Tolerability of Chemotherapy in HIV-Infected Women with Breast Cancer: Are There Prognostic Implications?

Abstract

Breast cancer is the most commonly diagnosed malignancy in women, but little is known about therapeutic outcomes in patients with both breast cancer and HIV. We performed a retrospective cohort study of women with or without HIV undergoing treatment for breast cancer from 1996 to 2011. Cases with HIV were 1:2 matched to non-HIV controls based on age, sex, race, and date of cancer diagnosis. Dose reduction and/or delay during chemotherapy, overall survival, and development of metastatic disease were studied outcomes. 156 (52 HIV, 104 non-HIV) subjects were analyzed. The majority of breast cancers in both groups were clinical stages 0, I, II, and III (73%). HIV infection preceded cancer diagnosis by a median of 13 years. Median CD4 count at time of cancer diagnosis was 417 cells/mcL. Approximately 87% (45/52) were on HAART, mostly protease inhibitor-based (57%) therapy. HIV-infected women needed more dose reductions and/or delays to chemotherapy due to toxicity (56% vs. 30%; p=0.03). Stage at diagnosis, triple negative receptor status, and dose reduction and/or delay were predictors of metastatic disease and death. HIV-infected women experienced more adverse events during breast cancer treatment, and a potential causative factor could be drug–drug interactions between HAART and chemotherapy.

Introduction

Since the advent of highly active antiretroviral therapy (HAART), patients with human immunodeficiency virus (HIV) infection have seen a significant improvement in their life expectancy. While the incidence of AIDS (acquired immunodeficiency syndrome)-defining cancers is on the decline, non-AIDS defining cancers (NADC) are currently emerging as important causes of mortality and morbidity in HIV-infected patients. NADC that appear to be increasing in incidence include lung, liver, kidney, anus, head and neck, and also Hodgkin's lymphoma.¹

Interestingly, breast cancer incidence has not been shown to be higher in HIV-infected women. Linked population-based AIDS and cancer registry data in eleven regions across the United States showed a decreasing risk of breast cancer in women with increasing time since AIDS diagnosis.² A meta-analysis with study cohorts across different continents showed decreased breast cancer incidence in HIV-infected women in both pre-HAART and HAART eras when compared to the general population.³ The observed low rates of breast cancer may be due in part to underreporting or possible competing mortality.⁴ Other hypotheses include the protective effect of immunodeficiency due to HIV against breast cancer^5,6 and CXCR4 receptor tropism of HIV.⁷

While the reason for lower incidence of breast cancer in HIV-infected women remains unclear, there has also been interest in documenting the clinicopathological presentation of breast cancer in this population. A study from a large urban medical center in New York City reviewed four cases of breast cancer and found more advanced presentation of breast cancer in HIV-infected women compared to those not infected.⁸ However, this finding could be secondary to poor access of care and/or racial disparities prevalent in diagnoses of HIV and breast cancer.⁹ In 2005, a case series of five HIV-infected women with breast cancer at a public hospital in New York City was published and did not find any differences in presentation or survival compared to those not infected.¹⁰ Similar findings were also reported by other authors.¹¹

From the limited information available, it appears that HIV-infected women experience more chemotherapy-associated adverse effects compared to non-HIV-infected women. The largest case series, which consists of twenty women with HIV-infection and breast cancer, showed poor tolerance of a doxorubicin-based regimen due to significant myelosuppression.¹² Myelosuppression necessitating dose reductions of chemotherapy has also been reported by other authors.^13,14 Recently, a review of nine cases of HIV-infection and breast cancer concluded that the clinical course of breast cancer treatment may be worse for HIV-infected women due to inability to receive full doses of chemotherapy.¹⁵

There are no current analyses available that methodically evaluate the effect of HIV infection on breast cancer therapy. The aims of our study were to: (1) identify differences in breast cancer presentation and outcomes between HIV and non-HIV-infected women, (2) describe factors necessitating dose reductions and/or delays of curative chemotherapy, and (3) evaluate risk factors for mortality.

