4.1.1. Clinical features

4.1.2. Staging criteria

Adults

The AIDS Clinical Trials Group (ACTG) staging classification was developed in the pre-ART era to establish a common language to use in describing AIDS-associated KS (Krown et al., 1989) and was later shown to be associated with survival (Krown et al., 1997). The classification is based on three criteria: tumour extent (T), the status of the patient’s immune system as measured by CD4 cell count (I) and presence of systemic symptoms (S). Each factor is further grouped into good risk (0) or poor risk (1). T0 disease is defined as KS that is limited to the skin but without oedema or ulceration, lymph node disease or minimal oral disease (e.g. flat palatal lesions). T1 disease is defined as the presence of tumour-associated oedema or ulceration, or extensive oral or gastrointestinal disease or other non-nodal visceral lesions (TABLE 2) (Krown et al., 1989; Krown et al., 1997).

TABLE 2

TUMOUR-IMMUNE SYSTEM–SYSTEMIC ILLNESS (TIS) STAGING SYSTEM FOR AIDS-RELATED KS AND PROGNOSIS.

The TIS criteria were developed prior to the availability of ART. A subsequent study performed early in the ART era showed the validity of these criteria with respect to survival, but with different weighting of the TIS factors (Nasti et al., 2003), However, they may not sufficiently reflect the significant impact of ART on the prognostic importance of these three criteria and do not include other factors that may affect prognosis and treatment decisions (Krown, 2005).

4.1.3. Treatment strategies

In the pre-ART era, chemotherapy was frequently used alone to treat patients with KS. Currently, however, ART initiation is recommended for all patients with KS, but there are no universally agreed criteria for identifying those patients with KS who also require chemotherapy, and practices vary widely. While the presence of symptomatic visceral KS would be considered an indication for systemic chemotherapy in most settings, patients with widespread or ulcerated cutaneous disease, symptomatic oedema or oral-cavity disease may also be considered suitable candidates for treatment with chemotherapy in many settings. Regimens commonly used in high-resource settings are single-agent liposomal anthracyclines (doxorubicin and daunorubicin) and single-agent paclitaxel. Other options, which are more widely used in lower-resource settings, include combination chemotherapy with BV, with or without non-liposomal doxorubicin. For those with less extensive skin involvement, ART alone may be sufficient to cause lesion regression. Localized therapies used both in the pre-ART era and together with ART for patients with symptomatic local disease or for cosmesis include radiotherapy, topical therapy and intralesional chemotherapy.

Kaposi sarcoma – classification/case definitions

(Note: these case definitions are modified from the original ACTG T0 and T1 definitions, and are intended as a general guide for treatment decision-making)

Mild/moderate disease

The systematic review was based on the PICO question: in HIV-infected ART-naïve adults diagnosed with mild/moderate KS, not requiring immediate anti-KS therapy (P), does the use of ART alone (I) compared to ART plus any additional therapy (chemotherapy, radiation, intralesional injections or topical therapy) (C), improve survival or clinical response (includes complete, partial, stable or progressive disease) (O). Secondary outcomes were time to response, KS IRIS, adverse events (including toxicity and/or worsening of co-existent disease), adherence and quality of life (Freeman et al., in press).

Chemotherapy with ART versus ART alone: No RCTs or observational studies have been specifically designed or powered to address the value of adding chemotherapy to ART compared to ART alone, in patients with mild to moderate KS. Two studies, one RCT in South Africa (Mosam et al., 2012) and one cohort from France (Dupin et al., 1999), compared ABV with ART versus ART alone, and a further two – one Spanish RCT (Martin-Carbonero et al., 2004) and one United Kingdom cohort (Bower et al., 2014) compared liposomal anthracyclines plus ART to ART alone. However, all these studies included patients at different stages of KS disease, and the number of participants with mild/moderate disease ranged from three to 213. None identified a significant mortality or treatment response benefit to the addition of either ABV or liposomal anthracyclines to ART in patients with mild to moderate ART-naïve KS.

