8.0 Antiretroviral therapy in specific populations

For some patient populations, specific considerations need to be taken into account when deciding when to start and the choice of ART. The following sections outline specific recommendations and the supporting rationale for defined patient populations. In parallel to guidelines on ART in adults, BHIVA also publishes guidelines on the management and treatment of specific patient populations, including coinfection with TB, coinfection with viral hepatitis B or C, and HIV-positive pregnant women. An outline of the recommendations for when to start and choice of ART, from the BHIVA guidelines for TB and viral hepatitis is summarized below. The reader should refer to the full, published guidelines for these patient populations for more detailed information and guidance on the BHIVA website (http://www.bhiva.org/publishedandapproved.aspx) and be aware that BHIVA clinical practice guidelines are periodically updated.

For these current guidelines, new guidance on when to start and choice of ART has been developed for HIV-related cancers, HIV-associated NC impairment, CKD, CVD and women. The guidance only considers specific issues concerning the initiation and choice of ART in these patient populations. Guidance on the management of pregnancy in HIV-positive women has not been included.

8.1  HIV with tuberculosis coinfection

This guidance provides a brief summary of the key statements and recommendations regarding prescribing ART in HIV-positive patients co-infected with TB. It is based on the BHIVA guidelines for the treatment of TB/HIV coinfection 2011 [1], which should be consulted for further information. The full version of the guidelines is available on the BHIVA website (http://www.bhiva.org/TB-HIV2011.aspx).

8.1.1 When to start antiretroviral therapy Recommendations

Timing of initiation of ART during TB therapy:

CD4 cell count (cells/μL)When to start HAARTGrade
<100As soon as practical within 2 weeks after starting TB therapy1B
100–350As soon as practical, but can wait until after completing 2 months TB treatment, especially when there are difficulties with drug interactions, adherence and toxicities1B
>350At physician's discretion1B
Auditable measure. 

Proportion of patients with TB and CD4 cell count <100 cells/μL started on ART within 2 weeks of starting TB therapy. Rationale

Most patients with TB in the UK present with a low CD4 cell count, often <100 cells/μL. In such patients, ART improves survival, but can be complicated by IRD and drug toxicity. Data suggest that ART can be delayed until the first 2 months of TB therapy has been completed but at CD4 cell counts <50 cells/μL the short-term risk of developing further AIDS-defining events and death is high, and ART should be started as soon as practicable and within 2 weeks of initiation of TB therapy [2-5]. Starting ART early in severely immunosuppressed HIV-positive patients presenting with TB is associated with decreased mortality and a lowering of the rates of disease progression but rates of IRD are high.

Patients with HIV and a CD4 cell count >350 cells/μL have a low risk of HIV disease progression or death during the subsequent 6 months of TB treatment, depending on age and VL [6]. They should have their CD4 cell count monitored regularly and ART can be withheld during the short-course of TB treatment.

One study performed in HIV-associated TB meningitis in the developing world, where 90% of the patients were male, the majority drug users, many with advanced disease and the diagnosis being made clinically, showed no difference in mortality starting ART early or late [7].

8.1.2 What to start Recommendations

We recommend EFV in combination with TDF and FTC as first-line ART in TB/HIV coinfection1B
We recommend that when rifampicin is used with EFV in patients over 60 kg, the EFV dose is increased to 800 mg daily. Standard doses of EFV are recommended if the patient weighs <60 kg1C
We recommend that rifampicin is not used with either NVP or PI/r1C
We recommend that where effective ART necessitates the use of PI/r, that rifabutin is used instead of rifampicin1C
Auditable measure. 

Proportion of patients with active TB on anti-TB therapy started on ART containing EFV, TDF and FTC. Rationale

Preferred antiretroviral therapy. 

HIV-related TB should be treated with a regimen, including rifamycin for the full course of TB treatment, unless there is rifamycin resistance or intolerance. Rifamycins frequently interact with ARV medications and can lead to similar toxicities, notably rash and hepatitis. We recommend EFV as the preferred therapy for ART because of its confirmed potency when used in TB/HIV coinfection [8-10], and its efficacy in RCT. We recommend that EFV be given with TDF and FTC due to the availability of a once-daily co-formulation, a reduced risk of rash compared with NVP and improved efficacy at higher HIV VLs (commonly occurring in this setting). ABC-3TC is an alternative acceptable NRTI backbone in patients with lower HIV VLs and that are HLA-B*57:01 negative (see Section 5.3 Which NRTI backbone).

There is significant variability in the effect that rifampicin has on EFV concentrations because of liver enzyme induction, especially of CYP450 3A4 [8,11–13]. Subtherapeutic EFV concentrations may occur among patients who weigh more than 60 kg who are taking standard dose EFV together with rifampicin, and increasing the dose of EFV from 600 mg daily to 800 mg daily may be necessary; however, there is a risk of increasing adverse effects. A cohort study and a small RCT have shown that the standard adult EFV dose (600 mg daily) together with two NRTIs is well tolerated and was efficacious in achieving complete viral suppression among adults on concomitant rifampicin-based TB treatment, although the majority of patients were of low body weight [9, 10, 14].

In summary, we recommend that when EFV is used with rifampicin, and in patients over 60 kg, the EFV dose is increased to 800 mg daily. Standard doses of EFV are recommended if the patient weighs <60 kg. We suggest that TDM be performed at about the week of starting EFV if side effects occur and the dose adjusted accordingly.


NVP taken with TB treatment is complicated by pharmacokinetic interactions and by overlapping toxicities, especially skin rash and hepatitis. One study showed that patients co-infected with HIV and TB who initiated NVP-based ART during TB treatment had a nearly twofold higher risk of having a detectable HIV VL after 6 months compared with those taking NVP who did not have TB. However, those patients who were established on NVP at the time of initiation of TB treatment did not have a higher risk of HIV virological failure [11]. Using a higher maintenance dose of NVP (300 mg bd) to overcome drug interactions has been associated with higher rates of hepatotoxicity [15]. In one randomized trial comparing NVP 200 mg twice daily at initiation with EFV 600 mg once daily among patients with TB and HIV and CD4 cell counts <250 cells/μL, non-inferiority of NVP was not demonstrated compared with EFV [16].

Protease inhibitors. 

When co-administered with rifampicin, concentrations of standard-dose PIs are decreased below therapeutic targets and cannot, therefore be recommended [17-19]. Changing the dosing of PI/r has resulted in unacceptable rates of hepatotoxicity [20-22].

Rifabutin has little effect on the concentrations of PI/r but rifabutin concentrations are increased when the drug is taken together with PIs. Current recommendations are to give rifabutin at a dose of 150 mg thrice weekly to adults taking PI/r. Some data suggest that 150 mg once daily can be given to reduce the theoretical risk of rifamycin resistance due to subtherapeutic rifabutin concentrations, but this strategy may be associated with increased side effects [23-30].

Other drugs. 

There are few clinical data to support the use of newer NNRTIs, INIs and CCR5 receptor antagonists with rifampicin or rifabutin. We recommend that physicians review pharmacokinetic and other data summarized in the current BHIVA guidelines for treatment of TB/HIV coinfection [1].

