It is has been well known from the earliest days of chemotherapy for tuberculosis that two or more drugs are required to retard the emergence of drug resistance [2]. Unfortunately, patients sometimes take only one drug (monotherapy) when more than one drug is prescribed. This occurs because of a temporary lack of a supply of one or more medications; mistakes in dispensing; patients' decisions to purchase only one medication to save money; supplies of one medication lasting longer than supplies of other medications; loss of one medication; misunderstandings; and patients' deliberate decisions to take only one drug because of perceived or real symptoms associated with the other drug, such as arthralgias with pyrazinamide or red urine with rifampin. When a patient takes monotherapy for one of these reasons, drug resistance is likely to follow.

Drug companies provide fixed-dose combinations of antituberculous medications with two or more medications in one capsule, such as isoniazid and rifampin (Rifamate [Marion Merrell Dow, Kansas City, Missouri], which has been available in the United States for many years), and isoniazid, rifampin, and pyrazinamide (Rifater [Marion Merrell Dow], which has recently been approved for distribution in the United States). Because these combined preparations make it impossible for the patient to receive monotherapy, it would follow that fixed-dose combinations would be preferable to individual drugs in the treatment of tuberculosis. However, only 15% to 18% of rifampin was sold in the form of fixed-dose combinations in the United States in 1993 (Parks D, Marion Merrell Dow. Personal communication). Why are these combined preparations used so infrequently?

In 1969, the National Academy of Sciences issued a strong statement condemning the use of fixed combinations of antimicrobial agents [3]. The statement did not address the treatment of tuberculosis and the prevention of monotherapy by the use of fixed-dose combinations, apparently because combinations of isoniazid and rifampin or isoniazid, rifampin, and pyrazinamide had not yet been introduced. With such a high-level authority opposing the use of fixed-dose combinations in general, it is understandable that this logical approach to the treatment of tuberculosis has not been widely recommended or widely used. Although such preparations have been mentioned in previous American Thoracic Society and Centers for Disease Control and Prevention (CDC) recommendations, they have never been strongly endorsed until recently [4].

Three controlled trials have compared the combined preparation of isoniazid, rifampin, and pyrazinamide with individual drugs [5-8]. These three trials showed more rapid rates of sputum conversion in the combined-preparation group; in one trial, this difference reached statistical significance [7, 8]. These trials also showed that the combined-preparation group had slightly higher rates of relapses, a difference that reached marginal statistical significance in one trial (P = 0.04) [5]; however, no difference in the rates of acquired drug resistance was found in these trials. Self-administered medication was used in only one trial, a trial that had an insufficient number of patients who failed treatment or relapsed to establish the value or lack of value of combined preparations in preventing the development of drug resistance [7, 8].

Considerable evidence shows that monotherapy is vastly more efficient than treatment with multiple drugs in creating drug resistance. In the following example, monotherapy was 30 times more efficient. Only 1.8% (3 of 170) of patients treated with isoniazid and rifampin given daily for 6 months developed one or more drug-resistant isolates in the fourth, fifth, or sixth month of chemotherapy [9]. By contrast, 54% of patients (38 of 70) developed drug resistance after isoniazid was given as monotherapy for the same duration (6 months) [10].

No one knows how much of the drug-resistance problem is due to monotherapy and how much is due to interrupted treatment in patients receiving both drugs when they actually take them. This is primarily because the proportion of poor compliance that takes the form of monotherapy is unknown. However, even if monotherapy occurs infrequently, it must still be a major cause of drug resistance because monotherapy is vastly more efficient than multidrug treatment in creating drug resistance.

