Content uploaded by Giedrius Barauskas
Author content
All content in this area was uploaded by Giedrius Barauskas
Content may be subject to copyright.
135
C-reactive protein in early prediction of pancreatic necrosis
Giedrius Barauskas, Saulius Švagždys, Almantas Maleckas
Clinic of Surgery, Kaunas University of Medicine Hospital, Lithuania
Key words: acute pancreatitis, pancreatic necrosis, C-reactive protein.
Summary. Aim of the study – to determine relation between the C-reactive protein and
pancreatic necrosis, and to estimate the prognostic value of C-reactive protein in early
diagnosis of pancreatic necrosis.
Material and methods. During 2001, 78 patients with acute pancreatitis were included
in the study. The clinical data, diagnostic procedures, and laboratory values were ana-
lyzed. According to severity of the disease patients were divided into two groups. Group I
consisted of 17 patients with necrotic pancreatitis, group II – of 61 patients with pancreatic
edema. Contrast-enhanced computed tomography scan was used to diagnose pancreatic
necrosis with subsequent fine-needle aspiration for microbiological evaluation. C-reactive
protein concentration in serum was measured on day 1, 2, 3, 5, 7 and 9 after admission.
The sensitivity, specificity, positive and negative predictive values for different C-reactive
protein concentration cut-off (100–150 mg/l) were calculated. Average C-reactive protein
values were compared between groups by t test for unpaired data. The difference was as-
sumed statistically significant when p<0.05.
Results. There was no significant difference in demographic data between the groups.
Necrosis of the pancreas was demonstrated on computed tomography scan in 17 cases. The
highest C-reactive protein values were detected on day 3 in group I patients. The difference
of average C-reactive protein concentration was significant between groups on all days
except day 7. The highest sensitivity and negative predictive value (94.1% and 95.7% re-
spectively) was obtained for C-reactive protein cut-off at 110 mg/l.
Conclusions. The results of our study show the C-reactive protein values increase sig-
nificantly in early stages of necrotic pancreatitis. C-reactive protein is an important prog-
nostic marker of pancreatic necrosis with the highest sensitivity and negative prognostic
value given the cut-off is 110 mg/l. The patients with C-reactive protein below 110 mg/l are
at low risk to develop pancreatic necrosis.
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt
Correspondence to S. Švagždys, Clinic of Surgery, Kaunas University of Medicine, Eivenių 2, 3007 Kaunas, Lithuania
E-mail: dak_sam@yahoo.com
Introduction
The majority of patients with acute pancreatitis
present with mild, uneventful course of the disease,
however, in 20-25% of cases disease may take a seri-
ous course with severe local and systemic complica-
tions and there is considerable mortality. In patients
with severe necrotic pancreatitis organ failure is com-
mon and may occur in the absence of infection. In the
natural course of the disease, infection occurs in a
considerable number of these patients. According to
Atlanta definition infected pancreatic necrosis is de-
fined as the presence of bacteria in diffuse or focal
areas of intrapancreatic or extrapancreatic necrotic
tissue (1). It is assumed that the most important prog-
nostic factor of acute pancreatitis is the development
of pancreatic necrosis with subsequent risk of pan-
creatic infection, multiple organ dysfunction, and death.
Of patients who die of acute pancreatitis, more than
60% of deaths are due to septic complications (2).
Conservative treatment of necrotizing pancreatitis is
associated with favorable results. There seems to be
evidence that prophylactic antimicrobial therapy may
reduce the rate of pancreatic infection and thus has a
positive impact on the mortality rate associated with
this disease. It is important to stratify patients early
for induction of antibiotic treatment (3); therefore it is
very important to use some marker, which could en-
able to diagnose the pancreatic necrosis at the very
onset of the disease (4). Ideally it should have high
sensitivity and positive prognostic value, and diagnose
pancreatic necrosis early (during the first 48 hours).
