Available via license: CC BY 4.0
Content may be subject to copyright.
Cytomegalovirus infection does not impact on survival
or time to first treatment in patients with chronic
lymphocytic leukemia
Helen Marie Parry,
1
* Sarah Damery,
2
Christopher Hudson,
3
Matthew J. Maurer,
4
James R. Cerhan,
4
Annette Pachnio,
1
Jusnara Begum,
1
Susan L. Slager,
4
Christopher Fegan,
5
Stephen Man,
5
Christopher Pepper,
5
Tait D. Shanafelt,
6
Guy Pratt,
1
and Paul A. H. Moss
1
Human cytomegalovirus (HCMV) is a widely prevalent herpes virus which establishes a state of chronic
infection. The establishment of CMV-specific immunity controls viral reactivation and leads to the accumulation
of very large numbers of virus-specific T cells which come to dominate the immune repertoire. There is
concern that this may reduce the immune response to heterologous infections and HCMV infection has been
associated with reduced survival in elderly people. Patients with chronic lymphocytic leukemia (B-CLL) suffer
from a state of immune suppression but have a paradoxical increase in the magnitude of the CMV-specific T
cell and humoral immune response. As such, there is now considerable interest in how CMV infection impacts
on the clinical outcome of patients with B-CLL. Utilizing a large prospective cohort of patients with B-CLL
(n5347) we evaluated the relationship between HCMV seropositivity and patient outcome. HCMV seropositive
patients had significantly worse overall survival than HCMV negative patients in univariate analysis (HR 52.28,
95% CI: 1.34–3.88; P50.002). However, CMV seropositive patients were 4 years older than seronegative
donors and this survival difference was lost in multivariate modeling adjusted for age and other validated
prognostic markers (P50.34). No significant difference was found in multivariate modeling between HCMV
positive and negative patients in relation to the time to first treatment (HR51.12, 95% CI: 0.68–1.84; P50.65).
These findings in a second independent cohort of 236 B-CLL patients were validated. In conclusion no
evidence that HCMV impacts on the clinical outcome of patients with B-CLL was found.
Am. J. Hematol. 91:776–781, 2016. V
C2016 The Authors. American Journal of Hematology Published by Wiley Periodicals, Inc.
䊏Introduction
Human Cytomegalovirus (HCMV) is a prevalent beta-herpes virus that is usually asymptomatic upon primary infection. HCMV maintains life-
long latency within cells of the myeloid lineage and its prevalence increases with age [1,2]. The age-adjusted prevalence in the United States is
50.4% by the age of 50 years [3] and continues to increase with each decade of life [4].
In health, a significant proportion of both CD41and CD81T cells are required to maintain viral latency and prevent HCMV viral reactiva-
tion [5,6]. This extreme expansion of HCMV-specific T cells, which is most marked in the CD81T cell repertoire, contributes to a reduction in
the number of na€
ıve T cells and leads to an inversion of the CD4:CD8 T cell ratio [7]. There is increasing concern that the burden of CMV infec-
tion can lead to health problems, particularly in older people, and CMV seropositivity has been associated with impaired responses to vaccination
[8], increasing levels of inflammatory cytokines [9] and an increase in overall morbidity and mortality in several studies [10,11].
Chronic lymphocytic leukemia (B-CLL) is characterized by the proliferation of mature B lymphocytes and has a median age at diagnosis of 71
years (https://www.hmrn.org/statistics/incidence; cited 1/6/15). The development of B-CLL leads to a state of immune suppression and patients
exhibit an increased susceptibility to infection from an early stage of disease [12]. The mechanisms that underlie the reduction in immune compe-
tence are multifactorial but include hypogammaglobulinemia and T cell dysfunction [13]. In addition, the burden of chronic infection and changes
in the T cell repertoire are also important for overall prognosis in B-CLL, and an inverted CD4:CD8 T cell ratio is associated with disease progres-
sion [14]. The CD4:CD8 ratio is markedly lower in CMV seropositive individuals compared with those who remain uninfected, and this effect is
particularly profound in patients with CLL. Indeed, one of the paradoxes of CMV infection in patients with CLL is that the magnitude of the
virus-specific immune response is actually higher than in healthy individuals, despite the state of immune suppression, in a mechanism that is
1
Institute of Immunology and Immunotherapy, University of Birmingham, Birmingham, B15 2TT, United Kingdom;
2
Institute of Applied Health Research, University
of Birmingham, Birmingham, B15 2TT, United Kingdom;
3
Faculty of Medicine & Health Sciences, University of Nottingham, Leicestershire, LE12 5RD, United King-
dom;
4
Department of Health Sciences Research, Mayo Clinic, Rochester, Minnesota;
5
Division of Cancer & Genetics, Heath Park, Cardiff, CF14 4XN, United Kingdom;
6
Department of Medicine, Mayo Clinic, Rochester, Minnesota
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Conflict of interest: Nothing to report
*Correspondence to: Helen M. Parry; Institute of Immunology and Immunotherapy, University of Birmingham. Birmingham, B15 2TT, United Kingdom.