Methods

We conducted a matched retrospective cohort study including patients with all stages of breast cancer seen at Memorial Sloan Kettering Cancer Center (MSKCC), a tertiary cancer care center located in New York City. IRB approval for the study was obtained through institutional protocol.

We searched institutional medical records to identify breast cancer patients with and without HIV infection and breast cancer from 1996 to 2011. The HIV-infected case cohort consisted of women with a diagnosis of HIV during or before their cancer treatment. The non-HIV-infected control cohort of patients consisted of women without HIV who were 2:1 matched by age, race, and date of diagnosis of cancer (within 1 year) to HIV-infected cases. Patients who received second opinions and/or did not receive any part of breast cancer treatment at our hospital were excluded.

Chart review was conducted to retrieve demographic and breast cancer clinicopathologic and treatment information. Breast cancer staging and receptor information were obtained and coded using American Joint Committee on Cancer (AJCC) standards. HIV viral load, CD4 receptor positive T-lymphocyte count, history of opportunistic infections, and treatment for HIV were obtained by review of charts, medication records, and available laboratory data. Chemotherapy dose reductions and/or delays were evaluated by the first author and a breast cancer medical oncologist (LP and TT). We considered a dose delay to be a chemotherapy postponement of greater than 2 days due to toxicity. Chart review was conducted to determine the reason for these changes, and when available, the severity of toxicity was noted. The final outcome of patients (survival and disease status) was recorded at the time of last follow-up.

The primary study outcomes were dose reduction and/or delay of chemotherapy, development of metastatic disease, and death. For subjects who did not reach these clinical endpoints, analysis time was censored at the date of last known follow-up. Clinical variables assessed for predictive value of outcomes consisted of age (years), race (white vs. other), HIV, breast cancer stage, and breast cancer receptor status (triple negative vs. other). In addition, CD4 (<200 cells/mcL) and HAART regimen (protease inhibitor (PI)-based vs. other, zidovudine (AZT)-based vs. other) were assessed within the HIV-infected cohort. Logistic regression was used to evaluate the predictors of dose reduction and/or delay of chemotherapy. Cox proportional hazards regression and Kaplan-Meier survival plots were used to evaluate the predictors of progression to metastatic disease and for death. Firth's penalized likelihood method was applied to all logistic and Cox proportional hazards calculations in order to avoid divergent parameter estimates due to monotone likelihood. For all clinical endpoints, association with predictors was assessed by univariate analysis, followed by multivariate analysis of predictors with a univariate p value of 0.20 or greater. A p value less than 0.05 was considered statistically significant. All analyses were performed using R version 3.02.¹⁶

Results

Demographics and breast cancer characteristics

Between 1996 and 2011, a total of 52 women with breast cancer and HIV infection were identified. The non-HIV-infected cohort was comprised of 104 patients who were selected as detailed in the previous section. Demographic characteristics and information on breast cancer staging and receptor status of patients are presented in Table 1. The majority of the women in the study were black. The vast majority of breast cancers in both groups were of clinical stages I, II, and III (73%). Stage 0 tumors were 17% of the group, and about 5% of cancers were stage IV at the time of initial diagnosis. Patients who had missing staging data (n=7) were excluded from analysis. A greater proportion of HIV-infected women had stage I tumor (35% vs. 22%), the non-HIV-infected group had more stage III tumors (21% vs. 10%). However, these differences did not reach statistical significance. The hormone receptor distribution between groups was also similar with some interesting findings: a higher proportion of ER (estrogen receptor) positive or PR (progesterone receptor) positive tumors was seen in non-HIV-infected women (64% vs. 48%), whereas HIV-infected women had a higher proportion of triple receptor negative tumors (19% vs. 9%). Though the study design did not include matching for stage or receptor status, the distribution of these attributes was overall similar between the two groups of patients, and the differences did not reach statistical significance.

Table 1.

