Immunopathology / MONOCLONAL BANDING IN HIV PATIENTS

Protein Electrophoresis and Immunoglobulin Analysis

in HIV-Infected Patients
Panagiotis A. Konstantinopoulos, MD, PhD,1 Bruce J. Dezube, MD,1 Liron Pantanowitz, MD,2
Gary L. Horowitz, MD,3 and Bruce A. Beckwith, MD3
Key Words: HIV; Protein electrophoresis; Monoclonal banding; Immunoglobulins
DOI: 10.1309/QWTQFGA9FXN02YME

Abstract
We studied the prevalence and nature of
immunoglobulin abnormalities in HIV-1infected
patients in the era of highly active antiretroviral
therapy. Protein electrophoreses (PEP) were performed
on and quantitative immunoglobulin levels obtained in
samples from 320 consecutive HIV-1infected patients.
Samples with possible PEP abnormalities underwent
immunofixation. The PEP pattern was normal in 83.8%
of samples, 8.1% had subtle oligoclonal banding, and
4.4% had a low-concentration (<5% of total protein)
monoclonal band. Hypogammaglobulinemia and
polyclonal hypergammaglobulinemia accounted for
1.9% each. In multivariate analysis, younger age (odds
ratio [OR], 1.06 with each decreasing year of life; 95%
confidence interval [CI], 1.02-1.11; P = .016), female
sex (OR, 2.4; 95% CI, 1.13-5.11; P = .02), viral load
(OR, 1.50 with each increasing logarithmic viral load
of 1.0; 95% CI, 1.14-1.98; P = .004), and CD4 cell
count (350 vs <350/L [0.35 109/L]) (OR, 2.71;
95% CI, 1.09-6.75; P = .032) were associated with
monoclonal or oligoclonal banding. These results
suggest that younger HIV-1infected patients with a
more robust immune system (higher CD4 cell count),
which is stimulated by uncontrolled viremia, are most
likely to have an augmented B-cell response to HIV
infection. One manifestation of this B-cell response is
low-concentration monoclonal banding in 4.4% of the
patients studied.

596
596

Am J Clin Pathol 2007;128:596-603

DOI: 10.1309/QWTQFGA9FXN02YME

Early in the investigation of AIDS, it was recognized that

many patients showed abnormal immunoglobulin patterns by
protein electrophoresis (PEP).1 These abnormalities included
polyclonal hypergammaglobulinemia, which may be striking,
hypogammaglobulinemia, and oligoclonal banding and monoclonal immunoglobulin bands.2 Over the years, there have
also been case reports of plasma cell dyscrasias occurring in
HIV+ patients.3-5 Although still uncommon, the incidence of
plasma cell neoplasia associated with HIV infection may be
increasing.6
The widespread adoption of highly active antiretroviral
therapy (HAART) since 1997 has dramatically altered the
treatment of and prognosis for HIV+ patients.7,8 As early as
1999, it was reported that HAART reduced monoclonal protein concentrations in HIV patients.8,9 Given these dramatic
changes in treatment, disease course, and life expectancy, coupled with the increased resolution and sensitivity of available
protein electrophoresis systems in clinical laboratories, we
sought to study the current electrophoresis patterns in a cohort
of HIV+ patients. The aim of the present study was to examine the prevalence and nature of PEP abnormalities present in
contemporary patients with HIV and to identify factors that
may be associated with these abnormalities.

Materials and Methods

Sample Selection
The study was designed to look at samples from a brief
cross-section of all patients with HIV at Beth Israel Deaconess
Medical Center, Boston, MA (our institution). During the
American Society for Clinical Pathology

