Amoxicillin/clavulanic acid-associated severe neutropenia

Abstract

© Author(s) (or their employer(s)) 2021. No commercial reuse. See rights and permissions. Published by BMJ. A 72yearold man presented with an abnormal blood count and was admitted. His history included hypertension and hyperlipidaemia with an old myocardial infarction and mild stroke; liver cirrhosis due to nonalcoholic steatohepatitis (NASH) with hypersplenism and oesophageal varices but no ascites, oedema or bleeding; and benign prostatic hypertrophy. His medications (unchanged for years) included furosemide, spironolactone, bisoprolol, rosuvastatin, alfuzosin and omeprazole. Two weeks prior, he was discharged from our department after left leg cellulitis and Streptococcus pyogenes bacteraemia responsive to parenteral clindamycin and ceftriaxone (later changed to penicillin G) continued for 10 days. On discharge, haemoglobin was 10.2 g/dL (mean corpuscular volume 96), white blood cells 3.4×10/μL (neutrophils 2.9×10/μL) and platelets 36×10/μL with Creactive protein (CRP) 80 mg/ dL. Five days before admission his primary physician prescribed amoxicillin 875 mg/clavulanic acid 125 mg (Augmentin) two times per day for ‘trace cellulitis’. On admission he was afebrile and stable with unremarkable examination, ECG and chest Xray. Haemoglobin was 7.8 g/dL (mean corpuscular volume 98), white blood cells 0.6×10/μL (neutrophils 0.3×10/μL) and platelets 33×10/μL with CRP 11 mg/dL. Other blood tests were essentially unchanged: albumin 2.9 g/dL, globulins 3.9 g/dL, prothrombin time 18.4 s (48%) and mild liver enzyme disturbance. There was no evidence of bleeding, haemolysis or infection, and peripheral blood smear confirmed agranulocytosis without other new findings. Hepatitis viruses, herpesviruses, parvovirus B19 and COVID-19 were negative, as was an autoantibody panel. He was treated with recombinant human granulocyte colonystimulating factor (GCSF, ‘Neupogen’) quickly regaining his usual absolute neutrophil count (ANC). Haemoglobin increased to 9.1 g/dL and he was discharged home on the third hospital day. No previous drug hypersensitivity was known and followup was uneventful. According to a recent Centers for Disease Control (CDC) report, amoxicillin is the most frequently prescribed antibiotic for outpatients in the USA, and amoxicillin/clavulanic acid is the third, with 27.5 million prescriptions in 2018. Nevertheless, acute idiosyncratic agranulocytosis or severe neutropenia (ANC<500 cells/μL) associated with its use is extremely rare. 3 Our patient’s baseline pancytopenia due to splenic sequestration (due to hypersplenism in portal hypertension) was a confounding factor, but (as supported by the patient’s recent baseline blood counts) entirely different from the haematological findings in the current presentation, days after starting amoxicillin/clavulanic acid. The temporal association with the start of the drug, absence of any alternative causes and rapid reversibility with drug discontinuation (and G-CSF), support an adverse drug reaction (Naranjo Scale 7, possible adverse drug reaction). Thus, amoxicillin/clavulanic acid use can be continued in lowrisk febrile neutropenia, but its potential to be (although very rarely) associated with severe idiosyncratic neutropenia must be recognised. Finally, we are reminded again of the potential risks in prescribing even commonly used and ‘innocent’ drugs, amplified when in all probability (‘trace cellulitis’; CRP 11 mg/dL) no further medication was actually required.