- Email: [email protected]
The association of subclinical infection with preterm labor: The role of C-reactive protein
Ho, Wong, and Ma
5. 6. 7.
8.
9.
complete and partial moles and in nonmolar abortuses. AMj 0BSTET GYNECOL 1984;149:129. Szulman AE, Surti V. The syndromes of hydatidiform mole. II. Morphologic evolution of the complete and par tial mole. AM J 0BSTET GYNECOL 1978; 132:20. Ma HK, Wong LC. Gestational trophoblastic disease in Hong Kong. Semin Oncol 1982;9:224. Korenman SG, Stevens RH , Carpenter LA, Robb M, Niswender GD, Sherman BM . Oestradiol radioimmuno assay without chromatography: procedure, validation and normal values. J Clin Endocrinol Metab 1974;38:718. Koligian KB, Stormshak F. Nuclear and cytoplasmic es trogen receptors in ovine endometrium during the es trous cycle. Endocrinology 1977; I01:524. Vaitukaitis J, Braunstein GD, Ross GT. A radioimmuno assay which specifically measures human chorionic go nadotropin in the presence of human luteinizing hor mone. AM j 0BSTET GYNECOL 1972; 113:75 1.
November 15, 1985 Am .J Obstet Gynecol
10. Miyake A, Aono T, Tanizawa 0, Kurachi H, Kurachi K. Restoration of the gonadotrophin response to LH-RH and oestrogen administration in patients after molar abor tion. Acta Endocrinol 1975 ;9 1:30. II. Miyake A, Aono T, Kurachi H, Tsutsumi H, Kurachi K. Restoration of the ovarian response to gonadotropins in patients after molar abortion. Obstet Gynecol 1981 ;58: 566. 12. Dawood MY, Ratnam SS, Teoh ES. Serum estradiol-17[3 and serum human chorionic gonadotropin in patients with hydatidiform moles. AM j 0BSTET GYNECOL 1974; 119:904. 13. Goldstein DP, Berkowitz RS. The diagnosis and manage ment of molar pregnancy. In: Goldstein DP, Berkowitz RS, eds. Gestational trophoblastic neoplasms. Philadel phia: WB Saunders, 1982: 143.
The association of subclinical infection with preterm labor: The role of C-reactive protein Ronald K. Potkul, M.D., Atef H. Moawad, M.D., and Kathryn L. Ponto, B.A. Chicago, Illinois The role of subclinical intrauterine infection in preterm labor was evaluated prospectively in 40 patients and appropriate control subjects. The 24 preterm labor patients (60%) with a negative C-reactive protein value responded to tocolysis 95.8% of the time, with a mean delay of delivery of 35.5 days and a mean gestational age of 36.9 weeks. The 16 patients (40%) with a positive C-reactive protein value responded to tbcolysis only 37.5% of the time, with a mean delay of delivery of 14.4 days and a mean gestational age of 33.2 weeks. Pathologic evidence of chorioamnionitis was present in 32.9% of 310 preterm deliveries as compared to only 22.3% of 1631 term deliveries. The presence of subclinical infection must be considered in cases of preterm labor, especially among patients for whom tocolytic therapy is unsuccessful. (AM J 0BSTET GYNECOL 1985;153:642-5.)
Key words: C-reactive protein , preterm labor, subclinical infection Despite the fact that it is the largest cause of perinatal morbidity and mortality, the cause of preterm delivery is poorly understood . Recently, several reports have suggested that a percentage of premature labor cases may be caused by subclinical chorioamnionitis and that, in addition, this type of infection may affect the uterine response to tocolysis. ,_, There are a number of indications that subclinical infections contribute to preterm labor. Bobitt et al.," in 1981 , isolated microorganisms from amniocentesis specimens of patients in preterm labor; the membranes From the Department of Obstetrics and G_vnecology and The Perinatal Center, The Pritzker School of Medicine and The Chicago Lying In Hospital, The University of Chicago. Supported in part b_v the Mother's Aid Research Fund, The Chicago Lying-In Hospital. Received for publication February 22, 1985; revised and accepted August 26, 1985. R eprint requests: Atef H. Moawad, M.D. , The Chicago Lying-In Hospital, 5841 South Maryland Ave., Chicago, IL 60637.
