Abstract
HIV-associated immune thrombocytopenic purpura (ITP) has decreased in incidence 10-fold since the advent of highly active antiretroviral therapy (HAART). For patients with detectable HIV viral loads, first-line treatment approaches involve optimizing HAART followed by standard ITP options used to treat those without HIV infection. In the general population, the thrombopoetin receptor agonists (TRAs), eltrombopag and romiplostim, are effective when used as salvage ITP therapy. In addition, eltrombopag has been used effectively in patients with thrombocytopenia secondary to hepatitis C—a virus seen commonly in HIV-infected patients, especially in those who also have a history of intravenous drug use. There are, however, few reports or studies of TRAs use in those with HIV infection. Herein, we describe 5 cases of refractory HIV-associated ITP managed with TRAs. Although platelet counts improved for all patients, 2 patients succumbed to thromboembolic complications. Our initial experience, as well as our findings from a Medline review, supports the potential utility of TRA as salvage therapy in the treatment of HIV-related ITP; however, we recommend caution in the use of these agents in those who are at highest risk of thrombosis. Additional studies are needed to determine the efficacy and, more importantly, the safety of TRAs in treatment of HIV-associated ITP.
Introduction
In the pre–highly active antiretroviral therapy (HAART) era, immune thrombocytopenic purpura (ITP) was a relatively common hematologic abnormality, which was seen in as many as 30% of patients infected with HIV. 1 –3 The incidence varied according to the definition of thrombocytopenia and the characteristics of the baseline population; for example, the disease was more common in HIV-infected intravenous drug users (IVDUs) compared to HIV-infected men who have sex with men and who are not IVDUs. 1 Although thrombocytopenia may occur at any time during the course of HIV infection, the incidence generally correlates with the degree of immunosuppression and is more prevalent in individuals with clinical AIDS. 2 –4
With the advent of HAART, the incidence of HIV-associated ITP has been reduced substantially, with some epidemiological studies showing contemporary rates as low as 1% to 3%. 5 –7 In cases where HAART does not lead to adequate improvement in platelet counts, next lines of therapy have traditionally consisted of corticosteroids, including dexamethasone and prednisone, intravenous immunoglobulin (IVIG), anti-Rho(D) immune globulin, rituximab and, in the current era, more sparingly, splenectomy. 8 –10 When platelet counts remain persistently <20 × 109/L despite medical and surgical interventions, patients with ITP retain a heightened risk of bleeding complications. 11 Once conventional options are exhausted, successes of alternative strategies for treatment of HIV-associated ITP are largely anecdotal.
The thrombopoetin receptor agonists (TRAs), eltrombopag (Promacta, GlaxoSmithKline; Brentford, England) and romiplostim (Nplate, Amgen; Seattle, Washington), were granted approval from the US Food and Drug Administration for the treatment of refractory ITP. These agents are effective in 59% to 88% of ITP cases, and loss of response while on continued therapy is uncommon. 12,13 Additionally, eltrombopag has been shown to be effective in the management of platelet counts in patients with hepatitis C, a virus seen commonly in HIV-infected patients. 14 The agents’ use in the treatment of refractory HIV-associated ITP has not, however, been sufficiently studied or reported. We describe the use of TRAs in 5 patients with HIV-associated ITP who were refractory to multiple lines of treatment. We also briefly review the current literature regarding HIV-associated ITP and comment on the use of these newly available drugs for those with refractory ITP.
