Landgren O, Turesson I, Gridley G, Caporaso NE. Risk of Malignant Disease Among 1525 Adult Male US Veterans With Gaucher Disease. Arch Intern Med. 2007;167(11):1189-1194. doi:10.1001/archinte.167.11.1189
Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007
Some, but not all, reports suggest that patients with Gaucher disease are at increased risk of developing malignancies, particularly hematopoietic tumors. The aim of this study was to assess the pattern of Gaucher disease and subsequent malignancies among male veterans admitted to US Veterans Affairs hospitals.
Among 832 294 African American and 3 668 983 white male veterans with at least 1 hospital admission in US Veterans Affairs hospitals and up to 27 years of follow-up, we identified a total of 1525 patients with Gaucher disease; 11.7% were African Americans. We used Poisson regression methods for cohort data to estimate relative risks (RRs) and 95% confidence intervals (CIs) after adjusting for attained age and calendar year, race, number of hospital visits, and latency.
When patients with Gaucher disease were compared with patients without Gaucher disease, the RR of any cancer was 0.91 (95% CI, 0.76-1.08 [n = 137]). When we stratified our analyses by race, risks were similar for whites (RR, 0.89; 95% CI, 0.74-1.07 [n = 120]) and African Americans (RR, 1.00; 95% CI, 0.61-1.64 [ n = 17]). Patients with Gaucher disease had an elevated risk for non-Hodgkin lymphoma (RR, 2.54; 95% CI, 1.32-4.88 [n = 9]), malignant melanoma (RR, 3.07; 95% CI, 1.28-7.38 [n = 5]), and pancreatic cancer (RR, 2.37; 95% CI, 1.13-4.98 [n = 7]). Among the remaining 19 cases involving defined solid tumors and 7 other hematologic malignancies, we found no statistical association with Gaucher disease.
We found 2- to 3-fold risks of non-Hodgkin lymphoma, malignant melanoma, and pancreatic cancer in patients with Gaucher disease, but no significant association between Gaucher disease and cancer in general or with other specific malignancies such as multiple myeloma.
Gaucher disease is the most common inherited lysosomal storage disease worldwide, a fact that is reflected in an incidence in the general population of 3 to 5 per 100 000 person-years in the United States.1,2 It is characterized by a deficiency of the lysosomal enzyme glucocerebrosidase,3 and its clinical manifestations include progressive hepatosplenomegaly and bone involvement.1 Biochemical abnormalities include anemia, thrombocytopenia, and elevated levels of alkaline phosphatase, angiotensin-converting enzyme, and immunoglobulins. Less frequently, the lymphatic system, lungs, skin, eyes, and kidneys are affected.4,5 The diagnosis is made by genetic testing of the β-glucosidase gene. Different mutations in the β-glucosidase gene determine the remaining glucocerebrosidase activity and, to a large extent, the phenotype.
Based on the presence or absence of neurologic symptoms, Gaucher disease is divided into 3 phenotypes. In type 1, which accounts for more than 90% to 95% of all cases in the United States,6 the symptoms range from mild to severe, may appear at any age, and include easy bruising and/or fatigue due to anemia and a low platelet count, hepatomegaly, splenomegaly, bone pain, bone degeneration and fractures, and lung impairment but no signs of brain involvement. In many cases, type 1 has a favorable prognosis and clinical course. Type 2 has an infantile onset of central nervous system involvement and is always fatal in early childhood. Type 3 has an onset of mild central nervous system involvement in adolescence or early adulthood; the clinical course is usually indolent.7
In 1982, Lee8 published the first study (to our knowledge) suggesting an association between Gaucher disease and subsequent risk of malignant disease. In a retrospective study of 239 patients with type 1 Gaucher disease, he found that 19 (54%) of the 35 deaths that occurred during follow-up were due to malignant disease; 3 of the patients had a diagnosis of multiple myeloma (MM).8 Inspired by these results, several other small hospital- or registry-based epidemiological studies,9- 11 as well as several case reports,12- 20 have examined the association between Gaucher disease and risk for malignancy, particularly hematopoietic malignancy.
