The relationship between prematurely ruptured membranes and fetal immunoglobulin production LARS
L.
CEDERQVIST,
IOANNIS LINDA
A.
ZERVOUDAKIS,
C. EWOOL,
STEPHEN New
M.D.
York,
D. I\‘PW
M.D.
B.S.
LITWIN.
M.D.
York
Cord blood and maternal sera were studied in a series of 227 cases of prematurely ruptured membranes (PRM) with respect to: (1) fetal immunogkbulin (Ig) synthesis associated with PRM, (2) the interrelationship behnreen different Ig classes during lnfeotion, and (3) the relationship be&veen lg values and the duration pf PRM prior to the onset of tabor. A preliminary report from this laboratory, which indicated that a humoral fetal immune response occurred in some but not all cases of PRM, and that significant increases in either IgA or IgM oould be found, was confirmed. There was both clinical and immunobgic evidence of one peak of infection one to 12 hours after onset of PRM and another after 72 hours after onset of PRM, suggesting that some patients were infected before the clinical onset of PRM symptomat&gy. Increased lgA and/or IgM was found in 16.3% of infants with clinical evidence of infection. This was comparable to the 18.5% of patients with PRM who had elevated IgA and/or IgM without clinical evidence of infection. Further, there was no correlation between the severity of infection and the presence of Ig elevation. Based on the data in the present series, lg determination in cord blood cannot be used to distinguish cases of PRM with and without fetal infection. (AM. J. OBSTET. GYNECOL. 134:784, 1979.)
THE DETERMINATION of immunoglobulin (Ig) concentrations in cord blood has been used to evaluate the fetus for intrauterine infection.‘. ’ IgM is frequentl) increased in congenital infections but not in all cases.’ Both the type and severity of the congenital infection determine the degree of IgM elevation. Elevated levels of IgA, on the other hand, have not been widel) applied to the diagnosis of intrauterine infections, mainly due to the lack of sensitive methods for determination of cord blood levels. We have used a hemagglutination inhibition method in which sheep erythrocytes are coated with monospecific myeloma proteins to From Department of Obstetrics and Gynecology and thr Division of Human Genetics, Department of Medicine, The New York Hospital-Cornell Medical Cmter. Recezvrd
for publication
Rruised
October
I, 1978.
Octobrr
2, 1978
,4ccepted
June
29, 1978.
Refirint requests: Dr. Lars L. Cederqvzct, Department o/ Obstetrics and Gynecology, The Ne; York Hospital-Cornell Medical Center, 525 E. 68th St., ~‘Vew York, New York 10021. 784
determine IgA.2 and we found measurable levels in all 176 samples of cord blood tested during the third trimester of gestation.3 Genetic typing has proved that some or all of the cord blood IgA is of fetal origin and that no maternal IgA passes through the placenta under nortnal circumstances: A prelitninary report of 16 cases ot prematurely ruptured membranes (PRM) frotn this institution indicated that the fetus showed higher tord blood Ig calues in some cases of PRM.” There was individual variability in the Ig produced by the fetus when infection occurred: a significant increase in either or both IgA or IgM or no increase at all was found. These changes were inttbrpreted as result of immune response to ascending infection from the maternal genital tract. The purpose of the present study was to determine: (1) the fetal Ig levels associated with PRM in an enlarged series comprising 227 cases, (2) the interrelationship between different Ig classes when au imtnune response occurred, and (3) the relationship between Ig values to the time interval from PRM to the onset of labor. 0002-9378:7Yi1~0784+05$00.~0i0
0 lYi!l
The (:. V. Mmbv
Co.
Volume Number
134 7
Material and methods Serum and cord blood from 227 mothers with a history of PRM between the thirty-second and forty-first weeks of gestation were tested for IgA, IgD, IgG, and IgM classes and IgA, and IgAz subclasses. The results were compared with data obtained from 104 women who had normal deliveries at comparable times of gestation without: evidence of infection. IgA, its subclasses IgA, and IgA*, and IgD were semiquantitatively measured by a hemagglutinationinhibition method, which has been described in detail elsewhere.’ The lower limit of detection was 0.5 mg/lOO ml for IgA and its subclasses IgA, and IgA2 and 0.02 mg/ 100 ml for IgD. When IgA, IgAl, or IgA2 were not detected (co.5 mg/lOO ml) the sample was concentrated by an Amicon concentratioh selector Model CDS10 (Amicon Corporation, Lexington, Massachusetts) 5 to 5.5 times using nitrogen with 60 pounds per square inch positive pressure. Thus concentrations of IgA, IgAl, and IgAz above 0.09 to 0.10 nigi 100 ml could be determined. The lowest IgD level tested was 0.1 mg/lOO ml for samples positive at this dilution; the :!l samples of the PRM group and the 13 samples of th,e control group that were negative were tested and found to be negative at 0.02 mgi 100 ml and may have had levels below this value. IgG and IgM were determined on single radial immunodiffusion plate? (Kallestad, Inc., Chaska, Minnesota). Precipitation rings were measured after 24 hours for IgG and after 72 hours for IgM. Sensitivity was 5.4 mg/ 100 ml for IgM and 3.6 mg/ 100 ml for IgG.
