Protein electrophoresis, immunoelectrophoresis and immunofixation electrophoresis as predictors for high-risk phenotype in familial Waldenström macroglobulinemia


  • Mary L. McMaster,

    Corresponding author
    1. Genetic Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
    • Genetic Epidemiology Branch, DCEG, NCI, NIH, DHHS, 6120 Executive Boulevard, MSC 7236 Bethesda, MD 20892-7236, USA
    Search for more papers by this author
    • Fax: +301-402-4489.

  • Gyorgy Csako

    1. Department of Laboratory Medicine, Clinical Center, National Institutes of Health, Department of Health and Human Services, Bethesda, MD
    Search for more papers by this author


Protein electrophoresis is used for the detection, evaluation and follow-up of monoclonal gammopathy (MG) conditions such as Waldenström macroglobulinemia (WM). Immunofixation electrophoresis (IFE) is currently the most common method for isotyping of monoclonal gammopathy because of its superior sensitivity relative to immunoelectrophoresis (IEP). We designed a study to evaluate the clinicobiological relevance of small monoclonal bands detected by serum protein electrophoresis, IEP, and IFE. Serum protein electrophoresis, IEP, and IFE were used to evaluate possible monoclonal gammopathy in 46 members (29 relatives and 17 nonbloodline spouses) from 3 families with multiple cases of WM. IFE identified small monoclonal bands initially missed by IEP in 5 individuals (2 blood relatives, 3 spouses) among 46 study participants. All bands were IgM type. Twenty-three individuals, including the 2 blood relatives and 2 of 3 spouses with monoclonal gammopathy, were then followed for a median of 17 years (range, 13–25). The monoclonal gammopathy progressed in the 2 relatives but disappeared in the spouses, and new IgM MG developed in 2 additional relatives with a prior history of IgM polyclonal gammopathy. Small monoclonal bands detected by IFE in a familial context may be biologically meaningful, both as phenotypic biomarkers and possibly as predictors of high risk for WM. Polyclonal IgM may also be a marker of genetic susceptibility in WM families. Larger studies are needed to confirm these observations. © 2007 Wiley-Liss, Inc.

Waldenström macroglobulinemia (WM) is a chronic lymphoproliferative disorder characterized by aberrant production of monoclonal immunoglobulin type M (IgM) in the setting of histological evidence of lymphoid malignancy, most commonly lymphoplasmacytic lymphoma.1 Although rare, familial clustering of WM is known to occur.2 In family studies, the ability to identify those individuals at high risk for cancer is desirable so that genetic analyses can be performed optimally.3 Apart from certain dysmorphic syndromes that confer increased risk for cancer (e.g., Fanconi anemia), few familial hematolymphopoietic cancer syndromes have phenotypic markers associated with cancer risk.

Serum protein electrophoresis followed by immunoelectrophoresis (IEP) or immunofixation electrophoresis (IFE) is integral to the recognition, diagnosis and clinical follow-up of WM.4 Methods to identify and type abnormal bands have evolved over time. Currently, IFE is the most common method for the detection and isotyping of monoclonal bands.5, 6 IFE is more sensitive than IEP7 for the detection of small monoclonal bands, but the clinical and biological significance of these small bands is unclear.5 Family studies in WM have identified a relatively high frequency of IgM monoclonal gammopathy (IgM MG) among relatives of WM patients.2 However, many questions remain regarding this finding, including whether small monoclonal bands have pathobiological significance in the familial context.

To address this question, we evaluated 3 WM families longitudinally using protein electrophoretic studies. Initially, we used IEP to identify and characterize monoclonal bands in relatives of WM patients. Because we routinely collected and cryopreserved serum, we were later able to use IFE to compare with the initial IEP results. We also obtained serial protein electrophoretic data over a span of up to 25 years to evaluate the clinical significance of any small monoclonal bands detected by IFE. The results of these evaluations are presented here.


IEP, immunoelectrophoresis; IFE, immunofixation electrophoresis; MG, monoclonal gammopathy; MGUS, monoclonal gammopathy of undetermined significance; PG, polyclonal gammopathy; QIg, quantitative immunoglobulins; WM, Waldenström macroglobulinemia.