Immunopathology / ORIGINAL ARTICLE

study period (October and November 2003), we tested a plasma sample (to be discarded) from each consecutive unique
patient with HIV for whom a specimen was submitted to our
laboratory for quantitative HIV viral load determination and a
concurrent CD4 cell count was performed. In our institution,
PEP is ordered for approximately 9% of known HIV+
patients, whereas 0.4% of all PEPs done in our institution are
performed on HIV-infected patients. The Committee on
Clinical Investigations at our institution approved the protocol
for this study. Patient demographics (age and sex) were
recorded. We were unable to obtain adequate information
regarding antiretroviral therapy. HIV viral load results (Cobas
Amplicor PCR, Roche Diagnostics, Indianapolis, IN) and
CD4+ T-cell counts (4-color flow cytometric analysis using a
FACSCalibur instrument, Becton Dickinson, San Jose, CA,
using their Multiset software) were obtained from the clinical
records of all patients.
Immunoglobulin Measurements
Measurement of total IgG, IgA, and IgM levels (Tinaquant, Roche Diagnostics) were performed on all samples.
Electrophoresis, Immunofixation, and Densitometry
Electrophoresis (Hydrasys, Sebia, Norcross, GA) was
performed on all samples. The electrophoretic gels were
reviewed independently by 3 pathologists (L.P., G.L.H., and
B.A.B.). Immunofixation (Hydrasys) was performed on samples when abnormalities were detected. The electrophoretic
patterns were categorized as normal, hypogammaglobulinemia, polyclonal hypergammaglobulinemia, oligoclonal banding (2 discrete bands of immunoglobulin confirmed by
immunofixation), or monoclonal band (1 discrete band of
immunoglobulin confirmed by immunofixation). For samples
with a single monoclonal band, the intensity of the band was
measured by densitometry (Phoresis software, Sebia), and the
total protein value was obtained (Hitachi 917, Roche
Diagnostics). The final determination of the pattern was made
by comparing the interpretations of all pathologists and, when

different, reviewing the gels together to reach a consensus

interpretation.
Statistical Analysis
Differences in immunoglobulin levels, age, CD4 cell
counts, and viral loads between males and females and
between patients with or without monoclonal and/or oligoclonal banding were investigated by using the independentsamples t test. For analyses using viral load, the logarithm of
the viral load was always used. Our method has a limit of
detection of 50 copies per milliliter. For the purposes of statistical analysis, samples with undetected viral loads were treated as 50 copies per milliliter. Associations between sex and
age (<40 vs 40 years), CD4 cell count (high or low using a
cutoff of 350 cells/L [0.35 109/L]), and detectable viral
load were evaluated by using the 2 test. Associations between
banding (monoclonal or oligoclonal) and sex, CD4 cell count
(high or low using the aforementioned cutoff), viral load
(detected or undetected), and age (<40 vs 40 years) were also
evaluated by using the 2 test. Logistic regression was performed using banding as the dependent variable and age, sex,
CD4 cell count, and viral load each as sole covariates.
After significant variables were identified in the univariate analysis, forward stepwise logistic regression was performed using combinations of these variables until the best
multivariate logistic regression model was determined.
Statistical calculations were performed by using SPSS software, version 9.0 (SPSS, Chicago, IL).

Results
Patient Characteristics
The patient population included 253 males and 67
females. Patient characteristics are given in Table 1. Ages
ranged from 7 to 67 years (median, 42 years). The mean age
of females (41.1 years) was slightly younger than that of

Table 1
HIV+ Patient Baseline Characteristics and Immunoglobulin Concentrations in 320 Cases
Characteristic

Female (n = 67)

Male (n = 253)

Mean age (y)

No. (%) age <40 y
Mean CD4 cell count/L ( 109/L)
Mean log HIV viral load (copies/mL)
No. (%) with undetected HIV viral load (copies/mL)
No. (%) with CD4 cell count <350/L ( 109/L)
Mean IgG, mg/dL (g/L)
Mean IgA, mg/dL (mg/L)
Mean IgM, mg/dL (mg/L)

41.1
29 (43)
563.9 (0.56)
2.8318
29 (43)
19 (28)
2,022.3 (20.2)
258.7 (2,587)
134.6 (1,346)

43.5
71 (28.1)
556.4 (0.56)
2.7634
116 (45.8)
76 (30.0)
1,534.9 (15.3)
305.23 (3,052)
119.2 (1,192)

.023*
.026
.867*
.704*
.881
.783
<.001*
.072*
.231*

Independent samples t test.

2 test.

Am J Clin Pathol 2007;128:596-603

American Society for Clinical Pathology

597

DOI: 10.1309/QWTQFGA9FXN02YME

597
597

Konstantinopoulos / MONOCLONAL BANDING IN HIV PATIENTS

males, and this difference was statistically significant.

