Circulating cytokines as markers of systemic inflammatory response

in severe community-acquired pneumonia

Anton A. Igonin,
a,
*
Victor W. Armstrong,
b
Maria Shipkova,
b
Natalya B. Lazareva,
a
Vladimir G. Kukes,
a
and Michael Oellerich
b
a
Department of Clinical Pharmacology, I.M. Sechenov Medical Academy, Moscow, Russia
b
Department of Clinical Chemistry, George-August University, Goettingen, Germany
Received 1 August 2003; received in revised form 4 November 2003; accepted 4 November 2003
Abstract
Objectives: To assess the influence of empirical antibacterial therapy on systemic inflammatory response in patients with severe
community-acquired pneumonia (CAP).
Material and methods: Thirty consecutive patients with CAP meeting systemic inflammatory response syndrome (SIRS) criteria were
recruited into this study. Blood samples for measurement of interleukin-6 (IL-6), interleukin-8 (IL-8), interleukin-10 (IL-10) and C-reactive
protein (CRP) concentrations were drawn on days 1, 3, 5, 7 and 10. For analysis, these patients were divided into two subgroups according to
British Thoracic Society (BTS) severity score and to clinical response to the initial antibacterial therapy.
Results: In the group with severe CAP (n= 15), serum concentrations of IL-6 ( P = 0.0001), IL-8, ( P = 0.001), IL-10 ( P = 0.0001) and
CRP ( P = 0.0001) were significantly higher compared to patients from the non-severe group (n= 15). IL-6 presented with a sharp decrease
between days 1 and 3 in non-responders with severe CAP ( P = 0.004). IL-6 concentrations on day 1 were significantly associated with a
response to empirical antibacterial treatment by day 3.
Conclusion: Despite the absence of a clinical response to empirical antibacterial treatment as assessed by conventional clinical parameters
on day 3 in patients with severe CAP meeting SIRS criteria, there was a marked reduction in the degree of the systemic inflammatory
response as reflected by IL-6 levels.
D 2003 The Canadian Society of Clinical Chemists. All rights reserved.
Keywords: Community-acquired pneumonia; Antibacterial therapy; Interleukin 6
Introduction
According to the results of meta-analyses, the average
mortality rate for community-acquired pneumonia (CAP)
ranges from 1% in hospitalized and ambulatory patients
to 36.5% in patients who need to be admitted to an
intensive care unit [1,2]. Diverse molecular mechanisms
of inflammation and cellular damage have been implicat-
ed in the pathogenesis of severe pneumonia, including
those related to overt generation of cytokines. Several
studies demonstrated that serum TNFa, IL-6 and IL-8 are
involved in the pathogenesis of organ injury and death
from severe infection [35]. Marchant et al. [6] reported
that plasma IL-10 was increased in patients with circula-
tory shock of septic and nonseptic origin. The role of
TNFa gene polymorphism was shown in a study with
CAP patients where septic shock and respiratory failure
were strongly associated with a specific haplotype of the
studied gene. The inflammatory response to local infec-
tion such as pneumonia has been shown to be largely
compartmentalized [7]. However, significantly elevated
systemic concentrations of TNFa and IL-6 have been
reported in patients with severe pneumonia [8,9]. An
enhanced anti-inflammatory response was also found in
severe CAP [10]. The relationship between the degree of
0009-9120/$ - see front matter D 2003 The Canadian Society of Clinical Chemists. All rights reserved.
doi:10.1016/j.clinbiochem.2003.11.001
Abbreviations: CAP, community-acquired pneumonia; ACCP/SCCM,
American College of Chest Physicians/Society of Critical Care Medicine;
BTS, British Thoracic Society; CRP, C-reactive protein; HR, heart rate;
ICU, intensive care unit; IL-6,-8,-10, interleukin-6,-8,-10; L, leukocytes;
ROC, receiver operating curve; RR, respiratory rate; SAPS II, simplified
acute physiology score; SIRS, systemic inflammatory response syndrome;
T, temperature; TNFa, tumor necrosis factor a.
* Corresponding author. Department of Clinical Pharmacology, I.M.
