Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.

com/ on December 23, 2020 at India:BMJ-PG Sponsored.

Archives of Disease in Childhood, 1986, 61, 24-29

C reactive protein in the evaluation of febrile illness

A PUTFO, 0 RUUSKANEN, 0 MEURMAN, H EKBLAD, H KORVENRANTA,
J MERTSOLA, H PELTOLA, H SARKKINEN, M K VILJANEN, AND P HALONEN
Departments of Pediatrics, Virology, and Medical Microbiology, University of Turku, and Children's
Hospital, University of Helsinki, Finland

SUMMARY We studied prospectively 154 febrile children to determine the diagnostic value of the
quantitative serum C reactive protein concentrations (CRP). Children with acute otitis media,
acute tonsillitis, or treated with antibiotics during the two previous weeks and infants less than 2
months of age were excluded. Ninety seven children were from private paediatric practice and 57
were patients who had been admitted to hospital. The comparison group consisted of 75 children
with confirmed bacterial infections whose CRP values were recorded retrospectively. In the study
group 35 (23%) children had a confirmed viral infection, 92 (59%) had a probable viral infection
as judged from the clinical picture and outcome of the illness, and 27 (18%) had a bacterial or
probable bacterial infection. When the duration of the disease was more than 12 hours and the
CRP value less than 20 mg/l, all children had a confirmed or probable viral infection. Nine

Protected by copyright.
children (one from the study group and eight from the comparison group) were found to have a
septic infection and a CRP value of 20 mg/l or less. In all these cases, however, the duration of the
symptoms was less than 12 hours. In addition CRP .20 mg/l was found in five (14%) children
with urinary tract infection in the comparison group. CRP values of 20-40 mg/l were recorded in
children with both viral and bacterial infections. A CRP value .40 mg/l detected 79% of
bacterial infections with 90% specificity. Our data show that determination of serum CRP
concentrations is a valuable tool in evaluating children who have been ill for more than 12 hours.

Febrile children comprise one of the most important IChildren with acute otitis media were excluded
problems in paediatric practice. The major diagnos- because they are routinely treated with antibiotics.
tic goal is to distinguish bacterial and treatable IChildren with tonsillitis were excluded because a
infections from viral ones. Recent studies have separate study was being carried out of tonsillitis.
shown serum C reactive protein (CRP) determina- There were 70 girls and 84 boys. Seventy eight
tion to be useful in the diagnosis of invasive bacterial (51%) children were less than 2 years old, 42 (27%)
infections. 1-3 We studied the value of CRP in between 2 and 6 years, and 34 (22%) over 6 years
evaluating febrile children with a special attention to old. The mean age was 3 years 4 months with a
respiratory virus infections. range of 2 months to 14 years.
The outpatients received a questionnaire for
Patients and methods follow up of the symptoms and signs of the disease.
Re-examination was performed in 83 of 97 out-
Study group. From March 18 to May 31 1984, 154 patients seven days after the first visit, and the
consecutive febrile children with respiratory tract questionnaire was returned. The follow up was
infection or without localising signs of infection were conducted by telephone in another 13 patients, and
studied prospectively. Of these, 97 were outpatients one patient returned the questionnaire by mail. The
from private paediatric practice and 57 were patients inpatients were treated in the hospital until afebrile
who had been admitted to our Department of and were not routinely re-evaluated.
Pediatrics. Fever were defined as an axillary Blood was taken for routine haematologic tests
temperature of 38°C (100.4°F) or higher. Children (total white blood cell counts (WBC) and differen-
with acute otitis media, acute tonsillitis, or treated tial counts, erythrocyte sedimentation rate (ESR),
with antibiotics during the two previous weeks and and CRP). Nasopharyngeal mucus aspirates for
infants less than 2 months of age were excluded. rapid diagnosis of virus was taken from 135 children.
24
 Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.com/ on December 23, 2020 at India:BMJ-PG Sponsored.
C reactive protein in the evaluation of febrile illness 25
Further studies were performed only on specific polystyrene microtitre plates (Linbro/Titertek, Flow
clinical indications. Urine analysis with culture was Laboratories, Hamden, Connecticut). The coupling
performed in 68 patients, throat culture for i buffer was 0-05M phosphate buffered saline (pH
haemolytic streptococci in 34 patients, blood culture 7.4). The same buffer supplemented with 1%
in 29 patients, chest roentgenogram in 27 patients, normal sheep serum was used for saturating the
bacterial antigen (Haemophilus influenzae type B, plates and as an assay buffer. Sera were diluted
Streptococcus pneumoniae, Neisseria meningitidis) 1:100 for incubation. After incubation conjugated
detection from the urine in 14 patients, lumbar antibodies to human IgM, IgA, or IgG (Orion
puncture in 11 patients, and bacterial antigen Diagnostica, Espoo, Finland) were added. P-nitro-
detection in serum in eight patients. Pneumococcal phenylphosphate (Orion Diagnostica) was used as
antibodies from paired sera were studied in 20 substrate, and the optical density was measured at
children. 405 nm by a vertical beam photometer (Titertek/
Multiskan, Eflab Oy, Helsinki, Finland). A stan-
Comparison group. The comparison group for CRP dard curve was prepared for each immuno-
concentrations, total WBC counts, and ESRs in globulin class. The results were expressed as relative
bacterial infections included 75 children with proved units (EIU), where one unit was 1/100 of cor-
bacterial disease treated in the same hospital from responding antibody concentration in the reference
January 1 1982 to February 28 1984. The case serum.
records of these children were reviewed retrospec-
tively. There were 22 boys and 53 girls, and the Statistics. Statistical analysis was carried out by the
mean age was 2 years 9 months with a range of 2 one way analysis of variance and the Tukey-Kramer
months to 16 years. Thirty six patients had urinary method using logarithmic CRP and ESR values. P
tract infection (bladder tap culture positive), 21 values less than 0*05 were considered significant.

