Nathan Geffen, TAC

Taraz Samandari of the CDC presented the results of a randomised double-blind placebo-controlled  Isoniazid Preventative Therapy (IPT) trial at CROI. This trial took place in Botswana and randomised HIV-positive patients to either six months of IPT, the standard of care, or continuous therapy, ie the 36 month duration of the trial. [1]

Botswana began implementing a country-wide IPT programme in 2001. The protocol provided for six months of IPT and no Tuberculin skin testing. The rationale for this trial was that evidence has been emerging that the effects of 6-month IPT decrease with time.

For the first six months, the entire cohort –after screening– of 1,995 patients was initiated on open-label 300mg once-daily isoniazid. After six months patients were randomised to either receive placebo or continue receiving isoniazid. There were 989 in the six-month arm (6H) and 1,006 in the continuous one (36H). Patients also received 25mg vitamin B6 daily.

Patients were excluded from the trial if they had current weight loss or an AIDS-defining illness, a history of IPT, TB in the last three years, an abnormal chest x-ray, signs of TB based on a physical examination or the following laboratory criteria: Hgb<6.6gm/dl, neutrophil count <1,000 cells/mm3, platelets <75,000/mm3, AST(SGOT) >2.5 times the upper limit of normal, ALT (SGPT) >2.5 times upper limit of normal, total bilurubin >1.5 times the upper limit of normal, creatinine >1.5mg/dl and negative beta-HCG.

There were no significant differences between the 6H and 36H arms at baseline.  Females comprised 72% of the cohort. Median age was 32. Median CD4 count was approximately 300 cells/mm3. Just under 4% had a history of TB. In 6H, 23% of participants were TST-positive (>=5mm) and in 36H, this was 26%. By six months about 45% of the cohort had initiated ART.

Only 11 people were lost to follow-up. There were 176 withdrawals and 36 deaths. Adherence was good: at 36 months, 78% of participants had attended more than 80% of their visits (where pills were provided).  In a random sample of 200 people on 36H, 74% had detectable INH in urine at 25-30 months.

In intention-to-treat analysis, there were 34 incident TB cases in the 6H arm and 20 in the 36H arm (HR:  0.57, p=0.047). In a modified intention-to-treat of 1,655 people who were enrolled after six months there were 25 TB cases in the 6H arm and 12 on the IPT arm (HR: 0.47, p=0.033).  In a per protocol analysis of people who had attended least 5 out of 6 clinic visits (ie >=80% adherent), there were 19 and 8 TB cases in 6H and 36H respectively (HR=0.043, p=0.045).

In an analysis of TST-positive patients, the results were even more compelling. There were 13 incident TB cases in 6H and 4 in 36H (HR:0.26, p=0.019).  In the modified intention-to-treat it was 11 cases in 6H and 1 case in 36H (HR: 0.08, p=0.015) and in the per-protocol analysis 8 and zero cases respectively (HR: 0.0, p=0.007).  However, in TST-negative patients the reduction was slight and not significant (incidence rate of 0.98 versus 0.78 per 100 patient years; HR: 0.86; p=0.69).

The effect of 6 months IPT remained the same as the 36H arm for nearly six months after the completion of the open-label part of the study, but then TB cases escalated in the 6H arm.

ART initiation was evenly distributed between the two arms. Increasing ART use decreased TB incidence for all participants. By one year, the risk was reduced by half.

Compared to 6H patients not on ART:

• 6H TST-negative patients on ART had a TB hazard ratio of 0.5 (95%CI: 0.26-0.97).
• This was identical for 6H TST-positive patients.
• 36H TST-negative patients not on ART had a TB hazard ratio of 0.92  (95%CI: 0.4-2.1).
• 36H TST-positive patients not on ART had a TB hazard ratio of 0.08 (95%CI: 0.01-0.06).
• 36H TST-negative patients on ART had a TB hazard ratio of 0.46 (95%CI: 0.16-1.3).
• 36H TST-positive patients on ART had a TB hazard ratio of 0.04 (95%CI: 0.01-0.36).

After month 6, there were seven possibly drug-related adverse events in the 6H arm and 12 in the 36H arm. The difference was not significant. There was one death possibly due to isoniazid in the 36H arm (hepatic encephalopathy), though the patient also had severe herpes. There were 20 cases of severe hepatitis in the first six months (across both arms) and 15 subsequently (6 on 6H and 9 on 36H).

DST was available for 24 samples in the 6H arm and 14 in the 36H arm. There were three cases of INH mono-resistance in 6H and 1 in 36H. One patient on each arm had MDR TB.

Mortality was 1.4% per annum.  Interestingly it was non-significantly higher in the 36H arm (HR: 1.55; p=0.17). However, for TST-positives, the HR was 0.28 (p=0.06) and for TST-negatives, the HR was 2.99 and this was significant (p=0.01).

Samandari concluded that:

• The benefit of 6 months of IPT was lost in less than 6 months after treatment completion.
• Continuous IPT reduced TB incidence by 43% compared to 6 months of IPT.
• TST-positive patients were at high risk of TB after six months of IPT, even on ART.
• Continuous IPT was 92% effective in reducing TB in TST-positive patients.
• ART’s effect was smaller than continuous IPT in reducing TB, but additive.
• However, only ART reduced TB in TST-negative people.
• Provision of IPT did not increase isoniazid resistance.
• TST-negative patients may be unduly exposed to harm by IPT.

comment

This is one of the largest and best-conducted IPT studies to date. The finding that the benefit of IPT can only be maintained with continuous therapy is important. This study makes a strong case for scaled up IPT rollout. The new South African treatment guidelines provide 6 months IPT in all patients who are not indicated for ART.

However, the significantly higher mortality in TST-negative patients in the intervention arm is cause for concern. It suggests that Tuberculin skin testing must be done before initiating patients on IPT.

Ref: Samandari T et al. Randomized, placebo-controlled trial of 6 vs 36 months isoniazid TB preventive therapy for HIV-infected adults in Botswana. Oral abstract 104LB. 17th Conference on Retroviruses and Opportunistic Infections, 16-19 February 2010, San Francisco.

http://retroconference.org/2010/Abstracts/39555.htm

http://retroconference.org/2010/Sessions/027.htm