Demographic Attributes and Chemotherapy Distribution in Study Cohorts

Demographic characteristics	HIV-BCA^a (n=52)	Non-HIV-BCA (n=104)
Matched variables
Age	48 years	49 years
Sex	Female (100%)	Female (100%)
Race	White 19 (38%)	White 38 (36%)
	Black 33 (63%)	Black 66 (63%)
Breast cancer characteristics at diagnosis
Stage 0	7 (13%)	20 (19%)
Stage I	18 (35%)	23 (22%)
Stage II	15 (29%)	31 (30%)
Stage III	5 (10%)	22 (21%)
Stage IV	4 (8%)	4 (4%)
Receptor status
ER/PR or both positive^b	25 (48%)	67 (64%)
HER2 positive	10 (19%)	17 (16%)
Triple negative	10 (19%)	9 (9%)
Early stage breast cancer therapy
Endocrine	19/20 (95%)	61/64 (95%)
Trastuzumab	2/5 (40%)	7/12 (58%)
Chemotherapy	22/27 (81%)	63/67 (94%)
Metastatic breast cancer therapy
Endocrine	2/20 (10%)	8/64 (12.5%)
Trastuzumab	3/5 (60%)	5/12 (42%)
Chemotherapy	6/27 (22%)	12/67 (18%)
Toxicity to chemotherapy	15/27 (55%)	20/67 (30%)

BCA, breast cancer.

Receptor positivity depended on standard criteria used at time of diagnosis.

HIV characteristics

The majority (47/52; 90%) of case patients were diagnosed with HIV prior to the breast cancer diagnosis by a median of 13 years; only three patients had HIV infection and breast cancer diagnosed concurrently. Nearly 86% (45/52) of the patients were on HAART. This was either PI-based (57%) or non-nucleoside reverse transcriptase inhibitor (NNRTI)-based (37%) therapy. The nucleoside or nucleotide reverse transcriptase inhibitors included zidovudine (AZT) in 40% and tenofovir (TDF) in 55% of patients, respectively. Data parameters of disease control such as HIV viral load and CD4 count were available only for about half the patients; the median CD4 count at the time of breast cancer diagnosis was 417 cells/mcL (range: 50–860 cells/mcL; data available for 22 patients) and the median HIV viral load was 667 copies/mL (range: not detected-197,000 copies/mL; data available for 16 patients). Only four patients had a CD4 count 200 cells/mcL at the time of breast cancer diagnosis.

Breast cancer treatment

Overall HIV and non-HIV cohorts were similar with respect to proportions undergoing systemic chemotherapy or radiation therapy (Table 1). About 20% of patients had radiation therapy palliatively (data not shown). A higher proportion of non-HIV-infected women received hormonal therapy (61% vs. 39%), possibly due to more ER positive tumors in this group.

Tolerance of chemotherapy and toxicity

The HIV-infected cohort experienced chemotherapy with more adverse effects than the non-HIV-infected cohort (55% vs. 30%; p=0.03), leading to more chemotherapy dose reductions and/or delays. For both groups, the majority of the dose reductions and/or delays occurred during adjuvant or neoadjuvant chemotherapy. Table 2 details the number and types of toxicities necessitating the dose reductions and/or delays. The most common toxicities necessitating the dose reduction and/or delay were anemia and neutropenia (with or without fever).

Table 2.

Types of Adverse Events in Study Cohorts

Toxicity table	HIV-BCA N=15	Non-HIV-BCA N=20
Neutropenia	2	9
Anemia	5	0
Thrombocytopenia	0	1
Neuropathy	0	3
Fever and neutropenia	4	0
Infections^a	1	2
Hypersensitivity	1	3
Other^b	3	1

Infections included one case each of oral herpes simplex, sinusitis, and pericolonic abscess.

Other significant toxicities included chest pain, myalgias, and liver function test abnormalities.

Primary outcomes

Overall, about 17% of patients in each group progressed to metastatic disease. The overall death rate (recorded at the time of last follow-up) was similar between HIV-infected and non-HIV-infected women (29% vs. 22%). Stage at diagnosis, triple negative receptor status, and dose reduction and/or delay were found to be predictors of metastatic disease and death (Tables 3 and 4); these predictors remained as independent predictors in multivariate analysis. Kaplan-Meier survival curves were computed for death stratifying by dose reduction and/or delay in chemotherapy and showed significant differences in survival between the groups (Fig. 1). For the whole group of patients (n=156), the average follow-up time was 1472.2 days; total follow-up time among all subjects was 228,191 days.

