Portal de Periódicos da CAPES

To eliminate tuberculosis, we must universally encourage treatment of its latent form

2015; Wiley; Volume: 20; Issue: 6 Linguagem: Inglês

10.1111/resp.12598

1440-1843

Dean E. Schraufnagel,

Diagnosis and treatment of tuberculosis

In September 2014, the World Health Organization (WHO) unveiled a new policy on latent tuberculosis, which expanded treatment to include persons living with human immunodeficiency virus (HIV), adult and child contacts with pulmonary tuberculosis, those beginning anti-tumour necrosis factor-α therapy, those receiving dialysis, those undergoing organ or haematological transplantation, and persons with silicosis. This is a laudable step, but needs to go further to foster a global attitude favouring treatment of latent tuberculosis. The value of treatment is an inescapable conclusion of widely accepted premises. There are 8.7 million new cases of tuberculosis and 1.4 million deaths in the world annually. Although the case rates are decreasing, the drop is too slow.1 The opportunity to save lives demands greater action sooner. It is commonly accepted that susceptible persons inhale viable bacilli that multiply and cause an immune reaction in the lung.2 Subsequently, primary disease results from a host immune reaction. The host's immune system usually then controls the infection. Persons with primary tuberculosis have no or minimal symptoms. Primary tuberculosis generally resolves, but leads to a latent phase. Although re-infection and progressive primary disease occur, the latent state in most persons persists for years and is the source of active cases. The presence of an immune response in an asymptomatic individual defines latent tuberculosis. Latent tuberculosis may develop into active disease at any time in the life of the infected person. About 10% of people with latent tuberculosis develop active disease, although the actual percentage does not reduce the importance of latency. Assuming tuberculosis is spread before the diagnosis is made, if the risk of a person with latent developing active disease is 5%, then to maintain a steady state of tuberculosis in the world, each affected individual must infect 20 persons. If the risk is 10%, then to maintain a global steady state, each individual must infect 10 persons and, if the risk is 20%, each individual would have to infect only 5. A main reason that tuberculosis is difficult to control is that its presentation is subtle and prolonged.3 Most persons have a cough for more than a month. By the time the diagnosis is made, close contacts often have new latent and sometimes active disease. Once persons with active tuberculosis are treated, the organisms are rapidly killed. Jindani et al. showed that giving isoniazid for 2 days killed 90% of the organisms in persons with active disease.4 Dharmadhikari et al. also confirmed that treated persons rapidly become non-contagious as assayed by their sensitive guinea pig model.5 Treating latent tuberculosis prevents active tuberculosis. A large International Union Against Tuberculosis trial showed that giving isoniazid for 1 year reduced the risk of subsequent tuberculosis by about 90% in adherent subjects.6 In Bethel, Alaska, the prevalence of tuberculin-positive persons under the age of 15 years was 75% in 1950; 8% of the population acquired new infections annually. An extensive isoniazid treatment programme begun in 1953. By 1960, annual new infection rate had dropped to 1%. Death rates from tuberculosis dropped from 650 per 100 000 in 1950 to 26 per 100 000 in 1960, and no deaths were reported in 1970.7 The main intervention was the treatment of latent disease in over 80% of the population, although other interventions, including effective therapy for active disease and education, occurred during this time. The isoniazid therapy was particularly effective in individuals under the age of 40 years. Those given isoniazid had a case rate of 1.75 compared with 5.46 for those given placebo 13 years after the intervention.8 Similar positive effects have been shown in HIV-infected persons.9, 10 Several other studies have also shown that treating latent tuberculosis is effective.11, 12 Treatment of latent tuberculosis should reduce the global burden of tuberculosis more than improved treatments given after patients have been diagnosed because they have already spread the infection. Shortening the treatment course to 4 or even 2 months would not shorten the contagious period. The past successes of treating latent tuberculosis have largely been with isoniazid, but laboratory studies show that isoniazid is not effective against dormant bacilli,13 which are the predominant organisms present in persons with latent tuberculosis. If isoniazid only kills replicating organisms, then bacillary resuscitation (switching to an oxidative metabolism) must occur frequently enough to eventually allow isoniazid to kill most of the bacilli over 9–12 months, but infrequent enough to require that long to do it. Despite many studies showing isoniazid is an effective treatment of latent disease, infected persons are not generally treated. Guidelines do not recommend it for most persons because the treatment duration is too long, too cumbersome and too toxic. Toxicity, adherence, cost and treatment failure are all related to the duration of therapy. Dormant bacilli can be defined by the metabolic changes they undergo in response to an environmental stress, such as hypoxia. The adaptation from a replicative to a dormant state consists of altering metabolism to use the reductive branch of the tricarboxylic acid cycle. About 50 genes in the DOS regulon control the changes that occur as oxygen or nutrients become scarce and cells must maintain essential metabolic activity for survival. It is reasonable to postulate that drugs that are effective against dormant bacilli would shorten the duration of therapy. In fact, the rifamycins are active against dormant bacilli and are effective in shortening the regimens. A Hong Kong trial showed that 3 months of daily rifampin was sufficient to reduce the number of new cases of active tuberculosis,14 and the United States TB Trials Consortium showed that weekly rifapentine and isoniazid for 12 doses was also effective.15 Pyrazinamide, nitrocompounds (such as metronidazole, niclosamide, nitazoxanide, pretomanid (PA-824) and delamanid), moxifloxacin, aminoglycosides, capreomycin, linezolid and clofazimine have activity against non-replicating bacilli, depending on the model used to study them. Bedaquiline is also very effective against dormant bacilli,16 but isoniazid and ethambutol are not.17, 18 The decision whether to treat latent tuberculosis is based on risk–benefit considerations that are a half century old. Complicated rules were developed that balanced the risk of tuberculosis occurring in an individual against the toxicity of taking isoniazid for 12 months. They did not consider the long-term protection8 versus the risk of side effects occurring only during the treatment period, Nor did they take into account individual