Protected by copyright.
meningitis (spinal fluid culture positive), 11 epiglot-
titis (blood culture positive), five septicaemia, and Results
two cellulitis with a positive blood culture.
Disease range. In the study group 35 children (23%)
Determination of CRP. The CRP values of the had a confirmed viral disease, 27 (18%) had a
patients who had been admitted to hospital were bacterial or probable bacterial disease, and 92
determined on admission quantitatively by turbi- (59%) had a probable viral infection. Sixty (65%)
dometry (Multistat III, IL) using standards from patients with probable viral infection had respira-
Behringwerke AG and antiserum from Dakopatts tory symptoms. Eighty one (88%) of the children
AS. In the outpatients the CRP was quantified with probable viral infection recovered within a
with a rapid liquid phase immunoprecipitation week without treatment with antibiotics. The
technique4 using reagents (CRP kit cat no D-147 remaining 11 patients were treated with antibiotics
and CRP Buffer cat no D-179) from Orion Diagnos- for various reasons.
tica, Finland. The viruses found by direct antigen detection
were parainfluenza type 3 (14 cases), influenza B
Enzyme immunoassay for viral antigens. Naso- (12), parainfluenza type 1 (7), adenovirus (3), and
pharyngeal secretion specimens were collected by respiratory syncytial virus (1). One patient had both
suction through nostrils with a disposable mucus parainfluenza type 3 and respiratory syncytial virus
extractor (Vygon, Ecouen, France). The specimen simultaneously. One child had parainfluenza type 3
volume obtained was usually 0-5-2-0 ml. The virus at the first visit and influenza B virus at the
specimens were tested parallel for respiratory follow up visit. Table 1 shows the diagnoses of the
syncytial virus, adenovirus, influenza A and B virus, patients. The aetiologic agent of all urinary tract
and parainfluenza virus type 1, 2, and 3 antigens by infections was Escherichia coli. The causative agent
analogous enzyme immunoassays. The indirect in the two cases of meningitis was Haemophilus
enzyme immunoassay was used as described influenza type B, cultured from the spinal fluid and
previously.5 the blood. Sixteen children had lobar pneumonia
with high fever, and they all responded to treatment
Enzyme immunoassay for pneumococcal antibodies. with antibiotics within 12-24 hours, fulfilling the
IgM, IgA, and IgG antibodies to pneumococci clinical criteria of bacterial pneumonia. Pneumococ-
were assessed by an enzyme immunoassay with a cal antibodies were raised in four of these patients
14 valent pneumococcal polysaccharide vaccine (IgM 7-4-13-1, IgM 7-5-30-1, IgM 12-5-24-5, and
(PneumovaxR; Merck, Sharp and Dohme) as anti- IgA 0*6-22.3, respectively). One patient was posi-
gen. The antigen was coupled (1-4 [ig/ml) onto tive for pneumococcal antigen in the urine and
 Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.com/ on December 23, 2020 at India:BMJ-PG Sponsored.
26 Putto, Ruuskanen, Meurman, et al
Table 1 Diagnoses of patients of the study group
Outpatients (n=97) Patients who had been admitted to hospital (n=57)
Disease No Disease No
Proved or suspected viral diseases
Upper respiratory tract infection 52 (19)' Fever without localising signs 8 (1)*
Fever without localising signs 32 (6)* Laryngitis 7 (4)'
Pharyngitis 3 Wheezy bronchitis 7
Urticaria 2 Upper respiratory tract infection 7 (4)'
Laryngitis t Febrile convulsions 3 (1)
Wheezy bronchitis t Exanthem subitum 2
Exanthem subitum t Purpura non ultra descriptus I

