Documente Academic
Documente Profesional
Documente Cultură
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
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
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)
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*
DOI: 10.1309/QWTQFGA9FXN02YME
597
597
Table 2
Electrophoretic Interpretation of 320 Plasma Samples From
320 HIV+ Patients
Protein Electrophoresis Pattern
Normal
Oligoclonal banding
Monoclonal band
Hypogammaglobulinemic
Polyclonal hypergammaglobulinemic
268 (83.8)
26 (8.1)
14 (4.4)
6 (1.9)
6 (1.9)
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.
Table 3
Characteristics of 14 Patients With Monoclonal Banding
Case No.
Stage
Therapy Status
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
DOI: 10.1309/QWTQFGA9FXN02YME
599
599
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
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
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
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
DOI: 10.1309/QWTQFGA9FXN02YME
601
601
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
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.
DOI: 10.1309/QWTQFGA9FXN02YME
603
603