Two studies (Letang et al., 2013; Mosam et al., 2012) examined the development of KS IRIS as an endpoint. Among patients with mild to moderate KS, none developed KS IRIS in either treatment arm in the trial from South Africa (Mosam et al., 2012). In the United Kingdom cohort within the multisite cohort study (Letang et al., 2013), six of the 77 patients with mild to moderate disease who received ART alone did so, in comparison to none of the seven observed in the chemotherapy plus ART arm (additional non-published data from authors). There were limited data on adverse events among those with mild to moderate KS, with no Grade III-V adverse events reported in the RCT (Mosam et al., 2012), and in only three patients out of 28 receiving chemotherapy in one cohort study (Martin-Carbonero et al., 2004; Freeman et al., in press).

Protease inhibitor (PI)-based ART versus non-nucleoside reverse transcriptase inhibitor (NNRTI)-based ART: A large RCT from Uganda compared PI versus NNRTI-based ART regimens in treatment-naïve patients with mild to moderate KS, and found no difference in overall survival or KS progression-free survival (Martin et al., 2013).

Radiation therapy, use of locally-injectable agents and topical therapy: Although there are many studies from the pre-ART era comparing different radiation dosages or examining the use of intralesional injections of chemotherapeutic agents (such as bleomycin, vincristine or vinblastine), none compared ART alone to ART plus radiotherapy, or intralesional chemotherapy. Similarly, none of three studies of topical retinoids met the criteria for inclusion in the review, because of a mix of different KS stages and/or lack of uniform receipt of ART (Bodsworth et al., 2001; Duvic et al., 2000; Walmsley et al., 1999).

Severe disease

The systematic review was based on the PICO question: in HIV-infected adults (receiving and not receiving ART) diagnosed with severe KS (defined as ACTG T1 disease, or when ACTG staging not available, as visceral KS, oedema affecting function, oral KS interfering with chewing or swallowing, tumours with ulceration, or life-threatening KS) (P), the interventions (I) and comparisons (C) were as follows: any chemotherapy regimen in combination with ART (I) compared with ART alone (C); any chemotherapy regimen alone (I) compared to ART alone (C); one chemotherapy regimen, e.g. liposomal anthracyclines (I), compared to another chemotherapy regimen (C), e.g. ABV, with or without the presence of ART. Outcome measures (O) were survival or clinical response (includes complete, partial, stable or progressive disease). Secondary outcomes were time to response, KS IRIS, adverse events (including toxicity and/or worsening of co-existent disease), adherence and quality of life (Gbabe et al., in press).

The review included six RCTs conducted in the USA (Cianfrocca et al., 2010; Cooley et al., 2007; Gill et al., 1996), Spain (Martin-Carbonero et al., 2004), South Africa (Mosam et al., 2012) and Zimbabwe (Olweny, et al., 2005); and three observational studies from the United Kingdom (Bower et al., 2009), Germany (Grünaug et al., 1998) and Venezuela (Hernández & Pérez, 1997). The number of patients with T1 stage disease ranged from 29 to 227 (Grünaug et al., 1998; Hernández & Pérez, 1997; Cianfrocca et al., 2010; Bower et al., 2009; Cooley et al., 2007; Gill et al., 1996). None of the studies had the same interventions, and therefore pooled analyses were not possible.

ART plus chemotherapy versus ART alone was compared in two trials. ABV plus ART versus ART alone was compared in an RCT from South Africa (Mosam et al., 2012) which excluded patients requiring urgent chemotherapy. No survival benefit was observed, but the overall clinical response (complete plus partial) favoured ABV plus ART (RR 1.78; 95% CI 1.16 to 2.72), and there was also less disease progression in this arm (RR 0.1; 95% CI 0.01 to 0.75). There were similar rates of adverse events. Pegylated liposomal doxorubicin (PLD) plus ART versus ART alone was examined in one very small RCT with five patients with severe disease in each arm, which appeared to favour liposomal anthracyclines in terms of treatment response but did not reach statistical significance (Martin-Carbonero et al., 2004)

ART plus chemotherapy versus ART plus a different chemotherapy regimen was examined in two RCTs. Cianfrocca and colleagues (2010) compared ART plus paclitaxel to ART plus liposomal anthracyclines, and found no difference in treatment response or survival between the two arms. Cooley and colleagues (2007) compared ART plus PLD to ART plus liposomal daunorubicin and similarly found no difference.

