Article Text
PDF
Abstract
Objectives The number of elderly patients diagnosed with hepatocellular carcinoma (HCC) is expected to increase. We compared the presenting features and outcome of HCC in elderly (≥70 years) and younger patients (<70 years).
Design Multicentre retrospective cohort study and nested case–control study.
Patients 614 elderly and 1104 younger patients from the ITA.LI.CA database, including 1834 HCC cases consecutively diagnosed from January 1987 to December 2004. Both groups were stratified according to treatment: hepatic resection, percutaneous procedures, transarterial chemoembolisation (TACE). Survival was assessed in the whole population and in each treatment subgroup. Age, sex, aetiology, cirrhosis, comorbidities and cancer stage (CLIP score) were tested as predictors of survival. In each subgroup, differences in patient survival were also assessed after adjustment and matching by propensity score.
Results Ageing was associated with a higher prevalence of comorbidities, better liver function and CLIP score. Regardless of age, two-thirds of patients underwent radical treatments or TACE. Elderly patients underwent more ablative procedures and fewer resections or TACE sessions. The survival of elderly and younger patients was comparable in each treatment subset, and was predicted by CLIP score. This result was confirmed by the propensity analysis.
Conclusions The overall applicability of radical or effective HCC treatments was unaffected by old age. However, treatment distribution differed, elderly individuals being more frequently treated with percutaneous procedures and less frequently with resection or TACE. Survival was unaffected by age and primarily predicted by cancer stage, assessed by the CLIP system, both in the overall population and in treatment subgroups.
- Hepatocellular carcinoma
- elderly
- cirrhosis
- liver
- staging
- cirhosis
- elderly
- liver
- staging
- CT
- computer tomography
- HBV
- hepatitis B virus
- HCC
- hepatocellular carcinoma
- HCV
- hepatitis C virus
- MRI
- magnetic resonance
- NASH
- non-alcoholic steatohepatitis
- PEI
- percutaneous ethanol injection
- RF
- radio-frequency ablation
- US
- ultrasonography
- TACE
- transarterial chemoembolisation
Statistics from Altmetric.com
- CT
- computer tomography
- HBV
- hepatitis B virus
- HCC
- hepatocellular carcinoma
- HCV
- hepatitis C virus
- MRI
- magnetic resonance
- NASH
- non-alcoholic steatohepatitis
- PEI
- percutaneous ethanol injection
- RF
- radio-frequency ablation
- US
- ultrasonography
- TACE
- transarterial chemoembolisation
Introduction
Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, and generally occurs in a cirrhotic liver.1 Its incidence has a marked geographical variability, being particularly high in South-East Asia and southern Africa, and lower in North America and northern Europe.1 Recent studies, however, have shown a rising incidence over the last decades in low-risk areas.1 2 Similarly, the peak age of HCC occurrence varies considerably worldwide, due to the different distribution of its three main aetiological factors; that is, hepatitis B (HBV) and C (HCV) virus infections and alcohol abuse, which have different age peaks of exposure and dissimilar intrinsic mutagenic properties requiring different times for the neoplastic transformation of the liver.3 Namely, in developed countries, where HCV and alcohol account for most HCCs, the median age of patients is over 60 years.4 Moreover, the number of elderly or even extremely elderly patients diagnosed with HCC is expected to increase in most developed areas owing to the following epidemiological phenomena: (1) the decreasing incidence of HBV infection, which is frequently acquired in the first two decades of life, thus anticipating the occurrence of HCC with respect to other aetiologies5; (2) the ‘delaying’ effect of anti-viral therapy on the occurrence of this tumour in HBV and HCV infected cases6; and (3) the rising incidence of ‘cryptogenic’ and non-alcoholic steatohepatitis (NASH)-related cirrhosis, which have a long-lasting and insidious development.7
Many elderly individuals can be considered clinically ‘fragile’ due to their lifetime accumulation of different diseases, making it possible that HCC treatments are less applicable or less useful in elderly patients than in younger ones. This assumption is contradicted, however, by some studies indicating that the outcome essentially depends on a correct selection of patients for HCC treatments.8–13 However, most of these results come from small series of patients managed in tertiary referral centres, making their ability to reflect what occurs in clinical practice somewhat questionable.14 15 Moreover, in all these studies younger and elderly patients differed in several prognostic factors. Therefore, compelling evidence that the results of HCC treatments are age-independent is still lacking.
