Familial Waldenstrom’s macroglobulinemia and relation to immune defects, autoimmune diseases, and haematological malignancies – A population-based study from northern Sweden

Abstract

Background: Waldenstrom’s macroglobulinemia (WM) is a rare lymphoprolipherative disorder with geographic and ethnic disparities in incidence. The cause of WM remains mostly unknown although a role for genetic, immune-related, and environmental factors has been suggested. Most cases of WM are sporadic although familial cases occur.

Aim: This study estimated the incidence of WM in northern Sweden and identified and described patients with familial WM in this area.

Patients and methods: The Swedish and Northern Lymphoma Registry, the Swedish Cancer Registry (1997–2011), and medical records were used to identify patients with WM in two counties (Norrbotten and Västerbotten) in northern Sweden and to calculate the overall age-adjusted incidence (2000–2012). We identified 12 families with a family history of WM, IgM monoclonal gammophathy (MGUS), and/or multiple myeloma (MM).

Results: In Norrbotten and Västerbotten, the age-adjusted incidence of WM/LPL is 1.75 and 1.48 per 100 000 persons per year, respectively (2000–2012), rates that are higher than the overall incidence of WM/LPL in Sweden (1.05 per 100 000 persons per year; 2000–2012). Autoimmune diseases and other haematological malignancies in the medical history (their own or in relatives) were reported in 9/12 and 5/12 families, respectively. A high proportion of abnormal serum protein electrophoresis was found in the relatives; 12/56 (21%) had a MGUS and 13/56 (25%) showed abnormalities in the immunoglobulin levels (i.e. subnormal levels and poly/oligoclonality).

Conclusion: The incidence of WM in Norrbotten and Västerbotten counties was higher than expected. We found a strong correlation between autoimmune/inflammatory diseases, other haematological malignancies, and familial WM and a high frequency of serum immunoglobulin abnormalities in the relatives of the WM patients, findings that strengthen the hypothesis that the aetiology of WM depends on both immune-related and genetic factors.

Waldenstrom’s macroglobulinemia (WM), a rare B-cell lymphoproliferative disorder, accounts for approximately 1–2% of all haematologic malignancies. WM is a disease of the elderly with a median age at diagnosis of around 70 years and with an incidence twice as high in men as in women. Diagnostic criteria for WM include the presence of an immunoglobulin M (IgM) monoclonal gammopathy in blood and infiltration of clonal lymphoplasmacytic cells in bone marrow. Accurate estimates of the incidence and prevalence of WM are hampered by the fact that several different diagnostic criteria are used. Both the international workshop on WM (iwWM) and the 2008 WHO classification require an infiltration in bone marrow of lymphoplasmacytic lymphoma (LPL) and an IgM monoclonal immunoglobulin (MI) in serum of any concentration [1], while the Mayo Clinic criteria require an IgM MI of >30 g/l in serum and/or an infiltration of >10% LPL-cells in bone marrow. In addition, estimates are difficult because the criteria often overlap for IgM monoclonal gammopathy of undetermined significance (IgM MGUS) and asymptomatic or smouldering WM. The aetiology of WM remains mostly unknown, but both genetic and immune-related factors have been suggested to contribute to the development of the disease. Most cases of WM are sporadic, but studies have shown that first degree relatives to WM patients have an increased risk of developing WM and other forms of lymphomas [2,3], suggesting that these relatives share some susceptibility genes. Familial MM is much less described and data of familial occurrence with WM and MM are sparse [3]. Patients with a personal or family history of autoimmune or inflammatory diseases such as rheumatoid arthritis and Sjögren’s syndrome also have a known risk of developing lymphoma, foremost DLBCL [4–7]. Thus, all types of chronic inflammation and/or antigen stimulation leading to B-cell activation might be pathogenetic for some lymphoma subtypes, including WM [8]. Patients with infections, such as hepatitis C and HIV, have a higher risk of developing WM, and antiviral therapy can lead to remission of the clonal disease [9,10].

