121
INTRODUCTION
Prematurity is the leading cause of perinatal morbidity and mortality.1 Despite the availability of tocolytic drugs, aggressive surveillance techniques, and other treatment and prevention strategies, no change has been observed in premature delivery rate since the 1 9 5 0 s.2 It is well known that twins have significantly higher perinatal mo rtality than singletons, with a rate 4 to 1 0 times greater in twins than in sing leto ns being repo rted.3 -4 The hig h perinatal mortality is associated with the increased rate of preterm delivery. Various approaches have been proposed in order to reduce the prematurity rate in twin preg nancies, such as pro phylactic cervical cerclage, monitoring of uterine activity, long term tocolysis and bed rest. However, most published studies5-9 have not been able to demonstrate effective metho ds that sig nificantly decrease the rate o f preterm deliveries in twin pregnancies. O ver the last seven years, the presence o f fetal fibro nectin in the cervix or vagina has been investigated as a possible marker for the risk of preterm birth. Fetal fibronectin in cervical fluid can pro vide direct evidence o f patho lo gic changes at the interface o f fetal and maternal tissues. Thus, fetal fibronectin as a screening test could assist the clinician in the identification of twin pregnancies at high risk for preterm birth.
Original Article
De te ction of fe tal fibrone ctin in twin
pre gnancie s in re lation to ge stational age
Hospital Maternidade “Leonor Mendes de Barros” and
Universidade Federal de São Paulo / Escola Paulista de Medicina, São Paulo, Brazil
Tenilson Amaral Oliveira Carla Muniz Pinto de Carvalho Eduardo de Souza Corintio Mariani-Neto Luiz Camano
ABSTRACT
Contex t: The presence of fetal fibronectin in the cervix or vagina has been investigated as a possible marker for the risk of preterm birth. Fetal fibronectin in cervical fluid can provide direct evidence of pathologic changes at the interface of fetal and maternal tissues.
O bjective: To evaluate the presence o f fetal fibro nectin as a predicto r o f premature delivery in twin preg nancies in relatio n to g estatio nal ag e.
Design: Acuracy study.
Setting: University referral unit.
Pa rticipa nts: 5 2 preg nant wo men with twin preg nancies and g estatio nal ag e o f between 2 4 and 3 4 weeks.
M a in m ea surem ents: Sensivity, specifity, predictive values and relative risk ratio s o f the co rrelatio n between fetal fibro nectin and preterm birth befo re 3 4 and 3 7 weeks using an immediate-reading membrane test o n
cervico vag inal secretio ns o btained fro m participants.
Result: The sensitivity varied between 6 6 .7 % and 8 5 .7 %, whereas the specificity was fro m 5 8 .3 % to 8 1 .8 %
acco rding to g estatio nal ag e at the time o f sampling . The relative risk o f spo ntaneo us preterm birth after a po sitive fetal fibro nectin test, as co mpared with a neg ative fetal fibro nectin test, ro se fro m 2 .8 at 2 4 2 6 weeks to 4 .1 at 2 7 -3 0 weeks. Analyses o f the risk o f delivery befo re -3 4 weeks were no t statistically sig nificant.
Conclusion: Fetal fibro nectin in the cervico vag inal secretio ns o f patients with twin preg nancies is a useful to o l fo r the early identificatio n o f twin preg nancies likely to deliver befo re 3 7 weeks. Ho wever, the clinical value o f the fibro nectin test is limited because o f lo w indices fo r predictio n o f delivery befo re 3 4 weeks. The best perio d fo r perfo rming the fetal fibro nectin test in twin preg nancies to predict preterm delivery is between 2 7 and 3 0 weeks.
Key w ords: Twins. Fetal fibro nectin. Preterm delivery.
122
The purpose of this study was to evaluate the efficacy of fetal fibronectin test in predicting preterm delivery in twin pregnancies in relation to gestational age at the time of sampling.
