Treatment

From DrugPedia: A Wikipedia for Drug discovery

Revision as of 10:54, 27 August 2008

Active tuberculosis will kill about two of every three people affected if left untreated. Treated tuberculosis has a mortality rate of less than 5% (or less in developed countries where intensive supportive measures are available).

The standard "short" course treatment for tuberculosis (TB), if it is active, is isoniazid, rifampicin, pyrazinamide, and ethambutol for two months, then isoniazid and rifampicin alone for a further four months. The patient is considered cured at six months (although there is still a relapse rate of 2 to 3%). For latent tuberculosis, the standard treatment is six to nine months of isoniazid alone.

If the organism is known to be fully sensitive, then treatment is with isoniazid, rifampicin, and pyrazinamide for two months, followed by isoniazid and rifampicin for four months. Ethambutol need not be used.

Drugs

All first-line anti-tuberculous drug names have a standard three-letter and a single-letter abbreviation:

• ethambutol is EMB or E,
• isoniazid is INH or H,
• pyrazinamide is PZA or Z,
• rifampicin is RMP or R,
• streptomycin is STM or S.

The US commonly uses abbreviations and names that are not internationally recognised: rifampicin is called rifampin and abbreviated RIF; streptomycin is commonly abbreviated SM.

Drug regimens are similarly abbreviated in a standardised manner. The drugs are listed using their single letter abbreviations (in the order given above, which is roughly the order of introduction into clinical practice). A prefix denotes the number of months the treatment should be given for; a subscript denotes intermittent dosing (so ₃ means three times a week) and no subscript means daily dosing. Most regimens have an initial high-intensity phase, followed by a continuation phase (also called a consolidation phase or eradication phase): the high-intensity phase is given first, then the continuation phase, the two phases divided by a slash.

So,

2HREZ/4HR₃

means isoniazid, rifampicin, ethambutol, pyrazinamide daily for two months, followed by four months of isoniazid and rifampicin given three times a week.

These standard abbreviations are used in the rest of this article.

There are six classes of second-line drugs (SLDs) used for the treatment of TB. A drug may be classed as second-line instead of first-line for one of two possible reasons: it may be less effective than the first-line drugs (e.g., p-aminosalicylic acid); or, it may have toxic side-effects (e.g., cycloserine); or it may be unavailable in many developing countries (e.g., fluoroquinolones):

• aminoglycosides: e.g., amikacin (AK), kanamycin;
• polypeptides: e.g., capreomycin, viomycin, enviomycin;
• fluoroquinolones: e.g., ciprofloxacin (CIP), levofloxacin, moxifloxacin (MXF);
• thioamides: e.g. ethionamide, prothionamide
• cycloserine (the only antibiotic in its class);
• p-aminosalicylic acid (PAS or P).

Other drugs that may be useful, but are not on the WHO list of SLDs:

• rifabutin
• macrolides: e.g., clarithromycin (CLR);
• linezolid (LZD);
• thioacetazone (T);
• thioridazine;
• arginine;
• vitamin D;
• R207910.

These drugs may be considered "third-line drugs" and are listed here either because they are not very effective (e.g., clarithromycin) or because their efficacy has not been proven (e.g., linezolid, R207910). Rifabutin is effective, but is not included on the WHO list because for most developing countries, it is impractically expensive.

The standard regimen

Rationale and evidence for the standard regimen

Tuberculosis has been treated with combination therapy for over fifty years. Drugs are not used singly (except in latent TB or chemoprophylaxis), and regimens that use only single drugs result in the rapid development of resistance and treatment failure.<ref name="MRC1948">Medical Research Council Streptomycin in Tuberculosis Trials Committee (1948). "Streptomycin treatment for pulmonary tuberculosis". Brit Med J ii: 769–82. </ref><ref name="Wang2006">Wang J-Y, Hsueh P-R, Jan I-S, et al. (2006). "Empirical treatment with a fluoroquinolone delays the treatment for tuberculosis and is associated with a poor prognosis in endemic areas". Thorax 61: 903–8. doi:10.1136/thx.2005.056887. PMID 16809417. </ref> The rationale for using multiple drugs to treat TB are based on simple probability. The frequency of spontaneous mutations that confer resistance to an individual drug are well known: 1 in 10⁷ for EMB, 1 in 10⁸ for STM and INH, and 1 in 10¹⁰ for RMP.<ref name="David1970">David H. L. (1970). "Probability Distribution of Drug-Resistant Mutants in Unselected Populations of Mycobacterium tuberculosis". Appl Microbiol 20 (5): 810–4. PMID 4991927. </ref> A patient with extensive pulmonary TB has approximately 10¹² bacteria in his body, and therefore will probably be harboring approximately 10⁵ EMB-resistant bacteria, 10⁴ STM-resistant bacteria, 10⁴ INH-resistant bacteria and 10² RMP-resistant bacteria. Resistance mutations appear spontaneously and independently, so the chances of him harbouring a bacterium that is spontaneously resistant to both INH and RMP is 1 in 10⁶, and the chances of him harbouring a bacterium that is spontaneously resistant to all four drugs is 1 in 10¹¹. This is, of course, an oversimplification, but it is a useful way of explaining combination therapy.

There are other theoretical reasons for supporting combination therapy. The different drugs in the regimen have different modes of action. INH are bacteriocidal against replicating bacteria. EMB is bacteriostatic at low doses, but is used in TB treatment at higher, bactericidal doses. RMP is bacteriocidal and has a sterilizing effect. PZA is only weakly bactericidal, but is very effective against bacteria located in acidic environments, inside macrophages, or in areas of acute inflammation.

All TB regimens in use were 18 months or longer until the appearance of rifampicin. In 1953, the standard UK regimen was 3SPH/15PH or 3SPH/15SH₂. Between 1965 and 1970, EMB replaced PAS. RMP began to be used to treat TB in 1968 and the BTS study in the 1970s showed that 2HRE/7HR was efficacious. In 1984, a BTS study showed that 2HRZ/4HR was efficacious,<ref name="BTS1984">British Thoracic Society (1984). "A controlled trial fo six months' chemotherapy in pulmonary tuberculosis. Final report: results during the 36 months after the end of chemotherapy and beyond". Brit J Diseases Chest 78 (4): 330–36. PMID 6386028. </ref> with a relapse rate of less than 3% after two years.<ref name="Ormerod1987">Ormerod LP, Horsfield N (1987). "Short-course antituberculous chemotherapy for pulmonary and pleural disease: five years' experience in clinical practice". Brit J Diseases Chest 81 (3): 268–71. doi:10.1016/0007-0971(87)90160-4. </ref> In 1995, with the recognition that INH resistance was increasing, the BTS recommended adding EMB or STM to the regimen: 2HREZ/4HR or 2SHRZ/4HR, which are the regimens currently recommended. The WHO also recommend a six month continuation phase of HR if the patient is still culture positive after 2 months of treatment (approximately 15% of patients with fully-sensitive TB) and for those patients who have extensive bilateral cavitation at the start of treatment.

Monitoring and DOTS

DOTS stands for "Directly Observed Therapy, Short-course" and is a major plank in the WHO global TB eradication programme. The WHO advises that all TB patients should have at least the first two months of their therapy observed (and preferably the whole of it observed): this means an independent observer watching tuberculosis patients swallow their anti-TB therapy. The independent observer is often not a healthcare worker and may be a shopkeeper or a tribal elder or similar senior person within that society. DOTS is used with intermittent dosing (thrice weekly or 2HREZ/4HR₃). Twice weekly dosing is effective<ref> "A 62-dose, 6 month therapy for pulmonary and extrapulmonary tuberculosis: A twice-weekly, directly observed, and cost-effective regimen" (1990). Ann Intern Med 112 (6): 407–415. PMID 2106816. </ref> but not recommended by the WHO, because there is no margin for error (accidentally omitting one dose per week results in once weekly dosing, which is ineffective).

Treatment with properly implemented DOTS has a success rate exceeding 95% and prevents the emergence of further multi-drug resistant strains of tuberculosis.

Some people recommend monthly surveillance until cultures convert to negative; this does not form any part of the UK or WHO recommendations for TB. If cultures are positive or symptoms do not resolve after three months of treatment, it is necessary to re-evaluate the patient for drug-resistant disease or nonadherence to drug regimen. If cultures do not convert to negative despite three months of therapy, consider initiating directly observed therapy.

Extra-pulmonary tuberculosis

Tuberculosis not affecting the lungs is called extra-pulmonary tuberculosis. Disease of the central nervous system is specifically excluded from this classification.

