Normal approximation diagnostics for the Cox model

We discuss two diagnostic methods for assessing the accuracy of the normal approximated confidence region to the likelihood-based confidence region for the Cox proportional hazards model with censored data. The proposed diagnostic methods are extensions of the contour measures of Hodges (1987, Journal of the American Statistical Association 82, 149-154) and Cook and Tsai (1990, Journal of the American Statistical Association 85, 770-777) and the curvature measures of Jennings (1986, Journal of the American Statistical Association 81, 471-476) and Cook and Tsai (1990). These methods are also illustrated in a study of hepatocyte growth factor in patients with lung cancer and a Mayo Clinic randomized study of participants with primary biliary cirrhosis.

Cancer risk in primary biliary cirrhosis: a study in northern England

BACKGROUND

Suggestions that breast cancer may be more common in patients with primary biliary cirrhosis (PBC) have been challenged. It has recently been proposed that total cancer rates may be higher in patients with PBC, as well as liver cancers.

AIMS

To investigate these proposals on a strictly defined case series.

SUBJECTS

A total of 769 prevalent or incident PBC patients with "definite" or "probable" disease detected in a defined area of the north-east of England during 1987-94.

METHODS

Cancer events and deaths were identified by obtaining information from one or more of the following sources: Office for National Statistics (ONS) Central Registers, Regional Cancer Registry, and clinical case records. Standardised cancer incidence (SIR) and mortality ratios (SMR) were calculated using the local region as the standard population.

RESULTS

There were 97 cancer events during 1987-96. SIR from cancer registrations for all cancers was 1.7 (95% confidence interval (CI) 1.3 to 2.2), for liver cancer was 74 (95% CI 32 to 146), and for breast cancer was 1.1 (95% CI 0.4 to 2.4). SMR for all cancers was 1. 8 (95% CI 1.4 to 2.4), for liver cancer was 39 (95% CI 20 to 68), and for breast cancer was 0.4 (95% 0.1 to 1.6). The results were similar after excluding the first year of follow up after PBC diagnosis.

CONCLUSIONS

There was some evidence of a small increase in overall cancer incidence and mortality in PBC patients. With the exception of liver cancer, it is unlikely that there is a high excess incidence for PBC patients from any cancer at a particular site, and specifically breast cancer.

Randomised controlled trials of ursodeoxycholic-acid therapy for primary biliary cirrhosis: a meta-analysis

BACKGROUND

Ursodeoxycholic acid (UDCA) is the only approved treatment for primary biliary cirrhosis, but its effect on disease progression and survival is uncertain. The aim of this study was to clarify the efficacy of UDCA in primary biliary cirrhosis.

METHODS

A systematic review, including the use of meta-analysis, was done for the randomised and switch-over phases of trials comparing UDCA with placebo, obtained from Medline and Embase databases, and from manual searches derived from review articles and abstracts of major international meetings. All trials had more than a mean of 6 months' follow-up and only included patients with primary biliary cirrhosis (PBC) according to established diagnostic criteria.

FINDINGS

17 relevant articles were identified: 11 randomised controlled trials, including 1272 patients, and six reports of the switch-over phases. UCDA had a favourable effect on liver biochemistry in most of the studies but not on symptoms or the progression of histological stage; two studies did not assess survival, liver transplantation, or complications of liver disease. Meta-analysis showed no difference between UDCA and placebo in the incidence of death (odds ratio 1.21, 95% CI 0.71-2.04), liver related death (0.72, 0.22-2.32), liver transplantation (1.27, 0.78-2.07), death or liver transplantation (1.26, 0.87-1.82), and in the development of complications of liver disease (1.11, 0.64-1.92). With the primary end point defined by the authors (a combined end point in three studies, and death or liver transplantation in the others) an odds ratio of 1.53 (0.97-2.42) was obtained. Assessment of the switch-over phases, during which there was a longer follow-up, did not change the results of the meta-analysis.

INTERPRETATION

Published randomised controlled trials of UDCA do not show evidence of therapeutic benefit in PBC and its use as standard therapy needs to be re-examined.

Hepatic retransplantation in cholestatic liver disease: impact of the interval to retransplantation on survival and resource utilization

The aim of our study was to quantitatively assess the impact of hepatic retransplantation on patient and graft survival and resource utilization. We studied patients undergoing hepatic retransplantation among 447 transplant recipients with primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC) at 3 transplantation centers. Cox proportional hazards regression analysis was used for survival analysis. Measures of resource utilization included the duration of hospitalization, length of stay in the intensive care unit, and the duration of transplantation surgery. Forty-six (10.3%) patients received 2 or more grafts during the follow-up period (median, 2.8 years). Patients who underwent retransplantation had a 3.8-fold increase in the risk of death compared with those without retransplantation (P <.01). Retransplantation after an interval of greater than 30 days from the primary graft was associated with a 6.7-fold increase in the risk of death (P <.01). The survival following retransplantations performed 30 days or earlier was similar to primary transplantations. Resource utilization was higher in patients who underwent multiple consecutive transplantations, even after adjustment for the number of grafts during the hospitalization. Among cholestatic liver disease patients, poor survival following hepatic retransplantation is attributed to late retransplantations, namely those performed more than 30 days after the initial transplantation. While efforts must be made to improve the outcome following retransplantation, a more critical evaluation may be warranted for late retransplantation candidates.

Primary biliary cirrhosis once rare, now common in the United Kingdom?

There is a widespread impression that the number of patients with the autoimmune liver disease primary biliary cirrhosis (PBC) is increasing, although its incidence and prevalence vary widely. Using thorough case-finding methods and rigorous definitions to assess changes in incidence and prevalence with time and to explore the symptomatology and mortality of the disease in a large group of unselected patients, we performed a descriptive epidemiological study of PBC in a well defined population over a fixed period of time using established diagnostic criteria and with clinical follow-up of all cases. In a population of 2.05 million in northern England 770 definite or probable PBC cases were identified. Prevalence rose from 201.9 per 10(6) in the adult population and 541. 4 per 10(6) women over 40 in 1987 to 334.6 per 10(6) adults and 939. 8 per 10(6) women over 40 in 1994. Incidence was 23 per 10(6) in 1987 and 32.2 per 10(6) in 1994. Three hundred patients died in median follow-up of 6.27 years (141 liver deaths); the standardized mortality ratio was 2.85. At presumed diagnosis, 60.9% had no symptoms of liver disease. By June 1994 62% of prevalent patients had liver symptoms. PBC is apparently increasing. It is still unclear whether this is because of a true increase, case finding, or increased disease awareness. The study draws attention to (1) high mortality from liver disease and non-liver-related causes even in patients initially with no liver symptoms and (2) apparently poor diagnostic awareness of the disease.

Increased survival in advanced primary biliary cirrhosis patients with regular albumin infusions?

Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease. Frequently, it can slowly progress to cirrhosis and, finally, death. Liver transplantation is the only way to avoid this endpoint. For patients not accepted for liver transplantation, only symptomatic treatment remains. In advanced PBC, one of the main concerns is a decreasing albumin level. As albumin is an important protein in the human body, we wondered whether regular albumin infusions may result in a higher quality of life and prolong survival. In this case report we describe three patients who survived for a remarkably long time on this treatment regimen before finally succumbing to the sequela of advanced liver disease. Some features of albumin are mentioned which may help explain this phenomenon. We conclude that regular albumin infusions may be a means of decreasing morbidity and prolonging survival in a subgroup of patients with advanced PBC. However, prospective evaluation of regular albumin infusions in a group of PBC patients (e.g. those rejected for liver transplantation) seems mandatory.