FIG. 1.

All-cause mortality in subjects with and without dose reduction and/or delay.

Table 3.

Predictors of Metastatic Disease (Cox Regression)

	Univariate		Multivariate
Variable	Hazard ratio (95% CI)	p Value	Hazard ratio (95% CI)	p Value
Age	0.99 (0.93–1.05)	0.721
Race (white)	1.82 (0.79–4.08)	0.156	1.61 (0.66–3.90)	0.288
HIV	0.96 (0.38–2.21)	0.927
Stage (2/3 vs. 0/1)	3.84 (1.50–12.28)	0.004	2.85 (0.97–11.17)	0.057
Triple negative receptor	5.06 (2.08–11.51)	0.001	3.97 (1.58–9.47)	0.004
Dose reduction	3.14 (1.36–7.38)	0.008	3.54 (1.44–9.08)	0.006

Table 4.

Predictors of Death (Cox Regression)

	Univariate		Multivariate
Variable	Hazard ratio (95% CI)	p Value	Hazard ratio (95% CI)	p Value
Age	1 (0.96–1.05)	0.842
Race (white)	1.15 (0.58–2.2)	0.682
HIV	1.6 (0.83–3)	0.156	1.34 (0.58–2.98)	0.486
Triple negative receptor status	3.17 (1.48–6.27)	0.004	3.46 (1.53–7.46)	0.004
Dose deduction	2.63 (1.31–5.33)	0.007	3.34 (1.5–7.68)	0.003
Stage (2/3 vs. 0/1)	4.08 (1.77–11.4)	0.000	3.19 (1.11–12.39)	0.029

Predictors of dose reduction and/or delay

These were analyzed in the subset of patients who received chemotherapy by logistic regression (Table 5). Univariate analysis showed HIV status and age to be predictors of dose reduction and/or delay. Multivariate analysis only showed age to be associated with increased likelihood of dose reduction and/or delay. Within the analysis of the HIV group alone, univariate analysis did not reveal significant predictors of dose reduction and/or delay (Table 6). However, multivariate analysis revealed that HIV duration (>5 years) was associated with decreased likelihood of dose reduction and/or delay, and use of a PI in the antiviral regimen was associated with increased likelihood of dose reduction and/or delay. This finding may be explained by an association between HIV duration and a PI-based regimen. Since our multivariate analysis reveals that both variables are strong predictors but in opposing directions, the lack of statistical significance in the univariate analysis is likely due to confounding effects due to an association between HIV duration and PI. Such an association is clinically plausible. In terms of the analysis, Firth's adjustment was utilized to avoid monotone likelihood estimates; the high upper bound of the confidence interval for PI in the HIV analysis suggests that this estimate may have suffered from this issue. Zidovudine (AZT), a known myelosupressing agent, was noted to be one of the drugs in antiretroviral regimens of 4/15 patients who suffered toxicity but did not show any association in our analysis (Table 6).

Table 5.

Predictors of Dose Reduction and/or Delay in Study Population (Logistic Regression)

	Univariate		Multivariate
Variable	Odds ratio (95% CI)	p Value	Odds ratio (95% CI)	p Value
Age	1.08 (1.01–1.15)	0.02	1.09 (1.02–1.16)	0.01
Race (white)	1.24 (0.52–2.96)	0.63
HIV	2.74 (1.06–7.08)	0.04	2.65 (0.95–7.33)	0.06
Stage (IV vs. I, II and III)	4.48 (0.82–24.58)	0.08	4.33 (0.72–26.18)	0.11
Receptor (triple negative vs. other)	1.54 (0.53–4.48)	0.43

Table 6.

