Sex and age differences in Mycobacterium tuberculosis infection in Brazil.

Globally, the prevalence of tuberculosis (TB) disease is higher in males. This study examined the effect of sex and age on Mycobacterium tuberculosis (Mtb) infection. Demographic and exposure data were collected on household contacts of sputum smear-positive pulmonary TB patients in Brazil. Contacts with tuberculin skin test induration ⩾10 mm at baseline or 12 weeks were considered Mtb infected. The study enrolled 917 household contacts from 160 households; 508 (55.4%) were female, median age was 21.0 years (range 0.30-87.0) and 609 (66.4%) had Mtb infection. The proportion infected increased with age from 63.3% in girls <5 years to 75.4% in women ⩾40 years and from 44.9% in boys <5 years to 73.6% in men ⩾40 years. Multivariable modelling showed the odds of infection increased between age 5 and 14 years among female contacts (OR 1.5 per 5-year age increase; 95% CI 1.1-2.2; P = 0.02) and between ages 0-4 and 15-39 years among male contacts (OR 2.7, 95% CI 0.83-8.9 and 1.1, 95% CI 0.99-1.3 per 5-year age increase; P = 0.10, 0.07, respectively). The study suggests that the age at which Mtb infection increases most is different in females compared with males. Studies are needed to explore whether these findings are due to differences in host susceptibility, exposure outside the household or other factors.

Defective interleukin 2 production and responsiveness in human pulmonary tuberculosis.

Patients with newly diagnosed, pulmonary tuberculosis had a tuberculin-specific defect in IL-2 production. Mean PPD-induced IL-2 activity was 81.2% lower in patients as compared with healthy tuberculin reactors. PPD-induced expression of T cell IL-2 receptors was 5.9 times less in peripheral blood mononuclear cells of patients with tuberculosis as compared with healthy tuberculin reactors. Furthermore, purified IL-2 failed to correct PPD-induced blastogenesis in patients. Suppression by adherent cells was operative in one group of patients; adherent cell depletion increased their T cell production of IL-2 7.2-fold. A second group of patients with low IL-2 production did not have suppressor adherent cells and were clinically distinct, with more extensive disease on chest x ray. The basis for low IL-2 production in such individuals is unknown. Disordered regulation of IL-2 metabolism may be a key feature in the depressed cellular immune response of tuberculosis.

Role of arginase in killing of schistosomula of Schistosoma mansoni.

Nonspecific resistance to the multicellular organism Schistosoma mansoni can be induced in mice by several infectious agents. We utilized the observed genetic restriction of such acquired resistance to study the mediators of killing of the larval stage of S. mansoni in vitro. Adherent peritoneal cell monolayers from Corynebacterium parvum-treated C57BL/6J but not from C. parvum-treated BALB/cJ mice killed an increased proportion of schistosomula in 24 h. Activated macrophages (Mphi) from both strains exhibited enhanced H(2)0(2) production after incubation with the parasites or phorbol myristate acetate. Thus H(2)0(2) production was not associated with schistosomula killing. Moreover, schistosomula killing was unaffected by catalase or superoxide dismutase. In contrast, activated C57BL/6J (but not BALB/cJ) Mphi released fourfold more arginase into supernates than control Mphi. Schistosomula killing by these Mphi correlated with arginase content of the supernates, was exaggerated in arginine-poor medium, and could be blocked by the addition of arginine. Exogenous bovine arginase added to Fischer's medium without macrophages produced comparable parasite mortality. Our data suggest that arginase is a critical mediator of in vitro killing of this multicellular organism by activated macrophages.

Modulation of the effector function of human monocytes for Mycobacterium avium by human immunodeficiency virus-1 envelope glycoprotein gp120.