We re-addressed this topic, analysing a large database generated by the recruitment and management of 1834 patients with HCC, in ten Italian centres with different levels of experience, and using the propensity score analysis to match younger and elderly patients for confounding factors.
Methods
Patients
At the time of data analysis, the ITA.LI.CA database included 1834 patients with HCC consecutively seen from January 1987 to December 2004 in ten medical institutions. These data were collected prospectively and updated every 2 years. After data collection and before statistical evaluation, the consistency of the dataset was checked by the group coordinator (F.T.) and, if clarifications or additional information were needed, the data were resubmitted to each centre.
The eligibility criteria for the present study were the availability of information regarding patient age, type of treatment and survival. These inclusion criteria were fulfilled by 1757 cases. Thirty-nine patients were excluded since they underwent liver transplantation, elderly patients being precluded from this treatment in Italy. Thus, the final analysis included 1104 younger individuals. We categorised patients as ‘elderly’ if aged ≥70 years (614 patients), and ‘younger’ if aged <70 years (1143 patients). This break-point was chosen because it permitted a comparison with the results of other relevant papers and, above this cut-off, the number of resections and TACE sessions would dramatically decrease (among the ‘extremely elderly’ patients (aged ≥75 years), only 14 and 46 individuals underwent surgery and TACE, respectively), diminishing the suitability of the comparison with the younger counterpart.
Aetiology of liver disease
The cause of liver disease was classified as follows: (1) HBV, if patients were HBsAg-positive; (2) HCV, if patients were serum antibody anti-HCV positive; (3) alcoholic, if the daily ethanol intake was >60 g for women and >80 g for men for more than 10 years; (4) multi-aetiology, if there was a combination of these causative factors; and (5) others, when the cause was different from those cited above.
Diagnosis and severity of cirrhosis
Cirrhosis was confirmed by histology in 555 patients and by laparotomy/laparoscopy in 40. Otherwise, the diagnosis was made unequivocally by clinical findings, nodular margins of the liver at ultrasonography (US), endoscopic and ultrasonographic signs of portal hypertension, and laboratory features. The severity of liver dysfunction was scored according to Child–Pugh classification.16
Diagnosis and staging of HCC
The diagnosis was based on histology or cytology in 603 patients. Otherwise, diagnosis was based on the diagnostic criteria of the Italian17 and, after 2000, the Barcelona conference guidelines.18
HCC was staged by CT scan and/or MRI and was classified as (1) solitary, (2) paucifocal (≤three nodules), (3) multifocal (>three nodules), and (4) massive/infiltrating.19
The size of expanding nodules was also measured (the largest one in multinodular tumours). All patients had a chest x-ray, while additional investigations to detect metastases were performed when extrahepatic involvement was suspected. Cancer stage was scored according to the latest version of the CLIP system.20 The BCLC system was not used since many cases lacked information on performance status, as this system was not widely used before its endorsement by the AASLD guidelines in 2005.21
Serological testing
Liver tests (prothrombin activity, plasma albumin, and bilirubin) and α-fetoprotein serum levels (AFP; reference range, 0–20 ng/ml) were determined by conventional methods. Hepatitis viral markers were tested using commercial kits.
Treatment
The criteria for qualifying patients for hepatic resection, percutaneous ethanol injection (PEI), radio-frequency thermoablation (RF) and transarterial chemoembolisation (TACE) have been reported in detail elsewhere.22 Patients excluded from these treatments were given systemic chemotherapy or anti-oestrogen agents (generally, tamoxifen until its inefficacy was demonstrated) or palliation.
Comorbid illnesses
Cardiovascular, renal, pulmonary, gastro-intestinal, metabolic, haematological and neurological diseases with potential impact on the prognosis were recorded.
Statistical analysis
Continuous data are expressed as mean±SD and discrete variables as absolute and relative frequencies. The Mann–Whitney U test and the Pearson χ2 test were used to compare continuous or discrete data, respectively.
Survival was calculated from the time of cancer diagnosis to death with values censored at the date of the last follow-up, and expressed as median and 95% CI. Life table estimates were calculated according to the Kaplan–Meier method and compared by the log rank test. For survival analyses, patients who underwent different treatments were allocated to the main treatment performed.
For the analysis of predictors of survival, we considered only variables available in more than 90% of patients. The tested variables were age, sex, aetiology, presence of cirrhosis, comorbid illnesses and CLIP score. Comorbidity was included in the statistical analysis as absent or present. Child–Pugh class, AFP level, gross pathology of the tumour and vascular invasion were not tested as they are collinear to the CLIP staging system, being its components. Variables associated (p≤0.10) with survival at univariate analysis were entered in the Cox multivariate stepwise regression model to identify the independent prognostic factors. The adjusted relative risk (hazard ratio, HR) and 95% CI were calculated for each independent predictive factor.