Predictors of Dose Reduction and/or Delay in HIV Group (Logistic Regression)

	Univariate		Multivariate
Variable	Odds ratio (95% CI)	p Value	Odds ratio (95% CI)	p Value
Age (years)	1.11 (0.99–1.31)	0.07	1.11 (0.94–1.44)	0.24
Race (white)	1.27 (0.27–6.31)	0.76
Receptor (triple negative vs. other)	1.86 (0.38–10.16)	0.45
Stage (IV vs. I, II and III)	2.13 (0.29–24.71)	0.46
HIV Duration (>5 years)	0.22 (0.03–1.17)	0.08	0.04 (0.00–0.64)	0.02
CD4 (<200)	4.09 (0.24–623.31)	0.35
HAART: (PI-based vs. other)^a	3.18 (0.66–17.64)	0.15	15.56 (1.15–2303.82)	0.04
HAART: (AZT-based vs. other)^b	1.04 (0.20–5.87)	0.96

PI, Protease Inhibitors.

AZT, Zidovudine.

Discussion

Our study is a retrospective cohort study of 52 women with HIV infection and breast cancer, focusing on presentation and outcomes of therapy. To our knowledge, this is the largest study of its kind and the only study that has systematically examined outcomes and tolerance to chemotherapy. While we did not find significant independent differences in outcome between HIV and non-HIV groups, we did find a greater prevalence of triple negative receptor tumors in the HIV-infected group (19% vs. 9%), which was in turn well-correlated with worse outcomes. Triple negative breast cancer is associated with poorer outcome and is reported to occur frequently in women of younger age and African Americans descent.¹⁷ Regarding HIV-infected women with breast cancer, Andrade et al.¹⁵ reported that four of nine women in their case series (44%) had negative hormone receptor status; other case series report 40–60% of patients having negative receptor status.^11
–13 The possibility that HIV-infected women may have a higher risk profile of tumor needs to be explored further in epidemiological studies.

While there is agreement that HIV-infected women tolerate chemotherapy to breast cancer poorly,^12,18 there is currently no systematic analysis available to identify if the subsequent dose reductions and/or delays lead to a difference in breast cancer treatment outcomes. This information is difficult to obtain as most cancer trials have excluded the HIV-infected, leading to a dearth of such data. The reasons for excluding HIV-infected patients range from concern for drug interactions or immunodeficiency due to HIV infection to just being “HIV positive” or “on HAART”. As Persad and colleagues have recommended, inclusion and exclusion criteria for clinical trials should be patient-centric and germane to the study, rather than the mass exclusion of all HIV-infected individuals.¹⁹

In our study, HIV-infected women had twice the rate of adverse events necessitating dose reductions and/or delays compared to non-HIV-infected women (56% vs. 30%). Perhaps more importantly, the HIV-infected group had more dose reductions of chemotherapy than dose delays (data not shown). Protracted neutropenia despite granulocyte colony-stimulating factor (GCSF) support and anemia dominated as the chief reasons requiring dose reductions and/or delays in chemotherapy.

Analyses of risk factors for dose reductions and/or delays in the HIV group showed being on PI class of drugs as an important risk factor in the multivariate analysis. PIs are a class of drugs that are notorious for causing drug interactions due to their close relationship with the cytochrome P450 system (CYP 450). In particular, PIs inhibit the CYP3A4 enzyme and may reduce the hepatic metabolism of alkylating agents (e.g., cyclophosphamide), anthracyclines (e.g., doxorubicin), and taxanes (e.g., paclitaxel) that are used in the treatment for breast cancer. The resulting increased levels of these cytotoxic agents may potentiate the risk for myelosuppression and in the case of taxanes, also the risk for peripheral neuropathy.²⁰

