Article Text

Download PDFPDF

Transient elastography as a screening tool for liver fibrosis and cirrhosis in a community-based population aged over 45 years
  1. Dominique Roulot1,2,
  2. Jean-Luc Costes3,
  3. Jean-François Buyck4,
  4. Ursula Warzocha1,
  5. Nicolas Gambier1,
  6. Sébastien Czernichow4,5,
  7. Hervé Le Clesiau3,
  8. Michel Beaugrand2
  1. 1Unité d'Hépatologie, Hôpital Avicenne, Assistance Publique, Hôpitaux de Paris, Bobigny and UPRES-EA 3406, Université Paris 13, Bobigny, France
  2. 2Service d'Hépatogastroenterologie, Hôpital Jean Verdier, Assistance Publique, Hôpitaux de Paris, Bondy and Université Paris 13, Bobigny, France
  3. 3Centre de Prévention Sanitaire et Sociale, Bobigny, France
  4. 4Département de Santé Publique, Hôpital Avicenne, Assistance Publique, Hôpitaux de Paris, Bobigny, France
  5. 5INSERM, U55; INRA, U1125; CNAM, EA3200; CRNH-IdF, Université Paris 13, Bobigny, France
  1. Correspondence to Dominique Roulot, Unité d' Hépatologie, Hopital Avicenne, 125 route de Stalingrad 93009 Bobigny, France; dominique.roulot{at}


Background Liver stiffness measurement (LSM) has been used to measure fibrosis in patients with various types of chronic liver diseases. However, its usefulness as a screening procedure in apparently healthy people had not been evaluated to date.

Methods 1358 subjects >45 years old from a general population attending for a medical check-up were consecutively enrolled in the study. All subjects were submitted to medical examination and laboratory tests in addition to LSM, performed on the same day by a single operator. Subjects with LSM values >8 kPa were referred to a liver unit for further investigations.

Results 168 subjects were not considered for analysis due to missing data (n=23), LSM failure (n=51) or unreliable LSM values (n=94). Among the 1190 remaining subjects, 89 (7.5%) had LSM >8 kPa including nine patients with LSM >13 kPa. Despite the fact that normal liver tests were observed in 43% of them (38 out of 89), a specific cause of chronic liver disease was found in all cases. Non-alcoholic fatty liver disease (NAFLD) was the likely cause of chronic liver disease in 52 patients, alcoholic liver disease (ALD) in 20, and both causes were associated in seven additional patients. Hepatitis C virus and hepatitis B virus chronic hepatitis was documented in five and four cases, respectively, and primary biliary cirrhosis in one. Liver biopsy was obtained for 27 patients, including the nine patients with LSM >13 kPa, who were diagnosed with liver cirrhosis due to ALD (n=5), chronic hepatitis C (n=3) or chronic hepatitis B (n=1). The 18 remaining biopsies showed liver fibrosis in all cases except one (isolated steatosis), with ALD and NAFLD being present in six and eight cases, respectively.

Conclusion LSM proved to be a useful and specific procedure to screen for cirrhosis in the general population and to detect undiagnosed chronic liver disease in apparently healthy subjects.

  • Fibroscan
  • liver stiffness
  • screening procedure
  • non alcoholic fatty liver disease
  • metabolic syndrome
  • cirrhosis
  • fibrosis
  • liver biopsy
  • nonalcoholic steatohepatitis

Statistics from

Significance of this study

What is already known about this subject?

  • The frequency of chronic liver disease is increasing due in particular to NAFLD, and it remains asymptomatic for a long time. Liver function tests (GGT and ALT) are currently the most common procedure to uncover liver involvement in apparently healthy individuals.

  • Liver stiffness measurement (LSM) is a recently developed method to detect cirrhosis and liver fibrosis in many types of hepatic disease. So far this procedure has been demonstrated to provide an excellent negative predictive value for cirrhosis.

  • Normal values of LSM in healthy subjects in the general population have been determined recently, allowing investigation of the positive predictive value of this procedure in apparently healthy subjects.

What are the new findings?

