Systematic investigation and comparison of diagnostic modalities in implant associated infections and infected non unions

Project No. FF-FR 0280


completed 08/2020


As of now, a reliable diagnosis of infected non-unions is not established. Especially low-grade infections are frequently caused by pathogens of low virulence and possibly clinically inapparent. Pathogen detection from operative tissue samples still stands as diagnostic golden standard. By reason of low sensitivity of the standard diagnostic procedures, false-negative results could be present in up to 30% of cases. Recently, a growing number of studies suggest a possible increase in sensitivity through the implementation of sonication.

Objective of this study were as follows: First part was the validation of sonication in an experimental (laboratory) setting. Second part was the comparison of sonication to other diagnostic methods regarding the determination of the most sensitive procedure of pathogen detection. Hereby, correct and timely diagnosis of infected non-union would be enabled and well-targeted therapy could be delivered while avoiding undertreatment.


  • Hypothesis 1: Sonication is a valid procedure for determination of pathogen load.
  • Hypothesis II: Filtration-based sonication is superior to conventional sonication with regards to quantitative results.
  • Hypothesis III: Sonication is superior to conventional tissue sample cultures in terms of sensitivity and specificity for diagnosis of low-grade infection in non-unions.
  • Hypothesis IV: Sonication is non-inferior to the other diagnostic procedures.


Part 1: Analysis of quantitative accuracy of sonication in an experimental setting (hypotheses I and II):
Sterilized titanium test specimen were inoculated with a bacterial suspension of a defined concentration and quantity (within a serial dilution of 100 - 109 bacteria ml).

Part II: Implementation of sonication for the diagnosis of infected non-unions (hypotheses III and IV):
Study design: prospective comparative
Patient collective: lntervention/study group (n = 100) patients with non-union of the lower extremity; control group (n = 100) patients scheduled for elective implant removal after uneventful healing.
Diagnostic procedures: (filtration-based) sonication, standard tissue sample cultures, enrichment in blood culture bottles after tissue homogenization, histopathological examination.
Definition of infected non-union/fracture-related infection: Detection of phenotypically identical pathogens in at least two samples or pathogen detection in one sample with simultaneous proof of osteomyelitis in histopathology.
Statistics: The different diagnostic procedures were compared and examined for statistically significant differences. Predictive values as well as sensitivity and specifity were calculated.


Part 1:
Reduction factors were acceptable for both with and without soiling.

Part II:
In 67 patients (33.5%; non-unions n = 27; control n = 40), all samples remained sterile. In 133 patients (non-unions n = 73; control n = 60) at least one positive sample (= pathogen detection) was present. Coagulase-negative staphylococci and cutibacteria were detected most frequently in both study and control group as well as among all diagnostic procedures. Sonication showed the highest overall number of positive results (n = 305) in 91 patients (non-unions n = 51; control n = 40). lnfection was confirmed in 72 patients (non-union n = 45; control n = 27), high suspicion of infection was diagnosed in 41 patients (non-union n = 29; control n = 12). Positive predictive value (PPV) for sonication was lowest among all diagnostic procedures (sonication PPV 31.4%; standard tissue sample cultures 50%; tissue homogenization 55.7%). Concerning sensitivity (68.5%) and specificity (47.8%), sonication also presented the lowest values. Standard tissue sample cultures and tissue homogenization were about equal with regards to sensitivity and specifity as well as positive and negative predictive values. In our study, sonication failed to prove to be the most sensitive diagnostic procedure for determination of infected non-union. The infection rate was high in the study group (45%). Further research is warranted with regards to the therapeutic consequences of the high frequency of pathogen detection: Which of the patients with infected non-unions have to be submitted to extensive septic treatment? The presence of infection in 27% of patients in the control group, who had all presented uneventful healing, has to be considered when working towards the ultimate goal, prevention of under-and overtreatment.

Last Update:



Financed by:
  • Deutsche Gesetzliche Unfallversicherung e. V. (DGUV)
Research institution(s):
  • BG Klinikum Duisburg gGmbH

-cross sectoral-

Type of hazard:




Description, key words:

diagnostic modalities, trauma surgery, sonication