Cancers in patients with hemophilia: a retrospective study from the Italian Association of Hemophilia Centers

Authors


Massimo Franchini, Immunohematology and Transfusion Center, Hospital of Mantova, Mantova, Italy.
Tel.: +39 376 201234; fax: +39 376 220144.
E-mail: massimo.franchini@aopoma.it

Abstract

Summary. Background: The increased life expectancy of the hemophilia population, primarily as a result of advances in factor replacement therapy, has enabled hemophiliacs to reach an older age. Consequently, age-related diseases, such as cardiovascular disorders and cancers, are being increasingly recognized in such patients. However, only few data are available on such co-morbidities, their management and impact on the primary bleeding disorders. Objectives: With the aim of investigating several still unclear issues regarding cancers in hemophilia patients, we conducted, on behalf the Italian Association of Hemophilia Centers (AICE), a study on cancers among Italian hemophiliacs. Patients: Data pertaining to 122 hemophiliacs with 127 cancers between 1980 and 2010 were retrospectively collected in 21 centers of the AICE which chose to participate. Results:  Sixty-nine percent of cancers were recorded during the decade 2001–2010. Eighty-three percent of patients were infected with hepatitis C virus (HCV) and 22% of them were also co-infected with human immunodeficiency virus (HIV). Forty-three percent of cancers were HCV-related, whereas 9% were HIVrelated. Virus-related cancers were more frequent and non-virus-related cancers less frequent in patients with severe hemophilia than in those with mild/moderate forms (P = 0.0004). The non-virus-related standardized mortality ratio (SMR) was 0.3. Hemorrhagic complications occurred more frequently in patients undergoing chemotherapy (14%) or radiotherapy (19%). Conclusions: The results of the present study confirm that cancers have become a new challenge for physicians working in hemophilia centers and underline the need for prospective trials to better assess the epidemiology and to optimize the management of hemophiliacs with cancer.

Introduction

Since the early 1970s there have been dramatic improvements in the availability and quality of treatment for people with hemophilia [1]. As a result of this progress, barring the consequences of the human immunodeficiency virus (HIV) and hepatitis C virus (HCV) epidemics in the 1980s, the life span of hemophiliacs has gradually become similar to that of males in the general population, at least in more developed countries [2]. However, with aging, hemophiliacs develop medical and surgical diseases (e.g. cardiovascular diseases, prostatic hypertrophy, cancers, renal disease) not previously seen in this group, which represent a new challenge for physicians working in hemophilia centers [3–5]. As a consequence, in the last few years, the interest of investigators has focused on the management of age-related clinical conditions in patients with inherited bleeding disorders.

The issue of malignancies in patients with hemophilia is particularly intriguing because of a presumed protective effect of this hereditary bleeding disorder against cancer spread and dissemination [6]. However, apart from anecdotal case reports, there is very little information available in literature on the epidemiology, clinical presentation and treatment options of this combination of conditions [6–9].

Thus, in order to elucidate various unclear issues regarding cancers in hemophiliacs, we conducted, on behalf of the Italian Association of Hemophilia Centers (AICE), a retrospective study among Italian hemophilia centers.

Patients and methods

Study design

This retrospective study was approved by the Institutional Review Boards of AICE and proposed at the beginning of 2009 to all member centers. The primary objective of the present study was to describe the cases of cancer (types of neoplasm, the patients’ outcome and the types of treatment used) in the Italian population of hemophiliacs, including carriers. A secondary aim was to investigate the influence, if any, that hemophilia had on the management and outcome of the malignancies. Furthermore, an epidemiological analysis of the incidence of cancers and survival of affected patients was conducted, comparing the results with those of a matched general population.

Cancer cases of any age – living or dead, with mild, moderate or severe hemophilia A (HA) or hemophilia B (HB) and carriers of HA or HB who had any type of malignancy, including HCV and HIV-related malignancies – were included. No restriction of the period of inclusion was required.

