Predictors of response of AIDS-associated Kaposi sarcoma to standard chemotherapy.
El-Koha, Omran Ali.
MetadataShow full item record
Predictors of response of AIDS-associated Kaposi-Sarcoma to standard chemotherapy Overview: Kaposi Sarcoma is the most common HIV-associated cancer. Its etiology and pathogenesis is not fully understood. Little is known about what predicts prognosis, survival and therapeutic response in HIV-KS. In South Africa given the high seroprevalence rates of HIV-l and human herpes virus 8 (HHV 8), Kaposi's sarcoma is a significant problem. The majority of patients have been treated solely with palliation due to the poor outcome associated with a diagnosis of HIV-KS, more so in the absence of highly active antiretroviral therapy (HAART). Since the national ARV rollout programme and the availability and accessibility of HAART to all patients with a diagnosis of HIV-KS, a new strategy has to be established to enable adequate patient selection for chemotherapy. There have been a few published studies addressing the predictors of response to chemotherapy in the first world. However, this is the first study of these factors in HIV-l infected African patients with Kaposi's sarcoma. Aim: To identify and assess the potential value of several parameters predictive of outcome, survival and therapeutic response in HIV- infected patients with KS. Clinical, hematological, biochemical, immunological and virological variables were evaluated. Methods: We collected data from 25 patients with AIDS-KS who were enrolled in a phase III randomized controlled trial comparing HAART alone with the combination of HAART and chemotherapy. All patients were from the combination therapy arm. The following variables were evaluated as predictors of prognosis and therapeutic response: age, gender, ethnic origin, Haemoglobin (Hb), white blood cells (WBCs), lymphocytes, neutrophils, platelets, S.albumin, ALP, GGT, CD4 count, HIV viral load. These variables were assessed in patients at baseline and month 6 of therapy. Patients were staged into good risk and poor risk according to the AIDS clinical trial group (ACTG) criteria. The outcomes assessed were response to treatment and mortality. Results: A total of 25 patients participated to the study. Of these 16(64%) were males and 9(36%) were females, with male: female ratio of 2.7:1. Median age was 34 years (24-47); all patients were of Black African origin. Of the 21 patients, 15 (71.4%) were of good prognosis and 6(28.6%) were of poor prognosis. At baseline the median values of the different variables were as follows: Hb 10.9 g/dl, WBCs 5.95x109/L, lymphocytes 1.7 x109/L, neutrophils 3 x10 9 /L, platelets 272 x10 9 /L, S.albumin 30 gil, total protein 88 gil, ALP 64 U/L, and GTT 21 U/L, CD4 count was 255 cells/mm 3 , HIV-RNA viral load was 42000( 4.610gs). At month 6, 22 patients remained alive, their median values were: Hb 12.2 g/dl, WBCs 4.65 x109/L, lymphocytes 1.5 x109/L, neutrophils 3 x10 9 /L, platelets 301 x109/L, S.albumin 36.5 gil, total protein 84.5 gil, ALP 78.5 U/L, GTT 44.5 U/L, CD4 count 288 cells/mm3 , HIV-RNA viral load was 50500( 4.6910gs). The baseline median CD4 and HIV-RNA viral load counts for the 3 patients who died before month 6 were 47 cells/mm3 and 31000(4.610gs); respectively. Response to therapy was evaluated in 21(84%) patients as 4(16%) patients were missing, of the 21 patients 3 (14.3%) had complete response and 18(85.7%) had partial response. With respect to sex 2(14.3%) males had complete response and 12(85.7%) had partial response, 1(14.3%) female had complete response and 6 (85.7%) had partial response. Non-parametric statistics were used because of the small sample size and the skewness of the data. Variables were described using medians and ranges, and compared between two independent groups using Mann-Whitney tests. Baseline and month 6 comparisons were done using Wilcoxon signed ranks tests. Receiver Operating Characteristic (ROC) curves were used to analyze cut points to optimize sensitivity and specificity of a quantitative variable for a dichotomous outcome. Discussion In the univariate analysis age and sex didn't influence prognosis and therapeutic response, the influence of ethnic origin couldn't be assessed as all patients were of the same ethnic origin. Baseline WBCs (P= 0.004) and lymphocytes (P=0.026) were significantly associated with complete response. Higher values of GGT (p=O.OOl); ALP (P=0.006) were associated with more deaths. Baseline CD4 count and HIV viral load were not of predictive value, lthough change CD4 (P=002) and VL (p=.OOO) over time was significant and most likely attributed to response to therapy. 90.9 % of patients reached undetectable HIV-l Viral loads at month 6. CONCLUSION: Neither CD4 count nor HIV viral load at baseline predicted prognosis or survival; however there was a borderline significance of CD4 (P=0.058) towards a better survival.