8.1.3 References

8.2  HIV and viral hepatitis coinfection

8.2.1 Introduction

The following guidance provides a brief summary of the key statements and recommendations regarding prescribing ART in patients with HIV/hepatitis B and C coinfection. It is based on the BHIVA guidelines for the management of coinfection with HIV-1 and hepatitis B or C virus [1], which should be consulted for further information and to the BHIVA website for latest updates (http://www.bhiva.org/publishedandapproved.aspx).

Where viral hepatitis B or C chronic infection has been diagnosed, individuals should be referred to a dedicated coinfection clinic or hepatologist for appropriate staging of their hepatitis B or C infection. The decision to initiate treatment for either HIV or viral hepatitis infections should ordinarily be made with agreement of the patient's HIV and viral hepatitis physicians.

In patients with cirrhosis (Child–Pugh grade B/C) certain ART should be used with caution and careful monitoring (including TDM) will be required by physicians experienced in the management of HIV and viral hepatitis coinfection. For further information on use of ART in patients with cirrhosis please refer to the BHIVA guidelines for the management of coinfection with HIV-1 and hepatitis B or C virus [1]. Summary of when to start recommendations

CD4 cell count (cells/μL)HBV requiring treatmenta HBV not requiring treatmentHCV with immediate plan to start HCV treatmenta HCV with no immediate plan to start HCV treatment
  1. aSee BHIVA guidelines for the management of coinfection with HIV-1 and hepatitis B or C virus [1] for indications to treat hepatitis B and C.
>500Start ART in some patients (2C) (include TDF and FTC)Defer ARTDefer ARTDefer ART
≤500Start ART (1B) (include TDF and FTC)Start ART (1B) (include TDF and FTC)
  • 350–500 cells/μL:
    • Start ART after HCV treatment commenced (1C)
  • <350 cells/μL:
    • Start ART before HCV treatment (1B)
    • Discuss with HIV and viral hepatitis specialist
Start ART (1C)

8.2.2 Hepatitis B When to start antiretroviral therapy

  • We recommend patients with HIV and HBV coinfection who have a CD4 cell count between 350 and 500 cells/μL start ART (1C).
  • We suggest patients with HIV and HBV coinfection who have a CD4 cell count >500 cells/μL and who require treatment for their hepatitis B start ART (2C).
Auditable measure. 

Proportion of patients with HIV and HBV coinfection with CD4 cell counts <500 cells/μL on ART.


Because of the negative effect of immune depletion on HBV disease progression, the availability of single drugs with high-level dual hepatitis B and HIV antiviral activity, and the increased risk of liver-related deaths in patients with CD4 cell counts below 500 cells/μL, co-infected patients with CD4 cell counts between 350 and 500 cells/μL should start ART and be treated with drugs active at suppressing both viruses [2]. Consideration can be given to some patients with CD4 cell counts between 350 and 500 cells/μL and HBV DNA of <2000 IU/L and no evidence of liver inflammation or fibrosis to close monitoring of their HIV and hepatitis B infections as an acceptable alternative strategy.

Individuals with a CD4 cell count >500 cells/μL who do not require hepatitis B therapy, should be monitored for HIV and hepatitis B disease progression and the need of therapy for either virus infection. Among individuals with a CD4 cell count >500 cells/μL who require treatment for hepatitis B infection there is the option to start ART with drugs active at suppressing both viruses. For indications to start treatment for hepatitis B infection, please refer to BHIVA guidelines on management of coinfection with HIV and hepatitis B or C virus [1]. What to start

  • We recommend patients with HIV and HBV coinfection who start ART include TDF and FTC as part of their ART regimen, if there are no contraindications for either drug (1A).
Auditable measure. 

Proportion of patients with HIV and HBV coinfection starting TDF and FTC as part of their first ART regimen.


TDF, FTC and 3TC are agents that have antiviral activity against both HIV and hepatitis B. The efficacy of these drugs against hepatitis B has been assessed in randomized trials extending out to 5 years in mono-infected patients [3]. They are recommended agents in these guidelines for the treatment of HIV-1 infection.

All hepatitis B coinfected individuals who start ART should commence a regimen containing TDF and FTC. Hepatitis B treatment options for patients declining ART are discussed elsewhere [1].

If an individual becomes intolerant or is unable to commence a TDF-containing regimen, TDF should be substituted with either adefovir or entecavir and an alternate ARV agent added to the regimen. No individual coinfected with hepatitis B should receive a regimen containing 3TC or FTC monotherapy as its use may result in the selection of the YMDD mutation [4, 5]. HBV resistance to TDF is rare and combination with 3TC and FTC has been demonstrated to be effective at suppressing HBV DNA and may induce hepatitis B e antigen seroconversion, and may reduce the risk of HBV breakthrough [6].

In individuals virologically failing hepatitis B therapy, a resistance test should be taken and new therapy for HIV and hepatitis B commenced only after close consultation with a specialist virologist or specialists in the HIV/viral hepatitis coinfection clinic. Co-infected individuals who need to start a new ART regimen for reasons such as ART virological failure should ensure that effective anti-hepatitis B therapy is continued in addition to their new ART regimen. Abrupt withdrawal of effective treatment may lead to a flare in hepatitis B replication with liver damage. This may be particularly severe in patients with cirrhosis.

8.2.3 Hepatitis C When to start antiretroviral therapy

  • We recommend patients with HIV and HCV coinfection be assessed for HCV treatment (GPP).
  • We recommend patients with HIV and HCV coinfection and CD4 cell count between 350 and 500 cells/μL start ART (i) immediately if HCV treatment is deferred, and (ii) after initiation of HCV treatment if this is started immediately (1C).
  • We recommend patients with HIV and HCV coinfection and CD4 cell count <350 cells/μL start ART before HCV treatment (1B).
Auditable measure. 

Proportion of patients with HIV and HCV coinfection and CD4 cell counts <500 cells/μL on ART.


HCV is believed to have a deleterious effect on HIV disease progression [7, 8]. In addition, HIV has an impact on hepatitis C infection. The rate of liver fibrosis progression is faster in HIV/HCV co-infected patients particularly among patients with low CD4 cell counts [9-11]. The estimated risk of cirrhosis was twofold higher in individuals with HIV/HCV coinfection compared with those with HCV mono-infection [12]. Liver mortality rates are reportedly higher in those with a low CD4 cell count [13] and hepatocellular carcinoma is believed to occur at a younger age and within a shorter time [14]. All patients with chronic hepatitis C should be assessed for their HCV-related liver disease as well as for their stage of HIV infection.

Cohort studies examining the effect of ART on the natural history of HCV infection have shown inconsistent results [12, 15]. A few studies have concluded that HIV VL, but not CD4 cell count, was directly related to fibrosis progression rate [16], a finding consistent with the role of HIV VL both as a predictor of AIDS survival and as a predictor of survival in HCV/HIV co-infected individuals [17, 18] and in HCV/HIV co-infected liver transplant recipients [19]. ART is not associated with serious histological liver disease [20].

