- Email: [email protected]
Mammary stimulation test predicts preterm birth in nulliparous women
Mammary stimulation test predicts preterm birth in nulliparous women Debra A. Guinn, M D , " ~"Thomas R. Wigton, MD," bJane A. James, MD," b Dorothy D. Dunlop, PhD, ~ Michael Socol, MD," ~ and Marilynn C. Frederiksen, MD" b Chicago, Illinois OBJECTIVE: This prospective clinical trial was designed to assess the ability of the mammary stimulation test to predict preterm birth In a private nulliparous population. STUDY DESIGN: The mammary stimulation test was performed between 26 and 28 weeks' gestation by 267 nulliparous women with singleton pregnancies, Test results were blinded to both patient and referring physician. Pregnancy outcome data were collected from the perinatal database and medical records, RESULTS: The mammary stimulation test was positive in 45 of 266 (17%) patients. Derivery occurred at <37 weeks in 27 patients (10,2%) and at <34 weeks in ~ive (1.9%). The mammary stimulation test demonstrated a sensitivity of 37%, a specificity of 84%, a positive predictive value of 20%, and a negative predictive value of 92% for delivery at <37 weeks' gestation. For delivery at < 34 weeks' gestation the mammary stimulation test had a sensitivity of 60%, a specificity of 82%, a positive predictive value of 6%, and a negative predictive value 99%. The odds ratio for delivery at <37 weeks was 3.0 (95% confidence Interval 1.3, 7.1), and for delivery at <34 weeks the odds ratio was 7.0 (95% confidence interval 1.1, 43.0). One third of preterm deliveries were secondary to idiopathic preterm labor, and the mammary stimulation test was positive in 77.8% (seven of nine) of these pregnancies. Patients with a positive test were more likely to require observation in labor and delivery for preterm uterine contractions (34% vs 4.3%, p < 0.01), and they were more likely to demonstrate change at cervical examination (14% vs 2%, p < 0.01). CONCLUSION: In this population traditionally considered to be at low risk for preterm birth a positive mammary stimulation test was useful in identifying patients at risk for preterm uterine activity and preterm birth. Equally important was the identification of women who were at low risk for preterm birth. (AM J OBSTETGYNECOL1994; 170:1 809-14,)
Key words: Mammary stimulation test, preterm birth, nulliparity Although preterm birth complicates only 7% to 10% of all pregnancies, prematurity is responsible for 70% to 80% of perinatal mortality. I~ Unfortunately, only 50% of patients who are delivered preterm have historical and demographic factors associated with preterm birth. Nulliparous patients account for 25% to 40% of all preterm births, and they are the most difficult pregnancies to prospectively identify as being at increased risk. 4-~'~ The mammary stimulation test, a test of uterine contractility in response to endogenous oxytocin, was From the Section of Maternal-Fetal Medicine, Departmentof Obstetrics and Gy~ecology," and the General Clinical Research Centerf Northwestern University Medical School, Northwestern Memorial Hospital, and the Centerfor Health Se~wicesand Policy Research, Northwestern University." Supported in part by National Center for Research Resources, National Institutes of Health, grant No. RR-O0048. Presented at the Sixty-first Annual Meeting of The CentralAssociation of Obstetricians and Gynecologists, White Sulphur springs, West Virginia, October 28-30, 1993. Repnnt requests: Debra Guinn, MD, 618 S. 20th St,, Birmingham, AL 35233. Copyright © 1994 by Mosby-Year Book, Inc. 0002-9378/94 $3.00 + 0 6/6/55032
developed by Eden et al. ~s to identify patients at increased risk for preterm delivery. In a medically indigent population the test demonstrated a sensitivity of 84% and a positive predictive value of 34%. To further test the efficacy of the mammary stimulation test, we conducted a prospective clinical trial in a private nul. liparous patient population, a population traditionally considered to be at low risk for preterm birth and largely without identifiable risk factors. Material and methods
Nulliparous patients receiving prenatal care from private practitioners affiliated with Northwestern Memorial Hospital were eligible for the study, which was conducted from January 1992 through February 1993. Patients with a placenta previa, multiple gestation, or ruptured membranes were excluded from participation. All patients were recruited by their attending health care provider. The research protocol was approved by the Northwestern University Institutional Review Board, and written informed consent was obtained from all patients before participation. The mammary stimulation test was performed be1809
1810
.June 199,l Am .] Obstct Gynecol
Guinn et al.
T a b l e I. Maternal demographic data
(N = 267) Age (yr) Gravidity Elective abortions 1
->2 (range 2-4) Spontaneous abortions 1 ---2 (range 2-3) Race White Black Hispanic Other Tobacco use Illicit drug use Education (yr) Occupation* Homemaker Trade Professional
31.7 +_ 4.2 1.6 +- .96 79 (30%) 59 20 48 (18%) 35 13 231 (86.5%) 14 (5.2%) 7 (2.6%) 15 (5.6%) 9 (3A%) 3 (1.1%) 16.5 "-: 1.8 22 (8.2%) 112 (41.9%) 133 (49.8%)
*Occupalion description based on Maim' Professional Codes, U.S. Census Report. 1¢'
tween 26 and 28 weeks' gestation in the Clinical Research Center of Northwestern Memorial Hospital by registered nurses trained in conducting the mammary stimulation test and in basic [~tal heart rate (FHR) interpretation. Gestational age assessment was based on last menstrual period, first-trimester physical examination, or ultrasonographic examination before 24 weeks of pregnancy. Patients were placed in the left lateral decubitus position and a Doppler ultrasonographic transducer and a uterine tocodynometer were placed across the maternal abdomen. A 10- to 20-minute baseline strip of FHR and uterine activity was obtained. In the absence of spontaneous uterine contractions the patient was asked to lightly rub one nipple through her clothing for 2 minutes, followed by a 2-minute rest period. This cycle was repeated three times, followed by unilateral nipple stimulation for 10 minutes and finally bilateral nipple stimulation for 10 minutes. All tests were interpreted by one investigator (D.A,G.) on the day that they were performed. The test was scored as positive if two or more contractions of ~ 4 0 seconds' duration occurred during any 15 minutes of monitoring, including the baseline strip. If this frequency of uterine contractions occurred at any time during the test, nipple stimulation was discontinued and the patient was monitored until uterine activity spontaneously subsided (i,e,, fewer than two contractions in 15 minutes). Both the patient and the attending health care provider were blinded to the results of the mammary stimulation test, except under the following circumstances: spontaneous contractions (two or more per 15 minutes) occurring before stimulation, contractions occurring for > 1 hour after stimulation, or the
presence of an abnol'mal FHR pattern. Pregnancy outcome data were retrieved through the perinatal database and medical records. Sample size calculations were based on the observed results froln Eden et al. ~'~ However, for this low-risk population it was assumed that the sensitivity would be lowered to 60%, the rate of positive tests would be lowered to only 25%, and the incidence of preterrn births would be only 7%. Theretbrc a sample size of 252 would have a 90% power of detecting a relationship between the mammary stimulation test and delivery outcome witb an ~ of 0,05. The odds ratio, sensitivity, specilicity, positive and negative predictive values, and tile associated 95% confidence intervals were calculated lbr the nmmmary stimulation test. Nhmiy-five percent conlldencc intervals Rn" the odds ratio were calculated with a logistic model to obtain a nlaxilnnln likelihood estimate of the log odds ratio with SYSTAT software. ''t A test that the log odds ratio was equal to zero (i.e., odds ratio = 1) was based on a l tesl) '~ Statistical signiticance was assumed at a level o f p < 0.05.