The UK and WHO recommendation is 2HREZ/4HR; the US recommendation is 2HREZ/7HR. There is good evidence from randomised-controlled trials to say that in tuberculous lymphadenitis<ref>Campbell IA, Ormerod LP, Friend JA, Jenkins PA, Prescott RJ. (1993). "Six months versus nine months chemotherapy for tuberculosis of lymph nodes: final results". Respir Med. 87 (8): 621–3. doi:10.1016/S0954-6111(05)80265-3. PMID 8290746. </ref> and in TB of the spine,<ref>Upadhyay SS, Saji MJ, Yau AC. (1996). "Duration of antituberculosis chemotherapy in conjunction with radical surgery in the management of spinal tuberculosis". Spine 21 (16): 1898–1903. doi:10.1097/00007632-199608150-00014. </ref><ref>Medical Research Council Working Party on tuberculosis of the spine.. "Five-year assessment of controlled trials of chort-course chemotherapy regimens of 6, 9 or 18 months' duration for spinal tuberculosis in patients ambulatory from the start or undergoing radical surgery". Int Orthopaed 23 (2): 73–81. </ref><ref>Parthasarathy R, Sriram K, Santha T, et al. (1999). "Short-course chemotherapy for tuberculosis of the spine: a comparison between ambulant treatment and radical surgery—ten-year report". J Bone Joint Surg Brit Vol 81B (3): 464–71. doi:10.1302/0301-620X.81B3.9043. PMID 10872368. </ref> the six month regimen is equivalent to the nine month regimen; the US recommendation is therefore not supported by the evidence.

Up to 25% of patients with TB of the lymph nodes (TB lymphadenitis) will get worse on treatment before they get better and this usually happens in the first few months of treatment. A few weeks after starting treatment, lymph nodes often start to enlarge, and previously solid lymph nodes may become fluctuant. This should not be interpreted as failure of therapy and is a common reason for patients (and their physicians) to panic unnecessarily. With patience, two to three months into treatment the lymph nodes start to shrink again and re-aspiration or re-biopsy of the lymph nodes is unnecessary: if repeat microbiological studies are ordered, they will show the continued presence of viable bacteria with the same sensitivity pattern, which further adds to the confusion: physicians inexperienced in the treatment of TB will then often add second-line drugs in the belief that the treatment is not working. In these situations, all that is required is re-assurance. Steroids may be useful in resolving the swelling, especially if it is painful, but they are unnecessary. Additional antibiotics are unnecessary and the treatment regimen does not need to be lengthened.

Tuberculosis of the central nervous system

Tuberculosis may affect the central nervous system (meninges, brain or spinal cord) in which case it is called TB meningitis, TB cerebritis, and TB myelitis respectively; the standard treatment is 12 months of drugs (2HREZ/10HR) and steroid are mandatory. Diagnosis is difficult as CSF culture is positive in less than half of cases, and therefore a large proportion of cases are treated on the basis of clinical suspicion alone. PCR of CSF does not significantly improve the microbiology yield; culture remains the most sensitive method and a minimum of 5 ml (preferably 20 ml) of CSF should be sent for analysis. TB cerebritis (or TB of the brain) may require brain biopsy in order to make the diagnosis, because the CSF is commonly normal: this is not always available and even when it is, some clinicians would debate whether it is justified putting a patient through such an invasive and potentially dangerous procedure when a trial of anti-TB therapy may yield the same answer; probably the only justification for brain biopsy is when drug-resistant TB is suspected. It is possible that shorter durations of therapy (e.g. six months) may be sufficient to treat TB meningitis, but no clinical trial has addressed this issue. The CSF of patients with treated TB meningitis is commonly abnormal even at 12 months;<ref name="Kent1993">Kent SJ, Crowe SM, Yung A, Lucas CR, Mijch AM. "Tuberculous Meningitis: A 30-Year Review". Clin Infect Dis 17: 987–94. PMID 8110957. </ref> the rate of resolution of the abnormality bears no correlation with clinical progress or outcome,<ref name="Teoh1986">Teoh R, O'Mahony G, Yeung VTF (1986). "Polymorphonuclear pleocytosis in the cerebrospinal fluid during chemotherapy for tuberculous meningitis". J Neurol 233 (4): 237–41. doi:10.1007/BF00314027. </ref> and is not an indication for extending or repeating treatment; repeated sampling of CSF by lumbar puncture to monitor treatment progress should therefore not be done.

Although TB meningitis and TB cerebritis are classified together, the experience of many clinicians is that their progression and response to treatment is not the same. TB meningitis usually responds well to treatment, but TB cerebritis may require prolonged treatment (up to two years) and the steroid course needed is often also prolonged (up to six months). Unlike TB meningitis, TB cerebritis often required repeated CT or MRI imaging of the brain to monitor progress.

CNS TB may be secondary to blood-borne spread: therefore some experts advocate the routine sampling of CSF in patients with miliary TB.<ref name="Chang1998">Chang AB et al (1998). "Central nervous system tuberculosis after resolution of miliary tuberculosis". Pediatr Infect Dis J 17 (6): 519–523. doi:10.1097/00006454-199806000-00019. PMID 9655548. </ref>

The anti-TB drugs that are most useful for the treatment of CNS TB are:

• INH (CSF penetration 100%)
• RMP (10–20%)
• EMB (25–50% inflamed meninges only)
• PZA (100%)
• STM (20% inflamed meninges only)
• LZD (20%)
• Cycloserine (80–100%)
• Ethionamide (100%)
• PAS (10–50%) (inflamed meninges only)

The use of steroids is routine in TB meningitis (see section below).

Steroids

The usefulness of corticosteroids (e.g., prednisolone or dexamethasone) in the treatment of TB is proven for TB meningitis and TB pericarditis. The dose for TB meningitis is dexamethasone 8 to 12mg daily tapered off over six weeks (for those who prefer more precise dosing should refer to Thwaites et al., 2004<ref name="Thwaites2004">Thwaites GE et al. (2004). "Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults". N Engl J Med 351 (17): 1741–51. doi:10.1056/NEJMoa040573. PMID 15496623. </ref>). The dose for pericarditis is prednisolone 60mg daily tapered off over four to eight weeks.

Steroids may be of temporary benefit in pleurisy, extremely advanced TB, and TB in children:

• Pleurisy: prednisolone 20 to 40mg daily tapered off over 4 to 8 weeks
• Extremely advanced TB: 40 to 60mg daily tapered off over 4 to 8 weeks
• TB in children: 2 to 5mg/kg/day for one week, 1mg/kg/day the next week, then tapered off over 5 weeks

Steroids may be of benefit in peritonitis, miliary disease, laryngeal TB, lymphadenitis and genitourinary disease, but the evidence is scant and the routine use of steroids cannot be recommended. Steroid treatment in these patients should be considered on a case by case basis by the attending physician. Thalidomide may be of benefit in TB meningitis and has been used in cases where patients have failed to respond to steroid treatment.<ref name="Roberts2003">Roberts MT, Mendelson M, Meyer P, Carmichael A, Lever AM (2003). "The use of thalidomide in the treatment of intracranial tuberculomas in adults: two case reports". J Infect 47 (3): 251–5. doi:10.1016/S0163-4453(03)00077-X. PMID 12963389. </ref>

Non-compliance

Patients who take their TB treatment in an irregular and unreliable way are at greatly increased risk of treatment failure, relapse and the development of drug-resistant TB strains.

There are variety of reasons why patients fail to take their medication. The symptoms of TB commonly resolve within a few weeks of starting TB treatment and many patients then lose motivation to continue taking their medication. Regular follow-up is important to check on compliance and to identify any problems patients are having with their medication. Patients need to be told of the importance of taking their tablets regularly, and the importance of completing treatment, because of the risk of relapse or drug-resistance developing otherwise.

One of the main complaints is the bulkiness of the tablets. The main offender is PZA (the tablets being the size of horse tablets). PZA syrup may be offered as a substitute, or if the size of the tablets is truly an issue and liquid preparations are not available, then PZA can be omitted altogether. If PZA is omitted, the patient should be warned that this results in a significant increase in the duration of treatment (details of regimens omitting PZA are given below).

