Posterior reversible encephalopathy syndrome with colitis in a patient treated with panitumumab

Posterior reversible encephalopathy syndrome (PRES) associates headaches, encephalopathy, seizures and visual disturbances with typical magnetic resonance imaging (MRI) findings. PRES is reversible and benign in most of cases, although sequels may exist, and fatal outcomes have been described in up to 19% [Alhilali et al. 2014; Fugate and Rabinstein, 2015]. Some PRES were associated to targeted therapies [Fugate and Rabinstein, 2015], including cetuximab [Kamiya-Matsuoka et al. 2016; Palma et al. 2011], a monoclonal antibody targeting the epidermal growth factor receptor (EGFR). Panitumumab is a fully human recombinant monoclonal antibody with a high binding affinity to human EGFR. Here we report and discuss the occurrence of a PRES in a patient presenting with colitis and treated with panitumumab. Patient consent was obtained and patient confidentiality was maintained.

A 44-year-old man without a past medical history was diagnosed with high-risk non-muscle invasive bladder cancer in 2010, and treated by transurethral resection of the bladder and Bacillus Calmette–Guerin therapy. Documented metastatic relapse involving lymph node and bone occurred in 2013. The patient was enrolled in a phase II clinical trial, testing the association of dose dense methotrexate, vinblastine, adriamycin, cisplatin (DD–MVAC) with panitumumab. Five courses of DD–MVAC were delivered, from October to December 2013. Cisplatin-based chemotherapy was discontinued due to a grade 3 acute renal failure, while panitumumab was continued alone at 6 mg/kg, bimonthly. The patient partially recovered from his renal failure and developed a chronic renal insufficiency (creatinine clearance [Cockroft] = 40 ml/min). On 27 October 2014, that is 12 days after the 18th cycle of panitumumab, and 10 months after cisplatin discontinuation, the patient was hospitalized for a temporary loss of consciousness. He complained of asthenia, headaches and diarrhea with abdominal pain. A physical examination noted a Glasgow score = 14, with high blood pressure (160/90 mmHg) and a soft abdomen. A few hours later, the patient had two tonic-clonic seizures and persistent high blood pressure (182/103 mmHg). A blood test showed a constant renal failure (creatinine clearance [Cockroft] = 38.5 ml/min) and an inflammatory syndrome (CRP = 194 mg/l). An MRI revealed symmetrical cortical and subcortical hyperintensity, involving the parietal lobes in T2 fluid attenuated inversion recovery (T2-FLAIR) sequence without restriction on diffusion weight imaging (DWI), arguing for vasogenic edema (Figure 1a, b). There were no arguments for an ischemic cause, an abscess or hemorrhage. Such an association of clinical symptoms with radiological findings was consistent with PRES. Electroencephalogram and cerebrospinal fluid, were normal, including infectious tests. Symptomatic treatment consisted in blood pressure control (continuous intravenous infusion of urapidil and nicardipine) and in prevention of seizures using levetiracetam. The patient received probabilistic antibiotherapy with ceftriaxone and metronidazole, justified by symptoms of colitis, confirmed by sigmoidoscopy, which showed ulcerative colitis without pathologic arguments for infectious cause. Panitumumab was discontinued. Pharmacovigilance and serious adverse event reports related to clinical trial were completed. The patient completely recovered without sequel and an MRI performed 3 days later showed disappearance of lesions (Figure 1c, d). The patient was then started on paclitaxel, as a second line regimen, after an assessment of bone progression. The patient died from the disease 6 months later without new occurrence of PRES. Panitumumab was never reintroduced.

Figure 1.

Typical MR early phase and delayed phase imaging of PRES syndrome: early phase MR imaging suggesting vasogenic edema. T2-FLAIR sequence (a) shows symmetrical cortical and subcortical hyperintensity involving the parietal lobes; DWI shows absence of restriction (b). An MR control performed 3 days later shows complete recovery (c and d).

