External validation of the ventral hernia risk score for prediction of surgical site infections

BACKGROUND

Previously, we reported that the Ventral Hernia Risk Score (VHRS) was more accurate in a Veterans Affairs (VA) population in predicting surgical site infection (SSI) after open ventral hernia repair (VHR) compared with other models such as the Ventral Hernia Working Group (VHWG) model. The VHRS was developed using single-center data and stratifies SSI risk into five groups based on concomitant hernia repair, skin flaps created, American Society of Anesthesiologists (ASA) score ≥3, body mass index ≥40 kg/m(2), and incision class 4. The purpose of this study was to validate the VHRS for other hospitals.

METHODS

A prospective database of all open VHRs performed at three institutions from 2009-2011 was utilized. All 436 patients with a follow-up of at least 1 mo were included. The U.S. Centers for Disease Control and Prevention (CDC) definition of SSI was utilized. Each patient was assigned a VHRS, VHWG, and CDC incision classification. Receiver-operating characteristic curves were used to assess predictive accuracy, and the areas under the curve (AUCs) were compared for the three risk-stratification systems.

RESULTS

The median follow-up was 20 mos (range 1-49 mos). During this time, 111 patients (25.5%) developed a SSI. The AUC of the VHRS (0.73; 95% confidence interval [CI] 0.67-0.78) was greater than that of the VHWG (0.66; 95% CI 0.60-0.72; p<0.01) and the CDC incision class (0.68; 95% CI 0.61-0.74; p<0.05).

CONCLUSIONS

The VHRS provides a novel, internally and externally validated score for a patient's likelihood of developing a SSI after open VHR. Elevating skin flaps, ASA score ≥3, concomitant procedures, morbid obesity, and incision class all independently predicted SSI. It remains to be determined if pre-operative patient selection and risk reduction, surgical techniques, and post-operative management can improve outcomes in the highest-risk patients. The VHRS provides a starting point for key stakeholders to discuss the management of ventral hernias.

Quality-of-life scores in laparoscopic preperitoneal inguinal hernia repair

BACKGROUND

Published support exists for using lightweight polypropylene mesh (PPM) to repair inguinal hernias with increased biocompatibility and decreased foreign body reaction and pain. However, quality of life (QOL) has not been assessed. We assess QOL in patients undergoing laparoscopic totally extraperitoneal hernia repair (TEP) with lightweight PPM.

METHODS

We performed an IRB-approved study of patients undergoing TEP hernia repair. Demographic information and hernia characteristics were collected perioperatively. Baseline Short Form-36 (SF-36), Carolinas Comfort Scale (CCS), and visual analog scale (VAS) for pain were performed preoperatively, and then after 1, 26, and 52 weeks.

RESULTS

Forty-eight patients undergoing TEP with mesh were selected. Average age was 43.2 years (SD = 13.2), and average BMI was 26.1 kg/m(2) (SD = 4.3). Procedures include bilateral hernia, right inguinal hernia, and left inguinal hernia repairs. Mean scores on the CCS(®) and VAS were low during the immediate post-op period and 1 year. SF-36 mean scores for body pain, physical function, and role physical showed decreases at the postoperative survey and then subsequent increases. Pain-associated scores increased during the immediate post-op period. CCS and SF-36 scores demonstrated improvement after 1 year. There was no significant difference in VAS. Bilateral repair patients reported more pain and reduced physical function versus unilateral repairs. Patients with larger mesh reported greater pain scores and reduced physical function scores.

CONCLUSIONS

Laparoscopic inguinal hernia repair is associated with initial declines in QOL in the postoperative period. Improvements appear in the long term. General health does not appear to be impacted by laparoscopic TEP. Smaller mesh and unilateral repairs are associated with improved QOL following laparoscopic TEP with PPM. Multiple metrics for QOL are required to reflect patient recovery.

Component Separation vs. Bridged Repair for Large Ventral Hernias: A Multi-Institutional Risk-Adjusted Comparison, Systematic Review, and Meta-Analysis

BACKGROUND

Repair of large ventral hernia defects is associated with high rates of surgical site occurrences (SSO), including surgical site infection (SSI), site dehiscence, seroma, hematoma, and site necrosis. Two common operative strategies exist: Component separation (CS) with primary fascial closure and mesh reinforcement (PFC-CS) and bridged repair (mesh spanning the hernia defect). We hypothesized that: (1) ventral hernia repair (VHR) of large defects with bridged repair is associated with more SSOs than is PFC, and (2) anterior CS is associated with more SSOs than is endoscopic, perforator-sparing, or posterior CS.