The other complaint is that the medicines must be taken on an empty stomach to facilitate absorption. This can be difficult for patients to follow (for example, shift workers who take their meals at irregular times) and may mean the patient waking up an hour earlier than usual everyday just to take medication. The rules are actually less stringent than many physicians and pharmacists realise: the issue is that the absorption of RMP is reduced if taken with fat, but is unaffected by carbohydrate, protein,<ref>Purohit SD, Sarkar SK, Gupta ML, Jain DK, Gupta PR, Mehta YR. (1987). "Dietary constituents and rifampicin absorption". Tubercle 68: 151–2. doi:10.1016/0041-3879(87)90034-1. </ref> or antacids.<ref>Peloquin CA, Namdar R, Singleton MD, Nix DE. (1999). "Pharmacokinetics of rifampin under fasting conditions, with food, and with antacids". Chest 115: 12–18. doi:10.1378/chest.115.1.12. PMID 9925057. </ref> So the patient can in fact have his or her medication with food as long as the meal does not contain fat or oils (e.g., a cup of black coffee or toast with jam and no butter).<ref>Sieger DI, Bryant M, Burley DM, Citron KM. (1974). "Effect of meals on rifampicin absorption". Lancet 2: 197–8. doi:10.1016/S0140-6736(74)91487-1. </ref> Taking the medicines with food also helps ease the nausea that many patients feel when taking the medicines on an empty stomach. The effect of food on the absorption of INH is not clear: two studies have shown reduced absorption with food<ref name="Peloquin1999">Peloquin CA, Namdar R, Dodge AA, Nix DE. (1999). "Pharmacokinetics of isoniazid under fasting conditions, with food, and with antacids". Int J Tuberc Lung Dis 3 (8): 703–10. PMID 10460103. </ref><ref>Joshi MV, Saraf YS, Kshirsagar NA, Acharya VN. (1991). "Food reduces isoniazid bioavailability in normal volunteers". J Assoc Physicians India 39: 470–1. </ref> but one study showed no difference.<ref>Zent C, Smith P. (1995). "Study of the effect of concomitant food on the bioavailability of refampicin, isoniazid, and pyrazinamide". Tubercle Lung Dis 76: 109–13. doi:10.1016/0962-8479(95)90551-0. </ref> There is a small effect of food on the absorption of PZA and of EMB that is probably not clinically important.<ref>Peloquin CA, Bulpitt AE, Jaresko GS, Jelliffe RW, James GT, Nix DE. (1998). "Pharmacokinetics of pyrazinamide under fasting conditions, with food, and with antacids". Pharmacotherapy 18 (6): 1205–11. PMID 9855317. </ref><ref>Peloquin CA, Bulpitt AE, Jaresko GS, Jelliffe RW, Childs JM, Nix DE (1999). "Pharmacokinetics of ethambutol under fasting conditions, with food, and with antacids". Antimicrob Agents Chemother 43 (3): 568–72. PMID 10049268. </ref>

It is possible to test urine for isoniazid and rifampicin levels in order to check for compliance. The interpretation of urine analysis is based on the fact that isoniazid has a longer half-life than rifampicin:

• urine positive for isoniazid and rifampicin patient probably fully compliant
• urine positive for isoniazid only patient has taken his medication in the last few days preceding the clinic appointment, but had not yet taken a dose that day.
• urine positive for rifampicin only patient has omitted to take his medication the preceding few days, but did take it just before coming to clinic.
• urine negative for both isoniazid and rifampicin patient has not taken either medicine for a number of days

In countries where doctors are unable to compel patients to take their treatment (e.g., the UK), some say that urine testing only results in unhelpful confrontations with patients and does not help increase compliance. In countries where legal measures can be taken to force patients to take their medication (e.g., the US), then urine testing can be a useful adjunct in assuring compliance.

RMP colours the urine and all bodily secretions (tears, sweat, etc.) an orange-pink colour and this can be a useful proxy if urine testing is not available (although this colour fades approximately six to eight hours after each dose).

Adverse effects

For information on adverse effects of individual anti-TB drugs, please refer to the individual articles for each drug.

The relative incidence of major adverse effects has been carefully described:<ref name="Yee2003">Yee D et al. (2003). "Incidence of serious side effects from first-line antituberculosis drugs among patients treated for active tuberculosis". Am J Resp Crit Care Med 167 (11): 1472–7. doi:10.1164/rccm.200206-626OC. PMID 12569078. </ref>

• INH 0.49 per hundred patient months
• RMP 0.43
• EMB 0.07
• PZA 1.48
• All drugs 2.47

This works out to an 8.6% risk that any one patient will need to have his drug therapy changed during the course of standard short-course therapy (2HREZ/4HR). The people identified to be most at risk of major adverse side effects in this study were:

• age >60,
• females,
• HIV positive patients, and
• Asians.

It can be extremely difficult identifying which drug is responsible for which side effect, but the relative frequency of each is known.<ref name="Ormerod1996">Ormerod L. P., Horsfield N. (1996). "Frequency and type of reactions to antituberculosis drugs: observations in routine treatment" 77 (1): 37–42. PMID 8733412. </ref> The offending drugs are given in decreasing order of frequency:

• Thrombocytopaenia: RMP
• Neuropathy: INH
• Vertigo: STM
• Hepatitis: PZA, RMP, INH
• Rash: PZA, RMP, EMB

Thrombocytopaenia is only caused by RMP and no test dosing need be done. Regimens omitting RMP are discussed below. Please refer to the entry on rifampicin for further details.

The most frequent cause of neuropathy is INH. The peripheral neuropathy of INH is always a pure sensory neuropathy and finding a motor component to the peripheral neuropathy should always prompt a search for an alternative cause. Once a peripheral neuropathy has occurred, INH must be stopped and pyridoxine should be given at a dose of 50mg thrice daily. Simply adding high dose pyridoxine to the regimen once neuropathy has occurred will not stop the neuropathy from progressing. Patients at risk of peripheral neuropathy from other causes (diabetes mellitus, alcoholism, renal failure, malnutrition, pregnancy, etc.) should all be given pyridoxine] 10mg daily at the start of treatment. Please refer to the entry on isoniazid for details on other neurological side effects of INH.

Rashes are most frequently due to PZA, but can occur with any of the TB drugs. Test dosing using the same regimen as detailed below for hepatitis may be necessary to determine which drug is responsible.

Itching RMP commonly causes itching without a rash in the first two weeks of treatment: treatment should not be stopped and the patient should be advised that the itch usually resolves on its own. Short courses of sedative antihistamines such as chlorpheniramine may be useful in alleviating the itch.

Fever during treatment can be due to a number of causes. It can occur as a natural effect of tuberculosis (in which case it should resolve within three weeks of starting treatment). Fever can be a result of drug resistance (but in that case the organism must be resistant to two or more of the drugs). Fever may be due to a superadded infection or additional diagnosis (patients with TB are not exempt from getting influenza and other illnesses during the course of treatment). In a few patients, the fever is due to drug allergy. The clinician must also consider the possibility that the diagnosis of TB is wrong. If the patient has been on treatment for more than two weeks and if the fever had initially settled and then come back, it is reasonable to stop all TB medication for 72 hours. If the fever persists despite stopping all TB medication, then the fever is not due to the drugs. If the fever disappears off treatment, then the drugs need to be tested individually to determine the cause. The same scheme as is used for test dosing for drug-induced hepatitis (described below) may be used. The drug most frequently implicated as causing a drug fever is RMP: details are given in the entry on rifampicin.

Drug-induced hepatitis

The single biggest problem with TB treatment is drug-induced hepatitis, which has a mortality rate of around 5%.<ref>Forget EJ, Menzies D (2006). "Adverse reactions to first-line antituberculosis drugs". Expert Opin Drug Saf 5 (2): 231–49. doi:10.1517/14740338.5.2.231. PMID 16503745. </ref> Three drugs can induce hepatitis: PZA, INH and RMP (in decreasing order of frequency).Template:Ref<ref name="Steel1991">Steel M. A., Burk R. F., DesPrez R. M. (1991). "Toxic hepatitis with isoniazid and rifampin: a meta-analysis". Chest 99 (2): 465–471. doi:10.1378/chest.99.2.465. PMID 1824929. </ref> It is not possible to distinguish between these three causes based purely on signs and symptoms. Test dosing must be carried out to determine which drug is responsible (this is discussed in detail below).

Liver function tests (LFTs) should be checked at the start of treatment, but, if normal, need not be checked again; the patient need only be warned of the symptoms of hepatitis. Some clinicians insist on regular monitoring of LFT's while on treatment, and in this instance, tests need only be done two weeks after starting treatment and then every two months thereafter, unless any problems are detected.

Elevations in bilirubin must be expected with RMP treatment (RMP blocks bilirubin excretion) and usually resolve after 10 days (liver enzyme production increases to compensate). Isolated elevations in bilirubin can be safely ignored.

Elevations in liver transaminases (ALT and AST) are common in the first three weeks of treatment. If the patient is asymptomatic and the elevation is not excessive then no action need be taken; some experts suggest a cut-off of four times the upper limit of normal, but there is no evidence to support this particular number over and above any other number. Some experts consider that treatment should only be stopped if jaundice becomes clinically evident.

If clinically significant hepatitis occurs while on TB treatment, then all the drugs should be stopped until the liver transaminases return to normal. If the patient is so ill that TB treatment cannot be stopped, then STM and EMB should be given until the liver transaminases return to normal (these two drugs are not associated with hepatitis).

Fulminant hepatitis can occur in the course of TB treatment, but is fortunately rare; emergency liver transplantation may be necessary and deaths do occur.