Results Two hundred sixty-seven patients were enrolled in the study. Maternal demographic data are presented in Table I. All patients started 1.n'enalal care in the first trimestm; and 257 (96.3%) of the patients had at least one ultrasonographic examination confirming gestational age. The test was well tolerated by the participants, and no one elected to withdraw from the study. Six patients, however, required observation in labor and delivery after termination of the test. Two patients were identified to have spontaneous uterine contractions during the baseline examination. Each patient demonstrated progressive change in cervical effacement and dilatation during observation in labor and delivery, and therefore intravenous magnesium sulfate was administered to arrest the uterine contractions. The patients were ultimately delivered at 32 and 33 weeks' gestation, respectively. Two patients with uterine activity (two or more contractions per 15 minutes) for more than 1 hour after nipple stinmlation was discontinued were observed in labor and delivery until spontaneous resolution of their contractions. Neither patient demonstrated change at cervical examination. Both were placed on modified bed rest by their health care providers and were subsequently delivered at ~ 37 weeks' gestation. The last two patients had F H R decelerations during the mammary stimulation test in the absence of uterine contractions. One fetus demonstrated recurrent spontaneous variable decelerations. Fetal hydrocephaly (ventricle/hemispheric ratio 70%) without additional congenital anomalies was identified on ultrasono-
Volume 17(I,Number 6 Am J Obstet Gynecol
Guinn et al.
1811
Table II. Sensitivity, specificity, and predictive measures of the mammary stimulation test*
I Sensitivity Specificity Positive predictive value Negative predictive value
<37wk 0,37 0.84 0.20 0.92
I
<34wk
(0.19-0.55) (0,79-0.88) (0.09-0.32) (0.89-0,96)
0.60 0.82 0.06 0.99
(0.17-1,00) (0,78.0.87) (0,00-0.13) (0.98-1,00)
95% Confidence intervals are in parentheses. *N = 265, one patient lost to follow-up and one patient excluded from analysis (see Results).
Table III. Cause of preterm birth
Cause
Mam~nary stimulation test (n = 10)
Idiopathic preterm labor Prcterm rupture of membranes Medically indicated preterm delivery Abruptio placentae
7 (70%) 2 (20%) 0 1 (10%)
graphic examination, This patient was not delivered preterm but was excluded from the outcome analysis because the mammm'y stimulation test was not completed. The other patient was completing the last 5 seconds of bilateral nipple stimulation when a 3-minute deceleration to 60 beats/min of the FHR occurred. The patient was subsequently monitored in labor and delivery for 4 hours without additional FHR decelerations. The pregnancy was otherwise uncomplicated, and a healthy neonate was delivered at 41 weeks' gestation. A positive mammary stimulation test was noted in 45 (17%) of tile 266 patients completing the test. Delivery occurred < 37 completed weeks'gestation in 27 (10.2%) patients and at < 3 4 weeks' gestation in five (1.9%) patients. The sensitivity, specificity, and positive and negative predictive values for the mammary stimulation test for delivery at < 37 and < 34 weeks are listed in Table II. The odds ratio for delivery at < 37 weeks was 3.0 (95% confidence interval 1.3, 7.1), and for delivery at < 34 weeks the odds ratio was 7.0 (95% confidence interval 1.1, 43.0). In addition, patients with a positive test were more likely to require observation in labor and delivery for complaints of regular uterine contractions (34% vs 4.3%, p < 0.01), and they were more likely to demonstrate change at cervical examination (14% vs 2%, p < 0.01). The demographic distribution of patients was similar to those who were delivered preterm versus those who were delivered at term and fbr those patients with a positive mammary stimulation test versus those with a negative test, T h e cause of preterm birth, however, was different between the positive and the negative groups, as enumerated in Table III. One third of preterm deliveries were secondary to idiopathic preterm labor; the mammary stimulation test was positive in 77,8% (seven of nine) of these pregnancies.
Mammary stimulation test (n = 17) 2 10 3 2
(11.8%) (58,8%) (17.6%) (11.8%)
Comment
Eden et al. ~ developed the mammary stimulation test to identify women at increased risk for preterm birth. They studied 92 medically indigent women at high risk for preterm delivery on the basis of past reproductive performance or a multiple gestation in the index pregnancy, These women were administered the mammary stimulation test at approximately 26 weeks' gestation and again at 30 weeks' gestation. A test was considered positive if any uterine contractions were noted, but the test results were not used for clinical management, The test was positive in 50% of patients, with a sensitivity of 84% and a positive predictive value of 34%. The negative predictive value in this high-risk population was 94%. The current study was designed to ascertain the utility of the mammary stimulation test in a population traditionally considered to be at low risk for preterm birth. The population tested was composed of privately insured, middle- to upper-class nulliparous patients with singleton gestations. The test was performed only once, to facilitate patient acceptance. The gestational age of 26 to 28 weeks was chosen in an effort to identify those women destined to be delivered before 34 weeks when the risk of neonatal morbidity and mortality is greatest. The definition of a positive test (i.e., two or more uterine contractions in 15 minutes) was set prospectively on the basis of published research experience with outpatient monitoring in low-risk patient populations]7. J~ The mammary stimulation test performed reasonably well in identifying patients at increased risk for preterm birth both at <37 and <34 weeks' gestation. The incidence of delivery at these gestational ages was 10.2% and 1,9%, respectively. The corresponding odds ratios for preterm birth after a positive test were 3.0
18t2
Guinn et al.