Proved or suspected bacterial diseases

Lobar pneumonia 2 (2)* Lobar pneumonia 14 (4)*
Streptococcal pharyngitis 2 (2)* Urinary tract infection 5 (5)**
Urinary tract infection t (1)** Meningitis 2 (2)"
Pneumococcaemia (1)**
*Confirmed viral infection (positive for nasopharyngeal secretion).
'Bacterial culture or antigen detection in serum/urine positive or serologically confirmed bacterial infection.

Table 2 C reactive protein values (CRP), white blood cell count (WBC), and erythrocyte sedimentation rate (ESR)
in patients with bacterial, viral, and probable viral infections

Protected by copyright.
CRP (mgll) WBC count (x/ 09/l) ESR (mm/h)
n Mean (SD) n Mean (SD) n Mean (SD)

Study group:
Bacterial infections 27 98 (67) 27 18-1 (6-7) 23 57 (3()
Viral infections 35 13 (12) 35 85 (45) 32 16 (11)
Probable viral infections 92 17 (20)) 92 9-2 (5-3) 84 16 (12)

Comparison group:
Septic infections 39 115 (71) 39 18-7 (7-8) 31 46 (32)
Urinary tract infections 36 89 (61) 36 19-8 (6-3) 33 53 (33)

serum, and one patient had a positive blood culture tive values of CRP, WBC count, and ESR for
for Haemophilus influenzae type B. bacterial infections in the study group. Of the
patients with viral or probable viral infection, 75%
CRP, WBC count, and ESR. Table 2 shows the had CRP less than 20 mg/l. Of the patients with CRP
mean (SD) values of CRP, WBC count, and ESR. >40 mg/l, 59% had bacterial infection, and 90%
The differences between bacterial and viral diseases of patients with viral or probable viral infection had
and probable viral infections were significant CRP <40 mg/l (Figure).
(p<0-05) in all these laboratory tests. There was no Of the patients with bacterial disease, 67% had
significant difference between viral and probable WBC count >15x109/l, and 87% of patients with
viral diseases, nor in bacterial infections between viral or probable viral infection had WBC count
the study group and the comparison group. <15x109/l. ESR >30 mm/h detected 91% of
Because CRP values may not begin to rise in some patients with bacterial infections, and 89% of
patients until 12-22 hours after onset of stimulus,6 patients with viral or probable viral infection had
patients who had been sick for less than 12 hours ESR <30 mm/h.
were excluded from the calculations of the sensitiv- In the comparison group there were eight patients
ity and specificity of CRP, WBC count, and ESR for with septic bacterial infections with duration of
bacterial infections. In three patients of the study symptoms for less than 12 hours and CRP <20 mg/l.
group the duration of symptoms was less than 12 In addition there were five children with urinary
hours; one had pneumococcaemia, one bacterial tract infection with CRP <20 mg/l.
pneumonia (confirmed serologically), and one
meningitis. One of these patients had CRP <20 mg/l. Outcome and treatment. Eighty three outpatients
Table 3 shows the sensitivity, specificity, and predic- attended the follow up visit. Eight children had otitis
 Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.com/ on December 23, 2020 at India:BMJ-PG Sponsored.
C reactive protein in the evaluation of febrile illness 27
Table 3 Sensitivity, * specificity, t positive predictive value,* and negative predictive value§ of C reactive protein
concentrations (CRP), white blood cell count (WBC), and erythrocyte sedimentation rate (ESR) for bacterial
infections in the study group of 151 children who had been ill for more than 12 hours (values are %)
Sensitivitv Specificity Positive predictive Negative predictive
value value
CRP (mg/I) >20 10( 75 43 100
>40 79 9( 59 96
>8() 63 98 88 93
WBC (x109/l) >1() 96 66 35 99
>15 67 87 50 93
>2(0 38 97 69 93
ESR (mm/h) >30) 91 89 61 98
CRP (mg)U/WBC (x1l9/1): -
>2(/> 15 67 95 70 94
>40/> 15 58 96 74 92
>80/>20 33 1()( 10( 89
ESR (mm/h)/WBC (x109/l): -
>30/>15 59 95 68 92
*Probability that test results will be positive when bacterial disease is present.
tProbability that test results will be negative wheit bacterial disease is not present.
*Probability that bacterial disease is present when results of test are positive.
§Probability that bacterial disease is not present when results of test are negative.-

