From Wiki Journal Club
Jump to: navigation, search
Sterling TR, et al. "Three months of rifapentine and isoniazid for latent tuberculosis infection.". The New England Journal of Medicine. 2011. 365(23):2155-66.
PubMedFull textPDF

Clinical Question

Among patients with latent TB, is a more convenient regimen of 3 months of rifapentine plus isoniazid noninferior to the standard-of-care 9 months of isoniazid?

Bottom Line

Weekly rifapentine plus isoniazid for 3 months is noninferior to daily single-agent isoniazid for 9 months in patients with latent TB.

Major Points

Isoniazid (INH) for 9 months under direct observation is a common treatment for latent TB. This regimen is 60-90% effective in reducing the rate of active TB, but compliance is suboptimal owing to the long duration of therapy and to its adverse effects, most notably hepatotoxicity. Many studies have thus evaluated alternative antibiotic regimens with shorter duration of therapy; these include 4 months of rifampin and 3 months of combined rifampin plus INH.

PREVENT TB randomized 7,731 patients with latent TB to either combination therapy with 3 months of rifapentine plus INH or to 9 months of INH alone, with both arms receiving directly observed therapy (DOT). Combination therapy was given once weekly, while monotherapy was administered daily. At a median follow-up of 2.5 years, PREVENT TB demonstrated the noninferiority of combination therapy: subjects receiving combination therapy developed active TB with similar rates as those receiving monotherapy (cumulative rate of 0.19% vs. 0.43% in the intention-to-treat analysis). This was attributed to fewer adverse drug reactions with combination therapy, shorter duration of therapy, and decreased loss to follow-up, which together led to an increased rate of combination therapy completion (82.1% vs. 69.0%).

The study was well received, but has several drawbacks. For example, this study enrolled predominantly subjects in the western world where rifamycin resistance is relatively uncommon; thus its results are not generalizable globally to countries that carry a greater burden of TB, where rifamycin resistance occurs frequently. In addition, few (2.6%) HIV-positive subjects were enrolled, which makes it difficult to comment on the durability of combination therapy among HIV-infected individuals who, prior studies have shown, develop TB at an alarming rate of 10% per year after discontinuation of anti-TB therapy.


CDC's 2011 treatment options include 12 dose, weekly, directly-observed treatment with rifapentine plus INH as an alternative therapy for non-pregnant, HIV-negative, patients >2 years of age who aren't suspected to be infected with resistant organisms[1]


  • Randomized, controlled noninferiority study
  • N=7,731 patients with latent TB
    • Isoinazid plus rifapentine (n=3,745)
    • Isoniazid alone (3,986)
  • Enrollment: 2001-2008
  • Setting: Brazil, Canada, Spain, and the US
  • Mean follow-up: 2.5 years
  • Analysis: Intention-to-treat, per-protocol
  • Primary endpoint: Culture-confirmed TB


Inclusion Criteria

  • Age ≥12 years (expanded to ≥2 years in 2005)
  • Close contact of patient with culture-confirmed TB (within 2 years before enrollment)
  • And one of:
    • Positive PPD;
    • HIV plus either positive PPD or close contact with patient with TB; or
    • Positive PPD with fibrotic changes on CXR

Exclusion Criteria

  • Confirmed or suspected TB
  • Resistance to INH or rifampin in the source case
  • Treatment with rifamycin or INH in prior 2 years
  • Prior completion of treatment for TB in HIV-negative persons
  • Sensitivity or intolerance to INH or rifamycin
  • Serum AST ≥5x ULN
  • Pregnancy or lactation
  • HIV therapy within 90 days after enrollment
  • Weight <10kg

Baseline Characteristics

Taken from the combined therapy group.


  • Age: 36 years
  • Males: 55.4%
  • White: 57.6%
  • Black: 24.5%
  • Asian or pacific islander: 12.4%
  • North american indian: 2.1%
  • Multiracial (in Brazil): 3.4%
  • US or Canada: 88.9%
  • Brazil or Spain: 11.1%
  • Completed high school: 56.9%

Reason for treatment:

  • Close contact with a patient with TB: 71.7%
  • Recent PPD conversion: 23.9%
  • Infection with HIV: 2.2%
  • Fibrosis on CXR: 2.2%

Medical characteristics:

  • HIV infection: 2.6%
  • Average BMI: 27 kg/m²
  • Previous incarceration: 5.5%
  • Unemployment: 10.6%
  • Self-reported alcohol use: 48.4%
  • Self-reported IV drug use: 3.7%
  • Homelessness: 5.9% INH vs. 7.4% combination (P<0.05)
  • Smoker: 27.9%
  • Hepatitis C: 2.5%
  • Hepatitis B: 1.1%


  • Randomized to either RIF+INH for 3 months or INH alone for 9 months
  • In members of a single household, randomization occurred only for the first person in the household; all other members of the household could receive the same treatment
  • Followed for 33 months, evaluated monthly
  • Adverse events reported up to 60 days following the last dose of study drug
  • After treatment completion, visits occurred every 3 months until 21st month, then months 27 and 33
  • Study was approved by the CDC


Comparisons are combination therapy vs. single-agent INH therapy.

Primary Outcomes

Rate of tuberculosis
Modified intention-to-treat analysis
0.19% vs. 0.43% (HR 0.38; 95% CI 0.15-0.99)
Per-protocol analysis
0.11 vs. 0.32 (upper limit of 95% CI for difference, 0.06%)

Secondary Outcomes

Therapy completion
82.1% vs. 69.0% (P<0.001)
Permanent drug discontinuation
17.9% vs. 31.0% (P<0.001)
Permanent drug continuation due to adverse event
4.9% vs. 3.7% (P=0.009)
0.8% vs. 1.0% (P=0.22)
Any serious adverse event
1.6% vs. 2.9% (P<0.001)
≥1 adverse event
14.7% vs. 17.6% (P<0.001)
Hepatotoxicity from study drug
0.4% vs. 2.7% (P<0.001)


  • Not possible to determine rates of resistance to either therapy arm due to small numbers of resistance that developed (one in combination arm, none in the single-agent INH arm)
  • Low incidence of HIV (approximately 2.6%) precludes generalization to HIV-positive cohorts


Supported by the CDC. Authors with multiple industry disclosures.

Further Reading

  1. CDC TB Treatment Recommendations