642
were intact in 25% of these cases. The significance of this finding is not clear since microorganisms can also be found in 10% of patients in normal term labor.' A study by Miller et al.' indicates that premature labor and delivery may frequently reflect undiagnosed sub clinical intrauterine infection. C-reactive protein is a serum globulin that forms a precipitate with the C-fraction carbohydrate of strep tococcal pneumonia ; it is frequently elevated in cases of bacterial infection and tissue injury and recovery. The purpose of this study was to examine the role of subclinical chorioamnionitis in premature labor and evaluate the use of C-reactive protein as a possible marker for this type of infection.
Material and methods The first phase of the study consisted of the micro scopic examination of placentas for evidence of cho rioamnionitis. All of the I 63 I deliveries between Feb
C-reactive protein in preterm labor
Volume 153 Number 6
643
Table I. Comparison of groups
I
Mean
3.30 2.42
0.62 0.23
1.50 1.00
1.3
2.71
0.38
1.33
1.0
2.80
0.32
1.32
Mean
23.6 22.7
1.8 0.9
24
23.8
16
23.1
Mean
Preterm control Term control
10 12
Negative C-reactive protein Positive C-reactive protein
I
SE
SE
N
Group
I
Gestational age (wk)
Abo1·tions
Parity
Gmvidity
Age (y1)
SE
0.80 0.42
0.33 0. 15
30.9 39.2
0.9 0.4
0.38
0.16
31.9
0.5
0.50
0. 16
31.2
0.5
0.48
0.33
I
M ean
Mean
0.21
I
SE
SE
0.29
Table II. C-reactive protein values correlated with groups Negative C-reactive protein (,;;;0.7 mgldl) N
n
Preterm control
10
10
Preterm labor
40
Term labor control
12
Group
l
%
Positive C-reactive protein (> 0.7 mgldl) 11
100
0
24
60
16
II
91.7
ruary 1, 1983, and October 31, 1983 , in which labor was neither induced nor complicated by premature rupture of the membranes( > 12 hours), were included. Pathologic evidence of chorioamnionitis was defined as: (1) infiltration of the extraplacental membranes by polymorphonuclear leukocytes, (2) accumu lation of polymorphonuclear leukocytes in the intervillous space immediately below the chorionic plate, (3) infiltration of the chorionic plate by leukocytes, and (4) angiitis of the umbilical vessels. The number of cases with patho logic evidence of chorioamnionitis was compared for both term (;. 37 weeks) and preterm (< 37 weeks) births by means of the x ~ test for statistical significance. Fifty patients with sin gleton pregnancies who were found to be in premature labor were enrolled in the second phase of the study after meeting several criteria. Each patient was required : (l) to be between 24 and 36 weeks' gestation; (2) to have intact membranes; (3) to be undergoing uterine contractions every 5 minutes for at least 1 hour, which were unresponsive to intra venous hydration; (4) to have a cervix judged favorable for tocolysis; (5) to manifest no evidence of fetal dis tress; (6) to be receiving no tocolytic drugs; (7) to have no evidence of present or recent clinical infection. A specimen of urine for culture was obtained from all patients at the time of admission in order to detect any current urinary tract infection. Of the 50 patients enrolled in the study, 10 were excluded for failure to meet the above requirements. These 10 cases included one twin gestation, two patients already receiving rito drine, one with a gestational age of < 24 weeks, one with a gestational age of > 37 weeks, two receiving am picillin for a recent urinary tract infection , two with urine cultures demonstrating the presence of> 10" bac
l
% 0 40 8.3
p
< 0.05 < 0.05
teria, and one patient with a fever of unknown origin. The control groups consisted of ( 1) 10 patients between 24 and 36 weeks' gestation who were not in labor and (2) 12 patients in labor at term with intact membranes. Blood for a quantitative C-reactive protein assay (a nephelometric assay with the Beckman Immunochem istry Analyzer) was obtained from the study group and the two control groups as part of the patient's routine admission laboratory tests .. The preterm patients were treated with intravenous ritodrine according to a stan dard obstetrics protocol by the obstetrics staff. Each case was managed without knowledge of the C-reactive protein results. Tocolytic therapy was abandoned if spontaneous rupture of the amniotic membranes oc curred or if advanced cervical dilatation and efface ment were encountered. Following delivery the pla centa was examined microscopically for evidence of chorioamnionitis. The C-reactive protein levels of the preterm labor patients were then compared to those of the control groups for both the numbers with patho logic evidence of chorioamnionitis and the degree to which gestation was prolonged by tocolysis . Statistical analysis was performed with Student's t test and the x ~ test; results were considered statistically significant if the confidence interval (p) was <0.05 .