The test should also be readily available in every clini-
136
cal laboratory, and it should be cheap and impersonal
(5). The aim of the study was to determine the prog-
nostic value of C-reactive protein (CRP) in early pre-
diction of pancreatic necrosis.
Material and methods
We analyzed case records of the patients with
acute pancreatitis, managed during the period of 2001–
2002 at the Department of Surgery, Kaunas Univer-
sity of Medicine Hospital. Acute pancreatitis was di-
agnosed according to the clinical symptoms and el-
evation of serum a-amylase more than three times.
CRP was measured every day for 5 consecutive days
as well as on day 7 and day 9 after admission. When
CRP value exceeded 120 mg/l or clinical picture of
severe acute pancreatitis was present, the contrast-
enhanced computed tomography (CT) scan was per-
formed. All the data was included into the database
created for this study. The patients were divided into
two groups: group I – patients with necrotic pancre-
atitis (NP), group II – patients with mild pancreatitis
(pancreatic edema – PE). The patients were included
into group I after demonstration of pancreatic necro-
sis on CT scan or at surgery. Fine needle aspiration
(FNA) was practiced in patients with pancreatic ne-
crosis and when infection was demonstrated, surgical
drainage of infected pancreatic necrosis was per-
formed. Patients with sterile pancreatic necrosis had
no surgery. Student’s t test for independent data was
used to compare concentrations of CRP between
groups. Sensitivity (S), specificity (SP), positive prog-
nostic value (PPV), and negative prognostic value
(NPV) for CRP cut-off concentrations from 100 to
150 mg/l were calculated to clarify the best cut-off
concentration for prognosis of pancreatic necrosis.
Results
Seventy-eight patients were treated because of
acute pancreatitis from January 2001 to January 2002.
None of the patients had attacks of acute pancreati-
tis previously. There were 17 (21.8%) patients with
NP and 61 (78.2%) patients with PE (Table 1). The
average age of patients in both groups was similar.
Men predominated in both groups, but the difference
was negligible in the NP group (52.9 vs. 47.8). Gall-
stones and alcohol were shown to have induced mild
pancreatitis in 24.6% and 29.6% of cases respec-
tively. Etiologic factor for necrotic pancreatitis in
52.9% of patients was gallstones. Evident causative
factor was not revealed in one third of patients in
both groups.
The contrast-enhanced CT scan was performed
in 27 cases, so all patients with pancreatic necrosis
were examined, some of them repeatedly. Low vol-
ume necrosis (<30%) was present in 10 (58.8%), 30-
50% necrosis – in 2 (11.8%), and subtotal necrosis
(>50%) – in 5 (29.4%) cases. The patients with ne-
crosis of the pancreas compounded 63% of those who
underwent CT scan. Six patients with pancreatic ne-
crosis underwent surgery: three of them on the first
day of hospitalization because of the signs of peritoni-
tis and uncertain diagnosis, suggestive of viscus per-
foration, and 3 – in the later course of the disease
after FNA demonstrated pancreatic and/or
peripancreatic necrosis.
The average concentrations of CRP in both groups
are shown in the Figure. Mean values of CRP dif-
fered in the groups significantly from the day of hospi-
talization except for day 7 (Table 2).
Sensitivity, specificity, positive, and negative
predictive values for various CRP cut-off
concentrations were determined (Table 3). With CRP
cut-off value of 100 and 110 mg/l all the parameters
were equal. Increasing CRP cut-off values resulted in
significant decrease of sensitivity and negative
predictive value, whereas only slight increase of
specificity and positive predictive value was
demonstrated.
Table 1. The characteristic of the patients in groups
FeatureGroup I (nekrotic pancreatitis) Group I (edemic pancreatitis)
Case number (n)1761
Male9 (52.9 %)37 (60.7 %)
Female8 (47.8 %)24 (39.3 %)
Average age (years)55.649.1
Etiology:alcohol17.6 %29.6 %
gallstones52.9 %24.6 %
other29.4 %36 %
Giedrius Barauskas, Saulius Švagždys, Almantas Maleckas
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt
137
Discussion
The concept of conservative management of ne-
crotizing pancreatitis originates from our understand-
ing of two-phase pattern in the natural course of the
disease. The first two weeks are characterized by a
systemic inflammatory response syndrome, which is
maintained by release of inflammatory mediators (6).