For these reasons, patients with HIV and hepatitis C infection with CD4 cell counts <500 cells/μL should start ART. This should be immediate if (i) CD4 cell count is <350 cells/μL, irrespective of whether HCV treatment is planned or not, and (ii) CD4 cell count is between 350 and 500 cells/μL and treatment for HCV has been deferred. For patients with CD4 cell counts between 350 and 500 cells/μL starting HCV treatment immediately, initiation of ART should be delayed until after the start of HCV treatment. Individual factors will determine the timing of ART after HCV treatment is commenced.

Individuals with a CD4 cell count >500 cells/μL who defer hepatitis C therapy, should be monitored closely for HIV or hepatitis C disease progression and the need for therapy for either virus. What to start

  • We recommend that potential pharmacokinetic interactions between ARVs and anti-hepatitis agents are checked before administration (with tools such as: http://www.hep-druginteractions.org) (GPP).
Auditable measure. 

Record in patient's notes of potential pharmacokinetic interactions between ARVs and anti-HCV agents.


Significant pharmacokinetic and pharmacodynamic interactions have been reported between ARV drugs and the newer anti-hepatitis agents. Boceprevir and telaprevir undergo extensive hepatic metabolism; boceprevir primarily by way of the aldoketoreductase system but also by the CYP450 enzyme system, whereas telaprevir is metabolized only by the CYP450 enzyme system, and the main route of elimination is via the faeces with minimal urinary excretion. Both boceprevir and telaprevir are potent CYP450 inhibitors. Therefore, DDIs are likely when used together with ARV drugs. Currently, studies have been completed for TDF, EFV, ATV/r and RAL with telaprevir and for TDF, DRV/r, LPV/r, ATV/r, EFV and RAL for boceprevir [21-26]. Other DDI studies are planned and currently information is available at http://www.hep-druginteractions.org.

Owing to the rapidly emerging data on the use of these newer agents and complexities of the drug interactions, we suggest that treatment of HCV infection in HCV/HIV co-infected patients should be carried out as part of a clinical trial. If a suitable clinical trial is not available, such treatment should only be carried out by physicians who have experience with the new HCV PIs and/or directly acting agents. Extreme care should be given to DDIs in patients who are taking HIV treatment. HIV treatment should be switched to agents where DDIs have been studied.

8.2.4 References

8.3  HIV-related cancers

8.3.1 Summary of auditable measures

Proportion of patients with an AIDS-defining malignancy on ART.

Proportion of patients with a non-AIDS-defining malignancy on ART.

Record in patient's notes of potential pharmacokinetic drug interactions between ARVs and systemic anticancer therapy.

8.3.2 When to start antiretroviral therapy: AIDS-defining malignancies

KS, high-grade B-cell NHL and invasive cervical cancer are all AIDS-defining illnesses and are thus indications to commence ART regardless of CD4 cell count or HIV VL. Kaposi sarcoma

  • We recommend starting ART in HIV-positive patients with KS (1A).

ART has been shown to reduce the incidence of KS in HIV cohort studies [1-4], to prevent KS in patients on ART [3], and, in addition, increases the time to disease progression in KS [5], improves prognosis in KS and prolongs survival in KS [6-8]. When initiating ART for KS, there appears to be no difference in response or outcome of KS between different HIV treatment regimens [3, 9]. Therefore, no recommendation can be made on choice of HIV therapy for patients with KS. Non-Hodgkin lymphoma

  • We recommend starting ART in HIV-positive patients with NHL (1B).

ART has been shown to reduce the incidence of NHL [1, 2, 10-18] and to improve the outcome [8, 19-22]. Before ART was available, the treatment of NHL with standard doses of chemotherapy produced marked toxicity and a high incidence of opportunistic infections [23]. In an attempt to decrease toxicity, modified-dose chemotherapy regimens were used by the AIDS Clinical Trials Group (ACTG). However, the reduced opportunistic infections were offset by the lower response rates [24]. Since the widespread availability of ART, two retrospective studies reported higher tumour response rates and overall survival in HIV seropositive patients with systemic NHL who were treated with CHOP chemotherapy and concomitant ART compared with those who were treated with CHOP alone [19, 20]. Similarly, in a separate study of liposomal doxorubicin in combination with cyclophosphamide, vincristine and prednisolone in HIV-associated NHL, improvement in survival was associated with HIV viral control, although complete remission rates were independent of HIV VL [25].

Further evidence to support the use of ART with chemotherapy in both KS and NHL is the finding from historical comparisons that the fall in CD4 cell count during chemotherapy is less profound when ART is prescribed concomitantly and that the duration of lymphocyte subset suppression is briefer [4, 26-28].

However, a number of US intergroup studies have either withheld ART during chemotherapy [29, 30] or delayed the initiation of ART [31]. The rationale for this approach includes avoiding adverse pharmacokinetic and pharmacodynamic interactions between ART and chemotherapy and the theoretical concern that PIs may inhibit lymphocyte apoptosis and thus contribute to chemoresistance of lymphomas [32]. Although no new HIV mutations were identified, these studies were small and ART was promptly reinstituted after abbreviated chemotherapy. Nevertheless, it took 12–18 months after completing chemotherapy for plasma HIV viraemia to become undetectable in many patients [30]. Importantly, patients with NHL frequently present with CD4 cell counts <200 cells/μL and thus the reduction in CD4 cell count associated with systemic chemotherapy and structured suspension of ART is not ideal. Cervical cancer

  • We suggest starting ART in HIV-positive patients with cervical cancer (2C).
  • We recommend starting ART in HIV-positive patients who are commencing radiotherapy or chemotherapy for cervical cancer (1D).

There is less clear evidence to support starting ART in women diagnosed with invasive cervical cancer, despite its status as an AIDS-defining illness. Co-registration studies have shown that ART has not reduced the incidence of cervical cancer [33-35], moreover the effects of ART on pre-invasive cervical dysplasia have been variable with some studies suggesting that ART causes regression of cervical intraepithelial neoplasia [36-42] and others showing no beneficial effect of ART [43-46]. The effects of ART on outcomes in HIV-positive women with invasive cervical cancer have not been reported but analogies with anal cancer may be drawn as the malignancies share common pathogenesis and treatment modalities. Combined chemoradiotherapy in anal cancer has been shown to cause significant and prolonged CD4 suppression even when ART is administered concomitantly [47-50]. Similarly the toxicity of chemoradiotherapy for HIV-associated anal cancer appears to be less profound among patients given ART compared to historical controls [48, 49, 51-56].

8.3.3 When to start antiretroviral therapy: non-AIDS-defining malignancies Recommendation

  • We suggest starting ART in HIV-positive patients with non-AIDS-defining malignancies (2C).
  • We recommend starting ART in HIV-positive patients who are commencing immunosuppressive radiotherapy or chemotherapy for non-AIDS-defining malignancies (1C). Rationale

While ART has little effect on the incidence of NADMs [2, 57-64] and there is no evidence that ART alone causes regression of NADMs, the immunosuppressive effects of both chemotherapy [4, 26-28] and radiotherapy [47-50] may justify starting ART in HIV-positive individuals who are commencing systemic anticancer therapy or radiotherapy.

