British Journal of Haematology, 2003, 121, 749­757
Criteria for the classification of monoclonal gammopathies,
multiple myeloma and related disorders: a report of the
International Myeloma Working Group
The International Myeloma Working Group*
Received 2 September 2002; accepted for publication 27 November 2002
Summary. The monoclonal gammopathies are a group of
increased calcium, renal insufficiency, anaemia, or bone
disorders associated with monoclonal proliferation of
lesions (CRAB) attributed to the plasma cell proliferative
plasma cells. The characterization of specific entities is an
process. Symptomatic myeloma requires evidence of ROTI.
area of difficulty in clinical practice. The International
Non-secretory myeloma is characterized by the absence of
Myeloma Working Group has reviewed the criteria for
an M-protein in the serum and urine, bone marrow plas-
diagnosis and classification with the aim of producing
macytosis and ROTI. Solitary plasmacytoma of bone,
simple, easily used definitions based on routinely available
extramedullary plasmacytoma and multiple solitary plas-
investigations. In monoclonal gammopathy of undeter-
macytomas (± recurrent) are also defined as distinct enti-
mined significance (MGUS) or monoclonal gammopathy,
ties. The use of these criteria will facilitate comparison of
unattributed/unassociated (MG[u]), the monoclonal protein
therapeutic trial data. Evaluation of currently available
is < 30 g/l and the bone marrow clonal cells < 10% with no
prognostic factors may allow better definition of prognosis
evidence of multiple myeloma, other B-cell proliferative
in multiple myeloma.
disorders or amyloidosis. In asymptomatic (smouldering)
myeloma the M-protein is 30 g/l and/or bone marrow
Keywords: classification, monoclonal, gammopathies, mul-
clonal cells 10% but no related organ or tissue impairment
tiple myeloma.
(ROTI)(end-organ damage), which is typically manifested by
The monoclonal gammopathies (paraproteinaemias) are a
group of disorders associated with monoclonal proliferation
of plasma cells. This group of disorders has also been
referred to as paraproteinaemias, dysproteinaemias or
Both serum and urine should be assessed for monoclonal
immunoglobulinopathies. They are characterized by the
proteins (M-proteins/paraproteins) when there is a clinical
secretion of electrophoretically and immunologically homo-
suspicion of B-cell neoplasia such as multiple myeloma, light-
geneous (monoclonal) proteins. Each monoclonal protein
chain amyloidosis (AL), Waldenstro¨m's macroglobulinaemia
(M-protein, myeloma protein or paraprotein) consists of two
or related B-cell lymphoproliferative disorders. Agarose gel
heavy (H) polypeptide chains of the same class and subclass
electrophoresis is preferred to screen for the presence of
and two light (L) polypeptide chains of the same type. The
M-proteins. Immunofixation is now the gold standard and
heavy polypeptide chains are IgG, IgA, IgM, IgD and IgE
should be performed to confirm the presence of an M-protein
(gamma, alpha, mu, delta, epsilon) while the light chain
and to distinguish its heavy chain and light chain type. It is
types are kappa (j) and lambda (k). The characterization of
essential to differentiate monoclonal gammopathies from
specific entities continues to be an area of potential difficulty
polyclonal gammopathies because monoclonal gammopa-
in clinical practice and there is a need to review classifica-
thies are neoplastic or potentially neoplastic, whereas
tion with a view to improving the definition of these
polyclonal gammopathies (with increase in both types of
light chain) result from an inflammatory or reactive process
such as chronic liver disease, connective tissue disorders,
chronic infections, etc. (Dispenzieri et al, 2001).
Correspondence: Robert A. Kyle, M.D., Mayo Clinic, 200 First Street
Serum protein electrophoresis should be performed when-
SW, Rochester, MN 55905, USA. E-mail:
ever multiple myeloma or related disorders are suspected or in
*Members of the Working Group are listed at the end of the paper.
the presence of unexplained weakness or fatigue, anaemia,
Ó 2003 Mayo Foundation

The International Myeloma Working Group
back pain, osteopenia, osteolytic lesions or spontaneous
The presence of a monoclonal light chain in nephrotic urine
fractures, elevation of the erythrocyte sedimentation rate or
is indicative of AL or light chain deposition disease in almost
plasma viscosity, hypergammaglobulinaemia, hypercalcae-
all instances.
mia, Bence Jones proteinuria, renal insufficiency, immuno-
Light chains may not be detectable in urine because of
globulin deficiency or recurrent infections. An M-protein is
reabsorption by the proximal renal tubules. For this reason,
usually seen as a narrow peak in the densitometer tracing or
variation in glomerular filtration and tubular function,
as a dense, discrete band on the agarose gel.
assay of light chains in serum can provide a more sensitive
Immunofixation should be performed when a peak or
method of detecting and monitoring light chain disease.
band is seen on protein electrophoresis or when multiple
This is particularly relevant in patients with non-secretory
myeloma or related disorders are suspected, despite a
myeloma, solitary plasmacytoma or primary systemic
normal serum protein electrophoretic pattern. Immunofix-
amyloidosis (Drayson et al, 2001; Katzmann et al, 2002).
ation is especially helpful when searching for a small
M-protein in patients with AL, solitary plasmacytoma of
bone, extramedullary plasmacytoma, heavy chain disease,
light chain deposition disease, or after successful treatment
of multiple myeloma or macroglobulinaemia. Immunofixa-
It is a challenge to develop fixed criteria for the diagnosis of
tion will detect a serum M-protein of 0Ć2 g/l and a urine
the monoclonal gammopathies because they often have
M-protein of 0Ć04 g/l. It must be kept in mind that an
overlapping features. If one attempts to cover all clinical and
M-protein may be present when the total protein concen-
laboratory parameters, the definitions are very complex and
tration, beta and gamma globulin levels, and quantitative
lengthy. This will result in definitions that will be difficult to
immunoglobulin values are all within normal limits. A
accept and follow in everyday practice. The International
small M-protein may be concealed in the normal b or c
Myeloma Working Group decided that the criteria should be
areas and easily overlooked. Furthermore, the presence of a
simple, easy to use and based upon routinely available
monoclonal light chain (Bence Jones proteinaemia) requires
laboratory tests rather than attempting to cover all
immunofixation because it is rarely seen in the agarose gel
diagnostic situations.
