Aortitis Associated with Prophylactic Short-acting Granulocyte Colony-stimulating Factor Administration: A Case Report and Review of the Literature

Abstract

We herein report an 83-year-old woman with filgrastim-associated aortitis during chemotherapy for relapsed diffuse large B-cell lymphoma. She had been treated with filgrastim as a prophylaxis for neutropenia during the fourth cycle of chemotherapy from day 9 to 18. On day 21, she developed a fever. Contrast-enhanced computed tomography revealed aortitis of the descending aorta. The fever abated with non-steroidal anti-inflammatory drug treatment. A literature review identified a small number of aortitis cases all caused by prophylactic use of granulocyte colony-stimulating factors (G-CSFs), among which short-acting filgrastim was rarely encountered. The present and previous findings imply a possible relationship between aortitis and prophylactic G-CSF usage.

Introduction

Since the first study reporting that granulocyte colony-stimulating factor (G-CSF) reduced the myelosuppressive effects of chemotherapy in 1988 (1), these agents have been widely used to treat chemotherapy-induced neutropenia (2). Both short-acting G-CSFs and pegylated long-acting G-CSFs are utilized to shorten the duration of neutropenia, thus reducing the incidence of febrile neutropenia (FN) in patients with non-Hodgkin lymphoma (NHL) (3). G-CSFs are safe and well-tolerated, but serious adverse events, such as acute lung injury, acute coronary syndrome, and disseminated intravascular coagulation, have been reported (4). Among these complications, aortitis is rare, and only a few cases have been reported during chemotherapy for NHL (5-10).

We herein report an 83-year-old woman with NHL who developed aortitis after chemotherapy and subsequent prophylactic filgrastim administration. Previous reports of G-CSF-associated aortitis in NHL are also reviewed, and the implications of the current case are discussed.

Case Report

An 83-year-old Japanese woman had been diagnosed at 67 years old with diffuse large B-cell lymphoma (DLBCL) transformed from marginal zone lymphoma in the spleen, bone marrow, and mammary gland of clinical stage IVA according to the Ann Arbor staging system (11). She had received 6 cycles of the R-CHOP regimen, which consisted of rituximab 375 mg/m², cyclophosphamide 750 mg/m², doxorubicin 50 mg/m², vincristine 1.4 mg/m², and 5 days of prednisolone 60 mg/m². Complete response (CR) was confirmed with ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT).

During chemotherapy, she was diagnosed with essential hypertension and dyslipidemia. She underwent endovascular abdominal aortic repair for abdominal aortic aneurysm (AAA) at 80 years old. After 5 years of sustained CR, ¹⁸F-FDG PET/CT demonstrated the accumulation of FDG at the pharynx, neck lymph nodes, and bone marrow at 82 years old. Relapse of DLBCL was proven with a tonsil biopsy.

Salvage chemotherapy was initiated with the R-modified ESHAP regimen, consisting of rituximab 375 mg/m² on cycle day 0, carboplatin 70 mg/body from day 1 to 4, etoposide 47.5 mg/m² from day 1 to 4, methylprednisolone 500 mg from day 1 to 5, and cytarabine 1,000 mg/m² on day 5 (12). As prophylaxis, levofloxacin, fluconazole, and inhaled pentamidine were prescribed during chemotherapy. Since the patient's neutrophil count decreased below 1,000 /μL, filgrastim [Filgrastim (Genetical Recombination) (Filgrastim Biosimilar 2); Nihon Kayaku, Japan; 75 μg/body/day] was administered subcutaneously for several days from day 18, 12, and 11, in cycles 1, 2, 3, respectively.