Ten-year survival in ursodeoxycholic acid-treated patients with primary biliary cirrhosis. The UDCA-PBC Study Group

Ursodeoxycholic acid (UDCA) treatment has been shown to increase survival without orthotopic liver transplantation (OLT) in patients with primary biliary cirrhosis (PBC) at 4 years. Whether this beneficial effect was maintained over the long term remained to be established. In a large cohort of UDCA-treated patients with PBC, we aimed to determine the 10-year outcome of these patients using two endpoints: (1) survival without OLT, and (2) survival. The cohort was comprised of 225 patients with PBC treated with UDCA (13-15 mg/kg/d) monitored from the beginning of treatment until time of last follow-up, OLT, or death. Because of the absence of a control group, survival without OLT was compared with survival predicted by the Mayo model (first 7 years), and observed 10-year survival with an estimation of survival of a standardized control cohort of the French population. Observed survival without OLT of UDCA-treated patients was significantly higher (P <.04) than survival predicted by the Mayo model. Observed survival was significantly lower (P <. 01) than survival predicted from the French population. Observed survival of noncirrhotic patients was not different (P >.9) from that of the French control population but survival of cirrhotic patients was significantly lower (P <.0001). Twenty-two patients died; 13 patients died of hepatic causes and 4 patients died after OLT. In conclusion, survival without OLT among patients treated with UDCA for PBC is higher than that of untreated patients, as predicted by the Mayo model. Ten-year survival among UDCA-treated patients is slightly lower than that of an age- and sex-matched general population, the difference mainly being explained by mortality among cirrhotic patients.

A global goodness-of-fit statistic for Cox regression models

In this paper, a global goodness-of-fit test statistic for a Cox regression model, which has an approximate chi-squared distribution when the model has been correctly specified, is proposed. Our goodness-of-fit statistic is global and has power to detect if interactions or higher order powers of covariates in the model are needed. The proposed statistic is similar to the Hosmer and Lemeshow (1980, Communications in Statistics A10, 1043-1069) goodness-of-fit statistic for binary data as well as Schoenfeld's (1980, Biometrika 67, 145-153) statistic for the Cox model. The methods are illustrated using data from a Mayo Clinic trial in primary billiary cirrhosis of the liver (Fleming and Harrington, 1991, Counting Processes and Survival Analysis), in which the outcome is the time until liver transplantation or death. The are 17 possible covariates. Two Cox proportional hazards models are fit to the data, and the proposed goodness-of-fit statistic is applied to the fitted models.

The geoepidemiology of primary biliary cirrhosis: contrasts and comparisons with the spectrum of autoimmune diseases

Recent data have suggested that the prevalence of many autoimmune diseases is higher than originally suspected. Indeed, the incidence of some autoimmune diseases may be increasing. Part of the problem in these latter two issues is that there is a dearth of well-designed and controlled epidemiologic studies, and often confounding variables in diverse populations and geographic areas that are not well controlled. Primary biliary cirrhosis (PBC) is a highly directed, organ-specific autoimmune disease that results in the destruction of intrahepatic bile ducts. It is primarily a disease of middle-aged women. Although there is no obvious association with MHC class I or class II alleles, the relative risk of a family member of a first-degree relative within a family having a member with PBC is a hundred-fold that of the general population. Unfortunately, most epidemiologic studies have been descriptive, providing incidence and prevalence rates with many methodologic problems, including lack of an appropriate case definition, varying criteria for inclusion of cases, and inaccurate estimate of the time period to which the rate applies. Because PBC is a very definable disease with significant clinical and serologic overlaps among patients throughout the world, we believe that a review of the geoepidemiology of PBC is not only specifically of value to workers interested in autoimmune liver disease, but also of generic interest in the study of autoimmune disease. In this review, we discuss the nature of the existing epidemiologic data and the possible roles of genetic and environmental factors in the etiology of the disease and compare such data to similar observations for multiple sclerosis, systemic lupus erythematosus, and rheumatoid arthritis.

Utilization of the Mayo risk score in patients with primary biliary cirrhosis receiving ursodeoxycholic acid

BACKGROUND/AIMS

Ursodeoxycholic acid (UDCA) is an effective therapy for most patients with primary biliary cirrhosis (PBC). During the management of these treated patients, a number of clinically important issues arose including which patients might be candidates for combined therapy, which patients require endoscopy for variceal bleeding, and how survival can be predicted during treatment. Our aims were: 1) to identify factors associated with a suboptimal response to UDCA in patients with PBC; 2) to define a simple, non-invasive method to predict those PBC patients most apt to have esophageal varices; and 3) to determine the reliability of the Mayo survival model in predicting the course of UDCA treated patients.

METHODS

We analyzed the prospectively collected data of 180 patients, who we continue to follow, with PBC who participated in a randomized, placebo-controlled trial of UDCA.

RESULTS

After six months of UDCA therapy, patients with serum alkaline phosphatase levels less than twice normal (p < 0.04), and/or a Mayo risk score < 4.5 (p < 0.04) were more likely to respond favorably to treatment over a two year period. The Mayo risk score was the single risk factor independently predictive of development of varices (p < 0.01); 93% of patients who developed varices had a Mayo risk score > or = 4. The Mayo survival model, recalculated after 6 months on UDCA therapy accurately predicted patient survival.

CONCLUSIONS

Suboptimal responders to UDCA can be identified by assessment of serum alkaline phosphatase levels, and/or Mayo risk score. A Mayo risk score above 4 helps in selecting patients for endoscopic surveillance for varices and the Mayo survival model accurately predicts the clinical course in patients with PBC receiving UDCA.

Cost-effectiveness of ursodeoxycholic acid therapy in primary biliary cirrhosis

Ursodeoxycholic acid (UDCA) is a safe and effective treatment for patients with primary biliary cirrhosis (PBC), but the cost of this drug has raised concerns regarding cost-effectiveness. The aim of our study was to determine the cost-effectiveness of UDCA in PBC. We compared the costs and outcomes of managing PBC patients with and without UDCA. From two previously published trials, the effectiveness of UDCA was determined by comparing the annual reduction in the development of ascites, varices, variceal bleeding, encephalopathy, liver transplantation, and death between the treatment groups. Average annual costs for each of these events were estimated based on literature and institutional data. Approximately twice as many major events occurred in the placebo group compared with the UDCA group. The relative risk (RR) of liver transplantation (1.95; 95% CI: 1.14-3.68) and development of esophageal varices (3. 11; 95% CI: 1.57-10.65) were significantly higher in the placebo group compared with the UDCA group. There were no significant increases in the RR of ascites, variceal bleeding, encephalopathy, or death between the two groups. Based on the estimated annual cost of managing these events and the annual costs of UDCA ($2,500), there was an annual cost savings per patient of $1,372. Compared with the placebo group, patients receiving UDCA had a lower incidence of major complications and lower medical care costs.

Clinical significance of serum bilirubin levels under ursodeoxycholic acid therapy in patients with primary biliary cirrhosis

We determined whether the normalization of serum bilirubin level (SBL) induced by ursodeoxycholic acid (UDCA) therapy was associated with an improved clinical outcome in patients with primary biliary cirrhosis (PBC). We estimated the prognostic values of SBL measured after 6 months of UDCA treatment for survival free of orthotopic liver transplantation (OLT). We used a database of 548 patients with PBC followed in three trials of UDCA. Among UDCA-treated patients, we compared survival free of OLT in patients with normalized SBL (</=17 micromol/L) with those who had persistently elevated SBL. Difference in survival was tested between UDCA-treated patients whose SBL normalized with treatment and placebo patients who had normal baseline SBL. We evaluated, in each treatment group, the prognostic value of 6-month SBL. Survival was estimated using the Kaplan-Meier method and compared by the Cox model. Survival free of OLT was significantly longer in patients who had normalized SBL (P <. 0001; relative risk [RR]: 3.7, UDCA group). Survival free of OLT was not significantly different between UDCA patients with normalized SBL and placebo patients with a normal baseline SBL (P =.69). For several cutoffs of 6-month SBL, RRs of OLT or death were similar in UDCA-treated and placebo patients: the RR of OLT or death associated with a 6-month SBL more than 30 micromol/L was 6.0 for UDCA and 5.7 for placebo groups. In conclusion, normalization of SBL during therapy is associated with improved clinical outcome. SBL under UDCA therapy is a prognostic factor in PBC. SBL under UDCA therapy should be interpreted as in untreated patients.