pattern. In addition, in the heavy chain diseases a discrete
A wide variety of diagnostic criteria have been utilized by
band or spike is often not apparent. IgD and IgE monoclonal
various groups of investigators such as the Medical
proteins are usually small and may be overlooked. Immu-
Research Council of the United Kingdom, Nordic Myeloma
nofixation should also be performed in patients with
Study Group, Spanish (Pethema) Study Group, Eastern
unexplained peripheral neuropathy, carpal tunnel syn-
Cooperative Oncology Group, Southwest Oncology Group
drome, refractory congestive heart failure, nephrotic syn-
and Chronic Leukaemia Myeloma Task Force of the
drome, orthostatic hypotension or malabsorption, because a
National Cancer Institute. A uniform approach would
monoclonal protein strongly suggests the possibility of
facilitate comparison of therapeutic trial data.
primary systemic amyloidosis (AL).
Quantification of the M-protein by densitometry gives
prognostic information in monoclonal gammopathy of
undetermined significance (MGUS) and a baseline for
monitoring disease activity in the malignant causes of
Since Waldenstro¨m's introduction of the term essential
hyperglobulinaemia in 1952, many similar terms have
Urine analysis is important in the evaluation of multiple
been used including benign, idiopathic, asymptomatic,
myeloma and related disorders. Immunofixation of an
non-myelomatous, discrete, cryptogenic, lanthanic and
aliquot from an adequately concentrated 24-h urine speci-
rudimentary monoclonal gammopathy, dysimmunoglobu-
men is recommended. Immunofixation can also be per-
linaemia, asymptomatic paraimmunoglobulinaemia and
formed on the first morning specimen or a random sample.
idiopathic paraproteinaemia. The term benign monoclonal
An M-protein appears as a dense, localized band on the
gammopathy is misleading because at diagnosis it is not
agarose gel or a tall, narrow, homogeneous peak in the
known whether the process will remain stable and asymp-
densitometer tracing. Its amount can be calculated on
tomatic or will develop into symptomatic multiple myeloma,
the basis of the size of the spike and the amount of total
macroglobulinaemia, amyloidosis or a related lympho-
protein in the 24-h specimen. It is not uncommon for a
plasma cell disorder.
patient to have a negative protein reaction and no spike on
The term MGUS denotes the presence of a monoclonal
electrophoresis but for immunofixation of the concentrated
protein in persons without evidence of multiple myeloma,
urine specimen to show a monoclonal light chain. Immuno-
macroglobulinaemia, amyloidosis, or other related plasma
fixation should be performed in all patients with multiple
cell or lymphoproliferative disorders. It was introduced over
myeloma, Waldenstro¨m's macroglobulinaemia, AL, solitary
a quarter of a century ago (Kyle & Bayrd, 1976; Kyle,
plasmacytoma or heavy chain disease. It should also be
1978). MGUS occurs in 3% of persons over 70 years of age
performed in patients with MGUS if the M-spike is greater
and 1% of persons older than 50 years (Axelsson et al,
than 15 g/l (Kyle, 1999). Immunofixation should also be
1966; Kyle et al, 1972; Saleun et al, 1982). The incidence
performed on the urine of every adult patient over 40 years
increases with advancing age and is higher in African-
of age who develops an unexplained nephrotic syndrome.
Americans than in Caucasians (Cohen et al, 1998).
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

Classification of Monoclonal Gammopathies
The original Mayo Clinic series of 241 MGUS patients,
twofold greater than the risk of progression with a value of
diagnosed in 1970 or earlier, was followed for 24­38 years.
5 g/l, while the risk of progression with an M-protein value
At the time of analysis, 10% were alive with a stable
of 25 g/l was 4Ć6 times that of the reference value of 5 g/l.
M-protein while more than half had died of unrelated causes
Uninvolved (normal or background) immunoglobulins were
without developing multiple myeloma or a related disorder.
reduced in 38% of cases, but this reduction did not identify
Approximately one-quarter of the patients developed mul-
patients in whom progression developed. In addition, the
tiple myeloma, macroglobulinaemia, amyloidosis or related
presence of small amounts of monoclonal light chain in the
lymphoproliferative disorders, with an actuarial rate of 16%
urine did not predict progression (Kyle et al, 2002).
at 10 years, 33% at 20 years and 40% at 25 years. The
Considerable debate occurred about the designation of a
interval from the time of recognition of the M-protein to the
specific serum M-protein level. Many patients with sympto-
diagnosis of serious disease ranged from two to 29 years
matic multiple myeloma will have a serum M-protein of less
(median, 10 years). In seven patients multiple myeloma
than 30 g/l. The size of the M-protein is not a prognostic
was diagnosed more than 20 years after detection of the
factor in multiple myeloma. However, most clinicians think
serum M-protein (Kyle, 1993).
of a serum M-protein of 30 g/l as a watershed number and
During a 20-year follow-up of 64 Swedish patients from a
one in which they look closely for the possibility of multiple
survey of 6995 persons, Axelsson (1986) reported that
myeloma. An M-spike of greater than 30 g/l usually
three patients had died of multiple myeloma or lymphoma
indicates the presence of multiple myeloma. While an IgA
while four of the 19 surviving patients had shown an
value of 20­25 g/l might be equivalent to an IgG M-protein
increase in the serum M-protein or had developed Bence
of 30 g/l, the use of a single value of 30 g/l, although
Jones proteinuria. In another group of 128 patients with
arbitrary, would make for simplicity and be more likely to be
MGUS, followed for 12­156 months (median 56 months),
adopted in practice.