Test dosing for drug-induced hepatitis

Drugs should be re-introduced individually. This cannot be done in an outpatient setting, and must be done under close observation. A nurse must be present to take patient's pulse and blood pressure at 15 minute intervals for a minimum of four hours after each test dose is given (most problems will occur within six hours of test dosing, if they are going to occur). Patients can become very suddenly unwell and access to intensive care facilities must be available. The drugs should be given in this order:

• Day 1: INH at 1/3 or 1/4 dose
• Day 2: INH at 1/2 dose
• Day 3: INH at full dose
• Day 4: RMP at 1/3 or 1/4 dose
• Day 5: RMP at 1/2 dose
• Day 6: RMP at full dose
• Day 7: EMB at 1/3 or 1/4 dose
• Day 8: EMB at 1/2 dose
• Day 9: EMB at full dose

No more than one test dose per day should be given, and all other drugs should be stopped while test dosing is being done. So on day 4, for example, the patient only receives RMP and no other drugs are given. If the patient completes the nine days of test dosing, then it is reasonable to assume that PZA has caused the hepatitis and no PZA test dosing need be done.

The reason for using the order for testing drugs is because the two most important drugs for treating TB are INH and RMP, so these are tested first: PZA is the most likely drug to cause hepatitis and is also the drug that can be most easily omitted. EMB is useful when the sensitivity pattern of the TB organism are not known and can be omitted if the organism is known to be sensitive to INH. Regimens omitting each of the standard drugs are listed below.

The order in which the drugs are tested can be varied according to the following considerations:

1. The most useful drugs (INH and RMP) should be tested first, because the absence of these drugs from a treatment regimen severely impairs its efficacy.
2. The drugs most likely to be causing the reaction should be tested as late as possible (and possibly need not be tested at all). This avoids rechallenging patients with a drug to which they have already had a (possibly) dangerous adverse reaction.

A similar scheme may be used for other adverse effects (such as fever and rash), using similar principles.

Deviations from the standard regimen

There is evidence supporting some deviations from the standard regimen when treating pulmonary TB. Sputum culture positive patients who are smear negative at the start of treatment do well with only 4 months of treatment (this has not been validated for HIV-positive patients); and sputum culture negative patients do well on only 3 months of treatment (possibly because some of these patients never had TB at all).<ref>Hong Kong Chest Service Tuberculosis Research Centre, British Medical Research Council. (1989). "A controlleed trial of 3-month, 4-month, and 6-moth regimens of chemotherapy for sputum smear-negative pulmonary tuberculosis: results at 5 years". Am Rev Respir Dis 139: 871–76. </ref> It is unwise to treat patients for only three or four months, but all TB physicians will have patients who stop their treatment early (for whatever reason), and it can be re-assuring to know that sometimes retreatment is unnecessary. Elderly patients who are already taking a large number of tablets may be offered 9HR, omitting PZA which is the bulkiest part of the regimen.

It may not always be necessary to treat with four drugs from the beginning. An example might be a close contact of a patient known to have a fully-sensitive strain of tuberculosis: in this case, it is acceptable to use 2HRZ/4HR (omitting EMB and STM) in the expectation that their strain will be INH susceptible also. Indeed, this was previously the recommended standard regimen in many countries until the early 1990s, when isoniazid-resistance rates increased.

TB involving the brain or spinal cord (meningitis, encephalitis, etc.) is currently treated with 2HREZ/10HR (12 months of treatment in total), but there is no evidence to say that this is superior to 2HREZ/4HR, it is merely that no-one has been brave enough to do the clinic trial that answers the question if the short course is equivalent.

Regimens omitting isoniazid

Isoniazid resistance in the UK accounts for approximately 6 to 7% of isolates at time of writing (25 Feb 2006). Worldwide, it is the most common type of resistance encountered, hence the current recommendation of using HREZ at the beginning of treatment until sensitivities are known. It is useful to know of current reported outbreaks (like the current outbreak of INH-resistant TB in London).

If a patient is discovered to be infected with an isoniazid-resistant strain of TB having completed 2 months of HREZ, then he should be changed to RE for a further 10 months, and the same thing if the patient is intolerant to isoniazid (although 2REZ/7RE may be acceptable if the patient is well supervised). The US recommendation is 6RZE with the option of adding a quinolone such as moxifloxacin. The level of evidence for all these regimens is poor, and there is little to recommend one over the other.

Regimens omitting rifampicin

It is rare for TB strains to be resistant to rifampicin without being resistant to isoniazid, but rifampicin intolerance is not uncommon (hepatitis or thrombocytopaenia being the most common reasons for stopping rifampicin). Of the first-line drugs, rifampicin is also the most expensive, and in the poorest countries, regimens omitting rifampicin are therefore often used. Rifampicin is the most potent sterilising drug available for the treatment of tuberculosis and all treatment regimens that omit rifampicin are significantly longer than the standard regimen.

The UK recommendation is 18HE or 12HEZ. The US recommendation is 9 to 12HEZ, with option of adding a quinolone (for example, MXF).

Regimens omitting pyrazinamide

PZA is a common cause of rash, hepatitis and of painful arthralgia in the HREZ regimen, and can be safely stopped in those patients who are intolerant to it. Isolated PZA resistance is uncommon in M. tuberculosis, but M. bovis is innately resistant to PZA. PZA is not crucial to the treatment of fully-sensitive TB, and its main value is in shortening the total treatment duration from nine months to six.

There is good evidence from UK trials that a regimen of 9HR is adequate for M. tuberculosis; this is also the first-line regimen used to treat M. bovis.

Regimens omitting ethambutol

EMB intolerance or resistance is rare. If a patient is truly intolerant or is infected with TB that is resistant to EMB, then 2HRZ/4HR is a perfectly acceptable regimen. EMB has no part to play in the treatment of TB that is sensitive to both INH and RMP, and the only reason for including it in the initial regimen is because of increasing rates of INH resistance. If INH resistance rates are known to be low, or if the infecting TB strain is known to be INH-sensitive, then there is no need to use EMB anyway.

Active tuberculosis will kill about two of every three people affected if left untreated. Treated tuberculosis has a mortality rate of less than 5% (or less in developed countries where intensive supportive measures are available).

The standard "short" course treatment for tuberculosis (TB), if it is active, is isoniazid, rifampicin, pyrazinamide, and ethambutol for two months, then isoniazid and rifampicin alone for a further four months. The patient is considered cured at six months (although there is still a relapse rate of 2 to 3%). For latent tuberculosis, the standard treatment is six to nine months of isoniazid alone.

If the organism is known to be fully sensitive, then treatment is with isoniazid, rifampicin, and pyrazinamide for two months, followed by isoniazid and rifampicin for four months. Ethambutol need not be used.

Drugs

All first-line anti-tuberculous drug names have a standard three-letter and a single-letter abbreviation:

• ethambutol is EMB or E,
• isoniazid is INH or H,
• pyrazinamide is PZA or Z,
• rifampicin is RMP or R,
• streptomycin is STM or S.

The US commonly uses abbreviations and names that are not internationally recognised: rifampicin is called rifampin and abbreviated RIF; streptomycin is commonly abbreviated SM.

Drug regimens are similarly abbreviated in a standardised manner. The drugs are listed using their single letter abbreviations (in the order given above, which is roughly the order of introduction into clinical practice). A prefix denotes the number of months the treatment should be given for; a subscript denotes intermittent dosing (so ₃ means three times a week) and no subscript means daily dosing. Most regimens have an initial high-intensity phase, followed by a continuation phase (also called a consolidation phase or eradication phase): the high-intensity phase is given first, then the continuation phase, the two phases divided by a slash.

So,

2HREZ/4HR₃

means isoniazid, rifampicin, ethambutol, pyrazinamide daily for two months, followed by four months of isoniazid and rifampicin given three times a week.

These standard abbreviations are used in the rest of this article.

There are six classes of second-line drugs (SLDs) used for the treatment of TB. A drug may be classed as second-line instead of first-line for one of two possible reasons: it may be less effective than the first-line drugs (e.g., p-aminosalicylic acid); or, it may have toxic side-effects (e.g., cycloserine); or it may be unavailable in many developing countries (e.g., fluoroquinolones):

• aminoglycosides: e.g., amikacin (AK), kanamycin;
• polypeptides: e.g., capreomycin, viomycin, enviomycin;
• fluoroquinolones: e.g., ciprofloxacin (CIP), levofloxacin, moxifloxacin (MXF);
• thioamides: e.g. ethionamide, prothionamide
• cycloserine (the only antibiotic in its class);
• p-aminosalicylic acid (PAS or P).

Other drugs that may be useful, but are not on the WHO list of SLDs:

• rifabutin
• macrolides: e.g., clarithromycin (CLR);
• linezolid (LZD);
• thioacetazone (T);
• thioridazine;
• arginine;
• vitamin D;
• R207910.

These drugs may be considered "third-line drugs" and are listed here either because they are not very effective (e.g., clarithromycin) or because their efficacy has not been proven (e.g., linezolid, R207910). Rifabutin is effective, but is not included on the WHO list because for most developing countries, it is impractically expensive.