Material and methods

Study subjects

This study was nested within a comprehensive ongoing investigation of familial WM. Three families were referred to the National Cancer Institute, National Institutes of Health (NIH) in the 1970s and 1980s because of the diagnosis of WM in at least 2 patients in each family. Among the 3 families, there were 8 known patients with WM. In addition, there was a single known case of IgM MG in each of two families and a single case of non-Hodgkin lymphoma (NHL) in one family. The diagnosis of WM was validated in 7 of the 8 cases (pathology report, n = 6, 75%; referring physician report, n = 1, 16%). We were unable to obtain confirmatory records for 1 patient.

All adult (age ≥18 years) living first- and second-degree blood relatives of WM patients, as well as spouses of patients or of patients' relatives, were eligible to participate. The study was conducted under Institutional Review Board (IRB) approval, and all patients gave informed consent for sample collection and analysis according to current IRB standards. Thirty-four of 36 eligible relatives (94%) and 21 spouses consented to participate. Relatives completed detailed medical history questionnaires.

Laboratory methods

At initial evaluation in the 1970s/1980s, all participants were requested to provide serum and urine for protein electrophoretic analysis and a concurrent serum specimen for cryopreservation at −70°C. Initial sample serum and urine protein electrophoresis and IEP were performed from 1977 to 1980 by Bio-Science Labs (Columbia, MD) and subsequently through 1990, by the NIH Department of Laboratory Medicine (DLM), MetPath (Kensington, MD) and/or SmithKline Beecham (Norristown, PA). Beginning in 2000, all original participants were recontacted for reevaluation with repeat protein electrophoresis and other laboratory studies. All protein electrophoresis analyses during this period were accomplished by IFE in agarose gel at the NIH DLM and reviewed by one of the authors (GC). Participants underwent routine immunoglobulin quantitation by immunoturbidimetry or immunonephelometry at both initial evaluation and follow-up. During the recontact phase, we retrieved cryopreserved serum whenever possible and performed IFE to confirm the original IEP results. We performed the immunofixation assay in duplicate whenever adequate cryopreserved serum was available. Most members from one family, originally referred in 1976, had been reevaluated in the 1980s when the other two families were referred. Because the population risk of developing a monoclonal band is age-dependent,8 we chose to use material contributed by those participants during their 1980s evaluation so as to minimize the likelihood of a false-negative result and to conserve biospecimen resources. For this group we planned that, should we detect an unsuspected IgM MG in the 1980s sample, we would then test serum from the earlier evaluation to determine whether the band(s) had been present at the earlier date.

Data analysis

Of a total of 55 eligible subjects, 9 were excluded from the analysis: 3 (all blood relatives) had an unequivocal monoclonal band identified by the initial IEP, 3 (2 relatives and 1 spouse) had insufficient cryopreserved serum for IFE, and 3 (all spouses) did not have IEP performed at baseline and were not available for follow-up. Among the remaining 46 subjects, all had both IEP and IFE performed on a baseline serum sample and of these, 29 had at least one follow-up IFE performed on fresh serum during the recontact phase. Thirty-seven subjects (80.4%) had parallel serum and urine protein electrophoresis on at least one occasion. Reasons for nonparticipation in the follow-up phase included death (n = 1), terminal illness (nonhematolymphoid malignancy, n = 1), patient refusal (n = 7), loss to follow-up (n = 4), and logistical obstacles (n = 4). The median interval between initial evaluation and follow-up IFE was 17 years (range, 13–25).

Immunoglobulin concentrations were converted to a percentage of the upper limit of the normal range (%ULN) reported by a given laboratory at the time of the assay to adjust for variations in methodologies and standardization across laboratories. Monoclonal immunoglobulins are named according to their heavy chain (IgG, IgM, or IgA) and light chain (kappa [κ] or lambda [λ]) class. Polyclonal gammopathy was defined as a polyclonal pattern by IEP or IFE in association with an elevated quantitative level of one or more immunoglobulins.