8.3.4 What to start Recommendation

  • We recommend that potential pharmacokinetic interactions between ARVs and systemic anticancer therapy are checked before administration (with tools such as: http://www.hiv-druginteractions.org) (GPP). Rationale

Significant pharmacokinetic and pharmacodynamic interactions have been reported between ARV drugs and systemic anticancer therapies. The mechanisms of the pharmacokinetic interactions include the inhibition and induction by ARV agents of enzymes, especially the CYP450 family and uridine diphosphoglucuronosyl transferase isoenzymes, involved in the catabolism and activation of cytotoxic chemotherapy agents. In addition, competition for renal clearance, intracellular phosphorylation and ABC (ATP-binding cassette) transporters, has been hypothesized to contribute to these drug interactions [65]. Similarly, pharmacodynamic interactions, in particular overlapping toxicities between ARVs and systemic anticancer therapy, suggest that some drug combinations should be avoided in patients with HIV-associated cancers. Much of the guidance on the use of individual ARV agents with systemic anticancer therapy comes from reviews of potential drug interactions rather than from clinical studies [65-67]. The pharmacokinetic interactions between ARVs and systemic anticancer therapy are not confined to cytotoxic chemotherapy agents and extensive interactions with newer targeted therapies such as imatinib, erlotinib, sorafenib, bortezomib and temsirolimus have been described [67]. Recommendation

  • We suggest avoiding ritonavir-boosted ART in HIV-positive patients who are to receive cytotoxic chemotherapy agents that are metabolized by the CYP450 enzyme system (2C). Rationale

In general, clinically important pharmacokinetic drug interactions with systemic anticancer therapies are most common with PI/r-based ART and most clinicians avoid these combinations where possible. For example, in a cohort study, the rates of severe infections and severe neutropenia following chemotherapy for AIDS-related NHL were significantly higher among patients receiving concomitant PI (mainly ritonavir boosted) than in those on NNRTI-based ART regimens, although there was no difference in survival between the groups [68]. Furthermore, case reports of clinically significant life-threatening interactions between ritonavir-boosted-based ART and docetaxel [69], irinotecan [70] and vinblastine [71] have been published. Recommendation

  • We recommend against the use of ATV in HIV-positive patients who are to receive irinotecan (1C). Rationale

The camptothecin cytotoxic agent irinotecan is extensively metabolized by uridine diphosphoglucuronosyl transferase 1A1 isoenzymes that are inhibited by ATV [72]. In patients with Gilbert's syndrome, who have a congenital deficiency of uridine diphosphoglucuronosyl transferase 1A1, irinotecan administration has led to life-threatening toxicity [73]. Recommendation

  • We suggest avoiding ARV agents in HIV-positive patients who are to receive cytotoxic chemotherapy agents that have overlapping toxicities (2C). Rationale

Both ARV agents and systemic anticancer therapies have substantial toxicity and where these overlap it is likely that the risk of toxicity is greater. For example, ZDV commonly causes myelosuppression and anaemia [74], which are also frequent side effects of cytotoxic chemotherapy and so these should not be co-prescribed where possible. Similarly, dideoxynucleosides cause peripheral neuropathy [75], a common toxicity of taxanes and vinca alkaloids, so co-prescribing should be avoided. Both ZDV and dideoxynucleosides are no longer recommended for initiation of ART but some treatment-experienced patients may still be receiving these drugs and alternatives should be considered.

8.3.5 References

8.4  HIV-associated neurocognitive impairment

8.4.1 Introduction

With the widespread use of effective combination ART, the incidence of severe HIV-associated cerebral disease has declined dramatically [1]; however, more subtle forms of brain disease, known as HIV-associated NC disorders are reported to remain prevalent [2]. This NC deficit may present with a wide spectrum of clinical symptoms, but typically includes patterns involving ineffective learning and problems with executive function, rather than pure difficulties in formulating new memory (the cortical defect typical of Alzheimer's disease [3]).

Given the changing picture of this disease, a revised nomenclature system has been proposed classifying subjects with abnormal neuropsychological testing results in to three categories based on patient's symptoms, measured via the activities of daily living scale [2]. Subjects with abnormal neuropsychiatric testing results, who are otherwise asymptomatic, are classified as having HIV-associated asymptomatic NC impairment; those who are mildly symptomatic are classified as having HIV-associated mild NC disorder; and those who are severely symptomatic are classified as having HIV-associated dementia. The clinical relevance of asymptomatic NC impairment, namely asymptomatic subjects with abnormal results on neuropsychological testing, remains unclear.

Reports describing rates of NC impairment vary with some groups describing that up to 50% of HIV-positive subjects meet the above diagnostic criteria [4]. However, such reports should be interpreted with caution as asymptomatic subjects are often included and not all reports correct for effective ARV use. A Swiss cohort has reported 19% of aviraemic HIV-positive subjects meet the classification for mild NC disorder or above [5].

Risk factors for the development of NC disorders are poorly understood and are likely to be multifactorial, including both HIV disease factors [6] and concomitant diseases [7]. Although it is possible the choice of combination ART a subject receives may influence NC function, this is a controversial area without definitive evidence. The following recommendations apply to patients with symptomatic HIV-associated NC disorders.

8.4.2 When to start antiretroviral therapy Recommendation

  • We recommend patients with symptomatic HIV-associated NC disorders start ART irrespective of CD4 lymphocyte count (1C).
Auditable measure. 

Proportion of patients with symptomatic HIV-associated NC disorders on ART. Rationale

Current evidence suggests NC function improves after commencing ART for the first time [8] in both cognitively symptomatic [9] and asymptomatic [10] subjects. However, these studies have been undertaken in individuals with other indications to commence ART, in general with CD4 lymphocyte counts in the designated range where treatment is recommended. For subjects with higher CD4 lymphocyte counts, the ongoing START study will prospectively assess NC function in HIV-positive subjects commencing ART at an earlier stage of HIV disease.

Therefore, ART is recommended in NC symptomatic subjects whose CD4 lymphocyte count itself is an indication to commence therapy.

In the absence of scientific data, in cognitively symptomatic subjects with higher CD4 lymphocyte counts in whom ART would not be otherwise indicated, a recommendation to consider commencing ART is based (i) on observed improvements in cognitive function reported in subjects with lower CD4 lymphocyte counts commencing therapy [8], and (ii) to avoid a future decline in CD4 lymphocyte count in such subjects, given the well-described association between low nadir CD4 lymphocyte count and NC impairment [6].

Suboptimal adherence to therapy may occur more frequently in subjects with NC impairment, hence adequate support services to optimize adherence are essential.

8.4.3 What to start Recommendation

  • We recommend patients with HIV-associated NC disorders start standard combination ART regimens (1C).
Auditable measure. 

Proportion of patients with HIV-associated NC disorders on ART containing two NRTIs and one of an NNRTI, a PI/r or an INI. Rationale

Although during the earlier years of ART, clear benefits on cerebral function of individual ARV drugs such as ZDV were reported [11] and the benefits of combination therapy overall are well described [8], data are sparse regarding any differences in these benefits between individual agents or combinations. Within cohort studies, the use of the NRTI class within ARV regimens has been associated with a reduced risk of severe HIV-associated dementia [12] compared with the use of other regimens; however, the confounders of a cohort study limit interpretation of these data.