The standard regimen

Rationale and evidence for the standard regimen

Tuberculosis has been treated with combination therapy for over fifty years. Drugs are not used singly (except in latent TB or chemoprophylaxis), and regimens that use only single drugs result in the rapid development of resistance and treatment failure.<ref name="MRC1948">Medical Research Council Streptomycin in Tuberculosis Trials Committee (1948). "Streptomycin treatment for pulmonary tuberculosis". Brit Med J ii: 769–82. </ref><ref name="Wang2006">Wang J-Y, Hsueh P-R, Jan I-S, et al. (2006). "Empirical treatment with a fluoroquinolone delays the treatment for tuberculosis and is associated with a poor prognosis in endemic areas". Thorax 61: 903–8. doi:10.1136/thx.2005.056887. PMID 16809417. </ref> The rationale for using multiple drugs to treat TB are based on simple probability. The frequency of spontaneous mutations that confer resistance to an individual drug are well known: 1 in 10⁷ for EMB, 1 in 10⁸ for STM and INH, and 1 in 10¹⁰ for RMP.<ref name="David1970">David H. L. (1970). "Probability Distribution of Drug-Resistant Mutants in Unselected Populations of Mycobacterium tuberculosis". Appl Microbiol 20 (5): 810–4. PMID 4991927. </ref>

A patient with extensive pulmonary TB has approximately 10¹² bacteria in his body, and therefore will probably be harboring approximately 10⁵ EMB-resistant bacteria, 10⁴ STM-resistant bacteria, 10⁴ INH-resistant bacteria and 10² RMP-resistant bacteria. Resistance mutations appear spontaneously and independently, so the chances of him harbouring a bacterium that is spontaneously resistant to both INH and RMP is 1 in 10⁶, and the chances of him harbouring a bacterium that is spontaneously resistant to all four drugs is 1 in 10¹¹. This is, of course, an oversimplification, but it is a useful way of explaining combination therapy.

There are other theoretical reasons for supporting combination therapy. The different drugs in the regimen have different modes of action. INH are bacteriocidal against replicating bacteria. EMB is bacteriostatic at low doses, but is used in TB treatment at higher, bactericidal doses. RMP is bacteriocidal and has a sterilizing effect. PZA is only weakly bactericidal, but is very effective against bacteria located in acidic environments, inside macrophages, or in areas of acute inflammation.

All TB regimens in use were 18 months or longer until the appearance of rifampicin. In 1953, the standard UK regimen was 3SPH/15PH or 3SPH/15SH₂. Between 1965 and 1970, EMB replaced PAS. RMP began to be used to treat TB in 1968 and the BTS study in the 1970s showed that 2HRE/7HR was efficacious. In 1984, a BTS study showed that 2HRZ/4HR was efficacious,<ref name="BTS1984">British Thoracic Society (1984). "A controlled trial fo six months' chemotherapy in pulmonary tuberculosis. Final report: results during the 36 months after the end of chemotherapy and beyond". Brit J Diseases Chest 78 (4): 330–36. PMID 6386028. </ref> with a relapse rate of less than 3% after two years.<ref name="Ormerod1987">Ormerod LP, Horsfield N (1987). "Short-course antituberculous chemotherapy for pulmonary and pleural disease: five years' experience in clinical practice". Brit J Diseases Chest 81 (3): 268–71. doi:10.1016/0007-0971(87)90160-4. </ref> In 1995, with the recognition that INH resistance was increasing, the BTS recommended adding EMB or STM to the regimen: 2HREZ/4HR or 2SHRZ/4HR, which are the regimens currently recommended. The WHO also recommend a six month continuation phase of HR if the patient is still culture positive after 2 months of treatment (approximately 15% of patients with fully-sensitive TB) and for those patients who have extensive bilateral cavitation at the start of treatment.

Monitoring and DOTS

DOTS stands for "Directly Observed Therapy, Short-course" and is a major plank in the WHO global TB eradication programme. The WHO advises that all TB patients should have at least the first two months of their therapy observed (and preferably the whole of it observed): this means an independent observer watching tuberculosis patients swallow their anti-TB therapy. The independent observer is often not a healthcare worker and may be a shopkeeper or a tribal elder or similar senior person within that society. DOTS is used with intermittent dosing (thrice weekly or 2HREZ/4HR₃). Twice weekly dosing is effective<ref> "A 62-dose, 6 month therapy for pulmonary and extrapulmonary tuberculosis: A twice-weekly, directly observed, and cost-effective regimen" (1990). Ann Intern Med 112 (6): 407–415. PMID 2106816. </ref> but not recommended by the WHO, because there is no margin for error (accidentally omitting one dose per week results in once weekly dosing, which is ineffective).

Treatment with properly implemented DOTS has a success rate exceeding 95% and prevents the emergence of further multi-drug resistant strains of tuberculosis.

Some people recommend monthly surveillance until cultures convert to negative; this does not form any part of the UK or WHO recommendations for TB. If cultures are positive or symptoms do not resolve after three months of treatment, it is necessary to re-evaluate the patient for drug-resistant disease or nonadherence to drug regimen. If cultures do not convert to negative despite three months of therapy, consider initiating directly observed therapy.

Extra-pulmonary tuberculosis

Tuberculosis not affecting the lungs is called extra-pulmonary tuberculosis. Disease of the central nervous system is specifically excluded from this classification.

The UK and WHO recommendation is 2HREZ/4HR; the US recommendation is 2HREZ/7HR. There is good evidence from randomised-controlled trials to say that in tuberculous lymphadenitis<ref>Campbell IA, Ormerod LP, Friend JA, Jenkins PA, Prescott RJ. (1993). "Six months versus nine months chemotherapy for tuberculosis of lymph nodes: final results". Respir Med. 87 (8): 621–3. doi:10.1016/S0954-6111(05)80265-3. PMID 8290746. </ref> and in TB of the spine,<ref>Upadhyay SS, Saji MJ, Yau AC. (1996). "Duration of antituberculosis chemotherapy in conjunction with radical surgery in the management of spinal tuberculosis". Spine 21 (16): 1898–1903. doi:10.1097/00007632-199608150-00014. </ref><ref>Medical Research Council Working Party on tuberculosis of the spine.. "Five-year assessment of controlled trials of chort-course chemotherapy regimens of 6, 9 or 18 months' duration for spinal tuberculosis in patients ambulatory from the start or undergoing radical surgery". Int Orthopaed 23 (2): 73–81. </ref><ref>Parthasarathy R, Sriram K, Santha T, et al. (1999). "Short-course chemotherapy for tuberculosis of the spine: a comparison between ambulant treatment and radical surgery—ten-year report". J Bone Joint Surg Brit Vol 81B (3): 464–71. doi:10.1302/0301-620X.81B3.9043. PMID 10872368. </ref> the six month regimen is equivalent to the nine month regimen; the US recommendation is therefore not supported by the evidence.

Up to 25% of patients with TB of the lymph nodes (TB lymphadenitis) will get worse on treatment before they get better and this usually happens in the first few months of treatment. A few weeks after starting treatment, lymph nodes often start to enlarge, and previously solid lymph nodes may become fluctuant. This should not be interpreted as failure of therapy and is a common reason for patients (and their physicians) to panic unnecessarily. With patience, two to three months into treatment the lymph nodes start to shrink again and re-aspiration or re-biopsy of the lymph nodes is unnecessary: if repeat microbiological studies are ordered, they will show the continued presence of viable bacteria with the same sensitivity pattern, which further adds to the confusion: physicians inexperienced in the treatment of TB will then often add second-line drugs in the belief that the treatment is not working. In these situations, all that is required is re-assurance. Steroids may be useful in resolving the swelling, especially if it is painful, but they are unnecessary. Additional antibiotics are unnecessary and the treatment regimen does not need to be lengthened.

Tuberculosis of the central nervous system

Tuberculosis may affect the central nervous system (meninges, brain or spinal cord) in which case it is called TB meningitis, TB cerebritis, and TB myelitis respectively; the standard treatment is 12 months of drugs (2HREZ/10HR) and steroid are mandatory. Diagnosis is difficult as CSF culture is positive in less than half of cases, and therefore a large proportion of cases are treated on the basis of clinical suspicion alone. PCR of CSF does not significantly improve the microbiology yield; culture remains the most sensitive method and a minimum of 5 ml (preferably 20 ml) of CSF should be sent for analysis. TB cerebritis (or TB of the brain) may require brain biopsy in order to make the diagnosis, because the CSF is commonly normal: this is not always available and even when it is, some clinicians would debate whether it is justified putting a patient through such an invasive and potentially dangerous procedure when a trial of anti-TB therapy may yield the same answer; probably the only justification for brain biopsy is when drug-resistant TB is suspected. It is possible that shorter durations of therapy (e.g. six months) may be sufficient to treat TB meningitis, but no clinical trial has addressed this issue. The CSF of patients with treated TB meningitis is commonly abnormal even at 12 months;<ref name="Kent1993">Kent SJ, Crowe SM, Yung A, Lucas CR, Mijch AM. "Tuberculous Meningitis: A 30-Year Review". Clin Infect Dis 17: 987–94. PMID 8110957. </ref> the rate of resolution of the abnormality bears no correlation with clinical progress or outcome,<ref name="Teoh1986">Teoh R, O'Mahony G, Yeung VTF (1986). "Polymorphonuclear pleocytosis in the cerebrospinal fluid during chemotherapy for tuberculous meningitis". J Neurol 233 (4): 237–41. doi:10.1007/BF00314027. </ref> and is not an indication for extending or repeating treatment; repeated sampling of CSF by lumbar puncture to monitor treatment progress should therefore not be done.