Asymptomatic primary biliary cirrhosis: a study of its natural history and prognosis

OBJECTIVE

To document the natural history of asymptomatic primary biliary cirrhosis and identify prognostic features that would predict the development of symptomatic disease.

METHODS

A retrospective chart review of all patients with abnormal liver biochemical tests and antimitochondrial antibody-positive, liver biopsy-compatible primary biliary cirrhosis who were seen in a single tertiary care center between 1983 and 1994 was performed. Statistical analysis using Cox regression was employed to compare survival of the study population with an age- and gender-matched control population and to identify potential prognostic variables.

RESULTS

Ninety-one patients were included. Median age at presentation was 53.2 yr. Ninety percent were female. Median follow up was 61.2 months (range 7-206 months). Thirty-six percent (33 patients) became symptomatic with 11% (10 patients) progressing to death or liver transplant. Median predicted length of survival from onset of disease for the entire cohort was 14 yr. Patient survival was less than that predicted for an age- and gender-matched control population (p < 0.05). Univariate and multivariate analysis on a broad spectrum of clinical, biochemical, and histological features at the time of initial presentation failed to reveal any prognostic variables that would distinguish those who would become symptomatic from those who would remain symptom-free. Specifically, three primary variables of interest (associated autoimmune disorders, hepatomegaly, and histological stage) were not found to predict prognosis.

CONCLUSION

Patients who present with asymptomatic primary biliary cirrhosis have a shorter life span than the general population. Presently, there are no prognostic features that identify the patients who will develop progressive disease from those who will remain symptom-free. Therefore, treatment should be offered to all patients.

Major clinical events, signs and severity assessment scores related to actual survival in patients who died from primary biliary cirrhosis. A long-term historical cohort study

BACKGROUND/AIMS

One of the prognostic methods for survival in primary biliary cirrhosis (PBC) is the Mayo model, with a time-scale limited to 7 years. The aim of our study was to assess how major clinical events, signs, several severity assessment methods and Mayo survival probabilities fit in with actual patient survival, by using yearly observations until 0.5 years before patient death from PBC.

METHODOLOGY

Data of 32 patients dying from PBC were collected prior to death at -0.5, -1, -2 etc. years (median: -5 years, range: -16 to -0.5 years). Major events registered were: first occurrence of ascites, upper gastrointestinal bleeding or manifest hepatic encephalopathy and signs, first observation of spider naevi or purpura. Severity assessment methods applied (all with scores and classes) were: Mayo (M), Child-Campbell (C), Pugh-Child (P), Pugh-Child-PBC (PP), 'Child-Pugh' (CP), and Ascites Nutritional State-Child (ANS). Fifty percent survival estimates were calculated from Mayo scores. Severity assessment method variables were: ascites (C, P, PP, CP, ANS), encephalopathy (C, P, PP, CP), nutritional state (C, ANS), edema (M), age (M), serum albumin (M, C, P, PP, CP), bilirubin (C, M, P, PP, CP), and prothrombin time (M, P, PP, CP).

RESULTS

In 27 out of 32 patients a major event occurred, always between -6 and -0.5 years (median: -1 year) and, never between -16 and -7 years (p < 0.0001). A sign was first observed in 30/32 between -14 and -0.5 years (median: -2 years). Compared to the total population, a sign, and even more so, an event indicated a shorter survival (p = 0.004 and p = 0.0002, respectively). The median 50% estimated survival (predicted by the Mayo model) fitted the actual survival from -6 to -0.5 years (r = -0.7, p < 0.0001), but not from -16 to -7 years (r = -0.1, p = 0.4). All -6 to -0.5-year severity scores correlated (p < 0.0001) both with actual survival (M, C, P, PP, and CP r = 0.7; ANS r = 0.5) and with estimated M 50% survival (C, P, PP, CP r = -0.9; ANS r = -0.6; M score: -0.99), but none with actual survival from -16 to -7 years, except for M, slightly (r = -0.3, p = 0.04). A nomogram for mean C, CP, M and ANS scores related to actual survival was constructed for the -6 to -0.5-year period. The C and CP classes A, B, and C did not appear to distinguish sufficiently into actual survival, whereas the M classes did.

CONCLUSIONS

The occurrence of a major event appeared to exclude survival over 6 years. In these final 6 years, Child-Campbell, Mayo and Pugh scores correlated equally well with actual survival and better than Ascites/Nutritional State score. In our PBC patients, Campbell was an excellent alternative for Pugh; for Pugh, the original Child-Turcotte variable limits were fully sufficient.

Experiments to determine whether recursive partitioning (CART) or an artificial neural network overcomes theoretical limitations of Cox proportional hazards regression

New computationally intensive tools for medical survival analyses include recursive patitioning (also called CART) and artificial neural networks. A challenge that remains is to better understand the behavior of these techniques in effort to know when they will be effective tools. Theoretically they may overcome limitations of the traditional multivariable survival technique, the Cox proportional hazards regression model. Experiments were designed to test whether the new tools would, in practice, overcome these limitations. Two datasets in which theory suggests CART and the neural network should outperform the Cox model were selected. The first was a published leukemia dataset manipulated to have a strong interaction that CART should detect. The second was a published cirrhosis dataset with pronounced nonlinear effects that a neural network should fit. Repeated sampling of 50 training and testing subsets was applied to each technique. The concordance index C was calculated as a measure of predictive accuracy by each technique on the testing dataset. In the interaction dataset, CART outperformed Cox (P < 0.05) with a C improvement of 0.1 (95% CI, 0.08 to 0.12). In the nonlinear dataset, the neural network outperformed the Cox model (P < 0.05), but by a very slight amount (0.015). As predicted by theory, CART and the neural network were able to overcome limitations of the Cox model. Experiments like these are important to increase our understanding of when one of these new techniques will outperform the standard Cox model. Further research is necessary to predict which technique will do best a priori and to assess the magnitude of superiority.

Liver transplantation for primary biliary cirrhosis and primary sclerosing cholangitis: does medical treatment alter timing and selection?

Liver transplantation is a highly effective treatment for patients with advanced primary biliary cirrhosis and primary sclerosing cholangitis. Transplantation is indicated when the patient's survival with transplantation is better than without or, earlier than this, if the patient's quality of life is intolerable from intractable fatigue or pruritus. Medical therapies for chronic cholestatic liver diseases are very limited. Ursodeoxycholic acid therapy in primary biliary cirrhosis reduces cholestasis and prolongs transplant-free survival; no other drugs are of proven efficacy in primary biliary cirrhosis, and none have any benefit on the disease progression of primary sclerosing cholangitis. Aggressive endoscopic therapy may produce symptomatic and biochemical improvement in primary sclerosing cholangitis but should be done without the expectation of retarding disease progression. Bilirubin is one of five criteria of the Child-Turcotte-Pugh score, which is necessary for the United Network for Organ Sharing listing for orthotopic liver transplantation. In addition, it is a major prognostic indicator in all the predictive models for primary biliary cirrhosis. Bilirubin reduction with ursodeoxycholic acid therapy in primary biliary cirrhosis appears to parallel disease severity, and prognostic models utilizing bilirubin retain their predictive power for survival even in treated patients. In summary, medical therapies for chronic cholestatic liver disease have very little effect on disease progression and, subsequently, on the timing or selection for transplantation. Liver transplantation is the only definitive therapy for primary biliary cirrhosis and primary sclerosing cholangitis.

Extrahepatic bile duct atresia: how efficient is the hepatoporto-enterostomy?

PURPOSE

Prospectively achieved findings of 151 children with EHBA operated between 1972-1997, of 108 patients in particular who were operated five and more years ago, should give evidence of the efficiency of HPE.