METHODS

Part I of this study was a review of a multi-center database of patients who underwent VHR of a defect ≥8 cm from 2010-2011 with at least one month of follow-up. The primary outcome was SSO. The secondary outcome was recurrence. Part II of this study was a systematic review and meta-analysis of studies comparing bridged repair with PFC and studies comparing different kinds of CS.

RESULTS

A total of 108 patients were followed for a median of 16 months (range 1-50 months), of whom 84 underwent PFC-CS and 24 had bridged repairs. Unadjusted results demonstrated no differences between the groups in SSO or recurrence; however, the study was underpowered for this purpose. On meta-analysis, PFC was associated with a lower risk of SSO (odds ratio [OR] = 0.569; 95% confidence interval [CI] = 0.34-0.94) and recurrence (OR = 0.138; 95% CI = 0.08-0.23) compared with bridged repair. On multiple-treatments meta-analysis, both endoscopic and perforator-sparing CS were most likely to be the treatments with the lowest risk of SSO and recurrence.

CONCLUSIONS

Bridged repair was associated with more SSOs than was PFC, and PFC should be used whenever feasible. Endoscopic and perforator-sparing CS were associated with the fewest complications; however, these conclusions are limited by heterogeneity between studies and poor methodological quality. These results should be used to guide future trials, which should compare the risks and benefits of each CS method to determine in which setting each technique will give the best results.

Suture, synthetic, or biologic in contaminated ventral hernia repair

BACKGROUND

Data are lacking to support the choice between suture, synthetic mesh, or biologic matrix in contaminated ventral hernia repair (VHR). We hypothesize that in contaminated VHR, suture repair is associated with the lowest rate of surgical site infection (SSI).

METHODS

A multicenter database of all open VHR performed at from 2010-2011 was reviewed. All patients with follow-up of 1 mo and longer were included. The primary outcome was SSI as defined by the Centers for Disease Control and Prevention. The secondary outcome was hernia recurrence (assessed clinically or radiographically). Multivariate analysis (stepwise regression for SSI and Cox proportional hazard model for recurrence) was performed.

RESULTS

A total of 761 VHR were reviewed for a median (range) follow-up of 15 (1-50) mo: there were 291(38%) suture, 303 (40%) low-density and/or mid-density synthetic mesh, and 167(22%) biologic matrix repair. On univariate analysis, there were differences in the three groups including ethnicity, ASA, body mass index, institution, diabetes, primary versus incisional hernia, wound class, hernia size, prior VHR, fascial release, skin flaps, and acute repair. The unadjusted outcomes for SSI (15.1%; 17.8%; 21.0%; P = 0.280) and recurrence (17.8%; 13.5%; 21.5%; P = 0.074) were not statistically different between groups. On multivariate analysis, biologic matrix was associated with a nonsignificant reduction in both SSI and recurrences, whereas synthetic mesh associated with fewer recurrences compared to suture (hazard ratio = 0.60; P = 0.015) and nonsignificant increase in SSI.

CONCLUSIONS

Interval estimates favored biologic matrix repair in contaminated VHR; however, these results were not statistically significant. In the absence of higher level evidence, surgeons should carefully balance risk, cost, and benefits in managing contaminated ventral hernia repair.

Doxycycline shows dose-dependent changes in hernia repair strength after mesh repair

BACKGROUND

Ventral hernia is a commonly occurring surgical problem. Our earlier studies have shown that a 30 mg/kg dose of doxycycline can significantly impact the strength of polypropylene (PP) mesh in a rat hernia repair model at 6 and 12 weeks. The objective of the present study was to investigate the dose dependence of doxycycline treatment on hernia repair strengths in rats.