Disseminated Mycobacterium avium infection in AIDS is associated with high tissue burdens (10(9)-10(10) mycobacteria/g tissue) of organism. The basis for the extraordinary susceptibility of AIDS to M. avium infection is unclear. HIV or its constituents may alter mononuclear phagocyte functions resulting in enhanced intracellular M. avium growth. The effects of an envelope glycoprotein (gp120), a transmembrane protein (p121), and core proteins of HIV-1 on M. avium infection of human monocytes were examined. Preculturing monocytes with gp120 inhibited M. avium phagocytosis and consistently enhanced intracellular growth of six M. avium strains. Pretreatment with p121, gag5, or p24 did not inhibit phagocytosis nor enhance intracellular growth of M. avium. Incubation of gp120 with soluble CD4 before addition to monocyte cultures or pretreatment of monocytes with OKT4A abrogated gp120 effects on M. avium phagocytosis and intracellular growth. gp120 also augmented cytokine production by infected monocytes. These results suggest that gp120, but not p121 or core proteins, modulate monocyte phagocytosis and enhance intracellular growth of M. avium at least in part through monocyte CD4 receptors. Direct effects of HIV-1 products may, therefore, contribute to the diathesis of AIDS to develop disseminated M. avium infection and to the extensive replication of the organisms within tissue macrophages.

Selective expansion of human gamma delta T cells by monocytes infected with live Mycobacterium tuberculosis.

Gamma delta (gamma delta) T cell receptor (TCR) expressing T cells comprise 3% of human peripheral blood lymphocytes, yet their role in the immune response remains largely unknown. There is evidence both in humans and in animal models that these cells participate in the immune response to mycobacterial antigens. In mice, exposure to mycobacterial antigens leads to the expansion of gamma delta T cells in draining lymph nodes and lungs. In humans, gamma delta T cell lines with reactivity to mycobacterial antigens have been derived from synovial fluid of a rheumatoid arthritis patient, skin lesions of leprosy patients, and peripheral blood of a healthy tuberculin reactor. Very little is known, however, about the factors which induce human gamma delta T cells to expand. In studies comparing the human T cell response to live and heat-killed Mycobacterium tuberculosis (MT), we have found that monocytes infected with live MT are very effective inducers of human gamma delta T cell expansion. After 7 d of exposure to live MT, gamma delta T cells were greatly increased in all healthy tuberculin reactors (PPD+) tested and frequently were the predominant T cell population. In contrast, heat-killed MT or purified protein products of MT induced a CD4+, alpha beta TCR+ T cell response with very little increase in gamma delta T cells. Furthermore, a similar selective induction of gamma delta T cells was observed when monocytes infected with live Salmonella were used to stimulate T cells. Heat-killed Salmonella, like heat-killed MT, induced a predominantly CD4+ alpha beta TCR+ T cell response. These findings suggest that human gamma delta T cells are a major reactive T cell population during the early stages of infection with living intracellular bacteria and are therefore likely to exert an important role in the initial interaction between host and parasite.

Thromboxane A2 mediates augmented polymorphonuclear leukocyte adhesiveness.