Several studies of patients treated for HIV-associated non-Hodgkin's lymphoma (NHL) allude to exacerbated cytopenias and neurotoxicity secondary to potential drug interactions between chemotherapy and PIs. A report in 2004 showed increased neutropenia in patients who received PIs along with infusional cyclophosphamide, doxorubicin, and etoposide therapy for AIDS-related non-Hodgkin's lymphoma (NHL).²¹ Vaccher and colleagues evaluated cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) plus HAART in AIDS-NHL compared to CHOP alone. While there were no differences in dose, number of cycles of chemotherapy, and response rates between the CHOP-HAART and CHOP arms, grade 3–4 anemia, and severe autonomic neurotoxicity were significantly greater in patients receiving concomitant HAART. In addition, while the grade of leukopenia was similar between the two groups, granulocyte-colony stimulating factor (GCSF) support was significantly greater in the CHOP-HAART group.²² This Italian group then specifically looked at doxorubicin pharmacokinetics and pharmacodynamics in 19 NHL patients who underwent CHOP with or without HAART. The investigators concluded that, while HAART had no significant effect on doxorubicin pharmacokinetics, toxicity may be mediated through pharmacodynamic interactions, particularly through P-glycoprotein (Pgp).²³ These proteins are involved in cellular efflux of a broad range of drugs including anthracyclines, vinca alkaloids, and paclitaxel.²⁴ A pilot study evaluating the interaction between PIs and paclitaxel in HIV-associated Kaposi's sarcoma showed that while there was increase in paclitaxel total exposure in patients on PI therapy, this did not translate into increased toxicity.²⁵ These findings highlight the complex methods of metabolism of chemotherapeutic agents and indicate that drug interactions with PIs may not be a straightforward phenomenon. For HIV patients who have undergone solid organ transplantation (SOT), there are convincing data to show that the pharmacokinetics of calcineurin inhibitors and mTOR inhibitors are significantly affected by PIs. In one study of 35 HIV-infected patients with liver and kidney transplants, patients taking PIs had marked increases in cyclosporine, sirolimus, or tacrolimus trough levels compared to those on NNRTIs alone or to patients not on antiretrovirals, necessitating either a dose reduction or an increase in dosing interval.²⁶ Management can be challenging, as evidenced by drug toxicity and also high rates of allograft rejection.²⁷ A recent retrospective review of kidney and liver transplant patients showed that a raltegravir-based regimen in the post-transplant period had minimal interactions with calcineurin inhibitors, and there were no cases of acute rejection.²⁸ Until newer immunosuppressants are available that could avoid serious drug interactions, replacing PIs with a non-PI-based regimen has been suggested for optimizing management in HIV-infected SOT recipients.²⁹

In the field of breast cancer, there is limited to no literature about chemotherapy and HAART interactions. In their case series, Hurley and colleagues noted grades 3 and 4 myelosuppression in HIV-infected women receiving doxorubicin-based chemotherapy, but is unclear about the role of HAART.¹² Out of nine women who received chemotherapy, five had an adverse event (55%), which is similar to the rate of adverse events (55%) in our study. Furthermore, the toxicities seemed unrelated to CD4 count.

Our results have shown that HIV-infected women are at greater risk for chemotherapeutic dose reduction and/or delay for breast cancer, and the reason for this may be related to the pharmacokinetic or pharmacodynamic interactions with PI and chemotherapy. Consistent with other literature in breast cancer, we also found that dose reduction and/or delay in curative chemotherapy is associated with increased risk of metastatic disease and mortality.³⁰ The HIV-infected women in our study had twice the rate of metastatic disease compared to the non-HIV-infected women (50% vs. 20%) and thus are likely to be at increased risk for mortality from breast cancer despite well-controlled HIV infection.

The limitations of our study are primarily those relevant to a retrospective observational study—ascertainment bias and misclassification. The study had relatively few numbers of patients who experienced adverse events, and this may have affected our ability to detect some of the associations. In addition, the toxicities were not graded by the standard adverse events criteria as in a prospective trial. The study findings also may not be generalizable since our center is primarily for the treatment of cancer and not for HIV care and most patients had well-controlled HIV.

Despite these disadvantages, this study is the first of its kind in reviewing systematically the tolerance to chemotherapy and outcomes in HIV-infected women with breast cancer. We have found that HIV infection is an independent risk factor for dose reductions and/or delays of curative chemotherapy, which may predispose them to a higher rate of metastatic disease and mortality. Factors such as poor bone marrow reserve, advanced AIDS, and opportunistic infections did not appear to be sufficient explanation for this finding, and drug–drug interactions may play a greater role than what was previously thought.