As proposed [Fugate and Rabinstein, 2015], a positive diagnosis of PRES was considered, due to the occurrence of neurologic symptoms (tonic-clonic seizures, headaches and confusion) with risk factors such as hypertension and renal failure associated with typical radiologic findings (Figure 1). Panitumumab was the patient’s only current treatment with a total of 15 months exposure, including >9 months used alone (post cisplatin discontinuation), and achieved a stable disease with only grade 2 skin toxicities. Concomitantly to PRES, the patient was diagnosed with colitis, responsible of the inflammatory syndrome. Even though biopsies did not confirm an infectious colitis, the patient received a short course of probabilistic antibiotherapy. Associations of PRES with severe sepsis have been previously reported [Bartynski et al. 2006], however, during this episode the patient did not have either clear symptoms of sepsis, or evidence of infection. These findings contrast with the occurrence 2 months later in the same patient of a severe sepsis due to Staphylococcus aureus without neurologic symptoms despite positive blood cultures and a CRP = 288 mg/l. Furthermore, in the litterature some associations of PRES with colitis were more imputed to treatment such as antibiotics or immunosuppressors rather than colitis itself [Gumus et al. 2010; Haddock et al. 2011; Kikuchi et al. 2016]. Here, symptoms of PRES preceded antibiotherapy and no additional immunosuppressive treatment was given. PRES was associated with hypertension, but it remains unclear whether this symptom may have caused PRES, and high blood pressure is not a common effect related to panitumumab [Gemmete and Mukherji, 2011]. Finally, cisplatin imputability could also be discussed as a risk factor responsible of renal insufficiency and hypomagnesaemia [Atsmon and Dolev, 2005; Xie and Jones, 2016]. While anti-angiogenic molecules may induce PRES by both the inhibition of nitric oxide synthase pathway necessary for endothelial homeostasis, and by high blood pressure [Fugate and Rabinstein, 2015; Levy et al. 2014; Tlemsani et al. 2011], pathophysiology for an anti-EGFR agent, especially panitumumab, remains unclear. However, three cases of PRES related to cetuximab were described [Kamiya-Matsuoka et al. 2016; Palma et al. 2011]. One patient was treated for metastatic penile cancer and the symptoms occurred few hours after the first infusion of the anti-EGFR. Interestingly, this patient was previously exposed to cisplatin and paclitaxel, both known to be involved in PRES. In our case, a first proposed hypothesis could be the effect of hypomagnesaemia, potentially initiated by cisplatin, and worsened by panitumumab [Abbas et al. 2015] and colitis. Of note, the last measurement of serum magnesium levels in this patient treated with panitumumab showed low magnesium levels – 1.07 mg/dl (1.7–2.4 mg/dl reference range). This hypomagnesaemia may have been associated with the long clinical benefit achieved by this treatment, as has recently been suggested in cases of colorectal cancer [Fujii et al. 2016]. Furthermore, an immune response induced by both panitumumab and colitis, may have promoted pro-inflammatory cytokine production and altered vascular permeability [Fugate and Rabinstein, 2015; Marra et al. 2014]. Finally, the induction of toxicity on blood brain barrier, as described for cytotoxic drugs, may have been responsible for hypoperfusion and hypoxia [Vaughn et al. 2008].

To our knowledge, this is the first case of PRES in a patient treated by panitumumab. The existence of PRES might be explained by a combined effect of hypomagnesaemia and immune response induced by both panitumumab and colitis. Using the Naranjo Scale [Naranjo et al. 1981], PRES was classified as a ‘possible’ (=3) adverse drug reaction to panitumumab, in a context of ulcerative colitis. Although scarcer than cases described using anti-angiogenic agents, this report constitutes the fourth case with an anti EGFR antibody, including at least two patients previously exposed to cisplatin treatment. It broadens the number of drugs and variety of clinical settings potentially involved in PRES, for which early diagnosis is key.