Article Text

Download PDFPDF

130 Fever in an intravenous drug user: if it isn’t endocarditis then what is it?
  1. Sam Straw1,
  2. Wazir Baig1,
  3. Jianhua Wu2,
  4. Jonathan Sandoe1
  1. 1Leeds Teaching Hospitals NHS Trust
  2. 2University of Leeds


Background Intravenous drug use (IVDU) is a predisposition for the development of infective endocarditis (IE). The incidence amongst people who inject drugs (PWID) is estimated to be 50–100 times greater than the general population. For this reason, PWID with symptoms of infection or bacteraemia are frequently referred to the IE team for assessment. We have previously reported on our experience of managing PWID with IE over more than a decade. Here we present the characteristics, final diagnoses and outcomes in all PWID admitted to hospital with other infections in whom IE was suspected.

Methods Patients aged 18 or over who had taken drugs intravenously within 90 days and referred to the IE team between 01/01/2006 and 31/12/2016 were eligible. Inclusion was dependent on the modified Duke criteria, we required patients to be either Duke criteria negative, or Duke possible and not treated for IE. PWID referred to the IE team who were either Duke definite or Duke possible and treated for IE acted as a comparison group. Patient demographics, microbiology, final diagnoses, survival and causes of death were recorded.

Results There were 112 episodes of other infections in 107 PWID during the study period. 69 (65%) patients were male, and the mean age was 36 ± 8.4 years. During the same period, we assessed and treated 105 episodes of confirmed IE in 92 PWID. There were positive blood cultures in 71 episodes, most commonly Staphylococcus aureus (39), coagulase negative Staphylococcus (3), beta-haemolytic Streptococcus (16), oral Streptococcus (7), Streptococcus angionosis (4), Enterococcus spp. (2), Pseudomonad (1) and other (7). In 10 episodes there was combined infection with two or more microorganisms, and in 41 episodes blood cultures were negative. Where IE was excluded, the final diagnosis was usually an infected deep vein thrombosis (DVT) (41) or skin and soft tissue infection (26). Other infections included osteomyelitis (9), septic arthritis (6), septic pulmonary emboli/lung abscess (5), central line infection (4), pneumonia (4), urinary tract infection (2), pericarditis (2) and in one episode there was a cerebral abscess. Survival following hospitalisation with an episode of infection was 91%, 86.4%, 83.5% and 70% at 1, 3, 5 and 10 years respectively. In univariate analysis, survival was significantly higher in PWID with other infections compared to those treated for IE during the same period (p=0.0002) (figure 1). During a median follow-up of 4.8 years there were a total of 26 deaths, with suicide or drug overdose being the cause of death in 9 cases.

Conclusion In approximately half of PWID referred to the IE team the final diagnosis was non-cardiac infection. In those with other infections the alternative diagnosis was usually an infected DVT or skin/soft tissue infection. Survival was favourable in those with other infections compared to those treated for IE. The implication of this finding is that cardiac infection has a negative impact on survival in PWID beyond infection elsewhere. It seems likely that damaged native heart valves, or indwelling prosthetic material poses additional risk of re-infection in the context of ongoing drug use. The continuing high mortality rate in both groups suggests that improving the support offered to PWID might affect outcomes in the longer term.

Conflict of Interest None

  • Infective endocarditis
  • Intravenous drug use
  • Valve disease

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.