Results Three hundred sixty-three (22.3%) of the 1631 term deliveries showed pathologic evidence of chorioam nionitis as compared to 102 (32.9%) of the 3 10 preterm deliveries during the 9-momh study period. This dif ference is statistically significant (p < 0.05). There was no significant difference in demographic characteristics (age, gravidity, parity, number of abor
644
Potkul, Moawad, and Ponto
November 15, 1985 Am J Obstet Gynecol
Table III. Correlation between C-reactive protein and pathologic evidence of chorioamnionitis Negative C-reactive protein Chorioamnionitis Group
N
n
Preterm labor T erm labor control
38* 12
Total
50
I
Positive C-reative protein
No chorioamnionitis
%
n
4 3
18.2
18 8
7
21.2
I
Chorioamnionitis
%
n
81.8
10 0
78.8
26
I
No chorioamnionitis
I
n
% 62.5
%
p
6
37.5
<0.01 NS
7
41.2
<0.01
I
10
58 .8
*Pathology reports were not available in two patients.
Table IV. C-reactive protein correlated with prolongation of gestation Days to delivery C-reactive protein
N
Mean
Negative Positive
24 16
35.5 14.4
I
Gestational age at delivery (wk) SE
Mean
4.1 5.0
36.9 33.2
1
0.6 0.7
< 0.01
p
SE
< 0.01
Table V. C-reactive protein correlated with results of tocolysis Successful C-reactive protein
N
n
Negative Positive
24 16
23 6
p
I
Failed
%
n
95.8 37.5
10
I
I
%
n
4.2 62.5
13 2
< 0.001
tions, and gestational age) between the study and the control group patients (Table I). In an attempt to apply the C-reactive protein assay as a highly sensitive but somewhat less specific clinical test, levels of >0.7 were considered positive for the remainder of the analysis. With the use of this value the C-reactive protein assay was found to be positive for 40% of the preterm labor group . There were no cases with positive C-reactive protein assays among pre term patients who were not in labor and only an 8% incidence among term patients in labor (Table II). The differences between the preterm labor group and each of the two control groups were significant. The C-reactive protein test results were found to cor respond reasonably well with the microscopic diagnosis of chorioamnionitis, which was present in 62.5% of those patients with a positive C-reactive protein assay (Table III). None of these patients exhibited any of the clinical signs of chorioamnionitis. In two of the pa tients with negative results for both the pathologic and C-reactive protein tests and in three patients with pos itive results in both cases, the membranes had ruptured more than 12 hours before delivery. Tocolysis was significantly more successful in the group of patients with negative C-reactive protein assay results. The mean number of days to delivery and the
Preterm birth
Term birth
1
%
n
54.2
II 14
12.5
1
% 45.8 87 .5
< 0.01
mean gestational age at delivery were 35.5 days and 36.9 weeks for the group with negative C-reactive pro tein versus 14.4 days and 33.2 weeks for the group with positive C-reactive protein (Table IV) . Tocolysis was successful in prolonging gestation for 1 week in 23 of the 24 patients (95.8%) witha negative C-reactive protein value (Table V) . The single failure occurred in a patient who required discontinuation of ritodrine therapy at 30.5 weeks because of severe ma ternal tachycardia. The contractions, which had ceased during ritodrine therapy, recurred after termination of the drug. Gestation was prolonged for > 7 days for only six of the 16 patients (37 .5%) with a positive C-reactive protein assay. Gestation was prolonged to term (~37 weeks) in 54.2% of the patients with a negative e-re active protein value. The same was true for only 12.5% of the patients with positive C-reactive protein values. Comment
The results of this study point to subclinical cho rioamniotic infection as a contributing factor in the cause of preterm labor. C-reactive protein was evaluated as a possible marker for subclinical chorioamnionitis because of past evi dence indicating that the C-reactive protein test is more sensitive than the appearance of fever, leukocytosis,
C-reactive protein in preterm labor 645
Volume 153 Number 6
