Despite the major advances in perioperative management, surgical site infection (SSI) still occurs in up to 25% of patients undergoing potentially contaminated surgical procedures like colorectal surgery, resulting in prolongation of hospital stay and increased hospital costs.1 Although some clinical retrospective studies have suggested the beneficial effects of neuraxial anesthesia in preventing SSI,2,3 the underlying mechanisms remain to be determined.
Lipocalin-2, known as neutrophil gelatinase-associated protein, is expressed in the neutrophil, kidney, prostate, and epithelium4 and has been reported to regulate bacterial growth during infection.5,6 This protein has been shown to exert bacteriostatic effects through inhibiting iron uptake, which is essential for the growth of Gram-negative bacteria such as Escherichia coli.5 In the normal skin, expression of lipocalin-2 is restricted to hair follicle compartments, especially in the inner root sheath and infundibulum, suggesting the putative role of lipocalin-2 as a guardian against skin infection, because the infundibulum is an open interface between an organism and its environment.7 One study has demonstrated that lipocalin-2 binds to enterochelin, a bacterial catecholate siderophore, and thereby limits bacterial growth by depriving bacteria of their iron-uptake ability.8 Another study showed that lipocalin-2 was required for protecting lung from Klebsiella infection.6 We, therefore, hypothesized that expression of lipocalin-2 is likely to be augmented around surgical sites or skin (especially hair follicle compartments), resulting in a reduced risk of infection with E coli after abdominal surgery.
Our previous study showed that epidural anesthesia diminished endotoxin-induced gut mucosal injury through modulating the immune status.9 In the present study, we tested the hypotheses that lipocalin-2 is upregulated at infected surgical sites with E coli and that epidural anesthesia suppresses bacterial growth around surgical sites by augmenting the expression of lipocalin-2.
This study protocol was approved by the Animal Care and Use Committee of Keio University School of Medicine (authorization number: 10261) in accordance with the National Institute of Health guidelines.
Preparation of the Animals and Bacterial Cells
Male Wistar rats, weighing 250 to 350 g, were used for the study after a 3- to 7-day acclimatization period in our laboratory. Rats reared under a 12-hour light-dark cycle were provided access to standard chow and water ad libitum. Under sevoflurane anesthesia, the jugular vein and carotid artery were cannulated (PE-50, Intermedic, Sparks, MD) under sterile conditions. The catheters were tunneled under the skin to the neck and connected to a dual-channel stainless swivel (Primetech Corporation, Tokyo, Japan) which allowed continuous use of these catheters, while enabling rats to move freely in the cage. After these procedures, an epidural catheter was placed by using a microsurgical technique. In brief, each rat was placed in the prone position, and the lumber vertebral column was flexed by placing a roll of gauze transversely under the lower abdomen. A midline skin incision was made over the spinous processes of the L2-L3 vertebrae. After dissection of the interspinous ligament between L2 and L3, a catheter (PE-10, Intermedic) was inserted into the lumber epidural space, and advanced 35 mm in the cephalad direction. In a pilot study, the tip of this catheter was located at approximately the Th10-Th11 level. After completion of the study protocol, the position of the epidural catheter was verified by autopsy.
At the end of the preparatory surgery, a pseudosurgical site (1-cm-long vertical skin incision reaching up to the subcutaneous tissue, 1 cm to the right of the spine) was made on the back of all rats at the level of Th11-Th13. After they emerged from general anesthesia, the animals were placed in a metabolic cage, allowed to move freely and have water and standard laboratory chow ad libitum. The arterial catheter was connected to a pressure transducer (Nihon Kohden, Tokyo, Japan) to monitor the mean arterial blood pressure and heart rate continuously on a polygraph recorder (Power Lab, AD Instruments, Mountain View, CA). Normal saline was infused at the rate of 6 mL/kg/h via the jugular venous catheter in all rats. The NBRC 3972 strain of E coli (isolated from human feces)10 was obtained from the Biological Resource Center of the National Institute of Technology and Evaluation, Tokyo, Japan.
