Literature Summary

Major Literature

2015 - Serum PCT as a Diagnostic Biomarker for Sepsis in Burned Pts (Burns, meta-analysis)

• Method

  • Looked at studies published up to March 2014

  • Study eligible if:

    • Provided sensitivity/specificity and true/false negative/positive data

    • Needed to be primarily adult population

• Results/Discussion

  • 9 eligible studies

    • Subject number in studies ranged from 25-175

    • Avg age 30s-50s

    • Avg TBSA 30-60%

    • Cutoff of "positive" PCT ranged from 0.5-3 (our cutoff here is 0.5)

  • Data analysis

    • Pooled sensitivity = 0.74 (fraction of pt with sepsis who had a positive PCT)

    • Pooled specificity = 0.88 (fraction of pt without sepsis who had a negative PCT)

    • Pooled positive likelihood ratio = 5.75 (pt with sepsis is 5.75 times more likely to have a positive PCT than healthy pt)

    • Pooled negative likelihood ratio = 0.33 (same concept as above)

    • Pooled diagnostic odds ratio = 22.6 (value of 1 means test can't differentiate pts with sepsis to pts without sepsis, 22.6 = moderate level of accuracy)

    • Area under the ROC = 0.92 (1 = perfect test to differentiate pts with sepsis to pts without sepsis, 0.92 = excellent test)

  • 1 of the study reported that PCT at the time of admission could serve as a prognostic marker, but not enough studies to perform a meta-analysis

• Conclusion

  • Authors of this paper gives relative affirmative view of PCT

    • "...one of the most promising sepsis biomarkers in burned pts"

  • Threshold at which PCT is positive remains unclear (this is a big issue!)

    • Evaluating the trend of PCT may be more important than absolute values

Since this meta-analysis, 1 other notable study has been published:

2014 - A New Marker of Sepsis Post Burn Injury? (Critical Care Med, prospective observational)

• Methods

  • Prospective observational study

  • 54 pts with =/>15% TBSA burn, intubated with no previous CV comorbidities

    • Got PCT and BNP samples daily to test accuracy for detecting new onset sepsis

      • Why BNP? BNP have been shown to be elevated in sepsis, likely due to myocardial depression that occurs in 50% of pts with septic shock and decrease in BNP clearance (complex pathophys)

  • If pt met the American Burn Association Criteria for Sepsis (see below, different from our usual SIRS + suspected/documented infection) --> cultures sent

• Results

  • 23 of 54 pts had confirmed sepsis

  • BNP (cutoff = 1200)

    • Sensitivity 96%

    • Specificity 99%

    • Positive likelihood ratio 166

    • Negative likelihood ratio 40

    • Diagnostic odds ratio 4030

    • No correlation between BNP levels and exceeding Parkland formula

  • PCT (cutoff = 1.4)

    • Sensitivity 20%

    • Specificity 64%

    • Positive likelihood ratio 0.6

    • Negative likelihood ratio 1.2

    • Diagnostic odds ratio 0.5

  • Multivariable model analysis adjusted for severity of burn --> BNP is an independent marker of sepsis

• Conclusion

  • Routine BNP that results in >1200 helpful in identifying pts with sepsis while PCT is not helpful in pts with no CV comorbidities