Case Reports
Case 1
A 32-year-old Caucasian male in otherwise good health presented for medical attention complaining of easy bruising. Save for ecchymosis involving his trunk and legs, his physical examination was unremarkable. The patient’s white blood cell and hematocrit were within normal limits, as were chemistry and hepatic blood panels; however, his platelet count was 8 × 109/L. Additional testing revealed an HIV viral load of 5000 copies/mL and a CD4 count of 261 cells/mm3. Ancillary studies included an erythrocyte sedimentation rate of 1 mm/h and negative serology for hepatitis B and C viruses, Helicobacter pylori, and rheumatoid factor and antineutrophil antibodies. The patient was initially prescribed 50 mg of prednisone daily and, in light of extensive ecchymosis and moderate mucosal bleeding, he was also given a single 1 g/kg dose of IVIG. Platelet counts improved rapidly to 129 × 109/L, but 1 week later they declined to 52 × 109/L. He began a HAART regimen consisting of efavirenz (EFV), tenofovir (TDF), and disoproxil fumarate as a once-daily combination pill (Atripla; Gilead, Foster City, California). With HAART, the patient rapidly achieved a nondetectable HIV viral load and his platelet counts transiently increased to 155 × 109/L, only to fall to 13 × 109/L 4 weeks later as corticosteroids were gradually tapered. Efforts to improve platelet counts with pulse dexamethasone (40 mg daily for 4 days) and rituximab (375 mg/m2 weekly × 4 weeks) proved ineffective (Table 1).
Characteristics of HIV-Related ITP Patients Given TRA.
Abbreviations: TRA, thrombopoetin receptor agonist; Atripla, efavirenz, tenofovir, and disoproxil fumarate; IVIGs, intravenous immunoglobulins; Pred, prednisone; Dex, dexamethasone; f/u, follow-up clinic visit; HAART, highly active antiretroviral therapy; ITP, immune thrombocytopenic purpura.
aPatient died prior to 6-month f/u due to myocardial infarction.
bPatient sustained response for 2 weeks after discontinuation of TRA but died post-second surgery.
cPatient to be reinitiated on maintenance therapy due to lack of compliance.
dInformation abstracted from retrospective chart review of 32 patients; information limited for HIV-ITP patient. However, a response to TRA was noted.
The patient had a platelet count of 20 × 109/L when he began romiplostim at a dose of 1 μg/kg. However, his compliance with HAART proved variable and his HIV viral load soon increased to >100 000 copies/mL. His platelet count again fell to <50 × 109/L. Over the course of 2 months, romiplostim was gradually increased in a stepwise fashion to 4 μg/kg/wk, which produced a platelet count of 143 × 109/L. Romiplostim was then reduced according to package insert guidelines. With further attention to achieving compliance with HAART, the patient was able to discontinue romiplostim after 7 months. Three years later, he has a normal platelet count, his CD4 count is 600 cells/mm3 and his HIV viral load remains <40 copies/mL.
Case 2
A 43-year-old African American female with a history of morbid obesity, hypertension, coronary artery disease, and congestive heart failure sought consultation for easy bruising. Her examination was notable only for a body mass index of 58 and lower extremity ecchymosis. Her white blood count was 7.3 × 109 cells/L, hemoglobin 12.1 g/dL, and her platelet count was 2 × 109/L. The peripheral blood smear was notable for a diminished number of platelets with an increased mean platelet volume. An HIV enzyme-linked immunosorbent assay and Western blot returned positive. Her CD4 count was 283 cells/mm3 and her HIV viral load was 74 743 copies/mL. A bone marrow biopsy showed megakaryocytic hyperplasia, and abdominal imaging studies did not suggest hepatic cirrhosis, splenomegaly, or lymphadenopathy. Three months after beginning Atripla, her HIV viral load was nondetectable and her CD4 count was 300 cells/mm3. The patient did not, however, achieve a consistent or sustained improvement in platelet count despite taking prednisone at a dose of 1 mg/kg. Other lines of ITP treatment followed, but her platelet counts remained <20 × 109/L (Table 1).