Oncologists, internists, and infectious disease specialists should be cognizant of drug–drug interactions with antiretrovirals and chemotherapeutic agents. Based on the preliminary results of this study, there are three aspects of patient care to consider. First, there should be close communication between the individual patient's breast oncologist and HIV provider before, during, and after administration of chemotherapy with close monitoring for adverse events. Second, there should be a low threshold to prophylactically use GSCF since neutropenia can be anticipated in HIV-infected women undergoing chemotherapy for breast cancer. Finally, changing HAART to a non-PI- and non-AZT-based regimen should be considered. Simply stopping antiretroviral therapy during receipt of chemotherapy is not a straightforward solution, as there is emerging evidence that recrudescent viral loads during HAART interruptions are detrimental.³¹ Additional studies by including HIV-infected patients in clinical trials of cancer therapy are needed to better understand the interplay between drug–drug interactions of antiretrovirals and chemotherapy and their effect on patient outcomes.

Footnotes

Acknowledgments

There are no collaborations, sources of funding, or other acknowledgments.

Author Disclosure Statement

There are no financial disclosures by any of the authors at this time.

References

Deeken

, Tjen-A-Looi

, Rudek

, Okuliar

, Young

, Little

, Dezube

. The rising challenge of non-AIDS-defining cancers in HIV-infected patients. Clin Infect Dis, 2012; 55:1228–1235.

Frisch

, Biggar

, Engels

, Goedert

; AIDS-Cancer Match Registry Study Group. Association of cancer with AIDS-related immunosuppression in adults. JAMA, 2001; 285:1736–1745.

Shiels

, Cole

, Kirk

, Poole

. A meta-analysis of the incidence of non-AIDS cancers in HIV-infected individuals. J Acquir Immune Defic Syndr, 2009; 52:611–622.

Herida

, Mary-Krause

, Kaphan

, et al. Incidence of non-AIDS-defining cancers before and during the highly active antiretroviral therapy era in a cohort of human immunodeficiency virus-infected patients. J Clin Oncol, 2003; 21:3447–3453.

Head

, Elliott

, McCoy

. Evaluation of lymphocyte immunity in breast cancer patients. Breast Cancer Res Treat, 1993; 26:77–88.

Pantanowitz

, Dezube

. Breast cancer in women with HIV/AIDS. JAMA, 2001; 285:3090–3091.

Hessol

, Napolitano

, Smith

, et al. HIV tropism and decreased risk of breast cancer. PLoS One, 2010; 5:e14349.

Voutsadakis

, Silverman

. Breast cancer in HIV-positive women: A report of four cases and review of the literature. Cancer Invest, 2002; 20:452–457.

Krown

. Breast cancer in the setting of HIV infection: Cause for concern?. Cancer Invest, 2002; 20:590–592.

10.

Oluwole

, Ali

, Shafaee

, DePaz

. Breast cancer in women with HIV/AIDS: Report of five cases with a review of the literature. J Surg Oncol, 2005; 89:23–27.

11.

Ruiz

, Davis

. Breast cancer in HIV-infected patients: A retrospective single-institution study. J Int Assoc Physicians AIDS Care (Chic), 2011; 10:30–34.

12.

Hurley

, Franco

, Gomez-Fernandez

, et al. Breast cancer and human immunodeficiency virus: A report of 20 cases. Clin Breast Cancer, 2001; 2:215–220.

13.

El-Rayes

, Berenji

, Schuman

, Philip

. Breast cancer in women with human immunodeficiency virus infection: Implications for diagnosis and therapy. Breast Cancer Res Treat, 2002; 76:111–116.

14.

Palan

, Shousha

, Krell

, Stebbing

. Breast cancer in the setting of HIV. Patholog Res Int, 2011; 2011:925712.

15.