In the healthy study, after preparation surgery, rats were assigned to the following 3 subgroups: control (group healthy control [HC]), saline (group healthy saline [HS]), and lidocaine (group healthy lidocaine [HL]). While normal saline was infused via an epidural catheter at the rate of 30 μL/h to group HC and group HS, group HL was given a 0.5% lidocaine infusion at the rate of 0.15 mg/h (30 μL/h). Our pilot study showed that dye solution spread to the Th6 level cephalad and L2 level caudad at this infusion speed. E coli (5.0 × 105 colony-forming units) in 10 μL of phosphate-buffered saline (PBS) solution was injected to the pseudosurgical site to evoke SSI in group HS and group HL.
In the disease study, we examined the effects of epidural lidocaine infusion on the progression of SSI under the disease condition evoked by lipopolysaccharide (LPS; from E coli 0111:B4, Sigma-Aldrich, St. Louis, MO) injection to elicit a systemic inflammatory response. After the same preparation surgery as the healthy study, 0.25 mg/kg of LPS was injected IV to all rats, which were then assigned to the same 3 subgroups: control group (group disease control [DC]), saline group (group disease saline [DS]), and lidocaine group (group disease lidocaine [DL]). The procedures for inducing SSI and for epidural catheter infusion were the same as those in the healthy study.
While we did not perform precise power analysis before commencing the present study, we determined to prepare 10 to 15 animals in each group to detect the differences of lipocalin-2 upregulation with or without epidural anesthesia.
At 1 and 72 hours after the pseudosurgery to make skin incision, 20-μL samples of arterial blood were obtained for measurement of peripheral white blood cell (WBC) counts and hemoglobin levels. At 1, 6, 24, and 72 hours, 300-μL samples of arterial blood were obtained for measurement of plasma concentrations of lidocaine, cytokines including tumor necrosis factor-α, interferon-γ, interleukin-1β (IL-1β), IL-4, and IL-6, and lipocalin-2. The WBC counts and hemoglobin levels were determined by an analyzer (Celltac, Nihon Kohden). Plasma cytokine concentrations were measured in duplicate by enzyme-linked immunosorbent assays using commercially available antibodies (Procarta Immunoassay Kit, Affymetrix, Santa Clara, CA) in accordance with the instructions provided with the kit. Plasma lidocaine concentration was assessed by fluorescence polarization immunoassay (Oriental Yeast Corporation, Tokyo, Japan).
Analyses of the Tissue Expression of Lipocalin-2 mRNA and E coli DNA Using Real-Time Polymerase Chain Reaction
At 72 hours, tissue specimens from the pseudosurgical site (80 mg of skin) were obtained for measurement of mRNA expression of lipocalin-2 and E coli DNA in situ. Samples were homogenized by using a tissue homogenizer (Micro Smash-100R, Tomy, Tokyo, Japan) and centrifuged at 5000 rpm for 10 minutes at 4°C.
Tissue mRNA was extracted by using a commercially available mRNA extraction kit (RiboPure™ Kit, Life Technologies, Carlsbad, CA) in accordance with the manufacturer’s protocol. Thereafter, mRNA was reverse-transcribed using a TaqMan RNA Reverse Transcriptase kit (ReverTra Ace qPCR RT Kit, TOYOBO, Tokyo, Japan) in accordance with the manufacturer’s protocol. Lipocalin-2 mRNA was amplified by real-time polymerase chain reaction (RT-PCR) using a specific primer (Table 1). Quantification of PCR products was performed by the so-called TaqMan probe method. The TaqMan probe, which is arranged so as to attach to a specific site on the DNA strand of the target substance, has a reporter dye at the 5′ terminus and a quencher at the 3′ terminus, which suppresses the emission of fluorescence from a reporter dye. Once the TaqMan probe is cleaved and the reporter dye is separated from quencher during PCR amplification, the reporter dye emits fluorescence. The resultant increase in fluorescence is proportional to the quantity of the PCR products. mRNA expression level was compared by the 2−ΔΔCt method using β-actin as the standard internal control and the sample from group HC as the calibrator sample.
