Characterization of HIV-Associated Hodgkin's Lymphoma in HIV-Infected Patients

Abstract

Introduction

Although the incidence and prevalence of AIDS-defining malignancies has decreased in the era of highly active antiretroviral therapy (HAART), the incidence and prevalence of Hodgkin's lymphoma (HL) in the HIV-infected population continues to rise. Compared with the general population, HIV-infected patients exhibit a 5-10-fold increased risk for developing HL.

Methods

A retrospective review of charts and electronic records from 2000-2010 at the HIV outpatient clinic (HOP)-Louisiana State University in New Orleans was conducted, and pathologically confirmed cases of HIV-HL were identified within this cohort.

Results

We found a prevalence of 6.3 cases per 1,000 patients per year of HIV-HL over a period of 10 years in our HIV outpatient clinic. The mean absolute CD4 count before treatment was 284 cells/mm³ and after treatment was 194 cells/mm³. The average time from the diagnosis of HIV infection to the diagnosis of HIV-HL was 7.6 years. The most common histopathologic type was mixed cellularity followed by lymphocytic predominance. The majority of patients had 6 cycles delivered. In terms of HL staging 87% presented with advanced stages (III B or IV). To the best of our knowledge 5 out of the 14 patients remain alive.

Conclusions

Patients in our cohort were older than most patients identified in other cohorts. All of our patients had coexisting chronic illnesses associated with inflammation, as well as detectable HIV viral loads and CD4 count >200, suggesting a role for both HIV- and non-HIV-associated inflammation in HIV-HL pathogenesis in this population. The role of HIV virus and other oncogenic viruses (EBV, HPV, and others) in the pathogenesis of Hodgkin's lymphoma in this group of patients needs to be elucidated.

Introduction

Although the incidence and prevalence of AIDS-defining malignancies has decreased in the era of highly active antiretroviral therapy (HAART), the incidence and prevalence of Hodgkin’s lymphoma (HL) in the HIV-positive population continues to rise^1–4 Compared with the general population, HIV-positive patients exhibit a 5- to 10-fold increased risk of developing HL. ^5,6 One recent study reported prevalence for HL of 6 cases per 1000 versus 6 cases per 100 000 in HIV-positive and HIV-negative populations, respectively. ⁷ Of note, HL may arise in the setting of well-controlled HIV infection. ⁵ Studies indicate a median (and non-AIDS-defining) CD4 count of 235 cells/mm ³ for patients at the time of diagnosis ⁵ and up to 50% of patients exhibit undetectable plasma HIV viral loads. ³ Therefore, even in the modern era of HAART, one would expect a relatively high incidence and prevalence of HL in US cities where there is a high prevalence of HIV infection.

Differences likely exist for pathogenesis for HL in HIV-positive and HIV-negative patients in the HAART era, although the magnitude and relevance of these differences is still somewhat unclear. Existing data suggest that HIV-initiated inflammation may play a role in the development of HL, ⁸ and an important role for Epstein–Barr virus (EBV) is indicated by its presence in nearly all HIV-associated HL (HIV + HL) cases as compared to only a minority of HL-associated HIV-negative patients. ⁶ In the pre-HAART era, unfavorable histologic subtypes—mixed cellularity (MC) and lymphocyte depletion (LD)—were predominant at the time of diagnosis. ⁹ In the HAART era, these histologic subtypes remain more common for HIV + HL, although more favorable subtypes, including nodular sclerosis (NS), are also observed for a substantial fraction of HIV + HL cases and HL arising in other settings of immune suppression. ^10–15 Some data suggest that the majority of patients with HIV + HL may be diagnosed at an advanced stage of disease,^11,16 while other data reveal no differences in the clinical presentation for HL arising in immune-suppressed and nonimmune-suppressed populations. ⁷ Immunohistochemical studies have also shown a different pathological evolution of HIV + HL compared with HIV-negative HL during treatment. ⁸

