Lymphopenia, Infectious Complications, and Outcome in
Spontaneous Intracerebral Hemorrhage
Andrea Morotti
1,2
, Sandro Marini
1,2
, Michael J. Jessel
2
, Kristin Schwab
2
, Christina
Kourkoulis
2
, Alison M. Ayres
2
, M. Edip Gurol
2
, Anand Viswanathan
2
, Steven M.
Greenberg
2
, Christopher D. Anderson
1,2
, Joshua N. Goldstein
1,2,3
, and Jonathan Rosand
1,2
1
Division of Neurocritical Care and Emergency Neurology, Department of Neurology,
Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
2
J. P. Kistler Stroke Research Center, Massachusetts General Hospital, Harvard Medical School,
175 Cambridge Street, Suite 300, Boston, MA 02114, USA
3
Department of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School,
Boston, MA, USA
Abstract
Background—Lymphopenia is increasingly recognized as a consequence of acute illness and
may predispose to infections. We investigated whether admission lymphopenia (AL) is associated
with increased risk of infectious complications and poor outcome in patients with spontaneous
intracerebral hemorrhage (ICH).
Methods—We retrospectively analyzed a prospectively collected cohort of ICH patients
ascertained between 1994 and 2015. We identified subjects with lymphocyte count obtained within
24 h from onset, and AL was defined as lymphocyte count < 1000/µL. Infectious complications
were assessed through retrospective chart review. Association between AL, infections, and
mortality was investigated using multivariable logistic regression.
Results—Of the 2014 patients meeting inclusion criteria, 548 (27.2%) had AL and 605 (30.0%)
developed an infectious complication. Case-fatality at 90 days was 36.9%. Patients with AL had
larger hematoma volumes, higher frequency of intraventricular hemorrhage, and lower Glasgow
Coma Scale score on presentation (all
p
< 0.001). AL was independently associated with increased
risk of pneumonia [odds ratio (OR) 1.97, 95% confidence interval (CI) 1.50–2.58,
p
< 0.001] and
multiple infections (OR 1.84, 95% CI 1.24–2.71,
p
= 0.003). AL was also an independent
predictor of 90-day mortality (OR 1.55, 95% CI 1.18–2.04,
p
= 0.002) after adjusting for
confounders.
✉
Andrea Morotti, amorotti@mgh.harvard.edu; a.morotti@ymail.com.
Compliance with Ethical Standards
Ethical Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of
the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable
ethical standards.
Conflict of interest Andrea Morotti, Sandro Marini, Michael Jessel, Kristin Schwab, Christina Kourkoulis, Alison M. Ayres, M. Edip
Gurol, and Steven M. Greenberg report no disclosures
HHS Public Access
Author manuscript
Neurocrit Care
. Author manuscript; available in PMC 2018 April 01.
Published in final edited form as:
Neurocrit Care
. 2017 April ; 26(2): 160–166. doi:10.1007/s12028-016-0367-2.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Conclusions—AL is common in ICH patients and independently associated with increased risk
of infectious complications and poor outcome. Further studies will be needed to determine
whether prophylactic antibiotics in ICH patients with AL can improve outcome.
Keywords
Stroke; Cerebral hemorrhage; Cerebrovascular disorders; Lymphopenia; Infection; Pneumonia
Introduction
Intracerebral hemorrhage (ICH) is still the deadliest type of stroke, with lack of an acute
phase treatment proven to reduce mortality and improve functional outcome [1].
Inflammation and lymphocyte migration into the brain appear to play a key role in
secondary brain damage and clinical deterioration following acute ICH [2]. Clinical trials of
fingolimod in ICH have, therefore, been guided by the theory that lymphocyte depletion may
improve outcome reducing edema and inflammation-mediated tissue damage [3]. Prior
studies, however, suggest that hemorrhagic stroke, like ischemic stroke or other acute severe
illnesses can be accompanied by immunodepression, manifested by lymphopenia [4–6].
Lymphopenia, in turn, may predispose to the development of infections [7, 8]. The role of
lymphopenia in ICH, and in particular, whether it predisposes to infectious complications
(IC) during the acute course is poorly understood and may have implications for the design
of future clinical trials of immunomodulation in ICH. In this study we investigated the
frequency and determinants of lymphopenia in subjects with ICH and whether it was
associated with increased risk of IC and poor outcome.
Methods
Study Design and Patient Selection
All aspects of the study were approved by the Institutional Review Board (IRB). Informed
written or verbal consent was obtained by patients or family members or waived by the IRB.