Follow-up studies

Participants who were found to have monoclonal gammopathy with elevated quantitative levels of the corresponding immunoglobulin on screening were offered additional evaluation. Consenting participants underwent bone marrow aspiration and biopsy, which was reviewed by an experienced hematopathologist to determine whether there was an asymptomatic hematolymphoid malignancy associated with the monoclonal band. Patients who declined bone marrow examination were categorized as having MGUS and were referred to their local physician for clinical follow-up.

Statistical analysis

Data were classified according to a variable's median and range when continuous or absolute and relative frequencies when the variable was categorical. All tests for significance of differences between variables were two-sided.


Characteristics of the 46 study subjects are presented in Table I. There were no significant differences in age or gender ratios between relatives and spouses. Spouses tended to be older because they included spouses of cases and of siblings of cases, whereas relatives were predominantly offspring of cases.

Table I. Characteristics at Ascertainment of 46 Study Subjects from three Multiple-Case WM Families
 Relatives (n = 29)  Spouses (n = 17)All participants (n = 46)
  • QIg, quantitative immunoglobulin levels; IgG, immunoglobulin type G; IgA, immunoglobulin type A; no., number.

  • 1

    The difference in median age between relatives and spouses is not statistically significant by two-tailed t-test.

  • 2

    The difference in gender proportions between relatives and spouses is not statistically significant by the χ2 test.

Median age at initial evaluation [range] (years)35 [24–69]43 [24–69]36 [24–69]1
 Male20 (69%)8 (47%)28 (61%)2
 Female9 (31%)9 (53%)18 (39%)
Relationship to cases   
 Bloodline relatives29 (100%)29 (63%)
  1st degree22/29 (76%)22/46 (48%)
  2nd degree7/29 (24%)7/46 (15%)
 Nonbloodline17 (100%)17 (37%)
Initial IEP results   
 Polyclonal with normal or decreased QIg22 (75%)15 (88%)37 (80%)
 Polyclonal with increased QIg7 (24%)2 (12%)9 (20%)
  IgG0/22 (0%)0/15 (0%)0/37 (0%)
  IgA2/22 (7%)0/15 (0%)2/37 (4%)
  IgM5/22 (17%)2/15 (12%)7/37 (9%)
No. with follow-up24 (83%)6 (35%)30 (65%)
Median age at follow-up [range] (years)52 [39–80]77 [52–85]54 [43–85]
Median follow-up interval [range] (years)17 [13–24]19 [16–25]18 [13–25]

At the initial evaluation, all subjects had a polyclonal Ig distribution in serum by IEP, without evidence of monoclonal bands. All urine IEP studies were normal. Performing IFE on cryopreserved serum from this initial evaluation, we identified a monoclonal band in 5 subjects (11%, Fig. 1a). In all 5 cases, the band was of IgM type and was too weak to be quantitated by densitometry. Three of the 5 had normal quantitative immunoglobulin (QIg) levels, while 2 had elevated quantitative IgM levels. The median age of these 5 subjects at detection of IgM MG was 41 years (range, 27–67). Two of the participants (Patients R1 and R2) were first-degree relatives of WM cases, and the others were spousal controls.

Figure 1.

Electrophoretic results in blood relatives of WM patients and non-bloodline spouses. Panel A shows results for all blood relatives and all spousal controls. At initial evaluation, 2 blood relatives and 3 spouses had monoclonal bands. During follow-up, existing monoclonal bands persisted in relatives and resolved in spouses (1 spouse was lost to follow-up), and 2 additional blood relatives developed new monoclonal bands. Median age at first detection of a monoclonal band was 46 years (range, 41–52) in blood relatives and 57 years (range, 27–67) in spouses. Abbreviations: PC, polyclonal; MC, monoclonal. Panel B shows results for all relatives, stratified by the presence or absence of PG. Among 8 blood relatives classified as having PG by IEP, 2 were discovered to have an IgM MG when their serum was tested using IFE. In both cases, the IgM MG persisted at follow-up. Among the remaining 6 blood relatives with IFE-confirmed PG, 3 had subclinical autoimmune thyroiditis (2 with IgA PG, 1 with IgM PG), and 1 each had ulcerative colitis (IgM), NHL (IgM) and HIV (IgM/IgG). Note that all patients with autoimmune thyroiditis were members of 1 previously described family9 and that 2 other members of that family had clinical thyroiditis in the absence of PG. The relative with IgM PG/HIV had a history of normal serum protein electrophoretic and quantitative assays 10 years previously. At follow-up, 1 of the 2 patients with IgA PG had persistently elevated IgA and the other had normal immunoglobulins. The patient with IgM PG/subclinical autoimmune thyroiditis had normal immunoglobulins at follow-up; the patient with IgM PG/ulcerative colitis had persistent IgM PG; and the patients with IgM PG/NHL and HIV each developed an IgM MG.

