Liver Failure/Cirrhosis/Portal Hypertension
Decreased survival of subjects with elevated liver function tests during a 28-year follow-up†
Article first published online: 14 SEP 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 51, Issue 2, pages 595–602, February 2010
How to Cite
Söderberg, C., Stål, P., Askling, J., Glaumann, H., Lindberg, G., Marmur, J. and Hultcrantz, R. (2010), Decreased survival of subjects with elevated liver function tests during a 28-year follow-up. Hepatology, 51: 595–602. doi: 10.1002/hep.23314
Potential conflict of interest: Nothing to report.
- Issue published online: 25 JAN 2010
- Article first published online: 14 SEP 2009
- Accepted manuscript online: 14 SEP 2009 12:00AM EST
- Manuscript Accepted: 3 SEP 2009
- Manuscript Received: 5 DEC 2008
- Swedish Research Council. Grant Number: 9127
- Bengt Ihre Foundation
- Karolinska Institutet
The long-term survival of subjects with nonalcoholic fatty liver disease (NAFLD) in comparison with both individuals with elevated transaminases attributable to other causes and the general poulation is poorly characterized. This study was undertaken to determine the frequency of NAFLD in a cohort of subjects who underwent liver biopsy from 1980 to 1984 because of elevated liver enzymes, and to assess mortality among subjects with NAFLD in comparison with the general Swedish population. The 256 subjects (61% men) had a mean age of 45 ± 12 years at the inclusion. Liver biopsies were blindly scored for NAFLD and nonalcoholic steatohepatitis (NASH). Causes of death were ascertained from the national Swedish Cause of Death Registry. Fatty liver was detected in 143 of the 256 subjects, including 25 (10%) with alcoholic fatty liver disease and 118 (46%) exhibiting NAFLD. Of those, 51 (20%) were classified as NASH and 67 (26%) as nonalcoholic bland steatosis. Cirrhosis was present in 9% at inclusion. During the follow-up period, 113 (44%) of the total population and 47 (40%) of the 118 subjects diagnosed with NAFLD died. Of the 113 deaths, 37 were of cardiovascular disease and 16 of liver diseases. Compared with the total Swedish population, adjusted for sex, age, and calendar period, subjects with NAFLD exhibited a 69% increased mortality (standardized mortality ratio [SMR] = 1.69; 95% confidence interval [CI], 1.24–2.25); subjects with bland steatosis, a 55% increase (SMR, 1.55; 95% CI, 0.98–2.32; P = 0.062); and subjects with NASH, 86% (SMR, 1.86; 95% CI, 1.19–2.76; P = 0.007). Conclusion: Patients with NASH are at increased risk of death compared with the general population. Liver disease is the third most common cause of death among patients with NAFLD. (HEPATOLOGY 2009.)
Although nonalcoholic fatty liver disease (NAFLD) is the most common cause of elevated serum levels of liver enzymes in the Western world, the long-term outcome of this condition is poorly characterized. In the early 1980s, increased determination of aminotransferase levels in connection with health surveys and screening programs led to improved detection of individuals with pathological liver function. Among adults, the most common abnormalities observed in the absence of symptoms are an elevated level of alanine aminotransferase (ALT) or gamma-glutamyltransferase activity. ALT levels are elevated in 2.8% of the general population,1 and in approximately 10% of these cases, no cause for this chronic hypertransaminasemia can be identified. The prognosis in connection with this condition remains unknown.2, 3
In two studies performed in the early 1980s, we found that 56% of asymptomatic subjects with elevated serum levels of hepatic enzymes who had undergone liver biopsy had fatty liver.4, 5 Nonalcoholic steatohepatitis (NASH) had not been characterized as an important entity at that time.6, 7 Recent evidence shows that some individuals with NAFLD have or develop NASH, which may lead to cirrhosis and an elevated risk for hepatocellular carcinoma (HCC).8
The long-term prognosis for individuals with NAFLD and NASH has been investigated in population-based studies9, 10 as well as in a cohort study in which NAFLD was considerd by biopsy,11 with a longest follow-up period so far of 14 years. Although the overall survival in connection with NAFLD was found to be slightly decreased,9 this reduction has been attributed to enhanced mortality, specifically among the subpopulation suffering from NASH,11 in which case bland steatosis might not alter the risk of death. The longer-term survival for subjects with NAFLD (with and without steatohepatitis) in comparison with both those with elevated serum levels of transaminases from other causes and the general poulation is thus incompletely characterized.