We retrospectively analyzed an ongoing prospective cohort of patients with spontaneous
ICH collected at a single academic hospital from January 1994 to April 2015 [9, 10]. The
inclusion criteria for the present study were: (1) diagnosis of spontaneous ICH on non-
contrast CT scan (2) complete white blood cell count obtained within 24 h from stroke
onset. Subjects were excluded if there was evidence of (1) traumatic intracranial bleeding,
(2) intracranial tumor, aneurysm, or other vascular malformation presumed to be the cause
of the hemorrhage, (3) hemorrhagic conversion of acute brain infarction, (4) missing data on
infectious complications, and (5) missing follow-up data on mortality at 90 days.
Image Acquisition and Analysis
All images were analyzed by study staff blinded to all clinical and laboratory variables. Non-
contrast CT images were acquired with an axial technique and 5-mm-thickness slices, 120–
140 kVp, 10–500 mA and reviewed for determination of ICH location and the presence of
intraventricular extension of the hematoma (IVH). Hematoma volume was calculated with
semi-automated computer-assisted technique (Analyze Direct 11.0 software).
Morotti et al.
Page 2
Neurocrit Care
. Author manuscript; available in PMC 2018 April 01.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Clinical Variables
Demographic and clinical data were systematically collected through patient and family
members’ interviews and retrospective review of hospital charts. We assessed the presence
of medical history of hypertension, diabetes mellitus, hypercholesterolemia, antiplatelet
therapy, and oral anticoagulant treatment (OAT) as previously described in detail [10]. We
also collected data on pre-stroke functional status, and functional dependence was defined as
requiring assistance in at least one instrumental activity of daily living before the index ICH
[11].
Because previous reports have demonstrated that the relationship between total lymphocyte
count and susceptibility to infections is not linear and that the odds of experiencing an IC
increase only when the lymphocyte count drops below a critical threshold [7, 12], we
analyzed admission lymphopenia (AL) defined as absolute lymphocyte count < 1000/µL [7,
13]. We also collected data on admission leukopenia (absolute leukocyte count < 4000/µL),
admission neutropenia (absolute neutrophil count < 1500/µL), and admission
monocytopenia (absolute monocyte count < 200/µL) [14, 15].
Infectious Complications and Mortality
IC were identified through a retrospective review of hospital charts, discharge reports,
laboratory, and radiological tests. The presence of an IC during the hospital stay was
established according to previously published criteria [16–18] by two investigators (AM,
SM), blinded to the presence of AL. In particular, we looked for evidence of the following
infections during the hospital stay: pneumonia, urinary tract infections, and sepsis. The
diagnosis of pneumonia was based on the combination of typical clinical presentation with
confirmatory chest X-ray changes [16]. A positive urine culture was required for the
diagnosis of urinary tract infection [17]. Sepsis was diagnosed in case of documented source
of infection associated with evidence of acute organ dysfunction [18]. Mucocutaneous
infections, gastrointestinal infections, and meningoencephalitis were grouped in the category
“other infections.” The case-fatality rate at 90 days was assessed via telephone interviews
and querying of the Social Security Death Index (SSDI) national database as previously
described [10].
Statistical Analysis
Categorical variables were expressed as count (%) while continuous variables as median
(interquartile range, IQR). Differences between patients with and without AL and with and
without infections were examined using the χ
2
test or Mann–Whitney
U
test as appropriate.
The association between AL and infectious complications was investigated with a
multivariable logistic regression analysis, adjusted for predictors of infections [19, 20]. All
the variables with
p
value <0.1 in univariate analysis were included in the regression model.
The relationship between AL and 90-day mortality was investigated with a multivariable
logistic regression, accounting for known predictors of outcome in ICH patients (age, ICH
volume, admission Glasgow Coma Scale score, presence on IVH, and infratentorial
location) [21].
p
values <0.05 were considered statistically significant. All analyses were
performed using the statistical package SPSS v. 21, 2012 (www.spss.com).
Morotti et al.
Page 3
Neurocrit Care
. Author manuscript; available in PMC 2018 April 01.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Results
A total of 2403 patients with ICH were screened and 2014 met the eligibility criteria for the
present analysis. The frequency of AL was 27.2% and a total of 605 (30.0%) patients
experienced an infection during the hospital stay. Overall mortality at three months was
36.9%. A total of 351 subjects were excluded because of lack of lymphocyte count, and 38
patients were excluded because of missing clinical or demographic data. Compared to the
study population, patients excluded from the analysis were older, more likely to be on
antiplatelet treatment, and less likely to have a medical history of hypertension, diabetes
mellitus, and hypercholesterolemia. The remaining demographic and clinical characteristics
were similar between the two groups (all
p
values >0.05).
