Fatal Stevens–Johnson syndrome/toxic epidermal necrolysis induced by allopurinol–rituximab–bendamustine therapy

Introduction

Stevens–Johnson syndrome (SJS), Stevens-Johnson syndrome/toxic epidermal necrolysis overlap (SJS/TEN), and toxic epidermal necrolysis (TEN) are a spectrum of acute hypersensitivity reactions based on body surface area skin detachment of <10%, 10–30%, and >30%, respectively. Compromised mucocutaneous tissues with extensive tissue sloughing and separation of the first epidermal layer are defining characteristics. Numerous medications with diverse indications have been reported as the potential causes of these life-threatening and potentially fatal skin reactions, yet no commonalities have been identified that predict the risk of reaction.^1,2 Allopurinol, bendamustine, and rituximab have all been demonstrated to cause variable skin toxicities, some of which have progressed to severe, life-threatening, and fatal toxicity.^3–13 Here, we present a case report where the concomitant use of allopurinol, rituximab, and bendamustine induced fatal SJS/TEN in a patient with non-Hodgkin’s B-cell lymphoma.

Case report

A 78-year-old Caucasian male with no pertinent past medical history was newly diagnosed with diffuse large B-cell lymphoma with bone marrow involvement. The patient received allopurinol at a dose of 300 mg by mouth daily to be taken until directed to discontinue as tumor lysis prophylaxis starting prior to his first cycle of rituximab and bendamustine (Table 1). Appropriate monitoring parameters including white blood cell count, absolute neutrophil count, platelet count, and kidney function were assessed, and all were within normal limits with respect to treatment parameters prior to initiating chemotherapy.

OPEN IN VIEWER

Table 1.

Rituximab + bendamustine chemotherapy regimen.

Drug	Dose (mg/m2)	Patient’s dose (mg)	Day	Cycle length
Rituximab	375	700	1	28 days
Bendamustine	90	175	2–3	28 days

Fourteen days after starting chemotherapy, the patient presented to his local hospital with a 3-day history of sore throat, extreme fatigue and weakness, severe pain throughout his skin that was aggravated by even slight touch, and a widespread non-pruritic macropapular rash. The patient had extensive skin sloughing of multiple body surfaces and mucous membranes as well as fragile skin throughout the remainder of his body. He received supportive care and, upon stabilization, was immediately transferred to our facility for treatment the day of his admission.

The patient was admitted to the burn surgery service for further assessment where tissue involvement was noted to include the patient’s face, anterior and posterior trunk regions, upper and lower extremities, oral mucosa, ocular conjunctiva, and genitalia, all of which were Nikolsky’s sign positive. The Lund–Browder Burn calculation was used to determine the percentage of total body surface area affected and estimated 14% involvement, consistent with a diagnosis of SJS/TEN.^2,14 The patient was also tachycardic, hypotensive, and febrile. Allopurinol was discontinued following confirmation that the patient had taken the medication as directed for 13 days with his last reported dose being the day before hospital admission. Supportive care measures including hydration, wound debridement, systemic and topical antibiotics, ophthalmic irrigation and lubrication, nutritional support, and pain management were provided. With further progression of symptoms, the patient requested palliative measures and died secondary to complications of SJS/TEN.

The patient had recently started palonosetron, dexamethasone, prochlorperazine, allopurinol, rituximab, and bendamustine as part of his treatment regimen and continued to take maintenance medications including vitamin D, aspirin, levothyroxine, and lorazepam. He had no known medication allergies and had not been previously treated with allopurinol, bendamustine, or rituximab.

To assess the relationship of medication use to adverse effect, the patient’s Naranjo probability algorithm score was calculated and supported a probable drug-induced reaction (Naranjo score = 5, Table 2). ¹⁵ This score was obtained because the adverse event appeared after the suspected drug was administered and there were no alternative causes that could explain the reaction, and because the reaction has been conclusively reported previously.

OPEN IN VIEWER

Table 2.

Naranjo ADR probability scale. ¹⁶

	Yes	No	Unsure, not done	Patient’s score
Are there previous conclusive reports on this reaction?	+1	0	0	+1
Did the adverse event appear after the suspected drug was administered?	+2	−1	0	+2
Did the adverse reaction improve when the drug was discontinued or a specific antagonist was administered?	+1	0	0	0
Did the adverse reaction reappear when the drug was readministered?	+2	−1	0	0
Are there alternative causes (other than the drug) that could, on their own, have caused the reaction?	−1	+2	0	+2
Did the reaction reappear when placebo was given?	−1	+1	0	0
Was the drug detected in the blood (or other fluids) in concentrations known to be toxic?	+1	0	0	0
Was the reaction more severe when the dose was increased or less severe when the dose was decreased?	+1	0	0	0
Did the patient have a similar reaction to the same or similar drugs in any previous exposure?	+1	0	0	0
Was the adverse event confirmed by any objective evidence?	+ 1	0	0	0
Total score				+5 (probable)