STUDY DESIGN

Fifty-six Sprague-Dawley rats underwent hernia repair with either PP mesh (n = 28) or sutures only (primary; n = 28); both groups were further divided into four doxycycline groups of seven animals each: control (0 mg/kg), low (3 mg/kg), medium (10 mg/kg), and high (30 mg/kg). One day before hernia repair surgery, animals received doxycycline doses by gavage and continued receiving daily until euthanasia. After 8 weeks, rats were euthanized and tissue samples from hernia repaired area were collected and analyzed for tensile strength using a tensiometer (Instron, Canton, MA, USA), while MMPs 2, 3, and 9, and collagen type 1 and 3 were analyzed by western blotting.

RESULTS

In mesh-repaired animals, medium and high doxycycline dose repaired mesh fascia interface (MFI) showed significant increase in tensile strength when compared to control. In the primary repaired animals, there was no significant difference in MFI tensile strength in any dose group. In medium-dose MFI, there was a significant reduction in MMPs 2, 3, and 9. In this animal group, MFI showed significant increase in collagen 1 and significant reduction in collagen type 3 when compared to control.

CONCLUSION

It is possible to improve the strength of mesh-repaired tissue by administering a significantly lower dose of the drug, which has implications for translation of the findings.

Development and validation of a risk stratification score for ventral incisional hernia after abdominal surgery: hernia expectation rates in intra-abdominal surgery (the HERNIA Project)

BACKGROUND

Ventral incisional hernias (VIH) develop in up to 20% of patients after abdominal surgery. No widely applicable preoperative risk-assessment tool exists. We aimed to develop and validate a risk-assessment tool to predict VIH after abdominal surgery.

STUDY DESIGN

A prospective study of all patients undergoing abdominal surgery was conducted at a single institution from 2008 to 2010. Variables were defined in accordance with the National Surgical Quality Improvement Project, and VIH was determined through clinical and radiographic evaluation. A multivariate Cox proportional hazard model was built from a development cohort (2008 to 2009) to identify predictors of VIH. The HERNIAscore was created by converting the hazards ratios (HR) to points. The predictive accuracy was assessed on the validation cohort (2010) using a receiver operator characteristic curve and calculating the area under the curve (AUC).

RESULTS

Of 625 patients followed for a median of 41 months (range 0.3 to 64 months), 93 (13.9%) developed a VIH. The training cohort (n = 428, VIH = 70, 16.4%) identified 4 independent predictors: laparotomy (HR 4.77, 95% CI 2.61 to 8.70) or hand-assisted laparoscopy (HAL, HR 4.00, 95% CI 2.08 to 7.70), COPD (HR 2.35; 95% CI 1.44 to 3.83), and BMI ≥ 25 kg/m(2) (HR1.74; 95% CI 1.04 to 2.91). Factors that were not predictive included age, sex, American Society of Anesthesiologists (ASA) score, albumin, immunosuppression, previous surgery, and suture material or technique. The predictive score had an AUC = 0.77 (95% CI 0.68 to 0.86) using the validation cohort (n = 197, VIH = 23, 11.6%). Using the HERNIAscore: HERNIAscore = 4(∗)Laparotomy+3(∗)HAL+1(∗)COPD+1(∗) BMI ≥ 25, 3 classes stratified the risk of VIH: class I (0 to 3 points),5.2%; class II (4 to 5 points),19.6%; and class III (6 points), 55.0%.

CONCLUSIONS

The HERNIAscore accurately identifies patients at increased risk for VIH. Although external validation is needed, this provides a starting point to counsel patients and guide clinical decisions. Increasing the use of laparoscopy, weight-loss programs, community smoking prevention programs, and incisional reinforcement may help reduce rates of VIH.

Laparoscopic Inguinal Hernia Repair With a Novel Hernia Mesh Incorporating a Nitinol Alloy Frame Compared With a Standard Lightweight Polypropylene Mesh

BACKGROUND

Numerous mesh materials are available for laparoscopic inguinal hernia repair. The role of fixation of mesh in laparoscopic inguinal hernia repair remains controversial. Mesh materials have been engineered to anatomically conform to the pelvis to potentially reduce or eliminate the need for fixation. This study evaluates the outcomes of laparoscopic inguinal hernia utilizing a device consisting of a lightweight polypropylene mesh with a nitinol frame (Rebound HRD) compared with repair with lightweight polypropylene mesh with permanent tack fixation.

