Optimal testing strategy for diagnosing latent tuberculosis infection and guiding isoniazid preventive therapy (IPT) is uncertain among HIV-infected patients. A 4-year prospective study was conducted among Thai HIV-infected patients who underwent simultaneous tuberculin skin test (TST) and QuantiFERON®-TB Gold In-tube Test (QFT-IT) at care entry. Based on baseline test results, patients were categorized into 4 groups: 1) QFT-IT-positive, TST-reactive 2) QFT-IT-positive, TST-non-reactive 3) QFT-IT-negative, TST-reactive, or 4) QFT-IT-negative, TST-non-reactive. The QFT-IT-positive patients were offered 9-month IPT and underwent yearly QFT-IT. Among 150 enrolled patients, there were 8, 12, 16 and 114 patients in group 1, 2, 3 and 4, respectively. Sixteen of 19 QFT-IT-positive patients (84%) completed IPT. The incidence of tuberculosis was significantly higher in patients who declined IPT than those completing IPT (11.11 vs. 0 case/100 patient-year; P<0.001). Among the 16 patients completing IPT, 11 (69%) and 2 (12%) had QFT-IT reversion at 1 year and 2 years after IPT, respectively. The remaining 3/16 (19%) had no reversion and their baseline interferon-γ levels were all above 1.2 IU/ml. Initial QFT-IT-guided IPT was effective in preventing tuberculosis. Serial QFT-IT for evaluating IPT effectiveness had limitations given delayed and no reversion, especially in those with high baseline interferon-γ levels.