METHOD

For five and more years the development of the disease was prospectively evaluated according to a cognitively conditioned program. The objective source parameter was the hepatic histology at the time of surgery. Departing from that parameter the structural hepatic changes were followed up for five years and more. They were related to the bile flow achieved and the age of the patient, at the time of surgery as well as to the present time of follow-up.

RESULTS

Of 108 patients who were operated five and more years ago, 62 are still alive (57%) 58 children are icterus-free. The following results can be stressed: 1) There is a close correlation between the age of the patient and the fibrosis level of the liver at the time of surgery. 2) There is a close correlation between the age and the fibrosis level at the time of surgery on the one hand and the long-term survival rate, the absence of icterus and the state of the hepatic function on the other. 3) There is a close correlation between the age and the fibrosis level at surgery on the one hand and the operatively feasible (continued) bile flow on the other. 4) With favorable conditions concerning age and fibrosis level at the time of surgery, the size of the bile flow achieved through surgery is influenced positively and negatively by the morphology and bio-physical parameters of the porta hepatis as well as by postoperative cholangitis episodes.

CONCLUSION

The efficiency of HPE depends on the age and (in correlation to that) on the level of hepatic fibrosis at the time of surgery, on the morphology and on bio-physical properties of the porta hepatis as well as on cholangitis episodes occurring after surgery. The fibrosis level of the Glisson's triads at the time of surgery, the morphology and bio-physical parameters of the porta hepatis are documented as causes for failing or discontinued bile flow after HPE and for the progression of fibrosis. In the absence of obvious causes it is postulated that the initially obliterating and continuing primary disease has progressed. The performance of the HPE is summarized in the results: of 108 children who were operated five and more years ago 62 are still alive, 58 are icterus-free, 25 show no progression of fibrosis and 46 of 82 patients who were initially cirrhosis-free have no cirrhosis at present.

Liver transplantation for primary biliary cirrhosis and primary sclerosing cholangitis: predicting outcomes with natural history models

In patients with primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), risk score models that reflect disease severity have been developed and can serve as an objective measurement to assess and evaluate the effect of the severity of liver disease on the outcome of liver transplantation. Thus, using the established Mayo risk scores for PBC and PSC, one not only can estimate survival for the individual patient but can measure disease activity as well. Indeed, several studies have suggested that the optimal timing of liver transplantation with use of the Mayo PBC model may be an important tool to improve survival, decrease morbidity, and decrease overall related costs. Likewise, studies in patients with PSC have yielded similar results. This review explores how prognostic mathematical survival models for PBC and PSC might be applied to individual patients in need of liver transplantation. The following question is addressed: How can the timing of liver transplantation be optimized to increase survival, decrease postoperative morbidity, and ultimately, decrease the overall resource utilization involved in this procedure?

Management of primary biliary cirrhosis and autoimmune cholangitis

Several drugs have been evaluated in the treatment of primary biliary cirrhosis over a number of years. These drugs have immunosuppressive, antiinflammatory, cupruretic, antifibrotic and bile acid properties. Ursodeoxycholic acid has been shown to improve survival free of transplantation in a conclusive fashion. This drug is the single agent that can be recommended for the treatment of primary biliary cirrhosis. Corticosteroid therapy and ursodeoxycholic acid have been evaluated in a few patients with autoimmune cholangitis. This article reviews a large number of studies that have been published assessing different drugs in the treatment of these two entities, particularly in the treatment of primary biliary cirrhosis.

Predictive models of natural history in primary biliary cirrhosis

Primary biliary cirrhosis is a slow, progressive disease. Although many years may elapse before asymptomatic primary biliary cirrhosis patients begin experiencing symptoms of liver disease, their overall survival is significantly lower than the normal population. The Mayo natural history model has been developed to depict patient survival in the absence of effective therapeutic intervention. Although there are a number of caveats in applying this model, it has been validated using external data sets and established as an accepted tool for clinical or research purposes. Furthermore, recent data suggest that the Mayo natural history model continues to provide useful, predictive information in the presence of ursodeoxycholic acid therapy, which has been shown to lower the serum bilirubin to the natural history model for patient survival. In addition to the natural history model for patient survival, mathematical models have been developed to describe histologic progression and development of esophageal varices.

A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazards model

Grønnesby and Borgan (1996) propose an overall goodness-of-fit test for the Cox proportional hazards model. The basis of their test is a grouping of subjects by their estimated risk score. We show that the Grønnesby and Borgan test is algebraically identical to one obtained from adding group indicator variables to the model and testing the hypothesis the coefficients of the group indicator variables are zero via the score test. Thus showing that the test can be calculated using existing software. We demonstrate that the table of observed and estimated expected number of events within each group of the risk score is a useful adjunct to the test to help identify potential problems in fit.

Targeting naproxen coupled to human serum albumin to nonparenchymal cells reduces endotoxin-induced mortality in rats with biliary cirrhosis

Endotoxin is thought to play a major role in cirrhotic liver disease. Cyclo-oxygenase inhibitors were shown to be partially protective against endotoxin but cannot be used in cirrhotic patients because of renal side-effects. We argued that administration of naproxen (NAP) linked to human serum albumin (HSA), which results in specific delivery of NAP to endothelial cells (EC) and Kupffer cells (KC) and exhibited hepatoprotective effects against lipopolysaccharide (LPS) in vitro, could protect cirrhotic rats from LPS toxicity while preserving renal function. The studies were performed in rats rendered cirrhotic by bile duct ligation (BDL); animals received LPS (Escherichia coli, 800 microg/kg) intravenously. Five groups were studied: LPS alone, rats pretreated with a conventional dose of NAP (50 mg/kg), NAP-HSA (22 mg/kg), NAP equimolar to NAP-HSA (1.5 mg/kg), or the HSA carrier. LPS induced significant mortality (55%); this was not affected by equimolar NAP (57%) but accentuated by conventional NAP (88%). In contrast, NAP-HSA provided significant protection (9%; P < .05). After conventional NAP treatment, significant renal toxicity was observed as evidenced by a marked reduction in sodium excretion (LPS vs. NAP-HSA vs. NAP [50 mg/kg] 33 +/- 22 vs. 50 +/- 39 vs. 4 +/- 3 micromol/h; P < .05). Renal prostaglandin E2 (PGE2) excretion was reduced by NAP in all groups, but most markedly at the conventional dosage (LPS vs. NAP-HSA vs. NAP [50 mg/kg] 132 +/- 115 vs. 39 +/- 19 vs. 9 +/- 8 ng/mL; P < .05). Successful targeting was evidenced by a significant hepatic enrichment of NAP in the NAP-HSA group compared with the equimolar untargeted group (30.16 +/- 9.33 vs. 1.13 +/- 1.95 nmol/g liver). Thus, targeting NAP to EC/KC results in improved survival, higher efficacy, and sparing of renal function in cirrhotic rats.

Monoethyl glycine xylidide (MEGX) test evaluation in primary biliary cirrhosis: comparison with Mayo score

OBJECTIVE

To evaluate the clinical and prognostic value of the monoethyl glycine xylidide (MEGX) test in patients with primary biliary cirrhosis (PBC) in comparison with the Mayo score (Mayo).

DESIGN

A prospective study.

METHODS

MEGX determinations at enrolment were compared to the Mayo score as well as to conventional clinical and laboratory parameters in 92 patients with PBC.

RESULTS

The MEGX test yielded higher basal values in long-term survivors compared to patients that were transplanted or died during the follow up; patients belonging to the last two groups displayed significantly higher Mayo scores at baseline. Although values for prothrombin time, serum albumin, alkaline phosphatase, cholesterol, cholinesterase, and gamma-glutamyltranspeptidase were significantly different in survivors compared to either transplanted or dead patients at univariate analysis, the multivariate analysis demonstrated an independent prognostic value for the MEGX and the Mayo score solely. The best discrimination between probability of death or survival was achieved with a cutoff value of 25 ng/ml for the MEGX test and of 6 for the Mayo score. When plotting both MEGX test and Mayo score, the point distribution displayed a bimodal trend, and the wide range of values given by the MEGX test was observed to supply a more precise assessment of liver reservoir and a better discrimination of progressive changes in liver function; the limited range of the Mayo score for values below 6 could only identify gross deteriorations.

