Letters

Current medical treatment for tuberculosis

Aspects of chemotherapy and management need clarifying

More weight was given to observational studies than randomised controlled trials

Authors' reply

Resources used for screening could be spent more usefully elsewhere

Aspects of chemotherapy and management need clarifying

EDITORWe take issue with some of the comments made by Chan and Iseman in their clinical review of current medical treatment for tuberculosis.¹ These points have already been covered in the evidence based guidelines on both the chemotherapy and management and the control and prevention of tuberculosis by the Joint Tuberculosis Committee of the British Thoracic Society, but we believe that they bear restatement. ^{2 3}

Supplementation with pyridoxine (up to 50 mg/day) as routinePeripheral neuropathy from isoniazid is very uncommon. Pyridoxine is appropriate for patients at increased risk of peripheral neuropathy.² However, pyridoxine antagonises isoniazid on a milligram for milligram basis. Studies have shown that only 6 mg/day prevents peripheral neuropathy, so only 10 mg (the smallest tablet size available) should be used, if indicated.

Routine liver function monitoringDetailed advice on liver function monitoring has been given and is incorporated in the 1998 code of practice, ^{2 4} but regular liver function measurements are not required in patients with normal pretreatment liver function, unless symptoms, in an informed patient, suggest hepatic upset.

Monthly visual acuity checks of patients taking ethambutolOcular toxicity is extremely rare in patients with normal visual acuity (Snellen chart) and renal function who are taking an ethambutol dose of 15 mg/kg, and there is no evidence that regular ocular checks prevent events.⁵ Detailed advice is again incorporated in the 1998 treatment guidelines.²

Negative pressure isolation for all suspected or confirmed cases of tuberculosis admitted to hospitalOf the 7300 patients with tuberculosis in England and Wales, some 4000 are culture confirmed; of these 1.3% have multidrug resistant tuberculosis (Scotland 0.3%, Northern Ireland 0). A detailed algorithm is given in the 2000 code of practice,³ with risk stratification for the patient, type of ward or room used, and risk of multidrug resistant tuberculosis.

Lawrence P Ormerod, consultant respiratory physician. 
Blackburn Royal Infirmary, Blackburn BB2 3LR pandpormerod{at}hotmail.com

Ian A Campbell, consultant respiratory physician. 
Llandough Hospital, Cardiff CF74 2XX

Peter D O Davies, consultant respiratory physician. 
Aintree University Hospital, Liverpool L9 7AL For the Joint Tuberculosis Committee of the British Thoracic Society

More weight was given to observational studies than randomised controlled trials

EDITORChan and Iseman review several aspects of the treatment of tuberculosis.¹ I was surprised, however, by their claim that directly observed treatment is an effective means of ensuring completion of treatment.

They even illustrate this assertion in figure 2, where they compare completion rates for patients whose treatment is directly observed and those whose treatment is not supervised. They do not say that this is based purely on observational data.² Although the review from which these figures were taken included several randomised controlled trials, none of the trials evaluated the effectiveness of directly observed treatment and non-supervised treatmentthey compared various types of directly observed treatment.

Since the review was published, randomised controlled trials of directly observed versus non-supervised treatment have been done and assessed in a systematic review.³ The conclusion is not in favour of believing in the effectiveness of directly observed treatment. This evidence must be taken seriously into consideration when making a statement on whether directly observed treatment is effective.

This seems to be yet another example of the tendency to overestimate the effectiveness of an intervention when relying on observational studies.⁴

Atle Fretheim, researcher. 
Department of Health Services Research, Norwegian Directorate for Health and Social Affairs, PO Box 8054, N-0031 Oslo, Norway atle.fretheim{at}shdir.no

Authors' reply

EDITOROrmerod et al make several important points.

Isoniazid neuritis is rare with adequate nutrition. But tuberculosis increasingly affects immigrants, alcoholics, substance misusers, and people with AIDS, in whom adequate nutrition cannot be assumed.

Excessive doses of pyridoxine can interfere with isoniazid, and 10 mg is an adequate daily dose. Unfortunately, smaller doses of pyridoxine are not consistently available, and doses up to 50 mg daily can be given safely.

Monitoring liver function during chemotherapy is not of proved utility; generally, symptomatic monitoring can detect early hepatitis and prevent fatal liver damage. Self monitoring for the groups above, in whom communication may not be reliable, is problematic. Hence biochemical monitoring remains an appropriate tool. An advantage of directly observed treatment (see below) is regular surveillance for drug toxicity.

Optic neuritis from ethambutol is rare with normal renal function and a dose of 15 mg/kg. The assertion that regular monitoring does not prevent events is, however, potentially misleading. Although neuritis may not be averted, early recognition can lessen the ultimate degree of impairment.

Algorithms are helpful but imperfect tools to identify patients with tuberculosis on admission to the hospital. Putting every suspect in negative pressure rooms is impractical and costly. However, nosocomial transmission in contemporary hospitals that house many immunocompromised patients remains a serious issue. Although perfect compliance may not be achievable, isolation of all likely candidates remains an appropriate goal.

Fretheim questions the utility of directly observed treatment, citing two studies in which such programmes did not perform well. This begs the question of what constitutes directly observed treatment. Adaptations to local conditions and adequate resources are essential to make this strategy work: it is not an automatic panacea. So long as non-adherence is the major impediment to effective treatment, directly observed treatment will remain a central strategic response.

Michael D Iseman, chief, clinical mycobacteriology service. 
isemanm{at}njc.org

Edward Chan, assistant professor of medicine. 
National Jewish Medical and Research Center, 1400 Jackson Street, Denver, CO 80206, USA

Resources used for screening could be spent more usefully elsewhere

EDITORChan and Iseman in their clinical review referred to the global resurgence of tuberculosis and the increase in case rates seen in North America and western Europe over the past decade.¹ Two causes mentioned for these increased rates are immigration and coinfection with HIV.

Both these factors apply in Croydon, south east England, where we have seen changing patterns of immigration with many refugees and asylum seekers arriving from countries with high incidences of tuberculosis. This has coincided with a 100% increase in notifications over the past 11 years, with an estimated HIV seroprevalence of 4.3% in new entrants.² Screening all new entrants from countries with high incidences of tuberculosis is recommended.³ However, no evidence shows that such screening is effective in areas with a mobile new entrant population.

We have set up a clinic for new entrants in Croydon, many of whom are not notified through the port of arrival system. This is a holistic clinic providing tuberculosis screening while also offering help with wider needs such as housing issues, finding schools, and registering with general practitioners. We have diagnosed only three cases of active tuberculosis (table). The patients had mild or no symptoms and were not infectious. Over the same period of time, some 110 cases of active tuberculosis were diagnosed in symptomatic new entrants presenting either through emergency departments or referred by general practitioners. Most were not known to the appropriate authorities and had therefore not been screened.

The considerable resources required for screening new entrants would be better spent on ensuring that all new entrants have rapid and easy access to primary care for assessment of all health needs, with a fast track referral system to local chest clinics for patients with suspected tuberculosis. Screening the children of new entrants should continue as people under 16 years benefit most from BCG and antituberculosis chemoprophylaxis.

Ruth J Whitfield, associate specialist. 
ruthwhitfield{at}btinternet.com

Rosemary Khan, senior TB specialist nurse. 
Anne Smith, senior TB specialist nurse. 
Mayday Hospital, Croydon CR7 7YE

Charlotte F J Rayner, thoracic and general physician. 
St George's Hospital, London SW17 0QT