Combining HAART with chemotherapy probably yields the most favorable outcomes for HIV + HL, ^11,13,17 and successful outcomes utilizing the combination of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) have been well characterized. ^11,17 However, concomitant therapy with HAART and ABVD may elicit unwanted toxicities, including neurotoxicities which have been reported in multiple studies. ^17–19 Moreover, individual case reports indicate that a successful outcome can be achieved for HIV + HL with HAART alone in select cases. ¹¹ The majority of data reveal a survival benefit with HAART in cases of existing HIV + HL, ¹⁷ and elevation of absolute CD4 counts with HAART correlates with more favorable outcomes. ¹⁷ The risk of various cancers arising in HIV-infected patients is reduced with HAART and preserved CD4 counts, ^3,15 but as previously mentioned, the incidence of HIV + HL has actually increased. ³ Paradoxically, existing data suggest that HAART increases the incidence of HL nearly 3-fold. ²⁰ Interestingly, multivariate analyses revealed that nonnucleoside reverse transcriptase inhibitors (NNRTI) were specifically associated with an increased incidence of HIV + HL. ²⁰ One proposed mechanism is that Reed-Sternberg (RS) cell proliferation in HL tumors is dependent upon CD4 count infiltration of lesions and production of proliferative signals for RS cells. ¹⁸

In summary, conflicting data exist for pathologic determinants and the role of HAART in HIV + HL progression. In addition, many of the data characterizing HIV + HL have been acquired outside of the United States. Therefore, additional characterization of HIV + HL cases in areas of the United States with high incidences of HIV infection may help clarify these relationships. New Orleans, Louisiana, has one of the highest incidences and prevalences of HIV infection of all major cities in the United States ²¹ (Department of Health and Human Services, 2010), and we conducted a retrospective study to define clinical and pathological characteristics of HIV + HL cases identified within an academic HIV outpatient clinic in New Orleans.

Methods

A retrospective review of charts and electronic records from 2000 to 2010 at the HIV outpatient clinic—Louisiana State University in New Orleans was conducted, and pathologically confirmed cases of HIV-HL were identified within this cohort. This was done by review of electronic and pathology records; information was collected for the following: patient demographics, CD4 counts at the time of diagnosis and after treatment of HL; plasma HIV viral loads before and after diagnosis and treatment of HL; comorbidities (hypertension, chronic obstructive pulmonary disease, diabetes mellitus, and any former AIDS opportunistic infections), duration of HIV infection; smoking history; history of illicit drug use; alcohol intake; choice and duration of antiretroviral therapy; HL stage at diagnosis; and HL treatment regimens.

A database was constructed to analyze this information, and univariate and multivariate analyses were performed to identify specific clinical correlates for the presence of HL. Statistical analyses were performed to determine correlation among variables and morbidity and mortality of Hodgkin's lympoma.

Results

We found a prevalence of 6.3 cases per 1000 HIV-positive patients per year over a period of 10 years in our HIV outpatient clinic, totaling 14 patients with pathologically confirmed HIV + HL. Eighty percent of the patients were African American, 15% Hispanic, and 5% White. In all, 85% were male. The average age at diagnosis was 43 years. Unlike what has been described for HIV-negative HL, we did not observe a bimodal distribution.

The mean absolute CD4 count before treatment was 284 cells/mm³ and after treatment was 194 cells/mm³. Plasma HIV viral load ranged from 89 to 210 000 copies/mL before HL treatment and from 344 to 302 000 copies/mL after HL treatment.

The average time from the diagnosis of HIV infection to the diagnosis of HIV + HL was 7.6 years. The mean hemoglobin was 9.8 g/dL, with an average of 2.8 g/dL for albumin. All patients demonstrated at least 1 comorbidity (hypertension, chronic obstructive pulmonary disease, and diabetes mellitus). Two patients (14%) had other diagnosed cancers, including cervical and lung cancers. Six (43%) of the patients were diagnosed with an AIDS-associated opportunistic infection before the diagnosis of HIV + HL. Twenty-one percent had family history of cancer.

The most common histopathologic type was mixed cellularity followed by lymphocytic predominance. It is important to note that in our small case series, 2 patients did not have any information in relation to histopathology. Eighty-six percent of the patients received chemotherapy treatment with ABVD. The majority of patients had 6 cycles delivered. In terms of HL staging, 87% presented with advanced stages (III B or IV). To the best of our knowledge, 4 of the 14 patients remain alive.

A great proportion of patients presented with weight loss, fever, and fatigue (10 of 14). Eighty-seven percent of patients were of antiretroviral treatment (mainly NRTI and NNRTI combination). In 13% of the patients, information was not available. Tables 1 and 2 show information on patients and their characteristics.

Table 1.