Chemotherapy versus a different chemotherapy regimen in the pre-ART era was compared in one RCT (Gill et al., 1996) of liposomal daunorubicin versus ABV in patients not on ART. The overall response and progression were similar in both arms. In one cohort study (Grünaug et al., 1998), mortality was three times higher in those patients who received either ABV or no chemotherapy, in comparison to PLD (survival time 11.8 months versus 4.4 months, p<0.01).

Two large RCTs that were part of a previous Cochrane review (Dedicoat et al., 2009) were excluded from this review because of a mix of KS stages. They compared PLD with ABV (Northfelt et al., 1998) or PLD with BV (Stewart et al., 1998) without ART (total 499 patients). Mortality was similar across the treatment arms (RR 1.26; 95% CI 0.83 to 1.91), but PLD had a superior clinical response rate (complete plus partial) (RR 2.16; 95% CI 1.68 to 2.78) without an increase in toxic side effects. Additionally, there were fewer withdrawals in the PLD arm (RR 0.57; 95% CI 0.48 to 0.68), but more OIs (RR 1.42; 95% CI 1.12 to 1.80).

No studies were identified that evaluated the optimal timing of ART in relation to chemotherapy (i.e. whether chemotherapy should be initiated prior to ART, and duration prior to ART initiation).

Type of chemotherapy

There are limited data on the relative benefits of ABV versus BV or paclitaxel regimens in children. No studies were identified on liposomal doxorubicin, although it is being used for the treatment of KS in children in higher-income settings. Given the small sample size in these studies, it was not possible to determine whether there was a difference in the proportion with complete or partial remission between use of ART plus vincristine versus ART plus BV (Gantt et al., 2010). The use of paclitaxel specifically in paediatric KS patients was reported in one study (Vaz et al., 2011). Cox and colleagues (2013) reported mortality among children treated with combination chemotherapy and ART versus children treated with single agent chemotherapy and ART. Of 36 children treated with combination chemotherapy and ART, 13 died by 12 months of follow-up; all 14 children treated with single agent chemotherapy and ART died (RR 0.38; 95% CI 0.24 to 0.58).

ART initiation

The GDG endorsed the prompt initiation of ART as the most important intervention in all patients with KS regardless of stage of disease, in order to improve survival and reduce morbidity through immune recovery and reduction in OIs. It was recognized that ART may result in KS lesion regression in some cases. Since 2006, WHO has recommended ART initiation in patients with KS or other WHO stage 3 or 4 disease in successive guidelines (WHO, 2013), and it has been associated with a marked improvement in survival and morbidity. ART is also widely available in HIV programmes at low cost (US$ 120 per patient per year), is well tolerated and requires minimal monitoring. This recommendation can therefore be readily implemented at all levels of service delivery, including at primary health care centres. However, where possible, patients with KS should first be evaluated at a site that has familiarity with KS and KS treatment.

PI and NNRTI regimens were equally effective in those with KS (Martin et al., 2013). Therefore the current preferred WHO recommended regimen of tenofovir disoproxil fumarate (TDF)+lamivudine (3TC)/emtricitabine (FTC)+efavirenz, which is available as a fixed dose combination of one pill once a day, can be used.

Use of chemotherapy

Although ART by itself may induce regression of KS lesions, the resolution of KS with ART may be a slow process over several months, as it usually follows the recovery of cell-mediated immunity in the host. It may also have little effect on severe disseminated disease. Furthermore, use of ART alone may be complicated by IRIS, especially in those with severe KS (Letang et al., 2013).