The SCORTEN (SCORe of TEN) severity of illness scale was also determined and predicted a 58.3% mortality rate in this patient (SCORTEN score = 4, Table 3). ¹⁶ The patient earned a point each for being older than 40 years old, having an associated malignancy, having serum blood urea nitrogen greater than 27 mg/dL, and having greater than 10% detached or compromised body surface. The patient’s heart rate was recorded at 116 on admission and likely would have been greater than 120 had he not received fluid resuscitation prior to arrival at our facility, which would given him a score of 5 with a predicted >90% mortality rate. The SCORTEN scale has been validated for not only TEN but rather for the full continuum of SJS, SJS/TEN, and TEN.

OPEN IN VIEWER

Table 3.

SCORTEN severity-of-illness index. ¹⁶

Risk factor	0	1	Patient’s score
Age	<40 years	>40 years	1
Associated malignancy	No	Yes	1
Heart rate (beats/min)	<120	>120	0
Serum BUN (mg/dL)	<27	>27	1
Detached or compromised body surface	<10%	>10%	1
Serum bicarbonate (mEq/L)	>20	<20	0
Serum glucose (mg/dL)	<250	>250	0
Total score			4 (58.3% predicted mortality rate)

To further assess risk FDA Medwatch adverse event, case reports over a 5-year period from 2008 through 2013 on allopurinol, bendamustine, and rituximab were examined (Table 4). SJS/TEN overlap case reports were initially assessed but only three cases were reported, possibly due to nomenclature confusion, while 167 TEN case reports were filed. SJS/TEN treatment and mortality have been shown to be much more similar to TEN than SJS, which is why we focused our analysis on TEN case reports. ¹⁶ It is interesting that approximately 30% of the TEN case reports did list concurrent SJS adverse reactions some of which may have been SJS/TEN overlap cases.

OPEN IN VIEWER

Table 4.

FDA Adverse Event Reporting System TEN case reports from October 30, 2008 to October 30, 2013—unless noted all data in parenthesis is % of TEN cases.

Drug (n = total ADR case reports)	Allopurinol (n = 3007)	Rituximab (n = 20,622)	Bendamustine (n = 4110)
TEN case reports (% of total ADR case reports)	134 (4.5)	19 (0.09)	14 (0.34)
Concomitant allopurinol	–	4 (19.0)	6 (42.9)
Concomitant rituximab	4 (3.0)	–	8 (57.1)
Concomitant bendamustine	6 (4.4)	8 (42.1)	–
Concomitant allopurinol and rituximab	–	–	4 (28.6)
Concomitant chemotherapy other than bendamustine	10 (7.5)	5 (26.3)	–
Resulting in death	49 (36.6)	5 (23.8)	4 (28.6)
Concomitant allopurinol	–	0/4	2/6
Concomitant rituximab	0/4	–	0/8
Concomitant bendamustine	2/6	0/8	–
Concomitant allopurinol and rituximab	–	–	0/4
Concomitant chemotherapy other than bendamustine	3/10	3/5	–
Resulting in hospitalization	91 (67.9)	14 (66.7)	6 (42.9)
Concurrent SJS	45 (33.8)	1 (4.8)	3 (21.4)
Median age (range)—yeara	64 (5–94)	57 (16–77)	61 (45–70)
Male sex—% of patientsa	40.5	68.8	50.0
Geographic regiona
Americas	38	8	9
Asia	28	1	1
Europe	57	11	4
Africa, Australia, or unknown	11	1	0

Allopurinol had the fewest total adverse drug reaction reports but the highest number of TEN reaction reports at 134, which was seven times higher than that reported for rituximab and approximately 10-fold higher than for bendamustine. Percentage of allopurinol TEN reactions resulting in patient death was also the highest at 36.6%, followed by bendamustine at 28.6% and rituximab at 23.8%.

Out of four TEN case reports of patients that received concomitant allopurinol, rituximab, and bendamustine, none died; however, two out of six patients that received concomitant allopurinol and rituximab died. TEN reactions due to allopurinol and rituximab resulted in a similar percentage of hospitalized patients with reactions due to bendamustine resulting in fewer hospitalizations. While eight of 19 rituximab TEN reactions occurred with concomitant bendamustine, another seven rituximab reactions occurred without concomitant allopurinol, bendamustine, or non-bendamustine chemotherapy (data not included in table). Three of 14 bendamustine reactions occurred without concomitant allopurinol or rituximab.

