Statistics from Altmetric.com
We read the BTS guidelines on the management of tuberculosis1 with great interest. For the most part the paper is an excellent summary of best practice and a good reference for a number of difficult situations. We were, however, less happy about the recommendation to move to a four drug regimen for most patients. We wonder if it is legitimate to generalise a practice which may be sensible in London with a significant refugee problem but which may be unnecessary in other parts of the UK.
The recommendation for a four drug regimen is graded as A (requires at least one randomised control trial). Two references are given for the statement. One, a conference report,2 lists ethnic risk factors for single and multidrug resistance drawn from the UK reference laboratory reporting service for tuberculosis. The other3is the report on the 1993 tuberculosis survey in England and Wales and draws attention inter alia to the small but rising incidence of drug resistance between 1988 and 1993. Neither is a controlled trial.
Single drug resistance has been with us from the earliest days of chemotherapy and three drug regimens have not been found wanting in the succeeding 50 years. Thus, the statement justifying a four drug regimen to counter this problem is surprising.
Multidrug resistance (to isoniazid and rifampicin) is another matter. However, data from the UK reference laboratory reporting system for tuberculosis (Mycobnet) give reason to pause. Of 93 isolates of multidrug resistant bacteria for 1997–8, 23 were also resistant to ethambutol, eight to pyrazinamide, and 15 to both. Thus, a four drug regimen for all these patients would either have been ineffective or would have led to further resistance developing in about half of the patients. Data for Leicestershire (which notifies about 200 cases each year) indicate that isolates from all seven cases of multidrug resistant tuberculosis identified in 1993–8 were resistant to at least one other drug.
In view of these results, it seems better to concentrate on obtaining bacteriological proof of resistance whenever possible, the use of more rapid methods for detecting resistance, and even withholding treatment in well non-infectious patients until sensitivity tests are available. A wholesale move to a four drug regimen will increase side effects, decrease compliance, and may not do much to counter the problem of multidrug resistance.
authors' reply Dr Cookson and colleagues question the recommendation to move to a four drug initial phase regimen for most patients in the 1998 guidelines on the management of tuberculosis.1-1 A four drug initial phase regimen has been advised in the UK since 1990, with the omission of ethambutol from the initial phase recommended to those at low risk of isoniazid resistance.1-2 The 1998 recommendations were made with the knowledge of the drug resistance rates and epidemiology from Mycobnet, both published and unpublished.
These data show that drug resistance is not confined to London but occurs in many geographical areas, with significant rates of isoniazid resistance in defined groups—that is, those with a history of prior treatment (irrespective of ethnic group), in ethnic minority groups, and in those who are HIV positive (irrespective of ethnic group).
The scientific and evidence base for short course chemotherapy was set out in detail in the recommendations1-1 in the section immediately preceding the detailed treatment recommendations commented on by Cookson et al. References 13–19 of the 1998 recommendations were controlled trials largely with a four drug initial regimen, although giving evidence that a three drug initial phase is satisfactory in those with fully sensitive organisms. The only controlled trial of six month short course chemotherapy for pulmonary disease in the UK (reference 13 of the recommendations) was with four drug initial regimens.1-3 The recommendation for an initial phase of four drugs is thus clearly based on multiple clinical trials and is therefore an A grade recommendation as stated. This recommendation is to cater for the significantly more common isoniazid resistance1-4 and is not based on current rates of multidrug resistant tuberculosis.1-4 The British Thoracic Society is not alone in recommending a four drug initial regimen for those at significant risk of isoniazid resistance. This policy is also advocated by the American Thoracic Society1-5 if the isoniazid resistance is >4% and the European Respiratory Society/International Union against Tuberculosis and Lung Diseases/World Health Organisation1-6 for most cases of tuberculosis. It is also advocated on the basis of bacteriological and controlled trial data by one of the pioneers of short course chemotherapy, Professor D Mitchison, who states: “If resistant strains are found more often, for instance in 3–10% of untreated patients, a fourth drug, usually ethambutol but sometimes streptomycin, is added”.1-7 There is also no evidence from published national audit studies or from programmatic data to support the statement by Cookson et al that a four drug regimen will “increase side effects or decrease compliance”.
The recommendations1-1 make explicit the need to obtain bacteriological confirmation and hence drug susceptibility whenever possible, and the need to be aware of rifampicin resistance and the use of molecular methods for detecting its presence. In this part, the recommendations meet the comments of Dr Cookson and colleagues.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.