University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Management Recommendations > Initiating ART in Hospitalized Patients

Ward 86 Management Recommendations

Initiating Antiretroviral Therapy in Hospitalized HIV-Infected Patients

updated June 2017

Contributors: Monica Gandhi, MD
Katerina Christopolous, MD
Mark Jacobson, MD

  1. Initiation of antiretroviral therapy (ART) should be strongly considered during hospitalization of any HIV-infected patient not currently taking ART.

    1. Hospitalization is an opportunity to directly observe ART initiation, monitor early adverse effects, and collaborate with the patient in determining a diurnal schedule that will optimize adherence and strategies to deal with any medication side effects. We have found that initiation of ART in the hospital reduces patients' fears about ART and often translates into good outpatient regimen adherence.

    2. Substance abuse is not a contraindication for starting ART. We have had success in achieving full virologic suppression among many patients whose methamphetamine, cocaine, heroin, or alcohol abuse is not well controlled.

    3. Patients with Axis 1 psychiatric disorders, such as major depression, bipolar disease, and schizophrenia, also should not be excluded from initiation of ART. Many of these patients will have some degree of insight about their HIV disease, and some are highly motivated to take ART.

    4. Likewise, patients who are homeless or marginally housed should not be automatically excluded from ART initiation. High rates of complete viral suppression have been observed in homeless populations using single-pill-combination regimens. A special consideration for this population may be to choose a regimen that does not need to be taken with food or spaced from food.

  2. All patients diagnosed with an AIDS-defining opportunistic infection (OI) or serious bacterial infection should have ART initiated as soon as possible, with the following caveats.

    1. For patients with ocular opportunistic infections (OIs), particularly cytomegalovirus (CMV) retinitis: We recommend waiting until 14 days of OI antimicrobial therapy have been administered before initiating ART.

    2. For patients with cryptococcal meningitis: In San Francisco, we generally wait until 14 days of antifungal therapy have been administered before initiating ART. In collaborations with our international colleagues in resource-constrained settings, such as Kenya and Uganda, we delay initiating ART until 5 weeks of antifungal therapy have been administered.

    3. For patients with central nervous system (CNS) OIs other than cryptococcal or tuberculous meningitis: For those who have evidence of brain edema or mass effect by CT or MRI imaging, we recommend waiting until 14 days of OI antimicrobial therapy have been administered before initiating ART. Waiting for up to 14 days may also be considered in patients with a CNS OI who have a neurologic deficit.

    4. For patients who are intubated or are hypotensive as a result of a non-CNS/nonocular opportunistic or bacterial infection: We generally initiate ART after 5-7 days.

    5. For patients with newly diagnosed tuberculosis: Timing of ART initiation in patients is discussed in the section Treatment of Tuberculosis and Latent Mycobacterium tuberculosis Infection in HIV-Infected Patients.

    6. ART regimen choice: Because HIV genotype assays typically require 2 weeks to complete, initial selection of an ART regimen must be based on clinical considerations. After the HIV genotype results return, the ART regimen can be modified, if necessary, with little risk of reducing a patient's subsequent treatment options. In selecting an empiric regimen, consideration should be given to the prevalence of primary drug resistance mutations in the local population, anticipated toxicities and drug-drug interactions of various antiretrovirals and other medications the patient requires, concomitant illnesses, and the degree of treatment experience and former viral resistance patterns, if known. For most patients who have never been on ART (and patients who were on ART in the past but whose virus is not thought to have resistance to ARVs), the following regimens are reasonable. For patients with known or suspected viral resistance to ARVs, the regimen should be individualized.

      Because dolutegravir, ritonavir (or cobicistat)-boosted darunavir, and tenofovir (as either tenofovir disoproxil fumarate [TDF] or tenofovir alafenamide [TAF]) are the well-tolerated and high-barrier-to-resistance drugs in their respective classes, and because drugs with a low barrier to resistance can contribute to the emergence of cross-class mutations in the setting of poor adherence, the regimen we most commonly chose before knowing genotype results is dolutegravir plus TDF/emtricitabine (FTC) or TAF/FTC OR ritonavir (or cobicistat)-boosted darunavir plus TDF/FTC or TAF/FTC (see Table below for details). However, potential toxicities and drug-drug interactions with concomitant therapies always must be taken into consideration before choosing an empiric ART regimen. Moreover, recent data from CROI 2017 suggest that dolutegravir must have at least one more active drug in the regimen to prevent the relatively rapid emergence of resistance, which suggests that its genetic barrier to resistance might not quite be as high as that of a boosted PI.