and 7.0. The sensitivity for delivery at < 3 4 weeks' gestation, the time o f greatest risk for neonatal morbidity, was 60%, which is h i g h c o m p a r e d with risk scoring systems2 H T h e positive predictive value o f 6%, however, was low and partially reflects the low incidence of p r e t e r m births at < 3 4 weeks. Perhaps the greatest value of the m a m m a r y stimulation test was its ability to p r e d i c t which patients were at low risk for p r e t e r m birth. The negative predictive value of 99% for p r e t e r m delivery at < 34 weeks is o n e of the highest r e p o r t e d in the literature. "~"~ T h e mammary s t i m u l a t i o n test is designed to predict patients at increased risk for idiopathic p r e t e r m labor. Potentially, if used as a screening test for p r e t e r m birth prevention p r o g r a m s , efforts could be directed at those patients at highest risk for p r e t e r m labor while obviating the need for i n c r e a s e d surveillance in others. T h e test has demonstrated efficacy in both a high-risk a n d a low-risk population a n d warrants further investigation. REFERENCES
1. Lee K, Paneth N, Gartner LM, Pearlman M. The very low birth weight rate: principal predictor of neonatal mortality in industrial populations. J Pediatr 1980;97:759-64, 2. Merkatz IR, Peter JB, Barden TP. Ritodrine hydrochloride: a beta-mimetic agent for use in preterm labor. II. Evidence of efficacy. Obstet Gynecol 1980;56:7-12. 3. Rush RW, Keirse MJ, Howat E et al. Contribution of preterm delivery to perinatal mortality. BMJ 1976; 2:965-9. 4. Lee KS, Paneth N, Gartner LM, Pearlman MA, Gruss L. Neonatal mortality: an analysis of recent improvement in the United States. Am J Public Health 1980;70:15-21. 5. Creasy RK, Gummer NE, Liggins GC. System for predicting spontaneous preterm birth. Obstet Gynecol 1980;55: 692-5. 6. Papiernik-Berkhauser E. Coefficient de risque d'accouchement pr6matur6. Presse Med 1969;77:793. 7. Nesbitt RE, Anbry RH. High risk obstetrics: value of semiobjective grading system in identifying the vulnerable group. AMJ OnSTET GVNECOL1969;103:972-5. 8. Akhtar J, Sehgal NN. Prognostic value of prepartum and intrapartum risk scoring method. South Med J 1980;73: 411-4. 9. Herron MA, Katz M, Greasy RK. Evaluation of preterm birth prevention program: preliminary report. Obstet Gynecol 1982;59:452-5. 10. Ross MG, Hobel CJ, Bragonier RJ, Bear MB. A simplified risk scoring system for prematurity. Am J Perinatol 1986; 3:339-44. 11. Main DM, Richardson D, Gabbe SG. Strong S, Weller SC. Prospective evaluation of a risk scoring system for predicting preterm delivery in black inner city women. Obstet Gynecol 1987;69:61-6. 12. Main DM, Gabbe SG, Risk scoring for preterm labor: where do we go from here? AM J OnSTET GwEeoL 1987; 157:789-93. 13. Eden R, Sokol RJ, Sorokin Y, Cook H, Sheeran G, Chik L. The MST (mammary stimulation test): a predictor of preterm delivery? AM J OBSTETGVNECOL1991;164:140919. 14. Wikinson L. Systat: the system for statistics. Evanston, Illinois: Systat, 1988. 15. Rat CR. Linear statistical inference and its applications. 2nd ed. New York: John Wiley, 1973. 16. Greene S, Prieve J, Morrison RR. The 1970 census of
June 1994 Am J Obstet Gynccol
population occupational classification system. Statistical Reporter 1969;77-84. 17. Main DM, Katz M, Chiu G, Campion S, Gabbe SG. Intermittent weekly contraction monitoring to predict preterm labor in low-risk women: a blinded study. Obstet Gynecol 1988;72:757-60. 18. Nageotte ME Dorchester W, Porto M, Keegan KA, Freeman RK. Quantification of uterine activity preceding preterm, term, and postterm labor. AM J O~STET GvNEeoL 1988;158:1254-9. Discussion DR. WASHINGTONC. HILL,Sarasota, Florida. T h e numb e r one p r o b l e m for obstetricians is p r e t e r m labor with subsequent p r e t e r m delivery. In 1993 our p r e d i c t i o n of which patients will have p r e t e r m labor and delivery is still n o t good. Guinn et al. have continued t h e work of Eden et al., ~ which was first r e p o r t e d at our 1990 meeting, and suggest a good, positive predictive value of the m a m m a r y stimulation test in a high-risk pregnant population. In my discussion of that p a p e r I felt further study of the m a m m a r y stimulation test was warranted. Guinn et al. have done just that by evaluating the predictive value o f the test for identifying mdliparous, private patients at high-risk f o r preterm birth. T h e overall value of the m a m m a r y stimulation test in this study is shown in Table I. Look particularly at its usefulness in predicting p r e t e r m delivery at < 34 weeks: good sensitivity, good specificity, p o o r positive predictive value, a n d great negative predictive value. These are encouraging results. Let's look closer at t h e results in predicting preterm birth at < 3 4 weeks. In their study the m a m m a r y stimulation test is excellent in predicting which patients will not deliver at < 34 weeks. This is good, because it is the < 3 4 - w e e k - o l d neonate who has significant morbidity and mortality, causing the neonatologist so much trouble. However, this study also showed that the positive predictive value for the same gestational age was p o o r at only 6%. This is not good. We n e e d a better system. However, the overall numbers of the patients destined to be delivered at < 34 weeks was only five. For this test to have better clinical usefulness, we need a higher positive predictive value so that some a p p r o p r i a t e intervention or t r e a t m e n t can be applied to this subgroup. Let's not d e b a t e what is " a p p r o p r i a t e intervention." We all have our own thoughts and biases. C o m p a r e d with other screening tests, the results of this test are not all bad. T h i n k about costly maternal Table I, Overall value of m a m m a r y stimulation test in nulliparous p o p u l a t i o n for p r e d i c t i n g preterln delivery <37 wk
Sensitivitiy (%) Specificity (%) Positive predictive value (%) Negative predictive value (%)
37 84 20 92
..
< 3 4 wk
60 82 6 99
Volume i70, Number 6 Am J Obstet Gynecol
Guinn et al, 1813
fulness of the m a m m a r y stimulation test? and (4) Why were there more patients with r u p t u r e of the membranes in the group with a negative test? REFERENCES
Fig. 1. Use of mammary stimulation test (MST) in combination with fetal fibronectin, cewical evaluation, and risk scoring in identifying women at increased risk for preterrn birth.