In 1962, Massari et al. identified the first case of familial WM (two brothers) [11], and since then more than 100 families have been reported with two or more cases of WM, MGUS, or MM. Typically, familial WM affects only a few family members, usually siblings. In addition, compared to sporadic WM, the age at diagnosis is lower and the incidence of autoimmune disease is higher [12–14].

This study estimated the incidence of WM in Norrbotten and Västerbotten counties, both located in northern Sweden, and compared this incidence to the overall incidence of WM and LPL in Sweden. In addition, this study identified WM patients in these northern counties with families with two or more cases of WM, IgM MGUS, and/or MM. A special focus was on medical history of autoimmune diseases and haematological malignancy.

Patients and methods

The incidence, standardised for age, was calculated for LPL (snomed3 = 96713) and WM (snomed = 97613) in Sweden and for the counties Norrbotten and Västerbotten. The number of patients diagnosed with LPL and WM was obtained from INCA (http://www.cancercentrum.se/inca/), which is a national IT-platform for managing different cancer registries, including the Swedish Lymphoma Registry (SLR). The general population size for Sweden and the two counties was obtained from Statistiska Centralbyrån (Statistics Sweden, http://scb.se/). The participants for this study were diagnosed with LPL or WM between 2000 and 2012. The incidence was calculated by dividing the number of patients diagnosed with LPL and WM for the different age groups, and counties with the general population size for the same age group, and county. To standardise age, a weight calculated from a standard Swedish population was multiplied by this ratio. The SLR does not require MI to be registered, so some of the patients registered as LPL might fulfil criteria for WM, if an IgM MI had been registered. Therefore, the incidence is calculated for WM and LPL together. In Västerbotten and Norrbotten, the diagnosis was verified by reviewing medical records and in this cohort all patients except one patient from Västerbotten had an IgM MI fulfilling the diagnosis of WM.

To identify patients with WM and LPL in Norrbotten and Västerbotten, we used the Northern and SLR and the Swedish Cancer Registry for the years 1997–2011. WM/LPL was classified according to the REAL classification until 2003 and thereafter according to the WHO classification. All living patients (n = 93) fulfilling diagnostic criteria for WM (Figure 1) were sent a letter of invitation to participate in the study. The letter included an initial screening questionnaire regarding heredity for WM, other haematological malignancies, and autoimmune diseases. In total, 75 WM patients (75/93 or 81%) responded to the questionnaire. Ten patients from eight families (11%) reported a family history of WM and/or IgM MGUS and three patients from three families (4%) reported a family history of WM and MM. Totally 13 patients from 11 families (15%) reported a family history of WM, IgM MGUS, and/or MM.]. We decided to include the families with both WM and MM because they are much less described in the literature. Family members with IgM MGUS were included due to their risk to develop MW. One additional prevalent patient/family (diagnosed before 1997) was included and is part of the following description (n = 12).The patients identified from the registries with a diagnosis of WM and a relative with WM, MM, and/or IgM MGUS are defined as index patients and the relatives to the index patients with the diagnosis of WM, IgM MGUS, and MM are defined as affected relatives. Their diagnoses were confirmed by reviewing medical records or by contacting the treating physician.

Figure 1; Flowchart over patient selection process.

Next, we contacted the index patients by telephone for additional information and written informed consent and asked them questions regarding availability of relevant relatives (primarily first degree relatives but other relatives were accepted) for the study. Written informed consent was achieved from 14 index patients (two families had two index patients each) and 63 relatives (45 first degree relatives and 18 other relatives, foremost children of the siblings but also grandchildren and cousins). All index patients and their relatives answered an extended questionnaire on heredity and medical history. They also received instruction for blood sampling and 11 index patients and 56 relatives (89%) accepted (40 first degree relatives and 16 other relatives). No blood samples were obtained from two index patients (deceased) and one index patient and seven relatives only agreed to complete only the questionnaire. In the 12 families there were 75 additional potential first degree relatives, 34 of them deceased. In most cases the parents to the index patients were dead and in some families the index patients did not want the children to participate in the study, particularly if they were younger. All reported haematological malignancies and autoimmune diseases were confirmed by medical records

Complete blood counts, serum protein electrophoresis (SPE) with immunofixation (IF), LDH, beta-2-microglobulin, creatinine, and free light chains in serum (FLC/s) were analysed in the local clinical laboratory at Sunderby Hospital, Luleå or in the University Hospital in Umeå. SPE was performed with agarose gel electrophoresis technique on Hydrasys 2 (Sebia, ILS Laboratories Scandinavia/Thermo Fisher). The serum proteins were quantified by immunological turbidimetric assay (Cobas 6000 c501, Roche).