Propensity analysis was carried out using logistic regression to create a propensity score for elderly and younger patients. Variables entered in the propensity model were sex, aetiology, presence of comorbid illnesses, cirrhosis and CLIP score. The model was then used to provide a one-to-one match between elderly and younger patients by using the nearest-neighbour matching method.23 Survival analysis was repeated in each matched subgroup in order to assess the impact of age on mortality amended from confounding factors.
A two-tailed p value <0.05 was considered statistically significant. All statistical analyses were performed using the SPSS 13·0 statistical package (SPSS Incorporated, Chicago, Illinois, USA).
Ethics
The clinical data were aggregated in the ITA.LI.CA database and managed according to the current Italian privacy legislation (all personal data were blinded in the general database). The general guarantor (F.T.) ensured protection and confidentiality of the original data and checked any information generated by their analysis. This procedure was notified to the Ethics Committee of the Institution of the database guarantor. The retrospective study conformed to the ethical guidelines of the Declaration of Helsinki.
Results
Whole population
Symptomatic HCCs were less common in elderly patients than in their younger counterparts (82 (13.3%) vs 232 (21%) cases, respectively (p<0.001)) since in the former the tumour was more frequently diagnosed under surveillance programmes (286 (46.6%) vs 564 (51.1%) patients, respectively (p=0.070)), or incidentally (246 (40.1%)) vs 308 (27.9%) cases, respectively (p<0.0001)).
The demographic and clinical characteristics of patients are summarised in table 1. Elderly patients were more likely to be females and infected by HCV, whereas HBV infection and combined aetiology were less common. Old age was also associated with a greater prevalence of comorbid illnesses, a better Child–Pugh class distribution and a higher prevalence of CLIP=0. Treatment allocation also differed since elderly patients underwent hepatic resection and TACE less frequently and ablative procedures more often than younger individuals.
Demographic and clinical characteristics of all patients
Over a median follow-up of 15 months (95% CI, 6 to 32 months), 1214 (70.7%) patients died, of whom 407 (66%) were elderly and 807 (73.1%) were younger subjects. The median survival was comparable in the two groups (elderly 25 months (95% CI, 22.3 to 27.7), younger 25 months (95% CI, 23.1 to 26.1), p=0.796). The survival rates at 1, 3, 5, 7 and 10 years were 70.1%, 34.6% 15.5%, 8.7% and 6.0% vs 71.1%, 35.3%, 18.7%, 9.1% and 6.7%, respectively (figure 1).
Overall survival of unselected elderly and younger patients.
Hepatic resection
Forty-three elderly and 142 younger patients underwent resection (table 2). They significantly differed in comorbid illnesses, which were more common in the former.
Demographic and clinical characteristics of all patients treated with hepatic resection
The median survival tended to be higher in elderly patients than in younger ones (52 months (95% CI, 40.1 to 63.9) vs 47 months (95% CI, 41.0 to 54.0), respectively, p=0.070). The survival rates at 1, 3, 5 and 7 years were 95.2%, 67.3%, 44.8% and 44.8% vs 86.5%, 61.6%, 32.4% and 20.1%, respectively (figure 2A).
Overall survival of elderly and younger patients treated with hepatic resection before (A) and after (B) matching with the propensity analysis.
At univariate analysis, CLIP score (p<0.001) and age (p=0.070) were associated with survival. At multivariate analysis only a CLIP score of 3 was independently associated with the risk of death (HR 4.81 (95% CI, 1.90 to 12.21) compared to CLIP=0).
The propensity analysis matched 32 elderly patients to 32 younger individuals (table 3). After matching, the median survival tended to be better in elderly than in younger patients (52 months (95% CI, 40.1 to 63.9) vs 42 months (95% CI, 37.1 to 46.8), p=0.087). The survival rates at 1, 3, 5 and 7 years were 96.8%, 68.3%, 41.9% and 21.0% vs 87.1%, 66.8%, 28.9% and 11.6%, respectively (figure 2B).
Demographic and clinical characteristics of patients in each therapeutic subgroup after matching with the propensity analysis score
Percutaneous ablation
The 195 elderly and the 230 younger patients treated with PEI or RF differed in several variables (table 4). Elderly patients were more likely to be female and infected with HCV, and had a better distribution of Child–Pugh classes and CLIP scores.