or fetal tachycardia in predicting the presence of in fection.6· 7 The level of C-reactive protein that was considered abnormal in this study was >0.7 mg/dl. This is in con trast to the cutoff point of 2 mg/dl chosen in two pre vious studies dealing with premature rupture of the membranes from our institution. 6 · 7 The 2 mg/dl level was arbitrarily chosen in an attempt to apply the C reactive protein assay as a highly specific but somewhat less sensitive clinical test of pending clinical infectious morbidity. This was necessary since decisions concern ing induction of labor are based largely on C-reactive protein levels. The 0.7 mg/dllevel chosen in this study was an at tempt to apply the C-reactive protein assay as a highly sensitive but somewhat less specific indicator of sub clinical infection rather than infectious morbidity. Quality control of the Beckman Automated Immu nochemistry system by the company and The Univer sity of Chicago Serology laboratory has confirmed that the C-reactive protein level is <0.6 mg/dl in a normal population. Several aspects of these results are noteworthy. Al though the number of cases is limited, we have shown that normal pregnancy and normal term labor do not raise C-reactive protein levels significantly. It is inter esting to note that 8% of the patients in labor at term had elevated C-reactive protein levels, a figure that ap proximates the 10% incidence of the presence of mi croorganisms in amniotic fluid obtained from patients in labor at term.' In contrast, 40% of our preterm labor study group had elevated C-reactive protein values . The histologic examination of placentas from preterm deliveries at our institution revealed that chorioam nionitis was present 32.9% of the time. Again, the in cidences of elevated C-reactive protein and histologic chorioamnionitis are fairly similar. It seems reasonable to argue that, in the absence of clinical infection of the urogenital or other systems, subclinical infection plays an etiologic role in the onset of preterm labor in ap proximately one third of preterm labor patients. We were also able to demonstrate that when the C reactive protein assay was negative, ritodrine therapy
was successful in virtually every case in which the drug was well tolerated . A recent study by Handwerker et al. 8 has also examined the success rate of tocolysis with respect to C-reactive protein levels. Their lower success rate may be due to the inclusion of cases of urinary tract infection. Our results indicate that it may become possible to predict the success or failure of tocolytic therapy on the basis of the results of the assay for C-reactive pro tein. We recommend, in cases of preterm labor with high C-reactive protein levels or in cases where 13-sym pathomimetic agents fail to stop preterm uterine con tractions, that the presence of infection as a contrib uting factor should be suspected, even in the absence of any clinical manifestation of the infectious process. The task of the prospective identification of causative agents, although a tedious and difficult one, is clearly called for when one considers that in our population approximately one third of cases of preterm labor may have an infectious cause. REFERENCES I. Bobitt JR, Ledger WJ. Unrecognized amnionitis and pre
maturity: a preliminary report.] Reprod Med 1977;19:1. 2. Bobitt JR, Ledger WJ. Amniotic fluid analysis. Its role in maternal and neonatal infection. Obstet Gynecol 1978; 51:56. 3. BobittJR, Hayslip CC, DamatoJD. Amniotic fluid infection as determined by transabdominal amniocentesis in patients with intact membranes in premature labor. AM J OBSTET GYNECOL 1981;140:947. 4. Miller JM , Pupkin MJ, Hill GB . Bacterial colonization of amniotic fluid from intact fetal membranes. AMJ OBSTET GYNECOL 1980; 136:796. 5. Lis twa HM, Dobek AS, CarpenterJ, et al. The predictability of intrauterine infection by analysis of amniotic fluid. Ob stet Gynecol 1976;48:31. 6. Ismail MA, Zinaman MJ, Lowensohn R, Moawad AH. The significance of C-reactive protein levels in women with pre mature rupture of membranes. AM J OBSTET GYNECOL 1985; 151 :541. 7. Evans MI , Hajj SN, Devoe LD, Angerman NS, Moawad AH . C-reactive protein as a predictor of infectious mor bidity with premature rupture of membranes. AMJ OBSTET GYNECO L 1980; 138:648. 8. Handwerker SM , Nergesh AT, Verma UL, et al. Correla tion of maternal serum C-reactive protein with outcome of tocolysis. Obstet Gynecol 1984;63:220.