The patient was started on romiplostim at a dose of 1 μg/kg/wk, but an adequate platelet response was not observed. Due to sporadic compliance with clinic visits, she was switched to eltrombopag at a dose of 50 mg daily. Two weeks later, her platelet count improved to 133 × 109/L. She continued to take this medication for 12 additional months, during which time her platelet counts varied from a low 60 × 109/L to high as 135 × 109/L. She subsequently presented with sub-sternal chest pain, ST-segment elevation, and elevated cardiac enzymes. An angiogram revealed a critical stenosis of both the proximal left-anterior descending and mid-right coronary arteries. She underwent successful double-vessel angioplasty and stent placement. Eltrombopag was temporarily discontinued and she was discharged to home with clopidrogrel, lisinopril, carvedilol, and 81 mg aspirin. Eltrombopag was reinitiated when she was seen in follow-up 10 days later in an effort to support adequate platelet counts while she remained on antiplatelet therapy. Her TRA was titrated to maintain a platelet count of 60 × 109/L. However, 5 months later she succumbed from complications of a second myocardial infarction.
Case 3
A 47-year-old Caucasian male had been on HAART (darunavir, raltegravir, ritonavir [RTV], emtracitabine, and TDF) for 3 months when he sought evaluation of diffuse bruising. He had isolated thrombocytopenia with a platelet count of 13 × 109/L. His CD4 count was 156 cells/mm3 and his HIV viral load was 60 copies/mL. To minimize his pill burden, the patient was switched to Atripla and, within 3 months, he achieved a nondetectable HIV viral load and his CD4 count increased to 300 cells/mm3. However, his platelet count did not improve. Six months after beginning HAART, the patient was hospitalized for recurrent ecchymosis, hematochezia, epistaxis, and a platelet count of 10 × 109/L. His peripheral blood smear was remarkable for large platelets and a bone marrow aspirate showed increased megakaryocytes. He was treated initially with prednisone (1 mg/kg), followed by several other lines of treatment, but without improvement in platelet count or in clinical course (Table 1). After beginning romiplostim at a dose of 1 μg/kg/wk, his platelet counts rose to 40 × 109/L; at a dose of 2 μg/kg/wk his platelets increased to 180 × 109/L. After an additional month of sustained normal platelet counts, romiplostim was tapered and ultimately discontinued. Three years later, the patient’s platelet count remains within normal limits, his HIV viral load is <40 copies/mL, and his CD4 count is 420 cells/mm3.
Case 4
A 32-year-old HIV-infected Caucasian male began HAART (zidovudine [ZDV], epivir, and nevirapine [NVP]) after he was diagnosed with pneumocystis jiroveci pneumonia. His initial CD4 count was 100 cells/mm3 and his initial HIV viral load was >100 000 copies/mL. For the next 15 years, his HIV viral load remained nondetectable, but his CD4 count remained <200 cells/mm3. During this time, the patient remained well and without a bleeding diathesis, though his platelet counts hovered between 40 and 60 × 109/L. In anticipation of a lumbar laminectomy for treatment of progressive back discomfort and lower extremity weakness, attention was placed on improving platelet counts. A detailed hematologic assessment, including serologic tests for parvovirus B-19, viral hepatitis, and autoimmune antibodies returned unremarkable and a bone marrow aspirate and biopsy showed normal cellularity with abundant megakaryocytes. Intravenous immunoglobulin, prednisone, and rituximab were sequentially prescribed, but each was unsuccessful in improving platelet counts (Table 1). With a platelet count of 39 × 109/L, the patient began romiplostim at a dose 1 μg/kg/wk. His platelet counts remained <50 × 109/L until romiplostim was increased to 4 μg/kg/wk. At this dose, his platelet count improved to 164 × 109/L. Romiplostim was tapered and then discontinued 3 weeks postsuccessful surgery. Four weeks later, his platelet count had returned to 53 × 109/L.
A year later, the patient required further lumbar surgery. His platelet count at that time was 49 × 109/L. Romiplostim was reinitiated at 2 μg/kg for 1 week, followed by an increase to 3 μg/kg for 2 more weeks. At the time of surgery 3 weeks later, his platelet count had increased to 120 × 109/L. He tolerated surgery without undue bleeding, but in postoperative care, 1 day following the procedure, he became acutely hypoxic and suffered a respiratory arrest from a presumed pulmonary embolus and died. The patient’s family declined a request for an autopsy.
