CORRESPONDENCE
Authors’ reply Sir—Elliot Berry’s desire to see each patient’s weight reduced to a bodymass index lower than 25 kg/m2 is the basis for so many doctors’ and patients’ despair because of constant supposed failure to achieve this goal. Slight losses in weight can lead to substantial benefits, which is why all current national guidelines, such as from the National Institutes of Health1 and WHO,2 underscore the value of even 5% and especially 10% weight losses. Berry’s second point, in which he suggests that sibutramine must have stopped working because of the weight plateau, neglects the fundamentals of energy balance. A plateau is inevitable if a person reduces energy intake by, say, 1·2 MJ per day. This reduction allows about a 10 kg weight loss before the basal metabolic rate falls and the energy cost of physical activity starts to match the reduced intake. To have continuous linear falls in weight requires a progressive and severe reduction in intake, which is not advised as a successful long-term weight-management strategy. The maintenance of a reduced bodyweight is a sign of the continued efficacy of sibutramine in limiting food intake, as shown by the rapid weight regain in patients randomly assigned placebo after 6 months. R Heine joins Berry in expressing their concerns about hypertension. We were careful to focus on this issue. 5% of patients with normal blood pressure were withdrawn because of the onset of hypertension; therefore, we proposed early monitoring of blood pressure to identify this subgroup. Blood pressure fell in patients with pre-existing hypertension on weight loss and they had no related difficulties in management. We also raised the issue of whether the overall benefit to be gained from sibutramine, given the metabolic improvements in the risk factors for cardiovascular disease and reduced insulin demands, were limited by the failure to reduce the average blood pressure. The analysis cited3 uses the classic Framingham equations that underlie all the accepted cardiovascular management strategies4 and is in keeping with the meta-analyses noted by Heine. Meta-analysis will show a decrease in the absolute risk of cardiovascular disease for sibutramine compared with placebo. The greater the risk, such as in patients with hypertension or diabetes, the greater the predicted benefit. Heine questions our emphasis on the lipid changes. These changes were
1288
unexpected, even in the sibutramineto-placebo group, and exceeded those predicted from dietary change and weight loss. The reason is unknown, but sibutramine might induce persistent enzyme changes involved in the conversion of VLDL to LDL and to HDL.5 The global epidemics of cardiovascular disease, diabetes, and obesity should be tackled by preventive measures involving changes in diet and physical activity, but pharmacological strategies, such as use of statins, hypotensives, oral hypoglycaemics and new weight management drugs, can still have a role in defined clinical disorders. *W Philip T James, Arne Astrup, Nick Finer, on behalf of the STORM group *International Obesity TaskForce, 231 North Gower Street, London NW1 2NS, UK; Research Department of Human Nutrition, Royal Veterinary and Agricultural University, Copenhagen, Denmark; and Centre for Obesity Research, Luton and Dunstable Hospital, Luton, UK 1
2
3
4
5
The practical guide: identification, evaluation, and treatment of overweight an obesity in adults, no 00-4084 Bethesda, MD: National Institutes of Health, 2000. Obesity: preventing and managing the global epidemic, no. 894. Geneva: WHO, 2000. Lauterbach KW, Evers T. The impact of sibutramine on CHD: a subgroup analysis. Int J Obes 2000; 24 (suppl 1): 99. Isles CG, ed. Consensus conference on lipid lowering to prevent vascular events. Proc R Coll Phys Ed 1999; 29 (suppl): 5. Packard CJ, Shepherd J. Lipoprotein heterogeneity and apolipoprotein B metabolism. Atheroscler Thromb Vasc Biol 1997; 17: 3542–56.