E-mail: H.m.parry@ bham.ac.uk
Contract grant sponsor: Wellcome Fellowship grant and also a Mayo partnership grant with the University of Birmingham.
Received for publication: 5 March 2016; Revised: 22 April 2016; Accepted: 25 April 2016
Am. J. Hematol. 91:776–781, 2016.
Published online: 28 April 2016 in Wiley Online Library (wileyonlinelibrary.com).
DOI: 10.1002/ajh.24403
V
C2016 The Authors. American Journal of Hematology Published by Wiley Periodicals, Inc.
776 American Journal of Hematology, Vol. 91, No. 8, August 2016 doi:10.1002/ajh.24403
RESEARCH ARTICLE
AJH
AJH
believed to be a response to increased levels of subclinical viral reacti-
vation [15,16]. As such there is now considerable interest in how
chronic CMV infection may influence the clinical profile of patients
with B-CLL [17]. Indeed, if the infection was shown to accelerate the
progression of disease then there may be a potential role for the use
of anti-viral medication in disease management [18].
We utilized two large independent prospective cohorts of newly
diagnosed patients (the “discovery” and “confirmation” cohorts) to
evaluate the relationship between HCMV seropositivity and disease
outcome in patients with B-CLL. In particular we examined the rela-
tionship between CMV serostatus and time to treatment (TTT) and
overall survival (OS) and show that infection status does not impact
significantly on these variables.
䊏Methods
Patients in the discovery cohort (n5347) were participants in the ongoing pro-
spective cohort study of NHL patients from the Molecular Epidemiology Resource
of the University of Iowa/Mayo Clinic Lymphoma Specialized Program of Research
Excellence (SPORE). This study was reviewed and approved by the Human Subjects
Institutional Review Board at the Mayo Clinic and the University of Iowa, and
written informed consent was obtained from all participants in accordance with the
Declaration of Helsinki. Since September 2002, enrollment was offered to consecu-
tive newly diagnosed patients with B-CLL, evaluated at Mayo Clinic Rochester or
the University of Iowa within 9 months of diagnosis [19]. All patients were US resi-
dents aged 18 years and older. Exclusion criteria included known HIV infection
and unwillingness or inability to provide written informed consent. Patients fulfilled
IWCLL criteria and/or fulfilled the World Health Organization criteria for the small
lymphocytic lymphoma variant (SLL) variant of B-CLL. Baseline clinical, laboratory,
and treatment data were abstracted from medical records and participants provided
peripheral blood serum samples. All participants were then followed every 6
months for the first 3 years, and then annually thereafter; disease progression and
deaths were verified through medical record review.
For the confirmation cohort, newly diagnosed patients with B-CLL were
enrolled from clinics at the University Hospital of Wales and Llandough Hospital.
Patients were diagnosed by an experienced hematologist and met the IWCLL crite-
ria for B-CLL. Informed consent was taken in accordance with ethical approval
obtained from the South East Wales Research ethics committee. Samples were taken
within 12 months of diagnosis and prior to any treatment. Data for disease pro-
gression and mortality was collected on an annual basis and verified through medi-
cal records.
Prognostic parameters. Prognostic testing, including immunoglobulin heavy
chain variable [IGHV] region gene mutation analysis, ZAP-70 status, CD38 status,
CD49d status, and cytogenetic abnormalities assessed by interphase FISH testing,
were all performed as part of clinical or research studies using methods previously
described [20–22]. Adverse FISH results were defined by the presence of 17p dele-
tion or 11q deletion. Serum immunoglobulin measurement was only available at
the time of sample collection in the discovery cohort.
CMV ELISA. HCMV seropositivity and HCMV IgG titer were determined using
the previously described CMV enzyme-linked immunosorbent assay (ELISA) [23].
Briefly, mock and viral-infected lysate was coated onto ELISA plates and incubated
overnight. Patient plasma samples and a series dilution of standards (derived from
three known HCMV-positive plasma donors) were added to the plate for 1 hr. The
plate was washed three times. An anti-human IgG-horseradish peroxidase second-
ary antibody was then added to the plate for 1 hr. After further washing, TMB
(3,30,5,50-tetramethylbenzidine) substrate was added and the plate kept in the dark
for 10 minutes before the addition of 1 M HCl. The sample was assessed using an
ELISA reader at 450 nm. To determine HCMV titers, mock values were first sub-
tracted from lysate values and titers then calculated against the standard curve. Val-
ues greater than 10 were considered to be seropositive. To ensure accuracy, all
samples were tested in duplicate.