In contrast to the major impact of ART on mortality, the evidence review demonstrated that the main added contribution of chemotherapy in KS patients who will initiate ART is to achieve more rapid regression and remission of significant KS lesions, and in turn provide potential relief of symptoms (such as lymphoedema), and improvement in quality of life. Studies have not yet shown a mortality benefit, although many were of limited sample size and were not powered to detect survival differences. An additional benefit may be to reduce the risk of developing life-threatening ART-associated KS IRIS. Potential harms of chemotherapy include significant toxicities in some cases. Other challenges include the limited access and relatively high cost of chemotherapy drugs in resource-limited settings, where there are also few specialist sites with the infrastructure to safely administer and monitor chemotherapy drugs.

Selection of patients for chemotherapy

There are currently no standardized or universally accepted criteria to guide which HIV-infected adults or children with KS would benefit most from concomitant chemotherapy in addition to receiving ART. Some practitioners (Bower et al., 2014) have advocated using T staging as the primary criterion for treatment planning. Using this approach, patients with T1 stage disease were treated with ART plus systemic chemotherapy, and those with T0 disease with ART alone. In other settings, symptomatic or functionally disabling KS has been used as the main indication for systemic chemotherapy in addition to ART. For example, in one study (Mosam et al., 2012), functionally disabling KS was considered an exclusion to participation in a randomized trial that compared ART alone to ART plus chemotherapy, since patients so affected were considered to require immediate treatment with combined chemotherapy and ART.

The GDG recognized the need for easily-recognized and objective clinical criteria that can be used by all levels of health care workers to guide which patients need prompt referral and assessment for chemotherapy, in addition to use of ART. Extent of tumour alone (based on the ACTG criteria) may be insufficient to guide who should be prioritized for urgent chemotherapy, as some patients with T1 disease may lack significant KS-related symptoms, whereas others with T0 disease may have extensive skin involvement that adversely affects quality of life and social functioning. Other factors to be considered include presence of distressing or functionally-disabling symptoms (Krown, 2004 & 2005). A simple categorization of KS into mild/moderate KS and severe symptomatic KS based in part on the original ACTG tumour extent criteria was used, with some additional clinical criteria to capture functionally-disabling complications of KS. These two categories do not capture all clinical scenarios, and, for example, do not consider asymptomatic visceral disease. Additional factors noted by the GDG that may help determine priority for immediate chemotherapy include rate of progression, quality of life, presence of other co-morbidities, contraindications to chemotherapy and the wishes of the patients and their families. The GDG recommended that those with severe KS and functionally-disabling symptoms merit urgent referral for chemotherapy in addition to ART. For those with less-extensive skin involvement, ART alone may be sufficient to cause lesion regression; if this fails, chemotherapy remains a future option. The GDG noted that the definitions of mild to moderate KS that may be managed by ART alone, and of severe KS requiring prompt chemotherapy, is a research gap that needs to be addressed, as are the potential benefits of earlier institution of chemotherapy in selected patients with mild to moderate KS.

Other treatments

No specific recommendations were made for the use of local therapies. Local therapies are limited by their inability to treat large areas or to affect the development of lesions in other areas, although they may sometimes be useful for managing localized or symptomatic KS lesions, or for cosmesis. Similarly, although radiotherapy was used in the past for the management of localized cutaneous KS, this was not felt to be a generally feasible intervention for resource-limited settings, as it was costly, as well as being less convenient for patients.

4.5.1. Drug availability and cost

There remains limited access to chemotherapeutic drugs in many parts of Africa, and access is generally limited to tertiary care facilities. The choice of drug regimen will be influenced by drug availability. Vincristine and bleomycin, and to a lesser extent doxorubicin, are available in major specialist tertiary care centres in sub-Saharan Africa. Liposomal preparations are not generally available in those settings at present, and are expensive. However, the entry of generic producers is likely to decrease costs and potentially allow for greater access. Problems with stock-outs of medications may occur, often due to forecasting rather than difficulty obtaining the drugs.

The number of treatment courses required will vary according to regimen type, and rapidity and completeness of treatment response. The total cost for six treatment cycles (given about every two to four weeks) of ABV is around US$ 400 (US$ 66 for vincristine, US$ 375 for bleomycin and US$ 33, US$ 125 and US$ 111 respectively for an ABV combination). The costs of liposomal doxorubicin treatment are high, ranging from US$ 150 per vial for the generic product to US$ 1000 for a branded one, or US$ 1800 to US$ 14 000 respectively for six treatment cycles.

