A 51-year-old man, healthy until 1 month ago, came to the hospital and reported dyspnoea on exertion and oedema in the lower limbs. He came from an urban environment, had no toxic habits, had not travelled abroad and had no contact with animals. Physical examination revealed a fever of 37.5oC and blood pressure of 175/76. Auscultation revealed a 3/6 diastolic-systolic murmur in the aortic focus and crackles in the left base. There was hepatomegaly at 5 cm from the costal margin and oedema in the lower limbs with fovea. Laboratory tests showed intense leukocytosis with left deviation. An echocardiogram showed a heart rate of 95 bpm, severe aortic insufficiency with a large vegetation on the aortic valve and a smaller one on the mitral valve, a slightly dilated and hypertrophic left ventricle with a normal LVEF. In view of the diagnosis of infective endocarditis, 3 blood cultures were taken, a serum sample was extracted for serological studies and treatment was started with diuretics, captopril, vancomycin and gentamicin. Blood cultures remained negative after 5 days of incubation, at which point ceftriaxone was added to the previous antibiotic treatment. Five days after admission, the patient required valve replacement surgery; the aortic valve was replaced with a mechanical prosthesis and the mitral valve was repaired. The culture of the valve was negative and its study with a technique recently incorporated into the microbiology laboratory provided the aetiological diagnosis the day after surgery. Several days later, the anatomo-pathological study of the valve confirms the results. After 20 days of incubation, the blood cultures were still negative.

The patient presented with infective endocarditis with negative blood cultures. Between 2.5% and 30% of cases of infective endocarditis have negative conventional blood cultures. In about half of these cases the cause is antibiotic treatment prior to blood culture and in the remainder it is caused by microorganisms that are difficult to culture. Depending on the clinical-epidemiological context, the most likely etiological agents are Coxiella burnetii, Brucella species, Bartonella species, HACEK group bacteria and to a lesser extent Tropheryma whipplei, Chlamydia species, Legionella pneumophila and others. In this case the patient had not taken antibiotics and we should suspect microorganisms that cannot be cultured under normal conditions. The serologies performed for Coxiella burnetii, Brucella, Bartonella, Chlamydia, Mycoplasma and L. pneumophila were negative, making it difficult to establish an aetiological diagnosis. Empirical broad-spectrum antibiotic treatment and special tests were therefore necessary to reach a diagnosis.

Currently, the main method for the aetiological diagnosis of infective endocarditis is still blood culture. Three blood cultures (two bottles per blood culture, aerobic and anaerobic) should be taken before antibiotic treatment is instituted. When blood cultures are negative, serology is the most effective diagnostic method if the endocarditis is caused by C. burnetii, Bartonella and Brucella as in these cases it has a chronic course and high antibody titres are produced and a single serum is necessary. In the case of other microorganisms such as Chlamydia, Mycoplasma and L. pneumophila, cross-reactions must be ruled out and their diagnostic value is lower.

In our case, a positive result was obtained on the same day of surgery for a universal 16S rRNA PCR performed with DNA extracted from aortic valve tissue. The amplification product was sequenced the following day and the sequence obtained was aligned in Genebank using BLAST software (http://www.ncbi.nlm.nih.gov/blast), presenting 100% identity with the T. whipplei sequence.

To confirm the diagnosis, specific PCR of the hsp65 region of T. whipplei was performed and was also positive, confirming the identity of the product obtained by sequencing. Therefore, in case of endocarditis with negative blood cultures, a serum should be obtained for serology for the mentioned microorganisms and, if possible, the heart valve should be sent to a reference laboratory for universal PCR of the 16S rRNA gene. A fragment of the valve or vegetation should be sent to the anatomical pathology laboratory to confirm the diagnosis of infective endocarditis and to perform special stains: Gram, Wartin-Starry (Bartonella), PAS (T.whipplei), etc., which have high sensitivity and specificity.