CONCLUSION

Our data show that the asymptomatic progressive functional deterioration occurring during the natural history of PBC can be monitored by the MEGX test because it appears to be able to identify abnormalities prior to the onset of alterations in conventional laboratory and/or clinical parameters which are likely to affect the Mayo score.

Serial quantitative liver function tests in patients with primary biliary cirrhosis: a prospective long-term study

Primary biliary cirrhosis (PBC) is a rare chronic cholestatic disorder of unknown origin that can now be treated effectively with ursodeoxycholic acid (UDCA). The clinical course of PBC is very variable, but a significant proportion of patients eventually die or undergo liver transplantation. In this single-center prospective long-term study, we analyzed the effect of UDCA therapy (10 mg/kg b.w./day) on conventional liver function tests and we also investigated whether serial quantitative liver function tests are useful in the clinical management of patients with PBC. Fifteen patients, most of them in an early disease stage, were followed up for either 4 (n = 7) or 5 (n = 8) years. In addition to regular conventional liver function tests, every 12 months quantitative liver function tests were performed. Thus we measured galactose elimination capacity, indocyanine green half-life and lidocaine half-life. Quantitative liver function tests were also performed once in healthy volunteers. Treatment with UDCA significantly improved conventional liver function tests, and this effect was maintained for several years (values in U/l before therapy and 4 years after therapy: AP = 1,346 +/- 317 vs. 516 +/- 93; gammaGT 378 +/- 80 vs. 144 +/- 30; LAP 122 +/- 10 vs. 71 +/- 9; AST 61 +/- 19 vs. 34 +/- 12; ALT 90 +/- 19 vs. 68 +/- 35; GLDH 14.3 +/- 1.9 vs. 8.2 +/- 1.9). Quantitative liver function tests were not significantly different between healthy volunteers and patients (GEC 6.8 +/- 0.3 vs. 7.0 +/- 0.3 mg/kg x min; ICG half-life 4.2 +/- 0.4 vs. 3.7 +/- 0.3 min; lidocaine half-life 75 +/- 8 vs. 79 +/- 6 min). In the patients, results of quantitative liver function tests (GEC, ICG and lidocaine half-lives) were not affected by UDCA therapy and remained constant over time. In the 1 patient who was transplanted, serial quantitative liver function tests did not indicate deteriorating liver function earlier than the patient's progressive symptoms or conventional liver function tests. Thus UDCA therapy markedly improved conventional liver function tests in patients with PBC, and this effect was maintained for at least 4-5 years. Possibly due to the fact that most of the patients were in an early disease stage, serial quantitative liver function tests provided little additional information that was relevant for planning therapy in the individual patient.

Combined analysis of randomized controlled trials of ursodeoxycholic acid in primary biliary cirrhosis

BACKGROUND & AIMS

Long-term ursodeoxycholic acid (UDCA) therapy slows the progression of primary biliary cirrhosis. This study examined the effect of UDCA therapy on survival free of liver transplantation in a large group of patients.

METHODS

Data from three clinical trials were combined in which patients with primary biliary cirrhosis were randomly assigned to receive UDCA (n = 273) or placebo (n = 275). After 2 years, patients from French and Canadian studies received UDCA for up to 2 years. Patients from the American study remained on their assigned treatment for up to 4 years.

RESULTS

Survival free of liver transplantation was significantly improved in the patients treated with UDCA compared with the patients originally assigned to placebo (P < 0.001; relative risk, 1.9; 95% confidence interval, 1.3-2.8). Subgroup analyses showed that survival free of liver transplantation was significantly improved in medium- and high-risk groups (serum bilirubin level, 1.4 to 3.5 or > 3.5 mg/dL; P < 0.0001 and P < 0.03, respectively) and histological stage IV subgroup (P < 0.01).

CONCLUSIONS

Long-term UDCA therapy improves survival free of liver transplantation in patients with moderate or severe disease. An effect in patients with mild disease is probably not found because they do not progress to end-stage disease in 4 years.

Survival probabilities of Pugh-Child-PBC classified patients in the euricterus primary biliary cirrhosis population, based on the Mayo clinic prognostic model. Euricterus Project Management Group

BACKGROUND/AIMS

Estimation of prognosis becomes increasingly important in primary biliary cirrhosis (PBC) with advancing disease and also with regard to patient management. The ubiquitous used Pugh scoring for severity of disease is simple while the Mayo model which has been validated for survival estimates is more sophisticated. We wanted to investigate whether Pugh and Mayo scores correlate (they have 3 of 5 variables in common) and if so whether a survival probability based on Mayo data could be affixed on Pugh classes and scores obtained in the same patients.

METHODOLOGY

All variables used for Mayo Clinic Prognostic Model (Mayo) scoring and Pugh-Child-PBC (Pugh) scoring were available in 143 PBC patients of the Pan European database Euricterus. Pugh scores P5-P15 and has classes A (P5-6), B (P7-9) and C (P10-15). We subdivided P5 in P5A (patients with albumin > 40 g/l plus prothrombin time < or = 12 secs) and P5B (the other patients in P5). We designed a category Pugh Early (PE) for patients with P5A characteristics and bilirubin < 17 mmol/l. Mayo scores R0-R15-with 1-7 years survival probabilities S-and has risk classes Low (L), Intermediate (Int), High (H) and Very High (VH).

RESULTS

The estimated survival probabilities of the 143 patients ranged from 88% at 7 years to 0% at 1 year, median 14% at 5 years. The Pugh and Mayo scores correlated r = 0.87 (p < 0.0001) and except age with P, all Mayo and Pugh variables correlated with both R and P at p < 0.0001. Survival in Pugh class A was median 43% at 7 years and was not different from survival in Mayo L+Int (p 0.58). In Pugh class B 7 years survival was 2%, not different from Mayo H (p 0.25). Survival in Pugh C was median 24% at 1 years and better than Mayo VH (p 0.02). Between P5A (survival 78% at 7 yr) and R 3-4; P5B-6 (40% at 7 yr) and R5; P7 (22% at 7 yr) and R6; P8-11 (12% at 5 yr) and R7-8; and P12-14 (5% at 1 yr) and R9-10 no significant differences were found. From P8 upward there was a steep increase in death rate. PE has a 7 year survival of at least 89%. Charts of projected survival estimates for Pugh scores and classes are presented.

CONCLUSION

It was possible (affixing Mayo to Pugh) to define 1-7 years survival probabilities to Pugh classes and scores for the last 7 years of the disease, i.e. the most important period for therapeutic decisions. These results need to be validated in other PBC populations.

Antimitochondrial antibody profiles in primary biliary cirrhosis distinguish at early stages between a benign and a progressive course: a prospective study on 200 patients followed for 10 years

In recent retrospective studies, it was shown that subtypes of antimitochondrial antibodies (AMA) can help to discriminate between a benign [only anti-M9 and/or anti-M2 positive by enzyme-linked immunosorbent assay (ELISA)] and a rather progressive course (anti-M2, -M4 and/or -M8 positive). According to different constellations of these AMA subspecificities in ELISA and complement fixation test (CFT), four AMA profiles (A-D) were defined. In 1984 we started a prospective study based on 200 PBC patients with known AMA profiles in order to correlate the antibody pattern with the clinical outcome. Progression was defined primarily as the necessity of liver transplantation and death due to hepatic failure or variceal bleeding. At entry, 18 (9%) of the 200 patients had AMA profile A (only anti-M9), 57 (29%) profile B (only anti-M2 with or without anti-M9), 74 (37%) profile C (anti-M2 in association with anti-M4/-M8 by ELISA), and 51 (26%) profile D (anti-M2/-M4/-M8 by ELISA and CFT). At the beginning of the study, 177 patients had PBC stage I/II. During the observation period of ten years, ten patients died and in 18 orthotopic liver transplantation (OLT) was performed; all these patients belonged to profile C/D. Furthermore, 44% of the patients with profile C and 31% of the patients with profile D progressed to late stages, as defined by histology and clinical manifestations such as portal hypertension and increase of bilirubin, while only one of the patients with profile B and none of the profile A-patients developed late stage PBC. A significant increase of bilirubin was observed only in C/D-patients. AMA profiles did not change during the follow-up. In conclusion, AMA profiles discriminate between a benign and a progressive course of PBC already at early stages.

