VALPROATE INDUCED ISOLATED THROMBOCYTOPENIA-AFTER ACUTE FEBRILE EPISODES OF UNKNOWN ORIGIN

Introduction: Valproate is the most used antiepileptic drug in children of Serbia. Thrombocytopenia is one of the rare but known adverse effects of valproate. Fever of unknown origin is a frequent clinical sign in children.

Case report: We present three cases of resistant thrombocytopenia in children on antiepileptic therapy (valporate). Thrombocytopenia appeared after a short febrile eppisodes of unknown origin, presenting with only few petechial skin spots. It lasts longer than 4 weeks, not reacting to intravenos imunoglobulins, nither to kortikotheroids. Thrombocythopenia responded dramatically to withdrawing of valproate. All of the children had valproate drug level in the lower third of therapeutic range.

Conclusion: Based on the known adverse drug reaction between valproate and the thrombocytopenia, complete platelet normalization after VPA withdrawal proved this assosiation.

Introduction

Valproate (VPA) belongs to a group of anticonvulsant drugs, which is effective in controlling a wide variety of seizures. It is a commonly used anticonvulsant. In Serbia VPA is available in slow-release tablets and syrup. VPA can cause haematological and liver toxicity, which is usually dose or blood level dependent (1). Valproate associated thrombocytopenia has been reported ranging from 1 to 21% but, severe clinically significant thrombocytopenia is rare (2). Significant negative correlation between VPA plasma levels and platelet counts is widely reported (3). During viral infection or acute febrile illness, such reaction may occur even with normal blood levels. This should be anticipated in children with acute febrile state of different origin (4).

We report three children on VPA monotherapy presenting with resistant thrombocytopenia after acute febrile episodes of unknown origin.

Case report

We report three children on VPA monotherapy developing resistant thrombocytopenia after acute febrile episodes of unknown origin, with prompt platellet normalisation after withdrawing VPA (Table 1).

Case 1 – A sixteen-month-old male child, treated with VPA (after a case of seizure disorder) for 127 days (in a dose of 22 mg/kg/day in three pided daily doses), was admitted to Institute for Child and Youth Health Care of Vojodina (IZZZDIOV) with rare skin petechias. Platelet level was 86G/l. Other haematological, biochemical and serological findings (IgM adenovirus, EBV, Coxackie virus) were within normal limits. Renal and liver functions were normal. Two days before admission child was febrile for about three days (up to 39oC axillary) without clear cause. He was using ibuprofen during two days.

The peripheral smear showed thrombocytopenia. Platelet function tests were normal. Serum valproate level was 436 (345–695 mmol/L). In view of persistently low platelet count, not reacting to intravenous immunoglobulin 400mg/kg/day during 5 days, and methilprednisolon 1mg/kg/day during 5 days, possibility of valproate induced thrombocytopenia was considered and a decision was made to withdraw valproate. VPA was completely withdrawn on 21st day of hospitalization. This was followed by a dramatic increase in the platelet count in three days.

Case 2 – A eighteen months old male child, on VPA monotherapy for atypical febrile seizures (21 mg/kg/day in three pided doses) for 183 days, was admitted to IZZZDIOV with petechial rash. Three days before admission child was febrile for about three days (up to 39oC axillary temperature) without clear cause. He was using ibuprofen during two days.

On admission the child had thrombocytopenia, platelet count 24G/l. Other investigations showed leucocytes 12G/l neutrophils 25%, lymphocytes 75%, haemoglobin 8 g/dl. All other haematological, biochemical findings were within normal limits. Serological analyses (IgM adenovirus, EBV, Coxackie virus) were negative. Renal and liver functions were normal. The peripheral smear showed lymphocytes and thrombocytopenia. Serum valproate level was 450 (345–695 mmol/L).  In view of persistently low platelet count, not reacting to intravenous immunoglobulin 400mg/kg/day during 5 days, and methilprednisolon 1mg/kg/day during 5 days, possibility of valproate induced thrombocytopenia was considered. Valproate was completely withdrawn on 17th day of hospitalization. This was followed by the platelet count normalization in five days.