Demographic and clinical characteristics of all patients treated with percutaneous ablative therapy
Nonetheless, the median survival was equivalent, being 42.0 months (95% CI, 35.4 to 48.6) in elderly subjects and 42 months (95% CI, 36.0 to 48.0) in younger patients (p=0.797). The survival rates at 1, 3, 5 and 7 years were: 90.1%, 53.4%, 29.0% and 16.6% vs 89.9%, 52.9%, 35.1% and 18.7%, respectively (figure 3A).
Overall survival of elderly and younger patients treated with percutaneous ablation techniques before (A) and after (B) matching with the propensity analysis.
CLIP score (p<0.001) and comorbid illnesses (p=0.050) predicted survival at univariate analysis. Moreover, CLIP score emerged as an independent predictor of mortality: compared with CLIP=0, the HR was 1·81 (95% CI, 1.31 to 2.49) for CLIP=1; 2.99 (95% CI, 1.97 to 4.54) for CLIP=2; and it was 3.82 (95% CI, 2.07 to 7.06) for CLIP=3.
The propensity analysis matched 119 pairs of cases (table 3). After matching, the median survival did not differ between the two groups (42 months in elderly patients (95% CI, 34.0 to 50.0) and 43 months in younger patients (95% CI, 36.4 to 49.6)), p=0.255). The survival rates at 1, 3, 5 and 7 years were 91.1%, 51.1%, 28.8% and 14.1% vs 91.1%, 58.2%, 37.3% and 22.1%, respectively (figure 3B).
TACE
The 158 elderly and 396 younger patients undergoing TACE differed for Child–Pugh class and CLIP score (table 5). Elderly patients had a better distribution of CLIP scores and more often belonged to Child–Pugh class A than their counterparts.
Demographic and clinical characteristics of patients treated with transarterial chemoembolisation combined
The median survival was similar in the two groups: 26 months (95% CI, 19.5 to 32.5) in elderly patients and 27 months (95% CI, 24.5 to 29.5) in younger individuals (p=0.730). The 1-, 3-, 5- and 7-year survival rates were 79.4%, 36.4%, 6.4% and 2.4% vs 78.9%, 32%, 13.3% and 4.5%, respectively (figure 4A).
Overall survival of elderly and younger patients treated with transarterial chemoembolisation before (A) and after (B) matching with the propensity analysis.
Sex (p=0.097), aetiology (p=0.001) and CLIP score (p<0.001) were associated with survival at univariate analysis. Only CLIP score was an independent predictor of mortality: compared with CLIP=0, the HR was 1·63 (95% CI, 1.25 to 2.13) for CLIP=1; 2.14 (95% CI, 1.78 to 3.28) for CLIP=2; 4.15 (95% CI, 2.82 to 6.12) for CLIP=3; and 6.16 (95% CI, 3.66 to 10.37) for CLIP 4–6.
The propensity analysis matched 134 pairs of patients (table 3). After matching, the median survival was comparable between the two groups (27.5 months (95% CI, 19.4 to 35.6) in elderly subjects and 27 months (95% CI, 22.7 to 31.3) in younger patients (p=0.391)). The 1-, 3-, 5- and 7-year survival rates were 81.4%, 38.5%, 6.9% and 2.6% vs 87.6%, 30.6%, 19.8% and 9.0%, respectively (figure 4B).
Untreated patients
The 218 elderly and 336 younger patients not qualified for radical or effective treatments differed for several variables. Elderly patients were more likely to be females (p<0.001) and infected by HCV (p<0.001), and less often infected by HBV (p=0.045), or alcoholics (p=0.039). Moreover, in elderly patients comorbidity was more common (p<0.001) and Child–Pugh class C less represented (p<0.001). Finally, solitary HCCs (p<0.001) and CLIP=0 (p=0.002) were more frequent, while CLIP=4–6 was less common (p=0.030) than in younger cases.
The median survival was 10 months (95% CI, 8.2 to 11.8) in elderly patients and 8 months (95% CI, 6.3 to 9.7) in younger individuals (p=0.999). The 1-, 3- and 5-year survival rates were 43.0%, 16.3% and 4.5% vs 38.1%, 16.3% and 8.3%, respectively figure 5A).
Overall survival of elderly and younger patients treated with palliation or other treatments before (A) and after (B) matching with the propensity analysis.