Sechenov Medical Academy, Bolshoy Tishinsky per. 43 App. 16, 123557
Moscow, Russia. Fax: +7-95-25-49-285.
E-mail address: igonin@rambler.ru (A.A. Igonin).
Clinical Biochemistry 37 (2004) 204209
illness severity assessed by clinical features and the
profile of the cytokines response in patients with CAP
has not been clearly established.
In this study, we have investigated the influence of
empirical antibacterial therapy on the illness severity as
assessed by the British Thoracic Society (BTS) score and on
the degree of systemic inflammatory reaction in patients
with CAP meeting SIRS criteria.
Materials and methods
The presence of clinical symptoms on admission, such as
pyrexia, productive or dry cough, dyspnea, pleuritic pain and
auscultatory findings (altered breath sounds, crepitations and/
or localised rales) corresponding to an acute infiltrate on a
chest radiograph were considered as indicators of pneumonia.
CAP patients met two or more of the ACCP/SCCM criteria
(1992) for SIRS: (temperature > 38 jC or < 36 jC, heart rate
> 90/min, respiratory rate > 20/min, PaCO
2
< 32 Torr
[ < 4.27 kPa], leukocytosis [>12000/Al] or leukopenia
[ < 4000/Al], or immature forms > 10%). Patients were further
assessed according to the BTS score. Major prognostic
features were: serum urea > 7.0 mmol/l, BP V 90/60 mm
Hg, respiratory rate z 30 min, new mental confusion; SaO
2
< 92% and bilateral or multilobar involvement on chest X-
ray were considered as additional parameters [8]. Patients
having two core features or 1 core plus 1 additional parameter
were allocated to group I (i.e., severe CAP). Those patients
who did not satisfy these criteria were included in group II.
Illness severity was also assessed using the SAPSII score
[11]. Exclusion criteria from the study were nosocomial
pneumonia, presence of immunosuppression, chronic heart
failure (NYHA III IV) or verified malignant disease. The
assessment of response to the current empirical therapy on the
3rd day of the study relied on clinical judgement and routine
markers such as respiratory rate (RR), heart rate (HR),
temperature (T) and leukocytes (L). Patients from both
groups were classified as responders or non-responders
according to their clinical reaction throughout 72 h of the
therapy (Fig. 1). A temperature higher than 38.3 jC on day 3
served as the key parameter indicating therapeutic failure
(non-responder) since it was among factors associated with
higher mortality risk [12].
All patients received empirical antibacterial treatment at
the beginning of the study according to the BTS recom-
mendations. Initial choice of antibiotics included amino-
penicillins, third-generation cephalosporins plus macrolides
or fluorquinolones. In all non-responders, the empirical
antibacterial regimen was changed on day 3. The selection
of the new antibacterial scheme relied on the results from
microbiological analysis and sensitivity tests if they proved
to be positive. Fluid infusion was prescribed if signs of
hypoperfusion occurred. Additional oxygen administration
was applied for PaO
2
< 60 mm Hg or SaO
2
< 90%.
Vital signs were documented and haematological and
biochemical assays were carried out on days 1, 3, 5, 7 and
10. Blood samples for determination of serum concentra-
tions of CRP and the cytokines TNFa, IL-6, IL-8 and IL-10
were also drawn on the above days. The first blood sample
was obtained before the initiation of standard antibacterial
treatment. Serum concentrations of CRP were determined
Fig. 1. Temperatures recorded on days 1 and 3 in patients allocated to (A) group I or (B) group II. k: Patients who had clinically responded to treatment by day
3; U: patients who did not clinically respond to treatment by day 3; - - - - - -: threshold temperature (38.3 jC) between responders and non-responders.
Table 1
General patient data. Patients were allocated to group I (i.e., severe CAP) or
group II according to the British Thoracic Society score
Group I Group II
Number of patients 15 15
Age (years) 50 F 15 47 F 17
Gender (M/F) 9/6 8/7
SAPS II scoreday 1 (points)* 27 (1834) 13 (618)
SAPS II mortality
riskday 1 (%)*
7.4 (2.9315) 1.5 (0.462.9)
Patients fulfilling all 4 SIRS
criteria on day 1

13/15 0/15
Two or multilobar involvement 11/15 5/15
Non-responders/Responders 10/5 4/11
*P (MannWhitney) < 0.05: group I versus group II.