13 developed malignant disease. The actuarial probability of
The presence of a monoclonal light chain in the urine is
developing malignant disease was 8Ć5% at 5 years and
generally considered as a finding that suggests the diagnosis
19Ć2% at 10 years (Blade´ et al, 1992).
of multiple myeloma. However, almost one-third of patients
Isaksson et al (1996) reported that 15 (26%) of 57
with well-documented MGUS had a small amount of
patients developed a malignant plasma cell process during
monoclonal light chain in the urine, which is not a risk
follow-up. In another study, 6Ć8% of 335 patients with
factor for progression to multiple myeloma (Kyle et al,
MGUS showed progression during a median follow-up of
2002). Some patients with a large amount of Bence Jones
70 months (Baldini et al, 1996). In a follow-up of 263 cases
proteinuria (> 1 g/24 h) may follow a benign course.
of MGUS, the actuarial probability of malignant transfor-
Although these patients with idiopathic Bence Jones
mation was 31% at 20 years (Pasqualetti et al, 1997).
proteinuria are at high risk for the development of multiple
Cesana et al (2002) reported that in 5Ć8% of 1104 patients,
myeloma or primary systemic amyloidosis, they may
MGUS evolved to multiple myeloma or related disorders.
remain stable for many years and should be observed
In a recent series, 1384 MGUS patients, residing in south-
indefinitely (Kyle & Greipp, 1982).
eastern Minnesota, were diagnosed at the Mayo Clinic
While the level of M-protein is a major risk factor for
between 1960 and 1994. During the 11 009 person years of
progression to multiple myeloma, consideration also needs
follow-up, 115 of the 1384 patients with MGUS showed
to be given to changes in the plasma cells even though a
progression to multiple myeloma or a related lymphoplasma
bone marrow examination is frequently not performed for
cell disorder. The relative risk (RR), when compared to the
small M-proteins. Approximately 5% of patients with symp-
white population of the Surveillance, Epidemiology and End-
tomatic myeloma will have a plasma cell content < 10%, but
Results Program (SEER), was as follows: multiple myeloma
this is usually due to an inadequate specimen or to the
(RR 25), IgM lymphoma (RR 2Ć4), primary amyloidosis (RR
possibility of uneven distribution of plasma cells in the bone
8Ć4), macroglobulinaemia (RR 46), chronic lymphocytic
marrow. Patients with a plasma cell content > 10% may be
leukaemia (RR 0Ć9) and plasmacytoma (RR 8Ć5). The overall
asymptomatic but they are likely to develop symptomatic
relative risk for progression was 7Ć3 times the expected rate.
multiple myeloma and require treatment. Plasma cells in
In 32 additional patients the M-protein level increased to
myeloma are phenotypically distinct from their normal
more than 30 g/l or the percentage of plasma cells in the
counterparts by virtue of the absence of CD19 expression or
bone marrow increased to more than 10%, but they have
expression of CD56. Using flow cytometry, phenotypically
not progressed to overt myeloma or a related disorder. The
`neoplastic' plasma cells are demonstrable with phenotypi-
cumulative probability of progression was 12%, 25% and
cally normal plasma cells (Ocqueteau et al, 1998; Rawstron
30% at 10, 20 and 25 years respectively. The risk of
et al, 2000). In the future, the monitoring of such changes
progression of MGUS to multiple myeloma or a related
may provide additional information with regard to progres-
disorder was about 1% per year. Patients with IgM and IgA
sion. In practice it is usual to use an arbitrary figure of less
M-proteins showed an increased risk of progression com-
than 10% plasma cells in the bone marrow aspirate and
pared with patients who had IgG M-proteins (principally to
minimal infiltration in a trephine biopsy.
lymphoproliferative diseases other than multiple myeloma in
The term MGUS has been widely adopted but, for
IgM MGUS). However, concentration of serum M-protein
patients, `undetermined significance' can be difficult to
was the most important risk factor for progression. Risk of
understand and accept. An alternative would be to use
progression with an M-protein value of 15 g/l was almost
the term primary monoclonal gammopathy, although this
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

The International Myeloma Working Group
Table I. Monoclonal gammopathy of undetermined significance
anaemia is present at the time of diagnosis in two-thirds
(MGUS) or monoclonal gammopathy, unattributed/unassociated
of patients with multiple myeloma but eventually occurs in
almost all patients. Anaemia is considered to be present
when the haemoglobin is 2 g/dl below the normal level for
M-protein in serum < 30 g/l
the laboratory or if the haemoglobin falls to 10 g/dl.
Bone marrow clonal plasma cells < 10% and low level of
Hypercalcaemia, which is present in 15­20% of myeloma
plasma cell infiltration in a trephine biopsy (if done)
patients at presentation, is a major but treatable cause of
No evidence of other B-cell proliferative disorders
renal insufficiency. A serum calcium level > 0Ć25 mmol/l
*No related organ or tissue impairment (no end organ damage,
above normal or a value of 2Ć75 mmol/l is considered as
including bone lesions)
representing hypercalcaemia. The serum creatinine is
173 mmol/l or more in one-fifth of multiple myeloma
*AL amyloid and the IgM paraprotein­related neurological syn-
patients at diagnosis. Conventional radiographs show
dromes would be instances of `MG associated with...'
abnormalities consisting of lytic lesions, osteoporosis or
fractures in nearly 80% of patients with multiple myeloma
would be more likely to convey the impression of a disease
at diagnosis. The vertebrae, skull, thoracic cage, pelvis, and
entity and also requires the introduction of `secondary
humeri and femori are the most frequent sites of involve-
monoclonal gammopathy'. Another approach would be to
ment and should be included in the films. Technetium-99m
simply delineate monoclonal gammopathy as a biochemical
bone scanning is inferior to conventional radiography
finding and qualify it as unattributed to or unassociated
because it detects only bone formation and should not be
with any disease entity, i.e. monoclonal gammopathy,
used. Computerized tomography (CT) or magnetic reson-
unattributed/unassociated or MG(u). There is a significant
ance imaging (MRI) is helpful in patients who have
risk of transformation to multiple myeloma or a related
compression fractures or skeletal pain without abnormalit-
disorder and consequently such patients should be monit-
ies on the radiographs or a neurological deficit from spinal
ored throughout their lives for evidence of progression.
cord compression. The MRI, if performed, shows no abnor-
The criteria recommended by the International Myeloma
malities in patients with MGUS (Bellai¨che et al, 1997).