Sex (p<0.001) and CLIP stage (p<0.001) were associated with survival at the univariate analysis. Again, at multivariate analysis only CLIP was independently associated with mortality. If compared with CLIP=0, CLIP=1 did not show any increase of risk, while the HR rose to 2.07 (95% CI, 1.42 to 3.02) for CLIP=2; to 3.0 (95% CI, 2.04 to 4.41) for CLIP=3; and to 6.41 (95% CI, 4.36 to 9.43) for CLIP=4–6.
The propensity analysis matched 129 younger patients to 129 elderly individuals (table 3). The median survival was comparable between the two groups (p=0.789), being 10.0 (95% CI, 8.5 to 11.5) months in elderly and 9.0 months (95% CI, 6.1 to 11.9) in younger patients. The 1-, 3- and 5-year survival rates were 40.8%, 15.1% and 4.6% in elderly patients, and 46.1%, 12.7% and 5.6%, respectively, in younger individuals (figure 5B).
Discussion
The prevalence of elderly subjects in our patient population was higher than in other previous unselected series.10 13 This was possibly due to both a lower proportion of HBV cases and the shift towards older ages of HCC occurrence in Italy, where the incidence of this cancer is declining due to the fading ‘cohort effect’ of HCV infection, mainly acquired during the 1950s and 1960s via iatrogenic routes.
Our elderly and younger HCC patients differed in several features. Elderly patients were more likely to be women and infected by HCV, whereas HBV and multi-aetiological cases were less common. This imbalance was expected since all these factors are known to influence the age at which HCC develops: (1) the peak age of women is delayed by 5 years compared to men24; (2) HCV infection is generally acquired during adult life,25 while HBV infection frequently occurs in early age26; and (3) a multifactorial aetiology accelerates the evolution of chronic liver disease, anticipating the appearance of cancer.27
Elderly patients also showed a better HCC stage, probably due the lower prevalence of symptomatic HCCs, which generally belong to advanced stages.28 The better tumour stage found in our elderly patients conflicts with previous reports describing no significant differences with respect to younger patients.9 10 12 13 As these studies did not report the circumstances leading to HCC diagnosis, a different proportion of cancers disclosed at a subclinical stage could explain the discrepancy.
Conflicting results have been reported as to whether aged patients develop HCC while their underlying liver disease is at a less advanced stage. A more preserved liver function and a lower grade of portal hypertension were found in patients aged ≥80 years than in those aged 50–60 years.12 Conversely, other studies did not find any difference in liver function tests and Child–Pugh grading across the cut-offs of 65 or 70 years.9 10 13 Our elderly patients included more Child–Pugh A cases at the expense of advanced cirrhosis as compared to younger individuals. As our series is the largest that has been published so far and includes unselected cases enrolled by ten centres scattered throughout Italy, it is conceivable that our results faithfully reflect what actually occurs in clinical practice.
Despite a higher prevalence of comorbidities and a mean age higher by 14 years, elderly patients showed a 5-year survival similar to that of their younger counterparts. The reason for such an unexpected result could be due to the low survival rate of both groups (<20% at 5 years) indicating that HCC occurrence outweighs the impact of both comorbidity and age per se on life expectancy.
Interestingly, about two-thirds of our patients underwent curative treatments or TACE, regardless of their age. This figure is close to the 64% reported by Poon et al in 199910 and is much better that those reported for aged patients recruited up to 1992 and 1997.9 13 This would confirm that the proportion of treatable HCCs is increasing, probably due to a growing use of surveillance programmes for patients at risk in clinical practice.29 However, the patient's age did affect the treatment strategy, as the elderly underwent percutaneous procedures more often and resections and TACE less frequently. There may be several reasons for the shift towards less aggressive and risky treatments in old patients, including hepatic resection and TACE being less feasible due to comorbidities, a more prudent therapeutic approach by physicians, and a more frequent refusal of aggressive treatments by the patient.
The key issue addressed by our study is whether the outcome of HCC treatments is affected by age. In fact, even for this topic there is no agreement since the outcome was found to be either age-independent8 10 or adversely affected by ageing.9 12 13 The latter evidence, however, could be biased by an ‘under-treatment’ of older individuals since, despite liver function and cancer stage that were no worse than in younger individuals, a conservative approach was chosen for most elderly patients. Lastly, none of the available studies adjusted survival rates for the confounding effect caused by the uneven distribution of prognostic factors between elderly and younger cases.
