Hepatitis C is caused by infection with HCV, an RNA flavivirus. In the haemophilia community, HCV was transmitted through clotting factor concentrates. Almost all haemophilia patients who were treated with large-pool coagulation factor concentrates before the mid-1980s, have been infected with HCV. In the majority, this has led to liver disease i.e. chronic hepatitis C.
A discussion of current treatment options for chronic hepatitis C is outside the scope of this monograph. A recent practice guideline can be found at http://www.aasld.org. Although it is commonly stated that haemophilia patients tolerate treatment for chronic hepatitis C less well than other patient groups (e.g., ), a recent meta-analysis of studies performed in people with haemophilia concluded that results (rates of sustained response) are similar to those in the general population .
The major complication of chronic hepatitis C is advanced fibrosis and cirrhosis, which develops in 20–30% of patients over 20–30 years . This long time between infection and development of complications means that, although the rate of new hepatitis C infection has sharply declined (to virtually zero in people with haemophilia after the introduction of viral inactivation steps for clotting factor concentrates), the prevalence of cirrhosis caused by chronic hepatitis C is still increasing.
A number of systems are used to grade fibrosis. Most of them score ‘no fibrosis’ as ‘0’ and ‘cirrhosis’ as ‘4’. Advancing grades of fibrosis are scored ‘1’, ‘2’ and ‘3’, but the exact definitions differ in the different systems .
The risk of cirrhosis is higher in patients who are older at the time of infection, co-infected with hepatitis B virus or HIV and who are male. Moreover, concomitant other liver disease, obesity, diabetes and alcohol consumption increase the risk of cirrhosis. The rate of development of cirrhosis seems to be lower in African Americans .
Haemophilia populations are very informative on the natural history of hepatitis C, because the time of infection is known (the first exposure to large pool concentrates). They confirm that progression to cirrhosis and overt liver disease is more common in patients who were older at the time of infection. A specific problem in haemophilia is co-infection with HIV, which strongly increases the risk of progression to cirrhosis and development of end stage liver disease. Hazard ratios for progression to end stage liver disease are as high as 8–14 in HIV-positive when compared with HIV-negative patients [5,6].
In HIV-negative patients, the prevalence of symptomatic liver disease was reported to be 3–14% after 16–35 years. The same cohort studies reported liver related death in 0–3% of patients [5–8]. A recent study in a Canadian cohort, using a Markov model incorporating the effect of treatment, predicted that at 20 years after infection, 37% of haemophilia patients would have cirrhosis, 12% hepatocellular carcinoma and 19% would have died from HCV . However, this is probably a major overestimate as in reality most haemophilia patient cohorts have been infected for almost 30 years and have not demonstrated these outcomes.
Although some studies saw more liver disease in patients with more severe bleeding disorders, this was not reported by all. This could be a methodological issue: patients with more severe bleeding disorders often had a longer duration of infection, because they were generally treated and thus infected at an earlier age.
Diagnosis of cirrhosis
The gold standard for the diagnosis of cirrhosis is a liver biopsy, although sampling error is a problem. In haemophilia, there is also the issue of bleeding risk and cost of substitution therapy. Both in persons with haemophilia and others, there is a lot of interest in non-invasive methods to diagnose fibrosis and cirrhosis . Cirrhosis may well be evident on ultrasound (US), if there is irregularity of the liver surface or nodularity of the liver , but can be missed. The most well-tested non-invasive options are FibroTest, which is a panel of five biochemical markers and FibroScan which uses transient elastography, an ultrasound-based technique, to measure liver stiffness. In a meta-analysis of diagnostic accuracy (vs. biopsy as the gold standard), the sensitivity of FibroScan was 64% for F2-4 fibrosis and 86% for cirrhosis. For FibroTest, data were difficult to summarize because different cut-off levels were used. If sensitivity above 80% was required, specificity dropped to 40–60% . FibroTest and FibroScan have been tested in haemophilia, but without comparison with liver biopsy [12–14]. When the two tests were compared with each other in haemophilia, concordance was not very good in F2–3 fibrosis, but reasonable (85%) in cirrhosis . Non-invasive tests overcome the issue of bleeding with liver biopsy, but not completely that of costs: FibroTest is only available commercially and cannot be performed in-house and FibroScan requires investment in expensive equipment. The American Association for the Study of Liver Diseases (AASLD) does not recommend the use of the currently available non-invasive tests instead of liver biopsy in routine clinical practice .