Data were collected on cancer cases through an ad hoc questionnaire survey. The case report form had three sections. The first part collected demographic data (year of birth and age at death, and cause of death), information on the type and severity of hemophilia (severe HA or HB, coagulation factor [F]VIII or IX deficiency < 0.01 IU mL−1; moderate HA or HB, factor deficiency between 0.01–0.05 IU mL−1; mild HA or HB, factor deficiency between 0.06 and 0.40 IU mL−1), type of mutation (if known), presence of inhibitors, presence of blood-borne viral infections and their complications (i.e. chronic hepatitis B, chronic hepatitis C, liver cirrhosis, hepatocellular carcinoma [HCC], HIV and acquired immunodeficiency syndrome [AIDS], response to antiviral treatment for HCV or HIV), concomitant diseases and therapeutic data (type of concentrate, treatment regimen, and mean annual consumption). The second section of the questionnaire dealt with the type and stage of neoplasia, the date of diagnosis, cancer-related risk factors, and date and type of metastases. According to Cáceres et al. [10], cancers that are HIV-associated or occur at an increased frequency in HIV patients include lymphomas, Kaposi’s sarcoma, lung cancer and seminoma. Among cancers in patients with HCV infection, HCC and non-Hodgkin’s lymphomas were considered to be virus associated, according to recent literature data [11]. The third part of the questionnaire collected data on the management of the neoplasias: invasive/surgical procedures, related replacement therapy (type, dosage and hemostatic efficacy) and hemorrhagic complications; type of chemotherapy/radiotherapy regimens, type of venous access, hemorrhagic complications, the occurrence of thrombocytopenia (platelet count < 50 × 109 L−1), response to treatment (date and relapses) and the period of follow-up. Finally, it was asked whether the patients’ hemophilia had influenced the choice of therapeutic protocol, contraindicated participation in experimental protocols, prevented completion of therapy or necessitated a reduction in the doses of chemotherapy.

Statistical analysis

The relationship between severity of hemophilia (mild, moderate and severe) and the origin of the neoplasia (viral-related or non-viral-related) was investigated. The null hypothesis of a uniform distribution of the frequencies within this contingency table was evaluated by means of a Pearson’s chi-square test (four degrees of freedom [d.f.]).

A time-to-event analysis was also performed. The dependent variable was death. The event rate in subjects with all combinations of HCV and HIV infections was investigated by Kaplan–Meier/log-rank analysis.

A standardized mortality ratio (SMR) was calculated to estimate the rate of cancer-related death of patients with hemophilia relative to that of the general male population adjusted for age and calendar period. The SMR is the number of observed deaths in the study population divided by the number of deaths expected if the mortality rate in the cohort, with its specific age-distribution, was the same as that in the general population. The number of expected deaths was calculated multiplying the specific age- and calendar-period number of persons with hemophilia-years of follow-up of the cohort by the corresponding cancer-specific age- and calendar-period mortality rates in the general Italian male population, and summing the products over the age groups. Ninety-five percent confidence intervals (95% CI) for the SMR were calculated, based on the Poisson distribution. Cancer mortality rates for the general Italian male population used to standardize mortality ratio for the study population were those for the year 1995. The reference dataset of Italian hemophilia patients was built pooling the number of patients provided by each of the 21 hemophilia centers who took part in the present study. All HA and HB patients registered during the study period were included.

Results

Twenty-four of the 50 Italian hemophilia centers that are members of AICE agreed to participate and 21 (see list in Appendix) provided information on 122 patients with 127 cancers during the period 1980–2010. This cohort corresponded to 5% (122/2496) of the hemophilia population followed by the enrolled hemophilia centers, which represented 59.5% of the whole Italian population of persons with hemophilia. Table 1 summarizes the characteristics (number of patients alive and dead, median age at diagnosis of cancer and death, HCV and HIV status) of the patients with cancer enrolled according to the severity of their hemophilia. Fifty-four patients (43 HA, 11 HB) had severe hemophilia, 21 (13 HA, 8 HB) had moderate disease, 45 (39 HA, 6 HB) had mild hemophilia and two were HA carriers. Eighty-three percent (101/122) of patients were infected by HCV and 22% (22/101) of those with HCV were co-infected by HIV. As expected, the rate of virus-infected patients was proportional to the severity of the hemophilia.