The present study provides compelling evidence that age does not affect the outcome of HCC treatments. Our data, obtained by means of a retrospective analysis of a large patient cohort, was confirmed by a nested case–control study matching elderly and younger cases for the main confounding factors. We therefore confirm that the adverse effect of ageing, observed in some series, can essentially be attributed to under-treatment of aged patients.
In conclusion, in patients diagnosed with HCC: (1) life expectancy is unaffected by age over 70 years; (2) the overall applicability of radical or effective therapies for HCC is also unaffected; (3) the distribution of these therapies, however, differs according to age, elderly individuals being more frequently submitted to percutaneous procedures and less frequently to hepatic resections or TACE; and (4) in each treatment subgroup, the prognosis is similar in elderly and younger patients and basically determined by cancer stage, assessed with the CLIP system.
This finding would rule out an intrinsic detrimental impact of advanced age on prognosis, and supports physicians in selecting treatment for HCC according to cancer stage and general condition of the patient, regardless of his/her age.
What is already known about this subject
The number of elderly or even extremely elderly patients diagnosed with hepatocellular carcinoma (HCC) is expected to increase in a number of developed countries
The results of HCC treatments in these patients are still controversial and, in particular, it is unclear whether old age per se affects the outcome of the available therapeutic options
Most of the available data come in fact from small and heterogeneous patient series, making their ability to reflect what occurs in clinical practice questionable
What are the new findings
This retrospective multicentre cohort study (refined by a nested case–control study) enrolled 614 elderly (≥70 years) and 1104 younger unselected patients. The study shows that:
The life expectancy of patients with HCC is unaffected by age ≥70 years
Although the overall application rate of curative or effective treatments is unaffected by age, their distribution differs according to age. Elderly patients are more frequently treated with percutaneous ablation and less frequently with hepatic resections or chemoembolisation
Patient survival is unaffected by age in each treatment subgroup, even after adjustment for the confounding factors with the propensity score analysis
Cancer stage, assessed by the CLIP scoring system, predicts patient outcome in all therapeutic subsets
How might it impact on clinical practice in the foreseeable future
The evidence that an age ≥70 years does not influence the outcome of curative and effective treatments of HCC, which is basically determined by the cancer stage, gives support to physicians in selecting the treatment according to the cancer stage and general condition of the patient, regardless of his/her age
Appendix
Other members of the ITA.LI.CA group: Dipartimento di Medicina Clinica, Alma Mater Studiorum – Università di Bologna: Paolo Caraceni, Marco Domenicali, Alice Grignaschi, Silvia Li Bassi; Dipartimento di Medicina Interna dell'Invecchiamento e Scienze Nefrologiche, Alma Mater Studiorum – Università di Bologna: Donatella Magalotti; Divisione di Medicina, Azienda Ospedaliera Bolognini, Seriate: Claudia Balsamo, Mariella Molaro; Divisione di Medicina, Ospedale Treviglio-Caravaggio, Treviglio: Lodovico Gilardoni, Mario Mattiello; Dipartimento di Medicina Clinica e Sperimentale, Clinica Medica V, Università di Padova: Alfredo Alberti, Angelo Gatta, Maurizio Gios; Cattedra di Medicina Interna II, Università Cattolica del Sacro Cuore di Roma: Luisa Siciliani; Cattedra di Malattie dell'Apparato Digerente, Università di Padova: Anna Baldan, Dario Marino, Adriana Sergio; Dipartimento di Medicina, Unità di Gastroenterologia, Ospedale Fatebenefratelli, Milano: Mariella Molaro, Maurizio Sala; Unità di Gastroenterologia, Ospedale Belcolle, Viterbo: Giorgia Ghittoni, Paola Roselli; Dipartimento di Urgenza/Emergenza Chirurgia generale e dei Trapianti, Alma Mater Studiorum-Università di Bologna: Gian Luca Grazi, Matteo Ravaioli, Alessandro Cucchetti, Antonio Daniele Pinna; Dipartimento di Malattie Apparato Digerente e Medicina Interna, Unità di Radiologia, Policlinico S. Orsola-Malpighi, Bologna: Emanuela Giampalma, Rita Golfieri.
References
Footnotes
Funding This study was partially funded by a grant (Ricerca Fondamentale Orientata 2006-2007) from the Ministero della Istruzione, della Università e della Ricerca (MIUR).
Competing interests None.
Ethics approval This study was conducted with the approval of the Ethics Committee of Policlinico S. Orsola-Malpighi, Bologna.
Provenance and peer review Not commissioned; externally peer reviewed.