Working Group for the diagnosis of MGUS are shown in
Additional evidence of end organ damage consists of
Table I. Obviously, patients must have no related organ or
symptomatic hyperviscosity, primary systemic amyloidosis or
tissue impairment (end organ damage including bone
recurrent bacterial infections (> 2 episodes in 12 months).
lesions) in order to be classified as MGUS. Because of the
Serum viscosity levels do not correlate well with the patient's
risk of developing a disease requiring therapy, these
symptoms or clinical findings. Examination of the fundus
individuals should be monitored at periodic intervals
reveals dilatation of the veins and haemorrhages. These
indefinitely for evidence of progression. The duration of
findings are more important in evaluating the patient than
follow-up for assigning the diagnosis is not, however,
the actual viscosity level. Primary systemic amyloidosis (AL)
included in the definition because patients with MGUS are
occurs in almost 10% of patients with multiple myeloma and
at risk indefinitely, although with a low risk plasma cell
is an indication for therapy, although the association of only
proliferative process.
an M-protein with AL amyloid occurs, requiring patient
management to be considered individually. The neurological
syndromes occurring in association with isolated IgM
monoclonal gammopathy also require separate considera-
tion. Susceptibility to recurrent bacterial infections may
The effect of end organ damage related to the plasma cell
result from impairment of antibody response due to the
proliferative process is a critical aspect in the diagnosis of
reduction of normal uninvolved immunoglobulins and
these disorders (Table II). A normocytic normochromic
neutropenia or a combination of these factors.
The evidence of myeloma-related organ or tissue damage
or functional impairment may frequently not be clear cut
and the diagnosis of transition from asymptomatic myeloma
Table II. Myeloma-related organ or tissue impairment (end organ
to multiple myeloma requiring treatment (symptomatic
damage) (ROTI) due to the plasma cell proliferative process.
myeloma) may need to be the subject of multidisciplinary
critical assessment.
*Calcium levels increased: serum calcium > 0Ć25 mmol/l above
the upper limit of normal or > 2Ć75 mmol/l
*Renal insufficiency: creatinine > 173 mmol/l
*Anaemia: haemoglobin 2 g/dl below the lower limit of normal
or haemoglobin < 10 g/dl
*Bone lesions: lytic lesions or osteoporosis with compression
The point of transition from MGUS to myeloma without
fractures (MRI or CT may clarify)
anaemia, renal insufficiency, or skeletal lesions attributable
Other: symptomatic hyperviscosity, amyloidosis, recurrent
to the neoplastic plasma cell proliferation is not sharply
bacterial infections (> 2 episodes in 12 months)
defined. If the M-protein level is 30 g/l and/or the bone
marrow contains 10% plasma cells, it would be reasonable
*CRAB (calcium, renal insufficiency, anaemia or bone lesions).
to apply the term asymptomatic myeloma (smouldering
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

Classification of Monoclonal Gammopathies
Table III. Asymptomatic myeloma (smouldering myeloma).
Table IV. Laboratory tests for multiple myeloma.
M-protein in serum 30 g/l
History and physical examination
Complete blood count and differential peripheral blood smear
Bone marrow clonal plasma cells 10%
Chemistry screen including calcium and creatinine
No related organ or tissue impairment (no end organ damage,
Serum protein electrophoresis, immunofixation
including bone lesions) or symptoms
Nephelometric quantification of immunoglobulins
Routine urinalysis, 24 h urine collection for electrophoresis and
Bone marrow aspirate and trephine biopsy (cytogenetics,
myeloma). These patients may have small amounts of
immunophenotyping and plasma cell labelling index if available)
M-protein in the urine and decreased concentration of
Radiological skeletal bone survey including spine, pelvis, skull,
normal immunoglobulins in the serum, but the plasma cell
humeri and femurs. An MRI may be helpful
labelling index is low (Kyle & Greipp, 1980). These patients
b2 microglobulin, C-reactive protein, and lactate dehydrogenase
must have no evidence of related organ or tissue impair-
Measurement of free monoclonal light chains if available
ment (end organ damage). Biologically, patients with
asymptomatic myeloma are similar to MGUS but it is
difficult to accept that diagnosis when the serum M-protein
Table V. Symptomatic multiple myeloma.
level is 30 g/l, or the bone marrow contains 10% plasma
cells. These patients must be followed up closely because
symptomatic myeloma develops in many of them. They
M-protein in serum and/or urine
Bone marrow (clonal) plasma cells* or plasmacytoma
should not be treated unless progression occurs. Although
Related organ or tissue impairment (end organ damage, including
the term `smouldering myeloma' has been frequently used
bone lesions)
to describe this state, the term can be difficult for patients to
understand and accept. The use of asymptomatic to describe
a stage of disease of which there are no symptoms and no
*If flow cytometry is performed, most plasma cells (> 90%) will
show a `neoplastic' phenotype.
related organ or tissue impairment is preferred. Patients
Some patients may have no symptoms but have related organ or
categorized as having Durie­Salmon stage I disease would
tissue impairment.
be included in this category, as would asymptomatic
patients with an apparently solitary plasmacytoma of bone
who have abnormalities detected by an MRI. The criteria
agreed upon by the International Myeloma Working Group
account for 10% or more of all nucleated cells but they
are presented in Table III.
