A meta-epidemiological study of bias in randomized clinical trials of open and laparoscopic surgery.

BACKGROUND: Blinding, random sequence generation, and allocation concealment are established strategies to minimize bias in RCTs. Meta-epidemiological studies of drug trials have demonstrated exaggerated treatment effects in RCTs where such methods were not employed. As blinding is more difficult in surgical trials it is important to determine whether this applies to them. The study aimed to investigate this using systematic meta-epidemiological methods. METHOD: The Cochrane Database of Systematic Reviews was searched for systematic reviews of RCTs that compared laparoscopic and open abdominal surgical procedures. Each review was then scrutinized to determine whether at least one of the included trials was blinded. Eligible reviews were updated and individual RCTs retrieved. Extracted data included the primary outcomes of interest (length of stay and complications), secondary outcomes and a risk of bias assessment. A multistep meta-regression analysis was then performed to obtain an overall difference in the reported outcome differences between trials that employed each bias-minimization strategy, and those that did not. RESULTS: Some 316 RCTs were included, reporting on eight different procedures. Patient-blinded RCTs reported a smaller difference in length of stay between laparoscopic and open groups (difference of standardized mean differences -0·36 (95 per cent c.i. -0·73 to 0·00)) and complications (ratio of odds ratios 0·76 (95 per cent c.i. 0·61 to 0·93)). Blinding of postoperative carers and outcome assessors had similar effects. CONCLUSION: Lack of blinding significantly altered the treatment effect estimates of RCTs comparing laparoscopic and open surgery. Blinding should be implemented in surgical RCTs where possible to avoid systematic bias.

Network meta-analysis of surgical management of gastro-oesophageal reflux disease in adults.

BACKGROUND: Proton pump inhibitors are the mainstay of treatment for gastro-oesophageal reflux disease, but are associated with ongoing costs and side-effects. Antireflux surgery is cost-effective and is preferred by many patients. A total (360o or Nissen) fundoplication is the traditional procedure, but other variations including partial fundoplications are also commonly performed, with the aim of achieving durable reflux control with minimal dysphagia. Many RCTs and some pairwise meta-analyses have compared some of these procedures but there is still uncertainty about which, if any, is superior. Network meta-analysis allows multiple simultaneous comparisons and robust synthesis of the available evidence in these situations. A network meta-analysis comparing all antireflux procedures was performed to identify which has the most favourable outcomes at short-term (3-12 months), medium-term (1-5 years) and long-term (10 years and more than 10 years) follow-up. METHODS: Article databases were searched systematically for all eligible RCTs. Primary outcomes were quality-of-life measures and dysphagia. Secondary outcomes included reflux symptoms, pH studies and complications. RESULTS: Fifty-one RCTs were included, involving 5357 patients and 14 different treatments. Posterior partial fundoplication ranked best in terms of reflux symptoms, and caused less dysphagia than most other interventions including Nissen fundoplication. This was consistent across all time points and outcome measures. CONCLUSION: Posterior partial fundoplication provides the best balance of long-term, durable reflux control with less dysphagia, compared with other treatments.

Network meta-analysis of the effect of preoperative carbohydrate loading on recovery after elective surgery.

BACKGROUND: Three meta-analyses have summarized the effects of preoperative carbohydrate administration on postoperative outcomes in adult patients undergoing elective surgery. However, these studies could not account for the different doses of carbohydrate administered and the different controls used. Multiple-treatments meta-analysis allows robust synthesis of all available evidence in these situations. METHODS: Article databases were searched systematically for RCTs comparing preoperative carbohydrate administration with water, a placebo drink, or fasting. A four-treatment multiple-treatments meta-analysis was performed comparing two carbohydrate dose groups (low, 10-44 g; high, 45 g or more) with two control groups (fasting; water or placebo). Primary outcomes were length of hospital stay and postoperative complication rate. Secondary outcomes included postoperative insulin resistance, vomiting and fatigue. RESULTS: Some 43 trials involving 3110 participants were included. Compared with fasting, preoperative low-dose and high-dose carbohydrate administration decreased postoperative length of stay by 0·4 (95 per cent c.i. 0·03 to 0·7) and 0·2 (0·04 to 0·4) days respectively. There was no significant decrease in length of stay compared with water or placebo. There was no statistically significant difference in the postoperative complication rate, or in most of the secondary outcomes, between carbohydrate and control groups. CONCLUSION: Carbohydrate loading before elective surgery conferred a small reduction in length of postoperative hospital stay compared with fasting, and no benefit in comparison with water or placebo.

Primary hepatocellular neoplasms in a MODY3 family with a novel HNF1A germline mutation.

