| |
Interferon-gamma May Reverse Cirrhosis
|
| |
| |
"Effect of Interferon-Gamma on Hepatic Fibrosis in Chronic Hepatitis B Virus Infection: A Randomized Controlled Study"
August 2005.
Clinical Gastroenterology and Hepatology
For patients unable to achieve a Sustained Viral Response using peginterferon plus ribavirin for hepatitis C therapy, the use of an antifibrotic therapy may reverse fibrosis or cirrhosis. Here is a preliminary study finding that interferon-gamma improved the fibrosis score for patients.
".....In summary, IFN-_ has significant beneficial effects on liver fibrosis in patients with chronic HBV infection. Moreover, our data indicate that IFN-_ antagonizes the TGF-_/Smad pathway, thereby providing a molecular explanation for its antifibrotic effects....".
"........Of 54 patients in the IFN-_ group, 34 (63%) exhibited significant improvement of semiquantitative fibrosis scoring (15 patients improved by 5 or more units and 19 patients by 2 or more units). In contrast, only 7 (24.1%) of 29 patients in the control group exhibited improvement in histology (4 patients improved ≥5 units and 3 patients improved ≥2 units). The difference between both groups was highly significant...."
Authors: Hong-Lei Weng_, Bao-En Wang, Ji-Dong Jia, Wan-Fen Wu, Jian-Zhong Xian, Peter R. Mertens_, Wei-Min Cai_ Steven Dooley
_ Institute of Infectious Diseases, First Affiliated Hospital, Medical School, Zhejiang University, Hangzhou, China
Liver Research Center, Beijing Friendship Hospital, Capital University of Medical Science, Beijing, China
Eighth Hospital of Guangzhou, Guangzhou, China
_ Department of Nephrology and Immunology, University Hospital, RWTH-Aachen, Aachen, Germany
Department of Medicine II, University Hospital of Mannheim, University of Heidelberg, Mannheim, Germany
ABSTRACT
Background & Aims: Hepatic fibrosis due to chronic HBV infection has enormous socioeconomic impact. Besides strategies targeting virus elimination, prevention or reversal of liver fibrosis is amenable. Given the antifibrotic activity of interferon-gamma (IFN-_), a randomized open-labeled multicenter trial was initiated to test IFN-_ in HBV infection.
Methods: HBsAg-positive patients with biopsy proven hepatic fibrosis (n = 99, stages 24, Scheuer criterion) were treated with diammone-glycyrrhizinate and potassium-magnesium aspartate. Sixty-six randomly assigned patients were treated with 50 _g IFN-_ intramuscularly on a daily basis for 3 months and on alternate days the subsequent 6 months. Efficacy was evaluated by liver biopsy and serologic markers.
Results:
Fifty-four patients in the IFN-_ group and 29 patients in the control group completed the study. The hepatic fibrosis score was significantly reduced in 63% of IFN-_ treated patients compared with 24.1% in the control group by using a semiquantitative scoring system evaluating both liver architecture and fibrotic deposits.
Mean values for the total fibrosis score decreased from 13.8 ± 5.8 to 10.1 ± 5.1 in the IFN-_ group (P = .0001), whereas they were unchanged in control subjects (13.2 ± 6.8 vs 12.6 ± 4.8, P = .937). The Scheuer system showed 12 out of 54 patients improved ≥1 stage(s) in the IFN-_ group compared with 1 of 29 in the control group.
Antifibrotic activity might be attributed to decreased transforming growth factor-beta signaling via phosphorylated Smad2 and reduced number of activated, _-smooth muscle actin positive hepatic stellate cells.
Conclusions: IFN-_ treatment for 9 months improves fibrosis scores in patients with chronic HBV infection most likely by antagonizing profibrogenic transforming growth factor-beta effects.