Statistical analysis. Chi-square testing (and Fisher’s exact test where appropri-
ate) was used to assess the association between HCMV seropositivity and demo-
graphic, clinical and prognostic factors in B-CLL patients, and both the discovery
and confirmation cohorts were characterized using descriptive statistics. For com-
patibility between datasets, Rai staging for the discovery cohort was converted to
Binet staging. Rai 0/1 representing Binet A, Rai stage II representing Binet B, and
Rai stage III/IV representing Binet stage C. TTT was defined as the time from diag-
nosis to disease progression requiring treatment. OS was defined as the time from
diagnosis to death due to any cause. Patients without an event or death were cen-
sored at time of last known follow-up. Kaplan–Meier survival curves and Cox pro-
portional hazards regression models were used to assess the association between
HCMV positivity and the outcomes of interest. Cox models were adjusted for
demographic and B-CLL prognostic factors which were found to impact signifi-
cantly on survival and TTT in univariate analysis. An analysis of the continuously
distributed CMV IgG ELISA values was also performed. HCMV titer was trans-
formed to the base 2 log and entered as a covariate into cox regression analysis.
Analyses were performed using SPSS Version 21 (Armonk NY: IBM Corp), and the
threshold for statistical significance was set at P50.05. CMV ELISA results were
analyzed using GraphPad Prism Version 5.03 (GraphPad Software, San Diego, CA),
and CMV titers were calculated with reference to the standard curve.
䊏Results
Patient cohorts
The discovery cohort consisted of 390 patients, of which 347
patients were identified as having serum available for CMV testing
and were used to explore the relationship between CMV and clinical
outcome in B-CLL. Serum samples were collected a median of 2.9
months following diagnosis (range 0–73).
The confirmatory cohort consisted of 236 patients. All patients had
serum drawn within 12 months of diagnosis, median 5.8 months
(range 0.5–11.75). The clinical characteristics for these two prospec-
tive observational cohorts are shown in Table I.
CMV status and clinical outcome
Discovery cohort. The median follow up time for this cohort was
3 years (IQR 2.4–4.2) and at the point of last data collection 68 par-
ticipants [20% (68/347)] had died. Of the 347 participants, 198 (57%)
were HCMV seropositive and 149 (43%) were HCMV seronegative.
HCMV positive patients were significantly older at the time of diag-
nosis with a median age of 64 years (range 37–87) compared with 60
years in uninfected individuals (range 37–91); P<0.0001. There was
no association between HCMV seropositivity and any other demo-
graphic or clinical characteristic (Binet stage, sex, ECOG performance
score or absolute lymphocyte count) or prognostic markers (ZAP-70,
CD38, IGHV, CD49d, FISH) (Table II). About 49 deaths (25%) were
recorded for the HCMV positive group and 19 deaths (13%) were
noted in the HCMV negative group. Of the 68 deaths, 33 were unre-
lated to CLL and 29 were documented as a CLL-specific cause of
TABLE I. Patient Characteristics of the Discovery and Confirmatory
Cohorts
Discovery cohort
(n5347)
Confirmatory cohort
(n5236)
Median age at diagnosis (years) 62 (37.0–91) 65.1 (24–99)
Follow up time (years 1IQR) 2.98 (2.41–4.17) 7 (4–10.2)
Male sex 237 (68%) 150 (64%)
Binet stage A 180 (52%) 182 (77%)
Binet stage B 155 (45%) 25 (11%)
Binet stage C 11 (3%) 27 (11%)
Missing 1 2
CD38 positive 82 (24%) 99 (42%)
CD38 negative 218 (63%) 117 (50%)
CD38 missing 47 20
Zap70 positive 93 (27%) 82 (35%)
Zap 70 negative 203 (59%) 134 (57%)
Zap 70 missing 51 20
IGHV mutated 176 (51%) 125 (53%)
IGHV unmutated 104 (30%) 40 (12%)
IGHV missing 67 71
CD49d positive 64 (18%) 102 (43.2%)
Cd49d negative 131 (38%) 69 (29.2%)
Cd49d missing 152 65
FISH
Normal 76(22%) 61(26%)
13q2120(35%) 37(16%)
12150(14%) 6(3%)
11q222(6%) 15(6%)
17p211(3%) 5(2%)
Other 5(1%) 2(0.8%)
Missing 63 110
RESEARCH ARTICLE Cytomegalovirus does not impact on CLL prognosis
doi:10.1002/ajh.24403 American Journal of Hematology, Vol. 91, No. 8, August 2016 777
death (7 due to secondary malignancies, 3 due to infection, and 19
from refractory disease). The cause of death was not available for 6
patients.
In univariate analysis HCMV positive patients had significantly
inferior overall survival, with the risk of death at any time point
more than twice that observed in HCMV negative patients (HR 2.28,
95% CI: 1.34–3.88; P50.0024) (Fig. 1a). CLL-specific cause of death
was also investigated in relation to CMV serostatus, HCMV positive
patients again had a significantly higher risk of dying from CLL-
related causes at any point than HCMV negative patients (HR 3.41,
95% CI: 1.37–8.47, P50.008).