Demographics and Characteristics of HIV + HL Patients

Characteristic	Number (%)
Male	12 (86)
Female	2 (14)
Average age for males at time of HL diagnosis, y	42.25
Average age for females at time of HL diagnosis, y	47
CD4 count before diagnosis of HL, cells/mm3	283
CD4 count after diagnosis of HL, cells/mm3	194
Comorbidities	14 (100)
AIDS opportunistic infections	6 (43)
Pneumocystis pneumonia	4 (29)
CNS toxoplasmosis	1 (7)
CNS cryptococcal meningitis	1 (7)
Other cancers	2 (14)
Smoking	10 (71)
Drug use (cocaine and marijuana)	6 (43)
Alcohol use (1-6 drinks daily)	6 (43)
Symptoms
Night sweats	10 (71)
Fever	5 (36)
Weight loss	4 (29)
Fatigue	3 (21)
Histopathology type
Mixed cellularity	9 (65)
Lymphocytic predominance	2 (14)
Nodular sclerosis	1 (7)
Unknown	2 (14)
Stage at time of HL diagnosis
IIIA	4 (29)
IIIB	2 (14)
IV	6 (43)
Unavailable pathology	2 (14)
Number of years with HIV infection at  the time of diagnosis	7.6

Abbreviations: HIV, human immunodeficiency virus; HL, Hodgkin's lymphoma; CNS, central nervous system.

Table 2.

Hodgkin's Lymphoma Stage and Type and Chemotherapy Regimen of HIV + HL Patients

Sex	Age at Dx of HL, y	Years with HIV Infection at Time of HL Diagnosis, y	AIDS Opportunistic Infection	HL Stage and Type	Chemotherapy Regimen	Survival
M	46	5	N	IIIA; mixed cellularity, lymphocytic predominance	ABVD, 6 cycles	Alive
F	59	5	N	Unknown	Unknown	Dead
M	39	11	Crytococcal meningitis	IIIB; mixed cellularity and lymphocytic predominance	ABVD, 5 cycles	Dead
M	40	Unknown	Pneumocystis pneumonia	IIIA; mixed cellularity	ABVD, 6 cycles	Dead
M	40	8	N	IIIA; mixed cellularity	ABVD, 6 cycles	Dead
M	45	5	N	IV; mixed cellularity	ABVD, 6 cycles	Dead
M	38	13	N	Unknown	Unknown	Dead
M	47	7	Pneumocystis pneumonia	IV; mixed cellularity	ABVD, 5 cycles	Dead
M	45	8	Central nervous system toxoplasmosis	IV; mixed cellularity	ABVD, 6 cycles	Dead
M	45	10	N	IIIB; mixed cellularity	ABVD, 5 cycles	Alive
M	35	10	N	IIA; mixed cellularity	ABVD, 8 cycles	Alive
F	41	16	Pneumocystis pneumonia	IV; mixed cellularity	ABVD, 3 cycles	Dead
M	52	1	Pneumocystis pneumonia	IV; mixed cellularity	ABVD, 6 cycles	Dead
M	29	1	N	IV; nodular sclerosis	ABVD, 6 cycles	Alive

Abbreviations: HIV, human immunodeficiency virus; HL, Hodgkin's lymphoma; CNS, central nervous system.

Discussion

Our small case series of patients with HIV + HL exhibited similarities and differences to previously documented series. The most common histopathology was mixed cellularity and lymphocytic predominance type which have also been found in other studies. The presence of these 2 subtypes has also been found in other studies.^7,22 The majority of our patients presented with stages IIIB or IV, which may indicate an increased aggressiveness of this condition in HIV-positive patients. ^7,23 Interestingly, a great deal of patients was symptomatic. This is a finding reported elsewhere. The fact that HIV-positive patients are more symptomatic than their HIV-negative counterparts is an issue that needs more understanding. The latter has also been found in other studies and perhaps the explanation for that might be the inflammatory background found in HIV + HL ⁷ .

 The prevalence of HL in our study is close to what has been reported previously. ¹¹ The majority of our patients was of African American race and male. These 2 findings have also been reported in other studies of HIV-positive patients. ¹¹ Our cohort average age was older compared to other reports. ¹³ One potential reason might be that our data captured just 1 wave of a bimodal distribution, which has been reported in cases of HIV-negative HL.