may range from less than 5% to almost 100%. Bone
marrow involvement may be focal rather than diffuse,
requiring repeated bone marrow examinations for diagno-
sis. Identification of a monoclonal immunoglobulin in the
Multiple myeloma (Kahler's disease, myelomatosis) is char-
cytoplasm of plasma cells by immunoperoxidase staining or
acterized by the neoplastic proliferation of a single clone of
immunofluorescence is helpful for differentiating monoclo-
plasma cells producing an M-protein. The clone of plasma
nal plasma cell proliferation from reactive plasmacytosis due
cells proliferates in the bone marrow and frequently invades
to connective tissue disorders, chronic liver disease, chronic
the adjacent bone, producing skeletal destruction that
infections or metastatic carcinoma. Conventional radio-
results in bone pain and pathological fractures. Anaemia,
graphs reveal abnormalities consisting of lytic lesions,
hypercalcaemia and renal insufficiency are other concom-
osteoporosis or fractures in nearly 80% of patients at
itant features. Bone pain, particularly in the spine or chest
diagnosis. The suggested laboratory tests for the diagnosis of
and less often in the extremities, is present at the time of
multiple myeloma are shown in Table IV.
diagnosis in two-thirds of patients. Weakness and fatigue
The criteria agreed upon by the International Myeloma
are common and are often due to anaemia. Anaemia is
Working Group for the diagnosis of symptomatic multiple
present initially in more than two-thirds of patients, while
myeloma are shown in Table V. No level of serum M-protein
renal insufficiency (creatinine greater than 173 mmol/l) is
or urine M-protein was included in the proposed diagnostic
present in one-fifth of patients at diagnosis. A serum M-spike
criteria. Approximately 40% of patients with symptomatic
or peak is seen in 80% of patients at diagnosis, while
multiple myeloma have an M-protein less than 30 g/l.
immunofixation reveals an M-protein in over 90%. An IgG
However, 97% of patients with multiple myeloma will have
M-protein is found in about one-half of patients, while one-
an M-protein in the serum or in the urine. No minimal level of
fifth have an IgA M-protein, and monoclonal light chain
clonal bone marrow plasma cells was designated because 5%
only (light chain myeloma) is found in almost 20% of cases.
of patients with symptomatic myeloma have fewer than 10%
The urine contains an M-protein in approximately 75% of
plasma cells in the bone marrow. The most critical criterion
patients. Ninety-seven per cent of patients with multiple
for symptomatic or treatable disease is the evidence of organ
myeloma have an M-protein in the serum or urine at the
or tissue impairment (end organ damage) manifested by
time of diagnosis. Hypercalcaemia is present in nearly 20%
anaemia, hypercalcaemia, lytic bone lesions, renal insuffi-
of patients initially and is a major but treatable cause of
ciency, hyperviscosity, amyloidosis or recurrent infections.
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

The International Myeloma Working Group
Definitions of multiple myeloma and criteria for treatment
Table VII. Solitary plasmacytoma of bone.
adopted by study groups in various countries have differed.
Agreement to adopt a uniform approach would be advan-
No M-protein in serum and/or urine*
tageous in collating data internationally and carrying out
Single area of bone destruction due to clonal plasma cells
treatment overviews and meta-analyses. An agreed-upon
Bone marrow not consistent with multiple myeloma
definition of symptomatic multiple myeloma requiring
Normal skeletal survey (and MRI of spine and pelvis if done)
treatment would also remove the need for the use of older
No related organ or tissue impairment (no end organ damage
other than solitary bone lesion)*
staging systems. Inevitably, the newer prognostic indicators
such as cytogenetic changes will result in more accurate
stratification of risk groups. The International Myeloma
*A small M-component may sometimes be present.
Working Group has now initiated a major investigation of
prognostic features with a view to producing an interna-
may be the presenting feature; a pathological fracture may
Response criteria have been adopted by the International
be the first symptom. Soft tissue extension of a plasmacy-
Myeloma Working Group (Blade´ et al, 1998). These criteria
toma, as in a rib, may result in a palpable mass. The axial
are currently undergoing review and some modification.
skeleton is more commonly involved than the appendicular
skeleton. Thoracic vertebrae are more commonly involved
than lumbar or cervical vertebrae. Involvement of the distal
appendicular skeleton below the knees or elbows is rare. The
Patients with non-secretory myeloma have no monoclonal
diagnosis of solitary plasmacytoma of bone is based on
protein in either the serum or the urine with immunofix-
histological evidence of a tumour consisting of monoclonal
ation (Table VI) (Cavo et al, 1985). This occurs in only 3%
plasma cells identical to those seen in multiple myeloma
of patients with symptomatic multiple myeloma. The
(Table VII). In addition, complete skeletal radiographs must
monoclonal protein should be identified in the plasma cells
show no other lesions of multiple myeloma, but the bone
by immunoperoxidase or immunofluorescence. More than a
marrow aspirate may contain a few plasma cells. A MRI of
dozen patients in whom no M-protein was found in the
the spine and pelvis may show unsuspected and asympto-
plasma cells have been described. These patients apparently
matic skeletal lesions. This finding would place the patient
do not synthesize an M-protein. Renal insufficiency is less
in the asymptomatic or smouldering myeloma category.
common than in patients with secretory multiple myeloma.