s e r u m e~-fetoprotein screening, which has a low positive predictive value, significant cost, and subsequent cost for evaluation o f the positive subgroup. T h e mammary stimulation test, on the other hand, has a low cost and could potentially save health dollars. Other advantages are it is easy to perform, safe, really not a bad reassuring sign for the n o r m a l patient, and identifies 1 7 % - a r e a s o n a b l e n u m b e r of p a t i e n t s - a s high risk. There are few disadvantages, b u t the problems are the number of patients studied to date with a positive test remains s m a l l - - 4 5 , or 17% of the total population in this study. T h e identification by a screening test of a 17% subg r o u p s is not b a d , b u t we need to study more patients with a positive m a m m a r y stimulation test. Another p r o b l e m is the test in this study was not a good screening test for abnormal results. It will b e difficult for any one test to predict preterm l a b o r or delivery; however; the use of this test in combin a t i o n with fetal fibronectin, '~'*cervical evaluation, ~' ~'and risk scoring ~' 8 in b o t h high- and low-risk patients may be its real role in identifying a subgroup of women who a r e at increased risk for preterm birth (Fig. 1). T h e m a m m a r y stimulation test appears in this study to be o f some value in predicting preterm birth in an e d u c a t e d , private, nulliparous population. I believe that it is of some value in a nulliparous population. This work does help further validate the role of the mamm a r y stimulation test in pretnaturity prevention. Yours is the largest series to date. The time has come for us to p e r f o r m a well-designed, prospective, collaborative, m u l t i c e n t e r e d study. I have several questions: (1) How would you suggest we use the test now? (2) What would you suggest we do with patients who have a positive test? (3) What is the n e x t step in the further evaluation of the clinical use-
1. Eden R, Sokol RJ, Sorokin Y, Cook H, Sheeran G, Chik L. The MST (mammary stimulation test): a predictor of preterm delivery? AMj OBSTETGVNECOL1991 ; 164:1409-19. 2. Lockwood CJ, Senyei AE, Dische MR, et al. Fetal fibronectin in cervical and vaginal secretions as a predictor of preterm delivery. N Engl J Med 1991;325:669-74. 3. Lockwood CJ, Wein R, Lapinski R, et al. The presence of cervical and vaginal fetal fibronectin predicts preterm delivery in an inner-city obstetric population. AM J OBST~-r G~'ECOL 1993; 169:798-804. 4. Morrison JC, Albert JR, McLaughlin BN, Whitworth NS, Roberts WE, Martin RW. Oncofetal fibronectin in patients with false labor as a predictor of preterm delivery. A~l J OBsaxr C,YNECOL1993; 168:538-42. 5. Leveno KJ, Cox K, Roark ML. Cervical dilatation and prematurity revisited. Obstet Gynecol 1986;68:434-5. 6. Stubbs TM, Van Dorsten P, Miller MC. The preterm cervix and preterm labor: relative risks, predictive values and change over time. AMJ OBsax'vGVN~COL1986;155:829-34. 7. Creasy RK, Gremmer BA, Liggins GC. System for predicting preterm birth. Obstet Cynecol 1980;55:692-5. 8. Main DM, Gabbe SG, Risk scoring for preterm labor: where do we go from here? AMJ OBSTETGYNEGOL1987;157:78993. DR. PAUL L. OCBIJRN,JS., Rochester, Minnesota. This study is the result of originaI r e s e a r c h a p p l i e d to a population of 267 nulliparous women from a private practice. The study design is adequate a n d appropriate. The results are presented with clarity and the statistical analysis is adequate. The length is reasonable and the tables add clarity without adding repetitiveness. Determining patients at risk for preterm labor and preterm birth is exceedingly important to American obstetrics. The methods, procedures, interpretation of results, and discussion are adequate. I am aware o f no previously published articles by these authors that use this research protocol. T h e key to the eventual usefulness of mammary stimulation test is whether it will be o f value. A definition that can be given for value in medicine in 1993 indudes one or all of the following: (1) something is of value if it saves lives, (2) something is o f value if it saves money, and (3) something is of value if it prevents morbidity (which indirectly saves money). With this definition the m a m m a r y stimulation test may prove to be o f vatue in two possible instances. (1) Negative tests may be used to avoid costly sm'veillance (such as home uterine monitoring) for preterm birth prevention. (2) Positive tests may help prevent preterm birth if these patients respond to s o m e therapy that prolongs pregnancy. I have several questions. The answers to some of these questions may serve to clarify the potential usefulness of the mammary stimulation test and improve the potential for understanding the results of this study, (1) How many patients with risk factors for preterm birth were included in this study? W e r e patients at high risk for preterm birth excluded? Was a risk scoring
1814
Guinn et al.
system applied to y o u r p o p u l a t i o n ? If so, which system? (2) Using the m a m m a r y stimulation test between 26 and 28 weeks' gestation cannot alert health care p r o viders to the risk o f p r e t e r m b i r t h between 20 weeks and 26 to 28 weeks. W h a t is the rate o f p r e t e r m births in your institution from 20 to 28 weeks' gestation? (3) A rate o f p r e t e r m births o f 10.2% seems somewhat high in a group of private patients. Can we b e sure the use o f the mammary s t i m u l a t i o n test d i d n o t stimulate this increase in p r e t e r m births? Is it possible that the fact that 91.8% of women in this study worked full time (see Table I) contributed to t h e a p p a r e n t increase in preterm births from 34 to 37 weeks' gestation? (4) If the mammary stimulation test (or any o t h e r test) can predict preterm labor, how should we use this knowledge to prevent preterm birth? W h a t t r e a t m e n t should we use if the mammary stimulation test is positive? In conclusion, I believe the authors have shown the mammary stimulation test to be useful in d e t e r m i n i n g which pregnancies are at risk for p r e t e r m l a b o r after 26 to 28 weeks' gestation. W h e t h e r the use o f this study can eventually l e a d to p r e t e r m birth prevention m u s t await furore studies. DR. GUINN (Closing). T h e m a m m a r y stimulation test performed reasonably well for the identification of nulliparous patients at risk f o r s p o n t a n e o u s p r e t e r m birth. It performed best in identi~/ing patients at risk for idiopathic p r e t e r m l a b o r who were delivered p r e t e r m (7/9) and worst in patients with p r e t e r m p r e m a t u r e rupture of the m e m b r a n e s (2/11) who were delivered preterm. Our study design i n c o r p o r a t e d the same nipple stimulation protocol u s e d in tile r e p o r t of E d e n et al. They defined a positive test as the presence of any uterine contractions. In that p o p u l a t i o n o f high-risk women, 50% of the p a t i e n t s h a d a positive test. In an effort to improve the positive predictive value of the mammary stimulation test we defined a positive test as the presence of two or m o r e contractions lasting _>40 seconds during any 15 minutes o f monitoring. With this criteria the incidence of a positive test in our low-risk population was 45 o f 265 (17%) a n d the positive predictive value was 20%. Future research efforts are needed to evaluate alternative n i p p l e stimulation p r o tocols and different criteria tbr a positive test. In regard to Dr. O g b u r n ' s questions, the only exclusions were multiple gestations, k n o w n placenta previa, and preterm p r e m a t u r e r u p t u r e of the membranes. Although not p r e s e n t e d here, all of o u r patients had
June 1994 Am J Obstet Gynecol
Creasy risk scores calculated at their first prenatal visits and again at 26 to 28 weeks' gestation. T h e mean score for the entire population at the first visit was 2.9 --- 2.0, and the mean score at 26 to 28 weeks was 3.2 ± 2,3. Only five patients (1.8%) at the first prenatal visit and 11 (4.2%) at 26 to 28 weeks' gestation had a Creasy risk score >_ 10. T h e r e were no significant differences in either score for those patients delivered preterm versus those patients delivered at term. T h e m a m m a r y stimulation test was performed once at 26 to 28 weeks' gestation to facilitate patient acceptance. This gestational age was chosen in an effort to identify those women destined to be delivered before 34 weeks, when the risk of neonatal morbidity and mortalit,/ is greatest. If our results had confirmed those of Eden et al., then we would have anticipated that no patients with a negative test would have been delivered within a month of p e r f o r m i n g the test. At Northwestern Memorial Hospital the incidence of p r e t e r m birth between 20 a n d 30 weeks' gestation is 1% to 2%, whereas the overall incidence o f preterm birth at < 3 7 weeks' gestation is 11% to 15%. In our private nulliparous study p o p u l a t i o n the incidence of preterrn birth was 10.2%, which was slightly higher than expected. T h e incidence of p r e t e r m birth in the patients who refused to participate in the study was 7.8%. It is conceivable that the patients who consented to the study did so because they felt that they were at an increased risk for p r e t e r m delivery. As Dr. Ogburn mentioned, 91.8% of our patients were employed in u p p e r level trade or professional occupations. It is possible that the type of e m p l o y m e n t or the fact that they were employed affected the rate of p r e t e r m delivery. We do not think that the m a m m a r y stimulation test increased the risk o f p r e t e r m birth. None of the patients in either our study or that of Eden et al. had p r c t e r m labor after the test. In our study only one patient was delivered within 3 weeks o f performing the test; h e r pregnancy was complicated by abruptio placentae. We believe that the m a m m a r y stimulation test can identify a group of patients at increased risk for spontaneous p r e t e r m birth. In spite of the obvious limitations o f risk assessment a n d the lack of proven treatments for those patients at risk for spontaneous preterm birth, the ability to identify high-risk patients is crucial. This knowledge should p e r m i t investigators to i m p l e m e n t clinical trials with significant power to thoroughly evaluate prematurity prevention initiatives.