  • The study based on a large series of individuals demonstrates that LSM can detect liver involvement in apparently healthy subjects, with a positive predictive value of 100% for cirrhosis, which was diagnosed in 0.7% of the population.

  • In all individuals with LSM values >8 kPa (7%) a cause of liver disease could be determined and, among the patients undergoing liver biopsy, almost all showed cirrhosis or liver fibrosis.

  • Although LSM scored normal in only 4% of the subjects with perturbed liver tests (increased levels of both GGT and ALT), liver tests were totally normal in 43% of the patients with liver disease, diagnosed based on elevated LSM.

How might this impact on clinical practice in the foreseeable future?

  • If future studies determine that the superior positive predictive value of LSM can be achieved at an acceptable cost, compared with common screening methods, LSM might represent a first-line procedure for the mass screening of liver disease in the general population.


Chronic liver diseases represent a major cause of morbidity and mortality in Europe1 and the USA, affecting 360 per 100 000 persons and ranking as the 12th leading cause of overall mortality.2 In addition to viral infections, non-alcoholic fatty liver disease (NAFLD) is recognised more and more as an important cause of chronic liver disease in the western world, with an estimated prevalence of up to 30%.3 A significant proportion of NAFLD cases progress to cirrhosis4 5 and hepatocellular carcinoma (HCC), particularly when associated with other causes of liver disease such as alcohol abuse or hepatitis C virus (HCV) infection.

Progressive hepatic fibrosis and subsequent development of cirrhosis is the hallmark of almost all chronic liver diseases. As many patients with chronic liver diseases remain asymptomatic for long periods of time, non-invasive screening methods for early identification of advanced fibrosis or cirrhosis in the general population may help to establish earlier diagnosis and treatment.

Liver stiffness measurement (LSM) by transient elastography (TE), a recent, easy to perform procedure, might represent a valuable screening tool in this context. This technique allows evaluation of liver stiffness,6 which is strongly correlated with liver fibrosis stages, assessed by simultaneous liver biopsies. Initially reported in patients with chronic hepatitis C,7 8 this correlation was confirmed in other liver diseases such as chronic hepatitis B,9 non-alcoholic steatohepatitis (NASH)10 and alcoholic liver disease (ALD).11 Moreover, the best diagnostic performance of TE has been reported in patients with severe fibrosis and cirrhosis, two conditions which may frequently cause complications.12 Although normal values in a large series of apparently healthy subjects have recently been established, allowing 8 kPa to be considered as the upper limit value of normal liver stiffness,13 so far no study has evaluated LSM by TE as a screening tool to detect liver fibrosis and cirrhosis in the general population.

The purpose of our study was to assess the performance of LSM as a screening procedure for liver diseases in a large unselected community-based population aged 45 years or above. All subjects with LSM values >8 kPa were referred to a liver centre to be investigated for liver fibrosis and cirrhosis.

Subjects and methods


All subjects aged 45 years or over, arriving for a scheduled free medical check-up in a social medical centre (Bobigny, France) between September 2005 and February 2008 were asked to participate in the study. Subjects accepting gave written and informed consent, and the local ethics committee (CCPPRB, Aulnay-sous-bois, France) approved the study protocol.

Subjects underwent LSM and had a complete clinical examination and laboratory tests on the same day. The following data were collected: age, gender, daily alcohol intake (numbers of glasses of alcohol per week) or previous history of excessive alcohol consumption (>30 g per day), tobacco use, history of hypertension, diabetes or dyslipidaemia, type and duration of antidiabetic and antihypertensive treatments, body mass index (BMI), waist circumference, and systolic and diastolic blood pressure. Laboratory tests included: platelet count, mean corpuscular volume, γ-glutamyltranspeptidase (GGT), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities, total cholesterol, high-density lipoprotein (HDL)-cholesterol, low-density lipoprotein (LDL)-cholesterol, triglycerides and fasting glucose.

