These authors contributed equally.

**Author details**

Juliane C. Finke1,2#

Berlin, Germany

**References**

Zürich, Switzerland

Michael O. Glocker<sup>1</sup>

#### **4.3. Full analysis set and per protocol set**

Data from the raw data set were inspected for completeness and categorized into three groups based on the detection rates. Group I contained those cytokines and growth factors (CGFs) for which plasma protein concentrations could be determined in over 80% of all samples. Group II included those CGFs for which plasma protein concentrations could be determined in over 60% but less than 80% of all samples. Group III contained proteins whose concentrations were determined in less than 60% of all samples and sCD40L which did not pass the QC test. In both, group I and group II proteins, the missing values were imputed using the lower limit of quantitation (LLOQ) except for MCP1 for which upper limit of quantitation (ULOQ) was imputed. After imputing, the "full analysis set (FAS)" contained a total of 1044 curated data points (data not shown). The "per protocol set (PPS)" was generated out of the "full analysis set (FAS)" by including only group I and group II CGFs, resulting in 754 data points.

#### **4.4. Biostatistical analysis**

solution was serially diluted by a factor of 5 and yielded in diluted human cytokine standards with the following concentrations: 10,000, 2000, 400, 80, 16, and 3.2 pg/mL. Next, each antibody-bead containing vial was sonicated for 30 s and then vortexed for 1 min. Sixty microliters of antibody-bead slurry from each of the 18 vials was added to the mixing bottle, and 1.92 mL "bead diluent" was added to achieve a final volume of 3.0 mL. Two 96-well plates were pre-wetted with 200 μL wash buffer, each. After sealing, the plates were fixed on the shaker (Heidolph® Promax 2020, Schwabach, Germany) and gently agitated for 10 min at room temperature. The wash buffer was decanted, and the residual amount was removed by inverting the plates and gently tapping onto absorbent towels for several times. Twenty-five microliters, each, of all six diluted human cytokine standards and the two quality control samples were added into their dedicated wells. Twenty-five microliters, each, of assay buffer were added to two "background" wells and to the wells that were dedicated to patient samples. Second, 25 μL of serum matrix was added into each of the diluted human cytokine standard wells, the background wells, and the quality control sample wells. Twenty-five microliters of each patient plasma was added into one of the patient sample wells. Twenty-five microliters, each, of antibody-bead slurry from the mixing bottle was added to all the wells. Afterwards, plates were sealed and incubated in the dark for 18 h at 4°C on the plate shaker. Solvents from each well were removed, avoiding loss of beads, and beads were washed twice with wash buffer (200 μL, for each well, 1 min incubation). After removal of wash buffer, 25 μL, each, of detection antibody solution was added to all wells. After 1 h incubation at 25°C, 25 μL of phycoerythrin-loaded streptavidin containing solution was added to all wells. The plates were sealed again, covered with aluminum foil, and then fixed on the plate shaker for 30 min at 25°C. Subsequently, beads were washed twice with wash buffer (200 μL, for each well, 1 min incubation). After removal of wash buffer, 150 μL of sheath fluid (Bio-Rad Laboratories, Hercules, CA, USA) was added to all wells and beads were resuspended on the plate shaker for 5 min. Plates were placed into the Bio-Plex suspension array 200 System (Bio-Rad Laboratories, Hercules, CA, USA), which had been calibrated with Bio-Plex® 200 calibration kit and validated with Bio-Plex® validation kit 4.0 (Bio-Rad Laboratories, Hercules, CA, USA). Measurement settings were as follows: data acquisition, 50 beads per region; sample size, 100 μL; and doublet discriminator gate, 5000–25,000 (low photomultiplier tube). The Bio-Plex suspension array 200 reader system contains a red laser for identification of the bead which is analyzed and a green laser for quantification of fluorescence intensity of phycoerythrinloaded streptavidin. All standards, controls, background, and plasma samples were prepared in duplicate and measured once. From the 2088 independent measurements, the Bio-Plex® manager 6.1 software (Bio-Rad Laboratories, Hercules, CA, USA) calculated median fluorescence intensity (MFI) and standard deviations of each duplicate recording. Fluorescence values of human cytokine standards were plotted as standard curves which were used for determining plasma concentrations (pg/mL) based on their fluorescence intensities of all proteins and all time points. In total, 1044 data points (raw data set) were stored as Excel files.

Data from the raw data set were inspected for completeness and categorized into three groups based on the detection rates. Group I contained those cytokines and growth factors (CGFs) for which plasma protein concentrations could be determined in over 80% of all samples. Group

**4.3. Full analysis set and per protocol set**

18 Issues in Flap Surgery

Statistical analyses were performed using the PPS with the IBM statistics software SPSS (version 20.0, SPSS Inc., Chicago, USA). Linear fit analysis between ischemia time and protein concentration was performed using the Origin statistics software (version. 8.1 G; OriginLab Corporation, Northampton, MA, USA). Linear regression was performed to calculate *R*<sup>2</sup> values, and ANOVA tests were performed to calculate *p* values to estimate whether protein concentration was related to the ischemia time [53]. Hierarchical cluster analysis and dendrogram presentation were performed on the Knowledge Discovery Environment (KDE) platform (InforSense Ltd., London, UK). Parameter settings were single linkage and Euclidean distance. CV% for each analyte was calculated as the ratio of the standard deviation to the plasma mean concentration [25, 26].

## **Author details**

Juliane C. Finke1,2# , Jingzhi Yang1# , Marius Bredell<sup>3</sup> , Uwe von Fritschen2 and Michael O. Glocker<sup>1</sup> \*

\*Address all correspondence to: michael.glocker@uni-rostock.de

1 Proteome Center Rostock, University Medicine and Natural Science Faculty, University of Rostock, Rostock, Germany

2 Division of Plastic Surgery and Hand Surgery, HELIOS Clinic Emil von Behring, Berlin, Germany

3 Department of Cranio-Maxillofacial and Oral Surgery, University Hospital of Zürich, Zürich, Switzerland