P (Fisher) < 0.05: group I versus group II.

A.A. Igonin et al. / Clinical Biochemistry 37 (2004) 204209 205
with an immunoturbidometric assay (Greiner, Flacht, Ger-
many) on a Hitachi 917 analyser. Concentrations of the
cytokines were measured using quantitative enzyme-linked
immunosorbent assays (Diagnostics Products Corporation,
Los Angeles, USA).
All patients gave their informed consent to participate in
this study. The study was approved by the Institutional
Ethics Committee.
Statistical comparisons were made by MannWhitney U
and Fisher tests and associations between variables were
assessed by the Spearman coefficient. Receiver operating
curve (ROC) analysis [13] was applied to estimate the
ability of the individual test to discriminate between the
two groups of patients who differed in their response to the
therapy on day 3.
Results
There were 30 patients with CAP meeting SIRS criteria
included in the study (Table 1). Fifteen patients were
classified as having severe CAP on admission (group I),
while the remaining 15 patients (group II) had non-severe
CAP. Eight patients from group I and 3 patients from group
II had received antibiotics before admission. Antibacterial
treatment as monotherapy was initiated in 67% (10/15) of
non-severe patients and in 13% (2/15) of patients from
group I. According to clinical data, 10 patients from group
I and four patients from group II did not show a positive
reaction to the initial treatment and were classified as non-
responders. Seven (7/30) patients (three from group I and
four from group II) were found to have chronic co-morbid-
ities such as coronary heart disease, cerebral vascular
disease, diabetes mellitus and COPD. None of the included
patients received drugs, which have the potential to produce
clinically significant anti-inflammatory effects either before
or after admission. Microbiological analysis was positive
only in 10/30 (33%) of patients. Three cases of gram-
negative bacteria (P. aeruginosa and K. pneumonia) were
found in sputum culture in group I. Staphylococcus pneu-
monia was identified in sputum of six patients and Staph-
ylococcus aureus in one patient.
None of the patients with non-severe CAP displayed all
four SIRS criteria on admission. RR and HR were the only
criteria, which showed significant differences between the
groups (Table 2). Patients from group I showed involvement
of at least one organ system other than respiratory; episodes
of arterial hypotension with systolic blood pressure less than
90 mm Hg and renal failure (blood urea nitrogen > 7 mmol/
L or serum creatinine z 2 Amol/l) accompanied by meta-
bolic acidosis, as well as signs of mental confusion were
among the most common signs of organ involvement in
group I. There was a significant difference in the severity
state according to SAPS II between groups I and II on the
day of admission ( P < 0.0001). The mortality risk in group
I on the first day of the study was 7.4% (2.9315%)
according to SAPS II. Two cases of mortality which
occurred after the 10-day study period were recorded in
group I. In the first case, a positive initial response to the
empirical antibacterial treatment was followed by the reac-
tivation of local and systemic inflammation because of
superinfection developed during mechanical ventilation
resulting in lung gangrene and progressive respiratory
failure. Aneurysm rupture with parenchymal cerebral hae-
morrhage was the reason for the second lethal outcome.
Serum IL-6, IL8 and IL-10 concentrations were signifi-
cantly higher in group I than in group II on admission as
were the concentrations of the acute phase protein CRP
(Table 3). TNFa was not found to be elevated on admission
Table 2
SIRS criteria on admission
SIRS criteria Group I Group II P
Temperature (jC) 39 (35.239.7) 38.5 (37.440) 0.34
HR (bpm) 112 (100160) 100 (100130) 0.02
RR (pm) 28 (2440) 22 (2032) 0.0005
Leukocyte (mm
3
) 10.5 (1.426) 11.5 (615) 0.8
Bands (%) 15.5 (246) 9 (158) 0.3
Table 3
Median values and ranges for cytokines and CRP during days 110 of the study period
IL-6 (Ag/l) IL-8 (Ag/l) IL-10 (Ag/l) CRP (mg/l)
Day/Group I II I II I II I II
1 Median 1510 108 283 40 82 12 230 126
Range 14942,452 4454 273864 13730 8882 4155 134495 30391
n 15 15 15 15 15 15 15 15
P 0.0001 0.02 0.001 0.006
3 Median 350 50 72 34 25 10 167 103
Range 45928 4345 231000 23614 5132 4151 76289 35213
n 14 15 14 15 14 15 14 15
P 0.002 0.3 0.01 0.0006
10 Median 79 45 120 42 14 9 50 38
Range 4980 4242 341915 23272 377 315 11148 2108
n 11 9 11 9 11 9 11 9
P 0.12 0.9 0.04 0.26
A.A. Igonin et al. / Clinical Biochemistry 37 (2004) 204209 206
either in group I (11Ag/l; 629 Ag/l) or group II (8.5 Ag/l;