Recently, attempts have been made to establish a relationship between cognitive function and CNS ARV drug delivery based on an ARV scoring system known as the clinical penetration effectiveness (CPE) score [13]. The CPE score aims to rationally score the cerebral effects of individual ARV agents. However, the system is predominantly designed around pharmacokinetic modelling rather than pharmacodynamic endpoints such as data describing changes in NC function. Studies that have assessed the correlation between the CPE scores of ARV regimens with NC function report conflicting findings with some cohorts reporting a positive association [14, 15], and others describing a negative association [16]. Given the potential flaws outlined in the design of the CPE score, a lack of prospective clinical data and discrepancies in findings within cohort studies, the CPE score should not influence therapeutic decisions in subjects with NC impairment commencing ART.

One small prospective study has assessed the cerebral effects of three different ARV regimens in neurologically asymptomatic subjects reporting greater improvement in NC function in subjects commencing a quadruple nucleoside regimen compared with an EFV- or ATV/r-containing regimen [17]. However, subjects were asymptomatic from a neurological point of view, limiting the relevance of these findings to neurologically symptomatic subjects.

The improvements in NC function observed with ZDV monotherapy [11] and the greater improvements in NC function observed with a ZDV-containing quadruple nucleoside regimen compared with other ART regimens [17], raise the possibility of selecting a ZDV-containing ARV regimen in subjects with NC impairment. Conversely, a lack of comparator data for ZDV monotherapy and potential toxicities arising from ZDV use may limit the relevance of these data. Of note, further to peripheral toxicities, which are well described with ZDV use, biomarker data suggest there may also be CNS toxicities associated with the use of ZDV-containing regimens [18].

In summary, we recommend patients with NC impairment start standard combination ART regimens and the choice should be determined, as with other patients, by different factors, including baseline VL, side effect profile, tolerability, DDIs and patient preference. Novel antiretroviral strategies and NC function

Novel ARV strategies, including protease-inhibitor monotherapy continue to be assessed in clinical trials as cost-beneficial treatment regimens with the potential for reduced long-term toxicities. Concerns have been raised regarding the cerebral effects of PI monotherapy [19], with such concerns based on the hypotheses that PI monotherapy comprises only one effective ARV agent that may not adequately suppress ongoing HIV replication in sanctuary sites such as the CNS, and on pharmacokinetic modelling that suggests that not all PIs have optimal penetration across the blood–brain barrier [13]. Furthermore, isolated cases describing the evolution of CNS disease in previously stable HIV-positive subjects receiving PI monotherapy have been reported [20].

One study was specifically designed to assess the cerebral effects of LPV/r monotherapy [21]; however, it was terminated early due to a lack of efficacy in the plasma compartment. Although cases of CNS disease were reported within this study, such results must be interpreted with caution as virological endpoints in the plasma compartment were not met and therefore such cases may be driven by poor ARV efficacy per se, rather than distinct CNS disease itself [22].

In the MONET study assessing DRV/r vs. standard therapy, no differences in patient-reported cognitive function are observed between the study treatment arms over 3 years of therapy [23]. Although reassuring, these data represent changes in patient-reported observations rather than observations from formal neuropsychological testing. Interestingly, in a small substudy within MONET, improvements in detailed neuropsychological testing and improvements in cerebral biomarkers measured via imaging techniques, were reported in both treatment arms [24].

In the ongoing UK PIVOT study, detailed neuropsychological testing is being assessed prospectively in subjects on PI monotherapy vs. standard therapy, the results of which will be of great interest to this field.

Given the above theoretical concerns regarding the CNS activity of PI monotherapy, and for the majority of HIV-positive subjects it may be possible to select other ARV regimens, we suggest this approach is currently avoided in neurologically symptomatic subjects.

8.4.4 Modification of antiretroviral therapy Recommendation

In patients with ongoing or worsening NC impairment despite ART, we recommend the following best practice management (GPP):

  • Reassessment for confounding conditions.
  • Assessment of CSF HIV RNA, CSF HIV genotropism and genotyping of CSF HIV RNA.
  • In subjects with detectable CSF HIV RNA, modifications to ART should be based on plasma and CSF genotypic and genotropism results. Rationale

Several published randomized controlled studies, assessing both intensification of ART with a new ARV agent [25] and with adjunctive therapies [26-29] have been published. Unfortunately, none of these studies describe improvements in cognition subsequent to the study interventions. Without evidence-based interventions, the Writing Group outlines below a best practice approach based on the current literature.

As HIV-associated NC disorders are a diagnosis of exclusion, re-evaluation of subjects with ongoing NC impairment despite ART for confounding conditions, with expert input from other clinical specialties such as psychiatry, neurology and neuropsychology, is recommended and, where possible, input from an HIV neurology service.

Assessment of CSF HIV RNA, CSF HIV genotropism and genotypic analysis of CSF RNA may be useful tools in the management of subjects with ongoing NC for the following reasons. First, data from cohorts of untreated HIV-positive subjects would suggest CSF HIV RNA to be greater in subjects with HIV-associated dementia and cognitive decline [30, 31] and therefore suppression of CSF HIV RNA may be beneficial for cognitive function. Secondly, in subjects with ongoing NC impairment, higher degrees of genetic diversity between HIV viral strains in the CSF and plasma compartment may exist [32], even in subjects with undetectable plasma HIV RNA [33]. Therefore, assessment for CSF HIV resistance may be worthwhile to tailor ART.

8.4.5 References

8.5 Chronic kidney disease

8.5.1 When to start antiretroviral therapy Recommendation

  • We recommend patients with HIVAN start ART immediately irrespective of CD4 cell count (1C).
  • We recommend patients with end-stage kidney disease who are suitable candidates for renal transplantation start ART irrespective of CD4 cell count (1C).
Auditable measure. 

Proportion of patients with HIVAN started on ART within 2 weeks of diagnosis of CKD. Rationale

The use of ART has been associated with a decline in the incidence of HIVAN in HIV cohort studies [1], with renal histological improvement in case reports [2, 3], and with delayed progression to end-stage kidney disease in case series [4, 5]. In the UK, most HIVAN cases are encountered in patients with advanced immunodeficiency who were not previously known to be HIV positive, or who disengaged from care or who declined ART [6]. HIVAN is rare in patients with CD4 cell counts >350 cells/μL or with undetectable HIV RNA levels [7]. Patients presenting with higher levels of proteinuria (urine albumin–creatinine ratio >70 mg/mmol or urine protein–creatinine ratio >100 mg/mmol or urine protein excretion >1 g/24 h) or proteinuria with haematuria (urine albumin–creatinine ratio >30 mg/mmol or urine protein–creatinine ratio >50 mg/mmol) or stage 4–5 CKD should be referred for specialist assessment and a renal biopsy considered; those found to have HIVAN should start ART immediately, irrespective of CD4 cell count.

For CKD other than HIVAN, there is limited information on the natural history per se and on whether ART confers renal benefit. Immunodeficiency is a potent risk factor for CKD [8, 9]. The majority of patients with CKD have (nadir) CD4 cell counts <350 cells/μL and thus qualify for ART as per current treatment guidelines. There are no data to provide guidance on whether HIV-positive patients with (or at risk of developing) CKD benefit from earlier ART initiation. None the less, HIV replication, immune activation and inflammation may play a role in the pathogenesis of kidney diseases or contribute to kidney disease progression in some patients [10]. For this reason, ART should be considered in those presenting with CKD other than HIVAN.