Secondary pancreatic infection develops usually in the
second phase of the disease, leading to multiple organ
failure with a twelve-fold increase in mortality rate
(4). As infection of the necrotic pancreas has been
considered a secondary phenomenon it should be at
least theoretically preventable by proper antibiotic
therapy. Early diagnosis of pancreatic necrosis is ex-
tremely important, as it is obvious that a beneficial ef-
fect of antibiotic treatment is limited to patients with
acute necrotizing pancreatitis (7). There is data show-
ing that pancreatic necrosis develops during the first
48–72 hours from the onset of acute pancreatitis (8).
Markers for necrosis are looked for among the sub-
stances that reach their peak concentrations in serum
or urine during the first 24–48 hours. The most often
explored are CRP, granulocyte elastase (9), tumor
necrosis factor (TNF) (10), interleukin 6 (6), a
1
–an-
titrypsin (11), trypsinogen (12), pancreatic ribonuclease
(13), trypsinogen activating peptide (TAP) (14), car-
boxypeptidase B activating peptide (CAPAP) (15),
human pancreas-specific protein/procarboxypeptidase
B (hPASP/PCPB) (16), and serum amyloid A (17, 18).
While CRP is readily available in all clinical laborato-
ries, all other discussed parameters are not. Their use
is still within the limits of clinical studies. The clinical
use of these tests in the present form is limited due to
drawbacks in terms of test performance and cost fac-
tors.
The CRP is non-specific mediator of inflamma-
tion, produced in hepatocytes. Its production and
excretion is stimulated by interleukin 1 and 6. The
CRP is considered to be quite a late indicator of pan-
creatic necrosis with peak concentrations in blood
serum detected after 72 hours (19). Similarly our
results revealed the highest average CRP concen-
tration in the group of patients with necrotizing pan-
creatitis on the third day after admission. It should
be mentioned that already on the first day of the dis-
ease average concentration of CRP differed statisti-
Table 2. Comparison of average CRP values between groups (necrotic pancreatitis vs pancreatic
edema)
NPPE
1159.833.929.4–222.30.014
2162.839.339.2–207.70.008
3377.7139.1118.4–358.10.001
5283.4107.769.0–282.50.005
7199.0104.0–101.3–291.10.3
9180.857.028.8–278.80.017
NP – nekrotic pancreatitis, PE – pancreatic edema, CRP – C-reactive protein.
Number of days from Average CRP mg/l95 % CIP value
beginning of disease
Table 3. Sensitivity, specificity, positive and negative prognostic values for various CRP cut-off
concentrations
CRB koncentracijaSensitivitySpecificityPositive PVNegative PV
(mg/l)(%)(%)(%)(%)
10094.164.757.195.7
11094.164.757.195.7
12088.264.755.691.7
13082.467.65688.5
14076.576.561.986.7
15070.676.56083.9
CRP – C-reactive protein, PV – prognostic value.
C-reactive protein in early prediction of pancreatic necrosis
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt
138
cally significantly between the groups of edematous
and necrotic pancreatitis. The only explanation for
this phenomenon might be delayed admission of pa-
tients with acute pancreatitis. Anyway this enabled
us to use CRP as the early marker of pancreatic ne-
crosis, having high sensitivity and negative prognos-
tic value. As some patients are not hospitalized on
the first day of the disease, it is very important that
elevated CRP remains long enough in the course of
the disease. On the day seven the difference of av-
erage CRP between groups was not statistically sig-
nificant. Possible explanation might be too small num-
ber of CRP tests performed in the PE group, but any-
way, the trend of increased CRP concentrations in
NP group is obvious. Similar persistence of elevated
CRP values up to 14-16 days was reported by M.