We examined the role of prostaglandins and thromboxanes as mediators of plasma-dependent increased polymorphonuclear leukocyte adhesiveness induced by Escherichia coli lipopolysaccharide. The cyclo-oxygenase inhibitors-indomethacin and d,l-6-chloro-alpha-methyl-carbozole-2-acetic acid (R020-5720)-reduced lipopolysaccharide-induced adherence of polymorphonuclear leukocytes by 74 and 62%, respectively. In addition, inhibitors of thromboxane synthetase-imidazole, 9,11-azoprosta-5,13-dienoic acid, and 1-benzylimidazole-suppressed the stimulation of adherence by 31, 66, and 83%, respectively. Exogenous prostaglandins E(1), E(2), and F(2)alpha did not increase polymorphonuclear leukocyte adherence, nor were they detected in significant quantities in supernates of polymorphonuclear leukocytes exposed to lipopolysaccharide. However, inhibitors of both cyclo-oxygenase and thromboxane synthetase reduced increases in adherence induced by arachidonic acid (10 mug/ml), suggesting that lipopolysaccharide-mediated increases in adherence were due to an arachidonic acid product other than prostaglandin E(2) or F(2)alpha. 8,11,14-Eicosatrienoic acid, a precursor of monoenoic prostaglandins, did not enhance polymorphonuclear leukocyte adhesiveness. We next demonstrated lipopolysaccharide-stimulated generation, by polymorphonuclear leukocytes, of a labile, low molecular weight, dialyzable substance capable of enhancing the adherence of unstimulated leukocytes. In parallel experiments, a 10-fold increase in immunoreactive thromboxane B(2) over basal levels was detected after exposure of leukocytes to lipopolysaccharide. The inhibition of lipopolysaccharide enhancement of adherence by specific rabbit antibodies to thromboxane B(2) strongly supported a primary role for thromboxane A(2) as the mediator of the observed increases in adherence. Lipopolysaccharide-stimulated purified platelets did not increase leukocyte adherence, whereas thrombin-stimulated platelets did increase adherence. These studies suggest that lipopolysaccharide stimulates polymorphonuclear leukocytes to produce thromboxane A(2), which enhances their adhesiveness to nylon.

Suppression of lymphocyte responses by tuberculous plasma and mycobacterial arabinogalactan. Monocyte dependence and indomethacin reversibility.

During tuberculosis, exposure of monocytes to circulating factors may induce the suppressor activity observed in some anergic patients. To explore this possibility, we examined the effects of plasma pooled from 28 untreated tuberculosis (TB) patients and the mycobacterial cell wall polysaccharide D-arabino-D-galactan (AG) on the in vitro function of peripheral blood mononuclear cells (PBMC) from healthy donors. In the [3H] thymidine incorporation assay, stimulated responses of PBMC incubated in culture medium supplemented with TB plasma or co-cultured with 3.0 microgram/ml AG were depressed significantly when compared with control responses. Cytotoxicity and altered kinetics of stimulated DNA synthesis did not contribute to the observed suppression. TB plasma and AG-induced suppression of the PBMC response to purified protein derivative was monocyte dependent and indomethacin reversible. In addition, TB plasma and AG directly inhibited the phytohemagglutinin-stimulated responses of T lymphocytes. In a quantitative assay of monocyte attachment to plastic, both TB plasma and AG significantly increased monocyte adherence from basal levels. These effects on monocyte adherence were reversed with indomethacin or antibody to mycobacterial polysaccharide. In addition, TB plasma passed over an immunoabsorbent column of Sepharose-linked antibody to mycobacterial polysaccharide was depleted of the suppressive and monocyte-adherence augmenting factors. 3.0 microgram/ml AG stimulated a fivefold increase in prostaglandin E2 production by cultured mononuclear cells. Our data suggest that AG circulating alone or bound in immune complexes may account for the observed effects of TB plasma. Similar in vivo exposure may contribute to the cell-mediated suppression of lymphocyte responses in tuberculosis.

Cell membrane is a major locus for ultraviolet B-induced alterations in accessory cells.

In vitro ultraviolet B (UVB) irradiation of human blood monocytes inhibits their accessory cell function for antigen- and mitogen-induced T cell responses. These studies were designed to characterize the nature of the UVB-induced defect in human monocyte accessory cell function. Irradiated monocytes were deficient in their ability to serve as accessory cells for OKT3-induced T cell activation. In vitro exposure of monocytes to 100 J/m2 UVB completely inhibited the T cell proliferative response (51502 cpm, non-UVB-irradiated; 302 cpm, UVB-irradiated). Analysis of the accessory signals altered by UVB indicated that irradiated monocytes were incapable of binding to OKT3 molecules attached to the CD3 antigen on T cells. Provision of an alternative mechanism for binding of OKT3 molecules by attaching anti-mouse IgG to the bottom of microtiter wells completely restored accessory cell function. Further characterization of the defect demonstrated that UVB radiation did not deplete p72 Fc receptors from the surface of irradiated monocytes. However, UVB exposure did produce a dose-dependent decrease in monocyte membrane expression of ICAM-1. It is proposed that UVB radiation leads to changes within the cell membrane that inhibit the ability of monocytes to express selected molecules necessary for binding of T cells.