Renal transplantation is the treatment of choice for those requiring renal replacement therapy. Patients to be considered for renal transplantation are required to have suppressed HIV RNA levels and to have CD4 cell counts >200 cells/μL [11], and should start ART, irrespective of CD4 cell count.

8.5.2 What to start Recommendations

  • We recommend against the use of ARV drugs that are potentially nephrotoxic in patients with stages 3–5 CKD if acceptable alternative ARV agents are available (GPP).
  • We recommend dose adjustment of renally cleared ARV drugs in patients with reduced renal function (GPP).
Auditable measure. 

Number of patients with CKD stages 3–5 on ARVs that are potentially nephrotoxic and a record of the rationale.

Record in patient's notes of calculated dose of renally cleared ARVs in patients with CKD stage 3 or greater. Rationale

There are no data from clinical RCTs to inform ART decisions in patients with CKD. The risk of CKD is increased with older age, reduced estimated glomerular filtration rate (eGFR), hypertension, diabetes and with cumulative exposure to indinavir, TDF, ATV and, to a lesser extent, LPV [12, 13]. Indinavir use is no longer recommended in view of the high incidence of renal complications: crystalluria and pyuria are reported in 20–67% [14-16] and nephrolithiasis, tubulointerstitial nephritis and gradual loss of renal function in 4–33% of patients [14, 17-20].

TDF has been associated with falls in eGFR [12, 21, 22], accelerated decline in eGFR [9], acute renal failure [23], tubulointerstitial nephritis [24], CKD [9, 12], renal tubular dysfunction [13, 25] and Fanconi syndrome [26, 27]. The incidence of TDF-associated renal toxicity is low in clinical trials and cohort studies of the general HIV population [28, 29]. Older age, pre-existing renal impairment, co-administration of didanosine or (ritonavir-boosted) PIs, advanced HIV infection and low body mass appear to increase the risk of renal complications [9, 13, 25, 27, 30, 31].

ATV has been associated with reductions in eGFR [32], nephrolithiasis and tubulointerstitial nephritis [13, 24, 33], and CKD [12]. The incidence of renal stones with ATV in one cohort was 7.3 per 1000 person-years, with almost half of those who developed renal stones having eGFR <60 at the time of ATV initiation [34].

The nephrotoxic potential of both TDF and ATV is low in patients with normal renal function. However, in patients with CKD and impaired renal function (eGFR <75 mL/min/1.73m2), alternative ARVs should be considered.

In patients undergoing renal transplantation, PIs give rise to challenging DDIs with calcineurin inhibitors (http://www.hiv-druginteractions.org). Post-transplantation, acute allograft rejection and impaired renal function are common [35]. We suggest TDF and ATV are avoided in patients who are waiting or who have undergone, renal transplantation, and that specialist advice is sought regarding choice and appropriate dose of ARVs.

NNRTIs, INIs, ABC and 3TC have not been associated with CKD and can be used in HIV-positive patients with CKD. In patients with impaired renal function, specific ARV drugs (all NRTIs except ABC) may need to be dose-adjusted [36]. Impaired survival has been reported with ART prescription errors in patients undergoing dialysis [37]. We recommend dose adjustment of renally cleared ARVs in patients with renal failure but caution against the risk of overinterpreting estimates of renal function for this purpose as true measures of renal function may be substantially higher in patients with mild–moderate renal impairment. Specific ARVs that require dose adjustment in patients with reduced renal function include 3TC, FTC, TDF, DDI, ZDV and MVC (depending on PI use). For further information and advice, the reader should refer to the summary of product characteristics for each ARV.

8.5.3 References

8.6 Cardiovascular disease

8.6.1 Introduction

CVD is a leading cause of non-AIDS morbidity and mortality among HIV-positive individuals [1, 2] and an increased risk of CVD events has been observed when compared with HIV-negative populations [3-8]. This has been attributed to the increased prevalence of surrogate markers of CVD (such as dyslipidaemia) and the proinflammatory state associated with HIV infection. However, because ART may not mitigate (or indeed may exacerbate) these effects, caution is required in extrapolating from these makers to effects on overall mortality. The following recommendations apply to patients with, or at high risk, of CVD.

8.6.2 Definition and assessment of cardiovascular disease risk

For the purposes of these guidelines, patients with an elevated CVD risk are as defined in the JBS2 guidelines [9] and include:

  • People with any form of established atherosclerotic CVD.
  • Asymptomatic people who have an estimated multifactorial CVD risk >20% over 10 years.
  • People with diabetes mellitus (type 1 or 2).
  • People with elevated blood pressure >160 mmHg systolic or >100 mmHg diastolic, or lesser degrees of blood pressure elevation with target organ damage.
  • People with elevated total cholesterol to high-density lipoprotein cholesterol ratio >6.0.
  • People with familial dyslipidaemia.

NICE does not recommend a specific CVD risk calculation for the UK population [10]. Cohort data have demonstrated that the observed myocardial infarction (MI) rates in HIV-seropositive people in developed countries paralleled those predicted by the Framingham risk equation [11] but the extent to which this can be extrapolated to women and men of non-European ethnicity is unknown. Therefore, there is insufficient evidence to recommend a specific CVD risk calculation for the population of HIV-positive adults in UK.

The Framingham CVD risk calculator works reasonably well in HIV-positive populations, although it is worth noting that it was not developed for use in non-white groups. Other algorithms may be better suited to these populations. A CVD risk calculator has been developed for use in HIV-positive populations (http://www.chip.dk/TOOLS) [12], although it should be noted that this provides 5-year risk estimates rather than the usual 10-year estimates. Alternatively, the QRISK calculator (http://www.qrisk.org) or the QIntervention tool (http://qintervention.org), which also provides an estimate of the risk of developing type II diabetes, can be used.

8.6.3 When to start antiretroviral therapy

There are insufficient data to inform whether CVD risk should affect the decision to start ART.

The SMART trial provides the only randomized data about the effect of ART on CVD risk, but was not powered for a CVD endpoint. Fewer major CVD events were observed in the viral suppression arm but the difference was not statistically significant [13]. In a post hoc analysis, HIV VL <400 copies/mL was associated with fewer CVD events suggesting that suppression of viraemia may have been protective; CD4 cell count was not significantly associated with CVD events [14, 15].

Several cohort studies have examined changes in rate of cardiovascular events in HIV-positive populations over time since the introduction of ART but no clear protective effect was found [16-19]. In the HIV Outpatients Study cohort, baseline CD4 cell count <350 cells/μL was associated with increased CVD risk, but 350–500 cells/μL and use of ART were not; in a parallel case–control study, cases were more likely to have a current (but not baseline or nadir) CD4 cell count of 350–500 cells/μL [20]. The Data Collection on Adverse events of Anti-HIV Drugs (D:A:D) study found that untreated patients had a lower incidence of MI than those on ART [21] and risk increased with longer exposure to combination therapy [22].