Key words: Mammary stimulation test, preterm birth, nulliparity Although preterm birth complicates only 7% to 10% of all pregnancies, prematurity is responsible for 70% to 80% of perinatal mortality. I~ Unfortunately, only 50% of patients who are delivered preterm have historical and demographic factors associated with preterm birth. Nulliparous patients account for 25% to 40% of all preterm births, and they are the most difficult pregnancies to prospectively identify as being at increased risk. 4-~'~ The mammary stimulation test, a test of uterine contractility in response to endogenous oxytocin, was From the Section of Maternal-Fetal Medicine, Departmentof Obstetrics and Gy~ecology," and the General Clinical Research Centerf Northwestern University Medical School, Northwestern Memorial Hospital, and the Centerfor Health Se~wicesand Policy Research, Northwestern University." Supported in part by National Center for Research Resources, National Institutes of Health, grant No. RR-O0048. Presented at the Sixty-first Annual Meeting of The CentralAssociation of Obstetricians and Gynecologists, White Sulphur springs, West Virginia, October 28-30, 1993. Repnnt requests: Debra Guinn, MD, 618 S. 20th St,, Birmingham, AL 35233. Copyright © 1994 by Mosby-Year Book, Inc. 0002-9378/94 $3.00 + 0 6/6/55032
developed by Eden et al. ~s to identify patients at increased risk for preterm delivery. In a medically indigent population the test demonstrated a sensitivity of 84% and a positive predictive value of 34%. To further test the efficacy of the mammary stimulation test, we conducted a prospective clinical trial in a private nul. liparous patient population, a population traditionally considered to be at low risk for preterm birth and largely without identifiable risk factors. Material and methods
Nulliparous patients receiving prenatal care from private practitioners affiliated with Northwestern Memorial Hospital were eligible for the study, which was conducted from January 1992 through February 1993. Patients with a placenta previa, multiple gestation, or ruptured membranes were excluded from participation. All patients were recruited by their attending health care provider. The research protocol was approved by the Northwestern University Institutional Review Board, and written informed consent was obtained from all patients before participation. The mammary stimulation test was performed be1809
1810
.June 199,l Am .] Obstct Gynecol
Guinn et al.
T a b l e I. Maternal demographic data
(N = 267) Age (yr) Gravidity Elective abortions 1
->2 (range 2-4) Spontaneous abortions 1 ---2 (range 2-3) Race White Black Hispanic Other Tobacco use Illicit drug use Education (yr) Occupation* Homemaker Trade Professional
31.7 +_ 4.2 1.6 +- .96 79 (30%) 59 20 48 (18%) 35 13 231 (86.5%) 14 (5.2%) 7 (2.6%) 15 (5.6%) 9 (3A%) 3 (1.1%) 16.5 "-: 1.8 22 (8.2%) 112 (41.9%) 133 (49.8%)
*Occupalion description based on Maim' Professional Codes, U.S. Census Report. 1¢'
tween 26 and 28 weeks' gestation in the Clinical Research Center of Northwestern Memorial Hospital by registered nurses trained in conducting the mammary stimulation test and in basic [~tal heart rate (FHR) interpretation. Gestational age assessment was based on last menstrual period, first-trimester physical examination, or ultrasonographic examination before 24 weeks of pregnancy. Patients were placed in the left lateral decubitus position and a Doppler ultrasonographic transducer and a uterine tocodynometer were placed across the maternal abdomen. A 10- to 20-minute baseline strip of FHR and uterine activity was obtained. In the absence of spontaneous uterine contractions the patient was asked to lightly rub one nipple through her clothing for 2 minutes, followed by a 2-minute rest period. This cycle was repeated three times, followed by unilateral nipple stimulation for 10 minutes and finally bilateral nipple stimulation for 10 minutes. All tests were interpreted by one investigator (D.A,G.) on the day that they were performed. The test was scored as positive if two or more contractions of ~ 4 0 seconds' duration occurred during any 15 minutes of monitoring, including the baseline strip. If this frequency of uterine contractions occurred at any time during the test, nipple stimulation was discontinued and the patient was monitored until uterine activity spontaneously subsided (i,e,, fewer than two contractions in 15 minutes). Both the patient and the attending health care provider were blinded to the results of the mammary stimulation test, except under the following circumstances: spontaneous contractions (two or more per 15 minutes) occurring before stimulation, contractions occurring for > 1 hour after stimulation, or the
presence of an abnol'mal FHR pattern. Pregnancy outcome data were retrieved through the perinatal database and medical records. Sample size calculations were based on the observed results froln Eden et al. ~'~ However, for this low-risk population it was assumed that the sensitivity would be lowered to 60%, the rate of positive tests would be lowered to only 25%, and the incidence of preterrn births would be only 7%. Theretbrc a sample size of 252 would have a 90% power of detecting a relationship between the mammary stimulation test and delivery outcome witb an ~ of 0,05. The odds ratio, sensitivity, specilicity, positive and negative predictive values, and tile associated 95% confidence intervals were calculated lbr the nmmmary stimulation test. Nhmiy-five percent conlldencc intervals Rn" the odds ratio were calculated with a logistic model to obtain a nlaxilnnln likelihood estimate of the log odds ratio with SYSTAT software. ''t A test that the log odds ratio was equal to zero (i.e., odds ratio = 1) was based on a l tesl) '~ Statistical signiticance was assumed at a level o f p < 0.05.