Identification of antigens of Mycobacterium tuberculosis using human monoclonal antibodies.

Pleural tuberculosis constitutes a human model of local protective immunity to mycobacterial infection as the disease is usually self-limited and recurrent pleurisy is rare. To identify potentially protective antigens of Mycobacterium tuberculosis, 37 human pleural fluid B cell clones were established using EBV and their supernatants assayed by ELISA and Western blot for antibody reactivity with M. tuberculosis sonicate and culture filtrate. One antibody identified 29,000, 31,000, and 33,000 bands in culture filtrate, and 31,000, 33,000, and 47,000 bands in sonicate; its species reactivity by ELISA was limited to M. tuberculosis. Eight antibodies identified a 31,000 band in culture filtrate and a 68,000 band in M. tuberculosis sonicate, suggesting recognition of a secreted antigen. The species crossreactivity of these eight antibodies extended to M. avium. Six antibodies identified multiple bands and had crossreactivity that included M. avium and M. kansasii. There was no reactivity with recombinant M. tuberculosis 65,000 antigen. Tuberculous pleurisy may prove useful in the identification of potentially protective mycobacterial antigens, particularly those secreted during active infection, and thus accessible to the human immune response.

Expression of functional interleukin 2 receptors by peripheral blood monocytes from patients with active pulmonary tuberculosis.

Peripheral blood monocytes from patients with active tuberculosis are "activated" by a number of criteria, including selective depression of T-lymphocyte responses to the mycobacterial antigen, tuberculin-purified protein derivative (PPD). We studied monocytes from patients with tuberculosis and healthy skin test reactive controls for expression and function of IL 2 receptors (IL 2R). Depletion of adherent monocytes increased the IL 2 activity of supernatants of PPD-stimulated T cell cultures from patients by 30-fold. When cultured with purified IL 2, adherent cells from the patients depleted IL 2 activity by 66%. Monocytes from patients displayed IL 2R on their surface as evidenced by anti-Tac reactivity, and released soluble IL 2R into the medium during culture. The release of soluble IL 2R was augmented when monocytes were cultured with PPD. Finally, freshly isolated adherent monocytes from patients but not healthy individuals expressed the gene encoding Tac protein. Thus, blood monocytes from patients with tuberculosis express functional IL 2R constitutively. This property may be important in the immunoregulatory and effector function of mononuclear phagocytes during tuberculosis.

Ultraviolet-irradiated monocytes efficiently inhibit the intracellular replication of Mycobacterium avium intracellulare.

The purpose of this study was to evaluate the effect of ultraviolet (UV) radiation on the antimicrobial activities of monocytes for the intracellular pathogen Mycobacterium avium intracellulare (MAI). UV radiation augmented monocyte antimicrobial activity for MAI in a dose-dependent fashion. UVB doses of greater than or equal to 25 J/m2 resulted in a 50-100-fold reduction in MAI growth 7 d after initiation of culture. The increased monocyte antibacterial effect could be blocked by a plate glass filter, indicating that wavelengths within the UVB were responsible for the effect. UV radiation did not stimulate monocyte phagocytosis, and enhanced inhibition of MAI growth was observed in populations of adherent mononuclear cells that were devoid of T cells. This suggested that UV radiation acted directly to augment intrinsic monocyte antimicrobial activities. The administration of 8-methoxypsoralen plus UVA radiation to monocytes also augmented their antimicrobial activities against MAI. UV radiation thus may serve as a unique agent by which to evaluate the mechanisms by which mononuclear phagocytes control the growth of MAI.

Destruction of the multicellular parasite Schistosoma mansoni by T lymphocytes.

