Temporal and geographic variations of Waldenstrom macroglobulinemia incidence

A large population-based study




Waldenstrom macroglobulinemia (WM) is a non-Hodgkin lymphoma (NHL) subtype. Little is known about the incidence and trends for this disease in the United States.


Twenty-year data from the Surveillance, Epidemiology, and End Results (SEER) program were used for this study. SEER*Stat was used for data analysis.


Of the 95,797 cases of NHL diagnosed between 1988 and 2007 in 9 SEER registries, 1835 (1.9%) were new cases of WM. Median age at diagnosis of WM was 73 years. The overall annual age-adjusted incidence was 0.38 per 100,000 persons per year, which increased with age, ranging from 0.03 in patients aged <50 years to 2.85 in patients aged ≥80 years. The incidence of WM was higher in men (0.54) than in women (0.27; P < .001) and was higher in whites (0.41) than in African Americans (0.18) or other races (0.21; P < .05). The annual percentage change for the whole population was 1.01% (P > .05). The annual percentage change was 1.21% for whites (P < .05) and 0.80% (P > .05) for nonwhites. Significant annual percentage change increases were seen in the group aged 70 to 79 years (1.24%; P < .05) and in 3 geographic registries (P < .001).


Although the overall incidence of WM remained steady over time, significant increases in incidence were seen over the past 20 years in whites, in those aged 70 to 79 years, and in 3 geographic registry areas. Cancer 2012. © 2011 American Cancer Society.


Waldenstrom macroglobulinemia (WM), a non-Hodgkin lymphoma (NHL) subtype, is a distinct B-cell lymphoproliferative disorder characterized by bone marrow infiltration with lymphoplasmacytic cells, along with evidence of an immunoglobulin M (IgM) monoclonal gammopathy.1 WM can be distinguished clinically from lymphoplasmacytic lymphoma (LPL) on the basis of detectable IgM monoclonal gammopathy in serum.2, 3 In the Surveillance, Epidemiology, and End Results (SEER) data from the National Cancer Institute, LPL and WM are coded separately. WM is a very rare disease and was reported first in 1944 as a malignant lymphoplasma proliferative disorder characterized by high levels of circulating IgM paraproteins and bone marrow infiltration by lymphoplasmacytoid cells.4

The consensus diagnostic criteria of WM were first presented at the Second International Workshop held in Athens, Greece in September 2002.2 The diagnostic criteria include IgM monoclonal gammopathy of any concentration; bone marrow infiltration by small B lymphocytes, plasmacytoid lymphocytes, and plasma cells in a diffuse, interstitial, or nodular pattern; and a surface Ig+CD19+CD20+CD5CD10CD23 immunophenotype.5-7 Patients might have symptoms attributable to tumor infiltration, circulating IgM, tissue deposition of IgM, or autoantibody activity of IgM. Symptoms of cytopenias, especially anemia associated with the replacement of the bone borrow with tumor cells, are common. Fatigue is another common presentation of WM. Patients may also present with symptoms of hyperviscosity associated with elevated IgM levels, including headache, blurred vision, and epistaxis.5-7 However, some patients with monoclonal IgM can be diagnosed by chance and are asymptomatic.8 These asymptomatic WM cases have also been referred to in the literature as smoldering WM.9

WM became reportable to the National Cancer Institute's SEER program in 1988.10 Other authors have used SEER data to report the incidence patterns and associated risk factors of WM, but their conclusions were based on limited data.10, 11 It was unknown whether the incidence of this disease had changed over the past 20 years or what factors were associated with these possible changes. In addition, no prior study compared incidence patterns for WM within a large population to confirm whether rate patterns differ according to age, race, sex, calendar time period, or geographic location. Therefore, this study took advantage of a large nationwide SEER program, covering 9 registries and accounting for 10% of the US population, to provide updated data on the temporal and geographic variations in the incidence of WM and the factors that may be associated with these changes.