Although TB meningitis and TB cerebritis are classified together, the experience of many clinicians is that their progression and response to treatment is not the same. TB meningitis usually responds well to treatment, but TB cerebritis may require prolonged treatment (up to two years) and the steroid course needed is often also prolonged (up to six months). Unlike TB meningitis, TB cerebritis often required repeated CT or MRI imaging of the brain to monitor progress.

CNS TB may be secondary to blood-borne spread: therefore some experts advocate the routine sampling of CSF in patients with miliary TB.<ref name="Chang1998">Chang AB et al (1998). "Central nervous system tuberculosis after resolution of miliary tuberculosis". Pediatr Infect Dis J 17 (6): 519–523. doi:10.1097/00006454-199806000-00019. PMID 9655548. </ref>

The anti-TB drugs that are most useful for the treatment of CNS TB are:

• INH (CSF penetration 100%)
• RMP (10–20%)
• EMB (25–50% inflamed meninges only)
• PZA (100%)
• STM (20% inflamed meninges only)
• LZD (20%)
• Cycloserine (80–100%)
• Ethionamide (100%)
• PAS (10–50%) (inflamed meninges only)

The use of steroids is routine in TB meningitis (see section below).

Steroids

The usefulness of corticosteroids (e.g., prednisolone or dexamethasone) in the treatment of TB is proven for TB meningitis and TB pericarditis. The dose for TB meningitis is dexamethasone 8 to 12mg daily tapered off over six weeks (for those who prefer more precise dosing should refer to Thwaites et al., 2004<ref name="Thwaites2004">Thwaites GE et al. (2004). "Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults". N Engl J Med 351 (17): 1741–51. doi:10.1056/NEJMoa040573. PMID 15496623. </ref>). The dose for pericarditis is prednisolone 60mg daily tapered off over four to eight weeks.

Steroids may be of temporary benefit in pleurisy, extremely advanced TB, and TB in children:

• Pleurisy: prednisolone 20 to 40mg daily tapered off over 4 to 8 weeks
• Extremely advanced TB: 40 to 60mg daily tapered off over 4 to 8 weeks
• TB in children: 2 to 5mg/kg/day for one week, 1mg/kg/day the next week, then tapered off over 5 weeks

Steroids may be of benefit in peritonitis, miliary disease, laryngeal TB, lymphadenitis and genitourinary disease, but the evidence is scant and the routine use of steroids cannot be recommended. Steroid treatment in these patients should be considered on a case by case basis by the attending physician.

Thalidomide may be of benefit in TB meningitis and has been used in cases where patients have failed to respond to steroid treatment.<ref name="Roberts2003">Roberts MT, Mendelson M, Meyer P, Carmichael A, Lever AM (2003). "The use of thalidomide in the treatment of intracranial tuberculomas in adults: two case reports". J Infect 47 (3): 251–5. doi:10.1016/S0163-4453(03)00077-X. PMID 12963389. </ref>

Non-compliance

Patients who take their TB treatment in an irregular and unreliable way are at greatly increased risk of treatment failure, relapse and the development of drug-resistant TB strains.

There are variety of reasons why patients fail to take their medication. The symptoms of TB commonly resolve within a few weeks of starting TB treatment and many patients then lose motivation to continue taking their medication. Regular follow-up is important to check on compliance and to identify any problems patients are having with their medication. Patients need to be told of the importance of taking their tablets regularly, and the importance of completing treatment, because of the risk of relapse or drug-resistance developing otherwise.

One of the main complaints is the bulkiness of the tablets. The main offender is PZA (the tablets being the size of horse tablets). PZA syrup may be offered as a substitute, or if the size of the tablets is truly an issue and liquid preparations are not available, then PZA can be omitted altogether. If PZA is omitted, the patient should be warned that this results in a significant increase in the duration of treatment (details of regimens omitting PZA are given below).

The other complaint is that the medicines must be taken on an empty stomach to facilitate absorption. This can be difficult for patients to follow (for example, shift workers who take their meals at irregular times) and may mean the patient waking up an hour earlier than usual everyday just to take medication. The rules are actually less stringent than many physicians and pharmacists realise: the issue is that the absorption of RMP is reduced if taken with fat, but is unaffected by carbohydrate, protein,<ref>Purohit SD, Sarkar SK, Gupta ML, Jain DK, Gupta PR, Mehta YR. (1987). "Dietary constituents and rifampicin absorption". Tubercle 68: 151–2. doi:10.1016/0041-3879(87)90034-1. </ref> or antacids.<ref>Peloquin CA, Namdar R, Singleton MD, Nix DE. (1999). "Pharmacokinetics of rifampin under fasting conditions, with food, and with antacids". Chest 115: 12–18. doi:10.1378/chest.115.1.12. PMID 9925057. </ref> So the patient can in fact have his or her medication with food as long as the meal does not contain fat or oils (e.g., a cup of black coffee or toast with jam and no butter).<ref>Sieger DI, Bryant M, Burley DM, Citron KM. (1974). "Effect of meals on rifampicin absorption". Lancet 2: 197–8. doi:10.1016/S0140-6736(74)91487-1. </ref> Taking the medicines with food also helps ease the nausea that many patients feel when taking the medicines on an empty stomach. The effect of food on the absorption of INH is not clear: two studies have shown reduced absorption with food<ref name="Peloquin1999">Peloquin CA, Namdar R, Dodge AA, Nix DE. (1999). "Pharmacokinetics of isoniazid under fasting conditions, with food, and with antacids". Int J Tuberc Lung Dis 3 (8): 703–10. PMID 10460103. </ref><ref>Joshi MV, Saraf YS, Kshirsagar NA, Acharya VN. (1991). "Food reduces isoniazid bioavailability in normal volunteers". J Assoc Physicians India 39: 470–1. </ref> but one study showed no difference.<ref>Zent C, Smith P. (1995). "Study of the effect of concomitant food on the bioavailability of refampicin, isoniazid, and pyrazinamide". Tubercle Lung Dis 76: 109–13. doi:10.1016/0962-8479(95)90551-0. </ref> There is a small effect of food on the absorption of PZA and of EMB that is probably not clinically important.<ref>Peloquin CA, Bulpitt AE, Jaresko GS, Jelliffe RW, James GT, Nix DE. (1998). "Pharmacokinetics of pyrazinamide under fasting conditions, with food, and with antacids". Pharmacotherapy 18 (6): 1205–11. PMID 9855317. </ref><ref>Peloquin CA, Bulpitt AE, Jaresko GS, Jelliffe RW, Childs JM, Nix DE (1999). "Pharmacokinetics of ethambutol under fasting conditions, with food, and with antacids". Antimicrob Agents Chemother 43 (3): 568–72. PMID 10049268. </ref>

It is possible to test urine for isoniazid and rifampicin levels in order to check for compliance. The interpretation of urine analysis is based on the fact that isoniazid has a longer half-life than rifampicin:

• urine positive for isoniazid and rifampicin patient probably fully compliant
• urine positive for isoniazid only patient has taken his medication in the last few days preceding the clinic appointment, but had not yet taken a dose that day.
• urine positive for rifampicin only patient has omitted to take his medication the preceding few days, but did take it just before coming to clinic.
• urine negative for both isoniazid and rifampicin patient has not taken either medicine for a number of days

In countries where doctors are unable to compel patients to take their treatment (e.g., the UK), some say that urine testing only results in unhelpful confrontations with patients and does not help increase compliance. In countries where legal measures can be taken to force patients to take their medication (e.g., the US), then urine testing can be a useful adjunct in assuring compliance.

RMP colours the urine and all bodily secretions (tears, sweat, etc.) an orange-pink colour and this can be a useful proxy if urine testing is not available (although this colour fades approximately six to eight hours after each dose).

Adverse effects

For information on adverse effects of individual anti-TB drugs, please refer to the individual articles for each drug.

The relative incidence of major adverse effects has been carefully described:<ref name="Yee2003">Yee D et al. (2003). "Incidence of serious side effects from first-line antituberculosis drugs among patients treated for active tuberculosis". Am J Resp Crit Care Med 167 (11): 1472–7. doi:10.1164/rccm.200206-626OC. PMID 12569078. </ref>

• INH 0.49 per hundred patient months
• RMP 0.43
• EMB 0.07
• PZA 1.48
• All drugs 2.47

This works out to an 8.6% risk that any one patient will need to have his drug therapy changed during the course of standard short-course therapy (2HREZ/4HR). The people identified to be most at risk of major adverse side effects in this study were:

• age >60,
• females,
• HIV positive patients, and
• Asians.