Table 1.   Characteristics of the patients observed during the study period (1980–2010)
DiagnosisPatients
n
Alive/dead
n
Median age at cancer diagnosis*Median age at death*HCV−/HIV−
n (%)
HCV+/HIV−
n (%)
HCV+/HIV+
n (%)
  1. *Years (range).

  2. For three patients, the HCV status was not available. These patients (one each with severe, moderate and mild hemophilia) were all HIV negative.

  3. HCV, hepatitis C virus; HIV, human immunodeficiency virus.

Severe hemophilia5417/3756 (13–74)56 (14–79)3 (6)34 (63)16 (30)
Moderate hemophilia2111/1059.5 (28–84)63.5 (45–84)15 (71)5 (24)
Mild hemophilia4525/2061.5 (33–78)64 (40–80)13 (29)30 (67)1 (2)
Carriers22/−42.5 (30–55)2 (100)
Total12255/6757 (13–84)61 (14–84)18 (15)79 (65)22 (18)

Among the 57 patients with non-virus-related cancers, 23 (40%) (8 with severe hemophilia, 4 with moderate hemophilia and 11 with mild hemophilia) are dead, whereas among the 65 patients with virus-related cancers, 44 (68%) are dead (29 with severe hemophilia, 6 with moderate hemophilia and 9 with mild hemophilia). Overall, the number of cancer-related deaths correlated with the severity of hemophilia, being higher among patients with severe hemophilia than among those with mild-moderate hemophilia (37 vs. 30). Nine patients (7%), all with severe HA, had concomitant FVIII inhibitors and seven of them (78%) are dead.

Table 2 shows the number of cancers according to the severity of hemophilia and virus-infection status. Interestingly, while the rate of virus-related cancers was higher in the more severe forms of hemophilia, the rate of non virus-related cancers was increased in the milder forms of hemophilia (= 0.0004 by Pearson’s chi-square). This finding was present in HA as well as in HB patients. Table 3 presents the type of cancers according to virus status. Solid cancers accounted for the great majority (82%) of the cases and approximately half of them were HCV-related HCC. On the other hand, approximately 40% (9/23) of hematological malignancies were HIV-related lymphomas. No cases of Kaposi sarcoma were observed among HIV-infected patients. Curiously, the four patients with multiple malignancies had mild or moderate HA (one HIV-infected patient with moderate hemophilia and seminoma and gastrointestinal carcinoma; one HIV-negative patient with mild hemophilia and bladder carcinoma, and two different gastrointestinal carcinomas; one HIV-negative patient with mild hemophilia and lung and pharyngeal cancers; and one HIV-negative patient with mild hemophilia and bladder and lung cancers).

Table 2.   Number of cancers according to hemophilia severity and infection status
DiagnosisPatients
n
Cancers
n (%)*
Total
n
Non-virus-related
n (%)
HCV-related
n (%)
HIV-related
n (%)
  1. *Increase of virus-related tumors and decrease of non-virus-related tumors in patients with severe hemophilia compared with milder forms (Pearson’s chi-square test, 4 d.f., χ2 = 24.9042, P-value = 0.0004).

  2. Hepatocellular carcinoma and non-Hodgkin’s lymphoma.

  3. Lymphomas (n = 9), Kaposi’s sarcoma (n = 0), lung cancer (n = 1), seminoma (n = 1).

  4. HCV, hepatitis C virus; HIV, human immunodeficiency virus.

Severe hemophilia
 Total545414 (26)31 (57)9 (17)
 Hemophilia A434311 (26)26 (60)6 (14)
 Hemophilia B11113 (27)5 (45)3 (27)
Moderate hemophilia
 Total212211 (50)9 (41)2 (9)
 Hemophilia A13147 (50)6 (43)1 (7)
 Hemophilia B884 (50)3 (38)1 (12)
Mild hemophilia
 Total454934 (69)15 (31)
 Hemophilia A394330 (70)13 (30)
 Hemophilia B664 (67)2 (33)
Carrier222 (100)
Total12212761 (48)55 (43)11 (9)
Table 3.   Types of cancer according to virus infection status
Type of cancerTotal
n (%)
Non-virus related
n (%)
HCV-related
n (%)
HIV-related
n (%)
  1. HCV, hepatitis C virus; HIV, human immunodeficiency virus.

Solid tumor10452502
 Urogenital tract18 (17)17 (33)1 (50)
 Hepatocellular carcinoma50 (48)50 (100)
 Gastrointestinal tract14 (14)14 (27)
 Respiratory system17 (16)16 (31)1 (50)
 Pancreas3 (3)3 (5)
 Thyroid2 (2)2 (4)
Hematological23959
 Non-Hodgkin lymphoma14 (61)2 (23)5 (100)7 (78)
 Multiple myeloma3 (13)3 (33)
 Chronic lymphocytic leukemia3 (13)3 (33)
 Hodgkin lymphoma3 (13)1 (11)2 (22)

Table 4 shows the treatment characteristics of the cancer patients enrolled in the study. Hemorrhagic complications were more commonly associated with chemotherapy and radiotherapy (14% and 19% of cases, respectively) than with invasive (i.e. biopsies or endoscopies) or surgical procedures (8% and 0.2% of cases, respectively). Anti-hemorrhagic prophylaxis was administered in 87 of 140 (62%) invasive or surgical procedures (factor concentrates in 83 cases, desmopressin in three cases and fresh-frozen plasma in one case). All the eight hemorrhagic complications in these subgroups occurred in patients treated with factor concentrates and were not related to the severity of hemophilia (three moderate, four mild and one carrier). Patients undergoing chemotherapy or radiotherapy did not receive anti-hemorrhagic prophylaxis.

Table 4.   Treatment characteristics of cancer patients enrolled in the study
ProcedurePatients evaluableProcedures/patientsAntihemorrhagic prophylaxisHemorrhagic complications (%)
  1. FC, factor concentrates; DDAVP, desmopressin; FFP, fresh-frozen plasma.

  2. *Patients evaluable.

  3. All hemorrhagic complications occurred in patients treated with factor concentrates.

Invasive procedures9888/7648 (45 FC, 2 DDAVP, 1 FFP)*7 (8)
Surgical procedures9452/4739 (38 FC, 1 DDAVP)*1 (0.2)
Chemotherapy10135/3205 (14)
Radiotherapy9316/1603 (19)

Figure 1 shows the number of cancers according to the different decades of the study. Fourteen cases (12 non virus-related and 2 HIV-related; median age at diagnosis: 56 years, range 14–70 years; incidence: 0.5 cases × 1000 person-years) were recorded during the period 1980–1990, 25 cases (13 non-virus-related, 6 HCC and 6 HIV-related; median age at diagnosis: 50 years, range 13–72 years; incidence: 1.0 cases × 1000 person-years) during the decade 1991–2000 and 88 cases (36 non-virus-related, 49 HCV-related [44 HCC and 5 NHL] and 3 HIV-related; median age at diagnosis: 60 years, range 16–84 years; incidence: 3.2 cases × 1000 person-years) during the decade 2001–2010.

Figure 1.

 Number of cancers (non-virus-related and virus-related) in the different decades of the survey.

Figure 2 shows the Kaplan–Meier survival estimates of cancer patients enrolled according to HIV/HCV infection status. The time-to-event analysis of all combinations of HCV and HIV infections shows that the incidence of deaths was significantly higher in the HIV–HCV co-infected population than in the HIV–HCV uninfected population (= 0.0323 by the log-rank test). With regards to the observed-to-expected mortality rate, the overall cancer-related SMR was 1.3 (95% CI, 0.6–2.1), whereas the SMR owing to cancers not related to HCV or HIV infections was 0.3 (95% CI, 0.1–0.7).