4.5.2. Toxicities

Most data on toxicity of different KS chemotherapy regimens are based on studies conducted in the pre-ART era. In general, most regimens are relatively well tolerated, but serious toxicities may occur. For example, vincristine is associated with neuropathy which may be irreversible. This is of particular significance in children where administration of vincristine may be associated with the development of gastric ileus. Differences in monitoring requirements exist between various regimens depending on the toxicities associated with different chemotherapeutic agents (see TABLE 4).

TABLE 4

MAJOR TOXICITIES OF CHEMOTHERAPY DRUGS.

Liposomal anthracycline (PLD and non-pegylated liposomal daunorubicin) treatment is the standard of care for adults in the USA and Western Europe. These regimens are of comparable efficacy and are better tolerated with less toxicity, and improvements in several domains of health-related quality of life, compared with BV or ABV. These include less alopecia and fewer gastrointestinal and neurological side effects, while Grade 3/4 myelosuppression, stomatitis and infusion reactions are more common with PLD. However, liposomal preparations are expensive, not widely available, remain under patent, and require cold storage. No formal pharmacokinetic (PK) drug-drug interaction studies have been conducted among contemporary ART agents and those used for the treatment of KS, whereas the PK of each agent has been evaluated. The data presently available are sparse and largely based on case reports. Yet there are reasons for potential concern, as many of the chemotherapeutic agents currently in use (i.e. liposomal and non-liposomal anthracyclines, paclitaxel, vincristine, etoposide) undergo metabolism in the liver by enzymes that have the potential to be inhibited or induced by various ART drugs (Rudek et al., 2011).

4.5.3. Site infrastructure and staff training

Site infrastructure: Provision of chemotherapy drugs requires adequate infrastructure (e.g. pharmacy, laminar flow) for preparation of the drugs, as well as a level of expertise with trained staff to ensure that the drugs are administered correctly and monitored appropriately to minimize potential side effects and ensure that serious toxicities are identified and managed. Since the drugs are largely available as liquid vials, they need to be reconstituted with calculation for a specific dose for each patient, which requires access to a laminar flow hood and appropriate training. In the absence of such facilities, reconstitution may result in aerosolization, exposure of staff to toxic drugs, and for pregnant women the risk of associated teratogenicity. In addition, most drugs can only be kept for a short time after reconstitution.

Staff training: Although treatment protocols are relatively straightforward to follow, they still require a level of expertise, infrastructure and follow-up to ensure that the drugs are given correctly and reliably in order to minimize potential side effects. Simplified treatment regimens and monitoring protocols are needed to optimize access.

4.5.4. KS diagnosis

Although definitive diagnosis of KS requires histological confirmation, in practice in many resource-limited settings, diagnosis is often based on clinical criteria alone, which carries with it the risk of both over- and underdiagnosis. There was overall support from the GDG for the importance of biopsy confirmation, especially in those patients being considered for chemotherapy (Amerson et al., 2012). Cutaneous biopsies using punch biopsy technique have been successfully performed by a range of health care providers in low-resource settings in Kenya and Uganda (Laker-Oketta et al., 2013). However, it was recognized that this will require clinical judgement; biopsy of pulmonary KS, for example, is associated with risk of haemorrhage. There is still very limited access to histopathology services, and most of these are in research settings. The requirement for biopsy should not significantly delay therapy.

4.5.5. Second-line chemotherapy

There is a need to address availability of second-line agents (as with ART) for patients that fail first-line therapy, although a formal evidence review on this topic was not considered for these guidelines. Possible second-line agents include paclitaxel, which has been used with some success in patients who have failed therapy with doxorubicin, bleomycin and vincristine, or doxorubicin in patients failing bleomycin/vincristine.

4.5.6. Need to optimize other aspects of general HIV care

Prompt access to ART and initiation of appropriate prophylaxis is critical, since the mortality of patients with KS appears to remain much higher in resource-limited settings than in high-income countries.