The aim of the current investigation was to examine the mortality and causes of death in a cohort of subjects with elevated serum levels of aminotransaminases. Furthermore, we wanted to determine the frequency of NAFLD and NASH in this population and compare the survival and causes of death in NAFLD subjects of those subjects with other liver diseases, and of the general population.
Patients and Methods
Between 1980 and 1984, 232 subjects with unexplained elevated serum levels of ALT, and therefore referred to our unit, have been characterized in a retrospective (n = 149) and a prospective (n = 83) study by Hultcrantz and coworkers.4, 5 Twenty-four additional subjects excluded from this previous prospective study because of a lack of radiological data were also included in the current follow-up study, giving 256 subjects altogether. The inclusion criteria were persistently elevated levels of aspartate aminotransferase and alanine aminotransferase (ALT) for longer than 6 months.
Subjects with symptoms or clinical signs of liver disease were excluded, as were those with serum levels of alkaline phosphate (greater than twice the upper normal limit, that is, >4.2 μkat/L) and those exhibiting clinical or laboratory signs of kidney disease. The mean age at the time of liver biopsy was 48.5 years for the men and 48.3 years for the women in the retrospective study, and the corresponding ages in the prospective study were 41 and 42 years, respectively (Table 1).
|Diagnosis Based on Biopsy||N (Total 256)||Patients with Cirrhosis (Total 23)||Men/Women||ALP (nv < 4.6)||AST (nv > 0.7)||ALT (nv > 0.7)|
Unless otherwise stated in the medical chart by the two physicians (and co-authors R.H. and G.L.), the patient was assumed not to overconsume alcohol. A great deal of effort was put into uncovering any such overconsumption at the time.
As noted in the medical records, testing for hepatitis C virus had been performed on 70 of our subjects, and 37 were positive. Of the 25 subjects with NAFLD who were tested, all were negative. Of the 47 in the latter group, 16 died before 1991, when such testing became generally available in Sweden. Of these 47 deceased subjects with NAFLD, two had been tested and were found to be negative for hepatitis C virus. No one in this group had any medical history of intravenous drug abuse or blood transfusions; nor did any exhibit the liver inflammation with portal infiltrates indicative of hepatitis C virus infection.
Follow-up and Identification as Still Living or Deceased.
All Swedish residents are assigned a unique 10-digit national registration number, and these identification numbers are recorded in the nationwide and virtually complete Cause of Death Registry.12 Through this registry, information concerning all deaths during the study period (1980 to July 9, 2008), including dates and causes of death (coded according to the International Classifications of Diseases versions 8, 9, and 1013–15) could be obtained. Similarly, through the national and continuously updated Population Registry, individuals still alive and residing in Sweden could be identified, so that the follow-up was complete.
Reevaluation of the Initial Liver Biopsies.
On an outpatient basis, liver biopsies had been performed on all of the subjects percutaneously with a 1.6-mm Menghini-type needle. All of these biopsies were reevaluated employing a modern classification by two of the authors (C.S. and R.H.), as well as by a third reference person (H.G.), all of whom were blinded to the patient details. Liver histology was scored in accordance with the system developed by Kleiner and Brunt et al.16 A classification of NASH was made on the basis of steatosis, lobular inflammation, and ballooning degeneration, and fibrosis was scored to determine the stage (progression) of this condition.
To assess the relative risk of death, we employed standardized mortality ratios (SMR), that is, the ratio between the observed numbers of deaths in the cohort compared with the number expected on the basis of mortality rates for the general population. The expected numbers of deaths were calculated by adding all person-years accumulated in the cohort into strata (sex, age [in 5-year groups) and calendar year of follow-up [in 5-year intervals]) and then multiplying the stratum-specific person-years by the corresponding stratum-specific incidence rates for the entire Swedish population. Ninety-five percent confidence intervals (CI) were calculated assuming that the observed events followed a Poisson distribution. Follow-up began at the date of the initial liver biopsy and ended on July 9, 2008, or, if earlier, the date of death. Kaplan-Meier curves are used to depict the mortality in the cohort graphically. All analyses were conducted using SAS statistical software. This study was approved by the regional ethics committee at the Karolinska Institutet in Stockholm.
Reevaluation of the Liver Biopsies.
Figure 1 depicts the distribution of diagnoses obtained by reevaluation of the liver biopsy specimens from our 256 subjects. Fatty deposition of greater than 5% was detected in the livers of 143 subjects (56%). According to the medical records, 25 of these 143 abused alcohol, and their condition was therefore classified as alcoholic fatty liver disease (AFLD). The remaining 118 had NAFLD, including 51 cases of NASH, and 67 of bland steatosis.