With the present options for treatment, the main question in clinical practice is whether there are signs of cirrhosis and this question can be answered by a combination of routine liver tests and a routine ultrasound in most patients.
Hepatitis C-related hepatocellular carcinoma
Once patients have developed cirrhosis, they are at risk for liver cancer, i.e. hepatocellular carcinoma (HCC). HCC is the leading cause of death in patients with cirrhosis because of hepatitis C. In the western world, the incidence of HCC is clearly increasing because of chronic hepatitis C. In 2000, 60–70% of HCC in Europe and 50–60% in North America was related to hepatitis C . In other parts of the world, the background incidence of HCC is higher because of chronic hepatitis B and exposure to toxins.
In large studies, the rate of development of HCC was 3–6% per year in patients with HCV cirrhosis. In patients with advanced fibrosis, the rate is approximately half of that [16,17]. In less advanced hepatitis C, the risk of HCC seems very low . An Italian study in persons with haemophilia, performed in the 1990s, reported six cases of HCC in 384 patients with chronic hepatitis C during 4 years of follow-up or 0.4% per year. All cases occurred in the 40 patients who had cirrhosis at baseline .
Risk factors for HCC in patients with HCV coincide with the risk factors for progression of HCV chronic hepatitis to cirrhosis. These factors include older age, older age at the time of acquisition of infection, male gender, heavy alcohol intake, co infection with HBV or HIV, a transfusion-related mode of HCV acquisition and possibly diabetes and obesity . A recent meta-analysis showed that infection with genotype 1b may also be associated with an increased risk of HCC (relative risk of 1.78) .
Strategies for early detection
Patients with an increased risk of HCC are candidates for surveillance: periodic examinations [most often US or alpha fetoprotein (AFP)] to look for early, asymptomatic HCC. The rationale behind surveillance is that early HCC can often be treated, whereas advanced, symptomatic HCC has a very poor prognosis. Surveillance has become routine practice, although scientific evidence for its benefit is scarce. A number of uncontrolled cohort studies in cirrhosis (not specifically hepatitis C) showed improved survival [21,22]. Only one randomized controlled trial has been performed, in hepatitis B. In that study, HCC related mortality was reduced by 37% (83 vs. 132/100 000), using US and AFP . The main problem with uncontrolled studies of surveillance is lead time bias: the earlier a tumour is found, the longer survival seems, simply because we start counting at an earlier time point. Moreover, it is not known if all small HCC progress to clinical disease. Thirdly, the usefulness of early diagnosis is limited in patients with advanced liver disease or co-morbidity, who might not be candidates for curative treatment (as discussed below).
The AASLD guidelines recommend surveillance in all patients with hepatitis C in whom the annual risk of HCC exceeds 1.5%. This threshold is based on cost effectiveness analyses [24,25]. With an annual risk of 3–6%, surveillance is recommended in all patients with hepatitis C cirrhosis. No clear recommendations were given in patients with late stage (F3) fibrosis, although literature indicates that HCC risk is not negligible. It seems to be at least half of that in cirrhosis [16,17], which would cross the threshold of 1.5% per year. In patients with F1 (mild) or F2 (moderate) fibrosis, the risk of HCC is probably much lower.
The risk of HCC decreases in patients with cirrhosis who are treated with interferon-based therapies, most strongly when there is a sustained virological response. A recent meta-analysis reported a relative risk of 0.43 in treated patients when compared with untreated controls, and 0.35 in patients with a sustained response when compared with treated patients without response . However, the remaining absolute risk is still not completely clear and AASLD recommends ongoing surveillance in those patients. Surveillance is not required in patients who had not developed cirrhosis at the time of successful HCV treatment. Ultrasono-graphy of the liver is the best available tool for surveillance for HCC, although sensitivity and specificity are limited at 65–80% and 90% respectively. Other limitations of the technique are operator dependency, decreased quality in obese patients and decreased sensitivity in patients with cirrhosis. Periodic measurement of serum AFP is only recommended if ultrasonography is not available: there is no single cut-off level that is both sensitive and specific enough for the presence of HCC. However, because of the limitations of US, many clinicians still favour the combination of US and AFP. A sudden rise of AFP and/or a high level of AFP deserves further radiological diagnostics (4-phase CT scanning) in case US is not conclusive.