Other evidence comes from examination of drug-resistance data and treatment practices in various communities. In the United Kingdom, which has low rates of drug resistance, 73% to 79% of rifampin is sold as fixed-dose combinations (Davies P. Personal communication). In the United States, which has high rates of drug resistance, 15% to 18% of rifampin is sold in this form. These data have been used as a basis for a claim that the low use of fixed-dose combinations in the United States is a reason for high rates of drug resistance [11]. In Los Angeles, where the health department has used Rifamate for the last 9 years, a survey in progress by one of the authors has found little drug resistance in patients who remained culture positive or who had a relapse after therapy if they received Rifamate for the entire duration of treatment. These two pieces of imperfect evidence and, more importantly, the fact that treatment with combined preparations precludes monotherapy and its highly efficient ability to create drug resistance form a persuasive argument that fixed-dose combinations should be used for treating tuberculosis whenever possible.

There are several concerns about the use of fixed-dose combinations. It is currently recommended that three drugs be used in the treatment of tuberculosis whenever the chance of drug resistance is small. In areas where the rate of drug resistance is greater than 4%, a fourth drug, ethambutol, should be added to the initial regimen [4]. This could be achieved by using Rifater and ethambutol. However, if a patient chose to take only Rifater and the patient's organisms were resistant to two of the three drugs in Rifater, monotherapy would result.

In addition, evidence suggests that pyrazinamide is modestly less effective than ethambutol in preventing drug resistance in a companion drug [12]. Consequently, a patient with organisms resistant to only one of the two major drugs in Rifater, that is, isoniazid or rifampin, might in theory develop resistance to the other major drug when treated with Rifater and ethambutol if he or she did not take the ethambutol. Therefore, in areas with high rates of drug resistance, it would be reasonable to use directly observed therapy until the results of susceptibility studies are available, especially for patients with advanced disease, coinfection with the human immunodeficiency virus, or a high probability of poor adherence to the treatment regimen.

Concern has been expressed that the dose of one or more of the individual drugs provided by fixed-dose combinations is not consistent with the dosages that are usually given. For Rifamate, this is not a problem because two capsules provide the conventional dose of 300 mg of isoniazid and 600 mg of rifampin, which is used for treating most adults. Each tablet of Rifater, however, contains 50 mg of isoniazid, 120 mg of rifampin, and 300 mg of pyrazinamide. The Rifater dosing schedule for adults weighing less than 65 kg is based on one tablet for each 10 kg of body weight. This schedule provides conventional doses of approximately 5 mg/kg of isoniazid, 12 mg/kg of rifampin, and 30 mg/kg of pyrazinamide. For patients weighing more than 65 kg, no further increase in dosage is recommended because of concern that rifampin toxicity might occur. These heavier patients are advised to take six tablets, which provide 720 mg of rifampin, 300 mg of isoniazid, and 1800 mg of pyrazinamide. These doses are identical to or close to the recently recommended maximum daily dose of 300 mg for isoniazid, 600 mg for rifampin, and 2000 mg for pyrazinamide [4]. After reviewing data from the manufacturer and the Food and Drug Administration, officials at the CDC have concluded that the 720-mg dose of rifampin in the fixed-dose combination Rifater is comparable to the 600-mg dose of rifampin that was recommended in the statement on the treatment of tuberculosis by the American Thoracic Society and the CDC (Geiter L. Personal communication).

Outside the United States, fixed-dose combinations with a much higher ratio of isoniazid to rifampin and pyrazinamide are available for intermittent twice-weekly or thrice-weekly therapy, which requires a higher isoniazid dose. Because these preparations are not available in the United States, individual drugs or combined preparations plus additional isoniazid must be used in the administration of intermittent regimens.

The issue of how one should manage drug reactions when using fixed-dose combinations has also been raised. These drug reactions are best handled exactly as is toxicity with individual drugs, that is, by stopping all drugs, in this case the fixed-dose combination, and reinstituting the drug or drugs that the patient can tolerate.