The role of cytotoxic T lymphocytes in host defenses against infectious agents is unknown as these cells have not previously been demonstrated to kill microorganisms directly. We studied the cytotoxicity of T lymphocytes purified from peripheral blood mononuclear cells of healthy subjects for the multicellular schistosomula of Schistosoma mansoni. Unstimulated and phytohemagglutinin (PHA)-stimulated T cells were cultured with schistosomula at a 5,000:1 effector/target (E:T) ratio for 18 h at 37 degrees C. Unstimulated T cells killed 2.1 +/- 0.6% of schistosomula as judged by dye uptake and did not change their infectivity for mice. In contrast, PHA-stimulated T cells killed 41.3 +/- 3.1% of schistosomula by dye uptake and 56.7 +/- 7.7% of these organisms could not mature to adult worms in vivo. Killing was associated with and dependent on increased binding of PHA-stimulated T lymphocytes to schistosomula. Significant schistosomula killing first was noted after 2 h of exposure to T cells to PHA and peaked at 24; enhanced killing by PHA-stimulated cells was observed at an E:T ratio of 500:1 and was maximal at 5,000:1. Exposure of T lymphocytes to oxidative mitogens, soluble antigens, and alloantigens also resulted in enhanced killing of schistomula. These studies show that T lymphocytes activated by a variety of stimuli develop the capacity to kill schistosomula of Schistoma mansoni. Direct killing of infectious agents by cytotoxic T cells may contribute to host resistance to infections.

Induction of monocyte expression of tumor necrosis factor alpha by the 30-kD alpha antigen of Mycobacterium tuberculosis and synergism with fibronectin.

Native 30-kD antigen, also known as alpha antigen, is a fibronectin-binding protein that is secreted by live Mycobacterium tuberculosis. This antigen may play an important biological role in the host-parasite interaction since it elicits delayed type hypersensitivity response and protective immunity in vivo and T lymphocyte blastogenesis and IFN-gamma production in vitro. In the present study, we show that, TNF-alpha protein is produced in monocyte culture supernatants in response to 30-kD antigen and the level is as high as that to purified protein derivative of M. tuberculosis. This stimulatory effect was not due to contamination with either bacterial lipopolysaccharide or mycobacterial lipoarabinomannan. The preincubation of monocytes with plasma fibronectin significantly enhanced the release of TNF-alpha into the culture supernatants in response to 30-kD antigen. This effect was blocked by polygonal antibody to plasma fibronectin. In contrast, the monocytic cell line U937 failed to release TNF-alpha protein in the culture supernatants in response to 30-kD antigen with or without preincubation with plasma fibronectin. To determine whether this observation was due to differential binding of the 30-kD to fibronectin on these cells, a cell based ELISA was used. Pretreatment of monocytes with fibronectin enhanced their binding of the 30-kD antigen. U937 cells bound the 30-kD antigen weakly with or without fibronectin pretreatment. These results indicate that 30-kD antigen which is a known secretary antigen of M. tuberculosis is a stimulus for human monocytes to express TNF-alpha and that stimulatory effect may be mediated through plasma fibronectin.

Comparison of intermittent ethambutol with rifampicin-based regimens in HIV-infected adults with PTB, Kampala.

BACKGROUND: The human immunodeficiency virus (HIV) is a key factor responsible for the high rates of tuberculosis (TB) in sub-Saharan Africa. Treatment of TB with rifampicin (R, RMP) containing short-course regimens is highly effective in HIV-infected adults. We conducted a study to compare the efficacy and safety of intermittent ethambutol (E, EMB) with two RMP-containing regimens to treat pulmonary TB in HIV-infected patients. SETTING: National Tuberculosis Treatment Centre, Mulago Hospital, Kampala, Uganda. DESIGN: This was a prospective cohort compared to two non-randomised control groups. The study group and the two control arms were treated with 2 months of isoniazid (H), RMP, pyrazinamide (Z) and EMB followed by 6 E3H3 for the study group and 4HR or 6HR for controls. RESULTS: Between April 1993 and March 2000, 136 patients were enrolled in the 2EHRZ/E3H3 arm, 147 in the 2EHRZ/4HR arm and 266 in the 2EHRZ/6HR arm. The relapse rate was 18.2 per 100 person-years observation (PYO) for the study regimen compared to 9.7/100 PYO (P = 0.0063) and 4.8/100 PYO (P = 0.0001) in patients treated with 2 EHRZ/4HR or 2EHRZ/6HR, respectively. CONCLUSION: The 2EHRZ/6E3H3 regimen is safe and effective but has a significant risk of relapse.