Data Source

The SEER Program of the National Cancer Institute has compiled incidence data since 1973 from population-based cancer registries in 5 states and in 4 metropolitan areas that make up approximately 10% of the US population.12 Each SEER registry includes codes for all cancers, including WM, that are entered according to a standard classification scheme based on diagnostic pathology reports in medical records.13 Only the first matching record for each SEER WM case that matched the selection criteria was included in our analysis. The International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) code that was used for WM was 9761.

Similar methods were used to calculate the incidence of other morphologically and clinically distinct B-cell malignancies reported to the SEER program during the same time period, including multiple myeloma (MM; ICD-O-3 9732), chronic lymphocytic leukemia (CLL; ICD-O-3 9823), mantle cell lymphoma (MCL; ICD-O-3 9673), and follicular lymphoma (FL; ICD-O-3 9690, 9691, 9695, 9698).14 The number of new cancers may include multiple primary cancers occurring in 1 patient. We examined data for all patients diagnosed with WM between 1988 (the year WM became reportable to the SEER program) and 2007 who came from 9 geographic areas: Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco-Oakland, Seattle-Puget Sound, and Utah.

Data Analysis

The WM annual incidence rates were calculated for population subgroups according to age, sex, race, registry area, and time period. Age-specific rates were calculated for 5 age groups, expressed as new cases per 100,000 persons per year, and then directly adjusted to the 2000 US standard population. Age-adjusted, age-specific incidence rates were also ascertained for 5 broader age groups (<50, 50-59, 60-69, 70-79, ≥80 years). The annual percentage change, the average annual rate of change over the time series selected, was used to measure trends or changes in rates over time. Trends in WM and other related disease incidence rates were examined by age group, sex, and race for the 9 registries combined. Frequencies of the variables of interest and age-adjusted incidence rates with 95% confidence intervals were calculated by using the 2010 SEER*Stat 6.6.2 client-server program developed by Information Management Services Inc. in Silver Spring, Maryland.15 The incidence rates were compared by confidence intervals and the annual percentage change statistic to assess incidence patterns over time with a 2-sided P value that was calculated using weighted least squares methods.


Of the 95,797 NHL cases diagnosed between 1988 and 2007 in 9 SEER registries, 1835 (737 male and 1098 female [1.9%]) were new cases of WM. The median age at diagnosis was 73 years. The overall annual age-adjusted incidence of WM was 0.38 per 100,000 persons per year (Table 1). There was a steep increase in incidence with advancing age until age 60 to 69 years, after which incidence rates increased more slowly. The incidence rate was 0.03 of 100,000 per year among those aged <50 years but increased to 1.07 of 100,000 among those aged 60 to 69 years (35-fold) and 2.85 of 100,000 among those aged ≥80 years (95-fold). The relative risk of WM was 36.09 among those aged 60 to 69 years and 96.10 among the population aged ≥80 years, compared with those aged ≤50 years. The age-adjusted incidence rate among men (0.54 of 100,000) was twice as high as that among women (0.27 of 100,000; P < .001). Incidence rates were higher among whites (0.41 of 100,000) than among African American (0.18 of 100,000) and other racial groups (0.20 of 100,000).

Table 1. Waldenstrom Macroglobulinemia Incidence Rates by Age, Sex, and Race, and Surveillance, Epidemiology, and End Results Registries From 9 Geographic Areas, 1988-2007
CharacteristicNo. of CasesAge-Adjusted Incidence Rate (95% CI)RR (95% CI)
  • Abbreviations: CI, confidence interval; RR, relative risk.

  • a

    Eleven patients were of unknown race.