627 Ag/l, P < 0.05).
To compare the ability of these tests to discriminate
between the patients responding to the empirical treatment
on day 3, ROC curves were constructed. Of the variables
tested, only IL-6 was able to significantly ( P = 0.03)
discriminate between the patients who responded to the
antimicrobial therapy and those patients who did not re-
spond (Table 4).
The response to the empirical antibacterial treatment did
not significantly depend on the CAP severity as assessed by
the BTS score ( P = 0.06, Table 1). A significant decrease in
T ( P = 0.0003), RR ( P = 0.02) and HR ( P = 0.03) reflecting
the positive response to antibacterial treatment was observed
among responders in the first 3 days of treatment. These
clinical parameters did not significantly change during the
first 72 h in non-responders from both groups, and initial
antibacterial therapy was therefore subjected to revision.
Despite the absence of clinical reaction in severe non-
responders, IL-6 ( P = 0.004) concentrations had markedly
decreased by day 3 after admission and initiation of the
treatment (Fig. 2). The concentrations of IL-8 ( P = 0.16)
and CRP ( P = 0.27) in severe patients assigned as non-
responders did not significantly differ on days 1 and 3,
whereas IL-10 ( P = 0.05) presented with a marginally
significant decrease on day 3. A further decline of IL-6
concentration in non-responders with severe CAP was
observed after day 3 when the second line of antibiotics
was applied (Fig. 2).
There was a highly significant correlation between the
SAPS II score and IL-6 (r = 0.61; P < 0.0001), IL-10 (r =
0.41; P < 0.01) or CRP (r = 0.57; P < 0.0001) during the
first 3 days of the study. The levels of IL-6 also showed a
positive correlation with IL-10 (r = 0.61; P < 0.0001).
Discussion
In the present study, we found that stratification of CAP
patients according to BTS could help to identify those
individuals with a severe state due to systemic inflammatory
response. These patients possibly require additional care, as
for patients with confirmed diagnosis of severe sepsis. The
data obtained by serial measurements confirmed that
patients with severe CAP meeting SIRS criteria had signif-
icantly increased levels of the circulating cytokines IL-6, IL-
8, IL-10 as well as CRP in comparison with non-severe
patients.
In the earlier study by Dehoux et al. [7] where lung and
serum cytokine levels were analysed in patients with unilat-
eral CAP, inflammatory response was localised since the
cytokine production was limited to the altered lung. The
levels of systemic pro-inflammatory (IL-6) and anti-inflam-
matory (IL-10) mediators measured in severe patients with
CAP and SIRS on admission were also increased in compar-
ison with those without SIRS [9,10]. Monton et al. [8] studied
serum and lung cytokines response to infection in patients
with severe pneumonia. TNFa, interleukin-1beta and IL-6
were higher in the bronchoalveolar lavage fluid than in the
corresponding serum. However, patients with severe pneu-
monia had a significantly greater elevation of circulating
TNFa and IL-6 levels than critically ill patients without
infection. In contrast, we did not observe any difference in
TNFa concentrations between the two groups. This might be
explained by the predominant local tissue activity of this
cytokine. On the other hand, TNFa is characterised by a short
half-life and the peak concentrations could, therefore, have
been missed due to delayed admission.