However, after adjusting for age, Stage, ZAP-70, FISH, IGHV,
CD38, and CD49d in a multivariate survival model, this risk was
attenuated, with only age (HR 1.12; 95% CI: 1.06–1.19; P<0.0001)
and unmutated IGHV (HR 2.78; 95% CI: 1.07–7.23; P50.036)
remaining significant as independent risk factors for OS. No differ-
ence in survival was seen between HCMV positive and negative par-
ticipants in the multivariate model (HR 50.61, 95% CI: 0.22–1.69;
P50.34).
We next went on to investigate the relationship between CMV
infection and the time to treatment, and this information was avail-
able on 322 of the original 349 patients in the cohort. The median
time to first treatment was 7.4 years and at the last follow up 115 of
the 322 patients (36%) were confirmed to have received treatment for
B-CLL. A log rank test for differences between HCMV positive and
negative participants demonstrated no difference in time to first treat-
ment in either a univariate model (HR 50.90, 95% CI: 0.62–1.30;
P50.560) or a multivariate model adjusted for age, Rai Stage, and
prognostic variables (HR 51.12, 95% CI: 0.68–1.84; P50.651). Binet
stage C compared with stage A (HR 50.33, 95% CI: 0.01–0.09;
P<0.0001) and stage B (HR 50.11, 95% CI: 0.04–0.28; P<0.0001),
as well as expression of CD49d (HR 2.05, 95% CI: 1.16–3.62;
P50.013) remained significant predictors of time to treatment after
adjusting for other variables in the multivariate model (Fig. 1b).
TABLE II. Association of CMV with Selected Characteristics of the
Discovery Cohort
CMV negative
(n5149) (43%)
CMV positive
(n5198) (57%) Pvalue
Age Median (range) 60 (37.0–91.0) 64 (37.0–87.0) P<0.0001
Gender
Female 43 (28.9) 67 (66.2) P50.32
Male 106 (71.1) 131 (33.8)
Binet stage
A 71 (47.7) 109 (55.3) P50.32
B 72 (48.4) 83 (42.1)
C 6 (4.0) 5 (2.5)
Missing 0 1
CD38
Positive 34 (25.8) 48 (28.6) P50.59
Negative 98 (74.2) 120 (71.4)
Missing 17 30
ZAP70
Positive 42 (32.8) 51 (30.4) P50.65
Negative 86 (67.2) 117 (69.6)
Missing 21 30
IGHV
Mutated 81 (35.2) 95 (61.3) P50.55
Unmutated 44 (64.8) 60 (38.7)
Missing 24 43
CD49d
Positive 23 (27.7) 41 (36.6) P50.19
Negative 60 (72.3) 71 (63.4)
Missing 66 86
FISH
Normal 30 (24.4) 46 (28.6) P50.21
13q 50 (40.7) 70 (43.5)
Trisomy 12 25 (20.3) 25 (15.5)
11q 13 (10.6) 9 (5.6)
17p 2 (1.6) 9 (5.6)
Other 3 (2.4) 2 (1.2)
Missing 26 37
Serum IgG
N9595P50.63
Mean (SD) 818.3 (281.3) 854.4 (316.0)
Median 820 806
Range 190.0–1750.0 189.0–2220.0
Missing 54 61
Figure 1.(A) Kaplan–Meier curve for overall survival by CMV status in the
discovery cohort is shown. Univariate analysis showed CMV positive
patients had significantly inferior overall survival, with a hazards ratio of
2.28 (P50.0024) However, this significance was lost on multivariate analy-
sis (HR 50.61, 95% CI: 0.22–1.69; P50.34) (B) demonstrates the time to
first treatment by CMV status in the discovery cohort. No significant differ-
ence in time to first treatment was observed on multivariate analysis
(HR 51.12, 95% CI: 0.68–1.84; P50.651).
Parry et al. RESEARCH ARTICLE
778 American Journal of Hematology, Vol. 91, No. 8, August 2016 doi:10.1002/ajh.24403
Confirmation cohort. In order to validate the results found in the
discovery cohort, a further large cohort with a longer follow up time
was assessed for the same outcome data. The median follow up time
of the confirmation cohort was 7 years (IQR: 4–10.2 years). Of 236
patients in this cohort, 109 (46%) had required treatment at the point
of last follow up and 93 had died (39%). 179 (76%) of the patients
were found to be HCMV seropositive and CMV serostatus was not
found to be associated with age at diagnosis within this cohort. No
association between HCMV seropositivity and any of the other demo-
graphic, clinical, or prognostic markers was found in the confirmation
cohort. Of the 93 deaths, 10 were unrelated to CLL and 83 were
documented as a CLL-specific cause of death (3 due to secondary
malignancies, 40 due to infection, and 40 from refractory disease).
As noted in the discovery cohort, a shorter median overall survival
of 10.6 years was noted in HCMV positive participants (95% CI: 8.4–
12.5) compared with 15.9 years (95% CI: 6.9–25.0) in HCMV nega-
tive participants although this did not reach statistical significance on
either univariate (HR 1.45, 95% CI: 0.86–2.43; P50.158) or multivar-
iate analysis (HR 0.96, 95% CI: 0.57–1.63; P50.882). Analysing CLL-
specific cause of death also failed to reveal any significant difference
between HCMV positive and negative patients 1.60 (0.91–2.81,
P50.100).