Correspondingly, there was a higher percentage of females
younger than 40 years compared with males.
There was no statistically significant difference in the
viral load and CD4 cell counts between male and female
patients (Table 1). Furthermore, there was no statistically significant difference between the percentages of male and
female patients who had an undetected viral load. Similarly,
there was no statistically significant difference between the
percentages of male and female patients who had a low CD4
count less than the cutoff.
Immunoglobulin Concentrations
The median concentrations of immunoglobulins among
all patients were as follows: IgG, 1,518 mg/dL (15.18 g/L;
range, 410-3,791 mg/dL [4.1-37.91 g/L]); IgA, 254 mg/dL
(2,540 mg/L; range, 8-1,332 mg/dL [80-13,320 mg/L]); and
IgM, 94 mg/dL (940 mg/L; range, 40-572 mg/dL [400-5,720
mg/L]). There were differences when the immunoglobulin
concentrations were analyzed by sex: Females had higher
average levels of IgG, but differences in IgM and IgA did not
reach statistical significance.
By using the reference ranges recommended by the assay
manufacturer (IgG, 7-16 g/L; IgA, 700-4,000 mg/L; and IgM,
400-2,300 mg/L), samples were also classified as having
decreased, normal, or increased levels of specific
immunoglobulins. Overall, 139 had an elevated IgG concentration, 72 had an increased IgA level, and 35 had an increased
IgM level. In 11 samples, levels of all 3 immunoglobulins
were increased, and in only 1 sample, levels of all 3
immunoglobulins were decreased. Two statistically significant
differences were noted in immunoglobulin levels. Among
females, 51 (76%) had elevated levels of IgG, whereas for
males, the rate was lower, 89 (35.2%; P < .001; 2). However,
63 males (24.9%) had increased IgA levels compared with
only 9 (13%) of females (P < .05; 2).
Electrophoretic Patterns
The consensus interpretations of the protein electrophoretic patterns are shown in Table 2. Of the 14 monoclonal bands, 13 were of the IgG type and 1 was IgG . All
bands represented less than 5% (average, 2.1%) of the protein
present by densitometric measurement. Multiplication of the
densitometric measurement by the total protein concentration
allowed an estimate of the absolute concentration to be calculated in 12 cases (1 case did not have a total protein measurement, and in 1 case, the monoclonal band overlaid a normal
band). The estimated concentration of monoclonal protein
ranged from 0.3 to 4.65 g/L, with an average of 1.85 g/L. One
sample with a monoclonal band showed concomitant polyclonal hypergammaglobulinemia, and only 1 case had visually apparent hypogammaglobulinemia Image 1. Among the
598
598

Am J Clin Pathol 2007;128:596-603

DOI: 10.1309/QWTQFGA9FXN02YME

Table 2
Electrophoretic Interpretation of 320 Plasma Samples From
320 HIV+ Patients
Protein Electrophoresis Pattern

No. of Patients (%)

Normal
Oligoclonal banding
Monoclonal band
Hypogammaglobulinemic
Polyclonal hypergammaglobulinemic

268 (83.8)
26 (8.1)
14 (4.4)
6 (1.9)
6 (1.9)

26 samples showing oligoclonal banding, 24 had only IgG

bands. Of these samples, 13 had and bands present, and 11
had only bands with a single light chain type identified. Two
oligoclonal samples had an IgA or IgM band in addition to an
IgG band.
Of 14 cases with monoclonal banding, data on stage and
treatment status were available for 12. Of the 12 patients represented, 11 received HAART at some point during their
infection. Eight were receiving HAART when the sample was
obtained, and 3 had received HAART in the past but were not
receiving it when the sample was obtained. Of the 12 patients,
2 had AIDS when the sample was obtained, and 10 had
asymptomatic HIV infection. The mean duration of HIV
infection in patients with monoclonal banding was 8.0 years
(the duration was known for 11 of 12 patients). These results
are given in Table 3.
Factors Associated With Banding
To look for associations with electrophoretic patterns, the
samples were grouped into 2 categories by PEP interpretation:
(1) oligoclonal or monoclonal bands present (banding) or (2)
no bands seen (no banding). The distribution of different clinicopathologic characteristics among patients whose samples
demonstrated banding on PEP and those whose did not are
given in Table 4 and in Figure 1. In univariate analysis, age,
sex, and viral load were statistically significantly associated
with banding. Specifically, patients with banding were
younger (mean, 38.9 vs 43.6 years; P < .001) and had higher
average viral loads (logarithmic viral load, 3.34 vs 2.69; P =
.004) than patients without banding. Furthermore, females
were more likely than males to have any form of banding seen
on PEP. There was no statistically significant difference in the
CD4 cell count among patients with or without banding.
We further studied the associations between age, viral
load, and CD4 cell count with banding. Specifically, patients
younger than 40 years were more likely to show banding on
PEP than patients 40 years or older and this association
approached statistical significance (Table 4). Patients with a
detected viral load were more likely to show banding on PEP
than patients with an undetected viral load (Table 4). Finally,
there was a trend observed in which patients with CD4 cell
American Society for Clinical Pathology

