- Email: [email protected]
Stillbirths and intrauterine infection, histologic chorioamnionitis and microbiological findings
GYNECOLOGY &OBSTETRICS International
Journal of Gynecology & Obstetrics 54 (1996) 115-123
Article
Stillbirths and intrauterine infection, histologic chorioamnionitis and microbiological findings S.R. Moyoa, I. HPgerstrand*b, L. Nystr6mc, S.A. Tswanaa, J. Blombergd, S. BergstCme, A. Ljunghd ‘Department of Medical Microbiology, University of Zimbabwe Medical School, Harare, Zimbabwe bDepartment of Pathology and Cytology, University Hospital, Land, Sweden =Departmentof Epidemiology and Public Health, University of LImeA. Ume& Sweden dDepartment of Medical Microbiology, University of Lund, Lund, Sweden cDepartment of International Health, Ullev~l University Hospital, Oslo, Norway Received 2 February
1996; revised 3 April 1996; accepted 18 April 1996
AbStrPCt
Objective: To analyze whether placental inflammation is associatedwith stillbirth in Zimbabwe. Method: Placentas from 66 stillbirths (>22 weeks’ gestation; patients with congenital malformations, diabetes or preeclampsia were excluded) and 66 term live births were studied for the presenceand severity of chorioamnionitis. The morphological results were compared with earlier presentedmicrobiological findings in the samematerial. Results: Chorioamnionitis was present in 79% of stillbirths and 30% of live births (O.R. 8.5, 95% C.I. 4.0-18). Nine percent of stillbirths but no live births presentedvasculitis of the chorionic plate, which verified an inflammatory responsefrom the infant (O.R. 14, 95% C.I. 2.8-72). The sametypes of microorganisms were isolated from stillbirths and liveborns, but Escherichia co/i and group B streptococci were more frequent among stillbirths. Conclusions: Morphological chorioamnionitis occurred 2.6 times more often in women with stillbirths than in women with live births. In 9% of stillbirths the infant showedan inflammatory response.Thus the infant was alive when the infection occurred and it is therefore reasonable to assumethat infection was the cause of death. Keywords: Chorioamnionitis; Vasculitis; Stillbirths; Developing country; Zimbabwe
1. Introduction Chorioamnionitis is characterized by. the presence of neutrophil leukocytes in the placenta. The inflammation starts in the fetal membranes l Corresponding 143307.
author, Tel.: +46 46 173418; Fax: +46 46
0020-7292/96/$.15.00 0 1996 International PII SOO20-7292(96)02705-l
and spreads through the space between the amnion and the chorion. It may reach the amniotic cavity and/or the basal and/or chorionic surface of the placenta and then the fetus, resulting in congenital pneumonia and/or funisitis [l]. The leukocytes originate from the mother [l]. However, in the walls of the chorionic and umbilical vessels and in
Federation of Gynecology and Obstetrics
116
S.R. Moyo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
the interstitial pulmonary tissue, the leukocytic infiltration is of fetal origin [l]. The leukocytes of the alveoli in congenital pneumonia were until recently regarded as maternal, but it has now been proven in male infants that many at least are fetal leukocytes 121. Whilst chorioamnionitis is an ascendinginfection reaching the placenta through the birth canal, villitis is a hematogenousinfection reaching the placenta from the blood vesselsof the mother and able to cause sepsisin the fetus [l]. In 1925,Johnson and Meyer [3] drew attention to the frequent combination of congenital pneumonia and inflammation of the placental membranes.Benirschke and Clifford 141studying 1300consecutive deliveries in the Boston Lying-in Hospital in 1959found inflammation of the membranes (chorioamnionitis) in 13% of cases. In a previous microbiological study of stillbirths [S] we found bacteria that had spread beyond the cervix in 33%of women with stillbirths as compared with 7% of women with live births. The difference was statistically significant (O.R. 6.3, 95% C.I. 2.8-14). This study concerns the morphological
signs of inflammation and their comparison with microbiological findings in the material from the mentioned study [5]. 2. Materials and methods
Placentas from 66 patients admitted to Harare Central Maternity Hospital with a confirmed diagnosis of intrauterine death in a singleton pregnancy of over 22 weeks’ gestation between March 1989and March 1991were included in the study. Patients with high blood pressure (diastolic ~90 mmHg, systolic L 140mmHg) or diabeteswere excluded from the study as well as those with stillborn infants with gross congenital abnormalities. For each case a referent was selectedamong women admitted during the same time period in spontaneouslabor at term prior to rupture of the membranes and before vaginal examination, and with the delivery resulting in a livebom infant. Specimensfor microbiological culture were taken aseptically and examined as described earlier. Specimenswere taken from the cervix of the moth-
Fig. 1. Section through half a placenta with the umbilical cord to the left (a), the free membranes to the right (b), (c) marks the basal plate (decidua basalis), (d) marks the chorionic plate and (e) the subchorionic space.