      Anticipated toxicities and drug-drug interactions that most influence our choice between dolutegravir and darunavir are:

      1. Dolutegravir may cause neuropsychiatric side effects. If these occur, dosing at bedtime on an empty stomach may help since food can increase dolutegravir levels.
      2. Darunavir, boosted with either ritonavir or cobicistat, can cause diarrhea or aggravate existing diarrhea, and the boosters (which mainly inhibit CYP3A4) can inhibit metabolism of other medications. A drug-drug interaction check should be performed before prescribing boosted darunavir.

    Table. Preferred empiric regimens for initiating antiretroviral therapy BEFORE genotype results are available*
    • Dolutegravir 50 mg PLUS tenofovir disoproxil fumarate 300 mg/emtricitabine 200 mg fixed-dose combination (Truvada) OR tenofovir alafenamide 25 mg/emtricitabine 200 mg fixed-dose combination (Descovy)
    • Darunavir (Prezista) 800 mg daily and ritonavir (Norvir) 100 mg daily OR darunavir 800 mg/cobicistat 150 mg fixed-dose combination (Prezcobix) PLUS tenofovir disoproxil fumarate 300 mg/emtricitabine 200 mg fixed-dose combination (Truvada) OR tenofovir alafenamide 25 mg/emtricitabine 200 mg fixed-dose combination (Descovy)
    * For patients initiating ART for the first time, and patients restarting ART whose virus has no expected resistance to ARVs.
    transparent gif


  1. Although substance abuse, psychiatric disorders, and homelessness are obvious barriers to optimal medication adherence, studies have demonstrated that high rates of adherence can be achieved in patients with these issues, particularly when counseling or antidepressant therapy can be provided.(1,2,3,4) Also, the availability of better-tolerated, higher-barrier-to-resistance drugs in recent years has lowered the risk of resistance developing and has provided more treatment options for patients who do develop drug resistance as a result of poor adherence.

  2. The question of when to start ART in the setting of a new AIDS-defining OI has been controversial in the past. Some retrospective studies suggested a benefit with initiation of ART early during treatment of an OI, whereas others raised concerns that early ART initiation might increase patients' risk of developing immune reconstitution inflammatory syndrome (IRIS). The potential for overlapping toxicities, drug-drug interactions, unacceptably high pill burden, and inadequate time for adherence counseling led some providers to defer ART initiation for a month or more after beginning treatment for an AIDS-related OI. The issue is less controversial now with the availability of better-tolerated ART regimens and the results of a randomized, multicenter trial (AIDS Clinical Trials Group A5164) that provide compelling evidence that early ART initiation is beneficial. A5164 enrolled 282 patients who had an absolute CD4 count of <200 cells/µL and presented with an acute AIDS-related OI or serious bacterial infection.(5) Subjects were assigned to a strategy of early ART, started within 14 days of initiating acute infection treatment, versus deferred ART, started at least 4 weeks after initiating acute infection treatment. Those randomized to the early ART arm had fewer deaths or clinical events signifying AIDS progression than the patients who were randomized to the deferred ART arm (14% vs 24%; odds ratio [OR] 0.51; p = .035). Although the actual difference in median time to ART initiation between the immediate and deferred arms of this trial was only 33 days, the total number of new AIDS-related events and deaths over the next 48 weeks differed by nearly 50%. In generalizing the results of the A5164 trial to clinical practice, two important features of the trial should be considered. First, the study population consisted predominantly of patients with Pneumocystis pneumonia (63%), cryptococcal meningitis (12%), and AIDS-associated bacterial infections (12%). Patients with tuberculosis and those who were intubated were excluded. Second, the actual timing of when the immediate and deferred arms started ART was a median 12 vs 45 days, respectively, after initiation of therapy for the acute infection. Thus, the results argue compellingly for initiating ART within 2 weeks (not necessarily within 24 hours) of initiating treatment for a new OI or serious bacterial infection. Of note, 50% of the subjects in the early arm initiated ART 9-13 days after initiating OI therapy, and the trial lacked the power to detect whether those who initiated ART earlier than that had better outcomes. The incidence of IRIS did not differ significantly between the immediate and deferred arms (6% vs 9%).

    1. Timing of initiation of ART in patients with newly diagnosed CMV retinitis is still controversial. Only 6 subjects with CMV disease were randomized in the A5164 trial, too few for results to be meaningful. In one retrospective study, initiation of ART on the same day as treatment was begun for acute CMV retinitis was associated with a significantly higher rate of sight-threatening immune recovery uveitis than among historical control patients who waited 2 to 13 months (average 6.3 months) before initiating ART.(6) On the other hand, delay in initiating ART could conceivably increase the risk of further irreversible retinal necrosis in this disease. Other ocular OIs (eg, zoster or toxoplasmic retinitis) are too rare for there to be evidence upon which timing of ART initiation can be based. However, waiting 14 days seems prudent to us considering the risk that any increase in ocular inflammation, as has been observed in CMV retinitis, could cause permanent visual impairment.