The presence of marrow involvement on the MRI is
The more carefully the serum and urine are examined for
associated with a higher rate of relapse. In a group of 23
evidence of an M-protein, the fewer cases of non-secretory
patients with solitary plasmacytoma of the thoracolumbar
myeloma will be found. The advent of accurate light-chain
spine, multiple myeloma developed in seven of eight patients
assays is likely to result in further definition of diagnosis; in
with a solitary lesion on plain roentgenographs alone but in
fact, two-thirds of patients with non-secretory multiple
only one of seven patients who also had negative results on
myeloma based upon immunofixation had an elevation of
MRI (Liebross et al, 1998). Typically, immunofixation of the
free monoclonal light chains in the serum (Drayson et al,
serum and concentrated urine should show no M-protein,
2001). Treatment for non-secretory myeloma is the same as
but approximately 50% of patients do have a small
for multiple myeloma. Response to therapy and survival of
M-protein in the serum or urine. Most patients with solitary
patients with non-secretory myeloma are similar to those in
plasmacytoma of bone have normal uninvolved immuno-
patients with a serum or urinary M-component.
globulin levels (Dimopoulos et al, 2000). There should be no
evidence of anaemia, hypercalcaemia or renal insufficiency
related to the plasmacytoma. Treatment consists of radi-
ation in the range of 40 Gy to 50 Gy. Radiotherapy of the
Solitary plasmacytoma of bone is uncommon and occurs in
solitary lesion usually results in disappearance of the
3­5% of patients with plasma cell neoplasms. It occurs more
M-protein. However, some patients may remain stable for
commonly in men than in women (65% vs. 35%) and the
long periods despite the persistence of an M-protein after
median age is about a decade younger than that of patients
tumoricidal radiation. The persistence of an M-protein
with multiple myeloma (55 vs. 65 years). The most
following tumoricidal radiation to an apparent solitary
common symptom at presentation is pain at the site of the
plasmacytoma is associated with an increased risk of
skeletal lesion. Severe back pain or spinal cord compression
progression (Wilder et al, 2002). Solitary plasmacytomas
> 5 cm have a greater incidence of progression (Tsang et al,
2001). There is no convincing evidence that prophylactic
chemotherapy affects the incidence of conversion to mul-
Table VI. Non-secretory myeloma.
tiple myeloma. Progression, when it occurs, usually does so
within 3­4 years, but the most uncertain criterion for
No M-protein in serum and/or urine with immunofixation
diagnosis is the duration of observation necessary before
Bone marrow clonal plasmacytosis 10% or plasmacytoma
Related organ or tissue impairment (end organ damage, including
deciding that the disease will not become generalized.
bone lesions)
Almost 50% of patients with solitary plasmacytoma are
alive at 10 years; 25­40% are surviving disease-free at 10
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

Classification of Monoclonal Gammopathies
years. Overt multiple myeloma occurs in almost 50% of
5% of patients with an apparently solitary plasmacytoma
patients with solitary plasmacytoma of bone. However,
(Table IX). Lesions may be in soft tissue (extramedullary)
progression may occur as long as 15 years later. Multiple
or bone (osseous). There is no evidence of bone marrow
solitary plasmacytomas without evidence of multiple myel-
involvement or other skeletal lesions. Multiple plasma-
oma occur in up to 5% of patients.
cytomas may be treated by tumoricidal radiation when
recurrent if there is no evidence of multiple myeloma.
Large numbers of solitary plasmacytomas or recurrent
lesions at short intervals are an indication for systemic
Extramedullary plasmacytoma is a plasma cell tumour that
therapy such as autologous stem cell transplantation.
arises outside the bone marrow (Table VIII). The upper
respiratory tract, including the nasal cavity and sinuses,
nasopharynx and larynx, is the most frequent location of
lesions. Epistaxis, rhinorrhoea and nasal obstruction are the
Plasma cell leukaemia has been previously delineated by
most frequent symptoms. Extramedullary plasmacytomas
the finding of a peripheral blood absolute plasma cell count
may also occur in virtually any organ including the
of at least 2Ć0 · 109/l and more than 20% plasma cells in
gastrointestinal tract, central nervous system, urinary
the peripheral blood differential white cell count. Plasma
bladder, thyroid, breasts, testes, parotid gland or lymph
cell leukaemia may be classified as primary when it
nodes. There is a predominance of IgA monoclonal protein.
presents in the leukaemic phase or as secondary when
The diagnosis is made on the basis of finding a monoclonal
there is leukaemic transformation of a previously recog-
plasma cell tumour in an extramedullary site and the
nized multiple myeloma. Approximately 60% of patients
absence of multiple myeloma on the basis of bone marrow,
with plasma cell leukaemia have the primary type.
radiography, and appropriate studies of blood and urine.
Patients with primary plasma cell leukaemia are younger
Treatment consists of tumoricidal radiation (40­50 Gy) and
and have a higher incidence of hepatosplenomegaly and
is often curative. The plasmacytoma may occur locally or
lymphadenopathy, a higher platelet count, fewer lytic bone
metastasize to regional nodes. In contrast to solitary
lesions, a smaller serum M-protein level, and longer
plasmacytoma of bone, symptomatic multiple myeloma
survival than do patients with secondary plasma cell
occurs in only 15% of patients (Alexiou et al, 1999).
leukaemia (Noel & Kyle, 1987; Garcia-Sanz et al, 1999). If
looked for, plasma cells may frequently be detected in the
peripheral blood. An alternative would be to use the term
multiple myeloma with peripheral blood involvement,
recognizing that appreciable numbers of circulating plasma
Multiple solitary plasmacytomas, which may be recurrent
cells would be regarded as a poor prognostic feature,
and without evidence of multiple myeloma, occur in up to
particularly when seen later in the course of disease.
Table VIII. Extramedullary plasmacytoma.
