A pilot study of a novel molecular host response assay to diagnose infection in patients after high-risk gastro-intestinal surgery

Abstract

AimSeptiCyte LAB measures the expression of four host-response RNAs in blood to distinguish sepsis from sterile inflammation. Sequential monitoring of this assay may have diagnostic utility in patients at high risk for postoperative infectious complications. MethodsIn this pilot study we studied esophagectomy patients who had developed a complication within 30 days following surgery as well as a random sample of 100 uncomplicated postoperative patients. PAXgene blood samples were collected postoperatively and whenever a complication occurred. SeptiCyte scores (ranging 0-10 with increasing likelihood of infection) were compared to post-hoc physician adjudication of infection likelihood using strict definitions. ResultsAmong 370 esophagectomy patients, 120 (32%) subjects developed a complication requiring ICU (re)admission, 63 (53%) of whom could be analyzed. Immediate postoperative SeptiCyte LAB scores were highly variable, yet similar for patients having a complicated and uncomplicated postoperative course (median score of 2.4 (IQR 1.6-3.3) versus 2.2 (IQR 1.3-3), respectively). Scores increased as complications developed, but this rise was higher for 34 subjects having confirmed infection (median difference 4.7 (IQR 4.1-5.8)) then for 12 subjects with a non-infectious complication (2.1 (IQR 0.4-3.6); p<0.0001). When 17 cases with undetermined infectious status were excluded, addition of SeptiCyte LAB to CRP resulted in improved diagnostic discrimination (AUC 0.88 (95%CI 0.77-0.99)) compared to CRP alone (AUC 0.76 (95%CI 0.61-0.91); p=0.04). ConclusionsSequential measurement of SeptiCyte LAB may have clinical utility in surgical patients at high risk of postoperative infection, but its diagnostic performance in this setting needs to be further evaluated.

Novel Approaches for Diagnosing Sepsis in the Critically Ill

“…Another study, explored the potential utility of longitudinal monitoring of SeptiScores™ in 67 patients who suffered from complications after elective esophagectomy [72]. SeptiCyte™ LAB results were analyzed pair-wise, comparing samples drawn immediately after surgery and at onset of a postoperative complication, which was then categorized as being of infectious, undetermined, or non-infectious origin [69].…”

“…The aforementioned clinical performance data suggest that SeptiCyte™ LAB has a sensitivity above 90%, which is high compared to an overall sensitivity of only 60% for culture-based methods in ICU patients with sepsis [75]. Yet specificity is highly variable and discriminative ability (as expressed by AUCs) for distinguishing sepsis from non-infectious causes of systemic inflammation, is at least equivalent to PCT and CRP (table 4) [41,72,73,74].…”

“…However, as SIRS and organ failure are virtually omnipresent and rather non-specific in ICU patients [7,10], these cohorts may not be fully representative of patients having a true diagnostic dilemma who would receive the test in clinical practice. Equally important, most studies used a case-control design and thus included only highly selected patients [39,41,54,70,72]. This results in a study population that is enriched for obvious diagnoses, which may lead to overestimation of diagnostic performance [58].…”

“…This is a generally accepted method and acknowledged by several guidelines [76][77][78]. Four studies [39,41,72,73] used definitions as proposed by the MARS consortium, which have demonstrated good concordance among observers for most types of infection [69]. Other studies used a similar approach [55], single expert-opinion [70], or an expert panel [74] to adjudicate patients.…”

“…As the patient flow might be an important source of selection bias it is important to fully report this in diagnostic studies, which was not always done, or showed a large proportion (~50%) of unavailable samples that was not corrected for [39,54,72]. Finally, the maximum time interval between performance of the index test (SeptiCyte™ LAB) and the reference standard (physician diagnosis) was explicitly stated to be between 24 and 72 hours in three studies [41,70,74].…”