The patient developed FN during the first and second cycles of chemotherapy. Contrast-enhanced CT (CECT) after two cycles of R-modified ESHAP showed shrinkage of the lymph nodes with no novel lesions, indicating partial response (13,14). During the third cycle, her D-dimer level was elevated (5.9 μg/mL), and CECT performed on day 22 revealed deep venous thromboembolism (DVT). Treatment with direct oral anticoagulant was initiated. During the second week of the fourth cycle, her neutrophil count was on decline at 3,100 /μL on day 9 (shown in Fig. 1). Since a further decrease was expected, filgrastim was administered daily from day 9 as secondary prophylaxis. The lowest neutrophil count was 20 /μL on day 15. On day 17, the neutrophil count recovered to 2,000 /μL, and filgrastim administration was then terminated.

Figure 1.

The clinical course of the present case. BT: body temperature, Neutro: neutrophil count, CRP: C-reactive protein, CBDCA: carboplatin, VP16: etoposide, mPSL: methylprednisolone, AraC: cytarabine

On day 21, she had a fever (body temperature 37.9 °C) and sinus tachycardia (heart rate of 119/min) with a normal respiration rate, oxygenation, and blood pressure. A physical assessment revealed no specific findings. Laboratory data showed an elevated serum C-reactive protein level (20.16 mg/dL) with the white blood cell count (5,800 /μL) and procalcitonin level (0.08 ng/mL) within the normal ranges. Tests on coagulation (prothrombin time, activated partial thromboplastin time, fibrinogen, and D-dimer) were stable before and after the fever onset. Her blood and urine culture were negative. βD-glucan and aspergillus antigen were negative in blood samples. Testing for autoimmunity in blood samples (rheumatoid factor and anti-nuclear antibody) was negative during the third cycle.

Since the physical and laboratory findings were all negative or nonspecific, and the fever did not abate, we performed CECT on day 22 in search for inflammation/infection sites or recurrent DVT. CECT revealed thoracoabdominal aortic wall thickening and surrounding fat infiltration, findings that had not been seen on CT before the fourth cycle of chemotherapy (day 22 of the third cycle) and not related to the site of endovascular abdominal aortic repair (shown in Fig. 2). Based on these findings, filgrastim-associated aortitis was diagnosed.

Figure 2.

Contrast-enhanced computed tomography (CECT) images. CECT performed on day 22 of the fourth R-modified ESHAP session (a) shows thoracoabdominal aortic wall thickening and surrounding fat infiltration (enlarged image, b). CECT on day 22 of the third cycle showed normal findings (c, enlarged image, d).

We started administration of 400 mg of celecoxib. The high fever and abnormal CRP value improved shortly thereafter, and she was discharged on day 28, accompanied by improvement of the aortic wall inflammation on CECT performed five months later.

Discussion

G-CSF is widely used for the treatment of chemotherapy-induced neutropenia during chemotherapy of various types of cancers. Previously, both short- and long-acting G-CSFs have been reported to cause aortitis (15). However, reports on whether or not G-CSFs are associated with aortitis in hematological malignancies are scarce.

Identifying the cause of a fever that develops during chemotherapy is often challenging. The differential diagnosis includes a tumor fever, a drug fever, infection, autoimmune diseases, and thromboembolism. The physical assessment, laboratory findings, and cultures were normal/negative or nonspecific for inflammation in our present case. CECT revealed aortitis, which necessitated a differential diagnosis of large-vessel vasculitis. However, overlap of patient demographics between Takayasu arteritis (TKA) and G-CSF-associated aortitis makes it difficult to differentiate one from the other solely by the patients' background (16). Our patient did not show the typical features or imaging findings of TKA (diminished or absent pulses, hypertension, and claudication of extremities; occlusion or narrowing of large vessels), and she had not developed symptoms or findings related to TKA before 40 years old. These features made TKA unlikely (17). Another large-vessel vasculitis, giant cell arteritis (GCA), was also unlikely because GCA manifests similar symptoms to TKA. Critically, aortitis in our current case was resolved by celecoxib alone, without steroids or immunosuppressive drugs, which distinguishes the case from TKA or GCA (18). Infectious aortitis was also a potential differential diagnosis. However, the blood culture was negative, CECT revealed no circumferential gas in or around the aorta (19), and the fever decreased without any antibiotics initiation, making infectious aortitis unlikely. The clinical development of the present case matched reported cases of G-CSF-associated aortitis, including unspecific inflammatory changes and a good prognosis. For these reasons, we concluded that the cause of the aortitis was filgrastim.

