Lamotrigine and anti‐seizure medications in HLH: A systematic review

Abstract

Lamotrigine and other anti‐seizure medications have been reported to cause both haemophagocytic lympho‐histiocytosis (HLH) and drug‐induced hypersensitivity syndrome (DIHS). HLH and DIHS have many characteristic clinical, laboratory, pathological findings and underlying mechanisms; we identified significant clinical overlap between these two syndromes in our systematic review. Of note, the presence of haemophagocytosis on bone marrow biopsy, while supportive, is not exclusive to or synonymous with a diagnosis of HLH.

To the Editor,

Haemophagocytic lympho‐histiocytosis (HLH) is a life‐threatening hyperinflammatory syndrome characterized by excessive activation of cytotoxic T cells, natural killer (NK) cells and macrophages producing hypercytokinaemia, multiorgan failure and high risk of death. ¹ Therapy‐associated HLH is typically attributed to immunomodulatory treatments, including cytotoxic chemotherapy for cancer, stem cell transplantation, chimeric antigen receptor T‐cell therapy and immune checkpoint inhibitors. ² , ³ However, other medications that are not used primarily as immunomodulators, such as antibiotics and anti‐seizure medications, have also been associated with HLH. ⁴ Particularly prominent among these is the anti‐seizure medication (ASM) lamotrigine; in 2018, the US Food and Drug Administration (FDA) released a drug safety announcement regarding the association of lamotrigine and HLH based on its findings of eight clinically suspected or confirmed cases of lamotrigine‐associated HLH. ⁵

Although lamotrigine and other aromatic ASMs are not used primarily as immunomodulatory agents, they also cause severe cutaneous adverse drug reactions including Stevens–Johnson syndrome/toxic epidermal necrolysis (SJS/TEN) and drug‐induced hypersensitivity syndrome/drug reaction with eosinophilia syndrome (DIHS/DRESS, here onwards referred as DIHS ⁶ ). Given the association of lamotrigine and other ASMs with both HLH and DIHS, we sought to review and synthesize the existing literature on ASM‐associated HLH and evaluate a potential relationship with DIHS.

Twenty‐nine anti‐seizure medications introduced between 1910 and 2020 were identified from a recent literature review of ASMs. ⁷ We interrogated the FDA Adverse Event Reporting System's (FAERS) database and separately searched EMBASE (1974‐) and MEDLINE/PubMed (1966‐) for records indexed as of 31 September 2024. ⁸ The first author completed the deduplication, exclusion of non‐English language studies and title and abstract screen, and then, the two separate searches were combined for final deduplication and full‐text screening. This was done to exclude duplicate cases or multiple publications of the same case/cases. All reports in which the authors clinically suspected HLH in the context of a recently initiated or increased dose of ASM were included. Studies with irrelevant cases, insufficient clinical data or attributed to a non‐medication trigger of HLH and editorials/commentaries without an associated case report were excluded.

The HLH‐2004 diagnostic criteria were used to establish the diagnosis of HLH: five of the eight criteria must be met ⁹ (Table S1). To account for cases where HLH was clinically suspected but a complete work‐up may not have been reported, we defined ‘suspected HLH’ as meeting four of the eight criteria. We also included the H‐score for each case by extracting the relevant outcome data and calculating it (for most cases, where it was not explicitly reported by authors ¹⁰ ). The H‐score was also used to define confirmed (H > 169) or suspected HLH (H > 138 in adult and H > 120 in paediatric patients); the H > 169 threshold was obtained from the original study and the lower thresholds were identified from a later validation study. ¹⁰ , ¹¹ The RegiSCAR score was used to define DIHS overlap with a score of 1 or less defined as ‘no DIHS’, 2–3 as ‘possible DIHS’, 4–5 as ‘probable DIHS’ and 6 or higher as ‘definite DIHS’. ¹² We extracted the clinical features involved in the RegiSCAR score and calculated and interpreted it except in cases where it was explicitly reported by authors.

We identified 31 cases of ASM‐associated HLH, of which 20 cases were attributable to lamotrigine as the sole ASM or new medication added to a stable regimen ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ (Figure S1: PRISMA flowchart; Table 1). Other aromatic ASMs including carbamazepine (3), phenobarbital (2) and oxcarbazepine (1) accounted for six cases, while valproic acid (3) and multiple medications (2) were implicated in the other five cases (Table 1). The median age of patients was 27 years, range: 3–66 years; there were 23 adult and 8 paediatric cases. Eighteen cases were in females and 13 in males. The most common concurrent conditions in our cases were DIHS (10) lamotrigine‐induced lupus (3) and EBV/HHV‐6 viraemia (whether deemed as active infection or asymptomatic, 3; Table 1).

TABLE 1.

Clinical characteristics of anti‐seizure medication‐associated HLH (N = 31).

Medications (N)	Median age, years [range]	Sex (female: male)	Concurrent conditions	Treatment regimen	Clinical outcome	Median H‐score [IQR] a
Lamotrigine (20)	31 [4–47]	11F:9M	DIHS (4) Drug‐induced lupus (3) EBV viraemia (2) Others (3)	Etoposide & dexamethasone (10) Steroids only (5) Steroids & IVIg (5) Adjuncts d (5) Supportive (1)	Resolution (17) Death (2) Relapsing course (1)	198.5 [175–234] b
Carbamazepine (3)	11 [4–66]	3F:0M	DIHS (2)	Steroids only (2) Steroids & IVIg (1)	Resolution (3)	158 [132–211]
Valproic Acid (3)	5 [3–19]	1F:2M	Atypical DIHS (1)	Steroids only (2) Steroids & IVIg (1)	Resolution (3)	223 [132–230]
Phenobarbital (2)	24.5 [24,25]	2F:0M	DIHS (1) HHV‐6 infection (1)	Steroids only (1) Etoposide & steroids (1)	Resolution (1) Relapsing course (1)	206.5 [160–253]
Oxcarbazepine (1)	3	1M	None	IVIg only (1)	Resolution (1)	190
Multiple Medications (2) c	35 [21–49]	1F:1M	DIHS (2) Other (2)	Etoposide & dexamethasone (1) Steroids (2) Adjunct (1)	Resolution (1) Death (1)	227 b
Total (31)	27 [3–66]	18F:13M	DIHS (10) Drug‐induced lupus (3) EBV/HHV‐6 viraemia (3) Others (9)	Etoposide & steroids (12) Steroids & IVIg (7) Steroids only (12) IVIg only (1) Adjuncts d (6)	Resolution (26) Relapsing course (2) Death (3)	198.5 b [159–230]

^a IQR reported for cases of n = 2 or more.

^b Excludes two lamotrigine‐associated cases and one multiple medication‐associated case where a precise H score was not calculable.

^c Multiple medication cases included: clobazam, levetiracetam, midazolam, phenytoin and valproic acid ²³ and lamotrigine, levetiracetam, phenobarbital and phenytoin. ³⁶

^d Other adjunctive treatments included: intrathecal methotrexate (2), anakinra (1), emapalumab (1), hydroxychloroquine (1), mycophenolate (1).

The majority of medication‐associated HLH presented with initial symptoms (such as fever, skin rash, cytopenias) within 3 weeks of drug initiation (19 of the 20 lamotrigine‐associated cases and 6 of the 11 of multidrug cases) with a median time of 2 weeks (Tables S2 and S3). In all cases, the drug was discontinued upon recognition of DIHS or HLH. The most common treatment regimens employed included the HLH‐1994 chemotherapy protocol in the form of etoposide and dexamethasone or other steroids (12), high‐dose corticosteroids only (12) or in combination with intravenous immunoglobulin (7). Other adjunctive treatments included intrathecal methotrexate (2), anakinra (1), emapalumab (1), hydroxychloroquine (1) and mycophenolate (1). In one case, drug discontinuation and otherwise supportive management only were employed. ²⁸ Outcomes were favourable: 26 of the 31 cases achieved complete clinical resolution, two had a chronic relapsing recurring course ¹⁶ , ¹⁷ and three patients died ¹³ , ³⁰ , ³⁶ (Tables S2 and S3).

Seven of the 31 ASM‐associated HLH cases also had probable or definite DIHS (Table 2), with one further case being formally diagnosed by study authors as DIHS, though per reported data, meeting the possible range with a score of 3. ⁴⁰ There were an additional 12 of the 31 cases, with a score of 2–3, meeting a diagnosis of possible DIHS, with the most common overlapping features being fever, transaminitis and a diffuse erythematous skin rash. Lamotrigine and other aromatic ASMs accounted for 18 of the 20 cases with possible to definite DIHS overlap.

TABLE 2.

HLH and DIHS overlap by anti‐seizure medications.

Medications (N)	REGIScar score for DIHS (points)
	≤1 (no case)	2–3 (possible)	4–5 (probable)	≥6 (definite)
Lamotrigine (20)	8	11 a	0	1
Carbamazepine (3)	1	0	2	0
Valproic acid (3)	1	1	1	0
Phenobarbital (2)	0	1	0	1
Oxcarbazepine (1)	1	0	0	0
Multiple medications (2)	0	0	0	2
Total (31)	11	13 a	3	4

^a One case with high clinical suspicion and diagnoses established with confirmatory biopsy. ⁴⁰

Serum ferritin was reported in 20 of the 31 cases and was quite elevated (median 18 821 μg/L) as expected in HLH. C‐reactive protein was reported in only 13 cases and was moderately elevated (median 99 mg/L) whereas soluble CD25 and eosinophil count were rarely reported (Table 3). These data are consistent with recent reports differentiating HLH from other cytokine storm syndromes; greater attention to these simple biomarkers in future reports of ASM‐associated HLH may be helpful in elucidating the relationship between HLH and DIHS. ⁴⁴

TABLE 3.

Inflammatory markers for HLH and DIHS.

Laboratory test a [N]	Median [IQR]
Ferritin b [20]	18 821 μg/L [3292–42 236]
sCD25 c [6]	4228 U/mL [3463–6352]
CRP [13]	99 mg/L [37.9–139.5]
Eosinophils [7]	1.328 × 109/L [0.12–1.5]

^a Maximum values available from studies included are reported.

^b Four studies reported ferritin >40 000, one—ferritin >21 000, one—ferritin >13 000; due to the lack of precise values, they are not included in the table.

^c Four additional studies reported sCD25 in pg/mL, for which there is no standard conversion formula to U/mL, and therefore, these studies have not been included in the table.

In our review, lamotrigine and other aromatic ASMs (carbamazepine, oxcarbazepine, phenobarbital, phenytoin) or a combination of aromatic ASMs accounted for 28 of 31 confirmed or suspected cases of HLH. The other three cases were attributable to valproic acid. Aromatic ASMs are also among the most common medications implicated in severe cutaneous adverse drug reactions including SJS/TEN and DIHS. ⁶

There is substantial clinical overlap between HLH and DIHS, with common clinical features of fever, diffuse skin rash, lymphadenopathy and multiorgan inflammation and laboratory markers including hyperferritinaemia (which may indicate the severity of adverse cutaneous drug reactions and is common in HLH ¹ , ⁴⁵ , ⁴⁶ ). As highlighted in Table 3, our cases of ASM‐associated HLH were accompanied by elevated inflammatory markers in the form of marked hyperferritinaemia, elevated CRP and elevated soluble CD25 levels.

We identified eight cases of HLH which developed in the setting of an established diagnosis of DIHS. Additionally, we identified 12 cases of ASM‐mediated HLH which carried substantial clinical overlap with DIHS, with two to three positive clinical features of DIHS, which were most commonly fever, diffuse skin rash and liver involvement. Peripheral eosinophilia was also reported in five of eight cases of definite or diagnosed DIHS and it was not reported in cases of possible DIHS (Tables S2 and S3). Given this overlap, further research is needed to identify whether there may be common mechanisms of immune dysregulation involved in these syndromes.

The most common treatment regimens used in ASM‐associated HLH are conventional or modified HLH‐1994 protocol‐based chemotherapy with etoposide with corticosteroids (12), corticosteroids alone (12) or in combination with intravenous immunoglobulin (7). Two cases were treated with the cytokine inhibitors anakinra (IL‐1) and emapalumab (interferon‐γ). Overall, we found that, with drug discontinuation and the use of immunosuppressive therapy, ASM‐associated HLH was associated with a good prognosis. Complete clinical resolution was reported in 26 of the 31 patients (84%), and only three deaths were reported (Table 1).

In our study, we were primarily limited by the extent to which cases reported findings which satisfied the REGIScar Score for DIHS, especially in cases where DIHS was not suspected as a coinciding diagnosis with HLH. For instance, we found the extent of lymphadenopathy, presence or degree of eosinophilia, the characteristics of the skin rash and the performance of a skin biopsy may not have been reported. Additionally, there was limited reporting of pertinent inflammatory markers associated with HLH, particularly with only a small minority of studies reporting soluble CD‐25 levels. This may understate the degree of HLH‐DIHS overlap in our reported cases.

This study confirms that, among ASMs, lamotrigine was the most commonly associated ASM with HLH. Furthermore, it identifies significant clinical overlap between ASM‐associated HLH and DIHS; further research is needed to understand the underlying mechanisms of immune dysregulation. As lamotrigine is increasingly prescribed in adults with epilepsy, healthcare practitioners must become cognisant of its propensity for these adverse immune reactions. ⁴⁷ Early recognition, prompt drug discontinuation and initiation of immunosuppressive therapy are associated with a favourable clinical prognosis.

AUTHOR CONTRIBUTIONS

Yaswanta K. Gummadi and Luke Y. C. Chen: project conceptualization, analysis, writing and editing. Yaswanta K. Gummadi: systematic review and data collection. Ashley E. Sutherland, Kristin M. Ikeda and Lori Connors: content expertise and revision. Luke Y. C. Chen: supervision of the project.

FUNDING INFORMATION

Luke Chen's research is supported by a philanthropic gift from the Hsu & Taylor Family through the VGH & UBC Hospital Foundation.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to disclose.

Supporting information

Figure S1.

Tables S1–S3.