circumstances and public health.19 Shorter courses and more effective medicine change these considerations. Rifapentine and isoniazid, once weekly for 12 weeks, was as effective as the standard isoniazid course and had less toxicity (0.5%) and a higher completion rate (82%).15 Studies are ongoing that test rifapentine by itself and with other drugs that are effective against dormant bacilli. The public health aspects of treating latent tuberculosis have largely been ignored. We vaccinate for polio. We treat hypertension. We treat high cholesterol. We treat latent syphilis. Why are we not treating latent tuberculosis? Both tuberculosis and syphilis are caused by intracellular bacteria and both produce granulomatous disease. Both have a latent and active disease phase. In both conditions, active disease develops from the latent infection in an unpredictable manner, although both can be related to decreased host immunity. The latent phases of these two diseases also have similarities, although the information on latent syphilis is old and controversial. In Norway, only 19% of 50 patients with untreated syphilis progressed to active disease.20, 21 In other studies of latent syphilis, 10% developed late cardiac disease, 6.5% developed late neurologic disease and 16% developed late ‘benign’ disease.22 Why, then, do most physicians treat latent syphilis and not latent tuberculosis? The answer is in the perceived risk–benefit ratio. The benefit of penicillin in treating syphilis is high, and the chance of adverse effects are low. A single dose of penicillin may be all that is required, whereas the current treatment of latent tuberculosis is prolonged. In treating latent tuberculosis, the risk was perceived to be high based on the hepatotoxicity and death of patients treated with isoniazid for 1 year. With the lower toxicity of the shorter therapy, considerations of the treatment of latent tuberculosis may approach those for the treatment of latent syphilis. The recent WHO document cited four potential concerns for treating latent tuberculosis. The concerns were that it could be unsafe, too expensive, not for developing countries and unethical.23 Nolan et al. treated 11 141 consecutive patients with isoniazid for latent tuberculosis, and had 11 cases of hepatotoxicity; 1 was hospitalized and none died. This amounted to 0.10% rate of hepatoxicity for those starting isoniazid and 0.15% of those completing therapy.24 LoBue and Moser reported giving isoniazid to 3788 patients and finding 0.3% had hepatotoxicity.25 The rates of side effects reported in these studies, however, are better than most other studies, probably because they were outstanding clinics that incorporated patient education, access, communication and clinical monitoring. It is important to note that other clinics may not achieve this level of safety and that a treatment programme should include strengthening clinics and attention to the care of these patients. The patients in the studies quoted above were given isoniazid for 6–12 months. Isoniazid is more toxic than rifampin26 owing to the longer duration of treatment. This paper emanates from a talk I gave at the 19th Congress of the Asian Pacific Society of Respirology held in Bali, Indonesia, in November 2014. The first response by an expert to my suggestion of universal treatment of latent tuberculosis was, ‘So you're going to treat a third of the world who have latent TB?’ The answer is no. First, it is unlikely that a ‘third of the world’ has latent tuberculosis. These estimates were published by Dye et al. in 1999 based on tuberculin testing27 rather than the more specific interferon-γ release assays. Interferon-γ release assays do not count those vaccinated with Bacillus Calmette–Guérin, which could cause a prevalence overestimation. Second, the call of this paper is to abandon the conservative attitude of non-treatment. The approach would be to offer treatment to those already known to have infection. Third, if the world can afford to offer anti-retroviral medicine for life to all who have HIV, it can afford to offer a short course of rifampin to cure latent tuberculosis of those who are infected. Last, not treating is too expensive because treating latent disease is a powerful tool to reduce the prevalence of tuberculosis. Treating tuberculosis before its spreads should be cost-effective in the global elimination process and, in fact, will be required for tuberculosis elimination.28 This objection is also not valid; the populations with the highest burden benefit most. Consider the Alaska and HIV experience. At one time, it was widely accepted that rifampin should not be used in developing countries because once resistance develops further treatment would be impossible. There was also a time when resistant tuberculosis was not effectively treated in developing countries because of the lack of capacity and cost, but the last decade has shown that the world can mobilize resources to treat drug-resistant tuberculosis in all countries. Although some have raised concerns about the ethics of treating persons with latent tuberculosis, in the rifapentine trial conducted by the US Centers for Disease Control and Prevention, it was felt to be unethical to have placebo group.15 There is a long and sad history about not treating latent syphilis, which we do not want to be repeated for tuberculosis. There is little concern about the ethics of giving someone standard public health vaccines. This argument is not for mass screening to detect latent tuberculosis. Mass screening has not been shown to be cost-effective because of the high false positive rates, especially in low prevalence settings. Mass therapy without testing is also less effective,29 and the detection tools have significant limitations. However, selective screening, of say health-care workers, for example, may be practical. A successful programme must have a well-functioning clinic with follow-up and follow-through, which includes patient education and access to health-care workers. Isoniazid is very effective at reducing the mycobacterial load in active disease, but sluggish for latent disease. We need a better drug. We should not need a companion drug because latent tuberculosis has a low bacterial burden. The agent should be active against dormant bacilli to shorten the time course. The duration of treatment still needs to be determined, but 2 months of pyrazinamide–rifampin and 12 doses of rifapentine–isoniazid have been shown to be effective. The drug should have an outstanding safety profile. The pharmaceutical industry should have interest in developing such a medication that could be used by a large segment of the world. The current best, widely available medications are the rifamycins. The strength of this argument is that it not only benefits the individual, but also public health. The argument will become even stronger as better regimens are developed. Universal treatment will decrease transmission and hasten elimination. In fact, as long as there is latent tuberculosis, there will be no elimination.