S. R. Moyo et al. /International
Journal of Gynecology & Obstetrics 54 (1996) 115-123
er, from the placental disk, the umbilical cord, and from the throat and ear of all the infants and in addition from the heart, blood, lung and liver of the stillborns. The results of serology in detecting the occurrence of viral (cytomegalovirus, rubella, herpes simplex virus 1 and 2, human immunodeficiency virus) and bacterial (syphilis) infection, and a Ch/amydia antigen test have been previously discussed [5]. The placentas were immersed in 4% formaldehyde for fixation. One piece was cut from the central part of each placenta including the chorionic as well as the basal plate. The tissue blocks were processed, embedded in paraffin and sectioned. Histologic slides were stained with hematoxylin and eosin. The presence of neutrophil leukocytes was estimated in the chorionic plate, the walls of the chorionic vessels(vasculitis), the subchorionic spaceand the basal plate (Fig. 1). The grading of positivity was as follows: grade 0, no leukocytes; grade I, <25 leukocytes/highpower field; grade II, 25-50 leukocytetiighpower field; and grade III, >50 leukocytes/high-
117
power field. Villous edema was scored according to the percentage of edematous villi [6], and sections with half or more of the villi with edematous stroma were classified as positive. The stroma of the chorionic villi were examined for the presenceof inflammatory cells (villitis). 2.1. statistical analysis
The data were analyzed using QUEST software [7]. The matching was dissolved to avoid loss of valuable information during analysis. This was possible since the ratio between matched and unmatched data was close to one [8]. The relative risks were estimated by odds ratios and 95% confidence intervals according to Miettinen [9].
3. Results
The prevalence of histologic and microbiological findings in casesand referents is summarized in Tables l-3 and Figs. 2-7.
Fig. 2. Chorioamnionitis: inflammation in the chorionic plate and subchorionic space. Bar = 40 my.
I18
S. R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
Table I Numbers (N) and prevalences (%) in cases(C) and referents (R) according to localization and grade of the inflammation Inflammatory grade
Chorionitis
Subchorionitis
Basal deciduit
Total case of
chorioamnionitis C N
0 I 11 III
30 26 9
Total
66
%
R N
%
C ~ N
46 39 I4
47 16 3
71 24 5
34 25 6
0
0
66
100
I2 100
%
R N
%
51 38 9
62 4 0
94 6 0
0
0
I2
66
100
65a
I2 100
66
C N
C N
%
R N
%
46
71
66
100
IS
23
0
0
3
4
0
0
0 0
66
100
100
R %
14 21 34 51 1523 3500 66
100
N
%
46 I7 3
69 26 5
66
100
*Basal plate was missing in one histological slide.
Table 2 Numbers (iV) and prevalences (%) ofsubjects wih chorioamnionitis in cases(C) and referents (R) by microbiology Total
Pathogens Not isolated
Isolated C N
R %
R
C
N
%
N
%
C
R
N
%
N
%
N
%
Chorioamnionitis No chorioamnionitis
52
79
20
30
I2
18
8
I2
46
70
II
27
14
I6 7
24
21
36 7
55
14
11
19
29
Total
66
100
66
100
43
66
39
59
23
35
27
41
Table 3 Prvalencesin cases(C; n = 66) and referents (R; n = 66). odds ratio (OR) and 95% confidenceintervals (CI) for histology and positive and negative microbiology Histology
Chorionitis
Sub-chorionitis
Basal deciduitis Chorioamnionitis, total
Microbiology
Prevalence in
OR
95 % CI
C
R
Positive Negative All
51.2 60.9 54.5
28.2
2.67
1.07-6.65
29.6
3.69
1.15-11.90
28.8
2.97
1.45-6.06
Positive Negative All
60.5 26. I
18.40
5.81-57.90
48.5
7.7 3.7 6.1
14.60
5.57-38.20
Positive Negative All
28.6 30.4
0.0 0.0
1.84-569.00
29.2
0.0
32.4 25.0 55.8
3.29-948.00
Positive Negative All
83.7 69.6 78.8
30.8 29.6 30.3
11.60
4.28-31.30
5.43 8.54
1.65-17.80 4.02-18.20
9.18
1.33-63.30
I .34-437.00
S.R. Mayo et al. /International
Fig. 3. Inflammation
Fig. 4. Chorioamnionitis: inflammation subchorionic space. Bar = 10 my.
Jourd
of Gynecology % Obstetrics 54 (19%) 115-123
119
grade III of the chorionic plate. Bar = 10 my.
grade II in the
Fig. 5. Deciduitis: inflammation grade III in the basal plate (trophoblasts, decidual cells and neutrophil leukocytes). Bar = 10 my.
120
S.R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
Fig. 6. Vasculitis grade III. Bar = 40 req.
3. I. Histology
Among stillbirths 52/66 (79%) had chorioamnionitis compared with 20/66 (30%) referents (O.R. 8.5, 95% CI 4.0-18) (Tables l-3). Thus chorioamnionitis occurred 2.6 times more often in women with stillbirths than in women with live births. Grade-II chorioamnionitis (25-50 leuko-
Fig. 7. Villitis: cytomegalovirus
10 my.
cytesibigh-power field) was found in one or more sites in 15 stillbirths and in three referents; gradeIII chorioamnionitis (> 50 leukocytes/high-power field) in three stillbirths, but in no referents. Inflammation in two sites (chorionic plate and subchorionic space) was seen in 26 casescompared with four in the referents, and inflammation in
inclusion seen in inset. Bar = 40 resp. 10 my.