Augmentation of natural killer cell activity by lipopolysaccharide through separable effects on the binding of nonadherent lymphocytes to tumor targets and tumor killing.

A single-cell assay was utilized to study the augmentation by Escherichia coli lipopolysaccharide (LPS) of the cytotoxicity of human lymphocytes for the human myeloid tumor K562. Preincubation with LPS at 20 micrograms/ml for 30 min at 37 degrees increased the binding of all nonadherent (NA) lymphocyte populations to K562 tumors [unseparated NA lymphocytes from 13.1 to 25.1%, immunoglobulin G Fc receptor-enriched lymphocytes from 27.6 to 42.9%, and immunoglobulin G Fc receptor-depleted lymphocytes from 14.0 to 23.7%, at p less than 0.001]. In contrast, interferon (IFN) at 10 units/ml had no effect on the overall binding of lymphocytes to K562 tumors. When lymphocyte-tumor conjugates were dispersed in agarose, cytotoxic activity of unseparated NA lymphocytes at 1 to 3 hr was markedly increased by preincubation with LPS (p less than 0.001). However, LPS did not enhance cytotoxicity if conjugates were formed in its absence. IFN, likewise, increased cytotoxic activity in unseparated NA lymphocytes at 1 to 3 hr (p less than 0.001). No synergistic cytotoxicity was seen with concurrent exposure to LPS and IFN. LPS increased cytotoxicity in the Fc receptor-enriched:tumor conjugates at 1 to 3 hr (p less than 0.001) and appeared to promote more efficient killing in individual conjugates over time. Cytotoxicity in the Fc receptor-depleted:tumor conjugates was not enhanced by LPS. Thus, LPS may enhance natural killer cell-like activity by increasing the binding of human lymphocytes to K562 tumors and by rearranging the population of binding cells to include more efficient killer cells. While the effects of LPS on binding appear independent of IFN, selective recruitment of more efficient killer cells could be through an IFN mechanism.

Cytokines and tuberculosis.

The evolution of Mycobacterium tuberculosis as an intracellular pathogen has led to a complex relationship between it and its host, the human mononuclear phagocyte. The products of M. tuberculosis-specific T lymphocytes are essential for macrophage activation for intracellular mycobacterial killing. However, some cytokines, including products of both lymphocytes and phagocytic cells, may contribute to enhanced mycobacterial survival and replication. In human immunodeficiency virus-associated tuberculosis, cytokine products may mediate enhanced susceptibility to tuberculosis as well as accelerated progression to AIDS. Better understanding of these interactions will allow the development of increasingly specific immune-based interventions for prevention and treatment of tuberculosis.

Impact of pulmonary tuberculosis on survival of HIV-infected adults: a prospective epidemiologic study in Uganda.