It can be extremely difficult identifying which drug is responsible for which side effect, but the relative frequency of each is known.<ref name="Ormerod1996">Ormerod L. P., Horsfield N. (1996). "Frequency and type of reactions to antituberculosis drugs: observations in routine treatment" 77 (1): 37–42. PMID 8733412. </ref> The offending drugs are given in decreasing order of frequency:

• Thrombocytopaenia: RMP
• Neuropathy: INH
• Vertigo: STM
• Hepatitis: PZA, RMP, INH
• Rash: PZA, RMP, EMB

Thrombocytopaenia is only caused by RMP and no test dosing need be done. Regimens omitting RMP are discussed below. Please refer to the entry on rifampicin for further details.

The most frequent cause of neuropathy is INH. The peripheral neuropathy of INH is always a pure sensory neuropathy and finding a motor component to the peripheral neuropathy should always prompt a search for an alternative cause. Once a peripheral neuropathy has occurred, INH must be stopped and pyridoxine should be given at a dose of 50mg thrice daily. Simply adding high dose pyridoxine to the regimen once neuropathy has occurred will not stop the neuropathy from progressing. Patients at risk of peripheral neuropathy from other causes (diabetes mellitus, alcoholism, renal failure, malnutrition, pregnancy, etc.) should all be given pyridoxine] 10mg daily at the start of treatment. Please refer to the entry on isoniazid for details on other neurological side effects of INH.

Rashes are most frequently due to PZA, but can occur with any of the TB drugs. Test dosing using the same regimen as detailed below for hepatitis may be necessary to determine which drug is responsible.

Itching RMP commonly causes itching without a rash in the first two weeks of treatment: treatment should not be stopped and the patient should be advised that the itch usually resolves on its own. Short courses of sedative antihistamines such as chlorpheniramine may be useful in alleviating the itch.

Fever during treatment can be due to a number of causes. It can occur as a natural effect of tuberculosis (in which case it should resolve within three weeks of starting treatment). Fever can be a result of drug resistance (but in that case the organism must be resistant to two or more of the drugs). Fever may be due to a superadded infection or additional diagnosis (patients with TB are not exempt from getting influenza and other illnesses during the course of treatment). In a few patients, the fever is due to drug allergy. The clinician must also consider the possibility that the diagnosis of TB is wrong. If the patient has been on treatment for more than two weeks and if the fever had initially settled and then come back, it is reasonable to stop all TB medication for 72 hours. If the fever persists despite stopping all TB medication, then the fever is not due to the drugs. If the fever disappears off treatment, then the drugs need to be tested individually to determine the cause. The same scheme as is used for test dosing for drug-induced hepatitis (described below) may be used. The drug most frequently implicated as causing a drug fever is RMP: details are given in the entry on rifampicin.

Drug-induced hepatitis

The single biggest problem with TB treatment is drug-induced hepatitis, which has a mortality rate of around 5%.<ref>Forget EJ, Menzies D (2006). "Adverse reactions to first-line antituberculosis drugs". Expert Opin Drug Saf 5 (2): 231–49. doi:10.1517/14740338.5.2.231. PMID 16503745. </ref> Three drugs can induce hepatitis: PZA, INH and RMP (in decreasing order of frequency).Template:Ref<ref name="Steel1991">Steel M. A., Burk R. F., DesPrez R. M. (1991). "Toxic hepatitis with isoniazid and rifampin: a meta-analysis". Chest 99 (2): 465–471. doi:10.1378/chest.99.2.465. PMID 1824929. </ref> It is not possible to distinguish between these three causes based purely on signs and symptoms. Test dosing must be carried out to determine which drug is responsible (this is discussed in detail below).

Liver function tests (LFTs) should be checked at the start of treatment, but, if normal, need not be checked again; the patient need only be warned of the symptoms of hepatitis. Some clinicians insist on regular monitoring of LFT's while on treatment, and in this instance, tests need only be done two weeks after starting treatment and then every two months thereafter, unless any problems are detected.

Elevations in bilirubin must be expected with RMP treatment (RMP blocks bilirubin excretion) and usually resolve after 10 days (liver enzyme production increases to compensate). Isolated elevations in bilirubin can be safely ignored.

Elevations in liver transaminases (ALT and AST) are common in the first three weeks of treatment. If the patient is asymptomatic and the elevation is not excessive then no action need be taken; some experts suggest a cut-off of four times the upper limit of normal, but there is no evidence to support this particular number over and above any other number. Some experts consider that treatment should only be stopped if jaundice becomes clinically evident.

If clinically significant hepatitis occurs while on TB treatment, then all the drugs should be stopped until the liver transaminases return to normal. If the patient is so ill that TB treatment cannot be stopped, then STM and EMB should be given until the liver transaminases return to normal (these two drugs are not associated with hepatitis).

Fulminant hepatitis can occur in the course of TB treatment, but is fortunately rare; emergency liver transplantation may be necessary and deaths do occur.

Test dosing for drug-induced hepatitis

Drugs should be re-introduced individually. This cannot be done in an outpatient setting, and must be done under close observation. A nurse must be present to take patient's pulse and blood pressure at 15 minute intervals for a minimum of four hours after each test dose is given (most problems will occur within six hours of test dosing, if they are going to occur). Patients can become very suddenly unwell and access to intensive care facilities must be available. The drugs should be given in this order:

• Day 1: INH at 1/3 or 1/4 dose
• Day 2: INH at 1/2 dose
• Day 3: INH at full dose
• Day 4: RMP at 1/3 or 1/4 dose
• Day 5: RMP at 1/2 dose
• Day 6: RMP at full dose
• Day 7: EMB at 1/3 or 1/4 dose
• Day 8: EMB at 1/2 dose
• Day 9: EMB at full dose

No more than one test dose per day should be given, and all other drugs should be stopped while test dosing is being done. So on day 4, for example, the patient only receives RMP and no other drugs are given. If the patient completes the nine days of test dosing, then it is reasonable to assume that PZA has caused the hepatitis and no PZA test dosing need be done.

The reason for using the order for testing drugs is because the two most important drugs for treating TB are INH and RMP, so these are tested first: PZA is the most likely drug to cause hepatitis and is also the drug that can be most easily omitted. EMB is useful when the sensitivity pattern of the TB organism are not known and can be omitted if the organism is known to be sensitive to INH. Regimens omitting each of the standard drugs are listed below.

The order in which the drugs are tested can be varied according to the following considerations:

1. The most useful drugs (INH and RMP) should be tested first, because the absence of these drugs from a treatment regimen severely impairs its efficacy.
2. The drugs most likely to be causing the reaction should be tested as late as possible (and possibly need not be tested at all). This avoids rechallenging patients with a drug to which they have already had a (possibly) dangerous adverse reaction.

A similar scheme may be used for other adverse effects (such as fever and rash), using similar principles.

Deviations from the standard regimen

There is evidence supporting some deviations from the standard regimen when treating pulmonary TB. Sputum culture positive patients who are smear negative at the start of treatment do well with only 4 months of treatment (this has not been validated for HIV-positive patients); and sputum culture negative patients do well on only 3 months of treatment (possibly because some of these patients never had TB at all).<ref>Hong Kong Chest Service Tuberculosis Research Centre, British Medical Research Council. (1989). "A controlleed trial of 3-month, 4-month, and 6-moth regimens of chemotherapy for sputum smear-negative pulmonary tuberculosis: results at 5 years". Am Rev Respir Dis 139: 871–76. </ref> It is unwise to treat patients for only three or four months, but all TB physicians will have patients who stop their treatment early (for whatever reason), and it can be re-assuring to know that sometimes retreatment is unnecessary. Elderly patients who are already taking a large number of tablets may be offered 9HR, omitting PZA which is the bulkiest part of the regimen.

It may not always be necessary to treat with four drugs from the beginning. An example might be a close contact of a patient known to have a fully-sensitive strain of tuberculosis: in this case, it is acceptable to use 2HRZ/4HR (omitting EMB and STM) in the expectation that their strain will be INH susceptible also. Indeed, this was previously the recommended standard regimen in many countries until the early 1990s, when isoniazid-resistance rates increased.

TB involving the brain or spinal cord (meningitis, encephalitis, etc.) is currently treated with 2HREZ/10HR (12 months of treatment in total), but there is no evidence to say that this is superior to 2HREZ/4HR, it is merely that no-one has been brave enough to do the clinic trial that answers the question if the short course is equivalent.

Regimens omitting isoniazid

Isoniazid resistance in the UK accounts for approximately 6 to 7% of isolates at time of writing (25 Feb 2006). Worldwide, it is the most common type of resistance encountered, hence the current recommendation of using HREZ at the beginning of treatment until sensitivities are known. It is useful to know of current reported outbreaks (like the current outbreak of INH-resistant TB in London).