Previous authors have diagnosed the disease with laboratory, culture, and imaging findings similar to our present case. Given that G-CSF-associated aortitis is self-limiting and difficult to diagnose, G-CSF-associated aortitis may have been overlooked or undiagnosed, resulting in unnecessary antibiotic usage. However, aortic dissection (20) and aneurysm formation (21) are reported as serious consequences of G-CSF-associated aortitis. Therefore, G-CSF-associated aortitis must be considered in the differential diagnosis of a fever during chemotherapy, and a careful diagnosis must be made.

Reports on G-CSF-induced aortitis during treatment of NHLs are limited, but those we identified are summarized in Table. In all cases, symptoms were not specific, including a fever or in some cases concomitant neck and chest pain. Of note, the long-acting G-CSF pegfilgrastim was used instead of a short-acting G-CSF in most previous cases (11/13 cases). Indeed, Sasaki et al. conducted a single-institution retrospective study of G-CSF-associated aortitis and reported that the incidence of aortitis was significantly higher when long-acting pegfilgrastim was administered (5 cases of aortitis out of 669 administrations, 0.75%) than when short-acting filgrastim was administered (no aortitis out of 490 total administrations, 0.0%) (8). These previous findings suggest two possibilities: 1) long-acting pegfilgrastim is more likely to induce aortitis than short-acting filgrastim, or 2) G-CSF usage as prophylactic treatment (i.e., G-CSF initiation before the onset of neutropenia) is more likely to induce aortitis than as reactive treatment (i.e., G-CSF initiation after the onset of neutropenia).

Table.

Previous Reported Cases of G-CSF-associated Aortitis in NHL Patients.

Ref	Country	Age	Sex	Cancer	Cancer treatment	Regimen	G-CSF	G-CSF usage	LVV symptoms	Vessels	LVV treatment
6	Japan	59	F	B cell lymphoma	Nd	R-CHOP	Pegfilgrastim	PP	Fever and neck and chest pain	Nd	Nd
5	Germany	Nd	M	B cell lymphoma	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
5	Japan	62	F	B cell lymphoma	Nd	Nd	Pegfilgrastim	PP	Fever	Nd	None
10	Japan	60†	M	B cell lymphoma	Nd	Nd	Lenograstim	Nd	Nd	Nd	Nd
7	Japan	69	M	Transformed DLBCL	Salvage	R-modified ESHAP	Pegfilgrastim	PP	Fever	Left subclavian artery	GC
7	Japan	62	F	DLBCL	Induction	R-CHOP	Pegfilgrastim	PP	Fever and the chest pain (two episodes in two cycles)	Descending aorta	First episode: none, second episode: GC
7	Japan	72	F	FL	Salvage	R-Bendamustine	Pegfilgrastim	PP	Fever and the chest pain	Aortic arch	GC
8	Japan	Nd	F	DLBCL	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
8	Japan	Nd	F	DLBCL	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
8	Japan	Nd	F	DLBCL	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
8	Japan	Nd	M	DLBCL	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
8	Japan	Nd	M	DLBCL	Nd	Nd	Pegfilgrastim	PP	Nd	Nd	Nd
9	Japan	78	F	DLBCL	Induction	R-THP-COP	Filgrastim	PP	Fever and right cervical pain	Right common carotid artery	GC
Present case	Japan	83	F	Transformed DLBCL	Salvage	R-modified ESHAP	Filgrastim	SP	Fever	Thoracoabdominal aorta	NSAIDs