The interval between ultrasounds is determined by the growth rate of the tumour: the aim is to diagnose HCC between its earliest visibility on US and the time it has reached 2 cm in diameter. From biological studies, this window is 6–12 months. Most clinical evidence does not show added benefit for surveillance intervals of 6 months over 12 months and AASLD’s recommendation to screen at 6 months’ interval is based on data in hepatitis B.
Evidence in haemophilia. Santagostino et al. performed a non-randomized, two-arm study in persons with haemophilia, in which they compared surveillance intervals of 6 and 12 months . They used both US and AFP. More cases of HCC were diagnosed in the 6-month group (0.40% vs. 0.14% per year), but in both groups tumours were multinodular and long-term survival was only seen in patients who had undergone orthotopic liver transplantation (OLT). Too few HCC were diagnosed for a meaningful comparison of the two strategies: in the 12-month group, two patients died and one was a long-term survivor; in the 6-month group, one patient was recently diagnosed, one died, one was on the waiting list for OLT and two were long-term survivors. The Santagostino study was designed after an earlier cohort study by the same group tested annual screening with US and AFP . In this study, all HCC were late stage disease without options for curative treatment. It should of course be noted that treatment options have increased after these two studies were performed.
Recommendation. We perform yearly US combined with twice-yearly AFP measurement in all haemophilia patients with chronic hepatitis C, including those in whom cirrhosis has not been diagnosed. We do this because fibrosis without cirrhosis is also associated with HCC and because present diagnostic methods (including non-invasive tests) cannot reliably exclude cirrhosis. We also continue surveillance in patients who have successfully been treated for HCV, as we did not exclude cirrhosis in most of them before treatment.
We discuss here the diagnosis of HCC in the setting of patients who have an a priori high risk of HCC: patients with cirrhosis due to hepatitis C. Strategies are different when HCC suspected in a patient with an a priori low risk, this is outside the scope of this monograph.
The diagnosis of HCC is different from most other cancers, as histology is not required if risk factors (i.e., cirrhosis) are present and imaging is typical. HCC exclusively receives arterial blood supply through the arterial tumour vessels, and accordingly most HCCs are hypervascular on angiography and as seen in the arterial phase of contrast-enhanced imaging. However, this hypervascularity is not present in dysplastic nodules, and in the majority of cases also absent in early well-differentiated HCC. It follows that the diagnosis of HCC in a cirrhotic liver can be reliably made when contrast-enhanced CT or MRI show enhancement of a nodule in the arterial phase and less enhancement in the venous phase (relative to the surrounding liver tissue). When compared with the gold standards of histological examination of an explanted/resected liver, biopsy or follow-up, CT had a sensitivity of 68% and specificity of 93%. In the same meta-analysis, MRI had a sensitivity of 81% and specificity of 85% . Ultrasonography, which is most valuable in surveillance, is not sufficiently specific for diagnosis. A high level of AFP (e.g. >500 U L−1) may help in establishing the diagnosis; however, the level is often only slightly raised which does not discriminate between tumour and active hepatitis. The role of FDG-PET scanning is limited in initial diagnosis as only about half of tumours are positive. However, FDG-PET might be useful in staging the disease .
The aim of surveillance programmes as discussed above is to diagnose HCC in a stage that curative treatment can be offered. If symptoms occur, HCC is most often in an advanced stage. Such patients present with decompensation of previous compensated liver disease, pain, weight loss or an upper abdominal mass or with metastases in intra-abdominal lymph nodes, lung, bone and adrenal glands .
The diagnostic approach to a suspected HCC depends on the size of the lesion. Lesions smaller than 1 cm are usually not malignant. Small nodules comprise a broad range of entities, some benign, some with malignant potential, some clearly malignant. Careful study of pathological and clinical features of small nodular lesions in cirrhotic liver has shown the evolution from premalignant lesions (low and high grade dysplastic nodules) to early, well-differentiated HCC to moderately differentiated HCC . AASLD recommends that nodules smaller than 1 cm should be followed-up by ultrasound, at 3–6 months intervals. If they remain stable for 2 years, standard surveillance can be resumed.