According to the National Cholesterol Education Program (NCEP) criteria,14 the existence of metabolic syndrome was established in subjects displaying at least three of the five following features: (1) waist circumference >102 cm in men and >88 cm in women; (2) triglycerides ≥150 mg/dl (1.69 mmol/l); (3) HDL-cholesterol <40 mg/dl (1.04 mmol/l) in men and <50 mg/dl (1.29 mmol/l) in women; (4) blood pressure ≥130/85 mm Hg; and (5) fasting glucose ≥110 mg/dl (6.1 mmol/l). Participants receiving antidiabetic (oral agents or insulin) or antihypertensive medication were classified for the study as subjects having increased fasting blood glucose or increased blood pressure, respectively.

All subjects with LSM values >8 kPa had serum samples frozen for further investigation and were referred to a liver centre. A complete screening for possible causes of liver diseases was then performed including determination of serum ferritin, the search for hepatitis B surface antigen (HbsAg), hepatitis C virus antibodies (anti-HCV) and autoantibodies (antinuclear, antismooth muscle and antimitochondrial antibodies). For the subjects who did not accept further investigations at a liver centre, complementary tests were performed on the frozen samples.

Liver stiffness measurements

The same physician (JLC), blinded to clinical and biological data, performed TE (FibroScan, Echosens, Paris, France) in all subjects. Details concerning the device and the examination procedure have been reported previously.6 Briefly, an ultrasound transducer probe is mounted on the axis of a vibrator. Vibrations of mild amplitude (1 mm) and low frequency (50 Hz) are transmitted by the transducer, inducing an elastic shear wave that propagates through the underlying tissues. Pulse-echo ultrasound acquisitions are used to follow the propagation of the shear wave and to measure its velocity, which is directly related to tissue stiffness (the elastic modulus): the stiffer the tissue, the faster the shear wave propagates. TE measures liver stiffness in a volume that approximates a cylinder 1 cm wide and 4 cm long, 25–65 mm below the skin surface. This volume is at least 100 times larger than that of a needle biopsy sample and therefore more representative of the hepatic parenchyma. The probe available at the time of the study was designed for subjects with normal morphology.

The success rate of LSM was calculated as the ratio between validated and total measurements. Results were expressed as the median value of the total measurements in kiloPascals (kPa), the values ranging from 2.5 to 75 kPa.7 Only procedures with at least 10 successful acquisitions and a success rate of at least 60% were considered reliable. In addition, the median value of successful measurements was considered representative of the liver stiffness in a given patient only if the IQR of all validated measurements was <30% of the median value. LSM was classified as failure (no acquisition), unreliable (acquisitions not meeting the above criteria) or validated.

A first cut-off of 8 kPa was selected since a previous study of LSM values in a healthy population reported that such a value (the 90th percentile) was a criterion sufficient for further investigations.13 The second cut-off of 13 kPa was chosen because a recent meta-analysis of LSM studies identified this value as the optimal cut-off for the diagnosis of cirrhosis.12

Liver histology and quantification of liver fibrosis and steatosis

Liver biopsy was proposed to all subjects with LSM values ≥8 kPa, according to the clinical background, and to all patients with LSM values >13 kPa. Liver biopsies were fixed in formalin and paraffin embedded. One experienced pathologist, blinded to the results of TE, analysed all biopsy specimens. The median biopsy length was 18 mm (range 13–25). Fibrosis was staged on a 0–4 scale. Activity was graded according to the Metavir score for viral hepatitis15 and the Brunt score for NAFLD and ALD.16 Steatosis was categorised by visual assessement as <10, 10–30, 30–60 and >60% of the hepatocytes. The presence of perisinusoidal fibrosis was also specified.

Statistical analysis

Results are expressed as means±SD or median (range), and percentages for continuous and categorical variables, respectively. Variables with skewed distributions were log-transformed, and geometric means (95% CIs) are presented in the tables.

The study population was stratified into three groups: LSM failure, unreliable LSM and validated LSM. Univariate and multivariate logistic regressions were used to evaluate sociodemographic, clinical and biological factors associated with LSM failure and with unreliable LSM: factors associated with outcome in univariate analyses (with a p value <0.20) were entered in the multivariate logistic regression. Factors with a p value <0.10 in the multivariate analysis were kept in the final model, using a backward selection procedure.