Immunopathology / ORIGINAL ARTICLE

PEP

IgG

IgA

IgM

PEP

IgG

IgA

IgM

Image 1 Immunofixation electrophoresis gels from 2 HIV+ patients. From left to right, the tracks represent all proteins (labeled
PEP [protein electrophoresis pattern]), IgG, IgA, IgM, and and light chains. A, Oligoclonal banding is present in the region
with at least 3-4 bands of IgG and IgG immunoglobulins. The total IgG, IgA, and IgM levels were 3,694 mg/dL (36.9 g/L), 190
mg/dL (1,900 mg/L), and 199 mg/dL (1,990 mg/L), respectively. B, A single monoclonal band of IgG is seen in the region,
marked by the arrows. The total immunoglobulin levels were decreased (IgG, 410 mg/dL [4.1 g/L]; IgA, <8 mg/dL [80 mg/L]; and
IgM, <4 mg/dL [40 mg/L]). This was the only case that showed decreased levels of all 3 immunoglobulins and was also the only
case with a single monoclonal band occurring concomitantly with decreased immunoglobulin levels.

counts of 350/L (0.35 109/L) or more were more likely to

show banding on PEP than patients with CD4 cell counts of
less than 350/L (0.35 109/L), but this difference did not
reach statistical significance (Table 4).
Logistic regression with banding as a dependent variable
and with age, logarithmic viral load, CD4 cell count, and sex
each as sole covariates demonstrated identical associations to
those with the 2 test or t test. It is important to note that we

looked for possible interactions between sex, age, CD4 cell

count, and viral load by including interaction terms (eg, sex *
viral load, sex * CD4 cell count, age * sex) in the logistic
regression, and there were no statistically significant interactions identified.
After all significant variables (age, sex, and viral load)
were identified and interactions were excluded, forward stepwise logistic regression was performed using combinations of

Table 3
Characteristics of 14 Patients With Monoclonal Banding
Case No.

Stage

Therapy Status

Duration of HIV Infection (y)

1
2
3
4
5
6
7
8
9
10
11
12
13
14

Asymptomatic
Asymptomatic
Unknown
AIDS
Asymptomatic
Asymptomatic
Asymptomatic
Asymptomatic
Asymptomatic
Asymptomatic
Asymptomatic
Asymptomatic
Unknown
AIDS

HAART
HAART
Unknown
HAART
HAART
None when sample obtained; received HAART in the past
None when sample obtained; received HAART in the past
None when sample obtained; received HAART in the past
HAART
No HAART
HAART
HAART
Unknown
HAART

4
6
Unknown
15
7
15
Unknown
6
2
3
11
16
Unknown
3

HAART, highly active antiretroviral therapy.

Am J Clin Pathol 2007;128:596-603

American Society for Clinical Pathology

599

DOI: 10.1309/QWTQFGA9FXN02YME

599
599

Konstantinopoulos / MONOCLONAL BANDING IN HIV PATIENTS

these variables until the best model was determined.

Immunoglobulin levels were not included in the multivariate
analysis because they are not independent variables (ie,
depend on other variables such as CD4 cell count and HIV
viral load). After we successfully included all 3 variables (age,
sex, and viral load) in the logistic regression model, we tried
to add the dichotomous variable CD4 cell count (ie, high or
low CD4 count corresponding to a CD4 count 350 or
<350/L [0.35 109/L]) to the model. The resultant model
Table 5 with these 4 variables (age, sex, viral load, and CD4
cell count) proved to be better than the 3-variable model (age,
sex, and viral load) with a nonsignificant goodness-of-fit test
result (P = .3563). According to the final best model, younger
age, female sex, viral load, and CD4 cell count (350 vs
<350/L [0.35 109/L]) were associated with banding.

Discussion

Table 4
Distribution of Different Clinicopathologic Characteristics
Among 320 Patients With Banding (Monoclonal or Oligoclonal)
and Without Bands in the Protein Electrophoresis Pattern*
Factor
Mean age (y)
Age (y)
<40 (n = 100)
40 (n = 220)
Sex
Female (n = 67)
Male (n = 253)
CD4 cell count/L ( 109/L)
<350 (0.35) (n = 95)
350 (0.35) (n = 225)
HIV viral load
Undetected (n = 145)
Detected (n = 175)
Mean CD4 cell count/L
Mean log HIV viral load
(copies/mL)

Banding
(n = 40)

No Banding
(n = 280)

38.9

43.6

18 (18.0)
22 (10.0)

82 (82.0)
198 (90.0)

15 (22)
25 (9.9)

52 (78)
228 (90.1)

8 (8)
32 (14.2)

87 (92)
193 (85.8)

9 (6.2)
31 (17.7)
570.5 (0.57)
3.34

136 (93.8)
144 (82.3)
556.2 (0.56)
2.69

P
<.001
.067
.011
.195
.002
.797
.004

Although most laboratory professionals are aware that a

large percentage of HIV+ patients may show striking polyclonal hypergammaglobulinemia and/or oligoclonal banding,

Data are given as number (percentage) unless otherwise indicated.