 <50 years1090.03 (0.02-0.04)1.00
 50-59 years2450.44 (0.39-0.50)14.93 (11.86-18.90)
 60-69 years4021.07 (0.97-1.18)36.09 (29.13-45.03)
 70-79 years6262.30 (2.13-2.49)77.79 (63.35-96.25)
 ≥80 years4532.85 (2.59-3.12)96.10 (77.78-119.59)
 Female7370.27 (0.25-0.29)1.00
 Male10980.54 (0.51-0.58)2.04 (1.86-2.25)
 White16590.41 (0.39-0.43)2.25 (1.76-2.90)
 African American740.18 (0.14-0.23)1.00
 Other910.21 (0.17-0.26)1.17 (0.84-1.63)
Geographic areas   
 Atlanta750.20 (0.15-0.25)1.00
 Utah720.25 (0.19-0.31)1.25 (0.89-1.77)
 Connecticut2430.34 (0.29-0.38)1.71 (1.30-2.26)
 New Mexico1100.34 (0.28-0.41)1.90 (1.37-2.64)
 Hawaii890.37 (0.30-0.46)1.88 (1.44-2.47)
 Detroit2850.37 (0.33-0.42)1.97 (1.51-2.60)
 San Francisco2950.39 (0.35-0.44)1.74 (1.27-2.38)
 Iowa3050.46 (0.41-0.51)2.33 (1.78-3.06)
 Seattle3610.52 (0.46-0.57)2.62 (2.02-3.43)
 1988-19923550.34 (0.30-0.38)1.00
 1993-19974270.37 (0.33-0.40)1.09 (0.94-1.26)
 1998-20025150.41 (0.38-0.45)1.22 (1.06-1.40)
 2003-20075380.39 (0.36-0.43)1.17 (1.02-1.34)
 Total18350.38 (0.36-0.40)1.00

There were substantial geographic variations in incidence rates across registries (Table 1). Incidence rates were lowest in Atlanta (0.20 of 100,000 persons per year) and highest in Seattle (0.52 of 100,000), with corresponding relative risks between 1.00 and 2.62 compared with the incidence rate in Atlanta. The WM incidence among the various registries can be divided into 3 levels: low (<0.30 for Atlanta and Utah), middle (0.30-0.45 for Connecticut, NM, Hawaii, Detroit, and San Francisco), and high (>0.45 for Iowa and Seattle).

Overall annual age-adjusted incidence rates from 1988 to 2007 appeared relatively stable with minimal variation, increasing from 1988 to 1992 through 1998 to 2002, but decreased somewhat thereafter. Incidence rates before and after publication of consensus diagnostics criteria were compared. The overall incidence rate from 1988 to 2002 (0.37 of 100,000) was not significantly different from the rate from 2003 to 2007 (0.39 of 100,000).

Because African American and other races had similar incidence rates, and each group had a small number of cases, we merged these 2 groups into “nonwhite” races. Males had higher incidence rates than females for all 9 SEER registries (Table 2). The highest rates were in Seattle for both males (0.71) and females (0.38). The lowest rates were in Atlanta for both males (0.30) and females (0.14). The male-to-female incidence rate ratio (IRR) ranged from 1.9 in Seattle to 2.3 in Utah. The highest and lowest rates for whites were in Seattle (0.54) and Atlanta (0.22). For nonwhites, the highest rate was in Hawaii (0.28), and the lowest was in Utah (0.00).

Table 2. Age-Adjusted Incidence Rates of Waldenstrom Macroglobulinemia in Registries by Sex, Race, and Age Groups
SEER RegistrySexRace
 Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.
  1. Abbreviation: CI, confidence interval.