While there is uncertainty as to whether treating HIV infection reduces CVD risk, there is good evidence from RCTs that interventions targeted at modifiable CVD risk factors are of benefit. For this reason, all HIV-positive adults should be assessed for CVD risk annually and interventions targeted at improving modifiable risk factors.

8.6.4 What to start Recommendations

  • We suggest avoiding ABC (2C), FPV/r (2C) and LPV/r (2C) in patients with a high CVD risk, if acceptable alternative ARV drugs are available.
Auditable measure. 

Number of patients with high CVD risk on either ABC or FPV/r or LPV/r and record of rationale. Rationale

Modifiable risk factors should be addressed in all patients with high CVD risk.

No RCT has been powered to assess the CVD risk associated with the use of individual ARVs and a history of CVD may be an exclusion criteria. A meta-analysis of all RCTs where ABC was assigned randomly found no association with MI, but the event rate in the population was low; the extent to which these findings can be extrapolated to a population with high CVD risk is unknown [23]. Although a post hoc analysis of the SMART study did find such an association, use of ABC was not randomized [24].

Two cohorts have found a strong association between recent ABC use and MI [25, 26] while another did not [27, 28]; all were limited in their ability to adjust for presence of CVD risk factors. An analysis of the manufacturer's trial registry found no association [29], but the trials only enrolled patients with low CVD risk. One case–control study, which did not adjust for important CVD risk factors, did find an elevated risk of MI associated with ABC use [7] but another did not [12]. Cerebrovascular events were more common in patients exposed to ABC in two cohort studies [8, 28] while another found a protective effect [27]. In view of the uncertainty about the safety of ABC in patients with a high CVD risk, we suggest the use of alternative agents where possible.

Early studies of PI exposure and risk of MI gave conflicting results, some reporting an increased risk [5, 30] while others did not [3, 16, 31]. The D:A:D cohort, with longer follow-up, reported an increasing risk of MI with years of PI exposure (independent of measured metabolic effects) [22]. Cumulative exposure to indinavir and LPV/r were associated with increasing risk of MI [adjusted relative risk per year for LPV/r 1.13 (95% CI 1.05–1.21); relative risk at 5 years 1.84] [26]. Case–control studies reported similar associations for LPV/r [7, 12] and FPV/r [12] but in one of these, important CVD risk factors were not included [7]. A further study found no association between PI exposure and all cerebrovascular events [8]. An updated analysis has recently reported no association between ATV/r use and an increased risk of MI [32]. Although there has been insufficient data to include DRV/r in these analyses, in patients with a high CVD risk, we suggest the use of alternatives to LPV/r and FPV/r where possible.

In the MOTIVATE studies for treatment-experienced patients, coronary artery disease events were only reported in the MVC arm (11 in 609 patient years), while there were none in the placebo arm (0 in 111 patient years); those affected generally had pre-existing CVD risk. No such signal was found in the MERIT study for treatment-naïve patients. MVC has also been associated with postural hypotension when used at higher than recommended doses in healthy volunteers; patients with a history of postural hypotension, renal impairment or taking antihypertensive agents may be at increased risk [33]. In view of the limited data available, special caution should be exercised in the use of MVC in patients with a high CVD risk and use of alternative agents, where possible, considered.

8.6.5 References

8.7 Women

8.7.1 Introduction

The following guidance considers issues concerning the initiation and choice of ART for HIV-positive women who are not currently pregnant. For guidance on the management of pregnancy in HIV-positive woman please refer to the BHIVA guidelines for the management of HIV infection in pregnant women 2012 [1].

There are few specific data on ART treatment in women other than in pregnancy. Data available are largely from a meta-analysis, post hoc analyses or derived from cohort studies. The majority of the randomized clinical trial data on ART comes from studies that have enrolled mostly male subjects. If RCTs do enrol women, the numbers are often too small to draw significant gender-based conclusions.

Approximately one-third of people diagnosed with, and accessing care, for HIV in the UK are women [2]. The majority are of childbearing age but the age range is increasing, adding the complexity of menopause and its sequelae to the management of HIV-positive women. Many HIV-positive women in the UK are of African heritage and face overlapping challenges to their health and well-being [3].

Women's experience of HIV reflects multiple social and cultural influences, which when combined with sex-specific biological factors influence individual responses to HIV.

8.7.2 When to start Recommendations

  • We recommend therapy-naïve HIV-positive women who are not pregnant start ART according to the same indicators as in men (see Section 4: When to start) 1A.
Auditable measure. 

Proportion of HIV-positive women with CD4 cell count <350 cells/μL not on ART. Rationale

Gender differences in HIV VL and CD4 cell count at different stages of infection have been observed [4] but have not been consistently associated with long-term clinical outcomes for HIV-positive women. Based on current data, the indications for starting ART do not differ between women who are not pregnant and men.

Gender-specific socio-economic and cultural factors may impact on women's ability to access care and manage their medication, compromising their ability to initiate and adhere to therapy, and they may require support from the multidisciplinary team.

8.7.3 What to start Recommendations

  • We recommend therapy-naïve HIV-positive women start ART containing two NRTIs and one of the following: PI/r, NNRTI or INI (1A), as per therapy-naïve HIV-positive men.
  • We recommend therapy-naïve HIV-positive women start ART with preferred or alternative NRTI backbone and third agent as per therapy-naïve HIV-positive men (See Section 5.1: What to start: summary recommendations) (1A).
  • Factors such as potential side effects, co-morbidities, drug interactions, patient preference and dosing convenience need to be considered in selecting ART in individual women.
  • We recommend both HIV-positive women of childbearing potential and healthcare professionals who prescribe ART are conversant with the benefits and risks of ARV agents for both the health of the HIV-positive woman and for that of an unborn child (GPP).
  • We recommend that potential pharmacokinetic interactions between ARVs, hormonal contraceptive agents and hormone replacement therapy are checked before administration (with tools such as: http://www.hiv-druginteractions.org) (GPP]). Rationale


There are few data to guide prescribing of initial ART specifically for women, as no RCT in patients starting ART has been powered to detect sex differences in efficacy. From the limited data available, virological outcomes within clinical trial settings generally appear to be no different between men and women.

A meta-analysis of FDA registrational RCTs analysed data from 22 411 HIV-positive patients participating in 43 trials for 16 ARVs. Overall, 20% of study participants were women. No significant differences in treatment response at week 48 were reported between men and women. Rates of ART discontinuation for virological failure were higher in men (8.15%) than in women (4.25%) [5].

A subanalysis of an RCT comparing ATV/r and LPV/r in ART-naïve patients of whom 31% were women, showed comparable virological efficacy at week 96 between the two treatment arms in women [6], although virological response rates were lower in women when compared with men.

In a study comparing ATV/r and EFV in 1857 ART-naïve patients of whom 17% were women, female sex was associated with increased virological failure on ATV/r compared with EFV [7]. No difference was seen with EFV between men and women.

The efficacy and tolerability of RAL were shown not to be different between men and women at 48 weeks in one study of a diverse cohort of both treatment-naïve and -experienced patients [8]. RPV in ART-naïve men and women showed no difference in rates of virological suppression at 48 and 96 weeks between men and women, but the number of women included was low and the study was not designed to investigate sex differences [9, 10].