G-CSF: granulocyte colony-stimulating factor, NHL: non-Hodgkin lymphoma, Nd: no data/not reported, M: male, F: female, DLBCL: diffuse large B-cell lymphoma, FL: follicular lymphoma, R-CHOP: rituximab, cyclophosphamide, doxorubicin, oncovin and prednisolone, R-modified ESHAP: rituximab, carboplatin, etoposide, methylprednisolone, and cytarabine, R-Bendamustine: rituximab and bendamustine, R-THP-COP: rituximab, pirarubicin, cyclophosphamide, vincristine, and prednisolone, LVV: large vessel vasculitis, MZL: marginal zone lymphoma, PP: primary prophylaxis, SP: secondary prophylaxis, GC: glucocorticoid, NSAIDs: non-steroidal anti-inflammatory drugs. †: rounded by truncation

Pegfilgrastim is a pegylated form of filgrastim with polyethylene glycol attached to filgrastim (22,23). The increased size hinders renal clearance, resulting in decreased plasma clearance with an increased serum half-time (24). The safety and efficacy of a single pegfilgrastim injection and daily filgrastim injections have been shown to be comparable in lymphoma (3,25,26) and other malignant tumors (27,28). However, a recent systemic review reported a higher frequency of vasculitis in patients treated with long-acting G-CSFs than in those with other types of G-CSFs (15), possibly because sustained high levels of G-CSF stimulate emergency granulopoiesis, in which neutrophils release inflammatory cytokines (29). Despite mounting reports of G-CSF-associated aortitis, short-acting G-CSF-associated aortitis during chemotherapy of NHLs remains scarce compared with long-acting G-CSF-associated aortitis. In this respect, the rarity of cases like the present one might be attributed to the difference in aortitis frequency between short- and long-acting G-CSF.

Furthermore, the timing of usage - prophylactic or reactive - should be taken into consideration. In our present case, filgrastim was administered when the neutrophil count was still 3,100 /μL because a further decrease was expected based on the laboratory findings of previous cycles. This is in response to recent guidelines that recommend prophylactic treatment with G-CSF for high-risk FN (30,31). A previous study reported no occurrence of filgrastim-associated aortitis in any of 490 injections over 5 years, with filgrastim administered after the neutrophil count dropped below 500 /μL (8). Since reported cases of G-CSF-associated aortitis in NHL patients, in which the usage of G-CSFs is delineated, were all caused by prophylactic G-CSFs (Table) (5-9), we may infer that our present case represents another aortitis case associated with prophylactic G-CSF administration.

The suggested pathology of G-CSF-associated aortitis, in which activated neutrophils release cytokines, such as interleukin-6 and tumor necrosis factor-α (32), might corroborate this assumption, as a larger number of neutrophils activated by G-CSF would result in exacerbated inflammation. Therefore, how and to what extent celecoxib, a selective cyclooxygenase inhibitor in the arachidonic acid cascade, played an anti-inflammatory role in the present case cannot be determined. Further studies will be needed to clarify the complications related to prophylactic G-CSF use compared with its reactive use. In addition, a filgrastim biosimilar, not the originator, was administered in our present case. The G-CSFs associated with aortitis in most previous reports were originators (10,33). Biosimilar products are not identical versions of the original biological drugs, so data on the frequency of adverse reactions will need to be accumulated going forward.

In conclusion, our present case highlights the notion that G-CSF-associated aortitis should be considered in the differential diagnosis of a fever that arises during chemotherapy. It also suggests that G-CSF-associated aortitis occurs not only after pegylated long-acting G-CSF injection but also after short-acting G-CSF injection. Our case implies the possibility that prophylactic treatment with G-CSF might be more likely to cause aortitis than reactive treatment.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Written consent was obtained from the case.

Author's disclosure of potential Conflicts of Interest (COI).

Kensuke Matsuda: Honoraria, Kyowa Kirin. Yosuke Masamoto: Honoraria, Kyowa Kirin; Research funding, Kyowa Kirin. Mineo Kurokawa: Honoraria, Kyowa Kirin; Research funding, Kyowa Kirin.