Independent samples t test.
2 test.

A
100

100
90

80

80
70
60
Percentage

Percentage

60

40

50
40
30
20

20

No banding
Banding

0
Undetected

Detected

Viral Load

10

No banding
Banding

0
Female

Male
Gender

Figure 1 Percentage of banding and no banding on the protein electrophoresis patterns in patients with detected or
undetected viral loads (A) and female or male patients (B).

600
600

Am J Clin Pathol 2007;128:596-603

DOI: 10.1309/QWTQFGA9FXN02YME

American Society for Clinical Pathology

Immunopathology / ORIGINAL ARTICLE

Table 5
Final Logistic Regression Model
95% Confidence
Interval
Variable

Odds Ratio

Age*
1.06
Female sex
2.4
Log viral load
1.50
CD4 cell count 350
2.71
vs <350/L (0.35 109/L)
*

Lower

Upper

1.02
1.13
1.14
1.09

1.11
5.11
1.98
6.75

.0163
.0232
.0043
.0327

With each decreasing year of life.

With each increasing logarithmic viral load of 1.0.

we found the vast majority of electrophoretograms in this population to be visually unremarkable or to show only subtle
abnormalities. Estimates of the prevalence of monoclonal
bands in healthy populations vary, but one study found that
5% of a cohort of adults (with unknown HIV status) between
22 and 65 years of age showed 1 or more electrophoretically
homogeneous (presumably monoclonal) bands using a highresolution agarose gel technique.10 The reported prevalence of
monoclonal gammopathy of undetermined significance
(MGUS) in healthy (non-HIV+) subjects is between 1% and
2% and varies by age.11 A recent report noted that only 2% of
a series of more than 1,300 patients with MGUS were younger
than 50 years,11 and the overall male/female ratio was
1.19:1.12 Our findings, in which 4.4% of HIV+ patients had
monoclonal and 8.1% had oligoclonal bands, show higher
rates than the aforementioned findings. This confirms earlier
studies that have shown an increase in oligoclonal and monoclonal banding in HIV+ patients Table 6.1,13-24 Many of the
prior studies involving HIV-infected persons have been relatively small, with only 3 studies reporting data for more than
70 patients. There is considerable variability in the findings

reported, which is likely due to diverse methods used for identifying patients, variable sensitivity of the electrophoretic
methods, and interobserver variation in classification of electrophoretic patterns.
Despite these factors, we can still draw some general conclusions from the literature. The reported prevalence of monoclonal bands in HIV+ patients ranges between 2.5% and 53%
and for oligoclonal banding, between 3% and 63%. Some
studies have grouped oligoclonal and monoclonal patterns
together, as we have, and the prevalence in these studies
ranges from 9% to 69% (Table 6). The reported prevalence of
oligoclonal bands in HIV+ patients may have decreased over
time, with 4 of 5 studies published in 1989 or before reporting
oligoclonal banding in more than 40% of patients, but 3 studies since 1990 have found oligoclonal banding in fewer than
40% of patients. Our finding of a 12% prevalence of oligoclonal or monoclonal bands in HIV+ patients is consistent
with this trend. We think that this decrease is probably real and
not an artifact of our study method. We used a high-resolution
semiautomated electrophoresis system that is in wide use in
clinical laboratories. In addition, this system has been reported to be very sensitive for the detection of low-concentration
bands.25 Our findings confirm this high sensitivity because the
average concentration of the monoclonal bands we identified
was only 2.1% (1.85 g/L) of total protein, and the smallest
band we detected was only 0.3 g/L.
Elevated levels of immunoglobulins have been described
in patients with AIDS since at least 1984.26 Our findings confirm that a significant proportion (43%) of HIV+ patients have
elevations of the level of 1 or more immunoglobulins, usually
IgG. The magnitude of the IgG elevation is modest in most
patients, but 10% of patients had IgG levels more than double
the upper limit of the reference range. By visual inspection of
the PEP, we interpreted only 6 cases to be hypergammaglobulinemic (these 6 cases had an average total IgG level of 2,651

Table 6
Studies of Prevalence of Monoclonal and/or Oligoclonal Banding in AIDS/HIV+ Patients by Protein Electrophoresis
Study