    2. Two randomized trials conducted in sub-Saharan Africa (one in Zimbabwe, the other in Uganda and South Africa) that studied early versus delayed initiation of ART after starting antifungal therapy for cryptococcal meningitis found significantly increased mortality with early ART. The actual median intervals between the diagnosis of cryptococcal meningitis and initiating ART were <3 days vs >10 weeks in the Zimbabwe trial (n = 54) and 9 vs 36 days in the Uganda/South Africa trial (n = 177).(7,8) In the former trial, antifungal therapy was limited to fluconazole,(7) while in the latter, the initial 2 weeks of treatment was combined amphotericin and fluconazole.(8) Of note, 1-year mortality was very high in both of these trials, 50% and 30% in the deferred treatment arms (in which survival was better) of the Zimbabwe and Uganda/South Africa trials, respectively.

      Although the A5164 trial lacked sufficient power to detect a clinically significant difference in outcomes for the subset of patients with cryptococcal meningitis randomized to early versus delayed ART, the overall 1-year mortality among the 35 cryptococcal meningitis patients was only 6%, and there were no deaths among the 13 assigned to early treatment. This dramatic difference in mortality raises the question of how generalizable results from African trials are to U.S., European, or Australian patients who typically receive fungicidal flucytosine in combination with amphotericin during the first 2 weeks of treatment for cryptococcal meningitis and who may have other characteristics (eg, earlier diagnosis, better supportive in-hospital care) that could have led to better outcomes. Because half of the early ART initiation arm in the A5164 trial did not initiate ART until 9-13 days after beginning OI treatment, and because neither African trial investigated a 2-week deferred ART initiation arm (the deferred arms were 5-10 weeks), we believe that 14 days is currently the appropriate interval between initiating antifungal therapy for cryptococcal meningitis and initiating ART in settings that are not resource-limited, whereas 5 weeks appears to be the optimal interval for those in resource-limited settings who receive initial amphotericin therapy.

    3. We recommend waiting until 14 days of antimicrobial therapy have been administered for a CNS OI before initiating ART in patients with brain edema, mass effect, or neurologic deficit because of the potentially catastrophic effects that further brain inflammation could have. This is our opinion. There is no available relevant evidence.

    4. We recommend waiting until 5-7 days of antimicrobial therapy have been administered before initiating ART for patients who are intubated or are hypotensive as a result of a non-CNS/nonocular opportunistic or bacterial infection because of the potentially catastrophic effects that further systemic inflammation could have. This also is our opinion. There is no available relevant evidence.


  1. Springer SA, Dushaj A, Azar MM. The impact of DSM-IV mental disorders on adherence to combination antiretroviral therapy among adult persons living with HIV/AIDS: a systematic review. AIDS Behav. 2012 Nov;16(8):2119-43.
  2. Bangsberg DR, Ragland K, Monk A, et al. A single tablet regimen is associated with higher adherence and viral suppression than multiple tablet regimens in HIV+ homeless and marginally housed people. AIDS. 2010 Nov 27;24(18):2835-40.
  3. Tsai AC, Karasic DH, Hammer GP, et al. Directly observed antidepressant medication treatment and HIV outcomes among homeless and marginally housed HIV-positive adults: a randomized controlled trial. Am J Public Health. 2013 Feb;103(2):308-15.
  4. Cooperman NA, Heo M, Berg KM, et al. Impact of adherence counseling dose on antiretroviral adherence and HIV viral load among HIV-infected methadone maintained drug users. AIDS Care. 2012;24(7):828-35.
  5. Zolopa A, Andersen J, Powderly W, et al. Early antiretroviral therapy reduces AIDS progression/death in individuals with acute opportunistic infections: a multicenter randomized strategy trial. PLoS One. 2009;4(5):e5575.
  6. Ortega-Larrocea G, Espinosa E, Reyes-Terán G. Lower incidence and severity of cytomegalovirus-associated immune recovery uveitis in HIV-infected patients with delayed highly active antiretroviral therapy. AIDS. 2005 Apr 29;19(7):735-8.
  7. Makadzange AT, Ndhlovu CE, Takarinda K, et al. Early versus delayed initiation of antiretroviral therapy for concurrent HIV infection and cryptococcal meningitis in sub-Saharan Africa. Clin Infect Dis. 2010 Jun 1;50(11):1532-8.
  8. Boulware DR, Meya DB, Muzoora C, et al; COAT Trial Team. Timing of antiretroviral therapy after diagnosis of cryptococcal meningitis. N Engl J Med. 2014 Jun 26;370(26):2487-98.