Of the 256 subjects, 41 (16%) exhibited chronic viral hepatitis, and of these 30 had hepatitis C alone and seven hepatitis B alone, the remaining four subjects being co-infected with both of these viruses. In addition to the 25 subjects with AFLD (see earlier discussion), 10 others (4%) among the 256 subjects had a history of alcohol abuse; however, the biopsy specimens of these 10 contained less than 5% fat, so they were classified as suffering from alcoholic liver disease without steatosis. Eleven subjects (4%) demonstrated autoimmune hepatitis; three (1%), hemochromatosis; six (2%), a deficiency in alpha1-antitrypsin; 13 (5%), unspecific histopathological findings; and 29 (11%), no apparent abnormality at all. At the time of biopsy, cirrhosis was present in 23 of the 256 subjects (9%) (including five with NASH, four with bland steatosis, and two with alcoholic steatohepatitis/AFLD, three with alcoholic liver disease, seven with viral hepatitis, one with hemochromatosis, and one other).
Relative Risk of Death.
We followed-up our 256 subjects for a total of 5,248 person-years (median period of follow-up, 24 years [range, 0.5–28 years]; mean period of follow-up, 21 years [standard deviation, 7.7]). The subjects with fatty liver were slightly overweight, with a body mass index of 27.2 ± 3.1 for those with bland steatosis and 28.52 ± 5.1 for those with NASH. The corresponding mean ages at the time of biopsy were 45.1 ± 11.5 and 49.4 years, respectively. The prognoses of the 256 subjects who underwent liver biopsy and the 118 subjects with NAFLD are given in Fig 2A-C and Tables 2 and 3. During this follow-up, 113 subjects died, which corresponded to an 80% increased risk of death in comparison with the general population (SMR, 1.8; 95% CI, 1.48–2.16; Table 3). Of the 118 subjects with NAFLD, 47 died (SMR, 1.7; 95% CI. 1.24–2.25); among the 67 with bland steatosis, 23 died (SMR, 1.6; 95% CI, 0.98–2.32; P = 0.062); among the 51 with NASH, 24 died (SMR, 1.9; 95% CI, 1.19–2.76; P = 0.007) (Table 3, Fig. 2A). When the nine patients with cirrhosis among those with bland steatosis or NASH were omitted from the calculation, the relative risks of death for those two groups were 1.3 (95% CI, 0.80–2.07) and 1.7 (95% CI, 1.04–2.63), respectively (Table 3). When subjects with stage 3 fibrosis were also omitted from these same two groups, the corresponding values were 1.1 (95% CI, 0.6–1.77) and 1.7 (95% CI, 1–2.74), respectively (Table 3). Furthermore, of the eight subjects with fatty liver but no inflammation, fibrosis, or ballooning at all, one was deceased (Fig. 2B). When the subjects were grouped for analysis according to their NASH Activity Score (0–2 [SMR, 1.01; 95% CI, 0.41–2.09], 3–4 [SMR, 2.05; 95% CI, 1.15–3.39], and 5–8 [SMR, 1.82; 95% CI, 1.16–2.71]), subjects in NASH Activity Score classes 3 to 4 and 5 to 8 had an increased risk of death (Table 2).
|NAS-Score||Stage of Fibrosis|
|No-NASH (1-2)||Alive (24)||12||2||0||0||10|
|Border line (3–4)||Alive (18)||10||3||0||0||5|
|NASH (5–8)||Alive (29)||11||12||2||1||3|
|Number of Deceased||RR||CI 95%||Approximate P Value (2-Tailed)|
|Excluding cirrhosis and fibrosis stage 3||0|
In the case of the AFLD/ASH subjects, a 294% increased risk of death was observed (SMR, 3.9; 95% CI, 2.41–6.09). Chronic infection with hepatitis C or B was associated with an increased risk of death (Table 3; Fig. 2C). By contrast, subjects with autoimmune hepatitis or another diagnosis did not exhibit reduced survival, although certain subsets contained too few subjects to allow valid comparison (Table 3). Of the 23 subjects with cirrhosis, 18 died during our follow-up.
Causes of Death.