Case 3 – A 27 months old female child, on VPA monotherapy (19 mg/kg/day in three pided doses) for 211 days for recurrent atypical febrile seizures was admitted to IZZZDIOV for thrombocytopenia registered in routine laboratory check up according to febrile episode of unknown origin. Two days before admission, the child was febrile for about three days (up to 39oC axillary temperature) without clear cause. She was using ibuprofen during two days.

On admission she had thrombocytopenia, platelet count 70G/l. Other investigations haematological and biochemical findings were within normal limits. Serological analyses (IgM adenovirus, EBV, enterovirus, Coxackie virus) were negative. Renal and liver functions were normal. The peripheral smear showed thrombocytopenia. Serum valproate level was 349 (345–695 mmol/L).  In view of persistently low platelet count, not reacting to intravenous immunoglobulin 400mg/kg/day during 5 days, and methilprednisolon 1mg/kg/day during 5 days, possibility of valproate induced thrombocytopenia was considered and a decision was made to withdraw VPA. VPA therapy was completely stopped on 28th day of hospitalization.  This was followed by normalization in the platelet count in seven days.

On follow-up for a three years period, our patients have had stable haematological parameters. First child got antiepileptic therapy (lamotrigine) after seizure reappearance. Second and third child are still without antiepileptic therapy.

Discussion

Valproate may cause thrombocytopenia via more than one mechanism. The most common factor is the presence of high VPA serum levels. Platelet counts should probably be monitored more carefully in patients known to have higher drug levels (5). Serum levels of our patients were in lower third of therapeutic levels.

Delgado et al., 1994. reported that 8 of 64 patients with VPA-associated thrombocytopenia developed signs of bleeding (2). Our patients did not bleed. They have only skin petechial spots. A bleeding diathesis associated with VPA use may include thrombocytopenia, as well as abnormal platelet function, megakaryocytic dysplasia, acquired Von Willebrand disease type I, and a deficiency of factor XIII (5,6,7). Abnormal platelet function, megakaryocytic dysplasia, acquired Von Willebrand disease type I, and a deficiency of factor XIII were excluded in all our patients.

Viral infections have been suggested as etiological agents for clinically significant episodic thrombocytopenia in patients treated with VPA. This may occur when the VPA serum levels are in the normal rang, as in our cases. The negative diagnostic work up and resistant thrombocytopenia which responded remarkably to VPA withdrawal suggested drug adverse reaction as the predisposing cause. Neither of our children had a clear recognizable cause for high fever. Haematological parameters done prior to starting valproate, as well as during treatment were documented to be normal in all presented children.

The mechanism of VPA-induced thrombocytopenia is unclear. The possibilities include a dose-dependent suppression of bone marrow production of platelets or peripheral platelet destruction due to the development of a platelet antibody brought about by VPA or one of its metabolites.  While the VPA concentration-dependent thrombocytopenia could be explained by the mechanism of bone marrow suppression, the rare occurrence of pancytopenia suggests that other mechanisms must be involved that renders the platelet cell lineage more vulnerable to VPA suppression or damage (8).

Previous studies have indicated that the interval between the initiation of VPA treatment and platelet nadir is variable among patients who develop thrombocytopenia, ranging from 8 days to 16 months (2). The mean time from VPA exposure to the first observed episode of thrombocytopenia in our patients was 174 days (range 127-211), but it was 3 days after febrile episode of unknown origin.

Women are described to be more likely to develop thrombocytopenia compared to men with females at approximately twice the risk (8). In our group boys were dominating (2 vs. 1).

VPA induced thrombocytopenia was reversible in all our patients, following VPA discontinuation, which is a consistent finding with a number of previous studies (2).

Based on the known adverse drug reaction between valproate and the thrombocytopenia, complete platelet normalization after VPA withdrawal proved this association.

Conclusion:

Need to consider valproate drug toxicity in the differentials for thrombocytopenia even in children with clinical presentations that may be more suggestive of an infective aetiology is emphasized.  In all children using VPA, the platelet count should be monitored during and after febrile episodes underpinning of its aetiology.

Table 1.  Characteristics of patients with thrombocytopenia

BIBLIOGRAPHY

Correspondence to:
Knežević-Pogančev Marija Institute for Child and Youth Health Care of Vojvodina, Department for developmental neurology and epileptology Novi Sad, Hajduk Veljkova 10, 21000, Novi Sad, Serbia mkp.marija@gmail.com