Atlanta0.30 (0.21-0.41), 430.14 (0.10-0.20), 320.22 (0.17-0.28), 600.13 (0.07-0.22), 15
Utah0.36 (0.26-0.48), 460.16 (0.10-0.23), 260.25 (0.20-0.32), 720.00 (0.00-0.33), 0
Connecticut0.48 (0.41-0.57), 1440.23 (0.19-0.29), 990.34 (0.30-0.39), 2330.16 (0.06-0.31), 8
Hawaii0.52 (0.39-0.67), 570.25 (0.17-0.35), 320.64 (0.45-0.89), 370.28 (0.21-0.37), 51
Detroit0.54 (0.46-0.63), 1680.26 (0.22-0.31), 1170.41 (0.36-0.46), 2480.20 (0.14-0.28), 35
San Francisco0.56 (0.48-0.65), 1810.26 (0.22-0.31), 1140.45 (0.40-0.51), 2510.19 (0.14-0.26), 41
New Mexico0.50 (0.39-0.63), 700.23 (0.16-0.31), 400.36 (0.30-0.44), 1080.08 (0.01-0.27), 2
Iowa0.65 (0.56-0.75), 1770.33 (0.28-0.40), 1280.46 (0.41-0.51), 3020.08 (0.00-0.44), 1
Seattle0.71 (0.61-0.81), 2120.38 (0.32-0.45), 1490.54 (0.49-0.60), 3480.19 (0.10-0.34), 12
All registries0.54 (0.51-0.58), 10980.27 (0.25-0.29), 7370.41 (0.39-0.43), 16590.20 (0.17-0.24), 165

A male predominance throughout all age groups was observed. A sharp increase in incidence rates was observed with advancing age in both sexes and both races until age 60 to 69 years (Table 3). Male-to-female IRR decreased from a high of 2 (<50 years) to 1.7 (60-69 years), then increased to 2.2 (70-79 years) and 2.4 (≥80 years) thereafter. The increase with age was greater among males than among females and among whites than among nonwhites. Incidence rates among males and females aged ≥80 years were 115 and 96× greater, respectively, than were rates among males and females aged <50 years. The incidence rate for whites aged ≥80 years was 100× greater than the rate was for whites aged <50 years; the incidence rate for other races aged ≥80 years was 54× greater than the rate for nonwhites aged <50 years. In males, the overall incidence rates were 0.59 (0.55-0.62) for whites and 0.29 (0.23-0.35) for nonwhites; in females, the overall incidence rates were 0.29 (0.27-0.31) for whites and 0.14 (0.10-0.17) for nonwhites (Table 3).The overall incidence rate among white males (0.59) was 4× higher than that among nonwhite females (0.14; Table 3).

Table 3. Age-, Sex-, and Race-Specific Age-Adjusted Incidence Rates of Waldenstrom Macroglobulinemia
Age GroupSexRaceSex and Race
 Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95% CI), No.Rate (95%CI), No.
  1. Abbreviation: CI, confidence interval.

<50 years0.04 (0.03-0.05), 640.02 (0.02-0.03), 450.03 (0.03-0.04), 890.02 (0.01-0.04), 180.04 (0.03-0.05), 520.03 (0.01-0.05), 100.03 (0.02-0.04), 370.02 (0.01-0.04), 8
50-59 years0.59 (0.50-0.68), 1580.31 (0.25-0.38), 870.47 (0.41-0.54), 2100.31 (0.21-0.44), 330.60 (0.50-0.71), 1320.51 (0.33-0.75), 250.34 (0.27-0.43), 780.14 (0.06-0.28), 8
60-69 years1.37 (1.20-1.55), 2400.81 (0.69-0.94), 1621.16 (1.05-1.29), 3590.61 (0.43-0.82), 411.48 (1.29-1.69), 2150.80 (0.51-1.19), 240.89 (0.75-1.04), 1440.45 (0.26-0.72), 17
70-79 years3.36 (3.03-3.71), 3861.53 (1.35-1.74), 2402.52 (2.32-2.73), 5761.06 (0.78-1.42), 463.64 (3.27-4.04), 3531.73 (1.17-2.46), 311.70 (1.48-1.94), 2230.59 (0.33-0.97), 15
≥80 years4.63 (4.07-5.24), 2501.93 (1.67-2.21), 2033.08 (2.80-3.39), 4251.28 (0.85-1.87), 275.13 (4.50-5.84), 2371.53 (0.79-2.71), 122.05 (1.77-2.37), 1881.13 (0.63-1.86), 15
All ages0.54 (0.51-0.58), 10980.27 (0.25-0.29), 7370.41 (0.39-0.43),16590.20 (0.17-0.24), 1650.59 (0.55-0.62), 9890.29 (0.23-0.35), 1020.29 (0.27-0.31), 6700.14 (0.10-0.17), 63

