During January 25, 2010–July 19, 2012, patients were prospectively enrolled in this study at 2 tertiary care referral hospitals located on the campus of the medical school of Khon Kaen University. Srinagarind Hospital is a 777-bed general hospital, and the Queen Sirikit Heart Center of the Northeast is a 200-bed specialized cardiac center in which ≈10 heart valve replacement surgeries are performed each month. Patients with suspected IE are referred from much of northeastern Thailand, a region of ≈21 million persons.

Transthoracic echocardiography was performed for patients suspected of having IE; consenting patients >16 years of age who met modified Duke criteria for endocarditis were enrolled in this study. Underlying cardiac conditions were assessed by cardiologists on the basis of patients’ medical records, history, physical examination, and echocardiographic findings. At admission to a hospital, 3 separate blood samples for culture were obtained in <90 minutes. Blood was inoculated into aerobic medium (BD BACTEC Plus Aerobic/F Medium; Becton Dickinson, Franklin Lakes, NJ, USA), and cultures were processed by using an automated system (BD Bactec Fx series; Becton Dickinson). Pathogens were identified to species level whenever possible, but some blood culture isolates were defined only to the genus level (e.g., viridans group streptococci). One month after admission, a convalescent-phase serum specimen was obtained from each study patient and these patients were evaluated by a cardiologist.

Acute- and convalescent-phase serum specimens were tested for C. burnetii and Legionella pneumophila by indirect immunofluorescence assay (IFA) as described (5). Phase 1 IgG reciprocal titers >800 for C. burnetii and total antibody reciprocal titers ≥256 for L. pneumophila on either serum specimen were defined as positive (5). IFA IgG reciprocal titers of ≥800 to Bartonella quintana and Bartonella henselae were deemed positive. Specific antibodies to Brucella melitensis and Mycoplasma pneumoniae were detected with a commercial immunoenzymatic antibody test (Brucella antibody and Platellia M. pneumoniae IgM kits, respectively; Bio-Rad, Marnes-la-Coquette, France). Reciprocal titers ≥200 were considered positive. DNA was extracted from surgically excised heart valves by using the QIAamp DNA FFPE Tissue Kit (QIAGEN, Courtaboeuf, France) as described by the manufacturer. Previously described broad spectrum PCR primers and amplification and sequencing conditions (4) were used to detect all bacteria (16S rRNA); all fungi (18S rRNA); Staphylococcus aureus, mitis and gallolyticus group streptococci, Enterococcus faecalis and E. faecium, Mycoplasma hominis, C. burnetii, Bartonella spp., and Tropheryma whipplei.

Table 1 describes characteristics of the 132 enrolled patients: the median age was 47 years (range 16–85) and 68.9% were male. Most of the study patients lived in rural areas, most had a history of animal contact, and most were farmers of rice or vegetables. All study patients had definite IE as determined by using modified Duke criteria. Patients had high fever and were severely ill at admission; more than half had congestive heart failure (Table 1). Underlying cardiac pathologic changes were identified in 96 (72.7%) study patients (Table 1); rheumatic heart disease (RHD) was the most common condition, as has been found in other resource-limited countries (1,2,6,7). RHD was identified in 37 (28.0%) of the 132 patients overall; the 37 represented 38.5% of the 96 patients with identified underlying cardiac pathologic changes. IE was detected on prosthetic heart valves in 12 (9.9%) of the study patients, including 2 patients whose blood cultures grew coagulase-negative staphylococci.

Of the 132 study patients, 100 (75.8%) underwent surgery. Such management reflects current international guidelines for complicated IE, which emphasize identifying high-risk patients, transferring them to a specialized medical–surgical center, and performing early valvular surgery (8). Within 1 month, 11 (8.3%) of the 132 patients died and 6 were still hospitalized; 5 died after discharge. Four additional patients are known to have died after leaving the study, but we did not systematically assess survival beyond the 1-month follow-up examination. The case-fatality proportion calculated on the basis of known deaths was 11.4%, but the possibility of additional fatalities cannot be excluded. The mortality rate was 18.8% for patients who did not undergo surgery and 9.0% for those who did.

A pathogen was identified for 60 (45.5%) cases (Table 2). The etiologic agent was identified for 7 (21.9%) of 32 nonsurgical cases, 5 by blood culture and 2 by IFA, compared with 53 (53.0%) of the 100 study patients who underwent surgery (Fisher exact test, p = 0.02). Among patients who underwent surgery, diagnosis was made by PCR of heart valve tissue for 29, by blood culture for 11, and by >1 method for the remaining 13; no discrepancies between results from different diagnostic modalities were found. Many (80.3%) patients had been referred from other hospitals for management of IE and were already being treated for that illness. We confirmed that 77 (58.3%) patients had received antibacterial drugs during the week before admission. We suppose that the low diagnostic yield was associated with antibacterial drug use, but more detailed information is needed to determine the accuracy of this supposition.

Among the 60 cases of IE for which a causative pathogen was identified, 15 (25.0%) were attributed to zoonotic bacteria (Table 2): C. burnetti (5 cases), B. henselae (4 cases), Streptococcus suis (4 cases), Erysipelothrix rhusiopathiae (1 case), and Campylobacter fetus (1 case). B. quintana causes ≈75% of Bartonella IE cases worldwide (9) but was not found in this study. Verification that C. burnetti (10), Bartonella spp. (11), E. rhusiopathiae, and C. fetus cause IE in Thailand highlight the noteworthiness of emerging zoonotic pathogens in this region. The identification of Q fever IE stimulated efforts by public health authorities of Thailand to characterize Q fever prevalence and transmission nationwide.

Streptococci generally cause the majority of infections of native heart valves (12), and 90.1% of patients in this study had native valve involvement. Viridans streptococci, including S. suis, accounted for 43.3% of cases of known etiology (Table 2) and included 3 species of streptococci not previously reported to cause IE in Thailand (S. gallolyticus, S. difficilis, and S. sinensis). In Thailand, as elsewhere, health care–associated IE and IE among intravenous drug users is often caused by S. aureus (13,14). None of the patients in this study were known to be intravenous drug users, and few had health care–associated infections; 5 (8.3%) of 60 cases were caused by S. aureus.

Zoonotic bacteria were detected by specialized testing of 11.4% of IE patients in northeastern Thailand, but an etiology could not be determined for more than half (54.5%) of the patients. Infections with viridans streptococci predominated among cases of known etiology, and RHD was the most common underlying pathologic change. To increase diagnostic yield, we recommend intensification of efforts to obtain blood cultures in the absence of antimicrobial drugs and increase of laboratory capacity to test for zoonotic bacteria (15).