Factors influencing delay in initiating antiretroviral therapy among HIV infected patients coinfected with tuberculosis.

The optimum time to start antiretroviral therapy (ART) in a HIV infected patient receiving antituberculosis therapy (ATT) is unknown. 2 Concurrent treatment of tuberculosis (TB) and HIV is complicated by poor adherence from high pill burden, overlapping toxicities, pharmacokinetic drug–drug interactions and risk of immune reconstitution inflammatory syndrome (IRIS). The risk of disease progression if ART is delayed has to be balanced against the risk of needing to discontinue treatment if ART is started soon after initiating ATT. We examined factors influencing delay in initiation of ART in patients taking ATT and compared our practice against British HIV Association (BHIVA) HIV/TB treatment guidelines. We retrospectively reviewed the case notes of coinfected ART naive patients attending a central London HIV/TB service between January 1998 and October 2007. Timing of ART initiation and reason for delay (defined as initiation out-with 2 weeks of ATT, if CD4 ,100 cells/mL, after induction phase of ATT, if CD4 = 100– 200 cells/ml, after completion of ATT, if CD4 .200 cells/ml) was recorded. Reasons for delay were patient determined (erratic clinic attendance, chaotic lifestyle—substance/alcohol misuse—or patient refused/ declined ART), physician determined (patient had severe TB or intercurrent (opportunistic) infection, severe (grade III/ IV) drug toxicity from ATT, concern about IRIS) or non-clinically determined (non-availability of an interpreter in the clinic). Outcomes in coinfected patients presenting before the introduction of the guidelines (February 2005) were compared with those presenting subsequently. Ninety-seven ART naı̈ve patients were identified (see web repository online for details of patients). Fourteen of 97 patients were lost to our service before they were eligible to start ART, according to BHIVA guidelines; six unexpectedly transferred care to another centre, five were lost to followup, two died and one was dispersed to another area. Initiation of ART was within the guidelines in 20/83 (24%) of the remaining patients. Prior to the introduction of the guidelines, ART was delayed in 41/63 (65%) and in 22/34 (64.7%) subsequently (table 1). Two patients had IRIS; both had CD4 ,100/ml, neither delayed ART initiation. Four patients died; no deaths could be attributed to delay in initiating ART. Reasons for delayed ART initiation varied according to CD4 group. In patients with CD4 ,100 cells/ml, ATT toxicity and TB induced multiorgan failure were common reasons. In this situation it may be unsafe to start ART and delay was appropriate. Among patients with CD4 .200 cells/ml, reasons for delay were often non-clinical. ART was commonly not considered as CD4 counts were high. Many patients declined ART among all CD4 groups, accounting for 21% of delay overall. Crucially, delays in ART initiation did not impact on adverse clinical outcomes and mortality rates were low at 4.8% (95% confidence interval (CI) 1.3% to 11.9%), lower than reported in a retrospective UK multicentre study (8.5%, 95% CI 4.9% to 13.5%) and by Lawn et al from South Africa (10%, 95% CI 5.8% to 15.7%). Rates of IRIS in our retrospective study were low compared with other studies ; this may be a result of underreporting of mild cases. In this specialist HIV/TB treatment centre, there were often cogent barriers to ART initiation; however, the resultant delay did not adversely affect outcomes. Randomised controlled trials are needed, in both developed and resource poor environments, to determine the optimal timing of ART initiation in patients on ATT.