Characteristics of subjects with a validated LSM were compared according to gender using the Student t test and χ2 or Fisher exact test, where appropriate. The study population was then stratified into two groups according to LSM value: <8.0 kPa and ≥8.0 kPa. Characteristics of the study population according to the categorical LSM variable were compared using analysis of covariance and χ2 or Fisher exact test in univariate analysis. Age-/sex-adjusted analysis and multivariate analyses were further performed using analysis of covariance and logistic regression models.

All statistical analyses were carried out with the use of SAS software version 9.1 (SAS Institute, Cary, North Carolina, USA). Statistical significance was judged at a p value <0.05 (two-tailed tests).



A total of 1358 subjects aged 45 years or above without known liver disease were enrolled in the study. Twenty-three of them (1.7%) were excluded because of missing clinical or biological data (figure 1). These subjects did not differ from the selected subjects in terms of age (59.7 vs 57.8 years, respectively; non-significant (NS)), gender (women 34.8% vs 40.8%, respectively; NS) and BMI (24.7 vs 26.7 kg/m2; NS). Fifty-one subjects (3.8%) were excluded because of LSM failure. Ninety-four subjects (7%) were excluded because of unreliable LSM. Factors influencing LSM failure and unreliable LSM are shown in table 1. The percentage of subjects with obesity and with a metabolic syndrome was significantly higher in the group of patients with LSM failure compared with the group with validated LSM (68.6% vs 17.1% and 47.1% vs 20.3%, respectively, p<0.0001). Factors associated with LSM failure and unreliable LSM results, in univariate and multivariate analyses, are shown in tables 2 and 3, respectively. LSM failure was independently associated with the following factors: BMI >30 kg/m2, elevated waist circumference, female sex and age older than 57 years. Unreliable LSM results were independently associated with the following factors: BMI >30 kg/m2 and ALT >40 UI/l.

Figure 1

Flow chart of the study. After exclusion of subjects with missing data, failure of liver stiffness measurement (LSM) or unreliable LSM, 1190 subjects were included in the study. In 1101 subjects, median values of LSM were <8 kPa; LSM values were between 8 and 13 kPa in 80 subjects and >13 kPa in nine subjects.

Table 1

Subjects characteristics according to LSM feasibility

Table 2

Factors associated with LSM failure in univariate and multivariate analyses

Table 3

Factors associated with unreliable LSM in univariate and multivariate analyses

Finally, a total of 1190 subjects were analysed. Their characteristics according to sex are shown in table 4. Among these 1190 subjects, the percentage of men was higher than that of women (60.6% vs 39.4%). The mean age according to gender was comparable (57.9±8.6 vs 57.3±9.2 years, for men and women, respectively, p=0.27). The mean BMI was 26.5±3.7 kg/m2 for men and 26.3±4.3 kg/m2 for women (p=0.50). The percentage of overweight subjects was higher in men than in women, but obesity was less frequent in men than in women (49.4% vs 40.2% and 15.4% vs 19.6%, respectively, p=0.006). In contrast, the percentage of subjects with metabolic syndrome was similar in men and women (20.6% vs 20.0%, p=0.82). Tobacco use was significantly higher in men than in women, but no influence of this parameter on liver stiffness could be demonstrated by multivariate analysis taking into account age and sex. Self-reported alcohol consumption was significantly higher in men than in women (2.7±4.4 vs 1.2±3.1 glasses/week, p<0.0001) but had no influence on LSM when considered as a continuous variable.

Table 4

Main characteristics of the 1190 evaluated subjects according to gender

Influence of age and gender on LSM

The median LSM value of the 1190 subjects was 5.3 kPa (range 1.8–35). Liver stiffness values tended to be higher with age, with a weak but significant correlation (Pearson coefficient correlation; ρ=0.10, p=0.0007).

Confirming previous results, LSM were significantly higher on average in men than in women (5.7 kPa (range 1.8–35) vs 4.9 kPa (range 2.4–20.7) p<0.0001).13 The percentage of women was higher in the group of subjects with LSM <8 kPa (95.5% vs 90.6%, p=0.001) (table 4).

Characteristics of the population according to LSM values

Of the total 1190 subjects, 1101 had LSM values <8 kPa (92.5%, group A), whereas 89 showed LSM values ≥8 kPa (7.5%, group B). In group B, nine individuals (0.76%) had LSM values >13 kPa. The characteristics of subjects according to the LSM values indicated above are summarised in table 5.