Recently, using the International Gaucher Registry,6 Rosenbloom et al10 assessed the cancer incidence among 2742 patients with Gaucher disease. They found 126 cases of cancer, actually suggesting an overall decreased risk of cancer (relative risk [RR], 0.79; 95% confidence interval [CI], 0.67-0.94). However, 10 of the patients had MM, yielding an estimated RR of 5.9 (95% CI, 2.82-10.82); all MM cases were diagnosed after the age of 50 years. Based on their findings, Rosenbloom and colleagues concluded that patients with Gaucher disease, particularly those older than 50 years, should undergo serial serum electrophoresis every 1 to 2 years to rule out MM and/or the precursor condition monoclonal gammopathy of undetermined significance (MGUS), which often precedes MM.21- 23
We were motivated by these previous investigations to conduct a study of cancer risk among patients with Gaucher disease using a database of adult male military veterans admitted to US Veterans Affairs (VA) hospitals. Advantages of the study include its large size (>4 million veterans) and long-term follow-up (up to 27 years).
An exemption from institutional review board review was obtained from the National Institutes of Health Office of Human Subjects Research, Bethesda, Md, because we analyzed existing data without personal identifiers. Informed consent was waived because there was no contact with study subjects. The cohort was identified from discharge records for all inpatient hospitalizations at 142 nationwide US VA hospitals between July 1, 1969, and September 30, 1996. Based on US census data, there were an estimated 30 million US veterans entitled to admission to VA hospitals during the study period. The target population included all African American (n = 832 294) and white (n = 3 668 983) male veterans hospitalized at least once at the age of 18 years or older. Other ethnic/racial groups and females were not included in this study owing to small numbers. Gaucher disease cases were defined as patients from the eligible population with an International Classification of Diseases, Eighth or Ninth Revision (ICD-8 or ICD-9), discharge diagnosis of 272.2 or 272.7, respectively (Table 1).
To estimate the risk of malignancy, all subjects without a prior discharge diagnosis of malignancy were followed up from 1 year after index hospital discharge (Gaucher disease diagnosis for Gaucher disease cases and first discharge for any reason for all others) until the diagnosis of a first malignancy, death, or the end of the observation period (September 30, 1996), whichever came first. Dates of death were ascertained from record linkage to Social Security Administration mortality files. The length of the time interval for the development of a defined cancer was estimated by subtracting the date of discharge for the first hospitalization that listed a discharge diagnosis of Gaucher disease from the date of admission for the first hospitalization that listed a discharge diagnosis of the defined cancer.
Estimates of RRs and 95% CIs were calculated using Poisson regression methods for cohort data24 after attained age and calendar year, race, number of hospital visits, and latency were adjusted for, and the cancer risks among men with Gaucher disease were compared with those among men who were not identified as having Gaucher disease. Variables for all diagnoses were based on dates of hospital discharge; therefore, all variables were time dependent with the exception of race. All P values and CIs were 2-sided, and P<.05 was considered statistically significant. Calculations were performed using AMFIT, a Poisson regression model (Epicure Version 1.4; HiroSoft International Corp, Seattle, Wash).
We identified 1346 white patients (88.3%) and 179 African American patients (11.7%) with Gaucher disease (Table 1). The subjects were followed up for as long as 27 years, with an average of 12.5 and 12.8 years for white and African American patients, respectively. On average, African American men tended to be younger than white men at study entry, and African American men were diagnosed as having Gaucher disease at a slightly younger age than white men. A total of 137 patients (9.0%) with Gaucher disease were hospitalized with a malignant disease. Patients with Gaucher disease tended to visit the hospital more often than patients without a diagnosis of Gaucher disease.
As shown in Table 2, Gaucher disease was not associated with an overall cancer risk (RR, 0.91; 95% CI, 0.76-1.08). When we stratified our analyses by race, risks were comparable for whites (RR, 0.89; 95% CI, 0.74-1.07) and African Americans (RR, 1.00; 95% CI, 0.61-1.64). We observed increased risks for non-Hodgkin lymphoma (NHL) (RR, 2.54; 95% CI, 1.32-4.88 [n = 9]), malignant melanoma (RR, 3.07; 95% CI, 1.28-7.38 [n = 5]), and pancreatic cancer (RR, 2.37; 95% CI, 1.13-4.98 [n = 7]) among patients with Gaucher disease.