Samples (50 mg of skin) of the pseudosurgical site were obtained, and the total genomic DNA of the bacterial precipitate was extracted by the DNeasy Blood & Tissue kit (Cat. 69506, Qiagen, Germantown, MD) in accordance with the manufacturer’s instructions. The appropriateness and effectiveness of this kit was confirmed in a previous study.11 The extracted DNA was quantified by using a RT-PCR detection kit for E coli (Primerdesign, Southampton, UK) in accordance with the manufacturer’s instruction. This kit containing the primer and probe mixture was designed for quantification of E coli-sp genomes based on the TaqMan probe method. The E coli copy number per mg tissue was compared between the groups.
Measurement of Tissue Myeloperoxidase Activity
Samples (50 mg of skin) from the pseudosurgical site and nonsurgical site (1 cm left side of supine) were taken and homogenized in 50 mM PBS after perfusion to eliminate myeloperoxidase (MPO) activity in the blood. Simultaneously, a sample from the right lung (50 mg) was also obtained and homogenized in 50 mM PBS. Homogenized tissues, stored at −80°C to lyse the cells, were thawed and centrifuged at 10,000 rpm for 30 minutes at 4°C. The supernatant was used for assay of MPO activity using a commercially available kit (NWK–MPO03, Northwest Life Science Specialties, Vancouver, WA) in accordance with the manufacturer’s protocol.
Immunohistochemical Staining for Lipocalin-2
Tissue samples from the pseudosurgical site and nonsurgical site (1 cm left side of supine) were obtained for preparation of paraffin-embedded tissue sections. Thereafter, the sections were deparaffinized with xylene, and rehydrated through a graded series of ethanol to water. Endogenous peroxidase activity was inhibited by 30-minute incubation in methanol with 0.3% hydrogen peroxide. The samples were then boiled in 0.01 mol/L citrate buffer (pH 6.0) at 120°C for 10 minutes in a microwave oven. Nonspecific staining was blocked with 2% normal swine serum in PBS for 60 minutes at room temperature. The blots were incubated with primary antibody (mouse monoclonal anti-lipocalin-2 antibody, dilution 1:100, ABS 039-08, Abcam, Cambridge, UK) at 4°C overnight. After washing in PBS, they were incubated with biotinylated secondary antibodies (horse anti-mouse IgG antibody, dilution 1:200, BA-2000, Vector Laboratories, Eching, Germany) for 30 minutes, and treated with peroxidase-conjugated biotin-avidin complex (Vectastain universal ABC-Elite kit, Vector Laboratories) for 30 minutes, and stained with 3,3′-diaminobenzidine tetrahydrochloride (Wako, Tokyo, Japan) in 0.15% hydrogen peroxide for 10 minutes. Finally, the sections were counterstained with hematoxylin for immunohistochemical analyses.
All data are expressed as mean ± SD, unless otherwise specified. Comparisons of plasma WBC counts, hemoglobin levels, lipocalin-2 gene expression levels, E coli DNA copy numbers, and MPO activities among groups were performed by 1-way analysis of variance (ANOVA), followed by Tukey-Kramer post hoc test. Log-transformed plasma cytokine and lipocalin-2 concentrations were compared by means of a linear mixed-effects model for repeated measures with heteroscedastic covariance matrices among groups which contains group, time, and group-by-time interaction as fixed effects and heterogeneous autoregressive or compound symmetry. Tukey-Kramer adjustment was used for between-group comparisons within the model. Lilliefor test was used to check normality of the residuals of the ANOVA and linear mixed-effects model, and the Levene test based on the means was used to check the homogeneity of error variance of the ANOVA (all P > 0.17, all P > 0.20, respectively, in the ANOVA). If a response was not normally distributed in the mixed model analysis, the rank-transformed response was also analyzed to evaluate the sensitivity of the results to the deviation from the assumption.
