This is a classic problem of clinical medicine where the desire to "do no harm" may conflict with the need to relieve patient suffering. In an ideal world, one would like to complete treatment for latent tuberculosis (TB) prior to instituting TNF-blockade. This would mean a 9-month course of isoniazid (INH) - or 4 to 6 months of rifampin, or 2 to 4 months of rifampin/pyrazinamide in INH-intolerant patients. Of course, this would be an intolerable delay for most patients with active rheumatoid arthritis (RA) who have already failed multiple DMARDs. The opposite approach, that is, starting treatment for latent TB at the same time that disease-modifiying therapy is instituted, is used in patients with lymphoma who require chemotherapy and are found to be PPD positive. However, patients with RA don't have an immediately life-threatening illness that would justify this approach.
Most infectious disease experts I have polled recommend 3 months of anti-TB treatment before intitiating TNF-blockade, but most also said that TNF-blockade could be started after 1 to 2 months if the patient's clinical condition demanded it. These are also the recommendations in a recently published review article. In addition to reducing the likelihood of TB reactivation, this period of pre-treatment allows one to be sure that anti-TB therapy is tolerated by the patient.
The risk of TB may be different with different TNF-inhibitors. The risk with infliximab, for example, appears to be higher than with etanercept. There are several possible reasons for this. First, infliximab is the TNF-inhibitor most commonly used in Europe, where TB is more endemic than in the U.S.. By contrast, 94% of etanercept users are in the U.S.. In addition, in the U.S., Medicare covers the cost of infliximab but not etanercept; so the elderly, who are more likely to have been exposed to TB, receive infliximab in greater percentages than younger patients.
In addition to these epidemiological differences, however, there may be biological differences between the two drugs which influence TB susceptibility in treated patients. Infliximab is a human-murine chimeric monoclonal antibody that binds to soluble and transmembrane TNF-alpha and that can cause complement-mediated cell lysis of macrophages and monocytes. Etanercept, a fusion protein consisting of the ligand portion of the p75 TNF receptor linked to the Fc portion of IgG1, does not produce cell apoptosis and monocytopenia. Hence, infliximab may have greater inhibitory effects on granuloma formation than etanercept and may therefore interfere to a greater extent with host defenses against TB. This may also explain infliximab's greater efficacy in treating Crohn's disease as compared to etanercept.
With regard to adalimumab, 8 cases of TB were identified in the first 542 patients treated in clinical trials (1.5%) according to the FDA. PPD screening was then instituted, and 5 TB cases were identified in the next 1,900 patients treated. TB was typically diagnosed after 3 to 8 months of treatment and was generally extrapulmonary. TB was seen with greater frequency in those receiving more than 40mg/2 weeks. Adalimumab is a fully humanized monoclonal antibody that binds to TNF-alpha. Like infliximab, it binds to transmembrane TNF-alpha and can cause cell lysis. Thus it may have similar effects on granuloma formation to infliximab. For this reason, and because etanercept has a shorter half-life than both infliximab and adalimumab, I favor the use of etanercept in patients with a positive PPD who require TNF-inhibition.
Anakinra is a recombinant IL-1 receptor antagonist (IL-ra) which competitively binds to IL-1. Like TNF,both IL-1beta and IL-1ra are expressed by cells in granulomas induced by M. tuberculosis. IL-1 receptor knock out (KO) mice develop larger granulomas than wild-type mice after infection with M. tuberculosis, suggesting that IL-1 modulates the host response to mycobacterial infection. IL-1-alpha/beta double KO mice also develop significantly larger granulomas than wild-type mice after infection with M. tuberculosis, but the granulomas are not necrotic. Thus, the impact of IL-1 on granuloma formation appears more nuanced than that of TNF. In practice, anakinra has been associated with bacterial rather than opportunistic infections. Infection with TB has not been reported to date in patients receiving anakinra; thus, it appears to be safe to use in patients with a positive PPD. Prudence dictates treatment of latent TB in patients known to be PPD positive when anakinra is started, but treatment can begin concomitantly.
 Gardam MA, Keystone EC, Menzies R, Manners S, Skamene E, Long R, Vinh DC. Anti-tumour necrosis factor agents and tuberculosis risk: mechanisms of action and clinical management. Lancet Infect Dis. 2003 Mar;3(3):148-55.
 Sugawara I, Yamada H, Hua S, Mizuno S. Role of interleukin (IL)-1 type 1 receptor in mycobacterial infection. Microbiol Immunol. 2001;45(11):743-50.
 Yamada H, Mizumo S, Horai R, Iwakura Y, Sugawara I. Protective role of interleukin-1 in mycobacterial infection in IL-1 alpha/beta double-knockout mice. Lab Invest. 2000 May;80(5):759-67.