For the whole population, the incidence of WM from 1988 to 2007 remained relatively stable, ranging from 0.32 in 1997 to 0.49 in 2002, with an annual percentage change of 1.01% (P > .05; Fig. 1). The incidence rates for men ranged from 0.68 in 2001 to 2002 to 0.44 in 1995 (Fig. 2, Top), and for women from 0.40 in 2003 to 0.18 in 1998. The annual percentage change was 0.84% (P > .05) for men and 1.22% (P > .05) for women. The annual percentage change was 1.21% (P < .05) for whites and 0.80% (P > .05) for nonwhites. The incidence for whites significantly increased over the years.

Figure 1.

Age-adjusted incidence trends of Waldenstrom macroglobulinemia and other related cancers are shown (1988-2007). CLL, chronic lymphocytic leukemia; FL, follicular lymphoma; MCL, mantle cell lymphoma; MM, multiple myeloma; NHL, non-Hodgkin lymphoma; WM, Waldenstrom macroglobulinemia.

Figure 2.

Age-adjusted incidence rates of Waldenstrom macroglobulinemia are shown by sex, race, and age groups (1988-2007).

The WM incidence increased with age from 1988 to 2007 (Fig. 2, Bottom). The annual percentage change was −1.61% (P < .05) for those <50 years of age, 1.69% (P > .05) for those aged 50 to 59 years, 0.92% (P > .05) for those aged 60 to 69 years, 1.24% (P < .05) for those aged 70 to 79 years, and 1.07% (P > .05) for those aged ≥80 years. The incidence rates for those aged <50 years compared with other age groups showed a decreasing trend, although the decrease was not significant. Other age groups' incidence rates had an increasing trend, but only the group aged 70 to 79 years had a significant increase.

The trends of WM incidence rate in 9 registries are displayed in Figure 3: Connecticut (3.04), Detroit (2.48), and Seattle (1.59) experienced significant upward trends (P < .001), whereas Atlanta (−0.26), Iowa (−0.66), Utah (−0.80, 1990-2007), and New Mexico (−2.56, 1988-2006) showed declines, although the changes were not statistically significant.

Figure 3.

Age-adjusted incidence trends of Waldenstrom macroglobulinemia at different registries are shown.

Figure 1 shows the age-adjusted incidence rate trends in WM, NHL, MCL, CLL, FL, and MM. The trends analysis supports the observation that the incidence rates of NHL and MCL experienced significant increases over the past 20 years, but the incidence rates of WM, MM, and CLL remained at a relatively stable level. The annual percentage change for NHL was 1.3% (P < .05); for MCL and FL, it was 8.3% (P < .05) and 1.10% (P < .05), respectively. The annual percentage change for MM was 0.04% (P > .05).


We used comprehensive 20-year data from the SEER program to provide broad descriptive analyses of incidence patterns and temporal trends for WM, with detailed evaluation according to age, sex, race, calendar year, and geographic locations. To our knowledge, the current large-scale study (N = 1835) is the first to describe such demographic data and geographic variations. We observed diverse incidence patterns and substantial racial and geographic heterogeneity in incidence of WM within 9 SEER registries during 1988 to 2007. Our analysis included nearly 3× as many cases as the next most recent US report of the incidence of WM, which was published >12 years ago.10

The overall annual age-adjusted incidence rate of WM in the 9 registries during 1988 to 2007 was 0.38 per 100,000 persons per year, with the incidence in women at 0.27 and in men at 0.54. Other US studies have estimated incidence of WM at 0.17 to 0.25 for females and 0.34 to 0.61 for males per 100,000 persons per year.10, 11 A later European population-based study reported an incidence of 0.42 for females and 0.73 for males per 100,000 persons per year.16 The incidence rates are different but with similar ratios between the sexes. The reason for this difference is not obvious but may be accounted for in part by the reference population used to standardize the incidence.