Andrade

, Luz

, Veloso

, Cardoso

, Moreira

, Grinsztejn

, Friedman

. Breast cancer in a cohort of human immunodeficiency virus (HIV)-infected women from Rio de Janeiro, Brazil: A cases series report and an incidence rate estimate. Braz J Infect Dis, 2011; 15:387–393.

16.

R Development Core Team. The R project for statistical computing. http://www.R-project.org/ (Accessed September 11, 2013 ).

17.

Dawson S

, Provenzano

, Caldas

. Triple negative breast cancers: Clinical and prognostic implications. Eur J Cancer, 2009; 45:27–40.

18.

Guth

. Breast cancer and human immunodeficiency virus infection: Issues for the 21st century. J Womens Health (Larchmt), 2003; 12:227–232.

19.

Persad

, Little

, Grady

. Including persons with HIV infection in cancer clinical trials. J Clin Oncol, 2008; 26:1027–1032.

20.

Pham

, Flexner

. Emerging antiretroviral drug interactions. J Antimicrob Chemother, 2011; 66:235–239.

21.

Bower

, McCall-Peat

, Ryan

, et al. Protease inhibitors potentiate chemotherapy-induced neutropenia. Blood, 2004; 104:2943–2946.

22.

Vaccher

, Spina

, di Gennaro

, et al. Concomitant cyclophosphamide, doxorubicin, vincristine, and prednisone chemotherapy plus highly active antiretroviral therapy in patients with human immunodeficiency virus-related, non-Hodgkin lymphoma. Cancer, 2001; 91:155–163.

23.

Toffoli

, Corona

, Cattarossi

, Boiocchi

, Di Gennaro

, Tirelli

, Vaccher

. Effect of highly active antiretroviral therapy (HAART) on pharmacokinetics and pharmacodynamics of doxorubicin in patients with HIV-associated non-Hodgkin's lymphoma. Ann Oncol, 2004; 15:1805–1809.

24.

Loo

, Clarke

. Merck Frosst Award Lecture 1998. Molecular dissection of the human multidrug resistance P-glycoprotein. Biochem Cell Biol, 1999; 77:11–23.

25.

Cianfrocca

, Lee

, Von Roenn

, Rudek

, Dezube

, Krown

, Sparano

. Pilot study evaluating the interaction between paclitaxel and protease inhibitors in patients with human immunodeficiency virus-associated Kaposi's sarcoma: An Eastern Cooperative Oncology Group (ECOG) and AIDS Malignancy Consortium (AMC) trial. Cancer Chemother Pharmacol, 2011; 68:827–833.

26.

Frasetto

, Browne

, Cheng

, et al. Immunosuppressant pharmacokinetics and dosing modifications in HIV-1 infected liver and kidney transplant recipients. Am J Transplant, 2007; 7:2816–2820.

27.

Roland

, Barin

, Carlson

, et al. HIV-infected liver and kidney transplant recipients: 1-and 3-year outcomes. Am J Transplant, 2008; 8:355–365.

28.

Tricot

, Teicher

, Peytavin

, et al. Safety and efficacy of raltegravir in HIV-infected transplant patients cotreated with immunosuppressive drugs. Am J Transplant, 2009; 9:1946–1952.

29.

van Maarseveen

, Rogers

, Trofe-Clark

, van Zuilen

, Mudrikova

. Drug–drug interactions between antiretroviral and immunosuppressive agents in HIV-infected patients after solid organ transplantation: A review. AIDS Patient Care STDs, 2012; 26:568–581.

30.

Wildiers

, Reiser

. Relative dose intensity of chemotherapy and its impact on outcomes in patients with early breast cancer or aggressive lymphoma. Crit Rev Oncol Hematol, 2011; 77:221–240.

31.

Strategies for Management of Antiretroviral Therapy (SMART) Study Group, Lundgren

, Babiker

, et al. Inferior clinical outcome of the CD4+cell count-guided antiretroviral treatment interruption strategy in the SMART study: Role of CD4+cell counts and HIV RNA levels during follow-up. J Infect Dis, 2008; 197:1145–1155.