Among the remaining 19 cases of defined solid tumors and 7 other hematologic malignancies, we found no statistical association with Gaucher disease (Table 2). It can be seen from the 95% CIs in Table 2 that this study had the statistical power to detect the common cancers with an RR of approximately 2. However, for rare cancers, such as MM, our study could detect only RRs greater than 5. Also, among the patients with Gaucher disease, we found no cases involving a subsequent diagnosis of MGUS. In a separate analysis of the data that we censored (ie, the first year of follow-up after index hospital discharge), we found no cases of cancer among the patients with Gaucher disease.
In this large registry-based cohort study, which included 1525 adult male patients with a diagnosis of Gaucher disease and a very long follow-up, we found no evidence of an increased general risk of malignant disease. When we considered individual solid and hematopoietic tumors, we found that the risks of NHL, malignant melanoma, and pancreatic cancer were elevated 2- to 3-fold among patients with Gaucher disease compared with those without Gaucher disease. The presence of Gaucher disease was not associated with a significantly elevated risk for any other defined malignancy.
Although there are only sparse data available, a general association between Gaucher disease and subsequent risk of malignant disease has been found in some,11,25 but not all,9 previous studies. In the largest study to date (n = 2742), Rosenbloom et al10 found a borderline decreased risk of cancer among patients with Gaucher disease. Because we had access to a large hospital-based investigation, with data retrieved from a nationwide record linkage of 142 VA hospitals, our results of a null general association between Gaucher disease and cancer add substantially to, and verify the results of, the restricted literature on this topic. Although type 1 Gaucher disease is a panethnic disorder, it is especially prevalent among persons of Ashkenazi Jewish descent (almost 1 per 1000), with a carrier rate of 1 in 17 Ashkenazi Jews.26 Thus, our results are particularly relevant for this ethnic group.
Regarding the association between Gaucher disease and specific types of malignant disease, there have previously been hospital-based cohorts9- 11 and case reports12- 20 with a major focus on hematopoietic tumors. Based on 9 patients with Gaucher disease, we found a 2.5-fold significantly elevated risk for NHL. The occurrence of lymphoma following Gaucher disease has been reported in prior studies,9- 11,27- 31 although, to our knowledge, elevated risk estimates for NHL have not previously been reported among patients with Gaucher disease. Our findings add new information and may have clinical implications for the treatment and follow-up of patients with Gaucher disease. If replicated, the observed elevated risk of NHL following Gaucher disease suggests that more aggressive workup should be undertaken for patients with Gaucher disease who have a new onset of symptoms, such as lymphadenopathy, fever, weight loss, and/or night sweats, that might be explained by a lymphoma diagnosis. It is particularly important to diagnose NHL at an early stage to improve the probability of cure, to avoid unnecessary trauma for the patient, and to minimize risks of secondary complications due to NHL treatment.
In contrast to previous hospital-based studies,9- 11 we failed to observe an elevated risk of MM, which is the malignancy that by far has drawn the most attention in the literature. As shown in Table 3, prior studies on Gaucher disease have typically included patient series using data from a single hospital or a limited number of hospitals. Although we had no information on clinical follow-up routines for patients with Gaucher disease in VA hospitals, we believe that skeletal and visceral assessment routines33 are probably more rigorous in specialist units that have a particular clinical and research interest in Gaucher disease than in VA hospitals with a broader medical approach. Follow-up procedures imply increased surveillance, a potential bias that could be the explanation for the previously reported association between Gaucher disease and MM. In their largest study to date, Rosenbloom et al10 found 10 MM cases, corresponding to an increased RR of 5.9. All cancers were registered whether they occurred before or after entry in the International Gaucher Registry.6 Life table analyses could not be used in that study because the date of cancer diagnosis was not known in most of the cases, and comparisons were made by calculation of cumulative lifetime probability of the cancer having developed by the attained age. In our study, although we found only 2 MM cases, a statistical excess of MM following Gaucher disease cannot be totally ruled out because both conditions are relatively rare. Although, to our knowledge, this is the second largest study of Gaucher disease that has been reported, the rarity of the 2 conditions imposes severe limitations on statistical power. Independently, even if the previously reported elevated risk of MM following Gaucher disease is true, it might not have strong clinical implications; therefore, population-based screening cannot be justified. Today, up-front therapy for patients with incident MM is based on chemotherapy, and, usually, treatment is initiated at diagnosis for individuals with stage 2 disease or higher.34 Currently, to our knowledge, there is no available preventive therapy for early-stage or smoldering MM,35 nor is there any treatment available for the precursor condition MGUS. Therefore, the recently proposed idea of Rosenbloom and colleagues10 that patients with Gaucher disease need to be screened with serum electrophoresis every 1 to 2 years to rule out MM and/or MGUS is probably not warranted. Instead, if there are clinical indications to suspect MM in an individual patient, a conventional workup for MM (including serum electrophoresis analysis, bone marrow biopsy, skeletal radiography, and conventional blood tests) appears to be the appropriate route of action.
