Study protocol: prediction of stroke associated infections by markers of autonomic control

Background: Infection is the most important complication after acute stroke. This is substantially based on a stroke-induced immunosuppression. Heart rate variability (HRV) represents the autonomic nervous system activity in connection with stroke-induced immunomodulation and infections. We demonstrated in a feasibility study that HRV indices obtained in patients without acute post-stroke infections can predict infections in the subsequent sub-acute phase. Methods/Design: The study PRED-SEP is a prospective observational study. Adult patients with acute ischemic infarction in the territory of the middle cerebral artery and severe neurological deficit (National Institutes of Health Stroke Scale: NIHSS >= 8) are recruited. Primary endpoint is the development of infections, secondary endpoints are SIRS and severe sepsis in the sub-acute phase (day 3-5) after stroke and the functional outcome after 3 months. Infection is defined according to the PANTHERIS study and comprises pneumonia, urinary tract infection and infections without determined focus. SIRS and severe sepsis are defined according to German Sepsis Society guidelines. Functional outcome is measured by lethality and neurological scores (modified Rankin Scale, Barthel Index). Prognostic factors are HRV risk indices calculated from selected intervals of 24 h ECG measurements within 48 hours after symptom onset. It is planned to recruit 240 patients. HRV risk indices (predictors) will be calculated according to standards and procedures previously developed and published by the authors. The predictive effects of HRV indices on infections will be estimated by fitting logistic regression models and estimating odds ratios with 95% confidence intervals. A prespecified modelling procedure will be applied to estimate unadjusted and confounder adjusted odds ratios. Secondary endpoints will be analysed in the same way. The functional outcome scales will be dichotomized. The association between HRV indices and pro-and anti-inflammatory markers will be quantified by calculating the appropriate correlation coefficients according to scale (Person or Spearman). Discussion: Since a general prophylactic antibiotic treatment after stroke is not recommended, results of this study could have essential implications for an early identification and hence, timely appropriate treating of stroke-induced infections.