Drug-resistant tuberculosis Sir—Standardised short-course chemotherapy (SCC) regimens with first-line agents seem to be an adequate treatment for some patients with drug-resistant tuberculosis, including multidrug resistance.1 These findings raise an important question: how can treatment of drug-resistant tuberculosis with SCC be successful? The successful treatment of isolated or any rifampicin or isoniazid resistance other than multidrug-resistance can be explained on the basis that, besides other less-efficient drugs (eg, ethambutol), at least one of the two most potent components of SCC— rifampicin or isoniazid—is active under these circumstances. This interpretation is supported by the finding that a roughly linear increase in the likelihood of treatment failures was seen as the number of drugs to which the strains were resistant increased.1 However, the successful treatment of
multidrug-resistant tuberculosis with SCC is difficult to grasp, the treatment, according to outcome, was adequate for 50–60% of the multidrug-resistant cases. The heterogeneity in treatment outcome is difficult to understand, as is the success of chemotherapy with apparent drug resistance. To resolve this seeming paradox, several issues need to be addressed. In contrast to standard bacteriological techniques, susceptibility testing of Mycobacterium tuberculosis does not involve determination of minimum inhibitory concentrations; susceptibility is defined by susceptibility or resistance to mostly a single drug concentration (termed critical concentration),2 which bears no relation to achievable serum concentrations.2 Besides a high-level drug-resistance phenotype, drug resistance in M tuberculosis frequently involves low-level-resistance phenotypes, possibly because of distinct mutational target alterations or changes in cell permeability.3–5 The hypothesis can be put forward that in the case of low-level resistance in vitro, the respective drug might still be active in vivo, especially given that SCC of tuberculosis is a combination treatment including cell-wall-active agents. In essence, this hypothesis implies that the seemingly paradoxical heterogeneity in treatment outcomes of multidrug-resistant tuberculosis may be due partly to limitations in the clinical predictive value of in-vitro susceptibility testing, based on the unique but mistakenly used techniques in diagnostic mycobacteriology. This hypothesis can be proven by prospective studies that define resistance more carefully by determining minimum inhibitory concentrations (especially for isoniazid or rifampicin resistance) to investigate the impact of different resistance levels on treatment outcome. Erik C Böttger Institute of Medical Microbiology, University of Zurich, CH-8028 Zurich, Switzerland (e-mail:
[email protected]) 1
2
3
Espinal MA, Kim SI, Suarez PG, et al. Standard short-course chemotherapy for drug-resistant tuberculosis: treatment outcome in 6 countries. JAMA 2000; 283: 2537–45. Inderlied CB, Salfinger M. Antimicrobial agents and susceptibility tests: mycobacteria. In: Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, eds. Manual of clinical microbiology. Washington, DC: American Society for Microbiology Press, 1995: 1385–404. Bodmer T, Zürcher G, Imboden P, Telenti A. Mutation position and type of substitution in the -subunit of the RNA polymerase influence in-vitro activity of rifamycin and rifampicin-resistant Mycobacterium tuberculosis. J Antimicrob Chemother 1995; 35: 345–48.
THE LANCET • Vol 357 • April 21, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
CORRESPONDENCE
4
5
Meier A, Sander P, Schaper KJ, et al. Correlation of molecular resistance mechanisms and phenotypic resistance levels in streptomycin resistant Mycobacterium tuberculosis. Antimicrob Agents Chemother 1996; 40: 2452–54. Yang B, Koga H, Ohno H, et al. Relationship between antimicrobial activities of rifampicin, ribabutin and KRM-1648 and rpoB mutations of Mycobacterium tuberculosis. Antimicrob Chemother 1998; 42: 621–28.
Efficacy of atovaquoneproguanil? Sir—Birthe Høgh and colleagues’ (Dec 2, p 1888)1 randomised trial of atovaquone-proguanil for malaria prophylaxis is a welcome step for travel medicine. However the study was wholly funded by the manufacturer of atovaquone-proguanil, and 18 named employees of the formerly known Glaxo Wellcome seem to have contributed to the paper. This contribution might explain why, in our view, the results are distorted by commercial bias. The investigators conclude that atovaquone-proguanil is not more efficacious than chloroquine-proguanil, since one patient had Plasmodium ovale infection in the atovaquone-proguanil group, compared with three with P falciparum in the chloroquineproguanil group. The difference was not significant. As a proxy measure of drug efficacy, Høgh and colleagues tested serum samples before and after travel for a rise in antibodies to circumsporozoite. They calculated the minimum protective efficacy of atovaquone-proguanil as 100% against P falciparum, and that of chloroquineproguanil as 70%. However we question the sensitivity of the serological test used as an index of malaria infection and the validity of the investigators’ calculations, since only one of the three travellers with confirmed P falciparum illness had raised antibodies to circumsporozoite. Høgh and colleagues assessed the tolerability of the two regimens by comparing the overall frequency of reported adverse events 7 days after travel. 61% of atovaquone-proguanil users reported one or more adverse event, compared with 64% of chloroquine-proguanil users, which is also a non-significant difference. Høgh and colleagues grouped the adverse events by body system and filtered out reported events that were treatment emergent. There seemed to be significantly fewer gastrointestinal adverse events (but not significantly fewer neuropsychiatric adverse events) in the atovaquone-proguanil group. We believe that lists of self-reported