Discussion

Allopurinol is a xanthine oxidase inhibitor, which blocks uric acid formation, and is utilized as prophylaxis in patients with an elevated risk of hyperuricemia and tumor lysis syndrome. It is generally well tolerated; however, it is known to cause hypersensitivity reactions including rash that may progress to more severe skin reactions including SJS, SJS/TEN, and TEN.^5–8 It is of interest that 4.5% of all reported FDA Medwatch allopurinol adverse effect reports were TEN reactions, a relatively large percentage. Only 12% of these cases occurred in the setting of chemotherapy, six of which were with concomitant bendamustine. It is possible that our patient may not have encountered this serious adverse effect or may have developed a more mild to moderate case had he taken a shorter course of allopurinol than 13 days. A retrospective study did show that immediate withdrawal of potential offending drug at the onset of SJS/TEN symptoms decreased the risk of death. ¹⁷ FDA Medwatch data do not routinely ascribe duration of therapy but this would have been an interesting contribution to our analysis.

Rituximab is a monoclonal antibody used as a chemotherapeutic agent to target CD20 + blood cells. Rituximab can cause severe mucocutaneous reactions, some with fatal outcomes. Case reports have linked rituximab with SJS, SJS/TEN, and TEN as well as paraneoplastic pemphigus, lichenoid dermatitis, and vesiculobullous dermatitis.^9–11 Our patient would be the first patient reported to FDA Medwatch that died secondary to a SJS/TEN reaction following allopurinol, rituximab, and bendamustine exposure. Of course, it is possible that rituximab use was underreported in patients receiving this treatment regimen as rituximab as a causative agent of SJS/TEN is rare, and some would argue controversial. ¹¹ It is noteworthy that no deaths occurred among the FDA Medwatch TEN cases of concomitant allopurinol and rituximab, bendamustine and rituximab, or a combination of the three drugs. Three of five patients that received rituximab with concurrent chemotherapy other than bendamustine died and were the majority of rituximab-related deaths. The first of these patients received cyclophosphamide, the second received fludarabine, cyclophosphamide, doxorubicin, and vincristine, and the third received cyclophosphamide, methotrexate, vincristine, and doxorubicin. Case studies of TEN have been reported for cyclophosphamide, fludarabine, methotrexate, and doxorubicin. A black box warning for TEN is included in the methotrexate package insert.

Bendamustine is a nitrogen mustard compound that functions as a chemotherapeutic agent with both DNA alkylation and anti-metabolite properties. Phase II clinical trial and post-marketing reports demonstrated rare cases of severe, life-threatening skin reactions including SJS and TEN.^12,13 To date, there are no published reports demonstrating that bendamustine alone causes SJS, SJS/TEN, or TEN; however, a single case reported that single agent bendamustine caused a severe cutaneous interface drug eruption. ⁴ In the bendamustine package insert, it is noted that SJS and TEN have been reported with concomitant allopurinol and other medications known to cause these severe skin reactions. ¹²

As allopurinol, bendamustine, and rituximab have all been known to cause severe skin reactions, questions remain regarding which agents are causative when, and whether or not the possibility of an additive risk of SJS/TEN when multiple drugs known to be associated with SJS/TEN are administered concomitantly. Given the severity of the reaction and the high-mortality risk, re-challenging one or more drugs in patients to fully elucidate the underlying cause of SJS/TEN is not acceptable. FDA Medwatch data have several limitations. There is no certainty that the reported adverse events are actually due to the drug identified, and submission of reports is voluntary and dependent on a broad variety of factors. The information from these reports cannot be utilized to calculate adverse event incidence. While difficult to draw strong associations with these limitations and the limited number of TEN case reports, the available Medwatch data on concomitant allopurinol, rituximab, and bendamustine use available at this time do not show additive risk.

Conclusion

While the incidence of SJS/TEN is rare, the overall mortality rate of patients with this adverse effect is approximately 30%. It is important to identify symptoms and disease progression early, remove potential offending agents immediately, and provide optimal supportive care to promote recovery and reduce mortality. Quantifying additive risk of this rare adverse effect from use of concomitant medications associated with this reaction such as allopurinol, rituximab, and bendamustine remains a challenge.

It is important to recognize that tumor lysis syndrome is a potential serious adverse effect that must be closely monitored. If allopurinol is deemed necessary as tumor lysis prophylaxis in treatment regimens containing rituximab and bendamustine, it is probably rational to utilize as short of course as is clinically required with close monitoring. Consideration of avoidance of allopurinol in patients with low risk for tumor lysis syndrome is likely clinically appropriate.