When investigating the influence of CMV infection on TTT in this
confirmation cohort, Binet stage, CD38, CD49d, ZAP-70, IGVH, and
adverse FISH status were all found to impact significantly on TTT in
univariate testing. Data was complete for prognostic factors in only
68 patients, of which 44 participants required treatment. No associa-
tion between HCMV status and TTT was found once these prognos-
tic variables were adjusted for in multivariate modeling (HR 1.13,
95% CI: 0.50–2.55; P50.766). Only Binet stage C compared with
stage A (HR 0.43, 95% CI: 0.12–0.14; P<0.0001) and stage B (HR:
0.29 (0.09–0.93); P50.037) were significant independent predictors
of time to treatment.
CMV titer and clinical outcome
High titers of CMV-specific antibody have been associated with
reduced overall survival in healthy elderly individuals and may repre-
sent a response to increased levels of subclinical viral load [24]. We
investigated this in our cohorts by calculating hazard ratios for the
association between patients with B-CLL and Log2 HCMV titer as a
continuous variable for both overall survival and time to treatment.
Data were available on 198 HCMV seropositive B-CLL patients
within the discovery cohort. About 49 deaths had occurred within
this cohort and although a trend was observed between increasing
antibody titer and reduced survival, this did not reach statistical sig-
nificance (HR 0.95, 95% CI: 0.82–1.09; P50.460). Similarly, in the
confirmation cohort, 75 deaths occurred amongst 179 HCMV positive
participants, but no relationship was seen between HCMV titer and
OS (HR 1.02, 95% CI: 0.88–1.19; P50.793). Therefore, no significant
relationship was found between level of CMV-specific antibody and
overall survival in CMV positive patients.
Similarly, no significant associations were seen in either cohort
between HCMV titer and TTT (Table III). Given the similar out-
come data for HCMV seropositivity on OS and TTT in both
cohorts, and in an attempt to increase the number of HCMV posi-
tive individuals available for statistical analysis, the datasets were
pooled and used to assess whether the titer of HCMV IgG as a con-
tinuous variable impacts on overall survival or TTT. Data from 337
HCMV positive individuals were analyzed, of which 124 deaths
were recorded during follow up, but again no relationship was
found between HCMV titer and overall survival (HR 1.01, 95% CI:
0.91–1.13; P50.800) or time to treatment (HR 1.00, 95% CI: 0.91–
1.10; P50.972).
䊏Discussion
CMV is recognized as an important pathogen in patients who are
immune suppressed and viral reactivation is a common occurrence in
patients with B-CLL who undergo treatment with alemtuzumab. How-
ever, there is now interest in the potential health associations of chronic
viral carriage in a range of clinical settings, particularly in patients with
B-CLL in whom the CMV-specific T cell immune response can expand
to a remarkably high frequency [15,16]. Despite this, no published epi-
demiological data exists investigating the relationship between CMV
and B-CLL prognosis. This work, reports the prevalence of HCMV
within the two cohorts was approximately 60%–70%, which is compara-
ble with studies of viral seroprevalence in healthy people at a similar
median age of 64 years [23]. As such we find no evidence that CMV
infection is associated with the development of B-CLL. In this regard it
is noteworthy that chronic viral infection has been suggested as a poten-
tial antigenic stimulus to account for the finding of shared immuno-
globulin genes sequences in tumors from different patients. Indeed, the
recurrent IGHV1-69 sequence that is common in B-CLL has been
shown to react with the pUL32 phosphoprotein from CMV [25].
In the discovery cohort HCMV seropositivity was associated with
a twofold increased risk of death on univariate analysis. However,
this association was lost in multivariable analysis after adjusting for
other prognostics variables, the most notable of which was age, as
HCMV positive patients were on average 4 years older than the
CMV seronegative patients. All other variables tested for within this
cohort were not statistically different between CMV positive and neg-
ative patients. The prevalence of CMV is known to rise with increas-
ing age and as such it may not be surprising that CMV seropositive
TABLE III. Impact of CMV IgG Titer on Overall Survival and Time to First Treatment
OVERALL
SURVIVAL
Median IgG titer
(inter-quartile range)
a
all patients
Median IgG titer
(inter-quartile range)
alive patients
Median IgG titer
(inter-quartile range)
deceased patients
Number of
deaths in CMV
positive
individuals (%)
Hazards
ratio
(95% CI) Pvalue
Discovery cohort 227.1 (113.8–503.4) 234.7 (114.6–556.1) 213.0 (105.2–440.7) 49/198 (25%) 0.95 (0.82–1.09) 0.46
Confirmatory cohort 202.1 (103.9–357.1) 196.9 (110.5–342.2) 209.5 (93.2–404.9) 75/179 (42%) 1.02 (0.88–1.19) 0.79
Combined cohorts 211.9 (106.8–434.6) 211.9 (114.6–448.8) 210.1 (98.3–433.1) 124/377 (33%) 1.01 (0.91–1.13) 0.80
TIME TO
TREATMENT
Median IgG
titer (inter-
quartile range)
a
all patients
Median IgG
titer (inter-
quartile range)
treated patients
Median IgG
titer (inter-
quartile range)
untreated patients
Number of
patients treated
in CMV positive
individuals (%)
Hazards ratio
(95% CI)
Pvalue
Discovery cohort 227.8 (1 18.5–506.5) 234.7 (123.2–455.0) 227.1 (107.0–522.7) 63/185 (34%) 0.99 (0.87–1.13) 0.93
Confirmatory cohort 202.1 (103.9–357.1) 204.0 (95.5–357.4) 185.7 (105.7–363.0) 86/179 (48%) 1.02 (0.89–1.17) 0.78
Combined cohorts 211.9 (107.1–445.6) 213.0 (1 15.0–418.6) 211.9 (107.1–445.6) 149/364 (41%) 1.00 (0.91–1.10) 0.97
a
Titer is the base 2 log, to take account of the non-linearity of the untransformed variable.