Case 5
An 18-year-old female was diagnosed with HIV at birth and with ITP and hepatitis C virus beginning at the age of 10 years. She was admitted to the hospital for treatment of persistent epistaxis. Her platelet count was 13 × 109/L, her CD4 count was 200 cells/mm3 and her HIV viral load was 54 000 copies/mL. As an infant and a young child, the patient’s HIV viral load had been well suppressed, but as an adolescent, she struggled with staying compliant with HAART and had acquired multidrug-resistant HIV infection. Over the previous 8 years, she received various treatments for HIV-associated ITP with minimal or no response to corticosteroids, anti-Rho(D) immune globulin, rituximab and, with modest benefit, to periodic doses of IVIG (Table 1). The patient began eltrombopag at a dose of 50 mg daily and 2 weeks later her platelet count improved to 107 × 109/L. Despite a persistently detectable HIV viral load, she has been able to maintain adequate platelet counts while on eltrombopag.
Discussion
Eltrombopag and romiplostim have been well studied in several clinical trials evaluating their effectiveness in treating various subsets of patients with problematic ITP. 12-13,18 However, these studies have not detailed the use and efficacy of these agents in the treatment of HIV-associated ITP. For patients with HIV-associated ITP, there is a substantial need for additional treatment options beyond the traditional ITP therapies, as platelet counts often remain inadequate despite multiple interventions. In 1 retrospective chart review, 23 (71%) of 31 patients with HIV-associated ITP in the HAART era required secondary or deeper lines of treatment for failure to respond to conventional ITP options. 8 Our initial experience with eltrombopag and romiplostim suggests that these drugs also improve platelet counts in the context of refractory HIV-associated ITP.
The pathophysiology of HIV-associated ITP has not been fully elucidated but is likely somewhat different than that of classic ITP. Severe T-cell depletion and consequent immune dysregulation occurs with HIV infection. In addition, megakaryocytes express HIV receptors CD4, CXCR4, and CCR5 and, therefore, may be vulnerable to direct viral cytotoxicity, which ultimately could lead to a decrease in platelet formation. 19 Antiplatelet immune complexes are also present on platelets of HIV-infected patients and may be an important contributor to ITP pathogenesis. 19 Both romiplostim and eltrombopag act on the thrombopoetin receptor on hematopoietic stem cells, the precursor to megakaryocytes and the formation of platelets. Although the precise etiology of classic and HIV-associated ITP may differ, the responses observed in our patients to these agents appear comparable to those seen in non-HIV-infected patients.
Our report is one of only a few detailing the beneficial use of romiplostim and eltrombopag for the treatment of refractory HIV-associated ITP. Utilizing a MEDLINE/PubMed search with the key words “HIV,” “immune thrombocytopenia purpura,” “romiplostim,” and “eltrombopag” and restricted to English language reports between the years 2008 and 2014, we identified 3 additional reports of successful use of TRAs in patients with HIV-related thrombocytopenia (Table 1). 15 –17 Similar to our experience with patients 1 and 3, in 2 instances patients responded well to TRAs and were later able to be weaned off these medications in the context of nondetectable HIV viral loads and normal CD4 counts.