Table 5

Main characteristics of subjects according to LSM values

Mean age was higher in group B than in group A (60.4±8.5 and 57.5±8.8 years, respectively, p=0.003). Gender distribution was different between the two groups, with 40.8% of women in group A and 23.6% in group B (p=0.001).

Tobacco use was not different between the two groups. Alcohol consumption was more important in subjects with LSM values >13 kPa, but the difference did not remain significant after adjustment for age and sex. The mean BMI was significantly higher in group B compared with group A (29.1±4.8 vs 26.2±3.8 kg/m2, respectively, p<0.0001). Waist circumference was also higher in group B compared with group A (97.2±12.0 vs 88.6±10.7 cm, respectively, p<0.0001). Higher mean values of systolic and diastolic blood pressure were observed in group B, compared with group A (139.3±14.1 vs 131.2±15.4 mm Hg, p=0.005, and 82.6±9.0 vs 79.6±9.5 mm Hg, p=0.005). Mean values of fasting glucose and triglycerides were higher in group B than in group A (6.52 (6.24 to 6.81) vs 5.51 (5.44 to 5.58) mmol/l, p<0.0001, and 1.31 (1.18 to 1.46) vs 1.12 (1.08 to 1.15) mmol/l, p=0.006). Metabolic syndrome was significantly more frequent in group B than in group A (41.6% and 18.6%, respectively, p<0.0001).

Mean GGT values were 24.7 (23.7 to 25.8) in group A and 47.0 (40.5 to 54.6) UI/ml in group B (p<0.0001). Mean ALT values were also significantly increased in group B compared with group A (35.2 (32.0 to 38.7) and 21.7 (21.2 to 22.3) UI/ml, respectively, p<0.0001). Nevertheless, 38 individuals in group B (43%) displayed totally normal liver tests (both GGT and ALT). In group A, only 42 individuals (4.2%) showed elevated values for both GGT and ALT. Platelet count was significantly lower in group B than in group A (234±66 vs 252±59 109/l, p=0.005), especially in the subgroup of patients with LSM ≥13 kPa (180±83 109/l), probably because of the presence of liver cirrhosis (see below).

Factors associated with LSM values >8 kPa in univariate and multivariate analyses are shown in table 6. LSM values >8 kPa were independently associated with the following factors: ALT >40 UI/l, GGT>45UI/l, BMI >30 kg/m2, elevated waist circumference, age older than 57 years, diabetes and triglycerides >150 mg/dl.

Table 6

Factors associated with LSM >8 kPa in univariate and multivariate analyses

Identified causes of liver disease in patients with LSM ≥8 kPa

For all 89 subjects with LSM values >8 kPa, a specific cause of chronic liver disease was either documented or highly suspected. NAFLD was suspected when subjects were overweight with or without a metabolic syndrome and in the absence of other causes of liver disease. Among the 89 subjects, 37 were overweight and 36 obese. Metabolic syndrome was present in 37 cases (41.5%) (14 overweight patients, 23 obese patients). Twenty-seven patients were alcoholic (alcohol consumption >3 glasses/day), 20 displayed isolated alcoholism and seven alcoholism associated with NAFLD. HbsAg was positive in five patients, anti-HCV antibodies were positive in eight patients and antimitochondrial antibodies were positive in one patient.

Liver histological findings according to LSM values

Liver biopsy was systematically proposed to patients with LSM >13 kPa and accepted each time after detailed information concerning the significance of elevated LSM. In all cases (n=9), cirrhosis was confirmed histologically. The causes of cirrhosis were alcoholism in five patients, chronic viral C infection in three and chronic viral B infection in one (table 7).