S.R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
three sites (also in the basal plate) occurred in four stillbirths but in no referents. In no referent did the basal plate show infiltration of neutrophil leukocytes. Chorioamnionitis thus tended to be more frequent, more severe and more extensive in stillbirths than in referents (Table 1). Inflammation was observed in the chorionic plate (Figs. 2 and 3), the subchorionic space(Fig. 4) and the basal plate (Fig. 5) in 36, 32 and 19 stillbirths compared with 19, 4 and 0 referents, respectively (Table 1). The corresponding odds ratios were 3.0, 20 and 27, all of which were statistically significant (Table 3). In no case was the inflammation restricted only to the basal plate. Vasculitis in the chorionic plate (Fig. 6) was seenin 6/66 stillbirths, but in none of the referents (O.R. 14, 95% C.1, 2.8-72). No microbial agent was found in two of the six cases,Escherichiu coli was found in one and group B streptococcusin another. In the remaining two casesthe mothers were seropositive for syphilis, which however should not have caused the purulent vasculitis. Villous edemawas found in 8166placentas from stillbirths (12O/o),but in no placenta from a live birth, corresponding to an odds ratio of 19 (95% C.I. 4.0-94). Six of the eight placentas showed chorioamnionitis and one showed villitis compatible with cytomegalovirus infection. Bacterial organisms were not isolated from the case with villitis nor from three additional subjects. E. co&, group B streptococci, Candida albicuns and Staphylococcusaureus were isolated from the four remaining placentas. Villitis (Fig. 7) was found in 5/66 stillbirths and l/66 referents (O.R. 5.3, 95% C.I. 0.57-256). Cytomegalovirus was detected by immunohistochemistry in one of the placentas with villitis among the stillbirths (the samecase as above). In two casesthe mother had syphilis, but no explanation could be found for the remaining three cases. Toxoplasma cysts were not found (no immunohistochemical method was performed). The cytomegalovirus case was cytomegalovirus-IgM negative, compatible with a reactivated infection in the mother. 3.2. Microbiology In 43/66 stillbirths bacteria were isolated from
121
the placenta and/or from the neonate compared with 39/66referents (O.R. 1.3,95% C.I. 0.60-2.8). Bacteria were recovered from the placentas in 18 stillbirths and three live births. Bacterial growth from the internal organs occurred in 27/66 stillbirths (41%). The main bacterial speciesisolated were E. coli, group B streptococci, coagulasenegative staphylococci, S. aureus and Streptocoo cus viridans. E. coli was isolated in 17166casesand in 9166 referents (O.R. 2.2, 95 % C.I. 0.83-5.9). Group B streptococci were isolated in only eight casesand three referents (O.R. 2.9, 95% C.I. 0.65-18). Coagulase-negativestaphylococci were isolated in nine casesand 10 referents (O.R. 0.88, 95% C.I. 0.30-2.6). S. uureus was isolated in six stillbirths and one referent (O.R. 6.5, 95% C.I. 0.75-65). In 18166stillbirths (27%), the same bacterial species was found in three or more sites. The referents cannot be compared, since no cultures from internal organs were available. 3.3. Comparison between histology and microbiology In 36 casesand 12 referents, the histologic and microbiological results from the placenta were in agreementas regards presenceof chorioamnionitis (i.e. bacterial growth in at least one location) (Table 2). In seven casesand 19 referents, histologic and microbiological findings were in agreement regarding absenceof chorioamnionitis. In 16 casesthere was histologic chorioamnionitis, but no microbial agent was detected. The same was true for eight referents. In sevencasesand 27 referents there was no histologic chorioamnionitis, but microbiological agents were found, however they were mostly of low virulence such as diphtheroids and cr-hemolytic streptococci [5]. 4. Discussion This study demonstrated inflammation of the chorionic plate, the subchorionic spaceand/or the basal plate in 30% of normal deliveries. It is probably too low a figure considering that only one histologic section was examined in each case. The prevalence is difficult to compare with figures presented in other materials since neither the
122
S.R. Moyo et al. /international
Journal of Gynecology & Obstetrics 54 (19%) 115-123
method of collection nor the diagnostic criteria agree. In South Africa, Miiller et al. [lo] found a prevalence of 28%; in the USA, Hillier et al. [ 1l] found a prevalence of 16% in Seattle, and Zhang et al. [12] found a prevalence of 10% in St Louis, whilst a prevalenceof 4% among normal deliveries has beenreported from Connecticut [ 131.A prevalence of 5% has been reported in Toronto, Canada [ 141.Thus the prevalencesof histologic chorioamnionitis tend to be higher in developing than in developed countries. In this study histologic chorioamnionitis was more severeand more extensive in stillbirths than in referents. The prevalence found (79%) was higher than in similar studies performed elsewhere [14-171. However, we have counted any presence of neutrophil leukocytes in the described locations as an expression of chorioamnionitis. Olding [ 151 in 1966reported 55% in a material of stillbirths in Sweden. In Canada and in the USA, prevalences of 58% [14] and 38% [16], respectively, have been reported, and in a recent study from Mozambique, 68% [17]. Using the samecriteria as in the present study but with more histologic sections in approximately 100 stillbirths from Lund in Sweden, Tolockienk and Hagerstrand (unpublished observations) found 97% with histologic chorioamnionitis. It is thus obvious that the prevalence of histologic chorioamnionitis is higher in stillbirths than in normal deliveries. In the present study, six (9%) stillbirths showed vasculitis in the chorionic plate. This indicates that the infant has reacted against the infection with diapedesisof focally increased leukocytes [l] and was alive when infected. In the rest of the material it cannot be ruled out that the chorioamnionitis developed after intrauterine fetal death. In the stillbirths examined in Lund, Sweden,6% had fetal vasculitis. It is reasonableto assumethat the cause of death was also chorioamnionitis in many subjects without vasculitis. Since we only diagnosed chorioamnionitis cases reaching beyond the free membranes, the disease process must have been present for some time. Regarding the pathogenesis of chorioamnionitis, the inflammation of the membranestends to spread to the chorionic plate and the subchorionic spacerather than to the basal plate. This
might indicate that bacterial invasion from the membranes into the amniotic fluid is more common than from the basal plate via the maternal blood to the infant [1,6,13]. Villous edema was present only in stillbirths (12%). The majority of these infants, however, were preterm, and this figure thus agreeswith the report by Shen-Schwartzet al. [ 181.These authors assumevillous edema to be a lesion of the immature intermediate villi, and found no correlation with chorioamnionitis. The villous edema might have been the final causeof death in those preterm stillbirths. Mild chorioamnionitis might also cause defecation of meconium and death of the infant due to vasoconstriction [19]. Unfortunately, we were unable to obtain any information on meconium defecation in our material. Histologic chorioamnionitis (72 cases and referents) was indicated by the microbiological findings (48 casesand referents) in approximately 70%,which is in accordancewith other studies [6]. Histologic chorioamnionitis without bacterial growth might indicate a failure to retrieve the microbes or an inability of the prevailing microbes to grow out due to the inflammatory response in the placenta. Microbiological chorioamnionitis without inflammation might mean that the microbial invasion was recent and that the sampling was done before the onset of the inflammatory process. Provided that no contamination occurred with microbiological cultures, any bacteria isolated from the placenta would presumably represent microbiological chorioamnionitis. Cultures from ear, throat and umbilical cord might have been contaminated by the vaginal flora. Considering only placental cultures, 18 (27%) stillbirths and three (4.5%) live births had microbiological chorioamnionitis. If we add to the stillborn group those with any bacteria isolated in cultures from internal organs, signs of feto-placental infection occurred in 28/66 cases(42%) (51. Cultures from the internal organs of the stillborn infants and from the placenta are vital in proving chorioamnionitis. However, without cultures from ear and throat, more casesof histologic chorioamnionitis would have been left without a microbiological explanation.
