A 59-year-old man came to the Emergency Department with a history of malaise in the last month and confusion in the last 3 days. He reported no fever or localised complaints. The patient had a history of poorly controlled diabetes, dyslipidaemia and schizoaffective disorder. He was agitated, his temperature was 38.3oC, cardiopulmonary auscultation was normal and he had mild ankle oedema. The most significant laboratory data were 18,700 leukocytes/mm3, 10.4 g/dL haemoglobin, plasma glucose 750 mg/dL and C-reactive protein 12.6 mg/dL. A blood culture was drawn, he was admitted with a presumptive diagnosis of hyperosmolar non-ketotic syndrome secondary to acute infection, and was treated with intravenous fluids, insulin and cefuroxime. Improvement was noted. On the third day, the clinical microbiology laboratory detected growth of small gram-negative bacilli in both blood culture bottles. One day later, rough, raised, non-hemolytic, non-hemolytic colonies of various sizes grew on the blood agar and chocolate agar plates, but not on the MacConkey agar plate. The colonies were positive for oxidase and negative for lacatalase, indole and urease. The isolate was identified, using the API 20E (bioMérieux) and RapIDNF Plus (Remel) systems, as Pasteurella haemolytica and reported as sensitive to cefuroxime. The patient was re-interviewed and reported that he had a cat at home that had bitten him several times in the last month. During the remainder of his hospitalisation the patient was afebrile but had several episodes of pulmonary congestion. A transthoracic echocardiogram (TTE) revealed moderate left ventricular dysfunction and mild mitral regurgitation. When stabilised, he was transferred to a rehabilitation unit where he completed a 10-day course of cefuroxime. One week later he was readmitted for increased dyspnoea and peripheral oedema. He was treated with diuretics with minimal improvement. He underwent a repeat TTE which this time showed significant aortic regurgitation and moderate mitral regurgitation. Due to persistent pulmonary congestion the patient was referred to surgery for replacement of his aortic and mitral valves. Small vegetations on the aortic valve were observed at the operation.
Blood culture criteria require the isolation of a microorganism of the Streptococcus viridans group, Streptococcus bovis, one of the HACEK group (Haemophilus species, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenellacorrodens and Kingella species) or Staphylococcus aureus in at least 2 blood cultures taken 12 or more hours apart. Therefore, the collection of more than one blood culture before starting antibiotic treatment and the correct identification of blood isolates are of paramount value in the diagnosis of IE.
Valvular vegetation was cultured, but no growth was obtained; there was also no growth in the additional blood cultures taken on the second admission. Culture-negative endocarditis can result from the use of inadequate microbiological techniques, infection by very demanding pathogens (such as Bartonella, Brucella or Chlamydia species), or antimicrobial administration prior to blood culture collection.
In the present case, the 16S rRNA amplicon sequence of the valve tissue had >99% concordance with that of Haemophilus aphrophilus. Subsequently, the blood isolate previously identified as Pasteurella haemolytica was identified as H. aphrophilus.
In the present case, the "chain of errors" started with insufficient inoculum to the tubes with oxidation/fermentation medium resulting in the erroneous characterisation of the isolate as "non-fermenting" which led to the selection of identification systems used for non-fermenting gram-negative bacteria, irrelevant in this case. Another possible cause of confusion for clinical microbiologists is that H. aphrophilus is one of only two Haemophilus species (the other being Haemophilus ducreyi) that does not require NAD (factor V) for growth and therefore can grow on blood agar. The genera Haemophilus and Pasteurella both belong to the same family (Pasteurelleae) and misidentification of H. aphrophilus as Pasteurella using commercial identification kits is well known.

In the present case, it is possible that correct identification and reporting of the isolate would have alerted the clinician to the possibility of IE and would have led to appropriate treatment at the first hospital admission. Therefore, it could have prevented the progression of valve destruction and the need for valve replacement.