5. 6. 7.
8.
9.
complete and partial moles and in nonmolar abortuses. AMj 0BSTET GYNECOL 1984;149:129. Szulman AE, Surti V. The syndromes of hydatidiform mole. II. Morphologic evolution of the complete and par tial mole. AM J 0BSTET GYNECOL 1978; 132:20. Ma HK, Wong LC. Gestational trophoblastic disease in Hong Kong. Semin Oncol 1982;9:224. Korenman SG, Stevens RH , Carpenter LA, Robb M, Niswender GD, Sherman BM . Oestradiol radioimmuno assay without chromatography: procedure, validation and normal values. J Clin Endocrinol Metab 1974;38:718. Koligian KB, Stormshak F. Nuclear and cytoplasmic es trogen receptors in ovine endometrium during the es trous cycle. Endocrinology 1977; I01:524. Vaitukaitis J, Braunstein GD, Ross GT. A radioimmuno assay which specifically measures human chorionic go nadotropin in the presence of human luteinizing hor mone. AM j 0BSTET GYNECOL 1972; 113:75 1.
November 15, 1985 Am .J Obstet Gynecol
10. Miyake A, Aono T, Tanizawa 0, Kurachi H, Kurachi K. Restoration of the gonadotrophin response to LH-RH and oestrogen administration in patients after molar abor tion. Acta Endocrinol 1975 ;9 1:30. II. Miyake A, Aono T, Kurachi H, Tsutsumi H, Kurachi K. Restoration of the ovarian response to gonadotropins in patients after molar abortion. Obstet Gynecol 1981 ;58: 566. 12. Dawood MY, Ratnam SS, Teoh ES. Serum estradiol-17[3 and serum human chorionic gonadotropin in patients with hydatidiform moles. AM j 0BSTET GYNECOL 1974; 119:904. 13. Goldstein DP, Berkowitz RS. The diagnosis and manage ment of molar pregnancy. In: Goldstein DP, Berkowitz RS, eds. Gestational trophoblastic neoplasms. Philadel phia: WB Saunders, 1982: 143.
The association of subclinical infection with preterm labor: The role of C-reactive protein Ronald K. Potkul, M.D., Atef H. Moawad, M.D., and Kathryn L. Ponto, B.A. Chicago, Illinois The role of subclinical intrauterine infection in preterm labor was evaluated prospectively in 40 patients and appropriate control subjects. The 24 preterm labor patients (60%) with a negative C-reactive protein value responded to tocolysis 95.8% of the time, with a mean delay of delivery of 35.5 days and a mean gestational age of 36.9 weeks. The 16 patients (40%) with a positive C-reactive protein value responded to tbcolysis only 37.5% of the time, with a mean delay of delivery of 14.4 days and a mean gestational age of 33.2 weeks. Pathologic evidence of chorioamnionitis was present in 32.9% of 310 preterm deliveries as compared to only 22.3% of 1631 term deliveries. The presence of subclinical infection must be considered in cases of preterm labor, especially among patients for whom tocolytic therapy is unsuccessful. (AM J 0BSTET GYNECOL 1985;153:642-5.)
Key words: C-reactive protein , preterm labor, subclinical infection Despite the fact that it is the largest cause of perinatal morbidity and mortality, the cause of preterm delivery is poorly understood . Recently, several reports have suggested that a percentage of premature labor cases may be caused by subclinical chorioamnionitis and that, in addition, this type of infection may affect the uterine response to tocolysis. ,_, There are a number of indications that subclinical infections contribute to preterm labor. Bobitt et al.," in 1981 , isolated microorganisms from amniocentesis specimens of patients in preterm labor; the membranes From the Department of Obstetrics and G_vnecology and The Perinatal Center, The Pritzker School of Medicine and The Chicago Lying In Hospital, The University of Chicago. Supported in part b_v the Mother's Aid Research Fund, The Chicago Lying-In Hospital. Received for publication February 22, 1985; revised and accepted August 26, 1985. R eprint requests: Atef H. Moawad, M.D. , The Chicago Lying-In Hospital, 5841 South Maryland Ave., Chicago, IL 60637.