METHODS
The study was done at Hospital Maternidade “Leonor Mendes de Barros” and Hospital São Paulo (Universidade Federal de São Paulo) between June 1 9 9 4 and August 1 9 9 6 . After informed consent was obtained from all patients, gestational age was established using the date of the last menstrual period and was confirmed by early ultrasonography.
Fifty-nine women were prospectively enrolled in the study and none were lost during it. Five of the m w e re e xc lud e d b e c a use o f me d ic a l co mplicatio ns that necessitated preterm delivery: severe pre-eclampsia (3 cases), eclampsia (1 case) and fetal distress (1 case). Two patients who delivered at term after using tocolytic agents to inhibit preterm labor were also excluded. Conversely, nine patients who underwent to c o lysis unsuc c essfully were included in the study. Thus, 5 2 patients remained for analysis.
Samples of cervicovaginal secretions were co llected every two weeks between 2 4 and 3 4 weeks of gestation. The presence of fetal fibronectin was determined qualitatively using a bedside membrane immunoassay. No attempt was made to control patient treatment and management during prenatal care based on fetal fibronectin test results. The fetal fibronectin results were not communicated to the obstetric team and therefore did not influence subsequent patient management. Preterm birth was defined as delivery befo re 3 7 weeks o f gestatio n.
Because the test results are qualitative and based o n the o bservatio n o f co lo red spo t, the re se a rc he r a nd tw o ind e p e nd e nt o b se rve rs co mpared the co lo r reactio n o f each test with a standard chart supplied with each kit, showing an example o f the reactio n intensity to define the test as positive or negative. After recording the results found by each observer, the samples were excluded from the stud y in c a se s o f d isc o rd a nt o p inio ns. Cervicovaginal samples with visible blood at swab were also excluded.
Statistical Metho ds. Correlation between the time of delivery and fetal fibronectin test result was performed by means of sensitivity, specificity and predictive values. The data was analyzed using relative risk ratios.
RESULTS
Twenty-eight (5 3 .8 %) delivered before 3 7 weeks. A total of 2 7 0 samples were collected from these patients during the study. The average number of samples obtained per patient was 5 .2 . Twenty three samples were excluded, because of discordant results between the observers (13 cases) or presence of blood at swab (10 cases). Therefore, we analyzed 2 4 7 samples for the prediction of preterm delivery. Table 1 presents the sensitivity, specificity, p o sitive a nd ne g a tive p re d ic tive va lue fo r sp o nta ne o us p re te rm d e live ry in re la tio n to gestational age at the time of sampling. The periods of gestational age considered were 2 4 to 2 6 , 2 7 to 3 0 and 3 1 to 3 4 weeks of gestation. Twenty-one women out of a total of 4 3 delivered preterm when the test was perfo rmed at 2 4 to 2 6 weeks o f gestation. Twenty-eight out of a total of 5 2 delivered preterm when the test was performed at 2 7 to 3 0 weeks of gestation. Twenty-six preterm births occurred out of a total of 5 0 women who underwent the test at 3 1 to 3 4 weeks. The sensitivity varied between 6 6 .7 % and 8 5 .7 %, the specificity rose from 5 8 .3 % to 8 1 .8 %, whereas positive and negative predictive values were equivalent in those three periods. Table 1 also shows that the relative risk (RR) for preterm birth was greater at 2 7 to 3 0 weeks than at 2 4 to 2 6 weeks and 3 1 to 3 4 weeks.
Because preterm birth before than 3 4 weeks is associated with the worst perinatal outcomes, we determined the value of the fetal fibronectin test in the prediction of spontaneous delivery before 3 4 weeks in those three different time periods. Nine patients out of a total of 4 3 delivered before 3 4 weeks when the test was performed at 2 4 to 2 6 weeks of gestation. Ten women out of a total of 5 2 delivered befo re 3 4 weeks when the test was perfo rmed at 2 7 to 3 0 weeks o f gestatio n. Six preterm births occurred out of a total of 5 0 women who underwent the test at 3 1 to 3 4 weeks. In Table
123
2 , it can be seen that the sensitivity and negative predictive values of the fetal fibronectin test at 2 7 to 3 0 weeks were better at predicting delivery before 3 4 weeks when co mpared with the o ther two intervals. Ho wever, the relative risk was no t statistically significant.