The study was approved from the Regional Ethical Review Board in Umeå, Sweden and conducted in accordance with the Declaration of Helsinki.

Results

Using data from the SLR (2000–2012), we calculated the age-adjusted overall incidence rate for WM in Norrbotten and Västerbotten to be 1.75 and 1.48 per 100 000 persons per year, respectively. The corresponding figure for WM/LPL in Sweden was 1.05 per 100 000 persons per year. Between 2000 and 2012, overall annual age-adjusted incidence rates for WM/LPL in Sweden appear to be relatively stable with small variations (0.88–1.21), although from 2004 to 2007 there was a decrease and after 2007 there was an increase.

Among the 75 patients in the northern region answering the questionnaire, we identified 11 families (15%) with two or more family members with WM, IgM MGUS, and/or MM. One additional family was included (in total 12 families; Table I). In 9/12 (75%) families, two or more relatives had WM and/or IgM MGUS. In 3/12 (25%) families, two or more relatives had WM, IgM MGUS, and/or multiple myeloma (MM). Pedigrees of two families are presented in Figures 2 and 3. The two most common relationships between the relatives were siblings and parent-child relationships. Only a few family members in each family were affected, mostly 2–4 persons (2.75 persons per family).

Figure 2. Pedigree of family D.

Figure 3. Pedigree of family I.

Table I. Characteristic of 12 families with WM, MM and/or IgM MGUS identified through Swedish and Regional Lymphoma Registry and the Swedish Cancer Registry (1997–2011).

Family ID	Number of affectedFamily members	Diagnosis	Gender and age at diagnosis	Relations of affected family members	Reported autoimmune disease	Reported haematological disease among relatives
A	3	WM, WM†, IgM MGUS	M76, M?, M48	Siblings, parent-child	–	–
B	4	WM, WM†, IgG MGUS,  monoclonal FLC/s	F52, M?, F33, F80	First cousins, parent-child, siblings	RA^b, Hypothyroidism, psoriasis, MS	CML
C	2	WM, WM	F29, F66	Parent-child	–	–
D	3	WM, WM, IgM MGUS	F62, M?, M52	Siblings	Hypothyroidism^c, PMR
E	2	WM,WM	M84, F85	Siblings	AHA^a	Acute leukemia
F	2	WM, IgM MGUS	M49, F80	Parent-child	ALS	–
G	3	WM, MM†, IgG MGUS	F80, M78, F80	First cousin, siblings	PMR^a	Hodgkin’s Lymphoma
H	4	WM, WM†, IgM MGUS,  IgM MGUS	F50, M78, F? F68	Parent-child, niece, first cousin	Hypothyroidism^a, ITP^a PMR^c, Bechterew’s disease^d	–
I	4	WM, IgM MGUS, IgM  MGUS, IgG MGUS	M66, M63, F72, F60	Siblings	RA^b, Systemic connective tissue disease^d	Follicular^c Lymphoma
J	2	WM, MM†	M71, M66	First cousin	Psoriasis, Psoriatic Arthritis, Arthritis of undetermined  significance	–
K	2	WM, WM	F73, M 85	Siblings	–	–
L	2	WM, MM	F69, F65	Siblings	Sjogren’s syndrome^b Hypothyroidism^a	Indolent B-cells lymphoma
Total 12					In 9/12 families	In 5/12 families

^aAutoimmune disease or hematological malignancy in index patient; Autoimmune disease or hematological malignancy in both index patient and affected or unaffected family member; Autoimmune disease or hematological malignancy in affected relative; ^dAutoimmune disease or haematological malignancy in affected and unaffected relatives. Not marked, Autoimmune disease or hematological malignancy in unaffected family member. AHA, autoimmune hemolytic anemia; ALS, amyotrophic lateral sclerosis; CML, chronic myeloid leukemia; F, female; ITP, immune thrombocytopenia; M, male; MS, multiple sclerosis; PMR, polymyalgia rheumatic; RA, rheumatoid arthritis. The index patients are marked with bold text. †, Deceased.