Outpatients Patients cdmitted Patients with septic Patients with urniary

to hospitcl infections tract infections

Protected by copyright.
180 160 130 80 140 190 170>200 140 >18C 75 151 169>200180>180 >200120 100 120 18C 85 95
0 000* *00*000000
* * * * * **
200 190 140 80 170 >200 >200 100 160>180 190 220>200 80 180>200180110 140 150
00 * * * * * *@
* * * * * * *
94 85 100 160>200180115 130 190 150 85 150 168 80 80 100
O0 @>200 0110
000
70-
0

0
c 40-
o
0 00 0@
.o_
°i0 0 ~~~0 0
c
0
- 00 -- ---
~~~~~~~-0-
c 40- .000 0
lio ~~0
30- 0o0 0
0~~~~~~~~~~~~~~

o oooooooo oo
I1U0 -~ 0~~~~~~~~~
0~~~~~~~~~
00000000000
0000000000000
0000000000000 00000
ooooooooooooo0 ooooooooooooo0 I 0000 _________
n=97 n=57 n=39 n=36
Study group ( n =154) Comparison group (n=75)
Figure Serum C reactive protein concentrations ofpatients in the study and the comparison groups (0= bacterial infection,
OI=confirmed viral infection, O=probable viral infection).
 Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.com/ on December 23, 2020 at India:BMJ-PG Sponsored.
28 Putto, Ruuskanen, Meurman, et al
media, six of whom had had a confirmed viral A CRP value exceeding 20 mg/l has been sug-
disease at the first visit. Fourteen children had gested as a screening limit for bacterial infections.'
minor respiratory symptoms, and one had wheezy In this study we found its sensitivity for bacterial
bronchitis. All but one of the children with a new infection to be 100%, but the false positive rate was
viral infection (influenza B) and otitis were afebrile. 25% and the positive predictive value only 43%.
The children contacted by phone or by mail had Our data prefer CRP value 40 mg/l as a screening
recovered uneventfully. limit, because its specificity (90%) and positive
Seven of the 97 outpatients were treated with predictive value (59%) are better and sensitivity is
antibiotics at the first visit. Of these, two had still 79%. Raised CRP values (.40 mg/l) were
alveolar pneumonia, one had urinary tract infection, found in one (3%) of 35 patients with confirmed
and two had influenza B virus infection. Two viral infection and in 13 (14%) of the 91 patients
patients with streptococcal pharyngitis were treated with probable viral illness. We have shown previously
after their throat culture proved positive. In the that in children who were admitted to hospital with
other eleven outpatients treated, antibiotics were adenovirus, influenza, parainfluenza, or respiratory
started during follow up or at the follow up visit. syncytial virus infection CRP values more than 40
Eight of the treated children had acute otitis media, mg/l were recorded in 38%, 20%, 0%, and 13% of
one had lymphadenitis, one had upper respiratory patients, respectively.9 Some of these patients may
tract infection, and one had a positive throat culture have undergone a concomitant bacterial infection,
for i haemolytic streptococci group G, although she which remained undiagnosed. On the other hand, it
did not have pharyngitis. All 22 bacterial infections is probable that some viral infections, especially
of patients who had been admitted to hospital were adenovirus infections,'( may induce so much tissue
treated with antibiotics. Altogether, 18 outpatients damage that CRP values will be raised.
(19%) and 32 patients who had been admitted to Many previous studies have evaluated the value of

Protected by copyright.
hospital (56%) were treated with antibiotics. WBC count, ESR, and CRP in the detection of
serious bacterial infections. 15- No test alone or in
Discussion combination has 100% sensitivity and specificity for
bacterial infections. Rasmussen and Rasmussen
This study differs from previous studies of CRP in have shown that WBC and differential counts have