A significance level was 2-tailed 5%. When the multiple-comparisons procedure was used, adjusted P value and corrected confidence intervals were shown. Analyses were performed using the SPSS/21.0J (SPSS Inc., Chicago, IL) and SAS version 9.3 (SAS Institute Inc., Cary, NC).
All animals in the healthy subgroups survived the 72-hour study period, while in the disease subgroups, 4 of 19 group DS animals and 3 of 18 group DL animals died. No significant changes in hemodynamic variables, such as the heart rate or mean arterial blood pressure, were observed within or between the disease and healthy subgroups throughout the study period (data not shown). In the disease subgroups (group DC, group DS, and group DL), WBC counts at 1 hour were reduced significantly compared with those in the healthy subgroups (group HC, group HS, and group HL, all P < 0.0001); however, the counts returned to normal by 72 hours. There were no significant changes in the hemoglobin level in any of the study subgroups (Table 2). Throughout the study period, the plasma lidocaine concentrations remained <0.8 μg/mL.
Plasma Inflammatory and Anti-inflammatory Cytokine Concentrations
Through the study periods, plasma concentrations of inflammatory cytokines, including tumor necrosis factor-α, interferon-γ, IL-1β, and IL-6, were significantly higher in the disease subgroups as compared with the healthy subgroups (all P < 0.0001, Fig. 1, A–D). However, there were no significant differences related to the presence/absence of epidural infusion of lidocaine (all P > 0.2515). On the other hand, no significant changes in the plasma concentration of IL-4, defined as one of the anti-inflammatory cytokines, were observed in either the healthy or disease subgroups at any time point during the study period (all P > 0.8743, Fig. 1E).
Changes of the Plasma Lipocalin-2 Concentrations and mRNA Expression Levels of Lipocalin-2 In Situ
In both the healthy and disease groups, alterations of the plasma lipocalin-2 concentrations during the 72-hour study period differed significantly among the subgroups. In all healthy subgroups, the plasma level of lipocalin-2 tended to increase slightly during the 72-hour study period. Although the value in group HS was significantly higher compared with group HC (P = 0.0019), the plasma concentration of lipocalin-2 was not markedly elevated in any of the subgroups during the study period (Fig. 2A). There were no statistically significant differences found between group HL and group HC or group HS and group HL (P = 0.0822 and P = 0.3116, respectively). In the disease subgroups, the plasma lipocalin-2 response behaved differently (Fig. 2B). The value was markedly higher in group DL, by approximately 3- and 6-fold compared with group DS and DC (all P < 0.0001).
Figure 3A depicts the relative mRNA expression levels of lipocalin-2 at the pseudosurgical site, normalized to the level in group HC at 72 hours. In the healthy subgroups, a significant and marked elevation of lipocalin-2 mRNA expression at the surgical site was found in both group HS and group HL ([4.69, 8.00], P = 0.0045, and [6.28, 12.10], P = 0.0031, respectively), whereas no significant difference was observed between group HS and group HL ([−19.7, 27.89], P = 0.9961). At the same time, in the disease subgroups, a distinct elevation of lipocalin-2 mRNA expression was found in the subgroups with SSI, that is, group DS and group DL compared with group DC ([12.45, 40.27], P = 0.0014, and [32.21, 92.13], P < 0.0001, respectively). Furthermore, in systemic disease conditions like endotoxemia, epidural infusion of lidocaine for 72 hours significantly augmented the expression of lipocalin-2 at the infected surgical site in comparison with saline infusion (group DL versus group DS: [15.32, 51.28], P < 0.0001).