For the index patients, the median age at diagnosis was 70 years (range 29–85) and the gender distribution was equal, seven men and seven women. In the two families with parent-child relationship, the younger generation was diagnosed at an earlier age. In one of the families the daughter was diagnosed at 29 years, earlier than the mother (66 years) who was accidentally diagnosed when hospitalised for other reasons. In the other family the father (78 years) had the diagnoses before his daughter (50 years) was diagnosed. The gender distribution between all affected family members (index patients and affected relatives) was equal, 16 men and 17 women.

Autoimmune and haematological diseases reported in the families are described in Table I. A large part of the index patients (9/12; 75%) reported autoimmune diseases in their own or in their family’s medical history. Two cases with serious rheumatologic diseases with biological treatment were observed. In family H, one of the affected relatives with IgM MGUS had four of nine siblings with Bechterew’s disease. In family I, four of five siblings had a rheumatic disorder (three with rheumatoid arthritis and one with systemic connective tissue disease unspecified). In two of the index patients, autoimmune thrombocytopenia and autoimmune haemolytic anaemia were observed (Table I). Five of 12 (42%) families reported haematological malignancies in their families, in most cases lymphoproliferative disorders but also one myeloid disorder (CML). None of the index patients had another haematological malignancy, but one of the affected relatives had both IgM MGUS and follicular lymphoma.

SPE with IF were analysed in 11 index patients and 56 relatives (Figure 4). A MI was observed in 24/67 (36%) of the samples. The distribution between kappa and lambda was 16:7, with one MI not classified. In the same family, we observed monoclonal MI of both kappa and lambda type. The abnormal SPEs of the relatives are presented in Table II. A high proportion (12/56; 21%) of the relatives had a MI and nine of the MI were observed in 17% of the previously healthy unaffected relatives. The majority of the MGUS were of IgM type (n = 8), but three were of IgG type and one showed monoclonal FLC/s. All previously unknown MI fulfilled criteria for MGUS, but three of the MI were small and could only be detected with IF. When we repeated the SPE with IF, all three MI were persistent. We also observed disturbances in the other immunoglobulin (IG) levels in six of eight patients with IgM MGUS. Five of these patients had decreased levels of IgG, but one patient with severe RA had an increased level of IgG. Some explanations for the decreased IG levels among the patients with MGUS could be that one patient with an IgM MGUS had a previous rituximab treated follicular lymphoma and two patients had the diagnosis of IgM MGUS for many years. Perhaps these patients would fulfil the diagnostic criteria for WM if a bone marrow examination had been repeated, but this was not performed due to the patient’s comorbidity.

Figure 4. Serum protein electrophoresis in 11 patients with familial WM and 56 relatives. Of the 12 MGUS were 9 unknown.

Table II. Immunglobulin abnormalities in 27 family members with abnormal serum protein electrophoresis.

Immunoglobulin (g/l)	Number	IgG (6.7–14.5)	IgA (0.88–4.5)	IgM (0.27–2.1)
Decreased IG level	6
IgG	4		N
IgA	1	N	↓	N
IgM	1	N	N	↓
Polyclonal	4
IgA	2	N	↑	N
gM	2	N	N	↑
Oligoclonal	1	N	N	N
MGUS	12
IgM	8	3 ↓, 1 ↑	3↓	MI
IgG	3	MI	N	N
FLC/s	1	-	-	-
Acute-reaction inflammatory  pattern	2	-	-	-
MM	1	MI	↓	N
Number	27

x, mean, (normal range, g/l); ↓, decreased; ↑, increased; MI, monoclonal immunoglobulin; N, normal immunglobulin level.