BACKGROUND
Chronic injuries to liver including viral hepatitides B and C, alcohol, metabolic or autoimmune diseases, and congenital abnormalities might lead to loss of tissue architecture and hepatic fibrosis.1 Worldwide, more than 350 million people are chronically infected with HBV, with deleterious sequelae such as cirrhosis and hepatocelluar carcinoma accounting for approximately 1 million deaths per year.2 With the advent of virostatic therapy, the primary goal of treatment is virus elimination.2,3 Even with improvement of treatment options, ie, development of long-lasting pegylated interferons and novel nucleoside analogues like lamivudine, elimination of HBV is only achieved in about 50% of patients. Treatment effectiveness largely depends on mutations within the HBV genome, with the poorest response rates for genotype 1c.2,3 Besides strategies to target virus replication, treatment approaches directing fibrosis development and diminution of inflammation are needed.4,5
Mechanisms of fibrogenesis encompass recruitment of inflammatory cells, cytokine-directed activation of hepatic stellate cells (HSCs) with acquisition of an _-smooth muscle actin (_-SMA) positive prosclerotic phenotype and subsequent accumulation of interstitial collagens.1 Both HSC transdifferentiation and apoptosis-dependent destruction of liver architecture are key steps for initiation and progression of fibrogenesis. Activated HSCs are the main source of extracellular matrix (ECM) components and release excess amounts of transforming growth factor-beta (TGF-_), initiating de novo activation of quiescent HSCs via paracrine effects and maintaining the transdifferentiated phenotype of myofibroblasts (MFBs) by an autocrine route. Furthermore, TGF-_ action is involved in hepatocyte apoptosis, indicating that "chronic" TGF-_ action leads to progressive disease, and anti-TGF-_ strategies were successful in experimental models.4,5
IFN-_, physiologically produced by natural killer cells and T lymphocytes, has pleiotropic functions, eg, modulates the immune response and tumor defense. In contrast to IFN-_, no virostatic activities of IFN-_ are reported in most clinical trials.6 However, IFN-_ was identified as inhibitor of ECM synthesis by antagonizing TGF-_.7 In chronic granulomatous disease and idiopathic pulmonary fibrosis, IFN-_ application was successfully implemented with few side effects.8,9 Given these strong indications for antifibrotic in vivo activity, we designed an open-labeled, randomized multicenter trial for patients with chronic HBV who received IFN-_ as an add-on therapy to diammone-glycyrrhizinate and potassium-magnesium aspartate.
Results
Patient Characteristics and Effects of Interferon-gamma
Six hundred seventy-eight patients were screened for eligibility. Three hundred fifteen patients fulfilled inclusion criteria, 99 for the biopsied group and 216 for the non-biopsied group, and were prospectively and randomly assigned to IFN-_ as add-on treatment at a ratio of 2:1 and 1:1, respectively. In the biopsied group, 54 patients in the IFN-_ group and 29 in the control group completed the study protocol. Dropouts mainly occurred as a result of noncompliance with a second liver biopsy (15 patients, 11 in the IFN-_ group and 4 in the control group) and fever (1 patient in the IFN-_ group). Baseline clinical characteristics of both groups (biopsied and non-biopsied patients) revealed no significant differences for age, sex, serum ALT and AST levels, serum albumin and bilirubin levels, HBV serology, and histologic changes between the 2 groups.. A second biopsy was performed after 9 months. Of 54 patients in the IFN-_ group, 34 (63%) exhibited significant improvement of semiquantitative fibrosis scoring (15 patients improved ≥5 units and 19 patients ≥2 units). In contrast, only 7 (24.1%) of 29 patients in the control group exhibited improvement in histology (4 patients improved ≥5 units and 3 patients improved ≥2 units). The difference between both groups was highly significant (P < .001).
Detailed Histologic Analysis
A commonly used assessment system of liver histology is the Scheuer scoring system. In this system, a significantly decreased inflammatory grade (mean of 2.29 after 9 months vs 3.13 at baseline, P < .001) and reduced fibrosis stage (2.96 after 9 months vs 3.26 at baseline, P < .001) were observed after IFN-_ treatment. In contrast, the mean inflammatory grade in the control group did not change (3.24 at baseline as well as after 9 months). The fibrosis stage increased from 3.00 to 3.28 (P < .05) after 9 months in the control group.
In the IFN-_ group, 12 patients (22.2%) improved at least 1 stage, 2 patients improved 2 stages, and 1 patient improved 3 stages. In 42 patients (77.8%), the fibrosis stage did not change, and no patient exhibited a worsened score after treatment, whereas in the control group, 1 patient (3.4%) improved 2 stages after 9 months, 21 patients (72.4%) exhibited no changes, and 6 patients (20.7%) were worse.