Maturity onset diabetes of the young type 3 (MODY3) and hepatocellular adenomas (HCAs) are associated with mutations in the HNF1A gene. HNF1A codes for the transcription factor HNF1α, which interacts with DNA as a homodimer or a heterodimer with HNF1ß, to regulate multiple cellular functions including glucidic metabolism, lipidic transport, and detoxication. We report three members of one family with a novel germline in-frame deletion of HNF1A exons 2-3 identified initially using array CGH and direct sequence analysis. All three family members have MODY3 in association with primary liver cell tumours (HCA, liver adenomatosis (LA), and hepatocellular carcinoma (HCC)). Additionally, a high rate of infant mortality was observed in the family. The described family demonstrates a novel HNF1A mutation associated with both benign and malignant primary liver cell tumours and MODY3.

Randomized controlled trial of preoperative oral carbohydrate treatment in major abdominal surgery.

BACKGROUND: Major surgery is associated with postoperative insulin resistance which is attenuated by preoperative carbohydrate (CHO) treatment. The effect of this treatment on clinical outcome after major abdominal surgery has not been assessed in a double-blind randomized trial. METHODS: Patients undergoing elective colorectal surgery or liver resection were randomized to oral CHO or placebo drinks to be taken on the evening before surgery and 2 h before induction of anaesthesia. Primary outcomes were postoperative length of hospital stay and fatigue measured by visual analogue scale. RESULTS: Sixty-nine and 73 patients were evaluated in the CHO and placebo groups respectively. The groups were well matched with respect to surgical procedure, epidural analgesia, laparoscopic procedures, fasting period before induction and duration of surgery. Postoperative changes in fatigue score from baseline did not differ between the groups. Median (range) hospital stay was 7 (2-35) days in the CHO group and 8 (2-92) days in the placebo group (P = 0.344). For patients not receiving epidural blockade or laparoscopic surgery (20 CHO, 19 placebo), values were 7 (3-11) and 9 (2-48) days respectively (P = 0.054). CONCLUSION: Preoperative CHO treatment did not improve postoperative fatigue or length of hospital stay after major abdominal surgery. A benefit is not ruled out when epidural blockade or laparoscopic procedures are not used. REGISTRATION NUMBER: ACTRN012605000456651 (http://www.anzctr.org.au).

Reverse portal flow after liver transplantation--ominous or acceptable?

Reversal of portal flow or hepatofugal flow after liver transplantation is a rare complication after liver transplantation. The available reports in the literature suggest that it is an ominous condition that requires immediate operative intervention, failing which prognosis would be grim. We report two children from two different centers who developed hepatofugal flow in the immediate post-operative period after liver transplantation. The possible etiologies in these patients were acute rejection in one and absence of an MHV causing inadequate hepatic venous outflow in the other. Both patients were treated non-operatively with steroids and immunosuppression. Spontaneous reversal to a normal hepatopetal flow occurred in both and the patients continue to be well six months after the transplant. Our experience contradicts the viewpoint that hepatofugal flow equates to mortality in the absence of surgical intervention. It remains to be defined as to which patients with hepatofugal flow will benefit from surgical intervention.

Systematic review of tumour number and outcome after radical treatment of colorectal liver metastases.

BACKGROUND: Resection of colorectal liver metastases (CLMs) is potentially curative but the effect of tumour number on prognosis is uncertain. This study compared the prognosis after resection and/or ablation of between one and three, or four or more CLMs. METHODS: A systematic literature review from January 2000 to June 2008 was performed. Study selection and data extraction were standardized, and analysis included assessment of methodological quality, heterogeneity and bias. Main outcomes were 3- and 5-year survival. A meta-analysis comparing radical treatment in the two groups was performed using the hazard ratio for overall survival. RESULTS: Of 1307 studies screened, 46 (9934 patients) were included in the analysis. Methodological quality was variable, and there was significant heterogeneity and reporting bias. The overall 5-year survival rate after radical treatment ranged from 7 to 58 per cent. Pooled hazard ratio for overall survival was 1.67 (95 per cent confidence interval 1.43 to 1.95; P < 0.001). Median reported 5-year survival for patients with four or more CLMs was 17.1 per cent. CONCLUSION: Radical treatment of more than three CLMs results in poorer overall survival. Nevertheless, 5-year survival is achievable and the number of lesions should not, of itself, be used to exclude patients from surgery.

Slow proliferation as a biological feature of colorectal cancer metastasis.