Buechler et al (20).
After years of discussion there is still no consen-
sus of the cut-off value for CRP test. Quite a wide
range of CRP cut-off values between 120 and 210
mg/l have been discussed by various authors (21, 22).
The lower the chosen cut-off, the higher is sensitiv-
ity and negative prognostic value of the test. On the
other hand the higher cut-off values are associated
with increasing specificity and positive prognostic
value.
It must be determined what is more important in
clinical setting – to attribute patients with edematous
pancreatitis to the group of necrotic pancreatitis or
vice versa. The chosen cut-off should let us diagnose
pancreatic necrosis with the highest accuracy, paying
no much attention to false positive result.
Hyperdiagnosis of pancreatic necrosis will result in an
early treatment with antibiotics in a subset of patients
who really do not have necrosis of the pancreas. This
would match the CRP cut- off at 110 mg/l according
to our data, showing the highest sensitivity and nega-
tive prognostic value, 94.1% and 95.7% respectively.
There is no doubt that in patients with edematous pan-
creatitis infectious complications are rare. Adminis-
tration of antibiotics will not have any positive impact
on the patient’s course. With regard to the develop-
ment of bacterial resistance, it might be even harmful.
Although with certain drawbacks, the latter approach
ensures all the patients with existing pancreatic ne-
crosis will receive beneficial treatment with antibiot-
ics (7).
The CRP test with its diagnostic characteristics is
far from ideal test for detection of pancreatic necro-
sis. According to our results, the CRP cut-off at 110
mg/l makes it possible to determine the group of pa-
tients with the lowest risk of pancreatic necrosis with
almost 96% overall accuracy. The latter group will
need neither antimicrobial therapy nor CT scan. Thus
the contrast-enhanced abdominal CT scan, the most
informative imaging technique up to date is reserved
to the cohort of high-risk patients to develop pancre-
atic necrosis (23).
377.7
159.8
180.8
283.4
199
162.8
33.9
139.1
107.7104
57
39.3
0
50
100
150
200
250
300
350
400
Laikas nuo susirgimo pradžios
CRP mg/l
NP EP
Fig. The range of mean CRP concentration in the groups
Giedrius Barauskas, Saulius Švagždys, Almantas Maleckas
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt
Number of days from the beginning of disease
PE
139
References
1.Bradley EL. A clinically based classification system for acute
pancreatitis. Arch Surg 1993;128:586-9.
2.Renner IG, Savage WT, Pantoia JL, Renner VJ. Death due to
acute pancreatitis: a retrospective analysis of 405 autopsy
cases. Dig Dis Sci 1985;30:1005-18.
3.Sharma VK, Howden CW. Prophylactic antibiotic administ-
ration reduces sepsis and mortality in acute necrotizing pan-
creatitis: a meta-analysis. Pancreas 2001;22:28-31.
4.Pezzilli R, Billi P, Miniero R, Fiocchi M, Cappelletti O,
Morselli-Labate AM, et al. Serum interleukin-6, interleukin-
8, and beta 2-microglobulin in early assessment of severity of
acute pancreatitis. Comparison with serum C-reactive pro-
tein. Dig Dis Sci 1995;40:2341-8.
5.Dervenis C, Johnson CD, Bassi C, et al. Diagnosis, objective
assessment of severity, and management of acute pancreatitis.
Int J Pancreatol 1999;25:195-210.
6.Buchler MW, Gloor B, Muller CA, Friess H, Seiler CA, Uhl
W. Acute necrotizing pancreatitis: treatment strategy accor-
ding to the status of infection. Ann Surg 2000;232:619-26.
7.Isenmann R, Beger HG. Bacterial infection of pancreatic
necrosis: role of bacterial translocation, impact of antibiotic
treatment. Pancreatology. 2001;1:79-89.
8.Isenmann R, Buchler M, Uhl W, Malfertheiner P, Martini M,
Beger HG. Pancreatic necrosis: an early finding in severe acute
pancreatitis. Pancreas 1993;8:358-61.
9.Dominguez-Munoz JE, Carballo F, Garcia MJ, de Diego JM,
Rábago L, Simón MA, de la Morena J. Clinical usefulness
of polymorphonuclear elastase in predicting the severity of
C reaktyviojo baltymo reikšmė ankstyvajai kasos nekrozės diagnostikai
Giedrius Barauskas, Saulius Švagždys, Almantas Maleckas
Kauno medicinos universiteto klinikų Chirurgijos klinika
Raktažodžiai: C reaktyvusis baltymas, kasos nekrozė, ūminis pankreatitas.
Santrauka. Darbo tikslas. Nustatyti C reaktyviojo baltymo koncentracijos priklausomumą nuo kasos nekrozės
ir nustatyti šio baltymo prognostinę reikšmę ankstyvajai kasos nekrozės diagnostikai.
Tyrimo medžiaga ir metodai. Atlikome Kauno medicinos universiteto klinikų Chirurgijos skyriuje ligonių,
gydytų 2001 metais dėl ūminio pankreatito, ligos istorijų analizę. Ligonius suskirstėme į dvi grupes: pirmąją
grupę sudarė 17 ligonių, sirgusių ūminiu nekroziniu pankreatitu; antrąją – 61 ligonis, kuriam buvo diagnozuotas
ūminis edeminis pankreatitas. Analizavome klinikinius, diagnostinių ir laboratorinių tyrimų duomenis. Kasos
nekrozei diagnozuoti naudojome pilvo organų kompiuterinę tomografiją su intraveniniu kontrastiniu tirpalu.
Nustačius kasos nekrozę, atliktos nekrozinio audinio punkcijos (plona adata) ir mikrobiologinis punktato įvertinimas.
C reaktyviojo baltymo koncentracija kraujo serume buvo tiriama 1, 2, 3, 5, 7 ir 9 gydymo stacionare parą.
Apskaičiuotas pastarojo tyrimo jautrumas, specifiškumas, teigiama ir neigiama prognostinė vertė parinkus įvairias
(nuo 100 iki 150 mg/l) C reaktyviojo baltymo koncentracijos ribas. Šio baltymo koncentracijos vidurkiams
grupėse palyginti naudotas t testas neporinėms imtims. Skirtumas laikytas statistiškai reikšmingas, kai p<0,05.
Rezultatai. Demografiniai rodikliai abiejų tiriamųjų grupių buvo panašūs. Kasos nekrozė, atlikus pilvo organų
kompiuterinę tomografiją, nustatyta 17 ligonių. Aukščiausios C reaktyviojo baltymo reikšmės nustatytos trečiąją
gydymo stacionare parą. Statistiškai reikšmingi C reaktyviojo baltymo koncentracijų vidurkių skirtumai grupėse
buvo 1–5 ir 9 dienomis. Aukščiausias jautrumas ir neigiama prognostinė C reaktyviojo baltymo reikšmė (94,1 ir
95,7 proc., atitinkamai), nustatyti parinkus 110 mg/l koncentracijos ribą.
Išvados. Šio tyrimo duomenimis, ligonių, sergančių ūminiu nekroziniu pankreatitu, C reaktyviojo baltymo
koncentracija būna žymiai didesnė. Tai yra svarbus ir paprastai nustatomas ankstyvasis kasos nekrozės žymuo.
Parinkus 110 mg/l C reaktyviojo baltymo koncentracijos ribą, tyrimas pasižymi dideliu jautrumu bei neigiama
prognostine verte. Ligoniams, kuriems C reaktyviojo baltymo koncentracija ligos eigoje neviršija 110 mg/l, yra
minimali kasos nekrozės rizika.
Adresas susirašinėjimui: S. Švagždys KMUK Chirurgijos klinika, Eivenių 2, 3007 Kaunas
El. paštas: dak_sam@yahoo.com
C-reactive protein in early prediction of pancreatic necrosis
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt
Conclusions
The results of our study show that the CRP values
increase significantly in early stages of necrotic pan-
creatitis. It is an important prognostic marker of pan-
creatic necrosis with the highest sensitivity and nega-
tive prognostic value given the cut-off is at 110 mg/l.
The patients with CRP values below 110 mg/l are at
low risk to develop pancreatic necrosis.
140
acute pancreatitis: results of a multicentre study. Br J Surg
1991; 78:1230-4.
10.Paajainen H, Laato M, Jaakkola M, Pulkki A, Niinikoski J,
Nordback I. Serum tumour necrosis factor compared with C-
reactive protein in the assessment of severity of acute panc-
reatitis. Br J Surg 1995;82:271-3.
11.Hedström J, Sainio V, Kemppainen E, Haapiainen R, Kivi-
laakso E, Schröder T, et al. Serum complex of trypsin 2 and
alpha 1 antitrypsin as diagnostic and prognostic marker of
acute pancreatitis: clinical study in consecutive patients. Br
Med J 1996;313:333-7.
12.Sainio V, Puolakkainen P, Kempainnen E, et al. Serum tryp-
sinogen-2 in the prediction of outcome in acute necrotizing
pancreatitis. Scand J Gastroenterol 1996;31:818-24.
13.Kemmer TP, Malfertheiner P, Buchler M, Kemmer ML,
Ditschuneit H. Serum ribonuclease activity in the diagnosis
of pancreatic disease. Int J Pancreatol 1991;8:23-33.
14.Tenner S, Fernandez-del Castillo C, Warshaw A, et al. Urinary
trypsinogen activation peptide (TAP) predicts severity in
patients with acute pancreatitis. Int J Pancreatol 1997; 21:105-
10.
15.Müller CA, Appelros S, Uhl W, Büchler MW, Borgström A.
Serum levels of procarboxypeptidase B and its activation pep-
tide in patients with acute pancreatitis and non-pancreatic
diseases. Gut 2002;51:229-35.
16.Rau B, Cebulla M, Uhl W, Schoenberg MH, Beger HG. The
clinical value of human pancreas-specific protein procarboxy-
peptidase B as an indicator of necrosis in acute pancreatitis:
comparison to CRP and LDH. Pancreas. 1998;17:134-9.
17.Rau B, Steinbach G, Baumgart K, Gansauge F, Grunert A,
Beger HG. Serum amyloid A versus C-reactive protein in
acute pancreatitis: clinical value of an alternative acute-phase
reactant. Crit Care Med 2000;28:736-42.
18. Mayer JM, Raraty M, Slavin J. Serum amyloid A is a better
early predictor of severity than C-reactive protein in acute
pancreatitis. Br J Surg 2002;89:163-71.
19.Triester SL, Kwodley KV. Prognostic factors in acute pancrea-
titis. J Clin Gastroenterol 2002;34:167-76.
20.Buchler M, Malfertheiner P, Schoetensack C, et al. Sensitivity
of antiproteases, complement factors and C-reactive protein
in detecting pancreas necrosis. Int J Pancreatol 1986;1:227-
33.
21.Wilson C, Heads A, Shenkin A, Imrie CW. C-reactive protein,
antiproteases and complement factors as objective markers
of severity in acute pancreatitis. Br J Surg 1989;76:177-81.
22.Larvin M, McMahon MJ. APACHE-II score for assessment
and monitoring of acute pancreatitis. Lancet 1989;2:201-5.
23.Biezenbos van de A R, Kruyt P M, Bosscha K. Added value
of CT criteria compared to the clinical SAP score in patient
with acute pancreatitis. Abdomi Imaging 1998;23:622-6.
Received 10 January 2003, accepted 20 September 2003
Giedrius Barauskas, Saulius Švagždys, Almantas Maleckas
MEDICINA (2004) Vol. 40, No. 2 - http://medicina.kmu.lt