If a patient is discovered to be infected with an isoniazid-resistant strain of TB having completed 2 months of HREZ, then he should be changed to RE for a further 10 months, and the same thing if the patient is intolerant to isoniazid (although 2REZ/7RE may be acceptable if the patient is well supervised). The US recommendation is 6RZE with the option of adding a quinolone such as moxifloxacin. The level of evidence for all these regimens is poor, and there is little to recommend one over the other.

Regimens omitting rifampicin

It is rare for TB strains to be resistant to rifampicin without being resistant to isoniazid, but rifampicin intolerance is not uncommon (hepatitis or thrombocytopaenia being the most common reasons for stopping rifampicin). Of the first-line drugs, rifampicin is also the most expensive, and in the poorest countries, regimens omitting rifampicin are therefore often used. Rifampicin is the most potent sterilising drug available for the treatment of tuberculosis and all treatment regimens that omit rifampicin are significantly longer than the standard regimen.

The UK recommendation is 18HE or 12HEZ. The US recommendation is 9 to 12HEZ, with option of adding a quinolone (for example, MXF).

Regimens omitting pyrazinamide

PZA is a common cause of rash, hepatitis and of painful arthralgia in the HREZ regimen, and can be safely stopped in those patients who are intolerant to it. Isolated PZA resistance is uncommon in M. tuberculosis, but M. bovis is innately resistant to PZA. PZA is not crucial to the treatment of fully-sensitive TB, and its main value is in shortening the total treatment duration from nine months to six.

There is good evidence from UK trials that a regimen of 9HR is adequate for M. tuberculosis; this is also the first-line regimen used to treat M. bovis.

Regimens omitting ethambutol

EMB intolerance or resistance is rare. If a patient is truly intolerant or is infected with TB that is resistant to EMB, then 2HRZ/4HR is a perfectly acceptable regimen. EMB has no part to play in the treatment of TB that is sensitive to both INH and RMP, and the only reason for including it in the initial regimen is because of increasing rates of INH resistance. If INH resistance rates are known to be low, or if the infecting TB strain is known to be INH-sensitive, then there is no need to use EMB anyway.

Tuberculosis and other conditions

Liver disease

It should be noted that patients with alcoholic liver disease are at an increased risk of tuberculosis. The incidence of tuberculous peritonitis is particularly high in patients with cirrhosis of the liver.

No dosing change needs to be made in the dosing of patients with known liver disease, unless the liver disease is thought to have been caused by TB treatment. Some authorities recommend avoiding PZA in patients with known liver disease, because of the five first-line drugs, PZA has the highest risk of producing drug-induced hepatitis.

Patients with pre-existing liver disease should have their liver function tests monitored regularly throughout TB treatment.

Drug-induced hepatitis is discussed in a separate section above.

Pregnancy

Pregnancy itself is not a risk factor for TB.

Rifampicin makes hormonal contraception less effective, so additional precautions need to be taken for birth control during tuberculosis treatment.

Untreated TB in pregnancy is associated with an increased risk of miscarriage and major foetal abnormality, and treatment of pregnant women. The US guidelines recommend omitting PZA when treating TB in pregnancy; the UK and WHO guidelines make no such recommendation. There is extensive experience with the treatment of pregnant women with TB and no toxic effect of PZA in pregnancy has ever been found. High doses of RMP (much higher than used in humans) causes neural tube defects in animals, but no such effect has ever been found in humans. There may be an increased risk of hepatitis in pregnancy and during the puerperium. It is prudent to advise all women of child-bearing age to avoid getting pregnant until TB treatment is completed.

Aminoglycosides (STM, capreomycin, amikacin) should be used with caution in pregnancy, because they may cause deafness in the unborn child. The attending physician must weigh the benefits of treating the mother against the potential harm to the baby, and good outcomes have been reported in children whose mothers were treated with aminoglycosides.<ref name="Drobac2005">Drobac PC et al. (2005). "Treatment of Multidrug-Resistant Tuberculosis during Pregnancy: Long-Term Follow-Up of 6 Children with Intrauterine Exposure to Second-Line Agents" (Template:Dead link – ^{Scholar search}). Clin Infect Dis 40 (11): 1689–92. doi:10.1086/430066. PMID 15889370. </ref>

Kidney disease

Patients with renal failure have a 10 to 30-fold increase in risk of getting TB. Patients with kidney disease who are being given immunosuppressive drugs or are being considered for transplant should be considered for treatment of latent tuberculosis if appropriate.

Aminoglycosides (STM, capreomycin and amikacin) should be avoided in patients with mild to severe kidney problems because of the increased risk of damage to the kidneys. If the use of aminoglycosides cannot be avoided (e.g., in treating drug-resistant TB) then serum levels must be closely monitored and the patient warned to report any side-effects (deafness in particular). If patient have end-stage renal failure and have no useful remaining kidney function, then aminoglycosides can be used, but only if drug levels can be easily measured (often only amikacin levels can be measured).

In mild renal impairment, no change needs to be made in dosing any of the other drugs routinely used in the treatment of TB. In severe renal insufficiency (GFR<30), the EMB dose should be halved (or avoided altogether). The PZA dose is 20mg/kg/day (UK recommendation) or three-quarters the normal dose (US recommendation), but not much published evidence is available to support this.

When using 2HRZ/4HR in patients on dialysis, the drugs should be given daily during the initial high-intensity phase. In the continuation phase, the drugs should be given at the end of each haemodialysis session and no dose should be taken on non-dialysis days.

HIV

In patients with HIV, treatment for the HIV should be delayed until TB treatment is completed, if possible.

The current UK guidance (provided by the British HIV Association) is

• CD4 count over 200—delay treatment until the six months of TB treatment are complete.
• CD4 count 100 to 200—delay treatment until the initial two month intensive phase of therapy is complete
• CD4 count less than 100—the situation is unclear and patients should be enrolled in clinical trials examining this question. There is evidence that if these patients are managed by a specialist in both TB and HIV then outcomes are not compromised for either disease.<ref>Breen RAM, Miller RF, Gorsuch T, et al. (2006). "Virological response to highly active antiretroviral therapy is unaffected by antituberculosis therapy". J Infect Dis 193 (10): 1437–40. doi:10.1086/503437. PMID 16619192. </ref>

If HIV treatment has to be started while a patient is still on TB treatment, then the advice of a specialist HIV pharmacist should be sought. In general, there is no significant interactions with the NRTI's. Nevirapine should not be used with rifampicin. Efavirenz may be used, but dose used depends on the patient's weight (600mg daily if weight less than 50kg; 800mg daily if weight greater than 50kg). Efavirenz levels should be checked early after starting treatment (unfortunately, this is not a service routinely offered in the US, but is readily available in the UK). The protease inhibitors must be avoided if at all possible.

Thioacetazone must not be used because the risk of potentially fatal exfoliative dermatitis.

Epilepsy

INH may be associated with an increased risk of seizures. Pyridoxine 10mg daily should be given to all epileptics taking INH. There is no evidence that INH causes seizures in patients who are not epileptic.

TB treatment involves numerous drug interactions with anti-epileptic drugs and serum drug levels should be closely monitored. There are serious interactions between rifampicin and carbamazepine, rifampicin and phenytoin, and rifampicin and sodium valproate. The advice of a pharmacist should always be sought.

Drug-resistant tuberculosis (MDR- and XDR-TB)

Main articles: MDR-TB and XDR-TB

Definitions

Multi-drug resistant tuberculosis (MDR-TB) is defined as TB that is resistant at least to INH and RMP. Isolates that are multiply-resistant to any other combination of anti-TB drugs but not to INH and RMP are not classed as MDR-TB.

As of Oct 2006, "Extensively drug-resistant tuberculosis" (XDR-TB) is defined as MDR-TB that is resistant to quinolones and also to any one of kanamycin, capreomycin, or amikacin.<ref>Template:Cite web</ref> The old case definition of XDR-TB is MDR-TB that is also resistant to three or more of the six classes of second-line drugs.<ref name="MMWR2006">Center for Disease Control (2006). "Emergence of Mycobacterium tuberculosis with Extensive Resistance to Second-Line Drugs — Worldwide, 2000–2004". MMWR Weekly 55 (11): 301–305. </ref> This definition should no longer be used, but is included here because many older publications refer to it.

The principles of treatment for MDR-TB and for XDR-TB are the same. The main difference is that XDR-TB is associated with a much higher mortality rate than MDR-TB, because of a reduced number of effective treatment options.<ref name="MMWR2006"/> The epidemiology of XDR-TB is currently not well studied, but it is believed that XDR-TB does not transmit easily in healthy populations, but is capable of causing epidemics in populations which are already stricken by HIV and therefore more susceptible to TB infection.<ref name="Reuters2006">Template:Cite web </ref>