Figure 2.

 Kaplan–Meier survival estimates of cancer patients enrolled according to hepatitis C virus/human immunodeficiency virus (HIV/HCV) infection status. Log-rank test (outcome: death) χ2(2) = 6.87; P-value = 0.0323.

Twelve of the 76 patients (16%) developed thrombocytopenia (platelet count < 50 × 109 L−1), complicated in two cases by a hemorrhage during chemotherapy. Finally, the physicians interviewed reported that chemotherapy and/or radiotherapy was not performed in two patients with severe hemophilia (2%) because of the underlying bleeding disorder.

Discussion

Thanks to the improvement of comprehensive health care provided by specialized hemophilia centers, the life expectancy and mortality rates in hemophilia patients are approaching values seen in the general population [3]. As a result, the hemophilia population is growing older and faces increasing numbers of age-related co-morbidities, such as cardiovascular diseases and cancers [12].

With regard to the latter, experimental data indicate that severe hemophilia reduces cancer metastasis in mice, perhaps because of less thrombin formation [13,14]. However, with the exception of the increased rates of HCC in HCV-infected patients and non-Hodgkin lymphomas in HIV-infected patients, it is uncertain whether the incidence of cancers in hemophiliacs differs from that observed in age-matched persons without hemophilia [15]. Unfortunately, only a few clinical studies have specifically addressed this issue after excluding malignancies related to viral infections [3].

For instance, on behalf the Association of Hemophilia Clinic Directors of Canada, Walker and Julian analyzed the causes of death among 2450 hemophiliacs during the years 1980–1995 and found an unexpected lower (0.3) SMR for cancer in HIV-negative patients [16]. More recently, on behalf the United Kingdom Haemophilia Centre Doctor’s Organization, Darby and colleagues analyzed mortality rates and causes of death in 6018 HIV-negative hemophiliacs during the period 1977–1998 [17]. The authors noted that while mortality from liver cancer and Hodgkin’s disease was increased compared with mortality in the general population (SMR of 13.51 and 4.95, respectively), for other cancers there was no evidence of increased mortality (SMR 0.90). On behalf of the AICE, we have recently presented the results of mortality and causes of death among Italian hemophiliacs during the period 1990–2007 [2]. The SMR for cancers other than HCC increased during the study period (0.34 during the years 1990–1999 vs. 0.67 during the years 2000–2007). In contrast with these findings, other studies have found a similar or even higher rate of non-HIV/HCV-related cancers in hemophiliacs than in non-hemophiliacs [18,19]. In a recent review on epidemiological studies regarding cancer in hemophilia, Miesbach and Seifried [6] found that non-HIV/HCV-related cancers accounted for 8–16% of all deaths in hemophiliacs with a SMR lower than one in all studies, indicating a lower cancer mortality in the virally uninfected hemophilia population than in the matched general male population. The low SMR due to non-HCC/HIV-related cancers found in the present study (0.3) is in keeping with the majority of literature data and confirms the results from our previous study [2]. However, we are aware that the great majority of published SMRs, including ours, are derived from retrospective studies and need to be validated by prospective trials on large populations of hemophilic patients, such as the ongoing European Association for Haemophilia and Allied Disorders (EAHAD) study [20].

Other interesting information emerged from an analysis of the present study. First, the rate of virus-related cancers was higher in the more severe forms of hemophilia whereas the rate of non-virus-related cancers unexpectedly decreased with the increase of severity of hemophilia. While the former finding is rather obvious and is clearly related to the higher prevalence of HIV/HCV infection in severe hemophiliacs, the latter is very intriguing and could somehow clinically confirm the in vitro studies documenting that FVIII enhances tumor growth and dissemination [13]. In addition, as we observed the same behavior in both HA and HB patients with non-virus-related cancers, we could hypothesize that this effect is indirectly mediated by thrombin rather than by the coagulation factor itself. The increased prevalence of cancers in patients with mild hemophilia is probably not an artifact simply as a consequence of the increased life-expectancy of hemophiliacs with milder forms of disease, as documented by the Kaplan–Meier survival estimates illustrated in Fig. 2, because the median age at death in severe hemophiliacs (63 years, range 50–79 years) after excluding those who died of HIV-related cancers was similar to that of patients with mild (64 years, range 40–80 years) and moderate (63.5 years, range 45–84 years) forms.