Other IG abnormalities observed in 13/52 (25%) of the previous unaffected relatives were hypoglobulinemia, polyclonal hyperglobulinemia, oligoclonality, and acute-reaction protein pattern (Table II). All the relatives with decreased IgG or IgA had IgM levels near the lower normal range. Half of these patients had a coincident medical history of autoimmune diseases or allergy. Coincident medical history of autoimmune diseases was also seen in half (2/4) of the patients with polyclonal hyperglobulinemia, and the patient with oligoclonality had a connective tissue disease. Blood counts, white blood cell counts, platelets counts, LDH, beta-2-microglobulin, and creatinine were normal in unaffected relatives and not reported otherwise. FCL/s were analysed in 45/67 patients. An abnormal ratio was only seen in patients with known disease (WM, MM, and MGUS).

Discussion

According to the SLR, the age-adjusted overall incidence of WM in Norrbotten and Västerbotten counties is relatively high: 1.75 and 1.48 per 100 000 persons per year, respectively (Table III). The overall incidence of WM is unclear in most countries. Only a few studies are published and the incidence seems to have geographical and ethnic disparities with an overall incidence of 0.38 and 0.55 per 100 000 persons per year for the US and southeast England, respectively. An overall age-adjusted incidence of WM/LPL for Japan and Taiwan is 0.043 and 0.031 per 100 000 persons per year, respectively [15–17]. As we cannot reliably distinguish between WM and LPL in the SLR, it is not possible to compare the overall annual incidence of WM with other countries. To date, we have validated the WM and LPL diagnosis for Norrbotten and Västerbotten, and differential diagnosis of WM and LPL for rest of the country is ongoing.

Table III. Incidence of WM/LPL in different countries.

	Diagnosis	No. of cases	Calendar period	Age-adjusted incidence(per 100 000 persons per year)	Crude incidence(per 100 000 persons per year)
Sweden	WM/LPL	1279	2000–2012	1.05	1.07
Sweden–Norrbotten	WM/LPL	63	2000–2012	1.75	1.93
Sweden–Västerbotten	WM/LPL	52	2000–2012	1.48	1.55
9 SEER registries in US	WM	1835	1988–2007	0.38	–
South-east England	WM	152	1999–2001	0.55	–
Japan	WM/LPL	280	1996–2003	0.043	0.078
Taiwan	WM/LPL	56	1996–2003	0.031	0.032

Population-based cancer registries are often used to determine cancer incidence. In a validation of the Swedish Cancer Registry for lymphomas reported between 1964 and 2003, comparing two study populations with lymphoproliferative diseases recruited from the Swedish Cancer Registry and Swedish Inpatients Registry (hospital-based population), the diagnostic accuracy for WM was high (92.5%), but the overall completeness (i.e. the number of individuals actually registered) was low (68.1%). Patients diagnosed at older ages, earlier years, and with early-stage disease were overrepresented in the non-registered group [18]. Another factor that complicates the estimation is that the classifications for WM have changed over the years, resulting in parallel classifications. Furthermore, there is an overlap and a gradual development of IgM MGUS to asymptomatic or smouldering WM and WM. In routine Swedish medical care, many patients with a small IgM MI (<10 g/l) and no other signs of lymphoproliferatives disorder are categorised as IgM MGUS, and WM is not diagnosed until a bone marrow examination is performed, mostly due to disease progression or symptoms. Another complicating factor is that MGUS is not reported to the Swedish population-based registries resulting in lack of consistent data and the limitation is seen in other population-based registries around the world. Some of the patients with WM were registered as indolent B-cell lymphoma UNS in the Lymphoma Registry, in most cases, due to the fact that the clinicians reported the bone marrow diagnosis and did not take into account that the patient had an IgM MI. In our experience, the incidence of WM is probably underestimated. The incidence in this report is calculated on the number of patients primary registered in SLG

One explanation for the high incidence of WM in the northern Sweden may be that the region has a small but stable and in some areas isolated population influenced by both genetic and environmental factors. In this area of the country, there is a high prevalence of some rare hereditary diseases, such as acute intermittent porphyria, hereditary transthyretin amyloidosis, and Gaucher’s disease type III. These diseases have another genetic background (often autosomal dominant mutations with variant penetrance). In addition, because there is a high incidence in our geographical area, there could be strong genetic mechanisms also in WM.