Results Two hundred sixty-seven patients were enrolled in the study. Maternal demographic data are presented in Table I. All patients started 1.n'enalal care in the first trimestm; and 257 (96.3%) of the patients had at least one ultrasonographic examination confirming gestational age. The test was well tolerated by the participants, and no one elected to withdraw from the study. Six patients, however, required observation in labor and delivery after termination of the test. Two patients were identified to have spontaneous uterine contractions during the baseline examination. Each patient demonstrated progressive change in cervical effacement and dilatation during observation in labor and delivery, and therefore intravenous magnesium sulfate was administered to arrest the uterine contractions. The patients were ultimately delivered at 32 and 33 weeks' gestation, respectively. Two patients with uterine activity (two or more contractions per 15 minutes) for more than 1 hour after nipple stinmlation was discontinued were observed in labor and delivery until spontaneous resolution of their contractions. Neither patient demonstrated change at cervical examination. Both were placed on modified bed rest by their health care providers and were subsequently delivered at ~ 37 weeks' gestation. The last two patients had F H R decelerations during the mammary stimulation test in the absence of uterine contractions. One fetus demonstrated recurrent spontaneous variable decelerations. Fetal hydrocephaly (ventricle/hemispheric ratio 70%) without additional congenital anomalies was identified on ultrasono-
Volume 17(I,Number 6 Am J Obstet Gynecol
Guinn et al.
1811
Table II. Sensitivity, specificity, and predictive measures of the mammary stimulation test*
I Sensitivity Specificity Positive predictive value Negative predictive value
<37wk 0,37 0.84 0.20 0.92
I
<34wk
(0.19-0.55) (0,79-0.88) (0.09-0.32) (0.89-0,96)
0.60 0.82 0.06 0.99
(0.17-1,00) (0,78.0.87) (0,00-0.13) (0.98-1,00)
95% Confidence intervals are in parentheses. *N = 265, one patient lost to follow-up and one patient excluded from analysis (see Results).
Table III. Cause of preterm birth
Cause
Mam~nary stimulation test (n = 10)
Idiopathic preterm labor Prcterm rupture of membranes Medically indicated preterm delivery Abruptio placentae
7 (70%) 2 (20%) 0 1 (10%)
graphic examination, This patient was not delivered preterm but was excluded from the outcome analysis because the mammm'y stimulation test was not completed. The other patient was completing the last 5 seconds of bilateral nipple stimulation when a 3-minute deceleration to 60 beats/min of the FHR occurred. The patient was subsequently monitored in labor and delivery for 4 hours without additional FHR decelerations. The pregnancy was otherwise uncomplicated, and a healthy neonate was delivered at 41 weeks' gestation. A positive mammary stimulation test was noted in 45 (17%) of tile 266 patients completing the test. Delivery occurred < 37 completed weeks'gestation in 27 (10.2%) patients and at < 3 4 weeks' gestation in five (1.9%) patients. The sensitivity, specificity, and positive and negative predictive values for the mammary stimulation test for delivery at < 37 and < 34 weeks are listed in Table II. The odds ratio for delivery at < 37 weeks was 3.0 (95% confidence interval 1.3, 7.1), and for delivery at < 34 weeks the odds ratio was 7.0 (95% confidence interval 1.1, 43.0). In addition, patients with a positive test were more likely to require observation in labor and delivery for complaints of regular uterine contractions (34% vs 4.3%, p < 0.01), and they were more likely to demonstrate change at cervical examination (14% vs 2%, p < 0.01). The demographic distribution of patients was similar to those who were delivered preterm versus those who were delivered at term and fbr those patients with a positive mammary stimulation test versus those with a negative test, T h e cause of preterm birth, however, was different between the positive and the negative groups, as enumerated in Table III. One third of preterm deliveries were secondary to idiopathic preterm labor; the mammary stimulation test was positive in 77,8% (seven of nine) of these pregnancies.
Mammary stimulation test (n = 17) 2 10 3 2
(11.8%) (58,8%) (17.6%) (11.8%)
Comment
Eden et al. ~ developed the mammary stimulation test to identify women at increased risk for preterm birth. They studied 92 medically indigent women at high risk for preterm delivery on the basis of past reproductive performance or a multiple gestation in the index pregnancy, These women were administered the mammary stimulation test at approximately 26 weeks' gestation and again at 30 weeks' gestation. A test was considered positive if any uterine contractions were noted, but the test results were not used for clinical management, The test was positive in 50% of patients, with a sensitivity of 84% and a positive predictive value of 34%. The negative predictive value in this high-risk population was 94%. The current study was designed to ascertain the utility of the mammary stimulation test in a population traditionally considered to be at low risk for preterm birth. The population tested was composed of privately insured, middle- to upper-class nulliparous patients with singleton gestations. The test was performed only once, to facilitate patient acceptance. The gestational age of 26 to 28 weeks was chosen in an effort to identify those women destined to be delivered before 34 weeks when the risk of neonatal morbidity and mortality is greatest. The definition of a positive test (i.e., two or more uterine contractions in 15 minutes) was set prospectively on the basis of published research experience with outpatient monitoring in low-risk patient populations]7. J~ The mammary stimulation test performed reasonably well in identifying patients at increased risk for preterm birth both at <37 and <34 weeks' gestation. The incidence of delivery at these gestational ages was 10.2% and 1,9%, respectively. The corresponding odds ratios for preterm birth after a positive test were 3.0
18t2
Guinn et al.
and 7.0. The sensitivity for delivery at < 3 4 weeks' gestation, the time o f greatest risk for neonatal morbidity, was 60%, which is h i g h c o m p a r e d with risk scoring systems2 H T h e positive predictive value o f 6%, however, was low and partially reflects the low incidence of p r e t e r m births at < 3 4 weeks. Perhaps the greatest value of the m a m m a r y stimulation test was its ability to p r e d i c t which patients were at low risk for p r e t e r m birth. The negative predictive value of 99% for p r e t e r m delivery at < 34 weeks is o n e of the highest r e p o r t e d in the literature. "~"~ T h e mammary s t i m u l a t i o n test is designed to predict patients at increased risk for idiopathic p r e t e r m labor. Potentially, if used as a screening test for p r e t e r m birth prevention p r o g r a m s , efforts could be directed at those patients at highest risk for p r e t e r m labor while obviating the need for i n c r e a s e d surveillance in others. T h e test has demonstrated efficacy in both a high-risk a n d a low-risk population a n d warrants further investigation. REFERENCES
1. Lee K, Paneth N, Gartner LM, Pearlman M. The very low birth weight rate: principal predictor of neonatal mortality in industrial populations. J Pediatr 1980;97:759-64, 2. Merkatz IR, Peter JB, Barden TP. Ritodrine hydrochloride: a beta-mimetic agent for use in preterm labor. II. Evidence of efficacy. Obstet Gynecol 1980;56:7-12. 3. Rush RW, Keirse MJ, Howat E et al. Contribution of preterm delivery to perinatal mortality. BMJ 1976; 2:965-9. 4. Lee KS, Paneth N, Gartner LM, Pearlman MA, Gruss L. Neonatal mortality: an analysis of recent improvement in the United States. Am J Public Health 1980;70:15-21. 5. Creasy RK, Gummer NE, Liggins GC. System for predicting spontaneous preterm birth. Obstet Gynecol 1980;55: 692-5. 6. Papiernik-Berkhauser E. Coefficient de risque d'accouchement pr6matur6. Presse Med 1969;77:793. 7. Nesbitt RE, Anbry RH. High risk obstetrics: value of semiobjective grading system in identifying the vulnerable group. AMJ OnSTET GVNECOL1969;103:972-5. 8. Akhtar J, Sehgal NN. Prognostic value of prepartum and intrapartum risk scoring method. South Med J 1980;73: 411-4. 9. Herron MA, Katz M, Greasy RK. Evaluation of preterm birth prevention program: preliminary report. Obstet Gynecol 1982;59:452-5. 10. Ross MG, Hobel CJ, Bragonier RJ, Bear MB. A simplified risk scoring system for prematurity. Am J Perinatol 1986; 3:339-44. 11. Main DM, Richardson D, Gabbe SG. Strong S, Weller SC. Prospective evaluation of a risk scoring system for predicting preterm delivery in black inner city women. Obstet Gynecol 1987;69:61-6. 12. Main DM, Gabbe SG, Risk scoring for preterm labor: where do we go from here? AM J OnSTET GwEeoL 1987; 157:789-93. 13. Eden R, Sokol RJ, Sorokin Y, Cook H, Sheeran G, Chik L. The MST (mammary stimulation test): a predictor of preterm delivery? AM J OBSTETGVNECOL1991;164:140919. 14. Wikinson L. Systat: the system for statistics. Evanston, Illinois: Systat, 1988. 15. Rat CR. Linear statistical inference and its applications. 2nd ed. New York: John Wiley, 1973. 16. Greene S, Prieve J, Morrison RR. The 1970 census of