RESEARCH ARTICLE Cytomegalovirus does not impact on CLL prognosis
doi:10.1002/ajh.24403 American Journal of Hematology, Vol. 91, No. 8, August 2016 779
patients were somewhat older [24,26]. However, samples were tested
at the point of diagnosis, which then raised an alternative possibility
that CMV infection could potentially serve to delay the diagnosis of
B-CLL. To address this question it would be important to obtain
information on CMV serostatus on a healthy cohort from the same
geographical area taken during the same time period but no appro-
priate cohort was available for comparison. Moreover, no difference
was found between age and CMV serostatus in the confirmation
cohort, which suggests that a potential protective effect of CMV
infection on the diagnosis of B-CLL is unlikely. This discrepancy
between the two cohorts in relation to age and CMV serostatus is
likely to reflect the different geographical locations of the two cohorts,
as the timing of primary exposure to CMV is known to vary dramati-
cally depending on geographical and socio-economic factors [27].
The lack of an association between survival and CMV infection
in B-CLL patients contrasts with studies which have investigated
the impact of CMV status in healthy elderly individuals. This may
be due partly to the younger age of the patients with B-CLL com-
pared with studies of healthy elderly people, which have focused
predominantly on those between 70 and 100 years of age
[11,28,29]. In addition, the detrimental impact of chronic HCMV
on the health of older people has been linked with elevated levels
of inflammatory markers, including IL-6 [23,30,31], which are often
already increased in patients with B-CLL and may therefore con-
found any impact of HCMV on further health outcomes [32,33]. It
may also suggest that any negative effect of HCMV infection in the
elderly is outweighed by a diagnosis of B-CLL. Another possibility
that should not be discounted is that CMV infection may actually
play a potential beneficial role on immune function in patients with
B-CLL. The virus stimulates a strong Th1 immune response and
leads to accumulation of large numbers of cytotoxic cells, including
NK and YdT cells. CMV infection has been associated with a
reduction in disease relapse after stem cell transplantation and may
therefore potentially play a role in the setting of other hemopoietic
malignancies [34,35].
In healthy elderly populations a low CD4:CD8 ratio and a high
CMV-specific T cell response both predict for poor survival [36] and
it has previously been suggested that CMV seropositivity may impact
on poorer survival in B-CLL, albeit using a small sample size (n557)
and based on univariate analysis only [16]. Our current study sug-
gests that HCMV does not impact on survival negatively and
although HCMV status was determined around the time of diagnosis
in this work, it is unlikely many (if any) patients who were CMV
negative at diagnosis would later seroconvert [37]. As age can act as a
confounding factor for HCMV seropositivity in survival analysis, a
larger patient cohort with longer follow up time may be needed to
identify if these factors act independently.
In addition, high HCMV-specific antibody titers have been corre-
lated with poor clinical outcome in elderly donors [28] but we were
unable to demonstrate this in patients with B-CLL. The development
of hypogammaglobulinemia is a feature of progressive CLL but unlike
other herpes viruses such as VZV and EBV, HCMV-specific IgG
titers have been shown to increase with disease progression and pro-
gressive hypogammaglobulinemia, Indeed, in a large cohort of
patients with CLL, almost 50% demonstrated an increase in their
CMV-specific IgG titer 4.6 years later and all CMV positive patients
were still found to have detectable CMV IgG [37]. As such it is
unlikely that any patients with hypogammaglobulinemia were incor-
rectly identified as seronegative for latent CMV. The increase in
CMV IgG observed over time may suggest episodes of reactivation
are occurring. However, patients with B-CLL rarely suffer from overt
clinical episodes of HCMV reactivation, except after treatment with
alemtuzumab or following allogeneic transplantation, suggesting that
CMV-specific T cell function is adequate to prevent clinical viral rep-
lication. Indeed, the function of CD81HCMV specific T cells is not
obviously impaired in patients with CLL [38,39].