BACKGROUND: We have recently reported an inverse relationship between colon cancer progression and tumour proliferative activity. Here, we extend our findings by evaluating the proliferative activity of liver metastatic lesions and primary colorectal cancers (CRC) that differ in their metastatic potential. METHODS: Using an earlier established multi-gene proliferation signature (GPS), proliferative levels were analysed in 73 primary CRCs and 27 liver metastases. RESULTS: Compared with primary CRCs, we observed a significantly lower expression of the GPS in liver metastases and confirmed their lower proliferative levels by quantitative RT-PCR and Ki-67 immunostaining. No difference could be detected in apoptotic indices as assessed by M30 immunostaining, indicating that the net growth rate is lower in metastases relative to primary tumours. Notably, relapsed primaries or those with established metastases had significantly lower proliferative activity than CRCs that were non-metastatic and did not relapse. CONCLUSION: Our results suggest that slow proliferation is a biological characteristic of both liver metastases and those primary tumours with the ability to metastasise. The delineation of the mechanisms underlying the inverse association between proliferation and CRC aggressiveness may be important for the development of new therapeutic strategies.

Reduced expression of a gene proliferation signature is associated with enhanced malignancy in colon cancer.

The association between cell proliferation and the malignant potential of colon cancer is not well understood. Here, we evaluated this association using a colon-specific gene proliferation signature (GPS). The GPS was derived by combining gene expression data obtained from the analysis of a cancer cell line model and a published colon crypt profile. The GPS was overexpressed in both actively cycling cells in vitro and the proliferate compartment of colon crypts. K-means clustering was used to independantly stratify two cohorts of colon tumours into two groups with high and low GPS expression. Notably, we observed a significant association between reduced GPS expression and an increased likelihood of recurrence (P < 0.05), leading to shorter disease-free survival in both cohorts. This finding was not a result of methodological bias as we verified the well-established association between breast cancer malignancy and increased proliferation, by applying our GPS to public breast cancer data. In this study, we show that reduced proliferation is a biological feature characterizing the majority of aggressive colon cancers. This contrasts with many other carcinomas such as breast cancer. Investigating the reasons underlying this unusual observation may provide important insight into the biology of colon cancer progression and putative novel therapy options.

Genome wide expression profiling identifies genes associated with colorectal liver metastasis.

Tumour cells have to undergo gene expression changes in order to metastasise and adapt to a new site. We investigated these changes in liver metastases of colorectal cancer by using genome-wide microarray analysis to profile the expression of 48 primary tumours and 28 liver metastases. Statistical analysis of these expression profiles using the significance analysis of microarrays (SAM) method identified 778 genes differentially expressed between primary tumours and metastases. Gene ontology analysis revealed that genes associated with tissue remodelling and immune response were upregulated in metastases relative to primary tumours, whereas genes associated with proliferation and oxidative phosphorylation were downregulated. Quantitative real-time PCR confirmed the differential expression of selected genes, osteopontin, versican, ADAM17, CKS2, PRDX1, CXCR4, CXCL12, and LCN2. The upregulation of genes associated with tissue remodelling and immune response are likely to be involved in metastatic invasion and colonisation of the new site because these genes can promote tumour progression. However, downregulation of genes associated with proliferation suggests that proliferation in metastases was reduced relative to primary tumours.

Sequential changes in the metabolic response to orthotopic liver transplantation during the first year after surgery.

OBJECTIVE: To quantify the sequential changes in the metabolic response occurring in patients with end-stage liver disease after orthotopic liver transplantation (OLT). SUMMARY BACKGROUND DATA: Detailed quantification of the changes in energy expenditure, body composition, and physiologic function that occur in patients after OLT has not been performed. Understanding these changes is essential for the optimal management of these patients. METHODS: Fourteen patients who underwent OLT for end-stage liver disease had measurements of resting energy expenditure, body composition, and physiologic function immediately before surgery and 5, 10, 15, 30, 90, 180, and 360 days later. RESULTS: Resting energy expenditure was significantly elevated after surgery (24% above predicted), peaking around day 10 after OLT, when it averaged 42% above predicted. A significant degree of hypermetabolism was still present at 6 months, but at 12 months measured resting energy expenditure was close to predicted values. Before surgery, measured total body protein was 82% of estimated preillness total body protein. During the first 10 days after OLT, a further 1.0 kg (10%) of total body protein was lost, mostly from skeletal muscle. Only 54% of this loss was restored by 12 months. Significant overhydration of the fat-free body was seen before OLT, and it was still present 12 months later. Although significant losses of body fat and bone mineral occurred during the early postoperative period, only body fat stores were restored at 12 months. Both subjective fatigue score and voluntary hand grip strength improved rapidly after OLT to exceed preoperative levels at 3 months. At 12 months grip strength was close to values predicted for these patients when well. Respiratory muscle strength improved less markedly and was significantly lower than predicted normal levels at 12 months. CONCLUSIONS: Before surgery, these patients were significantly protein-depleted, overhydrated, and hypermetabolic. After surgery, the period of hypermetabolism was prolonged, restoration of body protein stores was gradual and incomplete, and respiratory muscle strength failed to reach expected normal values. Our measurements indicate that OLT does not normalize body composition and function and imply that a continuing metabolic stress persists for at least 12 months after surgery.