One drawback of the Scheuer scoring system is that it does not reflect major changes occurring within 1 grade/stage. Therefore, a semiquantitative scoring system combining the previous descriptions of Chevallier et al12 and Knodell et al11 was applied. This system is designed to evaluate therapeutic efficacy, assessing 4 major sites for fibrotic deposits (portal tract, lobular and perisinusoidal space, together with width and number of septa) and 4 main sites of inflammation (portal tract, lobular inflammation together with piecemeal necrosis and bridging necrosis). With this system, inflammation and fibrosis scores decreased significantly in the IFN-_ group (P < .001), from 14.4 ± 5.2 to 7.6 ± 3.6 and 13.8 ± 5.8 to 10.1 ± 5.1. In the control group, both scores did not show remarkable changes (inflammation: 14.6 ± 5.4 to 14.8 ± 5.3; fibrosis: 13.2 ± 6.8 to 12.6 ± 4.8; P > .05). Inflammatory and fibrosis scores of individual patients are depicted in Figure 3. Most intriguingly, all patients in the IFN-_ group had an improved or unchanged histology. Fifty patients (92.6%) had decreased inflammatory scores, and 38 patients (70.4%) had decreased fibrosis scores. There were no increased scores after 9 months of IFN-_ treatment. In the control group, inflammatory scores decreased in 12 patients (41.4%), fibrosis scores decreased in 9 patients (31.0%), whereas an increase was found in 13 patients (44.8%) and 15 patients (51.7%), respectively.
Subsequently, a more detailed histologic analysis was performed, including portal and lobular inflammation, piecemeal necrosis, bridging necrosis, portal and lobular fibrosis, as well as width and number of septa. In the IFN-_ group, all indices decreased significantly after 9 months (P < .01), whereas in the control group, there were no remarkable differences.
Effect of Interferon-gamma on Compensated Cirrhotic Patients
Thirty-five patients with compensated cirrhosis (stage 4, Scheuer criterion; 26 in the IFN-_ group and 9 in the control group) were included in a subanalysis. Fibrosis scores decreased significantly in the IFN-_ group (17.5 ± 4.3 to 14.0 ± 3.3; P < .001), whereas it increased in the control group (13.3 ± 4.9 to 18.6 ± 8.5; P < .05). Five patients in the compensated cirrhosis group exhibited a significant improvement after IFN-_ treatment with narrowed fibrotic septa, vanished pseudolobules, regenerated hepatocytes, and normal lobular architecture. No compensated cirrhotic patient in the control group showed notable amelioration after 9 months.
Serum Hepatic Fibrosis Indices, Liver Function Tests, and HBV Serology Assay
Regarding antiviral effects of IFN-_, we identified only 3 patients who converted from HBeAg positive to anti-HBe positive and HBV-DNA positive to negative status after 9 months of IFN-_ treatment. In the control group, 1 patient exhibited seroconversion. These data indicate no significant antiviral activity of IFN-_. Subanalyses of HBV-DNA positive and negative patients (in situ hybridization) revealed no significant differences for inflammation and fibrosis after IFN-_ treatment (data not shown).
Hepatic fibrosis indices, including hyaluronic acid, laminin, collagen type IV, and procollagen type III, decreased significantly in the IFN-_ group after 9 months. In contrast, no remarkable alterations were found in the control group. In both groups ALT and AST levels were lowered to a similar extent (P < .01).
Safety Evaluation
During the clinical trial, no serious adverse events occurred. Side effects mostly occurred during the first 3 months of IFN-_ treatment, with fever, headache, muscular, skeletal and limb pain, nausea, and decreased white blood cell and platelet counts. However, discontinuation of treatment was necessary in only 5 patients. Dose reductions were not performed before discontinuation of treatment.
Interferon-gamma Decreases Activated Hepatic Stellate Cells in Patients With Chronic HBV Infection
_-SMA is a well-established marker for activated HSCs. Before treatment, _-SMA positive HSCs were found within inflamed and fibrotic regions of the damaged liver (Figure 4E). After 9 months of IFN-_ treatment, activated HSCs were scarcely found (Figure 4F). Numbers of activated HSCs decreased strongly after IFN-_ treatment, from 97 to 23 per observation field, especially in lobular regions. At the same time, liver damage-dependent infiltrates almost disappeared (Figure 4A, B), and the amount of interstitial collagens was markedly reduced (Figure 4C, D).