Table 4
ROC curve analysis reflecting the power of variables obtained on admission
in patients from group I and group II to discriminate between responders
and non-responders to the empirical antimicrobial treatment
Variable AUC SE 95% CI P
IL-6 0.73 0.094 0.540.91 0.03
IL-10 0.65 0.103 0.440.85 0.17
CRP 0.61 0.107 0.410.83 0.28
IL-8 0.59 0.116 0.360.81 0.41
Fig. 2. Serum Il-6 concentrations on days 1, 3 and 5 in patients allocated to either (A) group I or (B) group II. The bold lines represent the median values. k:
Patients who had clinically responded to treatment by day 3; U: patients who did not clinically respond to treatment by day 3; *: P < 0.005; : P < 0.05.
A.A. Igonin et al. / Clinical Biochemistry 37 (2004) 204209 207
The current recommendations indicate that the estima-
tion of adequacy of empirical antibacterial treatment in
CAP relies on respiratory symptoms (cough, dyspnea),
fever, and peripheral leukocyte count [14]. These signs
should be assessed 4872 h after initial therapy is begun
[15]. Temperature changes are likely to be an early marker
of response. For example, in pneumococcal pneumonia,
duration of fever in young adults in case of adequate
treatment is limited to within 2.5 days [14], whereas only
12.550% display radiological resolution of pneumonia
after 2 weeks [16]. However, in the group of severe CAP
patients who did not show a clinically significant response
to the empirical antibacterial therapy throughout the first 3
days of the study, we observed a significant decline in
serum IL-6 concentrations. The lethal episode of a patient
with severe CAP who developed lung gangrene was ac-
companied with an initial decrease of IL-6 on day 3 and its
subsequent elevation on the 5th day, which may be
explained by the superinfection due to mechanical ventila-
tion. Serum IL-6 concentration was the most sensitive and
specific measurement for the assessment of response to the
treatment among other studied parameters since it had the
highest discriminatory power. Several studies indicated that
IL-6 levels might be predictive of death, development of
nosocomial infection and rapidity of onset of shock [17,18].
In patients with pneumonia IL-6 levels correlated with the
illness severity as well as with the degree of multiple
system organ failure [810,19]. It was previously reported
that CRP may serve as an earlier and more sensitive marker
of pneumonia than temperature and white cell count [20].
However, when reaction to the treatment on day 3 in the
group of severe non-responders was assessed, CRP con-
centrations did not show the same significant discriminato-
ry power as circulating IL-6. Our data show that
inflammatory response in CAP is counterbalanced by the
anti-inflammatory reaction which is reflected by an en-
hanced IL-10 level on admission, which tended to decline
after 72 h of treatment. A significant correlation with
systemic IL-6 is consistent with the studies of other authors
[9,21]. Since the dynamics of the IL-6, IL-8 and IL-10 in
non-severe patients did not display statistically significant
changes, these parameters cannot be used for the assess-
ment of treatment strategy in this group.
In conclusion, patients with severe CAP particularly
those fulfilling all SIRS criteria displayed elevated levels
of inflammatory markers reflecting an enhanced inflamma-
tory response. The worsening of CAP severity is accompa-
nied by the enhancement of the systemic inflammation,
which is best reflected by elevated IL-6 serum concentra-
tions. The range of IL-6 and CRP values indicates that non-
severe patients according to BTS score could have more
controlled and less severe systemic inflammatory response
induced by CAP. Despite the absence of a clinical response
to empirical antibacterial treatment, as assessed by conven-
tional clinical parameters on day 3, there was a marked
reduction in the degree of the systemic inflammatory
reaction in patients with severe CAP meeting SIRS criteria
as reflected by IL-6 levels. This study suggests that moni-
toring of IL-6 concentrations in plasma on days 13 after
admission may serve as an additional marker to assess the
response to empirical antimicrobial treatment in patients
with severe CAP meeting SIRS criteria.
Acknowledgment
We gratefully acknowledge the skillful technical assis-
tance of Christa Scholz.
References
[1] Fine MJ, Smith MA, Carson CA, et al. Prognosis and outcomes of
patients with community-acquired pneumonia. JAMA 1996;275:
13441.