642
were intact in 25% of these cases. The significance of this finding is not clear since microorganisms can also be found in 10% of patients in normal term labor.' A study by Miller et al.' indicates that premature labor and delivery may frequently reflect undiagnosed sub clinical intrauterine infection. C-reactive protein is a serum globulin that forms a precipitate with the C-fraction carbohydrate of strep tococcal pneumonia ; it is frequently elevated in cases of bacterial infection and tissue injury and recovery. The purpose of this study was to examine the role of subclinical chorioamnionitis in premature labor and evaluate the use of C-reactive protein as a possible marker for this type of infection.
Material and methods The first phase of the study consisted of the micro scopic examination of placentas for evidence of cho rioamnionitis. All of the I 63 I deliveries between Feb
C-reactive protein in preterm labor
Volume 153 Number 6
643
Table I. Comparison of groups
I
Mean
3.30 2.42
0.62 0.23
1.50 1.00
1.3
2.71
0.38
1.33
1.0
2.80
0.32
1.32
Mean
23.6 22.7
1.8 0.9
24
23.8
16
23.1
Mean
Preterm control Term control
10 12
Negative C-reactive protein Positive C-reactive protein
I
SE
SE
N
Group
I
Gestational age (wk)
Abo1·tions
Parity
Gmvidity
Age (y1)
SE
0.80 0.42
0.33 0. 15
30.9 39.2
0.9 0.4
0.38
0.16
31.9
0.5
0.50
0. 16
31.2
0.5
0.48
0.33
I
M ean
Mean
0.21
I
SE
SE
0.29
Table II. C-reactive protein values correlated with groups Negative C-reactive protein (,;;;0.7 mgldl) N
n
Preterm control
10
10
Preterm labor
40
Term labor control
12
Group
l
%
Positive C-reactive protein (> 0.7 mgldl) 11
100
0
24
60
16
II
91.7
ruary 1, 1983, and October 31, 1983 , in which labor was neither induced nor complicated by premature rupture of the membranes( > 12 hours), were included. Pathologic evidence of chorioamnionitis was defined as: (1) infiltration of the extraplacental membranes by polymorphonuclear leukocytes, (2) accumu lation of polymorphonuclear leukocytes in the intervillous space immediately below the chorionic plate, (3) infiltration of the chorionic plate by leukocytes, and (4) angiitis of the umbilical vessels. The number of cases with patho logic evidence of chorioamnionitis was compared for both term (;. 37 weeks) and preterm (< 37 weeks) births by means of the x ~ test for statistical significance. Fifty patients with sin gleton pregnancies who were found to be in premature labor were enrolled in the second phase of the study after meeting several criteria. Each patient was required : (l) to be between 24 and 36 weeks' gestation; (2) to have intact membranes; (3) to be undergoing uterine contractions every 5 minutes for at least 1 hour, which were unresponsive to intra venous hydration; (4) to have a cervix judged favorable for tocolysis; (5) to manifest no evidence of fetal dis tress; (6) to be receiving no tocolytic drugs; (7) to have no evidence of present or recent clinical infection. A specimen of urine for culture was obtained from all patients at the time of admission in order to detect any current urinary tract infection. Of the 50 patients enrolled in the study, 10 were excluded for failure to meet the above requirements. These 10 cases included one twin gestation, two patients already receiving rito drine, one with a gestational age of < 24 weeks, one with a gestational age of > 37 weeks, two receiving am picillin for a recent urinary tract infection , two with urine cultures demonstrating the presence of> 10" bac
l
% 0 40 8.3
p
< 0.05 < 0.05
teria, and one patient with a fever of unknown origin. The control groups consisted of ( 1) 10 patients between 24 and 36 weeks' gestation who were not in labor and (2) 12 patients in labor at term with intact membranes. Blood for a quantitative C-reactive protein assay (a nephelometric assay with the Beckman Immunochem istry Analyzer) was obtained from the study group and the two control groups as part of the patient's routine admission laboratory tests .. The preterm patients were treated with intravenous ritodrine according to a stan dard obstetrics protocol by the obstetrics staff. Each case was managed without knowledge of the C-reactive protein results. Tocolytic therapy was abandoned if spontaneous rupture of the amniotic membranes oc curred or if advanced cervical dilatation and efface ment were encountered. Following delivery the pla centa was examined microscopically for evidence of chorioamnionitis. The C-reactive protein levels of the preterm labor patients were then compared to those of the control groups for both the numbers with patho logic evidence of chorioamnionitis and the degree to which gestation was prolonged by tocolysis . Statistical analysis was performed with Student's t test and the x ~ test; results were considered statistically significant if the confidence interval (p) was <0.05 .