METHODS

A prospective randomized trial evaluating the outcomes of laparoscopic inguinal hernia repair with a lightweight polypropylene mesh with a nitinol frame (N-LWM) compared with standard lightweight polypropylene mesh (LWM) was conducted. Randomization was performed at an N-LWM to LWM ratio of 2:1. Repairs were standardized to a laparoscopic extraperitoneal approach without fixation for N-LWM and titanium tack fixation for LWM repairs. Follow-up assessments were performed at 7 days, 6 months, and 1 year. Outcome measures include visual analog pain scale (VAS), Short Form 36 (SF-36), Carolinas Comfort Scale (CCS), operative details, complications, and recurrences.

RESULTS

There were 47 patients that underwent laparoscopic inguinal hernia repair and adhered to study protocol (31 N-LWM, 16 LWM). The groups did not differ significantly in age, body mass index, ethnicity, or employment. The N-LWM group had bilateral mesh placed in 51.6% and LWM 43.8% (P = .76). Operative duration was similar, 59.6 ± 23.1 minutes for LWM and 62.4 ± 26.7 minutes for N-LWM (P = .705) as was mesh handling time was 5.4 ± 3.1 minutes LWM versus 7.3 ± 3.9 minutes N-LWM (P = .053). VAS, CCS, and SF-36 survey results were similar between groups. There was one recurrence (0.03%) in the N-LWM group.

CONCLUSIONS

Nitinol-framed lightweight polypropylene mesh may be safely used during laparoscopic inguinal hernia repair with outcomes comparable to LWM at 1 year. N-LWM does not impact operating room time, mesh handling time, pain, recurrences, or complications.

Patient Symptoms Correlate Poorly with Objective Measures among Patients with Gastroesophageal Reflux Disease

Presentation of gastroesophageal reflux disease (GERD) varies among patients. To attempt to understand the patient's perception of the severity of their reflux symptoms, we developed a questionnaire on which patients rated symptom severity at each office visit. After receiving Institutional Review Board approval, we retrospectively reviewed patient charts of all patients seen by one surgeon for GERD symptoms and/or presence of hiatal hernia (HH) from September 2012 to April 2013. Data from patient questionnaires combined with objective findings from subsequent or prior workup and eventual operative information were recorded. A total of 144 questionnaires were reviewed from 108 patients. Frequencies were calculated for categorical variables. Patients were divided into four categories based on size of the HH on the endoscopic report; 10 patients had no HH, 15 had small HH, 20 had medium HH, and 31 patients had large HH. Size of HH was not available for three patients. Pre- and postoperative questionnaire responses were obtained for 15 patients. A combined reflux score was calculated using the median for each symptom. Patient perception of severity of symptoms does not necessarily predict presence of pathological reflux or HH nor is there a perfect combination of symptoms to predict the presence of pathological reflux or HH based on our sample. The workup of this pathology must be comprehensive, and the confirmation of reflux is imperative when the diagnosis is unclear.

Patient symptoms correlate poorly with objective measures among patients with gastroesophageal reflux disease

Presentation of gastroesophageal reflux disease (GERD) varies among patients. To attempt to understand the patient's perception of the severity of their reflux symptoms, we developed a questionnaire on which patients rated symptom severity at each office visit. After receiving Institutional Review Board approval, we retrospectively reviewed patient charts of all patients seen by one surgeon for GERD symptoms and/or presence of hiatal hernia (HH) from September 2012 to April 2013. Data from patient questionnaires combined with objective findings from subsequent or prior workup and eventual operative information were recorded. A total of 144 questionnaires were reviewed from 108 patients. Frequencies were calculated for categorical variables. Patients were divided into four categories based on size of the HH on the endoscopic report; 10 patients had no HH, 15 had small HH, 20 had medium HH, and 31 patients had large HH. Size of HH was not available for three patients. Pre- and postoperative questionnaire responses were obtained for 15 patients. A combined reflux score was calculated using the median for each symptom. Patient perception of severity of symptoms does not necessarily predict presence of pathological reflux or HH nor is there a perfect combination of symptoms to predict the presence of pathological reflux or HH based on our sample. The workup of this pathology must be comprehensive, and the confirmation of reflux is imperative when the diagnosis is unclear.

Complex ventral hernia repair with a human acellular dermal matrix

Purpose

The ideal approach to complex ventral hernia repair is frequently debated. Differences in processing techniques among biologic materials may impact hernia repair outcomes. This study evaluates the outcomes of hernia repair with a terminally sterilized human acellular dermal matrix (TS-HADM) (AlloMax^® Surgical Graft, by C. R. Bard/Davol, Inc., Warwick, RI, USA) treated with low-dose gamma irradiation.