Year

No. of Patients

Oligoclonal
Banding (%)

Monoclonal
Banding (%)

Oligoclonal or
Monoclonal Banding (%)

Heriot et al1
Papadopoulos et al13
Sala et al14
Crapper et al15
Sala et al16
Lefrere et al17
Papadopoulos and Costello18
Taichman et al19
Bratt et al20
Amadori et al21
Frankel et al22
Lefrere et al23
Pontet et al24
Present study

1985
1985
1986
1987
1987
1987
1987
1988
1989
1990
1993
1993
1998
2006

24
42
26
65
55
243
68
44
25
60
13
341
212
320

61

63
43
56
15
38

22
8

53
8
12
6

2.5

3.2
11
4

69

9
47

33
12

Am J Clin Pathol 2007;128:596-603

American Society for Clinical Pathology

601

DOI: 10.1309/QWTQFGA9FXN02YME

601
601

Konstantinopoulos / MONOCLONAL BANDING IN HIV PATIENTS

mg/dL [26.51 g/L]). This reflects the fact that we were mainly looking for qualitatively abnormal bands (M proteins) and
were relatively conservative in what we interpreted as a polyclonal increase in staining in the region. Although protein
electrophoretograms can be quantitated by densitometry,
clearly, our visual impression was not as sensitive as the quantitative measurement.
A number of studies21,27-30 have demonstrated that oligoclonal and monoclonal bands and even paraproteins in cases
of multiple myeloma in HIV+ patients can be directed against
HIV antigen epitopes. Other studies have shown that on further evaluation, even an apparently monoclonal paraprotein
band detected by PEP can often be shown to have more than
1 light chain type and reactivity against different HIV antigens, suggesting that there is an exuberant polyclonal immune
response against HIV that may manifest as monoclonal or
oligoclonal banding.28,29 These findings, coupled with our
own observation that patients with elevated total IgG levels
were more likely to show oligoclonal or monoclonal bands,
support the hypothesis that these bands are more than likely
part of an immune response directed toward HIV. Given these
lines of evidence, it may make sense to consider oligoclonal
banding and low-concentration monoclonal bands in HIV+
patients to be part of the same spectrum of immune response.
This is an important point for laboratory professionals
and clinicians. There is evidence of a slightly increased risk of
multiple myeloma in HIV+ patients,31-34 and there have been
a number of case reports of unequivocal multiple myeloma
occurring in HIV+ patients. Amara et al9 reported that 28% of
their HIV+ patients with monoclonal gammopathy developed
a malignancy (usually a B-cell/plasma cell malignancy) after
a mean follow-up of only 21 months. However, our findings
support the contention that most of the single, apparently
monoclonal, bands seen on PEP are not of a character that is
likely to be associated with multiple myeloma or other significant systemic plasma cell dyscrasia, but are more typical of
what might be seen with MGUS. Only 1 monoclonal band
was accompanied by a concomitant decrease in the other
immunoglobulins, and none of the bands was present in high
concentration (all were <5% of total protein). It is possible
that the cases in which we identified only a single band would
have shown additional bands if analyzed using more sensitive
techniques than are typically used in a clinical laboratory.
Although protein electrophoresis may not be routinely
ordered for HIV+ patients (ordered approximately in 9% of
the patients in our institution), if a PEP is performed and a
low-concentration monoclonal band is seen, it would be wise
to consider monitoring the PEP result over time.
Our multivariate analysis identified 4 factors that were
significantly associated with banding (monoclonal or oligoclonal): increased viral load, female sex, younger age, and
higher CD4 cell counts. Although these 4 factors can be seen
602
602