Recently, Ochoa and colleagues briefly described 5 patients with refractory HIV-related ITP who received romiplostim at the starting dose of 1 µg/kg/wk and titrated to maintain a platelet count of around 50 × 109/L per package guidelines. 20 The mean patient age was 48 years and 4 were males. All 5 patients had failed or relapsed after prednisone; 3 of the patients had also failed multiple treatments, including splenectomy, IVIG, Rho(D) immune globulin, rituximab, danazole, and vincristine. All but one of the patients were on HAART by the time of ITP treatment and 2 patients had CD4 counts <200 cells/mm3. One patient requested to be switched to eltrombopag and developed an acute lower extremity deep-vein thrombosis and symptomatic pulmonary embolism 1 week after starting the medication. Mean baseline platelet count was 27 × 109/L (13-40 × 109/L), and it increased to 66 × 109/L (45-86 × 109/L) within 1 week and 76 × 109/L (45-107 × 109/L) after 1 month of therapy. Four of the patients had durable responses for ≥12 months not requiring dose adjustments or modifications; the remaining patient was poorly compliant and missed several doses of romiplostim, but his platelet count improved every time the medication was restarted. Like our experience with patients 4 and 5, 2 of the their 5 patients had CD4 counts <200 cells/mm3, and 1 had a detectable HIV viral load. Their positive response to TRAs suggests that those agents may be effective even in patients with marked T-cell depletion and HIV viremia. This may be an important attribute in the subset of patients who are already heavily immunosuppressed and who may be particularly vulnerable to the added immunosuppression of conventional ITP regimens.
Patients with HIV-associated ITP may be unresponsive to 1 TRA but still respond to the other as was seen in patient 2. Aoki and colleagues also described a patient who was successfully switched from eltrombopag to romiplostim after failing to respond to the first agent. 21 Further investigation is necessary to determine why, in certain circumstances, 1 TRA is more effective than the other.
Side effects associated with TRAs are well documented. An increase in bone marrow reticulin fibrosis has been reported in 5% of patients who are treated with TRAs, but to date there are no disconcerting signals to suggest that the marrow fibrosis is progressive or irreversible. 22 –24 Eltrombopag also carries a warning because of increased risk of hepatotoxicity. 25 Additionally, portal vein thrombosis has been reported in patients with chronic liver disease receiving romiplostim. The risk of thrombosis associated with these agents has been estimated to be around 5.5%. 13,26,27 As evidenced by patient 2 who suffered successive myocardial infarctions while on eltrombopag and the patient described by Ochoa and colleagues who developed an acute lower extremity DVT and symptomatic pulmonary embolus within a week of starting eltrombopag treatment, care must be taken when prescribing these agents to patients with elevated thromboembolic risk factors including hypertension, smoking, coronary artery disease, and obesity. 20
Given that patient 2 had a history of congestive heart failure and coronary artery disease, it is difficult to establish a clear cause-and-effect relationship between the TRA and the cardiac events. Nonetheless, patients with ITP have a potential risk of thrombosis, particularly when treatment leads to platelet count increases above 400 × 109/L. 12,28
Patient 4, who died from complications of a presumed pulmonary embolus shortly after completing lumbar spine surgery, further emphasizes the importance of using these TRAs cautiously during periods of heightened risk of thrombosis such as postsurgery and during periods of immobility. However, due to insufficient evidence, a definitive association between TRA use and these 2 patient thrombotic events remains speculative. As a precautionary note, to minimize the risk for thrombotic/thromboembolic complications, do not use TRAs in an attempt to normalize platelet counts. Follow the dose adjustment guidelines to achieve and maintain a platelet count of roughly 50 × 109/L.
The patients described in this report and in our literature review presented with signs or symptoms of refractory HIV-associated ITP. Several of these individuals had been on a stable HAART regimen and all had undergone several standard treatments for ITP before receiving eltrombopag or romiplostim. Overall, they had a mean pretreatment platelet count of 15 × 109/L, and this increased to a mean of 110 × 109/L while on TRAs.
An important limitation of our analyses is that it is retrospective and is subject to inherent biases associated with such observations. 29 Due to this and the small number of patients yet described in the medical literature, we cannot make definitive conclusions about the efficacy or safety of these agents in HIV-related ITP.
While HIV-associated ITP remains an important clinical problem in the era of widespread HAART use, the availability of both romiplostim and eltrombopag gives medical providers additional options in treating this disease. Further experience is needed, however, to better determine the effectiveness and, most importantly, the safety of these drugs in the context of HIV-associated ITP.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.