Primary biliary cirrhosis: Dutch application of the Mayo Model before and after orthotopic liver transplantation

BACKGROUND/AIMS

A retrospective study of primary biliary cirrhosis (PBC) was performed to study the Original Mayo Model for predicting survival by a Dutch data-set of patients, presentation of disease progression; assessment of liver transplantation, prediction of post-transplantation survival; and the addition of two laboratory variables to the Original Mayo Model.

MATERIALS AND METHODS

Survival of 83 patients, 37 of whom underwent transplantation, were studied. Mean follow-up was 6.0 +/- 0.45 SEM years. Risk score at diagnosis, platelet count, and serum sodium were analyzed in a Cox model.

RESULTS

The Original Mayo Model estimated survival for low-, medium-, and high-risk groups accurately and it also presented disease progression. Baseline Mayo risk score in a Cox model had a regression coefficient of 1.01, indicating an excellent predictor p < 0.0001. Platelet count was a predictor of survival (p < 0.002), whereas serum sodium did not (p = 0.67). A new model combined of the Original Mayo risk score and platelet count predicted survival in high-risk patients somewhat better compared to the Original Mayo Model. With both models, liver transplantation had a significant beneficial effect on survival (p < 0.001). The scores revealed no significant influence (p = 0.47) for overall post-transplantation survival.

CONCLUSIONS

The Original Mayo Model remains the model of choice for patients with PBC for prognostication from 3-8 years, is a useful tool in the assessment of liver transplantation but not an indicator of post-transplantation survival. Platelet count showed to have additional prognostic value. A new model combined of platelet count and the Original Mayo risk score did predict survival in high-risk groups slightly better compared to the Original Mayo Model.

Transplantation for primary biliary cirrhosis

Primary biliary cirrhosis (PBC) remains one of the commoner indications for orthotopic liver replacement. The two main indications for transplantation are poor quality of life (because of the liver) or end-stage liver disease. A number of prognostic models have identified risk factors indicating poor prognosis, but in practice serum bilirubin greater than 150 mumol/L is used most commonly. Other indications for transplantation include progression of hepatopulmonary syndrome, increasing osteoporosis, evidence of malnutrition, and development of hepatocellular carcinoma. Postoperatively, patients do well. Recurrence of PBC remains controversial, but an increasing number of centers now report that a proportion of patients develop evidence of recurrent disease in the allograft. As yet PBC recurrence remains of little practical importance, although as survival increases beyond 10 years, this may become more relevant.

Survival algorithms and outcome analysis in primary biliary cirrhosis

The natural history of primary biliary cirrhosis (PBC) is one of slowly progressive cholestasis with liver damage, development of cirrhosis with its concomitant complications, and death unless the patient undergoes liver transplantation. Natural history studies have identified several variables associated with a decreased survival in patients with PBC. The course of the disease can be divided into three time periods: (1) a presymptomatic phase, probably lasting up to 20 years; (2) a symptomatic phase, with anicteric or mild jaundice, lasting up to 5 to 10 years; and (3) a preterminal or accelerated phase with marked jaundice, lasting up to 2 years. Since the course of the disease is one of slow progression leading to liver failure and death unless liver transplantation intervenes, several investigators have developed statistical models to predict survival. The ability to predict survival for individual patients with PBC has been valuable in the management of these patients, particularly in patient selection and timing of liver transplantation. In addition, survival estimates can be utilized in education and counseling patients and their families. These models may also be used to evaluate the efficacy of new treatments by comparing natural history survival with the survival achieved by therapeutic effect. Over the past several decades, the natural history models of PBC have been developed in the absence of effective medical therapy. The efficacy of liver transplantation and survival following liver transplantation has now been quantitatively established. Future efforts should be aimed at determining not only survival of patients with primary biliary cirrhosis in the presence of effective medical therapy but also at assessing the quality of life and cost-effectiveness of medical therapy and liver transplantation in the management of patients with primary PBC.

Use of Irwin's restricted mean as an index for comparing survival in different treatment groups--interpretation and power considerations

In the analysis of survival data from clinical trials and other studies, the censoring generally precludes estimation of the mean survival time. To accommodate censoring, Irwin (1949) proposed, as an alternative, estimation of the mean lifetime restricted to a suitably chosen time T. In this article we consider the use of Irwin's restricted mean as an index for comparing survival in different groups, using as an example published data from a randomized clinical trial in patients with primary biliary cirrhosis. Irwin's method, originally based on the actuarial survival estimator, is extended to incorporate covariates into the analysis through the use of piecewise exponential models. For comparing two survival curves, the logrank test is known to be optimal under proportional hazards alternatives. However, comparison of restricted means may outperform the logrank test in situations involving nonproportional hazard rates. We examine the size and power of these two procedures under various proportional and nonproportional hazards alternatives, with and without covariate adjustment. For survival curves that separate early in time the censored data generalization of the Wilcoxon test is known to exhibit high power, and we examine how the comparison of restricted means performs relative to this procedure also.

Identification of an autoantibody against alpha-enolase in primary biliary cirrhosis

BACKGROUND

Primary biliary cirrhosis is a chronic cholestatic liver disease in which autoreactive T cells may play an important role in the destruction of intrahepatic bile ducts. However, target antigens remain unknown. Alpha-enolase-derived peptide binds to human leukocyte antigen (HLA)-DR8, which is implicated in the development of primary biliary cirrhosis in Japanese patients. Partial homology between alpha-enolase and the inner lipoyl domain of E2 component of pyruvate dehydrogenase (PDH-E2) is also observed.

METHODS

Using alpha, beta and gamma enolase isozymes obtained from humans and/or rabbits, we examined serum samples of 56 patients with primary biliary cirrhosis, 19 autoimmune hepatitis, 38 acute and chronic viral hepatitis and 36 healthy subjects by immunoblotting.

RESULTS

Anti-alpha-enolase antibody was present in a significantly higher percentage of patients with primary biliary cirrhosis (16 of 56, 28.6%) and autoimmune hepatitis (6 of 19, 31.6%) than in normal subjects (p<0.005, p<0.01, respectively). Antibodies against beta and gamma-enolases were not detected in any serum sample. Although there was no significant correlation between the presence of anti-alpha-enolase antibody and clinical features of primary biliary cirrhosis, the mortality rate associated with hepatic failure in patients with positive autoantibody was significantly higher than that of antibody-negative PBC patients (6 of 16, 37.5% vs 5 of 40, 12.5%, p<0.05).

CONCLUSIONS

Since alpha-enolase is expressed on the cell surface, our data suggest that the immunological reaction to alpha-enolase might be involved in biliary epithelial destruction and be relevant to the disease progression.

Primary biliary cirrhosis in The Netherlands. An analysis of associated diseases, cardiovascular risk, and malignancies on the basis of mortality figures

BACKGROUND/METHODS

In 1979 death rate registration for primary biliary cirrhosis (PBC) became available in The Netherlands. In the 14-year period 1979-92, 417 persons died of and 179 with PBC. We investigated secondary causes of death using standardized mortality ratios (SMR) (1.0 as reference, P < 0.001 regarded as significant).