This review has considered the laboratory investigatory
No M-protein in serum and/or urine*
and clinical features that are used to characterize the
Extramedullary tumour of clonal plasma cells
monoclonal gammopathies. Criteria to define individual
Normal bone marrow
disorders are suggested with specific tabulated recommen-
Normal skeletal survey
dations for the following: monoclonal gammopathy of
No related organ or tissue impairment (end organ damage
undetermined significance (MGUS) for which the alternat-
including bone lesions)
ive term monoclonal gammopathy, unassociated/unattrib-
uted (MG[u]) is suggested; asymptomatic myeloma (to
*A small M-component may sometimes be present.
replace `smouldering myeloma'); symptomatic multiple
myeloma; non-secretary myeloma; solitary plasmacytoma
of bone; extramedullary plasmacytoma; and multiple
solitary plasmacytomas (± recurrent). Additional systems
Table IX. Multiple solitary plasmacytomas (± recurrent).
of staging are not necessarily required but it is anticipated
that prognostic groupings may be better defined as a result
No M-protein in serum and/or urine*
of the evaluation of prognostic factors that is currently
More than one localized area of bone destruction or
extramedullary tumour of clonal plasma cells which
may be recurrent
Normal bone marrow
Normal skeletal survey and MRI of spine and pelvis if done
No related organ or tissue impairment (no end organ damage
This study was supported by the International Myeloma
other than the localized bone lesions)
Foundation, which sponsored the three meetings (December
2000, May 2001 and December 2001) of the International
*A small M-component may sometimes be present.
Myeloma Working Group.
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

The International Myeloma Working Group
Group (1997) Magnetic resonance appearance of monoclonal
gammopathies of unknown significance and multiple myeloma.
Robert A. Kyle, Mayo Clinic
Spine, 22, 2551­2557.
J. Anthony Child, Leeds Teaching Hospitals
Blade´, J., Lopez-Guillermo, A., Rozman, C., Cervantes, F., Salgado,
Kenneth Anderson, Dana Farber Cancer Center
C., Aguilar, J.L., Vives-Corrons, J.L. & Montserrat, E. (1992)
Bart Barlogie, University of Arkansas
Malignant transformation and life expectancy in monoclonal
Regis Bataille, University of Nantes
gammopathy of undetermined significance. British Journal of
William Bensinger, Fred Hutchinson Cancer Research
Haematology, 81, 391­394.
Blade´, J., Samson, D., Reece, D., Apperley, J., Bjorkstrand, B.,
Gahrton, G., Gertz, M., Giralt, S., Jagannath, S. & Vesole, D.
Joan Blade´, IDIBAPS, Hospital Clinic
(1998) Criteria for evaluating disease response and progression
Mario Boccadoro, University of Torino
in patients with multiple myeloma treated by high-dose therapy
William Dalton, University of South Florida
and haemopoietic stem cell transplantation. Myeloma sub-
Meletios Dimopoulos, University of Athens
committee of the EBMT. European Group for Blood and Marrow
Benjamin Djulbegovic, University of South Florida
Transplant. British Journal of Haematology, 102, 1115­1123.
Mark Drayson, University of Birmingham (UK)
Cavo, M., Galieni, P., Gobbi, M., Baldrati, L., Leardini, L., Baccarani,
Brian Durie, Cedars-Sinai Cancer Center
M. & Tura, S. (1985) Nonsecretory multiple myeloma. Presenting
Thiery Facon, University of Lille
findings, clinical course and prognosis. Acta Haematologica, 74,
Rafael Fonseca, Mayo Clinic
Gosta Gahrton, Karolinska Institute
Cesana, C., Klersy, C., Barbarano, L., Nosari, A.M., Crugnola, M.,
Pungolino, E., Gargantini, L., Granata, S., Valentini, M. & Morra, E.
Philip Greipp, Mayo Clinic
(2002) Prognostic factors for malignant transformation in mono-
Jean Luc Harousseau, University of Nantes
clonal gammopathy of undetermined significance and smoldering
David Harrington, Harvard University
multiple myeloma. Journal of Clinical Oncology, 20, 1625­1634.
Mohamad Hussein, Cleveland Clinic
Cohen, H.J., Crawford, J., Rao, M.K., Pieper, C.F. & Currie, M.S.
Douglas Joshua, University of South Wales
(1998) Racial differences in the prevalence of monoclonal gam-
Heinz Ludwig, Wilhelmenia Hospital
mopathy in a community-based sample of the elderly. American
Gareth Morgan, Leeds Teaching Hospitals
Journal of Medicine, 104, 439­444.
Martin Oken, Virginia Piper Cancer Center
Dimopoulos, M.A., Moulopoulos, L.A., Maniatis, A. & Alexanian, R.
Raymond Powles, Royal Marsden Cancer Hospital
(2000) Solitary plasmacytoma of bone and asymptomatic mul-
Paul Richardson, Dana Farber Cancer Center
tiple myeloma. Blood, 96, 2037­2044.
Dispenzieri, A., Gertz, M.A., Therneau, T.M. & Kyle, R.A. (2001)
David Roodman, University of Pittsburgh
Retrospective cohort study of 148 patients with polyclonal
Jesus San Miguel, University of Salamanca
gammopathy. Mayo Clinic Proceedings, 76, 476­487.
Kazuyuki Shimizu, University of Nagoya
Drayson, M., Tang, L.X., Drew, R., Mead, G.P., Carr-Smith, H. &
Chaim Shustik, McGill University
Bradwell, A.R. (2001) Serum free light-chain measurements for
Bhawna Sirohi, Royal Marsden Cancer Hospital
identifying and monitoring patients with nonsecretory multiple
Pieter Sonneveld, University of Rotterdam
myeloma. Blood, 97, 2900­2902.
Guido Tricot, University of Arkansas
Garcia-Sanz, R., Orfa~o, A., Gonzalez, M., Tabernero, M.D., Blade´, J.,
Ingemar Turesson, Malmo General Hospital
Moro, M.J., Fernandez-Calvo, J., Sanz, M.A., Perez-Simon, J.A.,
Brian Van Ness, University of Minnesota
Rasillo, A. & Miguel, J.F. (1999) Primary plasma cell leukemia:
David Vesole, Medical College of Wisconsin
clinical, immunophenotypic, DNA ploidy, and cytogenetic char-
acteristics. Blood, 93, 1032­1037.
Donna Weber, M.D. Anderson
Isaksson, E., Bjorkholm, M., Holm, G., Johansson, B., Nilsson, B.,
Jan Westin, University of Lund
Mellstedt, H. & Osterborg, A. (1996) Blood clonal B-cell excess in
Keith Wheatley, University of Birmingham
patients with monoclonal gammopathy of undetermined sig-
nificance (MGUS): association with malignant transformation.
