman and Lichtigfeld doubt whether the patients represented a severe group of cases, because the mean intake CIWA-Ar score for the sample was below 10. The paper made the same point in its Discussion. Some patients may have entered the study with a CIWA-Ar score at which medication might not in many centers be prescribed. However, I do not think this invalidates the study’ conclusions for this following reason. Entry to the study was possible up to a breath alcohol concentration (BAC) of 1 ␮g/ml, and patients’ mean BAC at entry was 0.4 ␮g/ml. Some of these patients might well have developed a higher CIWA-Ar score in the next few hours, as their blood alcohol level fell to zero. It is worth noting too that Gillman and Lichtigfeld’s point about the slightly lower mean CIWA-Ar scores in the nitrous oxide group should be counter-balanced by that group’s slightly higher BAC at entry (0.4 ␮g/ml, compared to 0.3 ␮g/ml in the air group). I tend to agree with Alho et al. that the case for nitrous oxide in the treatment of alcohol withdrawal remains unproven. Jonathan Chick University of Edinburgh REFERENCES Alho H, Methuen T, Paloheimo M, Strid N, Seppa K, Tiainen J, Salaspuro M, Roine R (2002) Long-term effects of and physiological responses to nitrous oxide gas treatment during alcohol withdrawal: a double-blind, placebo-controlled trial. Alcohol Clin Exp Res 26:1816 –1822. Gillman MA, Lichtigfeld FJ (2004) An incorrect technique guarantees failure. Alchol Clin Exp Res 28:127.

Expert Response To Gillman/Alho Dear Editor: I have tried to summarize the arguments presented by both sides in the debate between Alho et al. (2002), and Gillman and Lichtigfeld (2004), respectively: The alcohol withdrawal syndrome (AWS) is a condition which may sometimes progress serious and in some cases life-threatening consequences. The typical signs and symptoms represent sympathetic nervous system hyperactivity, such as tachycardia, high blood pressure, tremor, sweating, nausea, anxiety and agitation. In treatment of withdrawal it is important to reduce these symptoms as well as any subjective suffering, although the most important goal is to prevent seizures, delirium tremens, and death. Treatment of AWS using N2O has mainly been developed in South Africa by Gillman and Lichtifeld (2004), and remains controversial there (Gagiano, 1994). The paper of Alho et al. (2002) reports data on physiologic responses to N2O from a clinical trial reported else-