S.R. Mayo et al. / Inremational Journal of Gynecology d Obstetrics 54 (19%) 115-123
Bacterial invasion of the amniotic cavity at term has been assumedto be a physiological phenomenon being either the consequenceor the cause of labor [20]. This assumption was basedon the tindings of microbial growth in 19% of amniotic fluid cultures during spontaneous parturition at term. The dominating organism was Ureaplasma urealyticum, an organism we did not try to isolate in our study. It might be argued that such a physiological process is offered sufficient time to develop into a manifest infection after the death of the infant. However, chorionic vasculitis in some of the infants provides evidence that they were alive at the onset of infection. Acknowledgments We would like to thank Mr J. Magwenzi, Dr Shylaja and Sr J. Moyo for their assistance.This study was supported by grant number 2.9999.10.3334from the University of Zimbabwe Research Board, grant number 26089.10 from the Black Medical Foundation Screening Committee, and by the Swedish Agency for Research Cooperation with Developing Countries (SAREC). References 111BenirschkeK, Kaufmann P. Pathology of the human placenta. 2nd ed. Springer Verlag, New York, 1990: 542-635. PI Scott JR, Peat D, Rhodes A. Investigation of the fetal pulmonary inflammatory reaction in chorioamnionitis using an in situ Y chromosome marker. Pediatr Pathol 1994; 14: 997-1003. [31 Johnson WC, Meyer JR. A study of pneumonia in the stillborn and newborn. Am J Obstet Gynecol 1925; 9: 151-167. 141 Benirschke K, Clifford ST. Intrauterine bacterial infection of newborn infants. J Pediatr 1959; 54: 11-18. PI Moyo SR, Tswana SA, Nystrom L. Bergstrom S, Ljungh A, Blomberg J. Intrauterine death and infections during pregnancy. Int J Gynecol Obstet 1995; 51: 211-218.
123
I61 Naeye RL. Acute bacterial chorioamnionitis. Trans-
placental effects on fetal health. New York: Liss, 1988: 73-86. [71 GustafssonL. QUEST. A program system for statistical and epidemiological data analysis. UmeH, Sweden: University of UmeH, 1990. VI Miettinen 0. Component of the crude risk ratio. Am J Epidemiol 1972;96: 168-172. 191 Miettinen 0. Estimability and estimation in case-referent studies. Am J Epidemiol 1976; 103: 226-235. 1101Mbller GH, Woods DL, Malan AF, Sinclair-Smith CC. Chorioamnionitis in relation to mode of delivery at term. S Afr Med J 1989;76: 201-202. PII Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, EschenbachDA. A case-control study of chorioamniotic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319: 972-978. 1121Zhang J, Kraus FT, Aquino TI. Chorioamnionitis: a comparative histologic, bacteriologic and clinical study. Int J Gynecol Path01 1985;4: I-10. 1131 Salatia CM, Weigl C, S&et-man L. The prevalence and distribution of acute placental inflammation in uncomplicated term pregnancies. Obstet Gynecol 1989;73: 383-389. H41 Quinn PA, Butany J, Taylor J, Hannah W. Chorioamnionitis: its association with pregnancy outcome and microbial infection. Am J Obstet Gynecol 1987; 156: 379-387. 1151 Olding L. Bacterial infection in casesof perinatal death. Acta Paediatr Stand Suppl 1966, 171. it61 Madan E, Meyer MP, Amortequi A. Chorioamnionitis: a study of organisms isolated in perinatal autopsies. Ann Clin Lab Sci 1988; 18: 39-45. iI71 Osman NB, Folgosa E, Gonzales C, Bergstrom S. Genital infections in the etiology of late fetal death: an incident case-referent study. J Trop Pediatr 1995; 41: 267-272. 1181Shen-SchwartzS, Ruchell E, Brown D. Villous oedema of the placenta: a clinicopathological study. Placenta 1989; lo: 297-307. P91 Naeye RL. Disorder of the placenta, fetus and neonate. Diagnosis and clinical significance. St Louis, MO: Mosby, 1992:257-261. 1201Romero R, Nores J, Mazor M, SepulvedaW, Oyarzun E, Parra M, et al. Microbial invasion of the amniotic cavity during term labor. Prevalence and clinical significance. J Reprod Med 1993;38: 543-548.
Journal of Gynecology & Obstetrics 54 (1996) 115-123
Article
Stillbirths and intrauterine infection, histologic chorioamnionitis and microbiological findings S.R. Moyoa, I. HPgerstrand*b, L. Nystr6mc, S.A. Tswanaa, J. Blombergd, S. BergstCme, A. Ljunghd ‘Department of Medical Microbiology, University of Zimbabwe Medical School, Harare, Zimbabwe bDepartment of Pathology and Cytology, University Hospital, Land, Sweden =Departmentof Epidemiology and Public Health, University of LImeA. Ume& Sweden dDepartment of Medical Microbiology, University of Lund, Lund, Sweden cDepartment of International Health, Ullev~l University Hospital, Oslo, Norway Received 2 February
1996; revised 3 April 1996; accepted 18 April 1996
AbStrPCt
Objective: To analyze whether placental inflammation is associatedwith stillbirth in Zimbabwe. Method: Placentas from 66 stillbirths (>22 weeks’ gestation; patients with congenital malformations, diabetes or preeclampsia were excluded) and 66 term live births were studied for the presenceand severity of chorioamnionitis. The morphological results were compared with earlier presentedmicrobiological findings in the samematerial. Results: Chorioamnionitis was present in 79% of stillbirths and 30% of live births (O.R. 8.5, 95% C.I. 4.0-18). Nine percent of stillbirths but no live births presentedvasculitis of the chorionic plate, which verified an inflammatory responsefrom the infant (O.R. 14, 95% C.I. 2.8-72). The sametypes of microorganisms were isolated from stillbirths and liveborns, but Escherichia co/i and group B streptococci were more frequent among stillbirths. Conclusions: Morphological chorioamnionitis occurred 2.6 times more often in women with stillbirths than in women with live births. In 9% of stillbirths the infant showedan inflammatory response.Thus the infant was alive when the infection occurred and it is therefore reasonable to assumethat infection was the cause of death. Keywords: Chorioamnionitis; Vasculitis; Stillbirths; Developing country; Zimbabwe
1. Introduction Chorioamnionitis is characterized by. the presence of neutrophil leukocytes in the placenta. The inflammation starts in the fetal membranes l Corresponding 143307.
author, Tel.: +46 46 173418; Fax: +46 46
0020-7292/96/$.15.00 0 1996 International PII SOO20-7292(96)02705-l
and spreads through the space between the amnion and the chorion. It may reach the amniotic cavity and/or the basal and/or chorionic surface of the placenta and then the fetus, resulting in congenital pneumonia and/or funisitis [l]. The leukocytes originate from the mother [l]. However, in the walls of the chorionic and umbilical vessels and in
Federation of Gynecology and Obstetrics
116
S.R. Moyo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
the interstitial pulmonary tissue, the leukocytic infiltration is of fetal origin [l]. The leukocytes of the alveoli in congenital pneumonia were until recently regarded as maternal, but it has now been proven in male infants that many at least are fetal leukocytes 121. Whilst chorioamnionitis is an ascendinginfection reaching the placenta through the birth canal, villitis is a hematogenousinfection reaching the placenta from the blood vesselsof the mother and able to cause sepsisin the fetus [l]. In 1925,Johnson and Meyer [3] drew attention to the frequent combination of congenital pneumonia and inflammation of the placental membranes.Benirschke and Clifford 141studying 1300consecutive deliveries in the Boston Lying-in Hospital in 1959found inflammation of the membranes (chorioamnionitis) in 13% of cases. In a previous microbiological study of stillbirths [S] we found bacteria that had spread beyond the cervix in 33%of women with stillbirths as compared with 7% of women with live births. The difference was statistically significant (O.R. 6.3, 95% C.I. 2.8-14). This study concerns the morphological
signs of inflammation and their comparison with microbiological findings in the material from the mentioned study [5]. 2. Materials and methods
Placentas from 66 patients admitted to Harare Central Maternity Hospital with a confirmed diagnosis of intrauterine death in a singleton pregnancy of over 22 weeks’ gestation between March 1989and March 1991were included in the study. Patients with high blood pressure (diastolic ~90 mmHg, systolic L 140mmHg) or diabeteswere excluded from the study as well as those with stillborn infants with gross congenital abnormalities. For each case a referent was selectedamong women admitted during the same time period in spontaneouslabor at term prior to rupture of the membranes and before vaginal examination, and with the delivery resulting in a livebom infant. Specimensfor microbiological culture were taken aseptically and examined as described earlier. Specimenswere taken from the cervix of the moth-
Fig. 1. Section through half a placenta with the umbilical cord to the left (a), the free membranes to the right (b), (c) marks the basal plate (decidua basalis), (d) marks the chorionic plate and (e) the subchorionic space.
S. R. Moyo et al. /International
Journal of Gynecology & Obstetrics 54 (1996) 115-123
er, from the placental disk, the umbilical cord, and from the throat and ear of all the infants and in addition from the heart, blood, lung and liver of the stillborns. The results of serology in detecting the occurrence of viral (cytomegalovirus, rubella, herpes simplex virus 1 and 2, human immunodeficiency virus) and bacterial (syphilis) infection, and a Ch/amydia antigen test have been previously discussed [5]. The placentas were immersed in 4% formaldehyde for fixation. One piece was cut from the central part of each placenta including the chorionic as well as the basal plate. The tissue blocks were processed, embedded in paraffin and sectioned. Histologic slides were stained with hematoxylin and eosin. The presence of neutrophil leukocytes was estimated in the chorionic plate, the walls of the chorionic vessels(vasculitis), the subchorionic spaceand the basal plate (Fig. 1). The grading of positivity was as follows: grade 0, no leukocytes; grade I, <25 leukocytes/highpower field; grade II, 25-50 leukocytetiighpower field; and grade III, >50 leukocytes/high-
117
power field. Villous edema was scored according to the percentage of edematous villi [6], and sections with half or more of the villi with edematous stroma were classified as positive. The stroma of the chorionic villi were examined for the presenceof inflammatory cells (villitis). 2.1. statistical analysis
The data were analyzed using QUEST software [7]. The matching was dissolved to avoid loss of valuable information during analysis. This was possible since the ratio between matched and unmatched data was close to one [8]. The relative risks were estimated by odds ratios and 95% confidence intervals according to Miettinen [9].
3. Results
The prevalence of histologic and microbiological findings in casesand referents is summarized in Tables l-3 and Figs. 2-7.
Fig. 2. Chorioamnionitis: inflammation in the chorionic plate and subchorionic space. Bar = 40 my.
I18
S. R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
Table I Numbers (N) and prevalences (%) in cases(C) and referents (R) according to localization and grade of the inflammation Inflammatory grade
Chorionitis
Subchorionitis
Basal deciduit
Total case of
chorioamnionitis C N
0 I 11 III
30 26 9
Total
66
%
R N
%
C ~ N
46 39 I4
47 16 3
71 24 5
34 25 6
0
0
66
100
I2 100
%
R N
%
51 38 9
62 4 0
94 6 0
0
0
I2
66
100
65a
I2 100
66
C N
C N
%
R N
%
46
71
66
100
IS
23
0
0
3
4
0
0
0 0
66
100
100
R %
14 21 34 51 1523 3500 66
100
N
%
46 I7 3
69 26 5
66
100
*Basal plate was missing in one histological slide.
Table 2 Numbers (iV) and prevalences (%) ofsubjects wih chorioamnionitis in cases(C) and referents (R) by microbiology Total
Pathogens Not isolated
Isolated C N
R %
R
C
N
%
N
%
C
R
N
%
N
%
N
%
Chorioamnionitis No chorioamnionitis
52
79
20
30
I2
18
8
I2
46
70
II
27
14
I6 7
24
21
36 7
55
14
11
19
29
Total
66
100
66
100
43
66
39
59
23
35
27
41
Table 3 Prvalencesin cases(C; n = 66) and referents (R; n = 66). odds ratio (OR) and 95% confidenceintervals (CI) for histology and positive and negative microbiology Histology
Chorionitis
Sub-chorionitis
Basal deciduitis Chorioamnionitis, total
Microbiology
Prevalence in
OR
95 % CI
C
R
Positive Negative All
51.2 60.9 54.5
28.2
2.67
1.07-6.65
29.6
3.69
1.15-11.90
28.8
2.97
1.45-6.06
Positive Negative All
60.5 26. I
18.40
5.81-57.90
48.5
7.7 3.7 6.1
14.60
5.57-38.20
Positive Negative All
28.6 30.4
0.0 0.0
1.84-569.00
29.2
0.0
32.4 25.0 55.8
3.29-948.00
Positive Negative All
83.7 69.6 78.8
30.8 29.6 30.3
11.60
4.28-31.30
5.43 8.54
1.65-17.80 4.02-18.20
9.18
1.33-63.30
I .34-437.00
S.R. Mayo et al. /International
Fig. 3. Inflammation
Fig. 4. Chorioamnionitis: inflammation subchorionic space. Bar = 10 my.
Jourd
of Gynecology % Obstetrics 54 (19%) 115-123
119
grade III of the chorionic plate. Bar = 10 my.
grade II in the
Fig. 5. Deciduitis: inflammation grade III in the basal plate (trophoblasts, decidual cells and neutrophil leukocytes). Bar = 10 my.
120
S.R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
Fig. 6. Vasculitis grade III. Bar = 40 req.
3. I. Histology
Among stillbirths 52/66 (79%) had chorioamnionitis compared with 20/66 (30%) referents (O.R. 8.5, 95% CI 4.0-18) (Tables l-3). Thus chorioamnionitis occurred 2.6 times more often in women with stillbirths than in women with live births. Grade-II chorioamnionitis (25-50 leuko-
Fig. 7. Villitis: cytomegalovirus
10 my.
cytesibigh-power field) was found in one or more sites in 15 stillbirths and in three referents; gradeIII chorioamnionitis (> 50 leukocytes/high-power field) in three stillbirths, but in no referents. Inflammation in two sites (chorionic plate and subchorionic space) was seen in 26 casescompared with four in the referents, and inflammation in
inclusion seen in inset. Bar = 40 resp. 10 my.
S.R. Mayo et al. /International
Journal of Gynecology & Obstetrics 54 (19%) 115-123
three sites (also in the basal plate) occurred in four stillbirths but in no referents. In no referent did the basal plate show infiltration of neutrophil leukocytes. Chorioamnionitis thus tended to be more frequent, more severe and more extensive in stillbirths than in referents (Table 1). Inflammation was observed in the chorionic plate (Figs. 2 and 3), the subchorionic space(Fig. 4) and the basal plate (Fig. 5) in 36, 32 and 19 stillbirths compared with 19, 4 and 0 referents, respectively (Table 1). The corresponding odds ratios were 3.0, 20 and 27, all of which were statistically significant (Table 3). In no case was the inflammation restricted only to the basal plate. Vasculitis in the chorionic plate (Fig. 6) was seenin 6/66 stillbirths, but in none of the referents (O.R. 14, 95% C.1, 2.8-72). No microbial agent was found in two of the six cases,Escherichiu coli was found in one and group B streptococcusin another. In the remaining two casesthe mothers were seropositive for syphilis, which however should not have caused the purulent vasculitis. Villous edemawas found in 8166placentas from stillbirths (12O/o),but in no placenta from a live birth, corresponding to an odds ratio of 19 (95% C.I. 4.0-94). Six of the eight placentas showed chorioamnionitis and one showed villitis compatible with cytomegalovirus infection. Bacterial organisms were not isolated from the case with villitis nor from three additional subjects. E. co&, group B streptococci, Candida albicuns and Staphylococcusaureus were isolated from the four remaining placentas. Villitis (Fig. 7) was found in 5/66 stillbirths and l/66 referents (O.R. 5.3, 95% C.I. 0.57-256). Cytomegalovirus was detected by immunohistochemistry in one of the placentas with villitis among the stillbirths (the samecase as above). In two casesthe mother had syphilis, but no explanation could be found for the remaining three cases. Toxoplasma cysts were not found (no immunohistochemical method was performed). The cytomegalovirus case was cytomegalovirus-IgM negative, compatible with a reactivated infection in the mother. 3.2. Microbiology In 43/66 stillbirths bacteria were isolated from
121
the placenta and/or from the neonate compared with 39/66referents (O.R. 1.3,95% C.I. 0.60-2.8). Bacteria were recovered from the placentas in 18 stillbirths and three live births. Bacterial growth from the internal organs occurred in 27/66 stillbirths (41%). The main bacterial speciesisolated were E. coli, group B streptococci, coagulasenegative staphylococci, S. aureus and Streptocoo cus viridans. E. coli was isolated in 17166casesand in 9166 referents (O.R. 2.2, 95 % C.I. 0.83-5.9). Group B streptococci were isolated in only eight casesand three referents (O.R. 2.9, 95% C.I. 0.65-18). Coagulase-negativestaphylococci were isolated in nine casesand 10 referents (O.R. 0.88, 95% C.I. 0.30-2.6). S. uureus was isolated in six stillbirths and one referent (O.R. 6.5, 95% C.I. 0.75-65). In 18166stillbirths (27%), the same bacterial species was found in three or more sites. The referents cannot be compared, since no cultures from internal organs were available. 3.3. Comparison between histology and microbiology In 36 casesand 12 referents, the histologic and microbiological results from the placenta were in agreementas regards presenceof chorioamnionitis (i.e. bacterial growth in at least one location) (Table 2). In seven casesand 19 referents, histologic and microbiological findings were in agreement regarding absenceof chorioamnionitis. In 16 casesthere was histologic chorioamnionitis, but no microbial agent was detected. The same was true for eight referents. In sevencasesand 27 referents there was no histologic chorioamnionitis, but microbiological agents were found, however they were mostly of low virulence such as diphtheroids and cr-hemolytic streptococci [5]. 4. Discussion This study demonstrated inflammation of the chorionic plate, the subchorionic spaceand/or the basal plate in 30% of normal deliveries. It is probably too low a figure considering that only one histologic section was examined in each case. The prevalence is difficult to compare with figures presented in other materials since neither the
122
S.R. Moyo et al. /international
Journal of Gynecology & Obstetrics 54 (19%) 115-123
method of collection nor the diagnostic criteria agree. In South Africa, Miiller et al. [lo] found a prevalence of 28%; in the USA, Hillier et al. [ 1l] found a prevalence of 16% in Seattle, and Zhang et al. [12] found a prevalence of 10% in St Louis, whilst a prevalenceof 4% among normal deliveries has beenreported from Connecticut [ 131.A prevalence of 5% has been reported in Toronto, Canada [ 141.Thus the prevalencesof histologic chorioamnionitis tend to be higher in developing than in developed countries. In this study histologic chorioamnionitis was more severeand more extensive in stillbirths than in referents. The prevalence found (79%) was higher than in similar studies performed elsewhere [14-171. However, we have counted any presence of neutrophil leukocytes in the described locations as an expression of chorioamnionitis. Olding [ 151 in 1966reported 55% in a material of stillbirths in Sweden. In Canada and in the USA, prevalences of 58% [14] and 38% [16], respectively, have been reported, and in a recent study from Mozambique, 68% [17]. Using the samecriteria as in the present study but with more histologic sections in approximately 100 stillbirths from Lund in Sweden, Tolockienk and Hagerstrand (unpublished observations) found 97% with histologic chorioamnionitis. It is thus obvious that the prevalence of histologic chorioamnionitis is higher in stillbirths than in normal deliveries. In the present study, six (9%) stillbirths showed vasculitis in the chorionic plate. This indicates that the infant has reacted against the infection with diapedesisof focally increased leukocytes [l] and was alive when infected. In the rest of the material it cannot be ruled out that the chorioamnionitis developed after intrauterine fetal death. In the stillbirths examined in Lund, Sweden,6% had fetal vasculitis. It is reasonableto assumethat the cause of death was also chorioamnionitis in many subjects without vasculitis. Since we only diagnosed chorioamnionitis cases reaching beyond the free membranes, the disease process must have been present for some time. Regarding the pathogenesis of chorioamnionitis, the inflammation of the membranestends to spread to the chorionic plate and the subchorionic spacerather than to the basal plate. This
might indicate that bacterial invasion from the membranes into the amniotic fluid is more common than from the basal plate via the maternal blood to the infant [1,6,13]. Villous edema was present only in stillbirths (12%). The majority of these infants, however, were preterm, and this figure thus agreeswith the report by Shen-Schwartzet al. [ 181.These authors assumevillous edema to be a lesion of the immature intermediate villi, and found no correlation with chorioamnionitis. The villous edema might have been the final causeof death in those preterm stillbirths. Mild chorioamnionitis might also cause defecation of meconium and death of the infant due to vasoconstriction [19]. Unfortunately, we were unable to obtain any information on meconium defecation in our material. Histologic chorioamnionitis (72 cases and referents) was indicated by the microbiological findings (48 casesand referents) in approximately 70%,which is in accordancewith other studies [6]. Histologic chorioamnionitis without bacterial growth might indicate a failure to retrieve the microbes or an inability of the prevailing microbes to grow out due to the inflammatory response in the placenta. Microbiological chorioamnionitis without inflammation might mean that the microbial invasion was recent and that the sampling was done before the onset of the inflammatory process. Provided that no contamination occurred with microbiological cultures, any bacteria isolated from the placenta would presumably represent microbiological chorioamnionitis. Cultures from ear, throat and umbilical cord might have been contaminated by the vaginal flora. Considering only placental cultures, 18 (27%) stillbirths and three (4.5%) live births had microbiological chorioamnionitis. If we add to the stillborn group those with any bacteria isolated in cultures from internal organs, signs of feto-placental infection occurred in 28/66 cases(42%) (51. Cultures from the internal organs of the stillborn infants and from the placenta are vital in proving chorioamnionitis. However, without cultures from ear and throat, more casesof histologic chorioamnionitis would have been left without a microbiological explanation.