Expression of E coli DNA and Immunohistochemical Analysis for Expression of Lipocalin-2 In Situ
Alterations of the copy number of E coli at the pseudosurgical site are illustrated in Figure 3B. In group HC, in which E coli was not applied to the pseudosurgical site, only a small copy number of E coli was detected (1.39 ± 2.34 copy number/mg tissue). While the copy number of E coli in group HS and group HL was significantly increased compared with group HC at 72 hours ([2.35, 5.66], P = 0.0282, and [1.99, 4.28], P = 0.0363, respectively), epidural infusion of lidocaine appeared to have no effects in the healthy subgroups ([−7.75, 12.33], P = 0.9080). In contrast, epidural infusion of lidocaine for 72 hours in the disease subgroups significantly suppressed the growth of E coli at the pseudosurgical site (group DL versus group DS: [12.21, 68.45], P = 0.0008).
Expression of lipocalin-2 was detected in the epidermis as shown in Figure 4 and root sheath (data not shown). Compared with the healthy subgroups, lipocalin-2 expression in the disease subgroups was apparently increased in both the epidermis and root sheath. However, this qualitative analysis suggests that epidural lidocaine infusion exerted no apparent effects on lipocalin-2 expression in the skin tissues.
Tissue MPO Activity in the Lung, Nonsurgical Sites, and Pseudosurgical Sites
Figure 5 shows tissue MPO activity in the lungs, and at nonsurgical and surgical sites. Tissue MPO activity was significantly higher in the disease subgroups than the corresponding healthy subgroups in all sampled tissues (group HC versus DC, group HS versus DS, and group HS versus DS, respectively, all P < 0.0091). In addition, the presence of SSI increased MPO activity in the infected tissues in both the healthy and disease subgroups, whereas no significant differences were found between group HS and group HL and between group DS and group DL, suggesting that epidural infusion of lidocaine did not modify the degree of accumulation of neutrophils at the infected surgical site or lungs. It is also noteworthy that MPO activity at the infected surgical site was lower than the lungs.
This study demonstrates that a neutrophil-derived protein, lipocalin-2, which has been reported to regulate bacterial growth in animal models of pneumonia and other infections,6,8,12 is significantly upregulated in both the blood and pseudosurgical site inoculated with E coli in the disease hosts accompanied by systemic inflammatory response, and that epidural anesthesia for 72 hours suppresses overgrowth of E coli in the pseudo-SSI field. Although we were unable to clarify any causal relationship between upregulation of lipocalin-2 and bacterial growth, the present study indicates a new profile of epidural anesthesia in the regulation of SSI.
In this study, we evaluated the effect of epidural anesthesia on the progression of SSI for 72 hours. Although SSI is considered to be instigated during surgery, its progression is unmasked only after a few days postoperatively. To maintain clinical relevance as much as possible, we used chronic placement of an epidural catheter for 72 hours in rats. Furthermore, we used an LPS injection model, in addition to a healthy model, to examine the expression of lipocalin-2 with and without epidural anesthesia. As shown in our animal model, systemic inflammatory responses, including a reduction of the WBC count and elevation of inflammatory cytokines, are obviously evoked after major surgery. To mimic the clinical situation of epidural anesthesia, we chose low-dose injection of LPS to induce a moderate, not lethal, level of inflammatory responses. A previous study indicated that 0.25 mg/kg IV LPS evoked a significant increase in blood neutrophil counts13; thereby, we chose the same dose in this study. Increased plasma inflammatory cytokine concentrations and marked reduction of the WBC count could reflect severity in this model as observed in patients who need epidural analgesia postoperatively and are at risk of SSI. Accordingly, our results demonstrated that epidural lidocaine infusion suppressed the growth of E coli only under the disease condition, indicating that epidural anesthesia could exert beneficial effects under disease conditions where systemic inflammatory response syndrome is evoked.