British Journal of Haematology, 92, 71­76.
Katzmann, J.A., Clark, R.J., Abraham, R.S., Bryant, S., Lymp, J.F.,
Alexiou, C., Kau, R.J., Dietzfelbinger, H., Kremer, M., Spiess, J.C.,
Bradwell, A.R. & Kyle, R.A. (2002) Serum reference and diag-
Schratzenstaller, B. & Arnold, W. (1999) Extramedullary plas-
nostic ranges for free kappa and free lambda immunoglobulin
macytoma: tumor occurrence and therapeutic concepts. Cancer,
light chains: relative sensitivity for detection of monoclonal light
85, 2305­2314.
chain. Clinical Chemistry, 48, 1437­1444.
Axelsson, U. (1986) A 20-year follow-up study of 64 subjects with
Kyle, R.A. (1978) Monoclonal gammopathy of undetermined
M-components. Acta Medica Scandinavica, 219, 519­522.
significance. Natural history in 241 cases. American Journal of
Axelsson, U., Bachmann, R. & Ha¨lle´n, J. (1966) Frequency of
Medicine, 64, 814­826.
pathological proteins (M-components) in 6,995 sera from an
Kyle, R.A. (1993) `Benign' monoclonal gammopathy ­ after 20­35
adult population. Acta Medica Scandinavica, 179, 235­247.
years of follow-up. Mayo Clinic Proceedings, 68, 26­36.
Baldini, L., Guffanti, A., Cesana, B.M., Colombi, M., Chiorboli, O. &
Kyle, R.A. (1999) Sequence of testing for monoclonal gammopa-
Damilano, I. (1996) Role of different hematologic variables in
thies. Serum and urine assays. Archives of Pathology and Labor-
defining the risk of malignant transformation in monoclonal
atory Medicine, 123, 114­118.
gammopathy. Blood, 87, 912­918.
Kyle, R.A. & Bayrd, E.D. (1976) The Monoclonal Gammopathies:
Bellai¨che, L., Laredo, H.-D., Loite´, F., Koeger, A.-C., Hamze, B., Ziza,
Multiple Myeloma and Related Plasma-Cell Disorders. C. C. Thomas,
J.-M., Pertuiset, E., Bardin, T. & Tubiana, J.-M. for the GRI Study
Springfield, IL.
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757

Classification of Monoclonal Gammopathies
Kyle, R.A. & Greipp, P.R. (1980) Smoldering multiple myeloma.
multiple myeloma. American Journal of Pathology, 152, 1655­
New England Journal of Medicine, 302, 1347­1349.
Kyle, R.A. & Greipp, P.R. (1982) `Idiopathic' Bence Jones pro-
Pasqualetti, P., Festuccia, V., Collacciani, A. & Casale, R. (1997)
teinuria: long-term follow-up in seven patients. New England
The natural history of monoclonal gammopathy of undetermined
Journal of Medicine, 306, 564­567.
significance. A 5- to 20-year follow-up of 263 cases. Acta Hae-
Kyle, R.A., Finkelstein, S., Elveback, L.R. & Kurland, L.T. (1972)
matologica, 97, 174­179.
Incidence of monoclonal proteins in a Minnesota community
Rawstron, A.C., Fenton, J.A., Ashcroft, J., English, A., Jones, R.A.,
with a cluster of multiple myeloma. Blood, 40, 719­724.
Richards, S.J., Pratt, G., Owen, R., Davies, F.E., Child, J.A., Jack,
Kyle, R.A., Therneau, T.M., Rajkumar, S.V., Offord, J.R., Larson,
A.S. & Morgan, G.J. (2000) The interleukin-6 receptor alpha-
D.R., Plevak, M.F. & Melton, III, L.J. (2002) A long-term study of
chain (CD126) is expressed by neoplastic but not normal plasma
prognosis in monoclonal gammopathy of undetermined sig-
cells. Blood, 96, 3880­3886.
nificance. New England Journal of Medicine, 346, 564­569.
Saleun, J.P., Vicariot, M., Deroff, P. & Morin, J.F. (1982) Monoclonal
Liebross, R.H., Ha, C.S., Cox, J.D., Weber, D., Delasalle, K. & Alex-
gammopathies in the adult population of Finiste`re, France.
anian, R. (1998) Solitary bone plasmacytoma: outcome and
Journal of Clinical Pathology, 35, 63­68.
prognostic factors following radiotherapy. International Journal of
Tsang, R.W., Gospodarowicz, M.K., Pintilie, M., Bezjak, A., Wells,
Radiation Oncology Biology Physics, 41, 1063­1067.
W., Hodgson, D.C. & Stewart, A.K. (2001) Solitary plasmacytoma
Noel, P. & Kyle, R.A. (1987) Plasma cell leukemia: an evaluation of
treated with radiotherapy: Impact of tumor size on outcome.
response to therapy. American Journal of Medicine, 83, 1062­1068.
International Journal of Radiation Oncology Biology Physics, 50,
Ocqueteau, M., Orfao, A., Almeida, J., Blade´, J., Gonzalez, M.,
Garcia-Sanz, R., Lopez-Berges, C., Moro, M.J., Hernandez, J.,
Wilder, R.B., Ha, C.S., Cox, J.D., Weber, D., Delasalle, K. &
Escribano, L., Caballero, D., Rozman, M. & San Miguel, J.F.
Alexanian, R. (2002) Persistence of myeloma protein for more
(1998) Immunophenotypic characterization of plasma cells from
than one year after radiotherapy is an adverse prognostic
monoclonal gammopathy of undetermined significance patients.
Implications for the differential diagnosis between MGUS and
Ó 2003 Mayo Foundation, British Journal of Haematology 121: 749­757