Results Ig measurements. IgA, IgA,, IgA2, and IgM were detected in all samples tested. IgD was measurable in 90.7% of the cord sera in the PRM group and in 84.7% of the cord sera in the control group. Cord blood values for IgM and IgG were consistent with normal ‘data obtained at other laboratories. In studies from other laboratories, measurable levels of IgA were reported for 5.4%’ to one third of cord blood’ when immunodiffusion technique was used and in all specimens when a radioimmunoassay technique was used.’ Our present methodology permitted us to measure IgA in all cord bloods. IgA data have been previously reported from this laboratory.4 IgD has previously been reported in only 4.5% to 8.0% of tested cord sera.“, ” The present data in this study are consistent with a recent study from this laboratory, where IgD was measurable in 84% of cord sera at term when the hemagglutination-inhibition technique was used.12 Cord Ig levels in the PRM group. Comparison of
785
Prematurely ruptured membranes and fetal immunoglobulin
Table I. Ig levels in cord blood obtained at delivery from patients with a history of PRM and from normal control subjects Control soup 4T
W W, IgA, I@’ 16 W
PRM
No.
mg/lOO
ml
NO.
104 104 104 104 104 104
1.9 -t 1.5 f 0.41 + 0.57 2 1,211 2 11.0 t
1.1 0.9 0.30 0.89 555 5.7
227 227 227 227 227 227
g&p mgllO0
3.0 2.3 0.41 0.43 1,135 17.1
Sip+ml
cance
p < 0.01 2 5.7 ‘-’ 4.7 p < 0.05 t 0.91 NS* _’ 0.4 1 NS* + 496 NS* + 34.3 p < 0.01
No. refers to the number of samples tested. Significance was determined between the control group and the PRM group by Student’s t test in unpaired samples. *NS = Difference not statistically significant. cord blood levels in patients with PRM to the control group demonstrated clear evidence of an immune response occurring in the PRM group (Table I). There were significantly increased levels of IgA (p < O.Ol), IgA, (p < 0.05), and IgM (p < 0.01) in the PRM group. No significant difference between the control group and the PRM group was found for IgAz, IgD, or IgG. The elevated levels of IgA in the PRM group were mainly due to an increase in the IgA, subgroup, while IgAz remained low. When the individual Ig levels in the PRM group were analyzed, a wide range of values was found for IgA and IgM, suggesting that some but not all subjects had elevated IgA and/or IgM in association with PRM. When two standard deviations on either side of the mean of the control values were used as the normal range, 28(12.3%x) had elevated IgA, 21 (9.3%) had elevated IgM, and 6 (2.6%) had increased levels of both IgA and IgM. Elevated IgA or IgM levels were found in 3% of the control group. Interrelationship between different Ig classes. When the correlation coefficient r for paired samples13 was applied, no correlation was found in the normal control group between IgA and IgG (r = 0.021), IgA and IgG (r = O.OSl), IgA and IgD (r = 0.017), IgM and IgD (r = 0.002), IgD and IgG (r = 0.008), IgG and IgM (r = 0.012), IgA, and IgD (r = 0.014), IgA, and IgG (r = 0.064), IgA, and IgM (r = 0.028), IgAz and IgD (r = 0.028). IgA, and IgG (r = 0.082), or between IgA and IgM (r = 0.084). In the group of 227 patients with PRM there was an inverse correlation between IgG and IgM (r = -0.231; t = 3.517; p < 0.001) and a positive correlation between IgG and IgD (r = 0.148; t = 2.242; p < 0.05). No correlation was found between IgA and IgG (r = 0.026), IgA and IgM (r = 0.028). IgA and IgD
Table II. Comparison Hours
between PRM
1-12 13-25 25-36 37-72 273
between and labor
Ig levels and different No.
142 31 19 15 20
time intervals
w
3.2 2.2 1.7 2.2 5.2
f 6.4* 2 1.4 ?z 1.3 + 1.6 -+ 7.3*
between W
15.9 9.7 14.7 55.7 25.9
No. refers to the number of samples tested. Significance was determined different PRM subgroups by Student’s t test in unpaired samples. *p < 0.05. tp < 0.01. (r = 0.149), or between IgM and IgD (r = 0.011). In the PRM subgroup of 43 patients with elevated IgA, IgM, or both in cord serum, there was no correlation between IgA and IgM (r = 0.234), IgA and IgG (r = 0.062), IgM and IgG (r = O.Oi3), IgM and IgD (r = 0.106), or IgA and IgD (r = 0.013). There was, however, as in the total PRM group, a positive correlation between IgG and IgD values (r = 0.558; t = 4.301; p < 0.001). Relationship between duration of PRM and cord blood Ig levels. There was immunologic evidence of one peak of infection one to 12 hours after onset of PRM and another after 72 hours after onset of PRM, suggesting that some patients were infected before the clinical onset of PRM symptomatology. When compared to normal control subjects, IgA was significantly elevated in the patients with one to 12 hours between PRM and labor (p < 0.05) and in the patients with more than 72 hours’ duration between PRM and labor (p < 0.05), while there was no significant difference in patients with 13 to 24, 25 to 36, and 37 to 72 hours between PRM and labor (Table II). IgM levels were also significantly elevated in the first group (one to 12 hours between PRM and labor; p < 0.05). Even though the Ig levels were elevated in the patients with 37 to 72 and over 73 hours between PRM and labor, they were not significantly raised due to the big standard deviation. The IgG levels were significantly decreased in patients with more than 72 hours between PRM and labor (p < 0.001). There was no difference in IgD levels in patients with different time intervals between PRM and labor. Relationship between cord blood Ig levels and fetal and maternal infections. Fifty-seven (25.1%) of the 227 women with PRM had clinical evidence of uterine infection as compared to seven (6.7%) of the control subjects. Of the 227 children born to mothers with PRM, 43 (18.9%) had clinical evidence of infection compared to 7 (6.7%) of the control subjects. Of these 43 infants,
PRM and label 49
2 28.9* -c 7.5 ” 19.0 f 90.5 -c 55.0
0.47 t 0.44 0.39 +- 0.3 1 0.36 c 0.28 0.50 d 0.42 0.3 1 lr: 0.33
between the control group
1
I&
1,216 1,183 1,059 993 718 (see Table
+ 526 t 395 It 224 i 456 I515? II
and the
seven (16.3%) had increased levels of IgA, IgM, or both in the cord blood. Seventeen of these 43 infants had PRM one to 12 hours, five had PRM 13 to 36 hours. five had PRM 37 to 72 hours, and five had PRM more than 72 hours prior to the onset of labor. There was no correlation between clinical evidence of infection associated with PRM and increased Ig levels. The microorganisms recovered from the placenta in 25 patients with evidence of maternal and/or fetal infection were Streptococcus viridans in 13 cases, Staphylococcus efidermidis in eight, Streptococcus f&calis in seven, peptostreptococci in five, Eschrrichia coli in five, bacteroides species in five, diphtheroides in 4, &streptococci group B in 4, klebsiella in 2, and Streptococcus breuis in two cases. In addition one case each of C~OJtridium perfiingens, Proteus mirabilis, Staphylococcus aureus, and fusobacterium species was isolated.
Case reports In the PRM group there were three perinatal deaths (1.32%) as compared to 1 (0.9%) in the control group. Case 1. A 2,800 gram infant was delivered after a 29 hour history of PriM prior to the onset of labor. It died after birth from septic shock with central nervous system damage. From the uterus and the fetus group B /3-hemolytic streptococci were cultured. Case 2. A 2,160 gram infant was delivered afker a 25 hour history oi‘ PRM. It died from interstitial pneumonitis after birth. Although peptostreptococci and bacteroides species were ide’ntified in uterine cultures, fetal cultures were negative. Case 3. A 2,500 gram infant died shortly after birth of septic shock after a history of nine days of PRhi. Group B /?-hemolytic streptococci were cultured from the fetal side of the placenta. All three children who died of infections following PRM had normal levels of IgA, IgM, IgD, and IgG. Case 4. One infant, weighing 2,840 grams, from the control group died of respiratory distress syndrome. At autopsy E. coli was identified in heatt blood cultures. This infant also had normal Ig levels.
Volume Number
134
Comment This enlarged series comprising 227 cases of PRM confirms a preliminary report from this laboratory indicating a humoral immune response by the fetus in some but nol: all patients with PRM. There is also individual variability in the type of immunoglobulins synthetized by the fetus when infection occurs: A significant increase in either or both IgA and IgM or no increase at all1 may be found.5 It is interesting to note that two patterns of infection can be identified when the time interval between PRM and the onset of labor is considered. First, there is both clinical and immunologic evidence of one peak after one to 12 hours of PRM. A second peak can be identified after 72 hours of PRM. The first peak suggests that the infection was present before the PRM, and it might have been the direct cause of PRM. The combination of intact membranes and amnionitis has been well established.14 Bobitt and Ledger15 noted significant bacterial counts at the time of membrane rupture in patients who subsequently were delivered of premature infants. Their study and the present series suggest fetal exposure to infectious agents in some patients before the clinical onset of symptomatology of membrane rupture. Thiz, fact is important for this group should, if identified, be delivered as soon as possible and not wait for spontaneous onset of labor or induction of labor at a later time. The second peak represents the expected increased infection rate after prolonged PRM. When the correlation between different Ig classes in the individual samples was tested, no correlation between IgA and IgM was found, indicating that these fetal immunoglobulins responded independently to an antigenic challenge. This finding is important as dem-
REFERENCES
1. Alford,
2.
3.
4.
5.
787
Prematurely ruptured membranes and fetal immunoglobulin
7
C. A., Schaefer, J., Blankenship, W. J., Straumfiord, N. V., and Cassadv. G.: A correlative immunolopic. microbiologic and cliniCa approach to the diagnosis-o; acute and chronic infection in newborn-infants, N. Engl. J. Med. 27’7:347, 1967. Cederqvist, L. L., Kimball, A. C., Ewool, L. C., and Litwin, S. D.: Fetal immune response following congenital toxoplasmosis, Obstet. Gynecol. 50:200, 1977. Cederqvist, L. L., Ewool, L. C., and Litwin, S. D.: The effect of fetal age, birth weight, and sex on cord blood immunoglabulin values, AM. J. OBSTET. GYNECOL. 131: 520, 1978. Cederqvist, L. L., and Litwin, S. D.: Production of alpha, and alpha* immunoglobulin heavy chains during fetal life, J. Immunol. 112:605, 1974. Cederqvist, L. L., Francis, L. C., Zervoudakis, I. A., Becker, C. (G., and Litwin, S. D.: Fetal immune response following prematurely ruptured membranes, AM. J. OBSTET. G’INECOL. 126:321, 1976.
onstration of increased IgM levels in cord blood is now used routinely to diagnose congenital or neonatal infections. IgA determination should also be included in cord blood studies, provided a sensitive method is available.16 The inverse relationship between IgM and IgG in the PRM group is hard to explain. It could be dependent on impaired placental transfer of IgG when infection occurs or be due to altered physiologic conditions in the placenta after membrane rupture. If these explanations were true, an inverse correlation between IgG and IgA would also be expected. The positive correlation between IgG and IgD is also difficult to explain ‘at the present time. The data, however, strongly indicate that IgD is not involved in the fetal immune response. Further studies are needed to clarify the relationship between maternal IgG and fetal immunoglobulins. Only 16.3% of infants with clinical evidence of infection had increased IgA and/or IgM. On the other hand 18.5% of the patients with PRM had elevated IgA and/or Ighl without clinical evidence of infection. This fact does not exclude that these latter children were exposed to infection in utero and overcame it with help of accelerated Ig synthesis. There was no association between the severity of infection and the presence of elevation of Ig levels. It is interesting to note that the three infants who died of infection associated with PRM had normal Ig levels. It can be speculated that this was due to either immature immune competence or immune paralysis but proof is lacking. Based on the data in the present series, Ig determination in cord blood does not lend itself to screening of perinatal infections associated with PRM.
6-.
Mancini, G., Carbonara, A. O., and Heremans, J. R.: Immunochemical quantitation of antigens by single radial
immunodiffusion, 7. Miller, M. J.,
Quantitation
Immunochemistry Sunshine, P. J.,
and
2:235, 1965. Remington,
Y. S.:
of cord serum IgM and IgA as a screening
procedure to detect congenital infection. infants, J. Pediatr. 75: 1287, 1969. Stiehm, E. R., and Fudenberg, H. H.:
immune
globulins
in health
Pediatrics
37:7 15, 1966.
Faulkner,
W., and Borella,
Results
in 5,006
Serum
levels
of
and disease. A survey, L.: Measurements
of IgA
levels in human cord serum by a new radioimmunoassay, J. Immunol. 105:786, 1970. Evans, H. E., Akpata, S. O., and Glass, L.: Serum immunoglobulin levels in premature and full-term infants, Am. J. Clin. Pathol. 56:416, 1971. Leslie, G. A., and Swate, T. R.: Structure and biologic functions of human IgD. I. The presence of immunoglobulin D in human cord sera, J. Immunol. 109:47, 1972.
788
Cederqvist et al.
1‘2. Cederqvist, I,. L., Ewool, L. C., and Litwin, S. D.: the fetal immune response, Scand. ,J. Immunol. 1977. 13. Batson, H. C.: An Introduction to Statistics in Science, Minneapolis, 1965, Burgess Publishing 14. Benirschke, I(.: Routes and types of infection in and newborn, Am. J. Dis. Child. 90:714. 1960.
IgD and 6:821, Medical Co. the fetus
15. Bobitt, J. K., and Ledger. u’. J.: Amn~c)tic Ilr~ul .1n;1114\ 11s role in maternal and neonatal inlfxrion, OfNet. ‘C;\necol. 51:56, 1978. 16. Ceder-c@, I.. I... Ewool. I.. C.,,, Bonsnes, K. M’. ,cnti l.ltwin, S. D.: Detectability and pattern of inttnurlogic~hlllins in normal amniotic fluid throughwit ~e~tatiw :\u J. c)BST.Fl. (;TNE( 01.. 130:220, t:)ix.
L.
CEDERQVIST,
IOANNIS LINDA
A.
ZERVOUDAKIS,
C. EWOOL,
STEPHEN New
M.D.
York,
D. I\‘PW
M.D.
B.S.
LITWIN.
M.D.
York
Cord blood and maternal sera were studied in a series of 227 cases of prematurely ruptured membranes (PRM) with respect to: (1) fetal immunogkbulin (Ig) synthesis associated with PRM, (2) the interrelationship behnreen different Ig classes during lnfeotion, and (3) the relationship be&veen lg values and the duration pf PRM prior to the onset of tabor. A preliminary report from this laboratory, which indicated that a humoral fetal immune response occurred in some but not all cases of PRM, and that significant increases in either IgA or IgM oould be found, was confirmed. There was both clinical and immunobgic evidence of one peak of infection one to 12 hours after onset of PRM and another after 72 hours after onset of PRM, suggesting that some patients were infected before the clinical onset of PRM symptomat&gy. Increased lgA and/or IgM was found in 16.3% of infants with clinical evidence of infection. This was comparable to the 18.5% of patients with PRM who had elevated IgA and/or IgM without clinical evidence of infection. Further, there was no correlation between the severity of infection and the presence of Ig elevation. Based on the data in the present series, lg determination in cord blood cannot be used to distinguish cases of PRM with and without fetal infection. (AM. J. OBSTET. GYNECOL. 134:784, 1979.)
THE DETERMINATION of immunoglobulin (Ig) concentrations in cord blood has been used to evaluate the fetus for intrauterine infection.‘. ’ IgM is frequentl) increased in congenital infections but not in all cases.’ Both the type and severity of the congenital infection determine the degree of IgM elevation. Elevated levels of IgA, on the other hand, have not been widel) applied to the diagnosis of intrauterine infections, mainly due to the lack of sensitive methods for determination of cord blood levels. We have used a hemagglutination inhibition method in which sheep erythrocytes are coated with monospecific myeloma proteins to From Department of Obstetrics and Gynecology and thr Division of Human Genetics, Department of Medicine, The New York Hospital-Cornell Medical Cmter. Recezvrd
for publication
Rruised
October
I, 1978.
Octobrr
2, 1978
,4ccepted
June
29, 1978.
Refirint requests: Dr. Lars L. Cederqvzct, Department o/ Obstetrics and Gynecology, The Ne; York Hospital-Cornell Medical Center, 525 E. 68th St., ~‘Vew York, New York 10021. 784
determine IgA.2 and we found measurable levels in all 176 samples of cord blood tested during the third trimester of gestation.3 Genetic typing has proved that some or all of the cord blood IgA is of fetal origin and that no maternal IgA passes through the placenta under nortnal circumstances: A prelitninary report of 16 cases ot prematurely ruptured membranes (PRM) frotn this institution indicated that the fetus showed higher tord blood Ig calues in some cases of PRM.” There was individual variability in the Ig produced by the fetus when infection occurred: a significant increase in either or both IgA or IgM or no increase at all was found. These changes were inttbrpreted as result of immune response to ascending infection from the maternal genital tract. The purpose of the present study was to determine: (1) the fetal Ig levels associated with PRM in an enlarged series comprising 227 cases, (2) the interrelationship between different Ig classes when au imtnune response occurred, and (3) the relationship between Ig values to the time interval from PRM to the onset of labor. 0002-9378:7Yi1~0784+05$00.~0i0
0 lYi!l
The (:. V. Mmbv
Co.
Volume Number
134 7
Material and methods Serum and cord blood from 227 mothers with a history of PRM between the thirty-second and forty-first weeks of gestation were tested for IgA, IgD, IgG, and IgM classes and IgA, and IgAz subclasses. The results were compared with data obtained from 104 women who had normal deliveries at comparable times of gestation without: evidence of infection. IgA, its subclasses IgA, and IgA*, and IgD were semiquantitatively measured by a hemagglutinationinhibition method, which has been described in detail elsewhere.’ The lower limit of detection was 0.5 mg/lOO ml for IgA and its subclasses IgA, and IgA2 and 0.02 mg/ 100 ml for IgD. When IgA, IgAl, or IgA2 were not detected (co.5 mg/lOO ml) the sample was concentrated by an Amicon concentratioh selector Model CDS10 (Amicon Corporation, Lexington, Massachusetts) 5 to 5.5 times using nitrogen with 60 pounds per square inch positive pressure. Thus concentrations of IgA, IgAl, and IgAz above 0.09 to 0.10 nigi 100 ml could be determined. The lowest IgD level tested was 0.1 mg/lOO ml for samples positive at this dilution; the :!l samples of the PRM group and the 13 samples of th,e control group that were negative were tested and found to be negative at 0.02 mgi 100 ml and may have had levels below this value. IgG and IgM were determined on single radial immunodiffusion plate? (Kallestad, Inc., Chaska, Minnesota). Precipitation rings were measured after 24 hours for IgG and after 72 hours for IgM. Sensitivity was 5.4 mg/ 100 ml for IgM and 3.6 mg/ 100 ml for IgG.
Results Ig measurements. IgA, IgA,, IgA2, and IgM were detected in all samples tested. IgD was measurable in 90.7% of the cord sera in the PRM group and in 84.7% of the cord sera in the control group. Cord blood values for IgM and IgG were consistent with normal ‘data obtained at other laboratories. In studies from other laboratories, measurable levels of IgA were reported for 5.4%’ to one third of cord blood’ when immunodiffusion technique was used and in all specimens when a radioimmunoassay technique was used.’ Our present methodology permitted us to measure IgA in all cord bloods. IgA data have been previously reported from this laboratory.4 IgD has previously been reported in only 4.5% to 8.0% of tested cord sera.“, ” The present data in this study are consistent with a recent study from this laboratory, where IgD was measurable in 84% of cord sera at term when the hemagglutination-inhibition technique was used.12 Cord Ig levels in the PRM group. Comparison of
785
Prematurely ruptured membranes and fetal immunoglobulin
Table I. Ig levels in cord blood obtained at delivery from patients with a history of PRM and from normal control subjects Control soup 4T
W W, IgA, I@’ 16 W
PRM
No.
mg/lOO
ml
NO.
104 104 104 104 104 104
1.9 -t 1.5 f 0.41 + 0.57 2 1,211 2 11.0 t
1.1 0.9 0.30 0.89 555 5.7
227 227 227 227 227 227
g&p mgllO0
3.0 2.3 0.41 0.43 1,135 17.1
Sip+ml
cance
p < 0.01 2 5.7 ‘-’ 4.7 p < 0.05 t 0.91 NS* _’ 0.4 1 NS* + 496 NS* + 34.3 p < 0.01
No. refers to the number of samples tested. Significance was determined between the control group and the PRM group by Student’s t test in unpaired samples. *NS = Difference not statistically significant. cord blood levels in patients with PRM to the control group demonstrated clear evidence of an immune response occurring in the PRM group (Table I). There were significantly increased levels of IgA (p < O.Ol), IgA, (p < 0.05), and IgM (p < 0.01) in the PRM group. No significant difference between the control group and the PRM group was found for IgAz, IgD, or IgG. The elevated levels of IgA in the PRM group were mainly due to an increase in the IgA, subgroup, while IgAz remained low. When the individual Ig levels in the PRM group were analyzed, a wide range of values was found for IgA and IgM, suggesting that some but not all subjects had elevated IgA and/or IgM in association with PRM. When two standard deviations on either side of the mean of the control values were used as the normal range, 28(12.3%x) had elevated IgA, 21 (9.3%) had elevated IgM, and 6 (2.6%) had increased levels of both IgA and IgM. Elevated IgA or IgM levels were found in 3% of the control group. Interrelationship between different Ig classes. When the correlation coefficient r for paired samples13 was applied, no correlation was found in the normal control group between IgA and IgG (r = 0.021), IgA and IgG (r = O.OSl), IgA and IgD (r = 0.017), IgM and IgD (r = 0.002), IgD and IgG (r = 0.008), IgG and IgM (r = 0.012), IgA, and IgD (r = 0.014), IgA, and IgG (r = 0.064), IgA, and IgM (r = 0.028), IgAz and IgD (r = 0.028). IgA, and IgG (r = 0.082), or between IgA and IgM (r = 0.084). In the group of 227 patients with PRM there was an inverse correlation between IgG and IgM (r = -0.231; t = 3.517; p < 0.001) and a positive correlation between IgG and IgD (r = 0.148; t = 2.242; p < 0.05). No correlation was found between IgA and IgG (r = 0.026), IgA and IgM (r = 0.028). IgA and IgD
Table II. Comparison Hours
between PRM
1-12 13-25 25-36 37-72 273
between and labor
Ig levels and different No.
142 31 19 15 20
time intervals
w
3.2 2.2 1.7 2.2 5.2
f 6.4* 2 1.4 ?z 1.3 + 1.6 -+ 7.3*
between W
15.9 9.7 14.7 55.7 25.9
No. refers to the number of samples tested. Significance was determined different PRM subgroups by Student’s t test in unpaired samples. *p < 0.05. tp < 0.01. (r = 0.149), or between IgM and IgD (r = 0.011). In the PRM subgroup of 43 patients with elevated IgA, IgM, or both in cord serum, there was no correlation between IgA and IgM (r = 0.234), IgA and IgG (r = 0.062), IgM and IgG (r = O.Oi3), IgM and IgD (r = 0.106), or IgA and IgD (r = 0.013). There was, however, as in the total PRM group, a positive correlation between IgG and IgD values (r = 0.558; t = 4.301; p < 0.001). Relationship between duration of PRM and cord blood Ig levels. There was immunologic evidence of one peak of infection one to 12 hours after onset of PRM and another after 72 hours after onset of PRM, suggesting that some patients were infected before the clinical onset of PRM symptomatology. When compared to normal control subjects, IgA was significantly elevated in the patients with one to 12 hours between PRM and labor (p < 0.05) and in the patients with more than 72 hours’ duration between PRM and labor (p < 0.05), while there was no significant difference in patients with 13 to 24, 25 to 36, and 37 to 72 hours between PRM and labor (Table II). IgM levels were also significantly elevated in the first group (one to 12 hours between PRM and labor; p < 0.05). Even though the Ig levels were elevated in the patients with 37 to 72 and over 73 hours between PRM and labor, they were not significantly raised due to the big standard deviation. The IgG levels were significantly decreased in patients with more than 72 hours between PRM and labor (p < 0.001). There was no difference in IgD levels in patients with different time intervals between PRM and labor. Relationship between cord blood Ig levels and fetal and maternal infections. Fifty-seven (25.1%) of the 227 women with PRM had clinical evidence of uterine infection as compared to seven (6.7%) of the control subjects. Of the 227 children born to mothers with PRM, 43 (18.9%) had clinical evidence of infection compared to 7 (6.7%) of the control subjects. Of these 43 infants,
PRM and label 49
2 28.9* -c 7.5 ” 19.0 f 90.5 -c 55.0
0.47 t 0.44 0.39 +- 0.3 1 0.36 c 0.28 0.50 d 0.42 0.3 1 lr: 0.33
between the control group
1
I&
1,216 1,183 1,059 993 718 (see Table
+ 526 t 395 It 224 i 456 I515? II
and the
seven (16.3%) had increased levels of IgA, IgM, or both in the cord blood. Seventeen of these 43 infants had PRM one to 12 hours, five had PRM 13 to 36 hours. five had PRM 37 to 72 hours, and five had PRM more than 72 hours prior to the onset of labor. There was no correlation between clinical evidence of infection associated with PRM and increased Ig levels. The microorganisms recovered from the placenta in 25 patients with evidence of maternal and/or fetal infection were Streptococcus viridans in 13 cases, Staphylococcus efidermidis in eight, Streptococcus f&calis in seven, peptostreptococci in five, Eschrrichia coli in five, bacteroides species in five, diphtheroides in 4, &streptococci group B in 4, klebsiella in 2, and Streptococcus breuis in two cases. In addition one case each of C~OJtridium perfiingens, Proteus mirabilis, Staphylococcus aureus, and fusobacterium species was isolated.
Case reports In the PRM group there were three perinatal deaths (1.32%) as compared to 1 (0.9%) in the control group. Case 1. A 2,800 gram infant was delivered after a 29 hour history of PriM prior to the onset of labor. It died after birth from septic shock with central nervous system damage. From the uterus and the fetus group B /3-hemolytic streptococci were cultured. Case 2. A 2,160 gram infant was delivered afker a 25 hour history oi‘ PRM. It died from interstitial pneumonitis after birth. Although peptostreptococci and bacteroides species were ide’ntified in uterine cultures, fetal cultures were negative. Case 3. A 2,500 gram infant died shortly after birth of septic shock after a history of nine days of PRhi. Group B /?-hemolytic streptococci were cultured from the fetal side of the placenta. All three children who died of infections following PRM had normal levels of IgA, IgM, IgD, and IgG. Case 4. One infant, weighing 2,840 grams, from the control group died of respiratory distress syndrome. At autopsy E. coli was identified in heatt blood cultures. This infant also had normal Ig levels.
Volume Number
134
Comment This enlarged series comprising 227 cases of PRM confirms a preliminary report from this laboratory indicating a humoral immune response by the fetus in some but nol: all patients with PRM. There is also individual variability in the type of immunoglobulins synthetized by the fetus when infection occurs: A significant increase in either or both IgA and IgM or no increase at all1 may be found.5 It is interesting to note that two patterns of infection can be identified when the time interval between PRM and the onset of labor is considered. First, there is both clinical and immunologic evidence of one peak after one to 12 hours of PRM. A second peak can be identified after 72 hours of PRM. The first peak suggests that the infection was present before the PRM, and it might have been the direct cause of PRM. The combination of intact membranes and amnionitis has been well established.14 Bobitt and Ledger15 noted significant bacterial counts at the time of membrane rupture in patients who subsequently were delivered of premature infants. Their study and the present series suggest fetal exposure to infectious agents in some patients before the clinical onset of symptomatology of membrane rupture. Thiz, fact is important for this group should, if identified, be delivered as soon as possible and not wait for spontaneous onset of labor or induction of labor at a later time. The second peak represents the expected increased infection rate after prolonged PRM. When the correlation between different Ig classes in the individual samples was tested, no correlation between IgA and IgM was found, indicating that these fetal immunoglobulins responded independently to an antigenic challenge. This finding is important as dem-
REFERENCES
1. Alford,
2.
3.
4.
5.
787
Prematurely ruptured membranes and fetal immunoglobulin
7
C. A., Schaefer, J., Blankenship, W. J., Straumfiord, N. V., and Cassadv. G.: A correlative immunolopic. microbiologic and cliniCa approach to the diagnosis-o; acute and chronic infection in newborn-infants, N. Engl. J. Med. 27’7:347, 1967. Cederqvist, L. L., Kimball, A. C., Ewool, L. C., and Litwin, S. D.: Fetal immune response following congenital toxoplasmosis, Obstet. Gynecol. 50:200, 1977. Cederqvist, L. L., Ewool, L. C., and Litwin, S. D.: The effect of fetal age, birth weight, and sex on cord blood immunoglabulin values, AM. J. OBSTET. GYNECOL. 131: 520, 1978. Cederqvist, L. L., and Litwin, S. D.: Production of alpha, and alpha* immunoglobulin heavy chains during fetal life, J. Immunol. 112:605, 1974. Cederqvist, L. L., Francis, L. C., Zervoudakis, I. A., Becker, C. (G., and Litwin, S. D.: Fetal immune response following prematurely ruptured membranes, AM. J. OBSTET. G’INECOL. 126:321, 1976.
onstration of increased IgM levels in cord blood is now used routinely to diagnose congenital or neonatal infections. IgA determination should also be included in cord blood studies, provided a sensitive method is available.16 The inverse relationship between IgM and IgG in the PRM group is hard to explain. It could be dependent on impaired placental transfer of IgG when infection occurs or be due to altered physiologic conditions in the placenta after membrane rupture. If these explanations were true, an inverse correlation between IgG and IgA would also be expected. The positive correlation between IgG and IgD is also difficult to explain ‘at the present time. The data, however, strongly indicate that IgD is not involved in the fetal immune response. Further studies are needed to clarify the relationship between maternal IgG and fetal immunoglobulins. Only 16.3% of infants with clinical evidence of infection had increased IgA and/or IgM. On the other hand 18.5% of the patients with PRM had elevated IgA and/or Ighl without clinical evidence of infection. This fact does not exclude that these latter children were exposed to infection in utero and overcame it with help of accelerated Ig synthesis. There was no association between the severity of infection and the presence of elevation of Ig levels. It is interesting to note that the three infants who died of infection associated with PRM had normal Ig levels. It can be speculated that this was due to either immature immune competence or immune paralysis but proof is lacking. Based on the data in the present series, Ig determination in cord blood does not lend itself to screening of perinatal infections associated with PRM.
6-.
Mancini, G., Carbonara, A. O., and Heremans, J. R.: Immunochemical quantitation of antigens by single radial
immunodiffusion, 7. Miller, M. J.,
Quantitation
Immunochemistry Sunshine, P. J.,
and
2:235, 1965. Remington,
Y. S.:
of cord serum IgM and IgA as a screening
procedure to detect congenital infection. infants, J. Pediatr. 75: 1287, 1969. Stiehm, E. R., and Fudenberg, H. H.:
immune
globulins
in health
Pediatrics
37:7 15, 1966.
Faulkner,
W., and Borella,
Results
in 5,006
Serum
levels
of
and disease. A survey, L.: Measurements
of IgA
levels in human cord serum by a new radioimmunoassay, J. Immunol. 105:786, 1970. Evans, H. E., Akpata, S. O., and Glass, L.: Serum immunoglobulin levels in premature and full-term infants, Am. J. Clin. Pathol. 56:416, 1971. Leslie, G. A., and Swate, T. R.: Structure and biologic functions of human IgD. I. The presence of immunoglobulin D in human cord sera, J. Immunol. 109:47, 1972.
788
Cederqvist et al.
1‘2. Cederqvist, I,. L., Ewool, L. C., and Litwin, S. D.: the fetal immune response, Scand. ,J. Immunol. 1977. 13. Batson, H. C.: An Introduction to Statistics in Science, Minneapolis, 1965, Burgess Publishing 14. Benirschke, I(.: Routes and types of infection in and newborn, Am. J. Dis. Child. 90:714. 1960.
IgD and 6:821, Medical Co. the fetus
15. Bobitt, J. K., and Ledger. u’. J.: Amn~c)tic Ilr~ul .1n;1114\ 11s role in maternal and neonatal inlfxrion, OfNet. ‘C;\necol. 51:56, 1978. 16. Ceder-c@, I.. I... Ewool. I.. C.,,, Bonsnes, K. M’. ,cnti l.ltwin, S. D.: Detectability and pattern of inttnurlogic~hlllins in normal amniotic fluid throughwit ~e~tatiw :\u J. c)BST.Fl. (;TNE( 01.. 130:220, t:)ix.