June 1994 Am J Obstet Gynccol
population occupational classification system. Statistical Reporter 1969;77-84. 17. Main DM, Katz M, Chiu G, Campion S, Gabbe SG. Intermittent weekly contraction monitoring to predict preterm labor in low-risk women: a blinded study. Obstet Gynecol 1988;72:757-60. 18. Nageotte ME Dorchester W, Porto M, Keegan KA, Freeman RK. Quantification of uterine activity preceding preterm, term, and postterm labor. AM J O~STET GvNEeoL 1988;158:1254-9. Discussion DR. WASHINGTONC. HILL,Sarasota, Florida. T h e numb e r one p r o b l e m for obstetricians is p r e t e r m labor with subsequent p r e t e r m delivery. In 1993 our p r e d i c t i o n of which patients will have p r e t e r m labor and delivery is still n o t good. Guinn et al. have continued t h e work of Eden et al., ~ which was first r e p o r t e d at our 1990 meeting, and suggest a good, positive predictive value of the m a m m a r y stimulation test in a high-risk pregnant population. In my discussion of that p a p e r I felt further study of the m a m m a r y stimulation test was warranted. Guinn et al. have done just that by evaluating the predictive value o f the test for identifying mdliparous, private patients at high-risk f o r preterm birth. T h e overall value of the m a m m a r y stimulation test in this study is shown in Table I. Look particularly at its usefulness in predicting p r e t e r m delivery at < 34 weeks: good sensitivity, good specificity, p o o r positive predictive value, a n d great negative predictive value. These are encouraging results. Let's look closer at t h e results in predicting preterm birth at < 3 4 weeks. In their study the m a m m a r y stimulation test is excellent in predicting which patients will not deliver at < 34 weeks. This is good, because it is the < 3 4 - w e e k - o l d neonate who has significant morbidity and mortality, causing the neonatologist so much trouble. However, this study also showed that the positive predictive value for the same gestational age was p o o r at only 6%. This is not good. We n e e d a better system. However, the overall numbers of the patients destined to be delivered at < 34 weeks was only five. For this test to have better clinical usefulness, we need a higher positive predictive value so that some a p p r o p r i a t e intervention or t r e a t m e n t can be applied to this subgroup. Let's not d e b a t e what is " a p p r o p r i a t e intervention." We all have our own thoughts and biases. C o m p a r e d with other screening tests, the results of this test are not all bad. T h i n k about costly maternal Table I, Overall value of m a m m a r y stimulation test in nulliparous p o p u l a t i o n for p r e d i c t i n g preterln delivery <37 wk
Sensitivitiy (%) Specificity (%) Positive predictive value (%) Negative predictive value (%)
37 84 20 92
..
< 3 4 wk
60 82 6 99
Volume i70, Number 6 Am J Obstet Gynecol
Guinn et al, 1813
fulness of the m a m m a r y stimulation test? and (4) Why were there more patients with r u p t u r e of the membranes in the group with a negative test? REFERENCES
Fig. 1. Use of mammary stimulation test (MST) in combination with fetal fibronectin, cewical evaluation, and risk scoring in identifying women at increased risk for preterrn birth.
s e r u m e~-fetoprotein screening, which has a low positive predictive value, significant cost, and subsequent cost for evaluation o f the positive subgroup. T h e mammary stimulation test, on the other hand, has a low cost and could potentially save health dollars. Other advantages are it is easy to perform, safe, really not a bad reassuring sign for the n o r m a l patient, and identifies 1 7 % - a r e a s o n a b l e n u m b e r of p a t i e n t s - a s high risk. There are few disadvantages, b u t the problems are the number of patients studied to date with a positive test remains s m a l l - - 4 5 , or 17% of the total population in this study. T h e identification by a screening test of a 17% subg r o u p s is not b a d , b u t we need to study more patients with a positive m a m m a r y stimulation test. Another p r o b l e m is the test in this study was not a good screening test for abnormal results. It will b e difficult for any one test to predict preterm l a b o r or delivery; however; the use of this test in combin a t i o n with fetal fibronectin, '~'*cervical evaluation, ~' ~'and risk scoring ~' 8 in b o t h high- and low-risk patients may be its real role in identifying a subgroup of women who a r e at increased risk for preterm birth (Fig. 1). T h e m a m m a r y stimulation test appears in this study to be o f some value in predicting preterm birth in an e d u c a t e d , private, nulliparous population. I believe that it is of some value in a nulliparous population. This work does help further validate the role of the mamm a r y stimulation test in pretnaturity prevention. Yours is the largest series to date. The time has come for us to p e r f o r m a well-designed, prospective, collaborative, m u l t i c e n t e r e d study. I have several questions: (1) How would you suggest we use the test now? (2) What would you suggest we do with patients who have a positive test? (3) What is the n e x t step in the further evaluation of the clinical use-
1. Eden R, Sokol RJ, Sorokin Y, Cook H, Sheeran G, Chik L. The MST (mammary stimulation test): a predictor of preterm delivery? AMj OBSTETGVNECOL1991 ; 164:1409-19. 2. Lockwood CJ, Senyei AE, Dische MR, et al. Fetal fibronectin in cervical and vaginal secretions as a predictor of preterm delivery. N Engl J Med 1991;325:669-74. 3. Lockwood CJ, Wein R, Lapinski R, et al. The presence of cervical and vaginal fetal fibronectin predicts preterm delivery in an inner-city obstetric population. AM J OBST~-r G~'ECOL 1993; 169:798-804. 4. Morrison JC, Albert JR, McLaughlin BN, Whitworth NS, Roberts WE, Martin RW. Oncofetal fibronectin in patients with false labor as a predictor of preterm delivery. A~l J OBsaxr C,YNECOL1993; 168:538-42. 5. Leveno KJ, Cox K, Roark ML. Cervical dilatation and prematurity revisited. Obstet Gynecol 1986;68:434-5. 6. Stubbs TM, Van Dorsten P, Miller MC. The preterm cervix and preterm labor: relative risks, predictive values and change over time. AMJ OBsax'vGVN~COL1986;155:829-34. 7. Creasy RK, Gremmer BA, Liggins GC. System for predicting preterm birth. Obstet Cynecol 1980;55:692-5. 8. Main DM, Gabbe SG, Risk scoring for preterm labor: where do we go from here? AMJ OBSTETGYNEGOL1987;157:78993. DR. PAUL L. OCBIJRN,JS., Rochester, Minnesota. This study is the result of originaI r e s e a r c h a p p l i e d to a population of 267 nulliparous women from a private practice. The study design is adequate a n d appropriate. The results are presented with clarity and the statistical analysis is adequate. The length is reasonable and the tables add clarity without adding repetitiveness. Determining patients at risk for preterm labor and preterm birth is exceedingly important to American obstetrics. The methods, procedures, interpretation of results, and discussion are adequate. I am aware o f no previously published articles by these authors that use this research protocol. T h e key to the eventual usefulness of mammary stimulation test is whether it will be o f value. A definition that can be given for value in medicine in 1993 indudes one or all of the following: (1) something is of value if it saves lives, (2) something is o f value if it saves money, and (3) something is of value if it prevents morbidity (which indirectly saves money). With this definition the m a m m a r y stimulation test may prove to be o f vatue in two possible instances. (1) Negative tests may be used to avoid costly sm'veillance (such as home uterine monitoring) for preterm birth prevention. (2) Positive tests may help prevent preterm birth if these patients respond to s o m e therapy that prolongs pregnancy. I have several questions. The answers to some of these questions may serve to clarify the potential usefulness of the mammary stimulation test and improve the potential for understanding the results of this study, (1) How many patients with risk factors for preterm birth were included in this study? W e r e patients at high risk for preterm birth excluded? Was a risk scoring
1814
Guinn et al.