Previous clinical studies presumed that perioperative epidural anesthesia and analgesia were effective for reducing the risk of SSI.2,3 In the present study, we hypothesized that epidural anesthesia suppressed the growth of bacteria, and found that the expression of lipocalin-2 was definitely augmented in both blood and in situ at the site of infection. Although we were unable to clarify the precise mechanisms underlying how epidural lidocaine infusion induced the expression of lipocalin-2 under disease conditions with inflammatory responses, several pathways could be proposed.14–16 Sympathetic blockade by epidural anesthesia could improve tissue perfusion, leading to a greater accumulation of neutrophils around the infected surgical site and augmentation of the oxidative killing function of neutrophils, followed by increased production of lipocalin-2. We, thereby, examined the accumulation of neutrophils at the surgical site by both functional and histological approaches. However, epidural lidocaine infusion did not affect the MPO activity level in situ, and immunohistochemical analysis did not reveal any change in the distribution of lipocalin-2 around the surgical site (Figs. 4 and 5), even though lipocalin-2 expression evaluated by RT-PCR, a more sensitive and confirmatory approach, was significantly upregulated around the surgical field. The augmentation by epidural anesthesia of lipocalin-2 expression evaluated by RT-PCR in this study could have been caused by some other mechanism not directly related to the accumulated number of neutrophils, or the epidural lidocaine infusion by itself could have preserved lipocalin-2 activity. A previous study demonstrated that attenuation of inflammatory responses by both direct and indirect pathways through the antinociceptive effects of epidural anesthesia could preserve the microbicidal activity of neutrophils at a more potent level.17 In the present study, however, epidural lidocaine infusion had no effects on plasma cytokine concentrations and neutrophil mobilization in situ. Besides, epidural lidocaine infusion had antibacterial effects through the upregulation of lipocalin-2 in the disease group, but not in the healthy group. To account for this discrepancy, it is plausible to postulate that another component including local level of inflammatory cytokines evoked by LPS injection and/or SSI which, in turn, upregulated lipocalin-2 expression, could be modulated by epidural anesthesia. Further studies are warranted to elucidate the precise mechanisms by which epidural anesthesia might augment the expression of lipocalin-2 at surgical sites and reduce the incidence of SSIs as found in the present study.
There are several limitations to interpreting the study data. First, although we examined the effect of epidural anesthesia on the development of SSI by evaluating the copy number of E coli around the pseudosurgical site using RT-PCR, it remains uncertain whether the difference in the copy number of E coli found is clinically relevant. However, the risk of infection is related to the number of bacteria in situ,17 that is, the higher the number of bacteria, the higher the risk of SSI. Second, we did not evaluate the effect of epidural anesthesia beyond 72 hours after surgical preparation. Thus, it is still unclear whether epidural anesthesia can also provide longer-term benefits over 3 days. However, because antibacterial prophylaxis during the early phase of surgery is of great importance against the development of SSI,18,19 epidural anesthesia during the perioperative period is also likely to alter long-term outcome, and this should be examined in clinical trials. Third, the anti-inflammatory effects of lidocaine itself, shown in a previous study,20 might be associated with the improved secretion of lipocalin-2 by neutrophils. However, a previous study demonstrated that <100 mg/dL lidocaine hardly influenced the multiplication of E coli.21 Because plasma lidocaine concentrations were in the undetectable range in both epidural lidocaine groups (<0.8 mg/dL), epidural lidocaine administration, not systemic infusion, could be crucial for the beneficial effect of epidural anesthesia. Fourth, our immunohistochemical analysis did not reveal any obvious changes of lipocalin-2 expression around the epidermis or the root sheath, both of which probably play a major role against pathogenic organisms in the surgical field. This result may imply that epidural injection of lidocaine modulated lipocalin-2 expression at some other location, which remained unknown in this study. Furthermore, we did not