Methods

A single-arm multi-center retrospective observational study of patients undergoing hernia repair with TS-HADM was performed. Data analyses were exploratory only; no formal hypothesis testing was pre-specified.

Results

Seventy-eight patients (43F, 35M) underwent incisional hernia repair with a TS-HADM. Mean follow-up was 20.5 months. Preoperative characteristics include age of 56.6 ± 11.1 years, BMI 36.7 ± 9.9 kg/m², and mean hernia defect size 187 cm². Sixty-five patients underwent component separation technique (CST) with a reinforcing graft. Overall, 21.8 % developed recurrences. Recurrences occurred in 15 % of patients repaired with CST. Major wound complications occurred in 31 % of patients overall. Based upon CDC surgical wound classification, major wound complications were seen in 26, 40, 56, and 50 % of Class 1, 2, 3, and 4 wounds, respectively. No grafts required removal.

Conclusions

Hernia recurrences are not uncommon following complex abdominal wall reconstruction. Improved outcomes are seen when a TS-HADM is utilized as reinforcement to primary fascial closure.

Economics of abdominal wall reconstruction

The economic aspects of abdominal wall reconstruction are frequently overlooked, although understandings of the financial implications are essential in providing cost-efficient health care. Ventral hernia repairs are frequently performed surgical procedures with significant economic ramifications for employers, insurers, providers, and patients because of the volume of procedures, complication rates, the significant rate of recurrence, and escalating costs. Because biological mesh materials add significant expense to the costs of treating complex abdominal wall hernias, the role of such costly materials needs to be better defined to ensure the most cost-efficient and effective treatments for ventral abdominal wall hernias.

Evaluation of ECHO PS Positioning System in a Porcine Model of Simulated Laparoscopic Ventral Hernia Repair

Purpose. Operative efficiency improvements for laparoscopic ventral hernia repair (LVHR) have focused on reducing operative time while maintaining overall repair efficacy. Our objective was to evaluate procedure time and positioning accuracy of an inflatable mesh positioning device (Echo PS Positioning System), as compared to a standard transfascial suture technique, using a porcine model of simulated LVHR. Methods. The study population consisted of seventeen general surgeons (n = 17) that performed simulated LVHR on seventeen (n = 17) female Yorkshire pigs using two implantation techniques: (1) Ventralight ST Mesh + Echo PS Positioning System (Echo PS) and (2) Ventralight ST Mesh + transfascial sutures (TSs). Procedure time and mesh centering accuracy overtop of a simulated surgical defect were evaluated. Results. Echo PS demonstrated a 38.9% reduction in the overall procedure time, as compared to TS. During mesh preparation and positioning, Echo PS demonstrated a 60.5% reduction in procedure time (P < 0.0001). Although a trend toward improved centering accuracy was observed for Echo PS (16.2%), this was not significantly different than TS. Conclusions. Echo PS demonstrated a significant reduction in overall simulated LVHR procedure time, particularly during mesh preparation/positioning. These operative time savings may translate into reduced operating room costs and improved surgeon/operating room efficiency.

Predictors of poor outcomes in functionally dependent patients undergoing ventral hernia repair

BACKGROUND

Preoperative functional health status has been shown previously to affect outcomes following ventral hernia repair. The objective of this study was to identify specific factors that contribute to morbidity and mortality in functionally dependent patients who undergo elective ventral hernia repair.

METHODS

We reviewed all patients in the ACS NSQIP database who underwent elective ventral hernia repair from 2005 to 2009. Patients were selected based on the following CPT codes: 49560, 49561, 49565, 49566, 49568, 49570, 49572, 49585, 49587, 49652, 49653, 49654, 49655, 49656, and 49657. Only patients classified as partially or totally dependent were included in this study. Thirty-day outcomes included mortality, wound occurrences, pulmonary occurrences, venous thromboembolism, development of sepsis/shock, renal failure/insufficiency, and cardiovascular events. We analyzed risk factors using multivariate analyses.