The AASLD recommendations are more complicated for lesions larger than 1 cm. For these, either contrast enhanced CT or MRI is advised. If the image is typical, HCC is diagnosed. If it is not, the other modality is also used. If this results in a typical image, HCC is diagnosed. If this is also inconclusive, biopsy is needed. This approach has been validated by a number of studies. The main limitation is that 30–40% of HCC is missed on fine needle biopsy . Repeated biopsies are often necessary. Other problems of biopsy include the risk of needle track seeding (2.7% in a recent meta-analysis ) and the difficulty to differentiate HCC from high-grade dysplastic nodules on small biopsy samples. In persons with haemophilia, the risk of bleeding and requirement of coagulation factor concentrates need to be considered .
The most widely used staging system for HCC is the Barcelona Clinic Liver Cancer (BCLC) staging scheme (Table 1) .
Table 1. The Barcelona Clinic Liver Cancer (BCLC) staging system (35).
|O, very early stage||Single, <2 cm||–||0|
|A, early stage||Single or max 3 |
nodules <3 cm
|B, intermediate stage||Multinodular||A, B||0|
|C, advanced stage||Portal invasion, N1, M1||A, B||1–2|
|D, end stage|| ||C||>2|
Recommendation. We follow the AASLD recommendations for diagnosis. The diagnostic work-up and indications for biopsy are not different from those in patients without haemophilia.
Treatment of HCC
Removal of the tumour(s), prior to spreading outside the liver is the only option for cure. This can be achieved by surgical resection, local ablation or liver transplantation. The first two can only be considered in selected cases with one or two nodules and relative adequate function of the cirrhotic liver. Impaired liver function and regenerative capacity in combination with the precancerous condition of the liver make the outcome less than optimal.
Liver transplantation is in itself the best option as it both cures the cirrhosis and removes both malignant and premalignant lesions. However, patient characteristics, donor shortage and (potential) tumour spread outside the liver may preclude this option. If local ablation or resection are not feasible, most liver transplant centres only accept patients for liver transplantation if the tumour load is not outside the so-called Milan criteria: one solitary HCC lesion ≤5 cm or maximum three lesions ≤3 cm, no gross vascular invasion and no regional node or distant metastases .
The different treatment modalities are discussed in more detail below.
Only a small minority of patients with HCC in the setting of HCV infection are good candidates for resection, because most will have cirrhosis and liver dysfunction. Patients with cirrhosis but still well-preserved liver function can be eligible, if their bilirubin and portal blood pressure are normal. In that case, 5-year survival can exceed 70%, while in less rigorously selected patients, 5-year survival is about 50% .
Recurrence of HCC, either a true recurrence of the same tumour or de novo HCC, is eventually seen in up to 70% of patients who undergo resection. Adjuvant therapy, either before or after surgery, does not reduce this rate . Data on the treatment of recurrence are scarce, although liver transplantation might be an option in some patients.
Evidence in haemophilia. In persons with haemophilia, only a few cases of partial hepatectomy have been described, including one in a patient with an inhibitor to factor VIII (FVIII) [39–41]. No complications were reported.
Practical recommendations for partial hepatectomy. Before surgery, the presence of an inhibitor to FVIII or FIX should be excluded. Throughout surgery, substitution is aimed at FVIII or FIX levels between 80 and 100%. We use continuous infusion in the postoperative period, aiming at levels between 50 and 80% in the first 5 days. After that, we either continue continuous infusion with levels between 30 and 50% until 2 weeks postoperatively, or we switch to bolus injections with trough levels of 30%. We prefer continuous infusion for as long as patients are hospitalized, especially on surgical wards where nurses are not used to work with coagulation factor concentrates. Unless there are contra-indications (i.e., arterial disease or otherwise increased risk of thrombosis), we also use tranexamic acid (1 g trice daily) for the first 7 days. We start low molecular weight heparin for thromboprophylaxis on the first day after surgery, if there have been no bleeding complications and continue as long as factor levels are above 50%. Compression stockings worn during surgery and until the patient has completely mobilized.