evaluate the relationship between tissue lipocalin-2 level and bacterial colonization. Even though a previous in vitro study8 showed that 1 μM lipocalin-2 inhibited the growth of E coli, no study to date has shown dose-dependent effects of lipocalin-2 against pathogenic bacteria. Further studies are needed to elucidate the exact mechanisms using recombinant lipocalin-2 or lipocalin-2-deficit animals. Fifth, it may be possible that lidocaine spreading into the epidural space was not constant among the rats, affecting the results of this study. However, an epidural placement technique has been established in our laboratory, as reported previously.9 We used a modification of Freise et al.’s method22 for epidural cannulation using a dual-channel stainless swivel, which enabled the tip of the epidural catheter to be placed far from the surgical incision site. In our previous study, we confirmed that lidocaine spreading into the epidural space remained constant without any change of efficacy for >3 days. Finally, we expected epidural anesthesia to affect the expression of inflammatory and anti-inflammatory cytokines, but no significant differences were found with or without epidural anesthesia. In this study, we applied a nonlethal LPS rat model to mimic the excessive inflammatory responses in patients after abdominal surgery. We also estimated that a small “pseudosurgical site with inoculation of E coli” had little influence on the expression of these cytokines; thereby, it might have masked the anti-inflammatory effect of epidural anesthesia. Further studies are needed to clarify the relationship between lipocalin-2 expression and anti-inflammatory effect of epidural anesthesia.
In conclusion, epidural anesthesia suppressed the growth of E coli at surgical wound sites, possibly associated with the upregulation of lipocalin-2 expression, especially under pathological conditions to evoke systemic inflammatory responses. The present study shed light on a new aspect of epidural anesthesia, that is, its potential to induce a bacteriostatic protein. Further studies are warranted to clarify the causal relationship between epidural anesthesia, bacterial growth, and expression of lipocalin-2.
Name: Toru Igarashi, MD.
Contribution: This author designed and conducted the study, collected and analyzed the data, and prepared the manuscript.
Attestation: Toru Igarashi approved the final manuscript and attested the integrity of the original data and the analysis reported in this manuscript. Toru Igarashi is the archival author.
Name: Takeshi Suzuki, MD.
Contribution: This author helped conduct the study, analyzed the data, and prepared the manuscript.
Attestation: Takeshi Suzuki approved the final manuscript and attested to the integrity of the original data and the analysis reported in this manuscript.
Name: Katsuya Mori, PhD.
Contribution: This author helped design the study, and collected and analyzed the data.
Attestation: Katsuya Mori attested to the integrity of the original data and the analysis reported in this manuscript.
Name: Kei Inoue, MD.
Contribution: This author helped conduct the study and collected the data.
Attestation: Kei Inoue approved the final manuscript.
Name: Hiroyuki Seki, MD.
Contribution: This author helped conduct the study.
Attestation: Hiroyuki Seki approved the final manuscript.
Name: Takashige Yamada, MD.
Contribution: This author helped conduct the study.
Attestation: Takashige Yamada approved the final manuscript.
Name: Shizuko Kosugi, MD.
Contribution: This author helped conduct the study.
Attestation: Shizuko Kosugi approved the final manuscript.
Name: Shizuka Minamishima, MD.
Contribution: This author helped conduct the study, performed the statistical analysis, and prepared the manuscript.
Attestation: Shizuka Minamishima approved the final manuscript.
Name: Nobuyuki Katori, MD.
Contribution: This author helped conduct the study.
Attestation: Nobuyuki Katori approved the final manuscript.
Name: Fumiya Sano, MS.
Contribution: This author analyzed the data.
Attestation: Fumiya Sano approved the final manuscript.
Name: Takayuki Abe, PhD.
Contribution: This author analyzed the data, performed the statistical analysis, and prepared the manuscript.
Attestation: Takayuki Abe approved the final manuscript.
Name: Hiroshi Morisaki, MD.
Contribution: This author helped conduct the study and helped design the study. This author analyzed the data and prepared the manuscript.
Attestation: Hiroshi Morisaki approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.
This manuscript was handled by: Markus W. Hollmann, MD, PhD, DEAA.