Results Three hundred sixty-three (22.3%) of the 1631 term deliveries showed pathologic evidence of chorioam nionitis as compared to 102 (32.9%) of the 3 10 preterm deliveries during the 9-momh study period. This dif ference is statistically significant (p < 0.05). There was no significant difference in demographic characteristics (age, gravidity, parity, number of abor
644
Potkul, Moawad, and Ponto
November 15, 1985 Am J Obstet Gynecol
Table III. Correlation between C-reactive protein and pathologic evidence of chorioamnionitis Negative C-reactive protein Chorioamnionitis Group
N
n
Preterm labor T erm labor control
38* 12
Total
50
I
Positive C-reative protein
No chorioamnionitis
%
n
4 3
18.2
18 8
7
21.2
I
Chorioamnionitis
%
n
81.8
10 0
78.8
26
I
No chorioamnionitis
I
n
% 62.5
%
p
6
37.5
<0.01 NS
7
41.2
<0.01
I
10
58 .8
*Pathology reports were not available in two patients.
Table IV. C-reactive protein correlated with prolongation of gestation Days to delivery C-reactive protein
N
Mean
Negative Positive
24 16
35.5 14.4
I
Gestational age at delivery (wk) SE
Mean
4.1 5.0
36.9 33.2
1
0.6 0.7
< 0.01
p
SE
< 0.01
Table V. C-reactive protein correlated with results of tocolysis Successful C-reactive protein
N
n
Negative Positive
24 16
23 6
p
I
Failed
%
n
95.8 37.5
10
I
I
%
n
4.2 62.5
13 2
< 0.001
tions, and gestational age) between the study and the control group patients (Table I). In an attempt to apply the C-reactive protein assay as a highly sensitive but somewhat less specific clinical test, levels of >0.7 were considered positive for the remainder of the analysis. With the use of this value the C-reactive protein assay was found to be positive for 40% of the preterm labor group . There were no cases with positive C-reactive protein assays among pre term patients who were not in labor and only an 8% incidence among term patients in labor (Table II). The differences between the preterm labor group and each of the two control groups were significant. The C-reactive protein test results were found to cor respond reasonably well with the microscopic diagnosis of chorioamnionitis, which was present in 62.5% of those patients with a positive C-reactive protein assay (Table III). None of these patients exhibited any of the clinical signs of chorioamnionitis. In two of the pa tients with negative results for both the pathologic and C-reactive protein tests and in three patients with pos itive results in both cases, the membranes had ruptured more than 12 hours before delivery. Tocolysis was significantly more successful in the group of patients with negative C-reactive protein assay results. The mean number of days to delivery and the
Preterm birth
Term birth
1
%
n
54.2
II 14
12.5
1
% 45.8 87 .5
< 0.01
mean gestational age at delivery were 35.5 days and 36.9 weeks for the group with negative C-reactive pro tein versus 14.4 days and 33.2 weeks for the group with positive C-reactive protein (Table IV) . Tocolysis was successful in prolonging gestation for 1 week in 23 of the 24 patients (95.8%) witha negative C-reactive protein value (Table V) . The single failure occurred in a patient who required discontinuation of ritodrine therapy at 30.5 weeks because of severe ma ternal tachycardia. The contractions, which had ceased during ritodrine therapy, recurred after termination of the drug. Gestation was prolonged for > 7 days for only six of the 16 patients (37 .5%) with a positive C-reactive protein assay. Gestation was prolonged to term (~37 weeks) in 54.2% of the patients with a negative e-re active protein value. The same was true for only 12.5% of the patients with positive C-reactive protein values. Comment
The results of this study point to subclinical cho rioamniotic infection as a contributing factor in the cause of preterm labor. C-reactive protein was evaluated as a possible marker for subclinical chorioamnionitis because of past evi dence indicating that the C-reactive protein test is more sensitive than the appearance of fever, leukocytosis,
C-reactive protein in preterm labor 645
Volume 153 Number 6