Am J Clin Pathol 2007;128:596-603

DOI: 10.1309/QWTQFGA9FXN02YME

in patients who are newly diagnosed and have not yet received
HAART, our findings do not support that possibility. In that
regard, the overwhelming majority of patients (11/12) with
monoclonal banding were receiving HAART therapy and had
already had HIV infection for a mean duration of 8.0 years.
These results are consistent with the hypothesis that as the HIV
viral load increases, the host B cells respond by making more
immunoglobulins (probably directed specifically at HIV epitopes), which can be detected on PEP as bands (monoclonal or
oligoclonal). Our findings are consonant with those of Redgrave
and colleagues,35 who reported that aviremic HIV-infected
patients receiving HAART have lower plasma levels of IgG and
IgA than viremic HIV-infected patients. The fact that patients
with CD4 cell counts of less than 350/L (0.35 109/L) were
less likely to show banding on PEP is also consistent with the
central role of CD4 cells in promoting terminal differentiation,
antibody secretion, and immunoglobulin isotype switching in
activated B cells. The increased banding associated with female
sex is unclear but may be related to the increased incidence of
autoimmune diseases in female patients.36,37
Our study of immunoglobulins and protein electrophoretic patterns in 320 HIV+ patients showed an increased prevalence of oligoclonal or monoclonal bands. The prevalence of
HIV-related banding seems to be lower than what has been
previously reported, perhaps in part owing to the more effective antiretroviral therapies that are currently available. HIV+
patients who were younger and female, had higher HIV viral
loads, and had CD4 cell counts more than 350/L (0.35
109/L) were more likely to have an abnormal banding pattern
detected on PEP. Although HIV+ patients may be at slightly
higher risk of developing multiple myeloma, all of the abnormal PEP patterns we identified were oligoclonal or low-concentration monoclonal bands.
From the 1Division of Hematology Oncology and 3Department of
Pathology, Beth Israel Deaconess Medical Center and Harvard
Medical School, Boston, MA; and 2Department of Pathology,
Baystate Medical Center, Tufts University School of Medicine,
Springfield, MA.
Address reprint requests to Dr Dezube: Beth Israel
Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215.
Acknowledgments: We acknowledge the invaluable assistance
of Shiva Gautam, PhD, with statistical analysis and of Carol Sklar
for technical laboratory assistance in this study.

References
1. Heriot K, Hallquist AE, Tomar RH. Paraproteinemia in
patients with acquired immunodeficiency syndrome (AIDS) or
lymphadenopathy syndrome (LAS). Clin Chem.
1985;31:1224-1226.
2. Le Carrer D. Serum Protein Electrophoresis and Immunofixation:
Illustrated Interpretations. Washington, DC: American
Association for Clinical Chemistry; 1994:94-108.

American Society for Clinical Pathology

Immunopathology / ORIGINAL ARTICLE

3. Kumar S, Kumar D, Schnadig VJ, et al. Plasma cell myeloma

in patients who are HIV-positive. Am J Clin Pathol.
1994;102:633-639.
4. Nosari AM, Landonio G, Cantoni S, et al. Multiple myeloma
associated to HIV infection: report of two patients. Eur J
Haematol. 1996;56:98-99.
5. Lallemand F, Fritsch L, Cywiner-Golenzer C, et al. Multiple
myeloma in an HIV-positive man presenting with primary
cutaneous plasmacytomas and spinal cord compression. J Am
Acad Dermatol. 1998;39:506-508.
6. Pantanowitz L, Dezube BJ. Multiple myeloma and HIV
infection: casual or causal coincidence [editorial]? AIDS Read.
2003;13:386-387.
7. Palella FJ, Delaney KM, Moorman AC, et al. Declining
morbidity and mortality among patients with advanced human
immunodeficiency virus infection. N Engl J Med.
1998;338:853-860.
8. Cauda R, Lucia MB, Marasca G, et al. Beneficial effect of
highly active antiretroviral therapy (HAART) in reducing
both HIV viral load and monoclonal gammopathy [letter].
Eur J Haematol. 1999;63:134-135.
9. Amara S, Dezube BJ, Cooley TP, et al. HIV-associated
monoclonal gammopathy: a retrospective analysis of 25
patients. Clin Infect Dis. 2006;43:1198-1205.
10. Papadopoulos NM, Elin RJ, Wilson DM. Incidence of globulin banding in a healthy population by high-resolution
electrophoresis. Clin Chem. 1982;28(4 pt 1):707-708.
11. Kyle RA, Therneau TM, Rajkumar SV, et al. A long-term
study of prognosis in monoclonal gammopathy of
undetermined significance. N Engl J Med. 2002;346:564-569.
12. Kyle RA, Rajkumar SV. Monoclonal gammopathies of
undetermined significance: a review. Immunol Rev.
2003;194:112-139.
13. Papadopoulos NM, Lane HC, Costello R, et al. Oligoclonal
immunoglobulins in patients with the acquired
immunodeficiency syndrome. Clin Immunol Immunopathol.
1985;35:43-46.
14. Sala PG, Mazzolini S, Tonutti E, et al. Monoclonal
immunoglobulins in HTLV IIIpositive sera [letter]. Clin
Chem. 1986;32:574.
15. Crapper RM, Deam DR, Mackay IR. Paraproteinemias in
homosexual men with HIV infection: lack of association with
abnormal clinical or immunologic findings. Am J Clin Pathol.
1987;88:348-351.
16. Sala PG, Mazzolini S, Tonutti E, et al. Monoclonal-oligoclonal
immunoglobulins in HTLV III infected subjects: HTLV
IIIinduced or HTLV IIIassociated? Boll Ist Sieroter Milan.
1987;66:14-17.
17. Lefrere JJ, Fine JM, Lambin P, et al. Monoclonal
gammopathies in asymptomatic HIV-seropositive patients
[letter]. Clin Chem. 1987;33:1697-1698.
18. Papadopoulos NM, Costello R. Oligoclonal immunoglobulins
in the sera of healthy subjects at risk for AIDS. Clin Biochem.
1987;20:257-258.
19. Taichman DB, Bayer K, Senior M, et al. Oligoclonal
immunoglobulins in HIV-antibodypositive serum. Clin Chem.
1988;34:2377.

