Treatment of MDR-TB

The treatment and prognosis of MDR-TB are much more akin to that for cancer than to that for infection. It has a mortality rate of up to 80%, which depends on a number of factors, including

1. How many drugs the organism is resistant to (the fewer the better),
2. How many drugs the patient is given (Patients treated with five or more drugs do better),
3. Whether an injectable drug is given or not (it should be given for the first three months at least),
4. The expertise and experience of the physician responsible,
5. How co-operative the patient is with treatment (treatment is arduous and long, and requires persistence and determination on the part of the patient),
6. Whether the patient is HIV positive or not (HIV co-infection is associated with an increased mortality).

Treatment courses are generally measured in months to years; it may require surgery, though death rates remain high despite optimal treatment. That said, good outcomes are still possible.

The treatment of MDR-TB must be undertaken by a physician experienced in the treatment of MDR-TB. Mortality and morbidity in patients treated in non-specialist centres is significantly inferior to those patients treated in specialist centres.

In addition to the obvious risks (i.e., known exposure to a patient with MDR-TB), risk factors for MDR-TB include male sex, HIV infection, previous incarceration, failed TB treatment, failure to respond to standard TB treatment, and relapse following standard TB treatment.

Treatment of MDR-TB must be done on the basis of sensitivity testing: it is impossible to treat such patients without this information. If treating a patient with suspected MDR-TB, the patient should be started on SHREZ+MXF+cycloserine pending the result of laboratory sensitivity testing.

A gene probe for rpoB is available in some countries and this serves as a useful marker for MDR-TB, because isolated RMP resistance is rare (except when patients have a history of being treated with rifampicin alone). If the results of a gene probe (rpoB) are known to be positive, then it is reasonable to omit RMP and to use SHEZ+MXF+cycloserine. The reason for maintaining the patient on INH is that INH is so potent in treating TB that it is foolish to omit it until there is microbiological proof that it is ineffective (even though isoniazid resistance so commonly occurs with rifampicin resistance).

When sensitivities are known and the isolate is confirmed as resistant to both INH and RMP, five drugs should be chosen in the following order (based on known sensitivities):

• An aminoglycoside (e.g., amikacin, kanamicin) or polypeptide antibiotic (e.g., capreomycin)
• PZA
• EMB
• a fluoroquinolones: moxifloxacin is preferred to ciprofloxacin or ofloxacin;
• rifabutin
• cycloserine
• a thioamide: prothionamide or ethionamide
• PAS
• a macrolide: e.g., clarithromycin
• linezolid
• high-dose INH (if low-level resistance)
• interferon-γ
• thioridazine

Drugs are placed nearer the top of the list because they are more effective and less toxic; drugs are placed nearer the bottom of the list because they are less effective or more toxic, or more difficult to obtain.

Resistance to one drug within a class generally means resistance to all drugs within that class, but a notable exception is rifabutin: rifampicin-resistance does not always mean rifabutin-resistance and the laboratory should be asked to test for it. It is only possible to use one drug within each drug class. If it is difficult finding five drugs to treat then the clinician can request that high level INH-resistance be looked for. If the strain has only low level INH-resistance (resistance at 1.0mg/l INH, but sensitive at 0.2mg/l INH), then high dose INH can be used as part of the regimen. When counting drugs, PZA and interferon count as zero; that is to say, when adding PZA to a four drug regimen, you must still choose another drug to make five. It is not possible to use more than one injectable (STM, capreomycin or amikacin), because the toxic effect of these drugs is additive: if possible, the aminoglycoside should be given daily for a minimum of three months (and perhaps thrice weekly thereafter). Ciprofloxacin should not be used in the treatment of tuberculosis if other fluoroquinolones are available.

There is no intermittent regimen validated for use in MDR-TB, but clinical experience is that giving injectable drugs for five days a week (because there is no-one available to give the drug at weekends) does not seem to result in inferior results. Directly observed therapy certainly helps to improve outcomes in MDR-TB and should be considered an integral part of the treatment of MDR-TB.

Response to treatment must be obtained by repeated sputum cultures (monthly if possible). Treatment for MDR-TB must be given for a minimum of 18 months and cannot be stopped until the patient has been culture-negative for a minimum of nine months. It is not unusual for patients with MDR-TB to be on treatment for two years or more.

Patients with MDR-TB should be isolated in negative-pressure rooms, if possible. Patients with MDR-TB should not be accommodated on the same ward as immunosuppressed patients (HIV infected patients, or patients on immunosuppressive drugs). Careful monitoring of compliance with treatment is crucial to the management of MDR-TB (and some physicians insist on hospitalisation if only for this reason). Some physicians will insist that these patients are isolated until their sputum is smear negative, or even culture negative (which may take many months, or even years). Keeping these patients in hospital for weeks (or months) on end may be a practical or physical impossibility and the final decision depends on the clinical judgement of the physician treating that patient. The attending physician should make full use of therapeutic drug monitoring (particularly of the aminoglycosides) both to monitor compliance and to avoid toxic effects.