Among the remaining 41 patients, polyclonality was confirmed by IFE of cryopreserved serum. Because we were interested in the possible significance of polyclonal gammopathy in the familial WM setting, we stratified the relatives based on their QIg results (Fig. 1b). Twenty-one had normal QIg levels, while 6 had quantitative elevations of either IgM (n = 4) or IgA (n = 2) and were classified as having polyclonal gammopathy (Table II).

Table II. Associated Conditions in Relatives with IgM Polyclonal Gammopathy at Initial Evaluation and follow-up Evaluation
PatientInitial evaluationFollow-up evaluation
IFE%ULNAssociated condition(s)1IFE%ULNAssociated condition(s)
  • ATG, anti-thyroglobulin antibody; ACA, anti-cardiolipin antibody; ATM, anti-thyroid microsomal antibody; ATP, anti-thyroperoxidase antibody.

  • 1

    Associated conditions include (i) clinical conditions (e.g., autoimmune thyroiditis) that the patient either reported on a detailed medical history questionnaire or was discovered to have during clinical evaluation; or (ii) abnormal laboratory values for autoantibodies in the absence of signs or symptoms of clinical disease.

  • 2

    The patient had no evidence for recurrent NHL throughout the study period.

1IgM PG133None↑ ATGNormal50Thyroiditis↑ ATG
2IgM PG110HIV+NormalIgM MG100HIV+↑ ATG
3IgM PG183NoneNormalIgM PG204None↑ ACA
4IgM PG136NHL (in remission)2NormalIgM MG212NHL (in remission)↑ ACA
5IgA PG106None↑ ATMIgA PG102Hypothyroidism↑ ATP
6IgA PG121NoneNormalNormal112None↑ ATG/ATP

Overall, 23 patients were available for follow-up, including 15 with normal electrophoresis and QIg and all 6 with polyclonal gammopathy. Among 15 who were normal initially, all continued to have normal IFE and QIg results. Among the 6 relatives with confirmed polyclonal gammopathy initially, 2 had normal IFE and QIg findings at follow-up, 2 had persistent polyclonal elevations of IgM or IgA, respectively, and 2 developed IgM(κ) monoclonal bands.

Details of all subjects found to have monoclonal gammopathy at some point during the study are presented in Table III and Figure 2. For Patient R1, IFE identified a monoclonal IgM(λ) band in cryopreserved serum. Follow-up IFE demonstrated persistence of a monoclonal band whose mobility matched the band detected in the cryopreserved specimen. Smoldering (asymptomatic) WM was subsequently diagnosed with bone marrow examination. At last follow-up, 39 months since diagnosis of WM, the patient remained untreated and asymptomatic with normal hematopoiesis, gradually increasing IgM levels, progressive immunoparesis (decrease in the levels of uninvolved Igs), and emergence of 2 additional small monoclonal bands [IgM(λ) and IgM(κ), respectively].

Figure 2.

Natural history of 7 individuals found to have small monoclonal bands by IFE during the study period. Solid symbols represent time points at which IFE demonstrated a monoclonal band. Open symbols represent time points at which IFE demonstrated a normal polyclonal pattern. All monoclonal bands were of IgM type. Patients R1–R4 are relatives; Patients S1–S3 are spouses. All relatives with IgM MG had quantitative IgM levels at or above the upper limit of normal, and their monoclonal bands were persistent. All spouses had IgM levels within the normal range, and the monoclonal bands were evanescent in the 2 spouses who were available for follow-up.