20. Bratt G, Waldenlind L, Von Krogh G, et al. A prospective

study of immunoglobulin changes in relation to human
immunodeficiency virus infection in a cohort of homosexual
men. Scand J Immunol. 1989;29:589-595.
21. Amadori A, Gallo P, Zamarchi R, et al. IgG oligoclonal bands in
sera of HIV-1 infected patients are mainly directed against HIV-1
determinants. AIDS Res Hum Retroviruses. 1990;6:581-586.
22. Frankel EB, Greenberg ML, Makuku S, et al. Oligoclonal
banding in AIDS and hemophilia. Mt Sinai J Med.
1993;60:232-237.
23. Lefrere JJ, Debbia M, Lambin P. Prospective follow-up of
monoclonal gammopathies in HIV-infected individuals.
Br J Haematol. 1993;84:151-155.
24. Pontet F, Gue X, Mazeron MC, et al. Qualitative
immunoglobulin abnormalities in HIV-positive patients: longterm follow-up. Clin Chem Lab Med. 1998;36:493-496.
25. Tormey WP. Low concentration monoclonal and oligoclonal
bands in serum and urine using the Sebia Hydragel Protein
Electrophoresis System [letter]. Clin Chem Lab Med.
1998;36:253-254.
26. Chess Q, Daniels J, North E, et al. Serum immunoglobulin
elevations in the acquired immunodeficiency syndrome (AIDS):
IgG, IgA, IgM and IgD. Diagn Immunol. 1984;2:148-153.
27. Papadopoulos NM, Costello R, Ceroni M, et al. Identification
of HIV-specific oligoclonal immunoglobulins in serum of
carriers of HIV antibody. Clin Chem. 1988;34:973-975.
28. Ng VL, Hwang KM, Reyes GR, et al. High titer anti-HIV
antibody reactivity associated with a paraprotein spike in a
homosexual male with AIDS related complex. Blood.
1988;71:1397-1401.
29. Ng VL, Chen KH, Hwang K, et al. The clinical significance
of human immunodeficiency virus type 1associated
paraproteins. Blood. 1989;74:2471-2475.
30. Konrad RJ, Kricka LJ, Goodman DB, et al. Brief report:
myeloma-associated paraprotein directed against the HIV-1
p24 antigen in an HIV-1seropositive patient. N Engl J Med.
1993;328:1817-1819.
31. Goeder T, Cote TR, Virgo P, et al. Spectrum of AIDSassociated malignancy disorders. Lancet. 1998;351:1833-1839.
32. Goedert JJ. The epidemiology of acquired immunodeficiency
syndrome malignancies. Semin Oncol. 2000;27:390-401.
33. Bonnet F, Lewden C, May T, et al. Malignancy-related causes
of death in human immunodeficiency virusinfected patients
in the era of highly active antiretroviral therapy. Cancer.
2004;101:317-324.
34. Dezube BJ, Aboulafia DM, Pantanowitz L. Plasma cell
disorders in HIV-infected patients: from benign gammopathy
to multiple myeloma. AIDS Read. 2004;14:372-374, 377-379.
35. Redgrave BE, Stone SF, French MA, et al. The effect of
combination antiretroviral therapy on CD5 B-cells, B-cell
activation and hypergammaglobulinaemia in HIV-1-infected
patients. HIV Med. 2005;6:307-312.
36. Cutolo M. Sex hormone adjuvant therapy in rheumatoid
arthritis. Rheum Dis Clin North Am. 2000;26:881-895.
37. Wilder RL. Hormones, pregnancy, and autoimmune diseases.
Ann N Y Acad Sci. 1998;840:45-50.

Am J Clin Pathol 2007;128:596-603

American Society for Clinical Pathology

603

DOI: 10.1309/QWTQFGA9FXN02YME

603
603