Interferon-gamma Inhibits Transforming Growth Factor-Beta Signaling in Liver Cells
Antiserum against a phosphorylated synthetic peptide of Smad2 was used to monitor TGF-_ signaling immunohistochemically. Fibrotic tissues of patients with chronic HBV infection displayed strong Phospho-Smad2 staining in hepatocytes and sinusoidal cells before IFN-_ treatment. Nuclear staining was almost absent after IFN-_ treatment, whereas Phospho-Smad2 was detected in the cytoplasm of some hepatocytes (Figure 4G, H). Semiquantitative analyses revealed a decreased number of Phospho-Smad2 positive nuclei (from 32 to 3 per observation field) at baseline compared with IFN-_ treatment. These results suggest that IFN-_ interferes with the profibrogenic TGF-_/Smad pathway in patients with chronic HBV infection.
Discussion
TGF-_, by virtue of HSC activation and ECM production, plays a key role in fibrogenesis,17 which is underlined by the observation that strategies abrogating TGF-_ signaling elicit antifibrotic effects in animal models.4,5 TGF-_ and IFN-_ have opposing effects on diverse cellular functions. TGF-_ signals through receptor serine/threonine kinases that phosphorylate and activate receptor Smads, whereas IFN-_ transduces its signal via a tyrosine kinase and the Jak/STAT pathway. Recently, transmodulation of IFN-_ and TGF-_ signaling pathways has been described with a negative regulatory action of IFN-_ on the TGF-_/Smad pathway.18,19
There is ample experimental evidence for a beneficial effect of IFN-_ on fibrotic disease.9,20 The action mode is primarily by antagonism of TGF-_ via inhibitory signaling events like Smad7.18 However, clinical trials evaluating IFN-_ as a treatment option in chronic liver disease, like chronic hepatitis B or C infections, are scarce. These have to take into consideration that a beneficial effect might only be seen after prolonged IFN-_ administration. Before the reported study, pilot investigations were performed to optimize IFN-_ dosage and treatment length. IFN-_ at 50 _g/day was administered to treat liver fibrosis in patients with chronic HBV infection for 6 and 9 months. In patients treated for 6 months, no significant improvement of liver fibrosis could be observed in biopsies, whereas a 9-month course indicated responsiveness. The chosen IFN-_ dosage was at the upper tolerable limit. Higher concentrations were not well tolerated mainly because of fever and ulcerations at the injection site.15 In the present study, no temporal side effects were apparent, including fever, headache, muscular, skeletal and limb pain, and nausea, necessitating discontinuation of treatment in only 5 patients.
The primary outcome of the study was assessed by histologic grading after 9-month IFN-_ treatment because liver histology remains the gold standard for fibrosis evaluation. However, variability in fibrosis distribution within the liver is a potential limitation. It is suggested that 1 cm or larger liver samples might be acceptable.12 A recent study suggests that a biopsy length of at least 2 cm might be necessary to evaluate fibrosis accurately by semiquantitative scores.21 The sample sizes in the present study were more than 2 cm in 102 specimens, more than 1.5 cm in 36, and more than 1 cm in 28. Thus, the vast majority of samples fulfilled the aforementioned criteria.
In the present trial, pathologists assessed the histology without knowledge of sampling time (first or second biopsy) and assignment to experimental or control group, although they were aware of clinical and biochemical data associated with the biopsy, which are important for diagnosis and assessment of disease advancement. Two scores were applied, the Scheuer criterion and a semiquantitative scoring system combining the ones described by Chevallier et al12 and Knodell et al11 with few modifications.13 In the latter, 4 main sites of fibrotic deposits and 4 major sites of inflammation were assessed. Considering the importance of numbers and width of septa, which indicate progression of disease and prognosis for patients with chronic liver fibrosis, mainly changes of septa were assessed and scored as effective (≥5 units) and improved (≥2 units) (Supplemental Table 2; see supplemental material online at www.cghjournal.org).
