Gaucher disease and cancer incidence: a study from the Gaucher Registry

Gaucher disease, caused by an inherited deficiency of glucocerebrosidase, is the most prevalent lysosomal storage disease worldwide. Because of an excessive accumulation of glucocerebroside in macrophages, progressive hepatosplenomegaly, anemia, thrombocytopenia, and bone involvement commonly occur.¹  It has been reported that the incidence of malignant disorders is increased in patients with Gaucher disease.^2,3  If this conclusion is accurate, the ramifications for patients and their physicians is quite significant and may imply a causal relationship, possibly due to some chronic stimulation of the immune system.⁴ 

The Gaucher Registry is a database that, as of September 30, 2003, has information on nearly 3000 patients with Gaucher disease of all types.⁵  More than 90% have type I disease and form the basis of this retrospective review of the incidence of cancer in patients with Gaucher disease. The majority of these patients are receiving enzyme replacement therapy (ERT) with imiglucerase for injection (Cerezyme; Genzyme, Cambridge, MA),⁶  but untreated patients are also included.

The Gaucher Registry tracks outcomes of routine clinical practice for patients with GD from more than 40 countries, regardless of severity and treatment status. All patients (or for those younger than 18 years of age, their parents or guardians) gave consent to participate in the Gaucher Registry at the time of enrollment. All physicians following patients with Gaucher disease are encouraged to submit clinical data on their patients to the registry. The study protocol was approved by institutional review boards at all sites. Data submissions are checked and queried regarding completeness and apparent errors but are not validated by review of either medical or pathologic records.

The Gaucher Registry data were reviewed for cancer occurrence by analyzing both adverse events and concurrent illnesses. The standard method of analysis for studies in which cancer incidence is followed in a defined population over time is a life table approach in which patients are followed from the time they enter the population (in this case, enter the registry) until their date of cancer diagnosis or death. Expected cases are derived from age-, sex-, and race-specific person-years at risk and incidence rates. However, in the Gaucher Registry, although age at first and last follow-up into the registry is known, the age at cancer diagnosis is unknown for most cases; therefore, the life table method cannot be used.

In this analysis, the numbers of selected cancers expected in the registry population was estimated by calculating cumulative lifetime probability of having developed the cancers by attained age. Attained age was defined in this analysis as the date of last registry follow-up or the date of diagnosis of cancer when known. The registry population was assigned to mutually exclusive 5-year age- and sex-specific subgroups. The probability of having developed cancer was calculated using DevCAN (Probability of Developing or Dying from Cancer), publicly available SEER (Surveillance, Epidemiology, and End Results) software (National Cancer Institute, Bethesda, MD), which allowed the calculation of the probability of developing cancers by specific attained ages. The number of cancers expected in this cohort were summed over all age groups and compared with the numbers observed in the patients in the registry. The standardized mortality ratio (SMR) relative risk (RR) was calculated by dividing the observed mortality by the expected mortality. For observed counts equal to 20, 95% confidence intervals (CIs) for the SMR were computed by using exact methods for the Poisson distribution. For observed counts greater than 20, an asymptotic method based on the log transformations was used.⁷ 

At the time of this analysis, there were 2742 patients in the registry. As shown in Table 1, the age distribution in the Gaucher Registry population in fact closely approximates the world standard population.⁸  Since the population age distributions are similar, and the Gaucher populations are stratified by age and sex to compare cancer rates, there should be no bias introduced as a result of difference in the population structure itself. Approximately 92% of the patients were type 1, 1% was type 2, 5% were type 3, and for 2% no type was identified by the participating physician. More than 80% of the patients had received ERT. The median age at last follow-up was in the third decade. There were 126 patients for whom at least one cancer event was reported (Table 2). One patient had 2 events of cancer reported. About 65% of all reported cancers occurred in patients older than 50 years of age. Cancer events were disproportionately greater in patients from the United States and in patients of Ashkenazi Jewish ethnicity. This observation is attributable to overrepresentation of older patients in the United States and Ashkenazi Jewish subpopulations. Cancer incidence was similar in patients ever treated with ERT (104 of 2241, 4.6%) and those never treated with ERT (22 of 495, 4.4%).

As shown in Table 3, the overall lifetime probability of developing cancer by attained age in this population was not greater than expected (relative risk, 0.79). This was also true for individual cancers, including breast cancer, prostate cancer, colorectal cancer, lung cancer, and all hematologic malignancies in aggregate, except multiple myeloma. The observed and expected incidence of myeloma was 10 and 1.7, respectively, with a relative risk of 5.9 (95% CI: 2.8, 10.8). With one exception, all patients with Gaucher disease with multiple myeloma were older than age 60, and the increase in myeloma incidence was observed in both US and European subpopulations (Table 2). Myeloma occurred in patients never treated with ERT (3 of 495, 0.6%) and in ERT-treated patients (7 of 2241, 0.3%).

In 1982, from a retrospectively collected population of 239 patients with type 1 Gaucher disease, Lee²  reported a review of 35 deceased patients with Gaucher disease, 19 of whom (54%) had evident malignancies. He postulated that, in patients with milder variants of Gaucher disease in whom life expectancy in the pre-ERT era was considerably greater than in severely affected patients, the accumulated glucocerebroside somehow contributed in the causation of cancer.²  Although detailed demographics for the 239 patients were not reported, assuming an age distribution similar to our study population (mean age, 33 years) resulting in 7887 risk-years, and an estimated cancer mortality for 1980 of 2 to 300 per 100 000,⁹  the predicted number of cancer deaths is 16 to 24. Thus, with the exception of multiple myeloma (3 of 19 deaths), the report of Lee²  does not support a hypothesis that patients with type 1 Gaucher disease (GD) have an increased lifetime cancer risk.