Our findings of higher incidence rates of WM throughout all age groups and calendar years and within each race are consistent with those of previous publications.10, 11 Although male sex is an important risk factor for most hematological malignancies, the causal factors underlying this association are unknown. Occupational and environmental exposures are always of potential concern in male-predominant cancers. At present, there is no compelling evidence that links WM to specific occupational or environmental exposures, tobacco, or alcohol,17 although higher risk in WM incidence (18.7%) among first-degree relatives was reported, suggesting the potential role of host susceptibility in WM etiology.18

Our finding of a dramatic increase in the incidence of WM with advancing age is consistent with a multistep model of carcinogenesis described previously.19 Our findings suggest that WM may evolve through a multistep process. In addition, further investigation into the biology of aging and age-related effects such as immune senescence have shown that age is particularly important in cancers whose rates increase sharply with age, such as WM, FL, and MM20; this is because chronic inflammation, DNA damage, and diminished immune surveillance, which occur more frequently in older persons, are also linked with cancer.21, 22

NHL is among the few neoplasms whose incidence and mortality rates have been rising in both sexes in Europe and North America since 1970.23-26 Our results showed a significant increase in NHL incidence rates from 1988 to 2007; but incidence rates of WM, a subtype of NHL, in the whole population did not follow a similar trend, suggesting that exposure to still-unknown causes of WM may have not changed.

WM may be misdiagnosed and/or underreported to population-based cancer registries. It has the features of indolent lymphomas such as lymphadenopathies, splenomegaly, and bone marrow infiltration.6, 7 The clinical presentation of WM is similar to that of MM as well: anemia and symptoms of hyperviscosity caused by elevated IgM levels.6, 7, 27 A diagnosis of WM requires the detection of increased IgM levels with monoclonal paraproteins, which are also associated with CLL, MCL, FL, diffuse large B-cell lymphoma, some T-cell lymphomas, and IgM-monoclonal gammopathy of undetermined significance. Identification of the paraprotein requires serum protein electrophoresis (SPEP), which is often not done in community hospitals. If these hospitals, which may not have the ability to measure IgM or to perform SPEP, contribute SEER data, this could lead to the underdetection of the IgM paraprotein and misdiagnosis of WM. Because WM is commonly diagnosed or managed outside of hospitals, it is possible that many of these cases are unreported to population-based registries.28 Therefore, the true incidence of WM may be underestimated, which may also partially explain the variations.

One potential limitation of our study is the lack of independent verification of diagnoses by expert hematopathologist review. Each SEER registry is responsible for finding every case of cancer in its defining registry area. Because IgM monoclonal gammopathies exist in other B-cell malignancies, and some patients were asymptomatic, misdiagnosis and under-registration in SEER registries are potential issues. Also, studies that included WM cases reported before 2001 may have used codes from earlier ICD-O versions, which require conversion to ICD-O-3 and could have resulted in higher proportions of unclassified cases (eg, lymphoma and not otherwise specified). Turner et al29 studied the agreement of computer-converted ICD-O codes with ICD-O-3 codes generated directly from diagnostic pathology reports and the reproducibility of unclassified cases and found that data from WM cases had a relatively low reliability. Another potential limitation was that some results that had been based on a small number of events were subject to random variation and were too statistically unreliable for presentation; furthermore, the incidence rates were too scattered to justify calculation of the trends in some registry areas. More data with a large sample size are needed before a meaningful explanation of this phenomenon can be generated.

In conclusion, our population-based study provides updated trends in the incidence of WM in the United States and describes patterns that have not been reported previously. The diverse incidence patterns by age, race, sex, and geographic location may be helpful in further exploring a specific etiology. These data may be useful in planning public health strategies, identifying high risk populations, and understanding the etiology of this cancer for future prevention.


This research was supported in part by the Crutchfield and Kimmel Research Funds.


The authors made no disclosures.