Immunoglobulin abnormalities have been associated with type 1 Gaucher disease. The most common finding is polyclonal hypergammaglobulinemia; however, an elevated occurrence of the precursor condition MGUS has also been reported.32,36- 39 In our study, it is not surprising that we observed no MGUS cases among 1525 identified patients with Gaucher disease. Because MGUS is generally asymptomatic and information on subsequent disease was derived from hospital discharge listings (ie, no detailed clinical or laboratory data were available), it is reasonable to conclude that the observed null association is the result of underascertainment of MGUS.40 Genetic studies have found that almost 50% of MGUS cases have primary translocations in the clonal plasma cells involving the immunoglobulin heavy-chain (IgH) locus on chromosome 14q32,41,42 a locus thought to be important for the initiation and support of clonal proliferation.43- 45 It has been suggested that hypermutability and proliferation of key cells associated with infections could be the precipitating events for these translocations.41 In light of the previously reported association between Gaucher disease and MGUS,32,36- 38 it may be hypothesized that abnormal macrophages in Gaucher disease could be especially vulnerable to such precipitating events. Results from a mouse model of Gaucher disease suggest that the inflammatory response in Gaucher disease is independent of glucocerebroside deposition,46 but, to our knowledge, the same relationship has not been explored in humans. Progressive glucocerebroside storage in Gaucher disease could plausibly act via chronic stimulatory mechanisms of the immune system, causing lymphoproliferation.32 Gaucher cells themselves might act (1) by inducing local inflammatory response in surrounding inflammatory cells,47 (2) by reducing T-cell functionality,48 or (3) by a combination of both processes. Interestingly, there are indications that decreased numbers and dysfunction of regulatory T cells are of pathogenetic importance in MGUS and MM.49 These and other potential underlying mechanisms remain to be explored in future preclinical studies. It would be of scientific interest to screen a cohort of patients with Gaucher disease to determine the prevalence of MGUS. If the rates for MGUS are found to be highly elevated compared with the currently reported 3.2% age-adjusted prevalence rate among adults older than 50 years,23 further studies may be required to explore underlying mechanisms.
Based on 5 cases, we found a 3.1-fold increased risk of malignant melanoma among patients with Gaucher disease, and based on 7 cases, we found a 2.4-fold increased risk of pancreatic cancer among patients with Gaucher disease. Interestingly, a host-related immunogenetic profile has recently been proposed to be of importance for susceptibility and tumorigenesis in malignant melanoma.50 Also, there are epidemiological indications that host-related immune response might play a role in the risk for pancreatic cancer.51 However, previous studies on Gaucher disease have not reported an excess risk for these cancers,9- 11 so chance is an alternate explanation.
The prevalence of Gaucher disease in our study is higher than previously reported.1,2 This finding is not surprising as the present study is hospital based (ie, it involves male US veterans with at least 1 hospital admission) rather than being based on the general population. For example, if we were to have estimated the prevalence of diabetes mellitus in this hospitalized population, it would also have been much higher than population-based rates. Similarly, the prevalence of Gaucher disease that we report is at the time of hospitalization (approximately at the age of 50 years) and not at the age of military service. With regard to Gaucher disease among blacks, there are only limited data in the literature.52- 56 As pointed out by the authors of a recent of investigation on African Americans with Gaucher disease (based on only 7 cases),52 the current understanding of Gaucher disease in African Americans has been derived from small numbers and is therefore inherently affected by ascertainment bias, with milder patients escaping diagnosis. To our knowledge, no population-based survey has been conducted to determine the age-adjusted prevalence of Gaucher disease among African Americans. In our study, based on 832 294 African American and 3 668 983 white male veterans with at least 1 hospital admission, we found 179 (1/1650) and 1346 (1/2726) individuals with a diagnosis of Gaucher disease, respectively. As discussed above, the general risk of cancer among Gaucher disease cases was very similar for the 2 races (RRs around 1.0).