RESULTS

Median age was 70-74 (35 to > 85) years. Secondary causes of death originated from the circulatory, digestive, and respiratory tracts and malignancies. Younger persons (< 60 years), dying of PBC, more often died with "toxicity related to immunosuppression' than older persons (P < 0.01). Younger persons (< 60) dying with PBC, more often died of hepatocellular carcinoma (HCC) than older ones (P < 0.05). In patients with PBC the frequency of HCC (SMR, 25.5; P < 0.0001) and diseases of the musculoskeletal system/connective tissue (SMR, 5.1; P < 0.0001) was higher than in the general population. Malignancies in general (SMR, 0.7), pancreatic carcinoma (SMR, 2.5), breast cancer (SMR, 0.1) and diseases of the circulatory system (SMR, 0.8) differed but not significantly (P < 0.05 - < 0.01). No difference existed in the localization of malignancies in patients dying of as compared with those dying with PBC.

CONCLUSIONS

Deaths occurred predominantly in the older age classes, with an age-related difference in some associated disorders. Patients with PBC showed an increased risk of HCC and diseases of the musculoskeletal system. Similar studies from different countries are needed.

Clinical significance of different antinuclear antibodies patterns in the course of primary biliary cirrhosis

BACKGROUND/AIMS

The significance of antinuclear antibodies (ANA) in primary biliary cirrhosis (PBC) patient is still controversial in the literature. The purpose of this paper is to investigate the clinical significance of ANA in PBC patients.

MATERIALS AND METHODS

Sixty-nine patients with PBC were investigated. Control groups included 21 patients with autoimmune hepatitis, 26 patients with alcoholic liver disease, 13 patients with systemic connective tissue disease and 27 healthy persons. ANA was detected by an immunofluorescence method on rat liver tissue sections and HEp-2 cells at serum dilution 1/40.

RESULTS

In 48 out of 69 PBC patients (70%), ANA was positive in HEp-2 cell line, but in rat liver tissue sections only 29% of patients had positive ANA reactions. Most frequent patterns were multiple nuclear dots (MND) in 42% and perinuclear in 16%. MND-ANA was also found in two autoimmune hepatitis patients and in one systemic lupus erythematosus patient. Survival from the moment of developing first symptom(s) attributable to liver disease was longer in the ANA positive patients than ANA negative ones (p < 0.02). Despite immunosuppressive treatment, in most of ANA positive patients (73%) ANA did not disappear. Most frequent ANA patterns in autoimmune hepatitis and systemic connective tissue diseases patients were homogeneous and anticentromere, respectively.

CONCLUSIONS

Immunofluorescence method on HEp-2 cell line for ANA detection is more sensitive than on rat liver tissue sections. In PBC patient's incidence of ANA, especially MND-ANA is a frequent immunological abnormality. ANA positive patients have better prognosis for survival.

Primary biliary cirrhosis (PBC) in an European country--a description of death rates in The Netherlands (1979-1992)

BACKGROUND/AIMS

In 1979, separate liver transplantation (LT) and primary biliary cirrhosis (PBC) death rate registration became available in The Netherlands (15 million inhabitants). The objective of this study was to investigate death rates from 1979-1992 and analyse the impact of LT.

PATIENTS AND METHODS

PBC was either a primary or secondary cause of death. Rates were expressed as absolute numbers or per million inhabitants in the corresponding age category. Age classes of 5 years were used. The Netherlands was divided in four regions, North, South, East and West. Standardized mortality ratios (SMR) were used for calculation of regional differences.

RESULTS

In the 14 year period between 1979-1992, 417 persons died from and 179 persons died with PBC, totaling 596 PBC patients (6.3 per million inhabitants > or = 35 years). No person younger than 35 died. Eighty-two percent were female, with a corresponding female/male ratio of 4.2 per million females/males inhabitants. In region South there were significantly fewer deaths (SMR 66%, p < 0.001) and in region North significantly more (SMR 141%, p < 0.05). The median age class at death was 70-74 (males and females alike). At age 35-59, death from PBC in 1992 per million was 1.2, and for > or = 65 years 15.7. In age class 80-84, the highest death rate from or with PBC was found with 28 deaths per million inhabitants and with a female/male ratio of 3.6. In 1992, with two deaths only, LT appeared to have nearly eliminated death from PBC in the age category 35-59 years.

CONCLUSION

Death from PBC mainly occurs in the old and very old, who may never seek a specialized center. This indicates a more specific management and therapy for this particular group is needed.

[Primary biliary cirrhosis: clinical features and survival of a Mexican population]

BACKGROUND

PBC progresses to cirrhosis and results in death due to liver failure or bleeding portal hypertension. Data of the clinical characteristics and survival of PBC patients allows the assessment of therapeutical alternatives as well as the establishment of inclusion criteria for liver transplantation.

AIMS

One hundred and twenty patients with histological diagnosis of PBC, admitted from 1972 to 1992, were selected with the purpose of studying the clinical and biochemical characteristics and survival.

METHODS

Patients who underwent liver transplant or those who had an incomplete follow-up were excluded.

RESULTS

Therefore only 80 patients were included: these were seventy five women and five men, with mean age 46 +/- 11 years (X +/- SD) to whom demographic data, biochemical analysis, liver function (Child-Pugh) and liver damage (Ludwig) were recorded at the time of histological diagnosis, which was considered zero for calculating the survival (Kaplan Meier). The most common symptoms at diagnosis were pruritus in 63 patients, jaundice in 48, asthenia and adynamia in 55 patients. Eight cases were asymptomatic. According to Child-Pugh's classification, patients were grouped as follows: forty in stage A, 29 in B, and three in C; and according to liver damage (Ludwig), 8 in grade I, 28 in grade II, 22 in grade III and 14 in grade IV. The most frequent clinical associations were Sjögren's syndrome, in 30% of patients, although one case was associated to progressive muscular dystrophy and another one to multiple myeloma and hypothyroidism; in 58.7% of the cases, antimitochondrial antibodies were negative. One year survival was 75%, five years 44%, and seven years 13%.

CONCLUSIONS

The most important characteristics of the studied patients were elevated percentage of negative antimitochondrial antibodies and short survival. it is important to impel the development of liver transplantation as the only mean to improve survival.

Is the Mayo model for predicting survival useful after the introduction of ursodeoxycholic acid treatment for primary biliary cirrhosis?

Treatment of patients with primary biliary cirrhosis (PBC) using ursodeoxycholic acid (UDCA) leads to a reduction in serum bilirubin. The first objective of this study was to assess the performance of certain prognostic indicators for PBC after the introduction of treatment with UDCA. Serum bilirubin is an important prognostic indicator for PBC and an important component of the Mayo model for grading patients into risk categories. In an analysis of patients enrolled in the Canadian multicenter trial, the Mayo score was calculated before and after treatment with UDCA. After treatment, the Mayo score continued to divide patients with PBC into groups with varying risk. In addition, the serum bilirubin alone was shown to do the same even after the introduction of treatment with UDCA. A second objective was to establish whether UDCA had an effect on long-term (2- to 6-year) survival in patients with PBC.

Effects of ursodeoxycholic acid on survival in patients with primary biliary cirrhosis

BACKGROUND & AIMS

Ursodeoxycholic acid (UDCA) has been shown to be a safe and effective treatment for patients with primary biliary cirrhosis; however, its effect on patient survival is less certain. To study this issue, the survival of patients receiving long-term UDCA treatment was compared with that of a control group, adjusting for their risk scores based on the Mayo model.

METHODS

One hundred eighty patients were randomized to receive either 13-15 mg.kg-1.day-1 UDCA (n = 89) of placebo (n = 91). After the study closure, the patients originally receiving placebo were switched to active drug, and prospective follow-up was continued for 3 years. Patients were censored at the time of transplantation, voluntary withdrawal, of crossover of the placebo group (efficacy analysis). The survival of the two groups was adjusted for risk scores at the time of entry to the study. A secondary analysis was an intent-to-treat analysis, whereby patients were followed up regardless of their voluntary withdrawal or crossover.