Factors Associated with Admission Lymphopenia
Patients with AL were older and had larger baseline hematoma volume and higher frequency
of intraventricular extension of the hemorrhage (Table 1). In addition, AL was associated
with infratentorial location of the hematoma and lower admission Glasgow Coma Scale
score.
Infectious Complications
The presence of AL was significantly higher in patients with IC (31.6 vs 25.3%,
p
= 0.004)
whereas the frequency of leukopenia, neutropenia, and monocytopenia was similar between
the two groups (Table 2).
After adjustment for potential confounders in multivariable regression, AL was
independently associated with increased risk of pneumonia, and multiple infections, but not
with urinary tract infection and sepsis (Table 3).
Mortality
AL was independently associated with increased case-fatality at 90 days (odds ratio 1.55,
95% CI 1.18–2.04,
p
= 0.002) after accounting for known predictors of mortality in ICH
(age, admission GCS, presence of IVH, infratentorial location of the hematoma and baseline
ICH volume) [21]. Figure 1 shows the adjusted survival analysis stratified by presence of
AL.
Duration of hospitalization was longer in patients experiencing at least one IC (11 vs 5 days,
p
< 0.001). While patients with IC had lower mortality in univariate analysis, their mortality
rate was significantly higher when the analysis was restricted to subjects surviving longer
than 72 h (mortality at 90 days: 27.0 vs 21.4%,
p
= 0.022), consistent with the presence of
survival bias underlying the unrestricted assessment of mortality [22].
All results were unchanged when pre-stroke functional status was also included in the
multivariate analysis (OR for infectious complications 1.46, 95% CI 1.44–1.87,
p
= 0.002;
OR for 90 days mortality 1.48, 95% CI 1.09–2.00,
p
= 0.011).
Morotti et al.
Page 4
Neurocrit Care
. Author manuscript; available in PMC 2018 April 01.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Discussion
We found that lymphopenia on admission is frequent in ICH patients and is independently
associated with increased risk of IC and long-term mortality. In particular, AL was an
independent predictor of pulmonary infections and multiple infections during the hospital
stay.
While it remains to be determined whether the immunosuppression manifested by
lymphopenia precedes the ICH or is an acute consequence of it, accumulating data suggest
that brain injury can influence the physiologic interplay between the central nervous system
and the immune system, leading to the onset of a systemic immunodepressive syndrome [5,
23, 24]. This immunodepressive state is characterized by the acute secretion of several stress
mediators, with secondary apoptotic loss of circulating lymphocytes that appears to be
proportional to the severity and extension of brain damage [5, 6, 23]. Indeed, in our study,
AL was associated with ICH of higher severity. Patients with AL had larger ICH volumes,
lower GCS, and higher frequency of IVH, and AL could theoretically be just a marker of
ICH severity. Pre-stroke functional status may also be an important confounder and the
presence of lymphopenia could simply reflect the degree of disability prior to the index ICH.
However, the association between lymphopenia, infections, and outcome remained
significant after adjusting for multiple potential confounders such as pre-stroke disability
and measures of ICH severity. In agreement with previous studies on stroke-associated
immunodepression [4–6, 8], our results suggest that the presence of lymphopenia correlates
with stroke severity but plays an independent role in predisposing to infections and
unfavorable outcome.
Consistent with prior studies [13], we also demonstrated that AL is an independent predictor
of poor outcome. Our results offer evidence for a mechanism through which AL may be
operating. Infections are indeed a major cause of mortality in stroke patients [19, 25–27],
and AL was associated with an increased risk of pneumonia and multiple infections in our
study.
In this regard, we observed a higher mortality in subjects with infectious complications only
restricting the analysis to patients surviving the first three days after stroke. One possible
explanation is the influence of survival bias. Multiple previous studies showed that
hematoma expansion, intraventricular bleeding, and limitation of care are the main
determinants of early mortality after ICH [28, 29]. Conversely, infections have a negative
influence on stroke outcome especially in the subacute phase, and therefore, patients that did
not experience early clinical deterioration and death are at higher risk of developing a
nosocomial infection [28, 29]. In line with this Katzan and Colleagues showed that stroke-
associated pneumonia increased mortality after exclusion of patients dying or receiving
withdrawal of care within 72 h from onset [30], again suggesting that infections have a
greater influence on late rather than early mortality after stroke.
It appears, therefore, plausible that an increased susceptibility to IC may be the link between
AL and worse outcome in patients hospitalized for ICH.
Morotti et al.
Page 5
Neurocrit Care
. Author manuscript; available in PMC 2018 April 01.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript