Secondary haemophagocytic lymphohistiocytosis vs methotrexate toxicity – a diagnostic conundrum

Abstract

Summary

A 74-year-old female with a background of rheumatoid arthritis, managed with weekly methotrexate, was admitted with; oedema, dyspnoea, cachexia and jaundice. Bloods revealed pancytopenia, hyperferritinaemia, hyperbilirubinaemia, hypoalbuminaemia and hypofolataemia. Imaging showed a large right-sided pleural effusion, requiring therapeutic aspiration, and splenomegaly. Bone marrow aspirate revealed haemophagocytosis. Differential diagnoses included methotrexate toxicity (MTXT) and haemophagocytic lymphohistiocytosis (HLH). Management was initiated for MTXT whilst ongoing investigation for possible HLH continued.

Introduction

Methotrexate (MTX) is a dihydrofolate reductase inhibitor used in the management of auto-immune disorders and some neoplastic conditions. ¹ Its anti-inflammatory effects occur primarily via the inhibition of purine and pyrimidine synthesis, resulting in increased sensitivity of T-cells to apoptosis and increased expression of long non-coding RNA that regulates a variety of immune responses. ² MTX requires concurrent folic acid and regular blood monitoring for complications, including hepatotoxicity, gastrointestinal disturbance, myelosuppression, and pulmonary fibrosis.^3,4 Methotrexate toxicity (MTXT) is usually mild at diagnosis, but when identified, treatment is via MTX withdrawal and folic acid replacement.

Haemophagocytic lymphohistiocytosis (HLH) is a hyperinflammatory condition resulting from an over-exaggerated immune response causing multi-organ failure via overproduction of T-cells producing a cytokine storm. HLH is divided into primary and secondary causes. Primary HLH is a genetic condition mostly seen in children, whereas secondary HLH, discussed here, is via a secondary process for example, viral infections, autoimmune conditions, or certain haematological malignancies, triggering an inappropriate immune response. The incidence is poorly recorded in literature, likely reflecting its under-diagnosis in clinical practice. ⁵

Treatment is primarily via identification and management of any underlying trigger condition. Interleukin-1 receptor antagonist (IL-1Ra) is involved in the body's natural anti-inflammatory response. ⁶ Anakinra, a recombinant form of IL-1Ra is used in rheumatoid arthritis (RA) where disease modifying anti-rheumatic drug therapy has proved inadequate. ⁷ Dexamethasone and anakinra have proven useful in some cases of secondary HLH. ⁸

Case presentation

A 74-year-old Caucasian female was admitted with peripheral oedema, dyspnoea, jaundice and diarrhoea over the preceding three months. Significant medical history included childhood tuberculosis (treatment unknown) and RA. Initial observations were: blood pressure of 106/48mmHg, oxygen saturations of 99% on air, respiratory rate of 13bpm, temperature of 36.3°C and heart rate of 64bpm.

One month prior to admission, without GP consultation, she had discontinued regular medications except for weekly MTX. Discontinued medications included; folic acid, prednisolone (5 mg daily) and sulfasalazine.

This patient was complex, with evidence of fluid overload, hyperbilirubinaemia and pancytopenia, and the differential diagnoses were wide. These included MTXT, cardiac failure and infection of unclear source. Treatment included antibiotics, diuresis, and blood product replacement. MTX was discontinued and haematology review was sought. They felt the presentation was of unclear aetiology, with MTXT most likely but HLH secondary to RA also possible.

Investigations

Admission key findings (Table 1):

Pancytopenia (i.e. leukopoenia, anaemia, thrombocytopenia), requiring packed red cells and platelet transfusions.

Hyperferritinaemia, elevated vitamin B12 and folate deficiency.

Elevated C Reactive Protein highlighting a systemic inflammatory process.

Deranged liver function tests including hypoalbuminaemia, hyperbilirubinaemia and raised alkaline phosphatase.

Normal renal function besides deterioration following diuresis.

Primary hypothyroidism, treated with levothyroxine as discussed with endocrinology.

Table 1.

Blood results.

Blood Test	Day 1	Day 2	Day 6	Day 12	Day 16	Day 23	Day 27	Normal Values
White Cell Count (×109/L)	2.4	2.6	1.6	1.9	3.1	20.1	17.0	3.6–9.2
Neutrophils (×109/L)	1.2	1.7	0.4	0.4	0.6	14.4	14.0	1.7–6.2
Lymphocytes (×109/L)	1.1	0.7	1.1	1.3	0.9	2.1	1.1	1.0–3.4
Monocytes (×109/L)	0.02	0.02	0.01	0.11	1.20	2.23	1.28	0.2–0.8
Eosinophils (×109/L)	0.05	0.03	0.05	0.04	0.19	0.18	0.03	0–0.4
Haemoglobin (g/L)	102	87	84	81	99	100	89	120–155
Platelets (×109/L)	46	23	7	30	54	118	65	140–400
C Reactive Protein (mg/L)	158	-	61	-	57	-	-	0.1–5
Aspartate Transaminase (iu/L)	-	50	-	-	49	-	-	5–34
Alkaline Phosphatase (iu/L)	119	108	110	168	283	331	336	30–130
Alanine Transaminase (iu/L)	22	22	20	32	31	49	46	1–55
Bilirubin (μmol/L)	109	111	66	76	72	60	59	1–20
Albumin (g/L)	28	24	23	32	-	28	34	35–50
Triglyceride (mmol/L)	-	1.2	-	-	0.9	-	-	0.4–1.7
Lactate Dehydrogenase (iu/L)	-	400	-	-	-	-	-	125–220
Ferritin (μg/L)	-	2296	-	-	1326	-	-	10–204
Vitamin B12 (ng/L)	-	1075	-	-	-	-	-	187–883
Folate (ug/L)	-	1.6	-	-	-	-	-	3.1–20
Haptoglobin (g/L)	<0.08	-	-	-	-	-	-	0.63–2.73
Reticulocyte Count (×109/L)	-	10.0	-	-	-	-	-	30–120
Thyroid Stimulating Hormone (miu/L)	-	5.7	-	-	12.0	-	-	0.35–4.94
Free Thyroxine (pmol/L)	-	9.1	-	-	7.9	-	-	9–19
Activated Partial Thromboplastin Time (secs)	-	38.1	-	-	33.4	-	37.2	25.5–34
International Normalised Ratio	-	1.3	-	-	1.3	-	1.5	-
Fibrinogen (g/L)	-	2.50	-	-	1.3	-	-	2.0–4.5
Estimate Glomerular Filtration Rate (ml/min/1.73 m2)	86	87	78	45	66	55	87	>90
Creatinine (μmol/L)	61	59	67	106	76	89	59	50–98
Urea (μmol/L)	5.2	5.5	10.8	13.9	9.9	11.7	10.9	2.5–7.8

Imaging:

Chest X-ray showed a large right-sided pleural effusion- a transudative aspirate with no malignant cells.

Computerised Tomography of thorax abdomen and pelvis demonstrated the pleural effusion with right lower and middle lobe collapse. Ascites and peripheral oedema were also noted. There was no evidence of solid-organ malignancy.

Echocardiography revealed normal ejection fraction and mild aortic stenosis, and normal right heart function.

Ultrasound abdomen revealed trace ascites and moderate splenomegaly (14 cm), but no abnormality of the liver parenchyma.

Other:

Bone marrow aspirate (BMA) as follows:

Formal Bone Marrow Aspirate Report:

There is a marked increase in macrophage activity, with macrophages present on almost every high-power field. Frequent haemophagocytosis. Myelopoiesis is markedly reduced, but with evidence of maturation to neutrophils and no gross morphological abnormality. Erythropoiesis relatively preserved but grossly abnormal with dramatic N-C dysynchrony, basophilic stippling, nuclear irregularity including complex forms, and binucleate forms. No excess of plasma cells. Modest increase in small mature lymphocytes likely relates to hypocellular nature of sample. There is one small discrete clump of cells which probably represents macrophages, but non-haematopoietic cells are not excluded.