Another interesting observation arises from the analysis of Figure 1 showing that approximately 70% of the cancer cases (virus-related and non-virus-related) were recorded during the decade 2001–2010. There are various possible explanations for this finding, with the major ones being: (i) the increasing life-expectancy, approaching that of general male population, of the hemophilia population observed during the last decade as recently documented by our survey [2]; (ii) the very long natural history of chronic hepatitis C, which reaches its final stage (i.e. HCC) after two to three decades which correspond, considering the time of the first exposure to non-virus-inactivated plasma-derived factor concentrates (i.e. 1970s and 1980s), to the last decade [21]; and (iii) the significant improvement in the comprehensive care provided by specialized hemophilia centers which results in a better clinical and laboratory monitoring of patients [22].

An analysis of Table 4 allows us to make some considerations on the management of cancer patients. Curiously, patients undergoing invasive procedures experienced hemorrhagic complications more frequently than those undergoing surgery, but the low numbers of evaluable patients do not allow us to draw reliable conclusions. However, a possible explanation of this finding could be that most invasive procedures were performed in an outpatient regimen without regular administration of replacement therapy after the procedures by some patients at home. In addition, the fact that patients usually did not receive anti-hemorrhagic prophylaxis during chemotherapy or radiotherapy but experienced the highest rate of bleeding is of note and underlines the need to consider providing continuous prophylactic replacement therapy in selected cases, especially when these treatments are used in patients with thrombocytopenia [15]. Finally, although the physicians declared that in all but two cases the underlying hemophilia did not preclude access to chemotherapy and/or radiotherapy regimens, the finding that approximately 80% of hemophiliacs with cancer and a concomitant inhibitor are dead raises the suspicion that such patients did not receive optimal anticancer therapy. In particular, concerns about bleeding complications may have delayed the implementation of diagnostic and/or therapeutic procedures and led to under-treatment.

In conclusion, on the basis of the results of the present study, the largest conducted so far on this topic, it is possible to foresee that, as a result of the improved life expectancy of hemophilia patients, an increasing number of cancers will be encountered in clinical practice and will represent a new challenge for physicians working in hemophilia centers. Prospective trials are needed to better characterize the epidemiology of cancers in hemophiliacs and to optimize the therapeutic approach to malignancies in these patients.

Addendum

A. Tagliaferri and M. Franchini designed the study; C. Di Perna, C. Santoro, Piercarla Schinco, R. Santoro, G. Rossetti, A. Coppola and M. Morfini collected the data; A. Tagliaferri, C. Di Perna and M. Franchini analysed the data and wrote the manuscript; M. Franchini was responsible for the statistical analysis; all the authors took part in the revision of the manuscript.

Acknowledgement

The authors thank Professor Pier Mannuccio Mannucci for the critical revision of the manuscript and Professor Carlo Mengoli for the statistical assistance.

Disclosure of Conflict of Interest

The authors state that they have no conflict of interest.

Appendix

A list of co-authors of the study with Hemophilia Center site between parentheses: M. G. Mazzucconi (Roma), G. Castaman (Vicenza), F. A. Scaraggi (Bari), S. Macchi (Ravenna), S. Pasca (Udine), L. Valdrè (Bologna), E. Zanon (Padaova), P. Radossi (Castelfranco Veneto), M. L. Serino (Ferrara), I. Cantori (Macerata), M. Napolitano, M. Lapecorella (L’Aquila), A. C. Giuffrida (Verona), A. Dragani (Pescara), G. Delios (Ivrea), M. C. Arbasi (Piacenza).

Ancillary