THE LANCET • Vol 357 • April 21, 2001
symptoms can be useful for documenting adverse-effect profiles of new drugs, but that such lists are too subjective to be used as a primary study endpoint in a tolerability trial.2 The number of drug users in each group who dropped out for any reason is a better measure of tolerability.3 In this study the number of dropouts was 49 of 511 for atovaquone-proguanil, and 43 of 511 for chloroquine-proguanil. Høgh and colleagues’ study is limited by its small sample size, and they show that atovaquone-proguanil is neither more nor less efficacious than chloroquine-proguanil in preventing malaria in non-immune travellers, and that the tolerability of the two drugs is similar. They do not mention that, for a typical 2-week visit to the tropics atovaquone-proguanil chemoprophylaxis is likely to be around four times more expensive than chloroquineproguanil. A cheaper drug than atovaquone-proguanil is doxycycline, which prevents P falciparum infection in non-immune users as effectively as mefloquine (currently the gold standard for malaria chemoprophylaxis),4 and more effectively than chloroquineproguanil.5 Doxycycline should be compared with atovaquone-proguanil in an adequately powered randomised trial. *Ashley Croft, Katie Geary, Liz Morgan Surgeon General’s Department, Ministry of Defence, Whitehall, London SW1A 2HB, UK (e-mail:
[email protected]) The views expressed in this letter are those of the authority, and do not necessarily reflect the opinion of the Ministry of Defence. 1
2
3 4
5
Høgh B, Clarke PD, Camus D, et al. Atovaquone-proguanil versus chloroquineproguanil for malaria prophylaxis in nonimmune travellers: a randomised, doubleblind study. Lancet 2000; 356: 1888–94. Croft AMJ, Garner P. Mefloquine for preventing malaria in non-immune adult travellers (Cochrane Review). In: The Cochrane Library, Issue 4. Oxford: Update Software, 2000. Greenhalgh T. Papers that report drug trials. BMJ 1997; 315: 480–83. Ohrt C, Richie TL, Widjaja H, et al. Mefloquine compared with doxycycline for the prophylaxis of malaria in Indonesian soldiers. a randomized, double-blind, placebocontrolled trial. Ann Intern Med 1997; 126: 963–72. Baudon D, Martet G, Pascal B, Bernard J, Keundjian A, Laroche R. Efficacy of daily antimalarial chemoprophylaxis in tropical Africa using either doxycycline or chloroquine-proguanil: a study conducted in 1996 in the French Army. Trans R Soc Trop Med Hyg 1999; 93: 302–03.
Authors’ reply Sir—We do not understand the allegation that our presentation of results is distorted by commercial bias. We designed the study as an equivalence (non-inferiority) trial to compare the
tolerability of atovaquone-proguanil with chloroquine-proguanil in short-term travellers. The primary endpoint (any adverse event, irrespective of relation to study drug) was discussed with European regulatory agencies before the study was undertaken. In contrast to the opinion of Ashley Croft and colleagues, these agencies agreed that the primary endpoint was appropriate to support registration of atovaquone-proguanil for malaria prophylaxis. As we describe, results for atovaquone-proguanil are equivalent to chloroquine-proguanil for the primary endpoint. The classification of adverse events by body system is standard in the premarketing and postmarketing phases of the medicines regulatory process.1 Classification of adverse events as treatment emergent, and as possibly drug related, was based on recommendations of the International Conference on Harmonization.2 Without knowledge of individual treatment assignments, investigators assessed whether there was a reasonable possibility that each adverse event was causally related to study drug, and whether they occurred before or after starting study drugs. Most adverse events were judged by the investigator to be unrelated to study-drug treatment, and many occurred while participants were not receiving study treatment. These secondary analyses provide relevant information that we believe is useful for clinicians. Although 49 participants in the atovaquone-proguanil group discontinued study treatment prematurely, 26 of these completed 7-day postexposure treatment with atovaquone-proguanil but not 28-day postexposure treatment with placebos for chloroquineproguanil; two did not travel; and one stopped study drug after a single dose of chloroquine placebo. In the chloroquine-proguanil group, three of 43 participants who discontinued study treatment prematurely did not travel. We do not agree that inclusion of these data in an analysis provides an unbiased measure of drug tolerability. We agree that assessment of a rise in antibodies to sporozites is an insensitive estimate of the number of travellers actually bitten by a malaria-infected mosquito, but we are unaware of any better method for estimating risk. Our study was not adequately powered to detect significant differences in efficacy. The 95% CI of the estimates of minimum efficacy are broad (59–100% for atovaquone-proguanil and 35–93% for chloroquineproguanil). As we discussed in the report, a study in travellers designed to show that a new antimalarial drug with
1289
For personal use. Only reproduce with permission from The Lancet Publishing Group.