The average CD4 counts in our study were very similar to other studies. ^2,13 Even though some authors refer that a range of 150 to 200 cells/mm³ is critical for a higher risk of HL, others have mentioned that it is the decrease in CD4 counts while patients are on antiretroviral therapy that increases the risk of developing HL. ²⁴ Another possibility may be declining CD4 counts below 200 at any time are a harbinger of HIV + HL, the same way that this decrease has been related to other cancers and opportunistic infectious illnesses in HIV-positive patients. It is interesting to note that in our cohort, the average CD4 count decreased after treatment of HL. This may reflect the effects of chemotherapy cycles, drug–drug interactions and as a result of this less availability of antiretrovirals, or noncompliance with medications. The same situation may apply for an increased HIV viral load that ranges before and after the treatment of HL. This increase is interesting and may be related to noncompliance with antiretroviral therapy, chemotherapy agents that may induce HIV replication, and chemotherapy interventions that may reactivate other viruses (such as EBV and other herpesviruses). No EBV studies were performed in our patients. To the best of our knowledge, no patient had coexisting HPV disease (anal/cervical carcinoma [CA] and warts).

Even though patients were on antiretrovirals, their viral loads were not under good control which may account for their less than optimal response to HL treatments. This may have affected survival. This issue has been previously mentioned by some reports^11,24

The issue of average number of years with HIV infection before starting the diagnosis of HIV-HL deserves further attention. This might be a proxy for CD4 nadir counts, length of immune suppression, or longer time of interaction among HIV virus, oncogenic viruses, and HL. Some studies have looked into the impact of CD4 counts at the time of diagnosis of HL. Interestingly, some studies showed a higher incidence of HL with CD4 counts >200 cells/mm³. More interesting is the fact that this correlation is lost when there is a profound immune suppression (less cases of HL in severely immune-suppressed patients), ²⁵ suggesting a supportive role for a partially existing immune system. ¹⁸

The role of comorbidities in this population has not been determined. In our study, 100% of our patients presented with other medical issues. The factoring of this and the presence of AIDS opportunistic infections in our cohort which in turn may predispose our patients to have a worse clinical course of HIV infection is unknown. At the same time, the relationship between comorbidities and chronic inflammation with resultant premalignant conditions in HIV-positive patients has not been elucidated.

The majority but not all of our patients was on antiretroviral therapy, which may reflect noncompliance, drug interactions, or side effects. The implications for lack of treatment might be crucial in response to HL treatment and prognosis in our cohort. Survival might have also been affected by this lack of compliance.

Limitations

An important limitation in this study is the small number of patients. The retrospective nature of this cohort is another limiting factor that would limit our ability to derive conclusions from this study. Information was collected from electronic medical records and as a result of that information at certain times was not available. An evaluation of pre- and post-antiretroviral therapy eras in relation to prevalence of HL could not be performed because of the lack of archived charts.

Conclusions

Our single-center cohort of patients with HIV + HL resembled other cohorts by prevalence, histopathology, more frequent constitutional symptoms, and diagnosis at advanced stages. Patients in our small case series were older than most patients identified in other series. ^{1 –4} All of our patients had coexisting chronic illnesses associated with inflammation, as well as detectable HIV viral loads and CD4 count >200, suggesting a role for both HIV infection and non-HIV-associated inflammation in HIV + HL pathogenesis in this population. The identity of specific risk factors for HIV + HL, which may be more prevalent in older patients harboring other chronic inflammatory conditions, needs to be determined in larger prospective trials, and this might lead to the development of more effective treatments and preventative measures for HIV + HL. Inflammatory conditions, including HIV replication at higher CD4 counts and presence of other diseases, might set up a rich tumor microenvironment for aberrant B cell development. Rising HIV viral loads while on chemotherapy seems to be a factor indicating suboptimal HIV + HL treatment and deserves further investigation.

The role of HIV virus and other oncogenic viruses (EBV, HPV, and others) in the pathogenesis of HL in this group of patients needs to be elucidated. In addition, the role of various comorbidities in the development of chronic inflammation or chronic immunogenic stimulation needs to be studied.

The most common chemotherapy regimen was ABVD. Side effects and drug–drug interactions have been studied in some reports, but overall this regimen seems to be safe and effective in HIV-positive patients. The role of antiretroviral therapy combined with chemotherapy regimens needs to be studied in prospective studies that include mortality and morbidity outcomes

Hodgkin’s lymphoma seems to be more aggressive in this population and survival appears to be shorter compared with HIV-negative patients. More prospective trials are needed to elucidate these issues

Finally, more prospective studies are needed to clarify issues such as population at risk, interactions between antiretroviral therapy and chemotherapy regimens, best means of diagnosis and follow-up, survival outcomes, and best ways to prevent it.