By these scoring systems, beneficial effects of IFN-_ on fibrosis could be observed. Sixty-three percent of the IFN-_ treated patients exhibited improved hepatic fibrosis degrees compared with 24.1% in the control group. For an intention to treat analysis, patients who did not undergo a repeat liver biopsy were assumed to have no change of biopsy score by treatment. Even with this assumption, inflammation and fibrosis scores significantly improved in the IFN-_ group (P < .001), from 13.9 ± 5.4 to 7.8 ± 4.0 (inflammation score) and 13.1 ± 6.2 to 10.4 ± 5.6 (fibrosis score), whereas in the control group both scores were unchanged (inflammation: 14.4 ± 5.5 to 14.9 ± 5.7; fibrosis: 13.5 ± 6.7 to 12.9 ± 5.1; P > .05). Beneficial histologic changes assessed according to the semiquantitative scoring system, ie, width, number of septa, and lobular fibrosis, were apparent in all fibrosis indices. With Masson trichrome and sirius red staining, changes of collagen deposits after IFN-_ treatment were most apparent. It is noteworthy that a beneficial effect was seen in patients with all degrees of liver fibrosis, even with compensated cirrhosis (stage 4). Although the determined serum fibrosis indices hyaluronic acid, laminin, procollagen type III, and collagen type IV demonstrated a tendency toward decreased values in the IFN-_ group, which reached statistical significance, the overlap between the values was high and did not indicate specificity.
Unexpectedly, the inflammatory indices, ie, bridging and piecemeal necrosis as well as portal and lobular inflammation, were similarly decreased after IFN-_ treatment. One possible explanation is IFN-_-dependent decrease of the number of activated HSCs, which produce and secrete both proinflammatory (eg, tumor necrosis factor-_ and interleukin-1) and profibrogenic (eg, TGF-_ and platelet-derived growth factor) factors.22 A similar beneficial effect has previously been observed in experimental models and in cell studies.2224
Furthermore, IFN-_ almost completely eliminated nuclear staining of Smad2 in HBV patients, indicating that efficient IFN-_-dependent inhibition of TGF-_ signaling occurs. Besides its role in maintaining the transdifferentiation stage of MFBs, TGF-_ is a strong inducer of hepatocyte apoptosis and, thus, a negative regulator of liver regeneration and liver cell mass. Thus, abrogation of TGF-_ signaling in HSCs and hepatocytes might contribute to improved conditions in patients with fibrosis.25
IFN-_ usage for chronic HBV infection is still considered controversial, with most clinical trials reporting no beneficial effects.6,26 In vitro studies indicate that local release of IFN-_ by antigen-stimulated T cells might be critical for anti-HBs formation.27 During the study period, only 3 patients in the IFN-_ group and 1 patient in the control group exhibited seroconversion, ie, eliminated HBeAG or lost HBV-DNA, indicating no effect of the 9-month treatment course on viral elimination.
Among patients in the biopsied group completing the study, 14 patients in the IFN-_ group were HBV-DNA negative (by in situ hybridization to liver tissue). It might not be excluded that confounding factors, eg, chronic alcohol ingestion, aggravate or perpetuate liver fibrosis development. Subanalysis regarding IFN-_ effectiveness in these patients did not reveal statistically significant differences.
Although the study was initially intended to be performed as placebo-controlled, objections from ethical committees necessitated the application of diammone-glycyrrhizinate and potassium-magnesium aspartate in both groups and an add-on study protocol. Our results indicate that these medications have only a slight effect on fibrosis development; however, they might have improved liver function.
In summary, IFN-_ has significant beneficial effects on liver fibrosis in patients with chronic HBV infection. Moreover, our data indicate that IFN-_ antagonizes the TGF-_/Smad pathway, thereby providing a molecular explanation for its antifibrotic effects.
Methods
Patient Recruitment
Patients were divided into 2 groups, one who accepted liver biopsy and one who did not. In the liver biopsy group, inclusion criteria were positivity for HBsAg in serum for at least 6 months, age from 1865 years, and performance of liver biopsy showing hepatic fibrosis stages 24 according to the Scheuer criterion.10 For patients without liver biopsy, inclusion criteria were positivity for HBsAg in serum for at least 6 months, age from 1865 years, and at least 2 values of serum hepatic fibrosis indices (including hyaluronic acid, laminin, collagen type IV, and procollagen type III) ranging above 2-fold the upper normal limit.
Patients with HCV or HDV infection, psychosis, pregnancy, lactation, or serious heart and/or renal failure were excluded, as were patients with the following therapies within the previous 3 months: IFN-_ or lamivudine, medications prescribed to reduce fibrogenesis such as colchicine, herbal recipe 861, and extractum semen percise. Patients with decompensated liver diseases, eg, serum albumin <3.5 g/dL, serum total bilirubin >85.5 _mol/L, serum ALT >400 U/L, and prothrombin activity ≦60%, history of hepatic encephalopathy, ascites, and/or white blood cell counts of less than 4 G/L and/or platelet counts of less than 50 G/L were excluded.