Of the 67 deaths among our subjects with fatty liver (AFLD + NAFLD), 22 (33%) were attributable to cardiovascular disease, 14 (21%) to extrahepatic malignancy, and 16 (24%) to liver diseases, including HCC. Among those with NAFLD, 14 (30%) died of cardiovascular disease, 13 (28%) of extrahepatic malignancies, and nine (19%) of liver disease. Eight AFLD subjects (40%) died of cardiovascular disease, and seven (35%) died of liver disease, whereas the subjects with alcoholic liver disease died primarily of malignancies (three subjects, 38%) (Table 4).
|Cause of Death||Diagnosis|
|Bland Steatosis (23)||NASH (24)||ASH/AFLD (20)||Alcohol (8)||Hep C (13)||Hep B (4)||Hep C & Hep B (2)||AIH (5)||AAT-defic (0)||HH (1)||Other (7)||Normal (6)|
|Bleeding from duodenum||1|
Subjects with liver cirrhosis died of their cirrhosis (n = 5), extrahepatic malignancies (n = 4), liver cancer (n = 3), cardiovascular disease (n = 3), external causes (n = 1), and alcohol (n = 1). The stages of fibrosis of those in the NAFLD group who died are documented in Table 5.
|Cause of Death||Stage of Fibrosis|
|Perforation of intestine||1|
The current report documents the longest follow-up of patients who originally demonstrated elevated serum levels of hepatic enzymes and were subsequently shown by biopsy to be suffering from NAFLD. This study has four major strengths: First, all of the subjects were enrolled consecutively during a defined period (1980–1984). Second, all underwent liver biopsy at the time of referral, so that the diagnoses of NAFLD are based on histological criteria. Third, reevaluation of the initial biopsy findings was performed in all cases. And finally, even deceased subjects could be followed-up through the Cause of Death Registry, so that there were no losses during follow-up.
The scoring system developed by Kleiner et al.16 combines the three parameters hepatic fat content, lobular inflammation, and ballooning, all of which contribute equally to the NASH score. Thus, a patient with pronounced steatosis and relatively moderate inflammation could receive the same NASH score as one with mild steatosis but intense inflammation and ballooning. One disadvantage to this scoring system is that it does not take fibrosis into consideration. Among our subjects with bland steatosis, we found eight with severe fibrosis or cirrhosis, which is unexpected in the absence of NASH. As a consequence, to evaluate the prognostic value of fatty liver without significant fibrosis, we also performed analyses omitting these eight subjects.
Subjects with bland steatosis without severe fibrosis or cirrhosis at the time of biopsy exhibited no increased risk of death. When patients with cirrhosis or severe fibrosis are excluded from the group with bland steatosis, deaths attributable to either cirrhosis or liver cancer are much less common. Both for subjects with fatty liver and for the whole study population, we found that persistently elevated serum levels of liver enzymes were associated with an increased risk of death during the study period. As a group, subjects with NAFLD demonstrated a significantly increased risk of death, but this risk was not as high as for patients with chronic viral hepatitis or alcoholic liver disease (Fig. 2C). In our population, those with NAFLD exhibited poorer survival than those with autoimmune hepatitis, hemochromatosis, or alpha-1-antitrypsin deficiency taken combined.
It is currently unknown why certain patients diagnosed by biopsy as suffering from NAFLD and elevated serum levels of liver enzymes develop inflammation and fibrosis (NASH). One of the main findings presented here is that survival among these subjects is lower than among the matched reference population. Recently, Adams and co-workers9, 11, 19 also reported that survival among NAFLD subjects is lower than expected in comparison with the general population. In most cases, their diagnosis of NAFLD was based on imaging rather than histological examination.
Moreover, in a 14-year follow-up, Ekstedt et al.11 demonstrated that survival of patients with NASH was lower than expected in comparison with the general population. In the current 28-year follow-up, we have confirmed these findings. The main causes of death among our subjects with NASH were cardiovascular disease, followed by extrahepatic cancers and next hepatic diseases.
As expected, cardiovascular diseases were found to be the major cause of death in the current investigation. Recent epidemiological studies indicate an increased incidence of major cardiovascular events in subjects with NAFLD, independent of traditional risk factors and aspects of the metabolic syndrome.19–21 A higher risk for cardiovascular mortality compared with the reference population has been shown in subjects shown to have NAFLD though biopsy11 as well as in a population-based cohort study.22
Liver diseases, including hepatocellular carcinoma (HCC), were found to be the third largest cause of death among subjects with NAFLD. HCC is a major health problem worldwide, with more than 500,000 cases diagnosed annually.23 Whereas the incidence of HCC has been increasing during the last 5 to 8 years, the survival of those affected has not changed significantly during the past two decades.
With NAFLD becoming more and more common, with enhanced prevalence among younger people, the associated rise in relative risk of mortality and terminal liver disease will be of considerable significance to public health in the future. In the current investigation, four subjects with NAFLD/NASH died of complications of cirrhosis, and five died of HCC. It thus seems reasonable to recommend changes in lifestyle for all subjects with NAFLD.