Skip metastases in colon cancer: assessment by lymph node mapping using molecular detection.

BACKGROUND: Colon cancer has been assumed to spread sequentially through the regional lymphatic bed, with skip metastases occurring in only 1% to 3% of cases. Molecular techniques allow the detection of occult metastases, but to date have not been applied to assess the pattern of regional lymphatic spread of colon cancer. METHODS: Fifty-five tumors from 54 patients with colonic adenocarcinoma were studied. Lymph node mapping was performed on fresh colonic specimens recording the position of each node on an anatomical diagram. Half of each lymph node was submitted for routine histology examination and half assayed for keratin 20 gene expression by reverse transcription-polymerase chain reaction. Logistic regression was used to analyze the distribution of histologic and occult metastases. RESULTS: A total of 1084 lymph nodes were dissected (median, 19 nodes; range, 4-52). Sixty-four lymph nodes from 20 tumors had histologically evident metastases and 76 lymph nodes from 13 tumors had occult metastases. There was no difference in the distribution of either histologic or occult metastases among paracolic, intermediate, and apical node groups. Ten patients had evidence of anatomical skip lesions after lymph node mapping and molecular analysis, only 1 of which was histologically detectable. CONCLUSIONS: This study demonstrates a higher incidence of skip metastases in colon cancer assessed by molecular techniques than has previously been reported, challenging the concept of sequential development of early lymph node metastases.

Diagnostic use of the sentinel node in colon cancer.

PURPOSE: The aim of this study was to compare the lymphatic drainage of colon cancer with the anatomic distribution of histologic and submicroscopic lymph node metastases. METHODS: Patients attending for colectomy were eligible to enter the study. At the commencement of surgery, 40 MBq of 99mTc colloidal antimony sulfide in 2 ml of Patent Blue dye was injected subserosally around the tumor. Resection was completed in a standard fashion. After resection, specimens were imaged with a gamma camera to determine the site of sentinel lymph nodes, and then dissected, recording the position of the lymph nodes on an anatomic diagram. Recovered lymph nodes were bisected, one-half for routine histology and one-half for assessment by keratin 20 (K20) reverse transcription polymerase chain reaction. The kappa measure of agreement was used to assess concordance between sentinel nodes and histologic and submicroscopic metastases. RESULTS: Four hundred fifty-six lymph nodes were dissected from 26 tumors and evaluated using lymphoscintigraphy and lymph node mapping. Sentinel nodes were evident in 23 tumors (88 percent). The sensitivity of sentinel nodes involvement as a predictor of metastatic disease was 55 percent (95 percent confidence interval, 23-83), with a false negative (nondiagnostic) rate of 45 percent. Sentinel nodes involved the apical group in four tumors, and represented anatomic "skip" lesions in four tumors. CONCLUSIONS: Direct lymphatic drainage to the apical group does occur in colon cancer; however, sentinel node mapping of colon cancer by this technique is of little clinical value because of the poor concordance between lymph node metastases and sentinel nodes.

Eosinophilic cholangitis: a case of 'malignant masquerade'.

BACKGROUND: A variety of causes of inflammatory bile duct stricture can masquerade as hilar cholangiocarcinoma. Eosinophilic cholangitis is a further example. CASE OUTLINE: A 50-year-old woman with epigastric pain and deranged liver function was found to have a stricture of the common hepatic duct on ERCP with an associated mass on MRl.The lesion was excised with reconstruction of the right and left hepatic ducts, and the patient recovered well. Histopathological examination of the resected gallbladder and bile duct showed diffuse inflammation with a predominant eosinophil infiltrate.The presence of Candida albicans in the bile duct lumen can probably be attributed to the preoperative biliary stent.There was a modest postoperative rise in peripheral eosinophil count. DISCUSSION: A literature search reveals only six previous cases of eosinophilic cholangitis, but similar infiltrates have also been seen in occasional cholecystectomy specimens. As the present patient did not have gallstones, the aetiology remains unclear. Peripheral eosinophilia is an unreliable clue to the diagnosis, which is usually likely to escape detection until the biliary stricture has been resected.