One other problem might occur if a patient consistently ingests less than the prescribed dose of a combined preparation. Would this lead to drug resistance? The ratio of blood levels and presumably tissue levels compared with the minimal inhibitory concentration is considerably higher for isoniazid than for rifampin and pyrazinamide. Therefore, if a patient consistently ingested Rifamate and Rifater at half or less than half of the prescribed dose, the isoniazid might be present in a sufficient concentration to kill or inhibit isoniazid-susceptible organisms, whereas the rifampin and pyrazinamide might not be present in adequate concentrations to kill or inhibit rifampin- and pyrazinamide-susceptible organisms. If this occurs, isoniazid monotherapy might be the result, which would lead to an isoniazid-resistant strain. We acknowledge this theoretical concern but believe that it is remote and should not outweigh the strong arguments in favor of using combined preparations.

A more practical concern is the similarity of the names Rifamate and rifampin, which sometimes leads to errors in prescribing and dispensing: Patients may receive monotherapy with rifampin when Rifamate is ordered. In other parts of the world, Marion Merrell Dow markets the same preparation under the name Rifinah. We are told that even this name has led to confusion, with patients being given rifampin when the fixed-dose combination, Rifinah, was ordered (Davies P. Personal communication). This problem could be overcome if the medical community, the government, and the drug company worked together to change the name Rifamate to a highly distinctive name such as Rif-Isoniazid. In addition, the physical similarity of rifampin and Rifamate capsules has resulted in patients receiving monotherapy with rifampin when Rifamate was ordered. In one case, the single drug rifampin was placed in a bottle labeled rifampin-isoniazid and was given to a patient (Contreras C. Personal communication). The chance for this type of error could be reduced if the drug company changed the color or some other physical characteristic of one of the preparations.

Concern has been expressed about a potential increase in drug costs if fixed-dose combinations are recommended as the standard treatment for tuberculosis [13]. This legitimate concern must be weighed against the much greater cost of allowing monotherapy to generate cases of drug-resistant tuberculosis and multidrug-resistant tuberculosis, which are far more expensive to treat [14] and may give rise to other multidrug-resistant cases. Furthermore, market forces should help to restrain a price increase because the patents on all the basic drugs have expired and because no patents on the concept of combining drugs have been granted.

Unfortunately, fixed-dose combinations supplied by some drug companies provide inadequate blood levels of the component drugs, especially rifampin [15, 16]. Although this is mainly a problem in developing countries, it might be a potential problem in the United States if other companies chose to introduce competitive fixed-dose combinations. This problem is easily avoided by requiring that all fixed-dose combinations meet strict bioavailability standards as recommended by the World Health Organization and the International Union Against Tuberculosis and Lung Disease [16].

A strong opinion has been expressed that all patients should be treated for tuberculosis with directly observed therapy [17]. If this recommendation is universally implemented, one could argue that fixed-dose combinations would not be needed. The recent recommendations of the American Thoracic Society and the CDC go a long way toward endorsing this practice by stating that "consideration should be given to treating all patients with directly observed therapy" [4]. However, this strong endorsement leaves open the option of using self-administered medication, the form of treatment in which combined preparations have their greatest value. Furthermore, patients who initially receive directly observed therapy may be changed to self-administered medication in the latter part of their treatment and frequently receive self-administered medication on weekends and holidays. These situations are best managed by treating all patients with combined preparations so that a change in product does not confuse the patients or staff, as long as the increase in the cost of the combined preparation is not excessive. Therefore, even with the increasing emphasis on directly observed therapy, the need for fixed-dose combinations of antituberculous drugs remains.

In short, at least three mechanisms contribute to the drug-resistance problem: 1) interrupted treatment, which is best controlled by increased supervision; 2) preexisting drug resistance, which requires a sufficient number of new drugs in the initial treatment of patients; and 3) monotherapy that occurs despite the prescription of multiple drugs, which is best prevented by using fixed-dose combinations of medication. This latter approach, which until recently has been largely neglected in the United States, should be widely promoted and accepted as a primary way to prevent the emergence of drug-resistant organisms.