Table III. Details of all Subjects Found to have a Monoclonal Band by IFE
SubjSexInitial evaluation, cryopreserved serumFollow-up evaluation, fresh serumLast follow-up
AgeQIgM1 (%)MC bandAgeQIgM (%)MC bandAgeQIgM(%)MC band
  • Subj, subject; MC, monoclonal; PC, polyclonal; MS, moderately strong; W, weak; n.m., not measurable; na, not available.

  • 1

    Because freezing can cause aggregation of IgM molecules and a falsely low value, the value from the original QIg assay is reported here.

  • 2

    Quantification reflects the size of only the largest band observed.

  • 3

    A weak monoclonal band was observed in only one of 2 duplicate samples.

  • 4

    This individual did not have sufficient cyropreserved material for a duplicate determination to confirm the presence of a weak IgM(κ) band.

 R1M41123IgM(λ)Wn.m.56256IgM(λ), IgM(λ)MS, W0.4 g/dl60304IgM(λ),IgM(λ), IgM(κ)MS, W, MS0.7 g/dl
 R2M39181IgM(λ)Wn.m.54430IgM(λ), IgM(κ), IgM(κ)MS, W, W0.4 g/dlnanananana
 R3M35110PC52100IgM(κ), IgM(κ)MS, W0.4 g/dl52137IgM(κ), IgM(κ)MS, W0.3 g/dl
 S1M6735IgM(κ)W3n.m.8430Not present8627Not present
 S2F5721IgM(κ)W4n.m.7228Not present7717Not present

Patient R2 was also found to have a small IgM(λ) monoclonal band present in cryopreserved serum, and IFE at 15-year follow-up identified an identical IgM(λ) band accompanied by 2 smaller IgM(κ) bands. He remained asymptomatic with normal hematologic and other laboratory findings and declined bone marrow evaluation and additional follow-up.

The 3 other participants were nonbloodline spouses. All 3 had polyclonal patterns and normal QIg at their initial evaluation. For Patient S1, duplicate IFE assays detected a small monoclonal band of IgM(κ) in one, but not both, aliquots. No abnormal bands were present in his follow-up sera collected 16 and 18 years later, and his IgM level remained stable over time. Patient S2 had a small monoclonal IgM(κ) band in a single specimen. At follow-up evaluations 15 and 19 years later, there was no detectable monoclonal component. Patient S3 had a small monoclonal IgM(λ) band detected using IFE, but she did not have enough archived serum to allow duplicate IFE assays and was not available for follow-up.

Apart from the relative who developed WM, no relative or control patient developed any other hematolymphoid malignancy.


As expected, we found that IFE was more sensitive than IEP7, 10–17 and detected monoclonal bands in 11% of individuals who appeared to have a polyclonal pattern by IEP. IFE has several other advantages over IEP18 (e.g., avoiding the “umbrella effect”19) and is thus particularly well suited for the identification of IgM monoclonal bands present in low concentration.

The expanding use of IFE has led to increasing recognition of small monoclonal bands.20 However, the clinicobiological significance of small bands detected by IFE is uncertain.5, 6 Several earlier studies have addressed the long-term clinical outcome of MGUS and have included a subset of patients with small monoclonal bands.21–26 Although patients with small monoclonal bands had a lower risk of progression to malignancy, their risk was 14% at 20 years,25 and did not appear to plateau.27 Therefore, patients who are younger at diagnosis of MGUS will likely have a higher lifetime risk of developing malignancy. Notably, all bands found in this study were of IgM type. Monoclonal IgM gammopathy is uncommon in the general population, accounting for 5–30% of all monoclonal gammopathies in various populations,23, 28–30 and has an overall estimated prevalence of ∼0.5%.31 Therefore, the exclusive occurrence of IgM MG in these WM families and the finding that all but one of the small bands detected in at-risk family members progressed to measurable levels are intriguing observations. Conversely, all of the bands detected in unrelated individuals resolved over time. This suggests that small bands occurring in the familial WM setting may be more likely to persist and progress than those found in the general population. Further long-term studies are needed to confirm these observations.