RESULTS

At the time of analysis, the patients receiving placebo had a significantly increased risk of death and/or requiring transplantation (relative risk, 2.6; P = 0.04) compared with the UDCA-treated patients

CONCLUSIONS

UDCA should be considered as a safe, effective, and life-extending treatment for patients with primary biliary cirrhosis.

A method for checking regression models in survival analysis based on the risk score

We propose to perform model check for the Cox and Aalen regression models using martingale residual processes grouped after the risk score. Asymptotic distributions of the grouped martingale residual processes are deduced, so both formal and graphical model check can be performed. The method is validated by stochastic simulation. A data example with patients with primary biliary cirrhosis of the liver is discussed.

Multilevel models for longitudinal variables prognostic for survival

The issue of modelling the joint distribution of survival time and of prognostic variables measured periodically has recently become of interest in the AIDS literature but is of relevance in other applications. The focus of this paper is on clinical trials where follow-up measurements of potentially prognostic variables are often collected but not routinely used. These measurements can be used to study the biological evolution of the disease of interest; in particular the effect of an active treatment can be examined by comparing the time profiles of patients in the active and placebo group. It is proposed to use multilevel regression analysis to model the individual repeated observations as function of time and, possibly, treatment. To address the problem of informative drop-out--which may arise if deaths (or any other censoring events) are related to the unobserved values of the prognostic variables--we analyse sequentially overlapping portions of the follow-up information. An example arising from a randomized clinical trial for the treatment of primary biliary cirrhosis is examined in detail.

Diagnostic plots to reveal functional form for covariates in multiplicative intensity models

We show how plots based on the residuals from a proportional hazards model may be used to reveal the correct functional form for covariates in the model. A smoothed plot of the martingale residues was suggested for this purpose by Therneau, Grambsch, and Fleming (1990, Biometrika 77, 147-160); however, its consistency required that the covariates be independent. They also noted that the plot could be biased for large covariate effects. We introduce two refinements which overcome these difficulties. The first is based on a ratio of scatter plot smooths, where the numerator is the smooth of the observed count plotted against the covariate, and the denominator is a smooth of the expected count. This is related to the Arjas goodness-of-fit plot (1988, Journal of the American Statistical Association 83, 204-212). The second technique smooths the martingale residuals divided by the expected count, using expected count as a weight. This latter approach is related to a GLM partial residual plot, as well as to the iterative methods of Hastie and Tibshirani (1990, Biometrics 46, 1005-1016) and Gentleman and Crowley (1991, Biometrics 47, 1283-1296). Applications to survival data sets are given.

Primary biliary cirrhosis in Estonia. With special reference to incidence, prevalence, clinical features, and outcome

BACKGROUND

Primary biliary cirrhosis (PBC) is a liver disease of unknown etiology, whose occurrence varies greatly between different regions. For a long time there have been no published data about the incidence and prevalence of PBC from Eastern Europe countries.

METHODS

The incidence and prevalence of PBC have been investigated in the Estonian population during the period 1973-92. Two sources of information were used: an information circular/questionnaire was sent to all district hospitals and gastroenterologists, and the case histories of all patients with a positive antimitochondrial antibody titer of 1:40 or more were reexamined.

RESULTS

During this period 69 cases of PBC were diagnosed. The male to female ratio was 1:22; 13% of the patients were asymptomatic. The mean annual incidence was 2.27 per million, and on 31 December 1992 the point prevalence was 26.9 per million. There were differences in prevalence among the various districts of Estonia. Associated autoimmune conditions were reported in 32% of the patients. Mean survival from the time of diagnosis was 52.5 months.

CONCLUSIONS

The incidence of PBC in Estonia is at the lower end of the range reported in the world literature. This has probably partly been caused by a low percentage of asymptomatic and male patients.

Prediction of prognosis of primary biliary cirrhosis in Japan

The clinical profile of primary biliary cirrhosis in Japan was clarified on the basis of data on 1066 patients attending 212 hospitals and institutions in this country. Six hundred and twelve patients (57.4%) were asymptomatic. The majority of the patients were middle-aged women. Pruritus was the most frequent initial symptom of symptomatic primary biliary cirrhosis. Antimitochondrial antibodies were positive in 877 patients (82.5%). Sjögren's syndrome was the most common associated autoimmune disease. Liver biopsy was performed in 753 patients at the time of diagnosis, and histological staging by Scheuer's classification indicated that 307 (43.7%) patients were in stage I and 222 (31.6%) were in stage II. The most frequent causes of death were hepatic failure and/or gastrointestinal bleeding, which affected 166 (78.3%) of the 212 patients who died. Statistical analysis using Cox's regression method revealed that the patient's age and the serum bilirubin, albumin, and total cholesterol concentrations were significant predictors of the prognosis. A prognostic index was also calculated that could be used to predict the duration of survival for patients with primary biliary cirrhosis.

Historical controls and modern survival analysis

Comparison of observed mortality with 'known,' 'background,' or 'standard' rates has taken place for several hundred years. With the developments of regression models for survival data, an increasing interest has arisen in individualizing the standardisation using covariates of each individual. Also, account sometimes needs to be taken of random variation in the standard group. Emphasizing uses of the Cox regression model, this paper surveys a number of critical choices and pitfalls in this area. The methods are illustrated by comparing survival of liver patients after transplantation with survival after conservative treatment.

Effect of distal splenorenal shunt on survival of patients with primary biliary cirrhosis

Distal splenorenal shunt is known to effectively control bleeding from esophageal and gastric varices; however, the effect of this selective shunt on liver function is less well understood. We examined retrospectively the effect of distal splenorenal shunt on the survival of 19 patients with primary biliary cirrhosis subjected to surgery for bleeding varices over a 20-yr period and had been followed for at least 1 yr. Actual Kaplan-Meier survival curve was compared with predicted survival curve based on the Mayo Clinic model using clinical data collected at the time of surgery. The patients median length of follow-up was 65.9 mo. Ten of the 19 patients died or underwent orthotopic liver transplantation during the period of observation. The actual Kaplan-Meier and predicted Mayo Clinic model survival curves were similar and did not differ significantly. Survival was best in patients with good liver function (i.e., low Mayo risk scores). Distal splenorenal shunt, therefore, did not appear to have an adverse effect on the survival of patients with primary biliary cirrhosis. We conclude that variceal bleeding in primary biliary cirrhosis patients with good liver function should not be considered an indication for liver transplantation. Instead, if treatment with sclerotherapy or beta-blockers fails then distal splenorenal shunt will prevent recurrent bleeding in 90% of patients and leave them with an excellent prognosis.

Primary biliary cirrhosis: prediction of short-term survival based on repeated patient visits

The progression of primary biliary cirrhosis was studied in 312 patients who were seen at the Mayo Clinic between January 1974 and May 1984. Follow-up was extended to April 30, 1988, by which time 140 of the patients had died and 29 had undergone orthotopic liver transplantation. These patients generated 1,945 patient visits that enabled us to study the change in the prognostic variables of primary biliary cirrhosis (age, bilirubin value, albumin value, prothrombin time and edema) from the time of referral. Also, using this database and the Cox proportional-hazards regression model, we developed an updated model for primary biliary cirrhosis that can be used to predict short-term survival at any time in the course of the disease. This model uses the values of the prognostic variables measured at the latest patient visit. Comparison of predicted survival from the update model and the natural history model of primary biliary cirrhosis showed that the updated model was superior to the original model for predicting short-term survival. This finding applied to both the Mayo Clinic patients and an independent set of 83 Dutch patients. The Mayo updated model is recommended for improving the accuracy of predictions of survival during the 2 yr after a patient visit.