BACKGROUND: Retrospective cohort studies of tuberculosis suggest that active tuberculosis accelerates the progression of HIV infection. The validity of these findings has been questioned because of their retrospective design, diverse study populations, variable compliance with anti-tuberculous therapy and use of anti-retroviral medication. To assess the impact of tuberculosis on survival in HIV infection we performed a prospective study among HIV-infected Ugandan adults with and without tuberculosis. METHODS: In a prospective cohort study, 230 patients with HIV-associated tuberculosis and 442 HIV-infected subjects without tuberculosis were followed for a mean duration of 19 months for survival. To assess changes in viral load over 1 year, 20 pairs of tuberculosis cases and controls were selected and matched according to baseline CD4 lymphocyte count, age, sex and tuberculin skin test status. RESULTS: During the follow-up period, 63 out of of 230 tuberculosis cases (28%) died compared with 85 out of 442 controls (19%), with a crude risk ratio of 1.4 [95% confidence interval (CI), 1.07-1.87]. Most deaths occurred in patients with CD4 lymphocyte counts < 200 x 10(6) cells/l at baseline (n = 99) and occurred with similar frequency in the tuberculosis cases (46%) and the controls (44%). When the CD4 lymphocyte count was > 200 x 10(6)/l, however, the relative risk of death in HIV-associated tuberculosis was 2.1 (95% CI, 1.27-3.62) compared with subjects without tuberculosis. For subjects with a CD4 lymphocyte count > 200 x 10(6)/l, the 1-year survival proportion was slightly lower in the cases than in the controls (0.91 versus 0.96), but by 2 years the survival proportion was significantly lower in the cases than in the controls (0.84 versus 0.91; P < 0.02; log-rank test). For subjects with a CD4 lymphocyte count of 200 x 10(6) cells/l or fewer, the survival proportion at 1 year for the controls was lower than cases (0.59 versus 0.64), but this difference was not statistically significant (P = 0.53; logrank test). After adjusting for age, sex, tuberculin skin test status, CD4 lymphocyte count, and history of HIV-related infections, the overall relative hazard for death associated with tuberculosis was 1.81 (95% CI, 1.24-2.65). In a nested Cox regression model, the relative hazard for death was 3.0 (95% CI, 1.62-5.63) for subjects with CD4 lymphocyte counts > 200 x 10(6)/l and 1.5 (95% CI, 0.99-2.40) for subjects with a CD4 lymphocyte count of 200 x 10(6)/l or fewer. CONCLUSION: The findings from this prospective study indicate that active tuberculosis exerts its greatest effect on survival in the early stages of HIV infection, when there is a reserve capacity of the host immune response. These observations provide a theoretical basis for the treatment of latent tuberculous infection in HIV-infected persons.

Duration of efficacy of treatment of latent tuberculosis infection in HIV-infected adults.

BACKGROUND: Treatment of latent infection is needed to protect HIV-infected individuals against tuberculosis. A previous report addressed short-term efficacy of three regimens in HIV-infected adults. We now report on long-term efficacy of the study regimens. METHODS: Three daily self-administered regimens were compared in a randomized placebo-controlled trial in 2736 purified protein derivative (PPD)-positive and anergic HIV-infected adults. PPD-positive subjects were treated with isoniazid (INH) for 6 months (6H), INH plus rifampicin for 3 months (3HR), INH plus rifampicin and pyrazinamide for 3 months (3HRZ), or placebo for 6 months. Anergic subjects were randomized to 6H or placebo. RESULTS: 6H initially protected against tuberculosis in PPD-positive individuals; however, benefit was lost within the first year of treatment. Sustained benefit was observed in persons receiving 3HR and 3HRZ. In a Cox regression analysis, the adjusted relative risk for tuberculosis compared with placebo was 0.67 [95% confidence interval (CI), 0.42-1.07] for 6H, 0.49 (95% CI, 0.29-0.82) for 3HR, and 0.41 (95% CI, 0.22-0.76) for 3HRZ. When the rifampicin-containing regimens were combined, the adjusted relative risk for tuberculosis compared with placebo was 0.46 (95% CI, 0.29-0.71). Among anergic subjects, a modest degree of protection with 6H was present (adjusted relative risk, 0.61; 95% CI, 0.32-1.16). Treatment of latent tuberculosis infection had no effect on mortality. CONCLUSION: Six months of INH provided short-term protection against tuberculosis in PPD-positive HIV-infected adults. Three month regimens including INH plus rifampicin or INH, rifampicin and pyrazinamide provided sustained protection for up to 3 years.