The strength of the current study includes its larger size in a patient population with relatively stable and standardized access to medical care that is provided to US veterans independent of socioeconomic status. Also, study subjects were older than 18 years at enrollment and they were followed up for intervals as long as 27 years. To evaluate incident rather than prevalent cancer, we estimated the risk of malignancy using all subjects without a prior discharge diagnosis of malignancy and censored the first year of follow-up. Furthermore, in a subanalysis, we assessed the occurrence of malignancy in that censored year (ie, the first year after the index hospitalization), and we found no tumors among patients with Gaucher disease. The limitations of our study include the lack of information about demographic, clinical, treatment, laboratory, or biomarker information for individual patients in the database as well as the lack of specific information on the Gaucher disease phenotype. However, based on the literature, 90% to 95% of the patients with Gaucher disease have type 1 disease. The identification of the cohort from hospital discharge diagnoses, rather than from screening, is likely to have led to underascertainment of Gaucher disease. Individuals with early-onset severe Gaucher disease are likely not included in this study because they were almost certainly disqualified from military service in their younger years. The use of a retrospective cohort might potentially have resulted in some underascertainment of cancer cases. Also, the restriction to male patients might limit the generalizability of our results. The lack of independent validation of cancer diagnosis is another limitation of our study; however, the ascertainment of cancers was similar among patients with and without Gaucher disease, so the chance of significant bias is reduced. In addition, we previously found more than 98% validity for cancer diagnoses in VA discharge records.57,58 Because MM is rare before the age of 40 years and the median age at diagnosis is approximately 70 years, the possibility that prolonged follow-up might have disclosed an increased incidence of MM in our study cannot be excluded. Another limitation involves the fact that available ICD-8 and ICD-9 codes for Gaucher disease are not entirely specific and include other lipid storage disorders, such as Niemann-Pick disease and mucolipidosis II. However, these non-Gaucher lipid storage disorders usually manifest at young ages. Therefore, individuals affected with these conditions would likely not serve in the military. Furthermore, Gaucher disease is the most common inherited storage disease,59 and other lipid storage disorders are not known to be associated with cancer. Although we had the statistical power to detect common cancers with RRs of approximately 2, for rare cancers (such as MM) we were unable to detect RRs of less than 5. Finally, because we studied multiple cancer outcomes and risk estimates were based on small numbers, positive results should be interpreted with caution.
In conclusion, among more than 1500 adult male US veterans with a diagnosis of Gaucher disease, we found no evidence of a generally elevated risk of subsequent hematopoietic or solid tumors. We observed 2- to 3-fold risks of NHL, malignant melanoma, and pancreatic cancer subsequent to Gaucher disease, but we were unable to confirm the previously reported association between Gaucher disease and MM.
Correspondence: Ola Landgren, MD, PhD, Genetic Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, 6120 Executive Blvd, Bldg EPS, Room 7110, National Institutes of Health, Department of Health and Human Services, Bethesda, MD 20892 (firstname.lastname@example.org).
Accepted for Publication: February 14, 2007.
Author Contributions:Study concept and design: Landgren, Turesson, Gridley, and Caporaso. Acquisition of data: Landgren and Gridley. Analysis and interpretation of data: Landgren, Turesson, Gridley, and Caporaso. Drafting of the manuscript: Landgren. Critical revision of the manuscript for important intellectual content: Turesson, Gridley, and Caporaso. Statistical analysis: Landgren and Gridley. Obtained funding: Landgren. Administrative, technical, and material support: Landgren and Gridley. Study supervision: Landgren.
Financial Disclosure: None reported.
Funding/Support: This research was supported by the Intramural Research Program of the National Cancer Institute, National Institutes of Health.
Acknowledgment: We thank Eric A. Engels, MD, for critical review of the manuscript; the Medical Administration Service of the Veterans Health Services and Research Administration for providing the data on which this study is based; and Heather Morris, Information Management Services, Silver Spring, Md, for assistance with data preparation.