Table 7

Liver histological findings according to LSM, clinical and biological data and causes of liver disease

For the patients with LSM between 8 and 13 kPa who accepted further investigation (n=66), liver biopsy was proposed according to the supposed cause of liver disease and the degree of LSM elevation. Liver biopsy was considered not necessary in 32 cases (principally because of patent metabolic syndrome and normal liver tests) and was refused by the patient in 16 cases (figure 2). The patients who did not undergo liver biopsy were not statistically different from those who underwent biopsy. Nevertheless, the percentage of women or patients with BMI >30 and with metabolic syndrome was higher in the group without liver biopsy (29.7% vs 8.0%, p=0.05; 43.8% vs 28.0%, p=0.17; 46.9% vs 28.0%, p=0.10 respectively). Eighteen patients with LSM between 8 and 13 kPa (22.5%) finally underwent liver biopsy (table 7). Histological examination showed liver fibrosis in all cases except one. Alcoholic steatohepatitis was found in six cases and non-alcoholic steatohepatitis in eight. One patient exhibited steatosis without fibrosis (F0) in the biopsy sample. Lesions of chronic viral C infection were present in one case, chronic viral B infection was diagnosed in two cases and primary biliary cirrhosis (PBC) in one case.

Figure 2

Flow chart for the investigation of patients with liver stiffness measurement (LSM) values ≥8 kPa.

It is noteworthy that hepatic lesions were present in all of the six patients with completely normal liver tests who underwent liver biopsy (Metavir score of A1F4 in one patient, A2F2 in 2, A1F2 in 1 and A1F1 in 2) (table 7).

All patients with histologically proven cirrhosis accepted periodical follow-up. Four patients with viral disease were treated: one patient with chronic hepatitis C and one patient with chronic hepatitis B infection achieved a sustained virological response. Among the five patients with ALD, three stopped drinking completely.


Despite the increasing popularity of TE to explore fibrosis in patients with chronic liver diseases, LSM has not been evaluated as a screening procedure for liver fibrosis in a large cohort of apparently healthy subjects. In our study of 1190 subjects we focused on individuals above 45 years of age because this fraction of the population is more at risk of having liver fibrosis.

Sample characteristics of our population were close to those of the French general population, with minor discrepancies. First, the prevalence of obesity in our study group was slightly higher than expected for a French population-based study. Indeed, 47.6% of the subjects showed a normal BMI, with 37.9% being overweight and 14.4% obese, whereas the last nationwide available statistics in 2003 reported 28% overweight and 10% obese.17 Secondly, the percentage of subjects with metabolic syndrome in our study was 20.7%, which is a much higher value than the 11% found in the French DESIR study.18 These discrepancies can probably be explained by the higher mean age of our group compared with the DESIR study and the lower socioeconomic status of the subjects who live in a relatively poor suburb of Paris, compared with the rest of the country.19 The average LSM and its distribution in the present group were similar, however, to those reported in a previous study conducted in the same district on a smaller number of individuals without overt cause of liver diseases and with normal liver enzymes.13 In particular, we could confirm the influence of gender on LSM, as well as the weak but positive correlation between age and LSM.

A technical limitation of TE in our study is the relatively high rate of failure (3.8%) or unreliable values (7%) of LSM, particularly in obese individuals (table 1). Although this high percentage could be due to the rather stringent criteria of validation used in the study, it is clearly linked to the high prevalence of overweight and obese people in our population. The factors which significantly influenced LSM failure and unreliable results were BMI, waist circumference and metabolic syndrome. Using the same criteria of validation, Castera et al reported the same rate of failure (3.1%) but a higher rate of unreliable results (15.8%).20 Here the superior reliability is probably explained by the fact that all LSMs were done by the same operator, who had already carried out >500 examinations before the beginning of the study. Failed or unreliable LSM probably caused the underestimation of patients with hepatic involvement due to metabolic syndrome. However, in the near future the use of probes specifically adapted to obese people, will probably improve the success rate and reliability of TE.

Since TE is a recent procedure, which has been tested in patients with various causes of chronic liver diseases and in smaller series of healthy individuals, the key issue of ‘cut-off’ LSM values for the diagnosis of liver fibrosis or cirrhosis has not been resolved yet. A recent meta-analysis of ∼50 reports suggested an optimal cut-off value of 13.0 kPa for cirrhosis.12 Confirming this hypothesis, all the examined individuals showing LSM values >13 kPa had histologically proven liver cirrhosis. In a study examining LSM values in normal individuals,13 upper normal limits of 7.8 and 8 kPa were proposed for non-obese healthy women and men, respectively. Consistent with these values, all subjects of the present study with LSM values >8 kPa probably had liver disease, which was confirmed histologically in patients who accepted liver biopsy. Thus, all individuals showing LSM values of ≥8 kPa should be investigated for causes of liver diseases, even in the presence of normal liver enzymes.