Discussion

Diagnosis of HLH is aided by a H-Score which provides a percentage probability of HLH by assigning points to 9 variables: 3 clinical, 5 biological, and 1 cytological (Table 2). The H-score was developed by Fadet et al. ⁹ to assess the probability of HLH in the adult cohort, unlike the HLH-2004 score which focused on paediatric patients. The probability can range from <1%, with ≤90 points, to >99% with ≥241 points. A H-Score of 169, provides a 93% sensitivity, an 86% specificity, with 90% accurate classification of non-intensive care patients. ¹⁰

Table 2.

H-Score for HLH.

Feature	Result	Points
Known Underlying Immunosuppression	No	0
	Yes	+18
Temperature (°C)	<38.4	0
	38.4–39.4	+33
	>39.4 (+49)	+49
Organomegaly	No	0
	Hepatomegaly OR splenomegaly	+23
	Hepatomegaly AND splenomegaly	+38
Number of cytopenias	1 lineage	0
	2 lineages	+24
	3 lineages	+34
Ferritin ng/mL (or μg/L)	<2000	0
	2000–6000	+35
	>6000	+50
Triglyceride (mmol/L)	<1.5	0
	1.5–4	+44
	>4	+64
Fibrinogen (g/L)	>2.5	0
	≤2.5	+30
Aspartate Transaminase (iu/L)	<30	0
	≥30	+19
Haemophagocytosis features on bone marrow aspirate	No	0
	Yes	+35

An initial H-Score gave a probability of 25–40%. Treatment with intravenous folinic acid and pulsed high-dose dexamethasone was commenced. Following BMA, a repeat H-Score was 80–88%. Anakinra was considered but given the patients clinical deterioration, this was abandoned.

Radiology reviewed the splenomegaly and felt this was similar to previous abdominal imaging and was deducted from the H-Score. The hyperferritinaemia also significantly reduced and recalculation gave a probability of only 3–5%.

Despite optimal management, the patient unfortunately passed away with the cause of death recorded as Methotrexate Toxicity.

The probability fluctuated considerably throughout as shown, until settling on 3–5% after specialist consultations. This case highlights the potential difficulties in distinguishing between MTXT and secondary HLH. RA, a classic cause of HLH, is commonly treated with MTX and so determining whether drug toxicity, or a new over-exaggerated immune response is the cause of symptoms is complex and requires early haematology input. Scoring tools aid diagnosis but having a high index of suspicion is integral.

MTXT is associated with many adverse effects which are poorly understood. This ranges from pancytopenia leading to immunosuppression along with other organ involvement such as renal impairment. This may be in part due to MTX being bound to albumin which is then excreted by the kidneys. It is also known to cause hepatotoxicity which is seen as a transaminitis on blood tests but again the mechanism behind this is not well known. Pulmonary involvement has also been documented, including lung fibrosis, interstitial pneumonitis and alveolar damage – this could be directly related to methotrexate or due to underlying RA. ³ The literature surrounding MTXT is scarce and therefore the incidence of these effects cannot be accurately assessed. As MTX is a folic acid antagonist, it is not uncommon for levels to be deplete and therefore it is essential for patients to take folic acid supplementation to avoid this.

Conclusion

MTXT is a rare but important differential in the acutely unwell patient taking methotrexate. Being mindful of whether a patient is taking the required concurrent folic acid is important, but also consider other potential causes of pancytopaenia. Scoring tools such as the H-score are helpful, but clinical judgement is needed to calculate and interpret this result, and the score can fluctuate as exemplified here. Early involvement of specialist haematology teams is essential.