Our population-based study confirms results from earlier family studies and our population has a high frequency of familiar WM (11%) and families with WM, MM, and/or IgM MGUS (15%). A large Swedish study that included 2144 patients with WM/LPL showed that first degree relatives had a 20 times higher risk of developing WM and an increased risk of developing other forms of non-Hodgkin’s lymphoma, chronic lymphocytic leukaemia, and MGUS, but no increased risk of developing MM or Hodgkin’s lymphoma [3]. A US study found that in 257 unrelated patients with WM 48 (18.7%) had at least one first degree relative with either WM (13 patients or 5.1%) or another B-cell disorder including NHL, CLL, myeloma (eight patients or 3.1%) or MGUS [2]. Thus, our families do not differ from other families described in the literature except that the gender distribution was equal for men and women [12,13]. In the families with both WM and MM 1/3 had a first degree relative with a small MI.

The WM families (75%) have a high prevalence of autoimmune diseases in accordance with the well established correlation between autoimmune diseases and lymphoma, including WM. Some autoimmune diseases, such as autoimmune thrombocytopenia and autoimmune haemolytic anaemia, are known manifestations of WM and were also seen in our families. Explanations for this correlation can be chronic inflammatory and/or antigen stimulation and genetic and/or environmental factors. The increased risk for WM is seen primarily in autoimmune diseases characterised by activation of B-cells and not of T-cells [8]. The chronic antigen-driven inflammation and lymphocyte activation are believed to contribute to the pathogenesis.

The initial screening questionnaire addressed heredity and autoimmunity and information bias cannot be excluded, since it is likely that the participants who answered the questionnaire were more motivated if they knew of heredity and/or autoimmune diseases in their relatives’ medical history. In addition, an increased diagnostic activity among the relatives can contribute to the increased risk for WM in some families. This detection bias is described in a study about increased risk for brothers to patients with prostate cancer with an especially high risk the first year after diagnosis. Moreover, in high socioeconomic groups prostate cancer was found in an earlier stage [19].

The prevalence for MGUS in the general population has significant geographic and ethnic disparities. For example, for Caucasians over 50 years of age the prevalence for MGUS is approximately 3% and the prevalence for IgM MGUS is approximately 0.5% [20–22]. In our study, there was a high frequency of MGUS (IgM, IgG, and monoclonal light chains) in the families (in 21% of the relatives of the index patients). An increased prevalence for MGUS in the relatives is seen also in other family WM studies [13,23].

MGUS is a known risk factor for developing WM (1.5% per year) and MM and indicates an excess risk for developing disease [24]. It is unknown whether the rate of progression to WM is higher in family cases. In our families, we also identified IgG MGUS and monoclonal FLC/s, which is observed in other studies, although more rarely. The high proportion of relatives with other IG abnormalities, such as hypoglobulinemia or polyclonal hyperglobulinemia, may indicate a higher risk of developing WM, MM, or other lymphophroliferative diseases. It is worth mentioning that in five of the affected patients (three of the index patients and two of the relatives with MGUS) a retrospective review of their medical records revealed that before they developed a MI they had a polyclonal hyperglobulinemia. In four of these, an underlying rheumatic disease might have been a contributing factor, but in one patient there was no underlying disease or other obvious explanation. Similar patterns are observed in other studies [14,25].

Some of our 12 families have a strong personal or family history of autoimmune diseases and disturbances in the IGs. Other families have no obvious comorbidity that could contribute to development of WM. These findings might reflect the heterogeneity of the disease and that there might be several causes of the pathogenesis. One limitation of this study, as well in other studies that examine familial WM, is the small number of families studied. To our knowledge, our study is one of the few population-based studies of familial WM.

In conclusion, we found a high incidence of WM in northern Sweden, a high prevalence of familiar WM, and strong relation between autoimmune/inflammatory diseases, haematological malignancies, and familial WM, a finding that strengthens the hypothesis that the aetiology of WM depends on both immune-related and genetic factors. More studies, however, are needed.