Improved Diagnosis of Peritoneal Dialysis Peritonitis by Calorimetry

Background:

Peritonitis is still considered the most important complication of peritoneal dialysis (PD) associated with high mortality, therapy failure, and healthcare expenses. We have recently demonstrated that peritonitis continuous to be the most common reason for technical failure in PD contributing equally to early and late failure; 12 of 279 patients died as a consequence of PD associated peritonitis. The diagnosis of peritonitis is typically based on clinical symptoms, which reflects an already advanced state of inflammation and disease. Since the implementation of the blood culture system for detection of bacterial growth in PD effluent more than ten years ago, no considerable progress has been made to improve diagnosis of PD associated peritonitis. In our own PD population 36 out of 219 (16%) infectious peritonitis episodes were culture-negative. Usually, detection time of microorganisms requires at least 12 hours and time to pathogen identification more than 48 hours. Rapidly dividing cells such as bacteria produce heat, which can be measured by calorimetry. As recently shown, by our research group, calorimetry allows rapid and accurate diagnosis of bacterial growth in cases of meningitis and in contaminated platelets. In pilot experiments, we could show the potential and the feasibility of calorimetry for early and accurate detection of microorganisms in PD fluid.

Hypothesis and aims:

We hypothesize that calorimetry of PD effluent can significantly improve the diagnosis of PD associated peritonitis by early and accurate detection of pathogens. As compared to traditional culture methods, calorimetry may (i) have a shorter time to positivity and (ii) a higher sensitivity without loss of specificity. Specific aims are to evaluate if calorimetry: (i) is more rapid in detection of microorganisms compared to conventional blood cultures, (ii) is superior, regarding sensitivity and specificity, in detection of PD associated peritonitis compared to the standard blood culture system and (iii) is a valuable tool for pathogen identification by specific heat signal "signatures" received by calorimetry. The Primary endpoint is the time to positivity of bacterial growth by calorimetry and conventional blood cultures. The secondary endpoints are the accuracy (i.e. sensitivity and specificity) of detection rates of microorganism in PD fluid and the rate of consistency of pathogen identification between calorimetry curves and conventional microbiological method.

Research plan:

After obtaining informed consent, we will prospectively include in this open-labeled comparative study all patients aged 16 years or older, who are performing PD at the University Hospital in the time period from January 2009 through December 2011. PD fluid will be submitted for conventional screening examination (cell count and differential) and calorimetry (sterile MonovetteR tube) during routine visit. If PD peritonitis is suspected, PD fluid will be also inoculated in aerobic and anaerobic culture bottles at bedside under sterile conditions, each bottle with 10 ml of PD fluid. For calorimetry, 1 ml PD fluid will be cultivated at 37°C in the calorimeter tubes, containing 2 ml of trypticase soy broth (TSB). Conventional culture bottles and calorimeter tubes are incubated for a total of 6 days, after which culture and/or calorimetry are regarded negative. Medical records will be prospectively abstracted for demographic characteristics, clinical, radiographic, laboratory and microbiological data using a standardized case report form. The sample size calculation was performed separately for primary and secondary endpoints on the following settings: 0.05, power 80%, two-sided test. To reach statistical significant differences for the primary endpoint, 286 samples are required for the study. For the secondary endpoint with p0 = 80% (sensitivity of conventional blood cultures), p1 = 95% (sensitivity of calorimetry cultures), 255 samples are required. With an expected drop-out rate of 5%, approximately 300 samples will be needed. With expected 120-180 samples/year study duration of approximately 2.5 to 3 years is needed.

Expected outcome:

We anticipate that calorimetry will significantly improve the diagnosis of PD-associated peritonitis. First (primary endpoint), we predict an earlier pathogen detection by calorimetry; second (secondary endpoint), we expect to increase the sensitivity by 15% without loss of specificity (>90%). Furthermore, we assume that the most important pathogens, seen in our PD population (i.e. S. aureus, coagulase-negative staphylococci, streptococci, enterobacteriaceae) may be identified within 12 hours (compared to > 48 hours by conventional microbiological methods) with an accuracy of 80%. The ultra-sensitive calorimetry technique may diagnose PD associated peritonitis in an early disease stage with high accuracy. This would be the basis of a rapid and targeted antibiotic therapy, which may significantly influence the outcome of these patients.