It has been shown previously that examination of a liver biopsy at entry into the healthcare system is a valuable predictor of future cirrhosis-related complications, exhibiting a positive predictive value of 18% in subjects with periportal fibrosis and a negative predictive value of 100% for those without established periportal fibrosis. One of four (25%) patients with cirrhosis at baseline, 3 of 12 patients (25%) with stage 3 fibrosis at baseline, and three patients of 22 (14%) with stage 2 fibrosis at baseline were found to have developed end-stage liver disease during follow-up.11 Of the nine subjects diagnosed with NAFLD and concomitant cirrhosis at the time of inclusion in the current study, three died of liver cancer, two of extrahepatic cancers, one of cirrhosis, one of cardiovascular disease, and one in an accident. Only one remains alive.
We conclude that in our cohort of subjects with elevated serum levels of liver enzymes who underwent consecutive liver biopsies 28 years ago, 46% could be diagnosed as suffering from NAFLD. At the time of the initial biopsy, 8% of those with NAFLD had cirrhosis, and 43% had NASH. Overall survival was reduced in subjects with NAFLD and NASH, whereas bland steatosis with or without severe fibrosis was not associated with any increase in mortality risk in comparison with the general population. Patients with NASH had a lower risk of death than those with alcoholic liver disease or chronic viral hepatitis but a higher risk than those suffering from autoimmune and metabolic liver diseases.
On the whole, patients with NAFLD die of liver-related causes to a greater extent than the general population, but we still see cardiovascular disease and extrahepatic malignancies to be the primary and secondary causes of deaths among these patients. The frequency of HCC is almost 1000-fold higher in this group than what has been reported for Sweden earlier. Our findings motivate a more active approach to the diagnosis and treatment of NASH, with more frequent use of liver biopsy for diagnosis. Thus, subjects with NASH have an increased risk of death; much emphasis should be put into treatment. New treatment strategies have to be sought.
- 1Nonalcoholic steatohepatitis: Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clin Proc 1980; 55: 434–438., , , .
- 2Serum alanine aminotransferase (ALT) elevation in asymptomatic US Air Force basic trainee blood donors. Dig Dis Sci 1993; 38: 2145–2150., .
- 3Evaluation of abnormal liver-enzyme results in asymptomatic patients. N Engl J Med 2000; 342: 1266–1271., .
- 4Liver investigation in 149 asymptomatic patients with moderately elevated activities of serum aminotransferases. Scand J Gastroenterol 1986; 21: 109–113., , , .
- 5Patients with persistent elevation of aminotransferases: investigation with ultrasonography, radionuclide imaging and liver biopsy. J Intern Med 1993; 233: 7–12., .
- 6Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999; 94: 2467–2474., , , , .
- 7Nonalcoholic steatohepatitis: an expanded clinical entity. Gastroenterology 1994; 107: 1103–1109., , , .
- 8Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002; 123: 134–140., , , , , , , et al.
- 9The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005; 129: 113–121., , , , , , et al.
- 10Increased overall mortality and liver-related mortality in non-alcoholic fatty liver disease. J Hepatol 2008; 49: 608–612., , .
- 11Long-term follow-up of patients with NAFLD and elevated liver enzymes. HEPATOLOGY 2006; 44: 865–873., , , , , , et al.
- 12Causes of death register. http://socialstyrelsen.se/register/dodsorsaksregistret. Accessed October 2009.
- 13World Health Organization. International Classification of Diseases. 1966.
- 14World Health Organization. International Classification of Diseases. 1987.
- 15World Health Organization. International Classification of Diseases. 1997.
- 16Design and validation of a histological scoring system for nonalcoholic fatty liver disease. HEPATOLOGY 2005; 41: 1313–1321., , , , , , et al.
- 17Chronic hepatitis: an update on terminology and reporting. Am J Surg Pathol 1995; 19: 1409–1417., .
- 18Liver biopsy interpretation, 2000. Eur J Hepatol Gastroenterol 1991: 79–85.. In:
- 19Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease. World J Gastroenterol 2007; 13: 1579–1584., , , , , , et al.
- 20Nonalcoholic fatty liver disease and risk of future cardiovascular events among type 2 diabetic patients. Diabetes 2005; 54: 3541–3546., , , , , , et al.
- 21Nonalcoholic fatty liver disease is independently associated with an increased incidence of cardiovascular events in type 2 diabetic patients. Diabetes Care 2007; 30: 2119–2121., , , , , , et al.
- 22Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study. Am J Gastroenterol 2008; 103: 2263–2271., , , , , , et al.
- 23Estimating the world cancer burden: Globocan 2000. Int J Cancer 2001; 94: 153–156., , , .