DISCUSSION
The majo r cause o f increased perinatal mortality in twin pregnancies is preterm delivery.10 The availability of a rapid search for the presence of cervical fetal fibronectin could improve our ability to efficiently identify patients at risk for preterm delivery. The rapid result membrane test was comparable to the standard fetal fibro nectin in cervical vaginal secretions, in accordance with Bittar et al.11
Goldenberg et al12 studied how various patterns of fetal fibronectin testing were related to spontaneous preterm birth in singleton pregnancies, and reported that the last test result seems to be the best predictor of subsequent spontaneous preterm delivery. In twin
pregnancies, however, it seems that the 27 to 30-week sampling period is the most appropriate moment to evaluate the risk o f preterm delivery in twin gestations because the number of false positives in this interval is less than in the other two.
N ag eo tte et al1 3 evaluated the predictive efficacy of weekly cervicovaginal fetal fibronectin determinations between 20 and 37 weeks in 102 asymptomatic singleton pregnancies. They reported that fetal fibronectin had high sensitivity (90.6%) but lo w specificity (4 4 .3 %) fo r predictio n o f preterm delivery. Acco rding to these autho rs, their large number of false-positive results may reflect degrees of cho rio nic-decidual cell activatio n insufficient fo r achieving preterm delivery.13 Tolino et al14 found high sensitivity (90.9%) and low specificity (68.5%) in 68 patients with multiple gestations who underwent weekly sampling of cervical and vaginal secretions.
W e believe that the fetal fibronectin test has great potential for improving the clinician’s ability to select patients fo r preventive appro aches and tocolytic drug therapy and for reducing the number
Ta ble 2 - Predictive va lues for sponta neous preterm delivery before 3 4 w eek s in a ccorda nce w ith the gesta tiona l a ge a t the time of the feta l fibronectin test
G estatio nal Sensitivity Specificity Po sitive predictive N eg ative predictive Relative Risk ag e (9 5 % CI) (9 5 % CI) value (9 5 % CI) value (9 5 % CI) (9 5 % CI) 2 4 to 2 6 5 5 .6 % 6 1 .8 % 2 7 .8 % 8 4 .0 % 1 .7
(2 2 .7 to 8 4 .7 ) (4 3 .6 to 7 7 .3 ) (1 0 .7 to 5 3 .6 ) (6 3 .1 to 9 4 .7 ) ( 0 .5 to 5 .6 ) 2 7 to 3 0 8 0 .0 % 4 5 .2 % 2 5 .8 % 9 0 .5 % 2 .7
(4 4 .2 to 9 6 .5 ) (3 0 .2 to 6 1 .2 ) (1 2 .5 to 4 4 .9 ) (6 8 .2 to 9 8 .3 ) ( 0 .6 to 1 1 .5 ) 3 1 to 3 4 6 6 .7 % 3 6 .6 % 1 8 .8 % 8 3 .3 % 1 .1
(3 0 .9 to 9 1 .0 ) (2 2 .6 to 5 3 .1 ) (7 .9 to 3 7 .0 ) (5 7 .7 to 9 5 .6 ) ( 0 .3 to 4 .0 ) CI = Co nfidence interval
Ta ble 1 - Predictive va lues for sponta neous preterm delivery before 3 7 w eek s in a ccorda nce w ith the gesta tiona l a ge a t the time of the feta l fibronectin test
G estatio nal Sensitivity Specificity Po sitive predictive N eg ative predictive Relative Risk ag e (9 5 % CI) (9 5 % CI) value (9 5 % CI) value (9 5 % CI) (9 5 % CI) 2 4 to 2 6 6 6 .7 % 8 1 .8 % 7 7 .8 % 7 2 .0 % 2 .8
(4 3 .1 to 8 4 .5 ) (5 9 .0 to 9 4 .0 ) (5 1 .9 to 9 2 .6 ) (5 0 .4 to 8 7 .1 ) (1 .4 to 5 .5 ) 2 7 to 3 0 8 5 .7 % 7 0 .8 % 7 7 .4 % 8 0 .9 % 4 .1