An important finding in this report is the validation of LSM as a reliable diagnostic method to screen the general population for fibrosis and cirrhosis. We could indeed detect previously unknown fibrosing liver disease in ∼8% of the individuals, including significant fibrosis in 7.6% and cirrhosis in 0.67% of the cases. A similar prevalence (2.8% for fibrosis and 0.4% for cirrhosis) was reported in a recent study, based on a biological approach in subjects of ≥40 years of age.21 The proportion of individuals with liver cirrhosis found here was slightly lower than that proposed in a recent survey of the literature (1%).22 However, our estimate of the prevalence of undiagnosed fibrosis and cirrhosis in the general population might be affected by potential biases with opposite consequences. On one hand, no patient with previously known liver disease was enrolled in our cohort. Moreover, as discussed above, some patients with liver involvement due to metabolic syndrome probably remained undiagnosed because of technical issues. On the other hand, most subjects were from a suburban low-income population with a high percentage of immigrants, where obesity and viral infection are more prevalent than in the rest of the French population.19 Finally, we studied only subjects >45 years of age and a majority of men, a subset of the population with higher prevalence of liver fibrosis.

In all supposedly healthy patients with LSM >8 kPa (n=89), a likely cause of liver disease could be documented or highly suspected. Interestingly 43% of them displayed normal GGT, AST and ALT values, whereas only 4.2% of the subjects with abnormal GGT and AST/ALT had LSM values of <8 kPa. These data indicate that LSM may have a superior screening performance for liver involvement compared with routine liver tests. Hepatic histological data were obtained in 27 patients, including almost one-third of subjects with LSM >8 kPa and all subjects with LSM >13 kPa, for which the aetiology of liver involvement was definitely established.

Among the 80 subjects with LSM values comprised between 8 and 13 kPa, the most frequent attributable cause was NAFLD (n=52, 58.4%), histologically confirmed in eight patients who showed liver fibrosis in all cases except one. NAFLD was the most frequent cause of intermediate fibrosis in our study. Our results confirm reports showing that NAFLD has become a leading cause of chronic liver disease in the western world, with an estimated prevalence of up to 30%.3 23 24 Advanced hepatic disease may be present in individuals with normal serum aminotransferase levels and normal ultrasound examination.25–28 So far, liver biopsy has been the gold standard for the assessment of fibrosis in patients with chronic liver disease. However, in cases of suspicion of NAFLD, liver biopsy is difficult to recommend for apparently healthy patients with normal liver tests, because of its possible complications. In one patient with LSM value of 10.3 kPa and normal liver enzymes, the fibrosis score was F2. In addition, fibrosis scores of F1–F2 were observed in all NAFLD subjects undergoing a liver biopsy. Although the data of the present study indicate that TE may detect NAFLD with fibrosis and isolated steatosis in apparently healthy subjects, the real value of TE in patients with early forms of NAFLD remains to be determined.

In conclusion, our study validated LSM as an effective procedure to screen for cirrhosis in the general population. It also suggests that a relatively high percentage of liver diseases remain undiagnosed in apparently healthy subjects and that LSM might contribute to referring these patients to hepatologists. Not only is the procedure well accepted, but also the results, when explained to patients, have an obvious psychological impact that help to convince them to accept further investigations. The cost-effectiveness of this type of screening should now be compared with the current screening procedures for specific causes of liver diseases.


The Fibroscan device was made available for the study by Echosens (Paris, France). The authors thank Céline Fournier and Linda Adara for assistance in setting up TE in the social medical centre and Christine Coulondou for technical assistance. The authors thank Professor J.C. Trinchet for his helpful comments on the manuscript.



  • See Commentary, p 883

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the CCPPRB, Aulnay-sous-Bois, France.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles