Rifamycin-containing tuberculosis preventive treatment regimens had the lowest pooled proportions of discontinuation due to adverse events among patients receiving or about to start immunosuppressive medications, according to a systematic review and meta-analysis published in Clinical Infectious Diseases. However, evidence for these regimens was limited to patients who began preventive treatment before starting immunosuppressive medications.
The authors reported that discontinuation due to adverse events was broadly comparable to that reported in the wider tuberculosis preventive treatment (TPT) literature.
Jonathon R. Campbell, PhD, an assistant professor in the Department of Medicine at McGill University in Montreal, said the study addressed an evidence gap in decisions about TPT.
“Deciding whether to offer TPT requires weighing risks of progression to tuberculosis disease against risks of adverse events, and risks of progression vary widely depending on the immunosuppressive medication someone is taking,” Dr. Campbell said. “We had no systematic estimates of TPT safety in these populations to support these decisions.”
Discontinuation varied by regimen
Researchers searched MEDLINE, EMBASE, Cochrane and Health Star for studies published from January 1952 through May 1, 2025, without language restrictions. Eligible studies included at least 10 patients receiving a guideline-recommended preventive regimen and reported a safety outcome.
The review included 24 observational studies and three randomized trials involving 3,184 adults. Populations included patients receiving biologics, corticosteroids or disease-modifying antirheumatic drugs, as well as solid organ or hematopoietic transplant candidates and recipients.
The primary outcome was permanent treatment discontinuation due to an adverse event, excluding temporary interruptions. Researchers used random-effects generalized linear mixed models to pool proportions by regimen group.
Twenty-five studies involving 2,802 patients contributed to the primary analysis. Pooled discontinuation proportions were:
2% with three-month isoniazid-rifamycin combination regimens, involving 990 patients.
4.4% with four-month rifamycin monotherapy, involving 228 patients.
5.3% with six- to 12-month isoniazid monotherapy, involving 1,511 patients.
6.9% with six- to nine-month fluoroquinolone-based regimens, involving 73 patients.
Combination regimens included isoniazid with rifampin or rifapentine; rifamycin monotherapy included rifampin or rifabutin. Fluoroquinolone regimens used moxifloxacin or levofloxacin, but estimates were highly imprecise because of small sample sizes. Of the 990 patients receiving combination regimens, 973 received isoniazid-rifampin and 17 received isoniazid-rifapentine.
Timing and population analyses
Subgroup analyses by population type showed no large differences in discontinuation proportions compared with the overall estimates and did not appreciably reduce heterogeneity.
Among solid organ transplant recipients, pooled proportions of treatment discontinuation due to adverse events were 6.1% with isoniazid monotherapy, 1.3% with four-month rifamycin monotherapy and 6.9% with fluoroquinolone-based regimens. The authors noted that small sample sizes, particularly for rifamycin-based and fluoroquinolone-based regimens, contributed to imprecision.
For isoniazid monotherapy, pooled discontinuation was similar between patients receiving biologics, corticosteroids or disease-modifying antirheumatic drugs and solid organ transplant recipients: 5.9% vs. 6.1%.
All studies evaluating regimens other than isoniazid monotherapy involved patients who started TPT before initiating immunosuppressive medications. Only studies of isoniazid monotherapy allowed analysis by treatment timing: Pooled discontinuation due to adverse events was 6.4% among patients who started TPT before immunosuppression and 5.5% among those who started while receiving immunosuppressive medications. Across all regimens, studies did not clarify whether patients who began TPT first subsequently started immunosuppressive medications during the preventive treatment course or after completing it.
In secondary analyses, hepatotoxicity was the only specific adverse event with sufficient data for meta-analysis. Its pooled proportion was 5.9% across 20 studies of isoniazid monotherapy. Other adverse event types and severity were reported too infrequently or inconsistently for comparable analyses.
Evidence gaps limit interpretation
The authors noted several limitations, including small samples for certain regimens and populations, substantial heterogeneity and inconsistent adverse event measurement. No study evaluated one month of daily isoniazid-rifapentine. Neuropathy and gastrointestinal symptoms reported during TPT can also occur with immunosuppressive medications.
Seventeen of the 24 observational studies were rated as having serious or greater risk of bias due to confounding. Two of the three randomized trials had a high risk of bias in adverse event evaluation domains. The researchers also detected possible publication bias.
The authors called for studies evaluating regimens other than isoniazid monotherapy during immunosuppressive treatment and for improved adverse event reporting.
“We need well-designed cohorts that include people on immunosuppressive therapy and report clearly on the medications they receive, including when immunosuppressive drugs were started relative to TPT,” Dr. Campbell said. “Without knowing how long someone was exposed to both TPT and the immunosuppressive medication, interpreting safety appropriately becomes very challenging.”
The authors declared no conflicts of interest.