system applied to y o u r p o p u l a t i o n ? If so, which system? (2) Using the m a m m a r y stimulation test between 26 and 28 weeks' gestation cannot alert health care p r o viders to the risk o f p r e t e r m b i r t h between 20 weeks and 26 to 28 weeks. W h a t is the rate o f p r e t e r m births in your institution from 20 to 28 weeks' gestation? (3) A rate o f p r e t e r m births o f 10.2% seems somewhat high in a group of private patients. Can we b e sure the use o f the mammary s t i m u l a t i o n test d i d n o t stimulate this increase in p r e t e r m births? Is it possible that the fact that 91.8% of women in this study worked full time (see Table I) contributed to t h e a p p a r e n t increase in preterm births from 34 to 37 weeks' gestation? (4) If the mammary stimulation test (or any o t h e r test) can predict preterm labor, how should we use this knowledge to prevent preterm birth? W h a t t r e a t m e n t should we use if the mammary stimulation test is positive? In conclusion, I believe the authors have shown the mammary stimulation test to be useful in d e t e r m i n i n g which pregnancies are at risk for p r e t e r m l a b o r after 26 to 28 weeks' gestation. W h e t h e r the use o f this study can eventually l e a d to p r e t e r m birth prevention m u s t await furore studies. DR. GUINN (Closing). T h e m a m m a r y stimulation test performed reasonably well for the identification of nulliparous patients at risk f o r s p o n t a n e o u s p r e t e r m birth. It performed best in identi~/ing patients at risk for idiopathic p r e t e r m l a b o r who were delivered p r e t e r m (7/9) and worst in patients with p r e t e r m p r e m a t u r e rupture of the m e m b r a n e s (2/11) who were delivered preterm. Our study design i n c o r p o r a t e d the same nipple stimulation protocol u s e d in tile r e p o r t of E d e n et al. They defined a positive test as the presence of any uterine contractions. In that p o p u l a t i o n o f high-risk women, 50% of the p a t i e n t s h a d a positive test. In an effort to improve the positive predictive value of the mammary stimulation test we defined a positive test as the presence of two or m o r e contractions lasting _>40 seconds during any 15 minutes o f monitoring. With this criteria the incidence of a positive test in our low-risk population was 45 o f 265 (17%) a n d the positive predictive value was 20%. Future research efforts are needed to evaluate alternative n i p p l e stimulation p r o tocols and different criteria tbr a positive test. In regard to Dr. O g b u r n ' s questions, the only exclusions were multiple gestations, k n o w n placenta previa, and preterm p r e m a t u r e r u p t u r e of the membranes. Although not p r e s e n t e d here, all of o u r patients had
June 1994 Am J Obstet Gynecol
Creasy risk scores calculated at their first prenatal visits and again at 26 to 28 weeks' gestation. T h e mean score for the entire population at the first visit was 2.9 --- 2.0, and the mean score at 26 to 28 weeks was 3.2 ± 2,3. Only five patients (1.8%) at the first prenatal visit and 11 (4.2%) at 26 to 28 weeks' gestation had a Creasy risk score >_ 10. T h e r e were no significant differences in either score for those patients delivered preterm versus those patients delivered at term. T h e m a m m a r y stimulation test was performed once at 26 to 28 weeks' gestation to facilitate patient acceptance. This gestational age was chosen in an effort to identify those women destined to be delivered before 34 weeks, when the risk of neonatal morbidity and mortalit,/ is greatest. If our results had confirmed those of Eden et al., then we would have anticipated that no patients with a negative test would have been delivered within a month of p e r f o r m i n g the test. At Northwestern Memorial Hospital the incidence of p r e t e r m birth between 20 a n d 30 weeks' gestation is 1% to 2%, whereas the overall incidence o f preterm birth at < 3 7 weeks' gestation is 11% to 15%. In our private nulliparous study p o p u l a t i o n the incidence of preterrn birth was 10.2%, which was slightly higher than expected. T h e incidence of p r e t e r m birth in the patients who refused to participate in the study was 7.8%. It is conceivable that the patients who consented to the study did so because they felt that they were at an increased risk for p r e t e r m delivery. As Dr. Ogburn mentioned, 91.8% of our patients were employed in u p p e r level trade or professional occupations. It is possible that the type of e m p l o y m e n t or the fact that they were employed affected the rate of p r e t e r m delivery. We do not think that the m a m m a r y stimulation test increased the risk o f p r e t e r m birth. None of the patients in either our study or that of Eden et al. had p r c t e r m labor after the test. In our study only one patient was delivered within 3 weeks o f performing the test; h e r pregnancy was complicated by abruptio placentae. We believe that the m a m m a r y stimulation test can identify a group of patients at increased risk for spontaneous p r e t e r m birth. In spite of the obvious limitations o f risk assessment a n d the lack of proven treatments for those patients at risk for spontaneous preterm birth, the ability to identify high-risk patients is crucial. This knowledge should p e r m i t investigators to i m p l e m e n t clinical trials with significant power to thoroughly evaluate prematurity prevention initiatives.