Pyomyositis in the upper Negro river basin, Brazilian Amazonia

álvaro H.D. Borges, Brian Faragher, David Lalloo

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)

Abstract

Pyomyositis remains poorly documented in tropical Latin America. We therefore performed a retrospective review of cases admitted to a hospital in the upper Negro river basin during 2002–2006. Seasonality was assessed by the cosinor model and independent predictors of outcome were identified by logistic regression. Determinants of time-to-fever resolution were analysed using Cox regression. No seasonal trend was observed (p=0.284) among 82 hospitalised patients. The disease predominated in young males and the most commonly affected part of the body was the lower limb (68 [63.5%] out of 107 lesions). Staphylococcus aureus was the only identified infecting organism (18 of 20 culture results, 90%). Complications occurred in 17 patients (20.7%) and the case fatality rate was 2.4%. Children were more likely to present with eosinophilia than adults (OR= 4.20, 95% CI 1.08–16.32, p=0.048), but no other significant differences regarding clinical presentation and outcomes were observed. The time-to-fever resolution was the only independent determinant of poor outcome (OR=1.52, 95% CI 1.22–1.92, p<0.001) and was significantly longer in patients treated with combined antibiotic therapy than in those treated with single antibiotics (HR=0.523, 95% CI 0.296–0.926, p=0.026). Further studies to determine the best antibiotic therapy modality for the treatment of pyomyositis are required.

Original languageEnglish
Pages (from-to)532-537
Number of pages6
JournalTransactions of the Royal Society of Tropical Medicine and Hygiene
Volume106
Issue number9
DOIs
Publication statusPublished - 1 Sept 2012

Keywords

  • Amazon
  • Diagnosis
  • Epidemiology
  • Outcome
  • Pyomyositis
  • Treatment

Fingerprint

Dive into the research topics of 'Pyomyositis in the upper Negro river basin, Brazilian Amazonia'. Together they form a unique fingerprint.

Cite this