Cohort studies in the UK have reported similar virological outcomes during the first year of treatment in heterosexual men and women [11]. An Italian cohort study reported no significant effect of gender on clinical progression or the risk of developing a clinical event [12]. Data from Spain, which included both naïve and ARV-experienced women patients, showed them with similar virological responses to men [13].

HIV-positive women starting ART should use ARVs from the list of preferred and alternative drugs outlined in Section 5.1 (What to start: summary recommendations). Factors, including potential for side effects, drug interactions, patient preference, co-morbidities and dosing convenience need to be taken into consideration when selecting ART regimens in individual women.

Toxicity, discontinuation and adherence. 

Adverse events and treatment discontinuations within ART clinical trials and cohort studies published between 2002 and 2007 have been systematically reviewed. The overall event rate is often the same but the adverse event profile may be different. Women were reported to be more likely than men to experience ART-related lipodystrophy, rash and nausea, and to discontinue therapy [4].

Data from the USA have shown that women are more likely than men to discontinue ART for poor adherence, dermatological symptoms, neurological reasons, constitutional symptoms and concurrent medical conditions [14]. UK cohort data found 88.6% of men compared with 80.7% of women spent 100% of the first year after starting HAART actually on therapy [11].

Comparison of ATV/r with LPV/r found poorer virological outcomes in treatment-naïve women compared with men. Gender differences in efficacy were due to higher discontinuation rates in women than men in both treatment arms [6]. CNS side effects of varying severity can occur with EFV, particularly at the initiation of therapy. This may be partly explained by the greater EFV exposure associated with a CYP2B6 variant, more commonly found in Africans and African Americans [15]. In the UK population, this is of particular relevance to women, the majority of whom are of African ethnicity. NVP-associated rash occurs more frequently in women than men [16]. Hepatotoxicity associated with NVP is more common in women with a CD4 cell count >250 cells/μL, restricts women's use of the drug [17].

A systematic review of studies on gender and ART adherence published between 2000 and 2011 in the resource-rich world concluded that overall reported adherence is lower in women than men [18]. However, of over 1000 studies initially identified for review, only 44 had adequate data on gender to allow any comparisons to be made. The authors identified the particular factors for lower adherence in women were depression, lack of supportive interpersonal relationships, young age, drug and alcohol use, black ethnicity, ART of six or more pills per day, higher numbers of children, self-perception of abdominal fat gain, sleep disturbances and increased levels of distress.

Fetal safety. 

Concerns about potential fetal toxicity of ARVs have influenced prescribing practice in HIV-positive women. Of note, other than ZDV in the third trimester, no ARV drug has a licence for use in pregnancy.

Pregnancy in women living with HIV who are already on effective therapy is increasing; 70% of HIV-positive pregnant women in the UK in 2010 were diagnosed before the current pregnancy, of which 60% were already on ART at conception [19]. Where newer drugs are available, women are conceiving on these agents, with ZDV now rarely used as first-line therapy for adults. European cohort data comparing pregnancies that were managed with ZDV-containing regimens vs. those without ZDV found no difference in risk of detectable VL at delivery, vertical transmission or congenital abnormality when comparing ZDV-sparing with ZDV-containing ART [20].

The most robust data on teratogenicity and first trimester ART exposure are from the Antiretroviral Pregnancy Registry (APR) [21]. This international prospective reporting system records rates of congenital birth defects in babies born to women with exposure to ART at any stage of pregnancy. Approximately 200 or more reports need to be received for a particular compound before data are reported for that compound by the APR. There are now over 200 prospective reports in the APR of first trimester exposure for ABC, ATV, EFV, FTC, 3TC, LPV, NVP, ritonavir, TDF and ZDV. No signal of increased risk of congenital abnormality has been demonstrated, and a greater than twofold higher rate than in the general population has been excluded. There are, so far, fewer than 200 prospective reports for DRV, RAL and RPV within the APR and hence no reports on these agents are yet available.

Despite previous concerns over the safety of EFV based on preclinical animal studies and retrospective case reports in human subjects, the current data do not provide evidence of excess teratogenicity above the expected baseline for infants exposed to EFV in the first trimester. Sufficient numbers of first trimester exposures of EFV have been monitored to detect at least a twofold increase in risk of overall birth defects within the APR, and no such increases have been detected to date [21].

Data from Côte d'Ivoire found no significant increased risk of unfavourable pregnancy outcome in women with first-trimester exposure to EFV compared with NVP [22]. A systematic review and meta-analysis of observational cohorts carried out in 2010 [23] and further updated in 2011 [24] reported birth outcomes among women exposed to EFV during the first trimester. No increased risk of overall birth defects among the babies of women exposed to EFV during the first trimester compared with exposure to other ARV drugs was found. The prevalence of overall birth defects with first-trimester EFV exposure was similar to the ranges reported in the general population.

A review of live births to women with HIV in a large unselected UK population between 1990 and 2007 found no increased risk of abnormalities in infants exposed to EFV in the first trimester, providing further reassurance that ART in utero does not pose a major risk of fetal anomaly [25]. Mathematical modelling using North American cohort data has demonstrated a theoretical loss of life expectancy in women who delay EFV at initiation of ARV [26].

Based on current evidence, EFV can be initiated in women of childbearing potential, can be continued in women who conceive on the drug and commenced in pregnancy but the data should be discussed in detail with the individual woman when deciding on her preferred treatment regimen. Given that no ARV drug is licensed for use in pregnancy apart from ZDV in the third trimester, a discussion regarding the potential unknown long- and short-term effects on an unborn child should be had with any woman of childbearing potential who commences any ARV drug regimen. Further details can be found in the BHIVA pregnancy guidelines [1].

Hormone interactions. 

Significant pharmacokinetic and pharmacodynamic interactions have been reported between ARV drugs and hormonal agents. Inducers of hepatic enzymes by ARVs may result in increased breakdown of ethinyl oestradiol and progestogens that can compromise contraceptive and hormone replacement therapy efficacy. Additional contraceptive measures or different ARV regimens may be required in these circumstances. Potential DDIs should be checked using various resources, including specialist HIV pharmacists, web-based tools such as the University of Liverpool website on HIV drug interactions and medical information departments in pharmaceutical companies. There is no significant interaction between ETV and the combined oral contraceptive pill, and no interaction is anticipated with RAL. Hormonal contraceptive agents, which have been shown not to have a significant interaction or where there is no anticipated interaction include depot medroxyprogesterone acetate, and the levonorgestrol IUS (Mirena coil).

8.7.4  HIV-positive women experiencing virological failure

There is very little evidence to guide prescribing ART in HIV-positive women experiencing virological failure on ART, with most studies recruiting approximately 10% of women. One study investigating DRV/r in ART-experienced patients recruited a large proportion of women and was powered to show a difference in virological efficacy between men and women; this showed higher discontinuation rates among women than men, with nausea being cited as a particular problem, but overall there was no difference in virological efficacy [27]. A further study has reported similar efficacy and tolerability of RAL in ART-experienced HIV-positive women [8].

In HIV-positive women experiencing virological failure on ART, the same principles of management and recommendations apply as per HIV-positive men experiencing virological failure (see Section 7: Management of virological failure).

8.7.5 References