Shiran et al³  reported a 20.8% incidence of cancer among 48 patients with GD from their hematology clinic compared with a 6.8% incidence among healthy patients throughout their institution. There was a 14.7-fold increased risk of hematologic malignancies, including 2 patients with multiple myeloma. There are also numerous case reports of various malignancies reported in association with Gaucher disease.^10-45  Notable is a striking incidence of multiple myeloma in these reports.^{10,15,16,19,20,30,34,36,38,42} 

In the large population of patients with Gaucher disease presented in our report, the overall incidence of malignancy is 3.6%, about half the incidence reported by Shiran et al.³  For hematologic malignancies other than multiple myeloma, the collective incidence is 0.9%, which is far less than the 10.4% incidence in the study of Shiran et al.³  The relative risk for developing hematologic malignancies is slightly increased, but this result is not statistically significant, and it also assumes that the 7 cases of hematologic malignancy reported as “unspecified” are indeed properly classified. Additionally, the one case of Waldenstrom macroglobulinemia might appropriately be better grouped with patients with myeloma.

For solid tumors in our patient population, the cumulative probability of developing cancer by attained age is actually less than expected. Because our study methodology relies on the SEER database as applied to a population with a US component of only 48.4%, our calculations of expected cancer incidence may be inaccurate. However, the majority of patients in the registry come from economically developed countries, where cancer patterns are generally similar to those in the United States.⁴⁶  Consequently, the application of US data worldwide may partly explain why the observed incidence of cancer in our patients was less than expected. However, even using country-specific cancer incidences, save for myeloma, the relative cancer risk is not increased.

The statistical methodology of this report assumes that the registry is likely to have identified all cancer cases whether they occurred before or after entry into the registry. This method possibly may underestimate the relative risk of cancer in these patients, because complete identification by the registry of all cancers is improbable. In addition, the registry population is young, with few patients having reached the ages of 60 and older, when cancer rates are highest in the general population. The study might have failed to detect some cancers occurring at young ages because any such patients from earlier cohorts would not have survived long enough to be enrolled when the Gaucher Registry was initiated in 1991. However, this is unlikely to be an important bias since the literature review reveals only very rare reports of malignancy in children or young adults with GD.^24,27 

In contradistinction to all other malignancies, the incidence of multiple myeloma in patients with type 1 Gaucher disease is almost 6 times higher than expected. The association of Gaucher disease with polyclonal and monoclonal gammopathies has previously been reported.^{16,19,34,36,38,42,47}  It has been theorized that persistent and progressive glucocerebroside storage may cause chronic stimulation of the immune system and consequent lymphoproliferation.⁴  However, the inflammatory response in Gaucher disease may in fact be independent of glucocerebroside deposition, as systemic inflammation is reported in glucocerebrosidase-deficient mice who demonstrate minimal storage of glucosyl ceramide.⁴⁸  Ferritin release from Gaucher cells has been implicated in reducing T-cell function and immunoglobulin M (IgM) release from B cells.⁴⁹  Interleukin 6 (IL-6) is a significant putative signal transduction cytokine in multiple myeloma,⁵⁰  and IL-6 levels are increased in some patients with Gaucher disease.⁵¹  Although recent histochemical analysis of GD splenic tissue suggests that classic Gaucher cells most closely resemble anti-inflammatory alternatively activated macrophages and do not produce typical proinflammatory molecules, there is nonetheless variable expression of IL-6 by Gaucher cells which also appear to induce a vigorous inflammatory response in surrounding classically activated red pulp macrophages.⁵² 

In conclusion, on the basis of this large database, the incidence of solid tumors and hematologic malignancies does not appear to be significantly increased in patients with Gaucher disease followed through early to middle age, except for multiple myeloma. This information should serve to alleviate inordinate fear of cancer in patients with Gaucher disease and their families that is attributable to frequent citation of the earlier literature on patient-accessible internet sites. Regarding the increased incidence of multiple myeloma, treating physicians should be cognizant of this finding and monitor patients with Gaucher disease accordingly. We recommend a baseline serum immunoelectrophoresis with initial diagnosis, at time of first evaluation, or prior to initiation of ERT, if such treatment is indicated. Abnormal results should be investigated following standard hematologic protocols. For patients with normal baseline results, we recommend serial follow-up at 1- to 2-year intervals, particularly in patients older than 50 years of age. Polyclonal gammopathies, which appear unrelated to emergence of monoclonal gammopathy and myeloma, are commonly identified in patients with type 1 Gaucher disease and often decrease in conjunction with the overall response to ERT.⁵³  Because myeloma occurs predominantly in elderly patients with Gaucher disease, long-term follow-up of patients in whom treatment is started in childhood or young adulthood is necessary to determine whether the increased incidence of monoclonal gammopathy and multiple myeloma in Gaucher disease will decrease with enzyme replacement therapy.

We thank the patients with Gaucher disease along with their physicians and health care personnel who submit data to the registry, and the US Regional Coordinators of the International Collaborative Gaucher Group (ICGG) for their review of the manuscript.