(6 6 .4 to 9 5 .3 ) (4 8 .8 to 8 6 .6 ) (5 8 .5 to 8 9 .7 ) (5 7 .4 to 9 3 .7 ) (1 .7 to 1 0 ) 3 1 to 3 4 8 4 .6 % 5 8 .3 % 6 8 .8 % 7 7 .8 % 3 .1
(6 4 .3 to 9 5 .0 ) (3 6 .9 to 7 7 .2 ) (4 9 .9 to 8 3 .3 ) (5 1 .9 to 9 2 .6 ) (1 .3 to 7 .6 ) CI = Co nfidence interval
124
of women who unnecessarily receive treatment for a perceived threat of preterm labor. However, despite the high sensitivity and negative predictive value towards delivery before 3 4 weeks, the period when the majority of perinatal mortality is likely to occur, the analysis of the data showed that the relative risk of a positive fetal fibronectin test is not statistically significant when compared with the negative test. Therefore, we found no evidence to support the idea that the fetal fibronectin test could reduce the perinatal mortality in twin pregnancies.
In conclusion, the fetal fibronectin test seems to be a useful to o l in clinical practice fo r early identification of twin pregnancies likely to develop preterm labor. However, the clinical value of the fibronectin test is limited because of low indices for predictio n o f delivery befo re 3 4 weeks. The best period for performing the fetal fibronectin test in twin pregnancies for prediction of preterm delivery is between 2 7 and 3 0 weeks.
REFERENCES
1. Creasy RK, Merkatz IR. Prevention of preterm birth: Clinical opinion. Obstet Gyneco l 1990;76:2s-4s.
2. Garite TJ, Lo ckwo o d CJ. A new test fo r diagno sis and predictio n o f preterm delivery. Co ntempo rary Ob/Gyn 1996;41:77-93.
3. Ho SK, Wu LYK. Perinatal facto rs and neo natal mo rbidity in twin preg-nancy. Am J Obstet Gyneco l 1975;122:979-87.
4. Spellacy WN, Ho ndier A, Fene CD. A case-co ntro l study o f 1253 twin preg-nancies from a 1982-1987 perinatal data base. Obstet Gynecol 1990;75:168-71.
5. Do r J, Shaley J, Masiach J, Blankstein J, Serr DM. Elective cervical suture o f twin pregnancies diagno sed ultraso nically in the first trimester fo llo wing o vulatio n. Gyneco l Obstet Invest 1982;13:55-60.
6. Cro wther CA, Neilso n JP, Ashurst HM, Verkuryl DA, Bannerman C. The effects o f ho spitalizatio n fo r rest o n fetal gro wth, neo natal mo rbidity and length o f gestatio n in twin pregnancy. Br J Obstet Gyneco l 1990;97:872-7.
7. Preventive Task Fo rce. Ho me uterine activity mo nito ring fo r preterm la-bo r: review article. JAMA 1993;270:371-6.
8. Ashwo rth MF, Spo o ner SF, Verkuryl DA, Waterman R, Ashurst HM. Failure to prevent preterm labo ur and delivery in twin pregnancy using pro -phylactic o ral salbutamo l. Br J Obstet Gyneco l 1990;97:878-82.
9. Maclennan AH, Green RC, O’Shea R, Bro o kes C, Mo rris D. Ro utine ho s-pital admissio n in twin pregnancies between 26 and 30 weeks gestatio n. Lancet 1990;335:267-9.
10. Gro the W, Ruttgers H. Twin pregnancies: an 11-year review. Acta Genet Med Gemello l 1985;34:49-58.
11. Bittar RE, Yamasaki AA, Sasaki S, Zugaib M. Cervical fetal fibro nectin in patients at increased risk fo r preterm delivery. Am J Obstet Gyneco l 1996;175(1):178-81.