Surgery for colorectal liver metastases with hepatic lymph node involvement: a systematic review.

BACKGROUND: Liver resection for colorectal metastases is the only known treatment associated with long-term survival; extrahepatic disease is usually considered a contraindication to such treatment. However, some surgeons do not regard spread to the hepatic lymph nodes as a contraindication provided that these nodes can be excised adequately. A systematic review of the literature was undertaken to address this issue. METHODS: An electronic search using Medline, Cancerlit and Embase databases was performed for studies reporting liver resection for colorectal metastases from 1964 to 1999. Data were extracted from papers reporting outcome for patients with positive hepatic nodes and analysed according to predetermined criteria. RESULTS: Fifteen studies were identified that gave survival data on 145 node-positive patients. Five patients were reported to have survived 5 years after liver resection; one was disease free, two had recurrent disease and the disease status was not described in the remaining two. Five studies containing 83 patients specified a formal lymph node dissection as part of the surgical procedure and four of the five node-positive 5-year survivors were from these studies. CONCLUSION: There are few 5-year survivors after liver resection, with or without lymph node dissection, for colorectal hepatic metastases involving the hepatic lymph nodes.

Keratin 20 is a specific marker of submicroscopic lymph node metastases in colorectal cancer: validation by K-RAS mutations.

Lymph node status has major prognostic importance in colorectal cancer and greater precision in the diagnosis of lymph node metastases should provide better prognostic and therapeutic guidance. Keratin 20 (K20) gene expression has been used as a marker of lymph node metastases, but the evidence for this remains circumstantial. This study has therefore sought to determine K20 specificity and to correlate K20 expression with mutant K-RAS expression, in order to provide direct evidence that K20 expression in lymph nodes of colorectal cancer patients genuinely reflects metastatic disease. Specificity of K20 expression was established against a range of tissue types and 289 lymph nodes from 41 non-cancer control patients. K20 expression was restricted to gastrointestinal epithelia and was only present in one of the 289 control lymph nodes, giving a calculated specificity of 97.6% (95% confidence limits: 87.1-99.9%). Forty-two tumour samples were analysed for the presence of K-RAS codon 12 gene mutations using a RT-PCR mutant allele-specific amplification (MASA) technique. Thirteen tumours (31%) had codon 12 mutations detected by MASA and these were further analysed to determine the exact nature of the mutation. MASA was then used to screen the lymph nodes from these patients for the presence of the tumour-specific K-RAS transcript and the results were compared with K20 RT-PCR and histopathology from the same samples. Whilst K-RAS MASA was not as sensitive as K20 RT-PCR, there was substantial agreement between the assays. There were no K20-negative lymph nodes which were found to be K-RAS MASA-positive, whereas seven nodes in four patients were K20-positive and K-RAS-negative, in keeping with the differences in assay sensitivity. These results further validate K20 as a marker by providing greater certainty that what is being detected represents occult metastatic disease.

Cholecystostomy: a review of recent experience.

BACKGROUND: Operative (OC) and percutaneous cholecystostomy (PC) are rarely undertaken for severe acute cholecystitis in patients in whom cholecystectomy is technically difficult or those with significant comorbidity. METHODS: A retrospective review was undertaken of the clinical, radiological and audit records of patients who were treated by cholecystostomy between 1988 and 1997 at Auckland Hospital. RESULTS: During the 10-year period 19 patients (eight male, 11 female; median age: 70 years, range: 35-90 years) had a cholecystostomy (OC: n = 8; PC: n = 11). The main indication for PC was high anaesthetic risk (10 cases). The main indication for OC was failed cholecystectomy (six cases). The patients undergoing PC tended to have a higher American Society of Anesthesiologists (ASA) grade than patients undergoing OC. The median delay from presentation to cholecystostomy was 3 days. More than half (11/19) were done during the 3 years (1992-94) after the introduction of laparoscopic cholecystectomy. The number of tube-related complications was significantly higher in PC patients (10/11 vs 3/8; P = 0.04), and the number of systemic complications was higher in the OC patients (4/8 vs 0/11; P = 0.018). The median duration of tube drainage was 17 days (range: 0-82 days) for OC and 24 days (range: 5-93 days) for PC. Four patients had stone extraction at the time of OC, including two who also had a partial cholecystectomy. One OC patient had stone extraction via the cholecystostomy tract. A cholecystectomy was performed in four patients. CONCLUSION: The data indicate that PC is a safe approach for high-risk patients. Operative cholecystostomy had a role following failed cholecystectomy. But PC might be safer in these patients if they could be identified pre-operatively.