RESULTS

We identified 75,865 patients who underwent elective ventral hernia repair, of which 1,144 were classified as functionally dependent. Overall, major morbidity was observed in 211 (18.4 %) patients. There was no statistically significant difference in any measurable outcomes between laparoscopic and open hernia repairs. Increasing age proved to be an independent predictor of mortality, with an odds ratio of 1.63 (95 % confidence interval (CI) 1.27-2.12) for each 10 years of age beyond the mean. Ascites and preoperative renal failure also were identified as independent predictors of mortality, with odds ratios of 9.7 and 11.5, respectively. Preoperative pulmonary compromise was shown to be an independent predictor of both mortality and major morbidity, with odds ratios of 4.1 and 2.47, respectively.

CONCLUSIONS

Elective ventral hernia repair in the functionally dependent patient population has significant morbidity and mortality. Increasing age, ascites, preoperative renal failure, and preoperative pulmonary compromise are independent predictors of 30-day mortality. Nonoperative management should be strongly considered in the presence of these risk factors. Further studies are needed to determine optimum management strategies for functionally dependent patients with ventral hernias.

Outcomes of simultaneous laparoscopic cholecystectomy and ventral hernia repair compared to that of laparoscopic cholecystectomy alone

BACKGROUND

Although incidental hernias frequently are found and repaired during laparoscopic cholecystectomy (LC), the outcomes of simultaneous LC and laparoscopic ventral hernia repair (LVHR) have not been scrutinized. In this study we evaluated short-term outcome data comparing simultaneous LC and LVHR against LC alone.

METHODS

The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database (2005-2009) was queried using primary procedure and secondary current procedural terminology (CPT(®)) codes for LC and LVHR. Outcomes analyzed included separate LC and LVHR and simultaneous laparoscopic cholecystectomy and ventral hernia repair (LC/LVHR). The 30 day clinical outcomes along with postoperative hospital length of stay (LOS) were assessed using the χ(2) test and analysis-of-variance test with p values < 0.01 set as significant. We also performed forward stepwise multivariable regression taking in to consideration over 50 ACS NSQIP risk factors to adjust for patient risk.

RESULTS

A total of 82,837 patients underwent LC and/or LVHR of which 357 (0.4%) underwent simultaneous LC/LVHR. Patients who underwent LC/LVHR were more likely to have surgical site infections, suffer sepsis or septic shock, and have pulmonary complications, including pneumonia, reintubation or prolonged ventilator requirements, than LC-alone patients. No difference was noted in 30 day mortality, rates of deep vein thrombosis/pulmonary embolism (DVT/PE), renal insufficiency, or stroke. After multivariable adjustment for over 50 ACS NSQIP risk factors, concurrent LC/LVHR continued to pose a higher risk for these outcomes relative to LC only.

CONCLUSIONS

Simultaneous LC/LVHR results in greater postoperative morbidity in terms of surgical site infections, sepsis, and pulmonary complications when compared to LC alone. In light of this increased short-term morbidity, consideration should be given toward performing LC and LVHR independently in patients requiring both procedures. Prospective studies with long-term follow-up are required to better understand the implications of simultaneous LC/LVHR.

Preoperative Functional Health Status Impacts Outcomes after Ventral Hernia Repair

Associated with the aging population is an increase in comorbidities and a decrease in the ability to perform basic daily activities. This is tracked within the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) as a patient's preoperative functional health status. Our goal was to evaluate the impact of preoperative functional status upon outcomes after ventral hernia repair. We reviewed all cases of patients that underwent ventral hernia repair from 2005 to 2010 in the ACS-NSQIP database. Patients were identified based on selected Current Procedural Terminology codes and grouped based on functional status as listed in the ACS-NSQIP database—independent, partially dependent, and totally dependent. Preoperative and operative variables were recorded for all patients. Clinical risk factors and short-term outcomes between groups were compared. Multivariable logistic regression was used to adjust for age, wound class, American Society of Anesthesiologists class, and case relative value units. A total of 76,397 patients were identified: 74,785 were independent (97.9%), 1,317 partially dependent (1.7%), and 295 totally dependent (0.4%). Totally dependent patients had an increased risk for all short-term outcomes after ventral hernia repair: wound occurrence, pneumonia, pulmonary embolism, urinary tract infection, myocardial infarction, deep venous thrombosis, sepsis, return to the operating room, and death ( P < 0.001 for all).