In conclusion, our data provide no evidence that CMV infection
can predispose toward the development of B-CLL. In addition, estab-
lished infection and HCMV-specific antibody titer do not appear to
have an impact on time to first treatment or overall survival in newly
diagnosed patients with B-CLL.
䊏Acknowledgments
The authors thank the Mayo Clinic and Cardiff trial coordinators
for their efforts in patient follow-up and data collection.
䊏References
1. Slobedman B, Mocarski ES. Quantitative analy-
sis of latent human cytomegalovirus. J Virol
1999;73:4806–4812.
2. Lopo S, Vinagre E, Palminha P, et al. Seropreva-
lence to cytomegalovirus in the Portuguese pop-
ulation, 2002-2003. Euro Surveill 2011;16:pii.
3. Bate SL, Dollard SC, Cannon MJ. Cytomegalovi-
rus seroprevalence in the United States: The
national health and nutrition examination sur-
veys, 1988-2004. Clin Infect Dis 2010;50:1439–
1447.
4. Vyse AJ, Hesketh LM, Pebody RG. The burden
of infection with cytomegalovirus in England
and Wales: How many women are infected in
pregnancy? Epidemiol Infect 2009;137:526–533.
5. Chidrawar S, Khan N, Wei W, et al. Cytomega-
lovirus-seropositivity has a profound influence
on the magnitude of major lymphoid subsets
within healthy individuals. Clin Exp Immunol
2009;155:423–432.
6. Derhovanessian E, Maier AB, Hahnel K, et al.
Infection with cytomegalovirus but not herpes
simplex virus induces the accumulation of late-
differentiated CD41and CD81T-cells in
humans. J Gen Virol 2011;92:2746–2756.
7. Olsson J, Wikby A, Johansson B, et al. Age-
related change in peripheral blood T-
lymphocyte subpopulations and cytomegalovirus
infection in the very old: The Swedish longitudi-
nal OCTO immune study. Mech Ageing Dev
2000;121:187–201.
8. Trzonkowski P, Mysliwska J, Szmit E, et al.
Association between cytomegalovirus infection,
enhanced proinflammatory response and low
level of anti-hemagglutinins during the anti-
influenza vaccination–an impact of immunose-
nescence. Vaccine 2003;21:3826–3836.
9. Weaver JD, Huang MH, Albert M, et al. Inter-
leukin-6 and risk of cognitive decline: MacAr-
thur studies of successful aging. Neurology
2002;59:371–378.
10. Strandberg TE, Pitkala KH, Tilvis RS. Cytomeg-
alovirus antibody level and mortality among
community-dwelling older adults with stable
cardiovascular disease. Jama 2009;301:380–382.
11. Savva GM, Pachnio A, Kaul B, et al. Cytomega-
lovirus infection is associated with increased
mortality in the older population. Aging Cell
2013;12:381–387.
12. Parry HM, Birtwistle J, Whitelegg A, et al. Poor
functional antibody responses are present in
nearly all patients with chronic lymphocytic leu-
kaemia, irrespective of total IgG concentration,
and are associated with increased risk of infec-
tion. Br J Haematol 2015;171:887–890.
13. Forconi F, Moss P. Perturbation of the normal
immune system in patients with CLL. Blood
2015;126:573–581.
14. Nunes CT, Wong R, Mason M, et al. Expansion
of a CD81PD-11replicative senescence pheno-
type in early stage CLL patients is associated
with inverted CD4:CD8 ratios and disease pro-
gression. Clin Cancer Res 2011;18:678–687.
15. Mackus WJM, Frakking FNJ, Grummels A,
et al. Expansion of CMV-specific
CD81CD45RA1CD27- T cells in B-cell
chronic lymphocytic leukemia. Blood 2003;102:
1057–1063.
16. Pourgheysari B, Bruton R, Parry H, et al. The
number of cytomegalovirus-specific CD41T
cells is markedly expanded in patients with B-
cell chronic lymphocytic leukemia and deter-
mines the total CD41T-cell repertoire. Blood
2010;116:2968–2974.
17. Katrinakis G, Kyriakou D, Alexandrakis M,
et al. Evidence for involvement of activated
CD81/HLA-DR1cells in the pathogenesis of
neutropenia in patients with B-cell chronic lym-
phocytic leukaemia. Eur J Haematol 1995;55:
33–41.
18. Pachnio A, Begum J, Fox A, et al. Acyclovir
therapy reduces the cd41t cell response against
the immunodominant pp65 protein from cyto-
megalovirus in immune competent individuals.
PLoS One 2015;10:e0125287.
19. Shanafelt TD, Drake MT, Maurer MJ, et al.
Vitamin D insufficiency and prognosis in
chronic lymphocytic leukemia. Blood 2011;117:
1492–1498.
20. Dewald GW, Brockman SR, Paternoster SF,
et al. Chromosome anomalies detected by
interphase fluorescence in situ hybridization:
Correlation with significant biological features
of B-cell chronic lymphocytic leukaemia. Br J
Haematol 2003;121:287–295.