Ninety-nine and 216 chronic HBV-infected patients in biopsied and non-biopsied groups, respectively, fulfilled the criteria and provided written informed consent.
Study Design
Hepatologists from 6 hospitals in China (Zhejiang, Beijing, Shanghai, Chongqin, Wuhan, and Guangzhou) participated in the study, which was performed from February 1999-November 2001 and approved by the ethical committees. Patients were randomly assigned to IFN-_ treatment or control groups at a 2:1 (biopsied group) or 1:1 ratio (non-biopsied group). The random allocation sequence and methods were designed and implemented by B.-H. Su (Department of Statistics, Shanghai Second Medical University). All patients obtained diammone-glycyrrhizinate (Foscarnet Sodium; Chiaking Pharmaceutical, Lianyungang, China) and potassium-magnesium aspartate (Panangin; Shangdongquancheng Pharmaceutical, Quancheng, China). Patients in the IFN-_ group received 50 _g (106 units) recombinant human interferon-gamma 1b (IFN-_; Clonbiotech, Shanghai, China) intramuscularly daily for the first 3 months and on alternate days for the following 6 months. Clinical and laboratory assessments were performed after 3, 6, 9, and 12 months. Liver function tests included ALT, AST, albumin, and bilirubin; HBV serology included HBsAg, HB surface antibody (anti-HBs), HBeAg, HB e antibody (anti-HBe), HbcAg, and HBV DNA copy number as well as whole blood cell counts.
Histology
Liver biopsies were performed before and after 9 months of treatment. Percutaneous puncture was guided by ultrasound. Sample length was at least 1 cm, mostly 1.5 cm. After fixation in 4% formaldehyde, paraffin sections were stained with hematoxylin-eosin, Masson trichrome, sirius red, and for reticular fibers. For inflammation and fibrosis assessment, the Scheuer criterion was applied.10 However, this criterion is not completely suitable to assess therapeutic efficacy, because it does not reflect major changes occurring within 1 grade/stage of inflammation/fibrosis. Therefore, a semiquantitative scoring system combining the Chevallier fibrosis and the Knodell inflammatory scoring systems was applied with a few modifications (Tables 1 and 2).1113
Three pathologists independently assessed histology without knowledge about sampling time points (first or second biopsy) and assignment to experimental or control group. Clinical data associated with the biopsy were available to the pathologist. A final consensus on one score was achieved by collective assessment of the results.
By immunohistochemistry _-SMA (Sigma-Aldrich, Taufkirchen, Germany) and phosphorylated Smad2 (Cell Signalling, Frankfurt, Germany) were detected. For semiquantitative analysis, from every specimen of 28 representative patients, 10 fields (magnification 200_) were selected randomly before and after IFN-_ treatment. Positive cells were counted under the light microscope by a pathologist who was blinded for the sequence of treatment. For in situ hybridization, an HBV-DNA probe was used as previously reported.14
From our pilot study, significant differences of primary outcome end points (fibrosis score changes ≥ ±2 according to the modified Chevallier system, histologic examination) were found. In addition, our prestudy power calculation (>80%) was made on the assumption that a difference of 30% between both groups exists, and the required population size to detect such a difference is 51 (34 in the IFN-_ group and 17 in the control group).15
Serum Hepatic Fibrosis Indices, Liver Function Tests, and HBV Serology Assay
Serum hepatic fibrosis indices included hyaluronic acid, laminin, collagen type IV, and procollagen type III, which were determined by radioimmunoassays purchased from Shanghai Navy Medical Institute. Serum values for ALT, AST, albumin, and bilirubin were determined by standard clinical chemical assays.
HBV copy numbers were determined by polymerase chain reaction as described,16 providing a linear range between 1 _ 1035 _ 107 copies (PG Biotech, Shenzhen, China).
Side Effects
Drug-related side effects were evaluated and graded as mild, moderate, severe, or life-threatening according to the World Health Organization guidelines.
Statistical Analysis
Results are expressed as means for rank data and include standard deviation for quantitative data. Analysis for statistical significance was performed according to Wilcoxon rank sum test with values of P <.05 considered significant.
|
|
| |
| |
|
|
|