[2] Requilme R, Torres A, El-Ebiary M, et al. Community-acquired pneu-
monia in the elderly. A multivariate analysis of risk and prognostic
factor. Am J Respir Crit Care Med 1996;154:14505.
[3] Pinsky MR, Vincent JL, Deviere J, Alegre M, Kahn RJ, Dupont E.
Serum cytokine levels in human septic shock: relation to multiple-
system organ failure and mortality. Chest 1993;103:56575.
[4] Casey LC, Balk RA, Bone RC. Plasma cytokine and endotoxin levels
correlate with survival in patients with the sepsis syndrome. Ann
Intern Med 1993;119:7718.
[5] Marty C, Misset B, Tamion F, Fitting C, Carlet J, Cavaillon JM.
Circulating interleukin-8 concentrations in patients with multiple or-
gan failure of septic and nonseptic origin. Crit Care Med 1994;22:
6739.
[6] Marchant A, Deviere J, Byl B, De Groote D, Vincent JL, Gold-
man M. Interleukin-10 production during septicemia. Lancet
1994;343:7078.
[7] Dehoux MS, Boutten A, Ostinelli J, et al. Compartmentalized cyto-
kine production within the human lung in unilateral pneumonia. Am J
Respir Crit Care Med 1994;150:7106.
[8] Monton C, Torres A, El-Ebiary M, Filella X, Xaubet A, de la Bella-
casa JP. Cytokine expression in severe pneumonia: a bronchoalveolar
lavage study. Crit Care Med 1999;27:174553.
[9] Glynn P, Coakley R, Kilgallen I, Murphy N, ONeill S. Circulating
IL-6 and IL-10 in CAP. Thorax 1999;54:515.
[10] Antunes G, Evans SA, Lordan JL, Frew AJ. Systemic cytokine levels
in community-acquired pneumonia and their association with disease
severity. Eur Respir J 2002;20:9905.
[11] Le Gall GR, et al. A new simplified acute physiology score (SAPS II)
based on European/North American multicenter study. JAMA 1993;
270(24):295763.
[12] American Thoracic Society. Guidelines for the initial management of
adults with community acquired pneumonia: diagnosis, assessment of
severity, and initial antimicrobial therapy. Am Rev Respir Dis
1993;148:141826.
[13] McNeil BJ, Keeler E, Adelstein SJ. Primer on certain elements of
decision making. N Engl J Med 1975;293:2115.
[14] Bartlett JG, Dowell SF, Mandell LA, et al. Fine Practice Guidelines
for the Management of Community-Acquired Pneumonia in Adults
Clinical Infectious Diseases 2000;31:34782.
[15] Huchon G, Woodhead M. Management of adult community-ac-
quired lower respiratory tract infection. Eur Respir Rev 1998;
8(61):391426.
[16] MacFarlane JT, Miller AC, Smith WH, et al. Comparative radio-
graphic features of community-acquired legionnaires disease, pneu-
A.A. Igonin et al. / Clinical Biochemistry 37 (2004) 204209 208
mococcal pneumonia, mycoplasma pneumoniae, and psittacosis.
Thorax 1984;39:2833.
[17] Martin C, Saux P, Mege JL, et al. Prognostic values of serum cyto-
kines in septic shock. Intensive Care Med 1994;20:2727.
[18] Damas P, Canivet JL, DeGroote D, et al. Sepsis and serum cytokine
concentrations. Crit Care Med 1997;25:40512.
[19] Igonin AA, Armstrong VW, Shipkova M, Kukes V, Oellerich M. The
monoethylglycinexylidide (MEGX) test as a marker of hepatic dys-
function in septic patients with pneumonia. Clin Chem Lab Med
2000;38(11):11258.
[20] Smith RP, Lipworth BJ, Cree IA, et al. C-reactive protein. A clinical
marker in community-acquired pneumonia. Chest 1995;108:128891.
[21] Taniguchi T, Koido Y, Aiboshi J, et al. Change in the ratio of interleu-
kin-6 to interleukin-10 predicts a poor outcome in patients with sys-
temic inflammatory response syndrome. Crit Care Med 1999;27:
12624.
A.A. Igonin et al. / Clinical Biochemistry 37 (2004) 204209 209