21. Jelinek DF, Tschumper RC, Geyer SM, et al.
Analysis of clonal B-cell CD38 and immuno-
globulin variable region sequence status in rela-
tion to clinical outcome for B-chronic
Parry et al. RESEARCH ARTICLE
780 American Journal of Hematology, Vol. 91, No. 8, August 2016 doi:10.1002/ajh.24403
lymphocytic leukaemia. Br J Haematol 2001;115:
854–861.
22. Shanafelt TD, Geyer SM, Bone ND, et al.
CD49d expression is an independent predictor
of overall survival in patients with chronic lym-
phocytic leukaemia: A prognostic parameter
with therapeutic potential. Br J Haematol 2008;
140:537–546.
23. Kilgour AH, Firth C, Harrison R, et al. Seropo-
sitivity for CMV and IL-6 levels are associated
with grip strength and muscle size in the
elderly. Immun Ageing 2013;10:33.
24. Dollard SC, Staras SAS, Amin MM, et al.
National prevalence estimates for cytomegalovi-
rus igm and igg avidity and association between
high igm antibody titer and low igg avidity !.
Clin Vaccine Immunol 2011;18:1895–1899.
25. Steininger C, Widhopf GF, Ghia EM, et al.
Recombinant antibodies encoded by IGHV1-69
react with pUL32, a phosphoprotein of cyto-
megalovirus and B-cell superantigen. Blood
2012;119:2293–2301.
26. Hecker M, Qiu D, Marquardt K, et al. Continu-
ous cytomegalovirus seroconversion in a large
group of healthy blood donors. Vox Sang 2004;
86:41–44.
27. Cannon MJ, Schmid DS, Hyde TB. Review of
cytomegalovirus seroprevalence and demo-
graphic characteristics associated with infection.
Rev Med Virol 2010;20:202–213.
28. Roberts ET, Haan MN, Dowd JB, et al. Cyto-
megalovirus antibody levels, inflammation, and
mortality among elderly Latinos over 9 years of
follow-up. Am J Epidemiol 2010;172:363–371.
29. Hadrup SR, Strindhall J, Kollgaard T, et al. Lon-
gitudinal studies of clonally expanded CD8 T
cells reveal a repertoire shrinkage predicting
mortality and an increased number of dysfunc-
tional cytomegalovirus-specific T cells in the
very elderly. J Immunol 2006;176:2645–2653.
30. Wikby A, Nilsson B-O, Forsey R, et al. The
immune risk phenotype is associated with IL-6
in the terminal decline stage: Findings from the
Swedish NONA immune longitudinal study of
very late life functioning. Mech Ageing Dev
2006;127:695–704.
31. Trzonkowski P, Mysliwska J, Godlewska B, et al.
Immune consequences of the spontaneous pro-
inflammatory status in depressed elderly
patients. Brain Behav Immun 2004;18:135–148.
32. Adami F, Guarini A, Pini M, et al. Serum levels
of tumour necrosis factor-alpha in patients with
B-cell chronic lymphocytic leukaemia. Eur J
Cancer 1994;30A:1259–1263.
33. Callea V, Morabito F, Luise F, et al. Clinical sig-
nificance of sIL2R, sCD23, sICAM-1, IL6 and
sCD 14 serum levels in B-cell chronic lympho-
cytic leukemia. Haematologica 1996;81:310–315.
34. Ito S, Pophali P, Co W, et al. CMV reactivation
is associated with a lower incidence of relapse
after allo-SCT for CML. Bone Marrow Trans-
plant 2013;48:1313–1316.
35. Elmaagacli AH, Steckel NK, Koldehoff M, et al.
Early human cytomegalovirus replication after
transplantation is associated with a decreased
relapse risk: Evidence for a putative virus-
versus-leukemia effect in acute myeloid leuke-
mia patients. Blood 2011;118:1402–1412.
36. Wikby A, Johansson B, Olsson J, et al. Expan-
sions of peripheral blood CD8 T-lymphocyte
subpopulations and an association with cyto-
megalovirus seropositivity in the elderly: The
Swedish NONA immune study. Exp Gerontol
2002;37:445–453.
37. Vanura K, Rieder F, Kastner MT, et al. Chronic
lymphocytic leukemia patients have a preserved
cytomegalovirus-specific antibody response
despite progressive hypogammaglobulinemia.
PLoS One 2013;8:e78925.
38. te Raa GD, Pascutti MF, Garcia-Vallejo JJ, et al.
CMV-specific CD81T-cell function is not
impaired in chronic lymphocytic leukemia.
Blood 2014;123:717–724.
39. Riches JC, Davies JK, McClanahan F, et al. T
cells from CLL patients exhibit features of T-cell
exhaustion but retain capacity for cytokine pro-
duction. Blood 2013;121:1612–1621.
RESEARCH ARTICLE Cytomegalovirus does not impact on CLL prognosis
doi:10.1002/ajh.24403 American Journal of Hematology, Vol. 91, No. 8, August 2016 781