S.R. Mayo et al. / Inremational Journal of Gynecology d Obstetrics 54 (19%) 115-123
Bacterial invasion of the amniotic cavity at term has been assumedto be a physiological phenomenon being either the consequenceor the cause of labor [20]. This assumption was basedon the tindings of microbial growth in 19% of amniotic fluid cultures during spontaneous parturition at term. The dominating organism was Ureaplasma urealyticum, an organism we did not try to isolate in our study. It might be argued that such a physiological process is offered sufficient time to develop into a manifest infection after the death of the infant. However, chorionic vasculitis in some of the infants provides evidence that they were alive at the onset of infection. Acknowledgments We would like to thank Mr J. Magwenzi, Dr Shylaja and Sr J. Moyo for their assistance.This study was supported by grant number 2.9999.10.3334from the University of Zimbabwe Research Board, grant number 26089.10 from the Black Medical Foundation Screening Committee, and by the Swedish Agency for Research Cooperation with Developing Countries (SAREC). References 111BenirschkeK, Kaufmann P. Pathology of the human placenta. 2nd ed. Springer Verlag, New York, 1990: 542-635. PI Scott JR, Peat D, Rhodes A. Investigation of the fetal pulmonary inflammatory reaction in chorioamnionitis using an in situ Y chromosome marker. Pediatr Pathol 1994; 14: 997-1003. [31 Johnson WC, Meyer JR. A study of pneumonia in the stillborn and newborn. Am J Obstet Gynecol 1925; 9: 151-167. 141 Benirschke K, Clifford ST. Intrauterine bacterial infection of newborn infants. J Pediatr 1959; 54: 11-18. PI Moyo SR, Tswana SA, Nystrom L. Bergstrom S, Ljungh A, Blomberg J. Intrauterine death and infections during pregnancy. Int J Gynecol Obstet 1995; 51: 211-218.
123
I61 Naeye RL. Acute bacterial chorioamnionitis. Trans-
placental effects on fetal health. New York: Liss, 1988: 73-86. [71 GustafssonL. QUEST. A program system for statistical and epidemiological data analysis. UmeH, Sweden: University of UmeH, 1990. VI Miettinen 0. Component of the crude risk ratio. Am J Epidemiol 1972;96: 168-172. 191 Miettinen 0. Estimability and estimation in case-referent studies. Am J Epidemiol 1976; 103: 226-235. 1101Mbller GH, Woods DL, Malan AF, Sinclair-Smith CC. Chorioamnionitis in relation to mode of delivery at term. S Afr Med J 1989;76: 201-202. PII Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, EschenbachDA. A case-control study of chorioamniotic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319: 972-978. 1121Zhang J, Kraus FT, Aquino TI. Chorioamnionitis: a comparative histologic, bacteriologic and clinical study. Int J Gynecol Path01 1985;4: I-10. 1131 Salatia CM, Weigl C, S&et-man L. The prevalence and distribution of acute placental inflammation in uncomplicated term pregnancies. Obstet Gynecol 1989;73: 383-389. H41 Quinn PA, Butany J, Taylor J, Hannah W. Chorioamnionitis: its association with pregnancy outcome and microbial infection. Am J Obstet Gynecol 1987; 156: 379-387. 1151 Olding L. Bacterial infection in casesof perinatal death. Acta Paediatr Stand Suppl 1966, 171. it61 Madan E, Meyer MP, Amortequi A. Chorioamnionitis: a study of organisms isolated in perinatal autopsies. Ann Clin Lab Sci 1988; 18: 39-45. iI71 Osman NB, Folgosa E, Gonzales C, Bergstrom S. Genital infections in the etiology of late fetal death: an incident case-referent study. J Trop Pediatr 1995; 41: 267-272. 1181Shen-SchwartzS, Ruchell E, Brown D. Villous oedema of the placenta: a clinicopathological study. Placenta 1989; lo: 297-307. P91 Naeye RL. Disorder of the placenta, fetus and neonate. Diagnosis and clinical significance. St Louis, MO: Mosby, 1992:257-261. 1201Romero R, Nores J, Mazor M, SepulvedaW, Oyarzun E, Parra M, et al. Microbial invasion of the amniotic cavity during term labor. Prevalence and clinical significance. J Reprod Med 1993;38: 543-548.