We were also intrigued to find that an IgM MG emerged in 2 relatives of WM patients in the setting of a preexisting IgM PG. The question of whether IgM PG is biologically significant when found in relatives of WM patients will be difficult to answer. Polyclonal gammopathy is associated with a wide variety of conditions, many of which are characterized by chronic immune stimulation32 and are hypothesized to contribute to lymphomagenesis in some cases.33, 34 However, very little is known about whether polyclonal gammopathy is a risk factor for subsequent development of a monoclonal gammopathy. In a study of moderate-to-marked polyclonal gammopathy, no cases of WM or multiple myeloma developed.32 In our study, all cases of IgM PG were associated with normal total levels of gamma globulins, and most occurred in patients with a coincident or prior history of a condition (e.g., autoimmune disease) that is sometimes associated with polyclonal gammopathy.32 However, not all family members with similar conditions exhibited polyclonal gammopathy and in at least 1 patient, autoimmune disease was detectable only by serology and appeared after resolution of the PG. While the population frequency of IgM MG has been estimated, the prevalence of IgM PG is largely unknown. In most conditions associated with polyclonal gammopathy, selective elevation of IgM appears to be distinctly uncommon,35 though it occurred in 11% of NHL patients in one study.36 Taken together, the relatively frequent occurrence of an uncommon event (IgM PG) and subsequent development of a rare event (IgM MG) in family members but not spouses suggests that IgM PG may also be a phenotypic biomarker in the familial WM context. However, until larger studies confirm or refute these observations, it remains unknown whether IgM PG in these patients is a manifestation of immune dysregulation related to WM familiality, a discrete clinical condition, both, or neither.

Our data obtained over a 25-year period are compatible with the notion that IgM MG may be a phenotypic marker for genetic susceptibility in WM families. This hypothesis is consistent with our recent observation suggesting that evidence for genetic linkage is strengthened by inclusion of IgM MG in the familial WM phenotype37 and has important implications for future genetic studies. The ability to determine affection status (i.e., whether an individual is affected by the condition of interest and therefore likely to carry the disease-related mutation) is critical for the design of genetic studies aimed at elucidating the genetic determinants of disease. In the absence of a biomarker such as a chromosomal rearrangement, most genetic mapping studies rely on large numbers of families and/or families having many members affected with the cancer of interest. WM is rare, and unlike such syndromes as familial breast cancer, WM families typically have few members who are diagnosed with the disease. Thus, recognition of a phenotypic marker of genetic risk would confer an important advantage because it may permit more accurate assessment of affection status.

The major limitation of this study is small numbers, which affects our ability to make statistical inferences. Should adequate numbers be available for analysis, conditional models would account for familial effects. Unfortunately, given the low frequency of MGUS in the general population, longitudinal studies of cohorts sized to provide adequate power to answer these questions definitively are impractical. Secondly, it would have been preferable to have all IEP assays performed by a single laboratory. However, the primary question of this study concerned the clinicobiologic significance of IFE, which was performed in a single laboratory and reviewed by a single experienced pathologist. Finally, the families studied here may not be representative of all familial WM. Nonetheless, to our knowledge, these data represent the longest systematic longitudinal evaluation of protein electrophoresis and IFE in familial WM. Our results provide important clues that will inform future clinical and genetic studies of familial WM. We plan to continue to obtain follow-up data on these and other families to refine our understanding of the clinical spectrum of this distinctive disorder.


The authors thank Dr. W. Blattner, Dr. G. Shaw, and Dr. J.F. Fraumeni, Jr., for their invaluable and far-sighted contributions to family ascertainment and initial evaluation, and Ms. T. Giambarresi and Ms. L. Vasquez for their untiring work in patient care and data collection and management. We especially thank our patients and their families, whose generosity of spirit and selfless participation over many years have made this research possible. MLM is member of the Commissioned Corps, U.S. Public Health Service, Department of Health and Human Services.