or fetal tachycardia in predicting the presence of in fection.6· 7 The level of C-reactive protein that was considered abnormal in this study was >0.7 mg/dl. This is in con trast to the cutoff point of 2 mg/dl chosen in two pre vious studies dealing with premature rupture of the membranes from our institution. 6 · 7 The 2 mg/dl level was arbitrarily chosen in an attempt to apply the C reactive protein assay as a highly specific but somewhat less sensitive clinical test of pending clinical infectious morbidity. This was necessary since decisions concern ing induction of labor are based largely on C-reactive protein levels. The 0.7 mg/dllevel chosen in this study was an at tempt to apply the C-reactive protein assay as a highly sensitive but somewhat less specific indicator of sub clinical infection rather than infectious morbidity. Quality control of the Beckman Automated Immu nochemistry system by the company and The Univer sity of Chicago Serology laboratory has confirmed that the C-reactive protein level is <0.6 mg/dl in a normal population. Several aspects of these results are noteworthy. Al though the number of cases is limited, we have shown that normal pregnancy and normal term labor do not raise C-reactive protein levels significantly. It is inter esting to note that 8% of the patients in labor at term had elevated C-reactive protein levels, a figure that ap proximates the 10% incidence of the presence of mi croorganisms in amniotic fluid obtained from patients in labor at term.' In contrast, 40% of our preterm labor study group had elevated C-reactive protein values . The histologic examination of placentas from preterm deliveries at our institution revealed that chorioam nionitis was present 32.9% of the time. Again, the in cidences of elevated C-reactive protein and histologic chorioamnionitis are fairly similar. It seems reasonable to argue that, in the absence of clinical infection of the urogenital or other systems, subclinical infection plays an etiologic role in the onset of preterm labor in ap proximately one third of preterm labor patients. We were also able to demonstrate that when the C reactive protein assay was negative, ritodrine therapy
was successful in virtually every case in which the drug was well tolerated . A recent study by Handwerker et al. 8 has also examined the success rate of tocolysis with respect to C-reactive protein levels. Their lower success rate may be due to the inclusion of cases of urinary tract infection. Our results indicate that it may become possible to predict the success or failure of tocolytic therapy on the basis of the results of the assay for C-reactive pro tein. We recommend, in cases of preterm labor with high C-reactive protein levels or in cases where 13-sym pathomimetic agents fail to stop preterm uterine con tractions, that the presence of infection as a contrib uting factor should be suspected, even in the absence of any clinical manifestation of the infectious process. The task of the prospective identification of causative agents, although a tedious and difficult one, is clearly called for when one considers that in our population approximately one third of cases of preterm labor may have an infectious cause. REFERENCES I. Bobitt JR, Ledger WJ. Unrecognized amnionitis and pre
maturity: a preliminary report.] Reprod Med 1977;19:1. 2. Bobitt JR, Ledger WJ. Amniotic fluid analysis. Its role in maternal and neonatal infection. Obstet Gynecol 1978; 51:56. 3. BobittJR, Hayslip CC, DamatoJD. Amniotic fluid infection as determined by transabdominal amniocentesis in patients with intact membranes in premature labor. AM J OBSTET GYNECOL 1981;140:947. 4. Miller JM , Pupkin MJ, Hill GB . Bacterial colonization of amniotic fluid from intact fetal membranes. AMJ OBSTET GYNECOL 1980; 136:796. 5. Lis twa HM, Dobek AS, CarpenterJ, et al. The predictability of intrauterine infection by analysis of amniotic fluid. Ob stet Gynecol 1976;48:31. 6. Ismail MA, Zinaman MJ, Lowensohn R, Moawad AH. The significance of C-reactive protein levels in women with pre mature rupture of membranes. AM J OBSTET GYNECOL 1985; 151 :541. 7. Evans MI , Hajj SN, Devoe LD, Angerman NS, Moawad AH . C-reactive protein as a predictor of infectious mor bidity with premature rupture of membranes. AMJ OBSTET GYNECO L 1980; 138:648. 8. Handwerker SM , Nergesh AT, Verma UL, et al. Correla tion of maternal serum C-reactive protein with outcome of tocolysis. Obstet Gynecol 1984;63:220.