that rapid detection of respiratory virus antigens in little value in distinguishing bacterial infections from
nasopharyngeal mucus permitted us to diagnose the viral infections." McCarthy et al found a clinically
specific viral aetiology within 30 hours in 35 (23%) useful association between a WBC count >15x109/l
patients. The results show that the quantitative CRP and ESR >30 mm/h and pneumonia or bacteraemia
test is a valuable tool for distinguishing viral infec- in children less than 2 years old with temperatures
tions from bacterial infections in febrile children. If >40'C.16 In a further study they showed that ESR
the duration of the illness was more than 12 hours >30 mm/h was more sensitive than a positive slide
and the CRP value was less than 20 mg/l all children test for CRP (1:50) in bacteraemia, pneumonia, and
investigated had viral or probable viral infection. other bacterial infections, but it was less specific
Seventy six (78%) of the 97 febrile children studied than a positive CRP test or a high WBC count.17
in private practice and 19 (34%) of the patients who In this study the best single indicator of bacterial
had been admitted to hospital belonged to this infection was ESR >30 mm/h with a sensitivity of
group. There were nine children (one in the study 91% and a specificity of 89% (Table 3). In con-
group and eight in the comparison group) with CRP firmed viral and probable viral infections ESR was
<20 mg/l who had an invasive bacterial infection. In less than 30 mm/h in 89% and 88%, respectively.
all cases the duration of the disease was 12 hours or The differences between ESR >30 mm/h and CRP
less, which suggests that the CRP test is good for >40 mg/I, however, were small, and the CRP test
screening only in patients who have been sick for has several other advantages over ESR. CRP can be
more than 12 hours. McCabe and Remington re- measured quantitatively with a simple and inexpen-
ported CRP values less than 20 mg/l in 14% of 36 sive method in 15 minutes.4 It increases within hours
patients with bacteraemia.7 The study, however, did and, most important, CRP decreases during success-
not report the exact duration of the symptoms of the ful treatment with a half life of three days.'8 The
patients. CRP was <20 mg/l in five (14%) of the decrease is much more rapid than that of ESR.'2
patients with urinary tract infection in the compari- The rapid decrease of CRP permits the use of this
son group of this study, suggesting that the infection test during the follow up of patients with bacterial
would have been in bladder level. Jodal et al have infections. Peltola et al have recently shown that a
shown that serum CRP concentrations will not reincrease of CRP during the treatment of bacterial
increase in cystitis.8 meningitis reflects complications. 19
 Arch Dis Child: first published as 10.1136/adc.61.1.24 on 1 January 1986. Downloaded from http://adc.bmj.com/ on December 23, 2020 at India:BMJ-PG Sponsored.
C reactive protein in the evaluation of febrile illness 29
The determination of CRP combined with a response. Serum C-reactive protein kinetics after acute myocar-
careful clinical examination was useful when the dial infarction. J Clin Invest 1978;61:235-42.
necessity of treatment with antibiotics was being
7 McCabe RE, Remington JS. C-reactive protein in patients with
bacteremia. J Clin Microbiol 1984;20:317-9.
considered. Only seven of 97 outpatients were Jodal U, Lindberg U. Lincoln K. Level diagnosis of symptoma-
treated with antibiotics at the first visit. The tic urinary tract infections in childhood. Acta Paediatr Scand
uneventful recovery of 79 (88%) of the untreated 1975;64:201-8.
9 Ruuskanen 0, Putto A, Sarkkinen H, Meurman 0, Irjala K. C-
outpatients justifies our antibiotic policy. reactive protein in respiratory virus infections. J Pediatr
In conclusion, our data show that determination 1985 ;107:97-100.
of CRP is a valuable tool in evaluating febrile "' Ruuskanen 0, Sarkkinen H, Mcurman 0, et al. Rapid diagnosis
children. When the duration of the illness is 12 hours of adenoviral tonsillitis: a prospective clinical study. J Pediatr
or more and serum CRP concentration less than 20 1984;104:725-8.
Rasmussen NH, Rasmussen LN. Predictive value of white blood
mg/l and there is no identifiable focus of bacterial cell count and differential cell count to bacterial infections in
infection-that is, otitis, tonsillitis, cystitis, etc-the children. Acta Paediatr Scand 1982;71:775-8.
disease is most probably a benign viral infection. 12 Peltola H, Rasanen JA. Quantitative C-reactive protein in
CRP values of 20-40 mg/l may be recorded both in relation to erythrocyte sedimentation rate, fever, and duration
of antimicrobial therapy in bacteraemic diseases of childhood.
viral and bacterial infections. Most febrile children Journal of Infection 1982;5:257-67.
with CRP >40 mg/l have a bacterial infection. 13 Bennish M, Beem MO, Ormiste V. C-reactive protein and zeta
Finally, it should be emphasised that no laboratory sedimentation ratio as indicators of bacteremia in pediatric
test will replace careful clinical judgment of a febrile patients. J Pediatr 1984;104:729-32.
4 Hanson LA, Jodal U, Sabel KG, Wadsworth C. The diagnostic
child. value of C-reactive protein. Pediatr Infect Dis 1983;2:87-90.
'1 Todd JK. Childhood infections. Diagnostic value of peripheral
white blood cell and differential cell counts. Am J Dis Child
This research was supported by grants from the Academy of 1974;127:810-6.
Finland, the Sigrid Juselius Foundation, and the Foundation for McCarthy PL, Jekel JF, Dolan TF, Jr. Temperature greater

Protected by copyright.
Pediatric Research. than or equal to 40°C in children less than 24 months of age: a
prospective study. Pediatrics 1977;59:663-8.
References 7 McCarthy PL, Jekel JF, Dolan TF, Jr. Comparison of acute-
phase reactants in pediatric patients with fever. Pediatrics
Peltola HO. C-reactive protein for rapid monitoring of infec- 1978;62:716-20.
tions of the central nervous system. Lancet 1982;i:980-3. lx Schofield KP, Voulgari F, Gozzard DI, Leyland MJ,
2
Peltola H, Holmberg C. Rapidity of C-reactive protein in Beeching NJ. C-reactive protein concentration as a guide to
detecting potential septicemia. Pediatr Infect Dis 1983;2:374-6. antibiotic therapy in acute leukemia. J Clin Pathol 1982;35:
3 Peltola H. C-reactive protein in rapid differentiation of acute 866-9.
epiglottitis from spasmodic croup and acute laryngotracheitis: a 9 Peltola H, Luhtala K, Valmari P. C-reactive protein as a
preliminary report. J Pediatr 1983;102:713-5. detector of organic complications during recovery from child-
4 Harmoinen A, Perko M, Gronroos P. Rapid quantitative hood purulent meningitis. J Pediatr 1984;104:869:72.
determination of C-reactive protein using LKB 8600 reaction 201 Griner PF, Mayewski RJ, Mushlin Al, Greenland P. Selection
rate analyzer. Clin Chim Acta 1981;111:117-20. and interpretation of diagnostic tests and procedures. Principles
Sarkkinen HK, Halonen PE, Arstila PP, Salmi AA. Detection and applications. Ann Intern Med 1981;94(Part 2):555-600.
of respiratory syncytial, parainfluenza type 2, and adeno virus
antigens by radio-immunoassay and enzyme-immunoassay on Correspondence to Dr Anne Putto, Department of Pediatrics,
nasopharyngeal specimens from children with acute respiratory Turku University Hospitals, 20520 Turku, Finland.
disease. J Clin Microbiol 1981;13:258-65.
6 Kushner 1, Broder ML, Karp D. Control of the acute phase Received 29 July 1985