Percutaneous ablation is a curative option for patients with small HCC who cannot undergo resection. It is also used in patients on the waiting list for liver transplantation. Tumours are injected with a chemical substance (most often ethanol) or with a transducer that either heats (radiofrequency ablation, RFA) or freezes the malignant cells. Most centres routinely use percutaneous ethanol injection (PEI) and/or RFA .
For PEI, multiple sessions of injections are often required to achieve maximal control of the HCC, and it performs less well in larger tumours. By contrast, RFA is effective in larger tumours, but has a higher rate of complications (including bleeding) and is more expensive .
A number of studies have compared PEI and RFA. A recent meta-analysis summarized the evidence as follows: in HCC <2 cm, there is no significant difference in survival using both methods. In larger HCC, RFA has lower recurrence rates and better survival . Likewise, the AASLD recommends RFA over PEI in tumours >2 cm.
Debate is ongoing whether RFA might even be non-inferior to resection in patients with HCC <2 or 3 cm [44,45]. In Western countries, resection is still the first option, although in Japan RFA may be offered as first choice.
Evidence in haemophilia. Published data on percutaneous ablation in persons with haemophilia are limited to a series of five PEIs. Before the procedure, coagulation factor concentrates were administered to achieve levels of 65–100% FVIII or FIX. After the procedure, levels >40% were maintained for another 2 days. There were no early complications, but one patient presented with gastrointestinal bleeding on day 4, for which no source was found .
Practical recommendations for percutaneous ablation. Our practice is to aim for 80–100% of FVIII/FIX during the procedure, and afterwards keep levels above 50% for 3 days and above 30% for another 2 days.
Liver transplantation in haemophilia has the added bonus of, in addition to potentially cure the HCC and cirrhosis, curing the coagulation defect. However, the indications for liver transplantation are exactly the same in persons with haemophilia as in others, including the above mentioned Milan criteria for acceptable tumour load. In the study that introduced these criteria, survival was 75% at 4 years . In a large multi-centre retrospective review, patients who satisfied the Milano criteria had a 5-year survival of 73%, compared to 54% in those who had larger tumours or macrovascular invasion .
In liver transplantation, the question is not just what the optimal treatment for an individual patient is. Given the scarcity of donor organs, the optimal use of available cadaveric livers must also be considered. To achieve fair allocation, livers are allocated based on objective criteria (serum bilirubin, serum creatinin, INR), which are combined in the Model for End-Stage Liver Disease (MELD) score . The MELD score is not easily calculated, as it uses logarithms, but calculators are available online (for instance on the United Network for Organ Sharing website, http://www.unos.org). After some discussion on the relative weight of HCC, patients are now given 22 MELD points. The waiting time for transplantation is considerable, depending on blood group, local waiting list and local availability of organs. A proportion of patients has progression of HCC or dies while on the waiting list. This has prompted the use of living donor transplantation. In this procedure, the right hepatic lobe of a healthy volunteer donor (close family member or spouse) is used . The advantages of a living donor are a shorter waiting period and elective surgery. A modelling study showed that a living transplantation increases life expectancy and cost-effectiveness when compared with cadaveric transplantation, as soon as the waiting time for a cadaveric transplant exceeds 7 months . There is one major downside: the risk to the donor. Estimated risk of complications is 20–40% and mortality is 0.3–0.5% .
Evidence in haemophilia. The first successful liver transplantation in haemophilia was performed in 1985 . The Birmingham haemophilia and liver centres reported a series of 11 liver transplants in haemophilia patients between 1990 and 2001. Five-year survival was nine of 11 (82%). Data on HCV recurrence were available in eight. Two developed cirrhosis at 1 and 3 years post-transplantation respectively. Four others had histological evidence of HCV hepatitis. Coagulation factor substitution was managed by continuous infusion and could be stopped at a median of 36 h after transplantation . Transplantation has also been performed in patients with inhibitors to FVIII . In 2005, the first case of living donor transplantation was reported in haemophilia .
Liver transplantation has now become standard practice in persons with haemophilia who have an indication for this procedure. This requires close collaboration between liver surgeon, hepatologist, anaesthesiologist and haematologist.