12. Go ldenberg RL, Mercer BM, Iams JD, Mo awad AH, Meis PJ, Das A, McNellis D, Mio do vnik M, Menard MK, Caritis S, Thurnau GR, Bo tto ms SF. Na-tio nal Institute o f Child Health, Human Develo pment Maternal-Fetal Medicine Units Netwo rk. The preterm predictio n study: Patterns o f cervico vaginal fetal fibro nectin as predicto rs o f spo ntaneo us preterm delivery. Am J Obstet Gyneco l 1997;177:8-12.
13. Nageo tte MP, Casal D, Senyei AE. Fetal fibro nectin in patients at increased risk fo r premature birth. Am J Obstet Gyneco l 1994;170(1pt1):20-5.
14. To lino A, Ro nsini S, Zullo F, Pellicano M, Regine V, Nappi C. Fetal fibro nectin as a screening test fo r premature delivery in multiple preg-nancies. Int J Gyneco l Obstet 1996;52(1):3-7.
RESUMO
Objetivo: Avaliar a presença da fibro nectina fetal co mo fato r de predição do parto prematuro na prenhez gemelar de aco rdo co m a idade gestacio nal. Tipo de Estudo: Estudo de acurácia. Loca l: Centro de Referência Universitário . Pa rticipa ntes: 5 2 gestantes co m prenhezes gemelares fo ram submetidas ao teste de imuno ensaio de membrana entre 2 4 e 3 4 semanas para verificar a co rrelação entre a presença de fibro nectina fetal na secreção cérvico -vaginal e o risco de parto prematuro antes de 3 4 e 3 7 semanas de gestação . Va riá veis estuda da s: Sensibilidade, especificidade, valo res preditivo s e risco relativo da co rrelação entre a presença de fibro nectina fetal na secreção cévico -vaginal e o risco de parto prematuro antes de 3 4 e 3 7 semanas de gestação .
Resulta do: A sensibilidade vario u entre 6 6 ,7 % e 8 5 ,7 %, enquanto a especificidade fo i de 5 8 ,3 % a 8 1 ,8 % de aco rdo co m a idade gestacio nal em que o exame fo i realizado . O risco relativo para o parto prematuro espo ntâneo apó s um teste po sitivo co mparado ao teste negativo vario u de 2 ,8 entre 2 4 -2 6 semanas a 4 ,1 entre 2 7 -3 0 semanas. O s dado s analisado s para predição do parto antes de 3 4 semanas não mo straram resultado s estatisticamente significantes. Conclusã o: O teste da fibro nectina fetal é útil para identificação preco ce das gemeligestas co m risco elevado para o parto prematuro antes de 3 7 semanas. Co ntudo , o valo r clínico do teste é limitado devido ao s baixo s valo res de predição para o co rrência do parto antes de 3 4 semanas. O melho r perío do para a realização do teste para predição do parto prematuro é entre 2 7 e 3 0 semanas de gestação .
Tenilson Amaral Oliveira - MSc in O bstetrics from Univ. Federal de São Paulo and Supervisor of the high-risk pre-natal unit at Hospital Maternidade “Leonor Mendes de Barros”
Carla M uniz Pinto de Carvalho - MD. Post-graduate student at Universidade Federal de São Paulo
Eduardo de Souza - MD, PhD. Assistant professor in the discipline of obstetrics at Universidade Federal de São Paulo
Corintio M ariani-N eto - MD. Scientific Director at Hospital Maternidade “Leonor Mendes de Barros”
Luiz Camano - MD, PhD. Professor and Chair in the discipline of O bstetrics at Universidade Federal de São Paulo
Sources of Funding: Not declared
Conflict of interest: Not declared
Last received: 3 1 july 1 9 9 8
Accepted: 1 0 november 1 9 9 8
Address for correspondence:
Tenilson Amaral O liveira
Rua Azevedo Júnior 1 4 3 , apto. 7 3 CEP 0 3 0 4 0 -9 0 0 - São Paulo/ SP - Brazil