Practical recommendations for liver transplantation: In our centre, we formulate a plan for factor substitution before patients are placed on the waiting list. This plan is available to all team members, in the electronic patient file. An inhibitor is excluded at this time point, with repeat measurements at least every 6 months (in low risk patients with generally >1000 exposure days). Shortly before transplantation, FVIII or FIX concentrate is infused, aiming for levels of 100 and 80% respectively. After this initial bolus, a continuous infusion of 4 units per kg bodyweight per hour is started. A FVIII or FIX level is measured before the start of surgery. During transplantation, laboratory staff is available for repeat measures if surgery is complicated or haemostasis is insufficient. At the end of surgery and at least daily afterwards, factor levels are again monitored. Decrease of substitution is guided by these measurements.
Palliative options with proven efficacy (increased survival) are limited to trans-catheter arterial chemoembolization (TACE) and sorafenib. The AASLD recommends TACE in BCLC intermediate (B) stage HCC, and sorafenib in advanced (C) stage.
In TACE, chemotherapy (either doxorubicin or cisplatin in lipiodol emulsion) is infused directly in the hepatic artery. Subsequently, the blood vessel is embolized using small particles, thus combining cytotoxic and ischaemic damage to the tumour. A recent advance is combining both steps in the use of embolic particles that elute cytotoxic drugs .
Extensive tumour necrosis is seen after TACE in most patients, with objective responses in 20–60% and very rare complete responses. Necrosis causes fever, abdominal pain and ileus, from which patients normally recover in 2 days. TACE has been shown to improve survival, but the size of the gain depends heavily on patient characteristics. In patients with more advanced disease (i.e., BCLC stage C, especially those with portal invasion) the benefits do not outweigh the risk of complications .
Evidence in haemophilia. Four cases of TACE in persons with haemophilia have been described in the same paper quoted earlier for PEI . Here too, substitution was used for 2 days after the procedure. Moreover, no early complications were seen but 2/4 patients had late gastrointestinal bleeding.
We have used TACE twice, in a single patient with severe haemophilia A who had a longstanding inhibitor. He was treated with recombinant factor VIIa, 90 μg kg−1, for 3 days. During the procedure and the first 12 h afterwards, dosing was every 2 h. Afterwards, we decreased the interval between doses. The procedure was uncomplicated.
In advanced HCC (BCLC grade C), sorafenib offers a survival advantage of about 3 months [55,56], but little effect on time to symptomatic progression. Sorafenib is a multikinase inhibitor that inhibits, among others, the vascular and platelet-derived growth factors. It is taken orally twice daily. The main side-effects are a skin hand-foot syndrome and diarrhoea.
No other therapies than those discussed above have proven efficacy in HCC. Quite some studies have dealt with the use of Chinese herbal medicine, mostly as an adjunct to other therapies . Most of these studies have found a positive effect, but the quality of the evidence is generally low and these drugs are not used in main-stream medicine. A fair number of good quality trials have tested tamoxifen. A meta-analysis concluded that it has no effect survival .
The choice of treatment modality for HCC is not influenced by the presence of haemophilia. However, many options are invasive and require the expertise of a Haemophilia Treatment Centre.
Prognosis. Patients with early stage HCC (BCLC stage 0 and A) who can be treated with curative intent (resection, liver transplantation or ablation) can expect a 5-year survival of 50–70% .
Although transplantation cures the underlying cirrhosis, it does not cure HCV. So far, there are no successful strategies to prevent recurrent HCV infection after transplantation. In most patients, HCV recurs in the transplanted liver, with a faster development of fibrosis and cirrhosis than in a native liver . Current practice is to start a course with pegylated-interferon plus ribavirin when significant fibrosis has developed . Once cirrhosis post-transplant develops, mortality is high (26% after 1 year in a study in 39 patients .
There are no surveillance guidelines for recurrent HCC after transplantation. In our centre, we perform twice yearly ultrasound, AFP measurement and chest X-ray. The utility of this follow-up is probably limited, because there are few therapeutic options for recurrent or metastatic HCC.
For patients who are treated palliatively (TACE or sorafenib), 3-year survival is 10–40% . Patients who were candidates for palliative treatment, but did not receive it (the control arm in RCTs) had 1- and 2-year survivals of 18 and 7% respectively . Median survival of patients with advanced HCC, with only symptomatic treatment, is <3 months .