- Email: [email protected]
A prospective random trial of home uterine activity monitoring in pregnancies at increased risk of preterm labor
Reece et al.
4. Goldstein I, Reece EA, Pilu G, Bovicelli L, Hobbins JC. Cerebellar measurements with ultrasonography in the evaluation of fetal growth and development. AMJ 0BSTET GYMCOL I987;I56:I065-9. 5. Usher R, McClean F. Intrauterine growth of live born Caucasian infants at sea level: standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr I969;74:90 I. 6. Battaglia FC, Lubchenco LO. A practical classification of newborn infants by weight and gestational age. J Pediatr I967;7I:I59. 7. Behrman RE, Lees MW, Peterson ED, Lannoy CS, Seeds AE. Distribution of the circulation in the normal and as phyxiated fetal primate. AM j 0BSTET GYNECO!. I970; I08:956. 8. Campbell AGM, Dawes GS, Fishman AP, eta!. Regional distribution of blood flow in mature fetal lamb. Circ Res I967;2I:229. 9. Rudolph AM, Haymann MA. The circulation of the fetus in utero: methods for studying distribution of blood flow, cardiac output and organ blood flow. Circ Res I967; 2I: I63.
September 1987 Am ] Obstet Gynecol
IO. Lockwood C, Weiner S. Assessment of fetal growth. Clin Perinatal I986;I3:335. II. Giezsson RT, Persson PH. Diagnosis of intrauterine growth retardation using ultrasound. Clin Obstet Gynae col I982; II :457. I2. Crane JP, Kopla MM. Prediction of intrauterine growth retardation via ultrasonically measured head/abdominal circumference ratios. Obstet Gynecol I979;54:597. 13. Vintzileos AM, Lodeiro JG, Feinstein SJ, Campbell WA, Weinbaum PJ, Nochimson DJ. Value of fetal ponderal index in predicting growth retardation. Obstet Gynecol I986;67:584. I4. Gohari P, Berkowitz RL, Hobbins JC. Prediction of intra uterine growth retardation by determination of intra uterine volume. AM J 0BSTET Gvr.;ECOL I977; I27:255. 15. Philipson EH, Sokol RJ, Williams T. Oligohydramnios: clinical associations and predictive value for intrauterine growth retardation. AMJ 0BSTET GYNECOL I983; I46:271. I6. Jeanty P, Romero R. Obstetrical ultrasound. New York: McGraw-Hill, I984.
A prospective random trial of home uterine activity monitoring in pregnancies at increased risk of preterm labor Jay D. lams, M.D., Francee F. Johnson, R.N., B.S.N., Richard W. O'Shaughnessy, M.D., and Lou C. West, R.N., B.S.N. Columbus, Ohio In a prospective trial we enrolled 157 women at increased risk of preterm birth, randomly assigning women in a ratio of 1 :2 to receive either frequent (;;.5 days/wk) nursing contact, education in preterm labor symptoms, and self-palpation of uterine activity (group E, n = 50), or daily nursing contact, preterm labor education, and the Term Guard home uterine activity monitor (group EM, n = 107). Comparison of the rate of preterm birth, the incidence of preterm labor and successful tocolysis, and the mean birth weight and gestational age revealed no significant differences and suggested that beneficial effects previously attributed to monitored contraction data may in fact be the result of frequent (five or more times per week) nursing contact and careful attention to preterm labor symptoms and perceived contractions. (AM J OesTET GYNECOL 1987;157:638-43.)
Key words: Term Guard uterine activity monitor, preterm birth, preterm labor Early detection of increased uterine activity may lead to more effective tocolysis in women at risk of preterm labor. Self-palpation of uterine activity as part of an educational program for women at risk of preterm la bor has been reported by Herron et al.' to increase the percentage of patients with preterm labor who are canFrom the Department of Obstetrics and Gynecology, The Ohio State University. Supported by a grant from the Tokos Medical Corporation and by March of Dimes Birth Defects Foundation Grant No. 2-187/C 185. Presented at the Seventh Annual Meeting of The Society of Perinatal Obstetricians, Lake Buena Vista, Florida, February 5-7, 1987. Reprints not available.
638
didates for tocolysis, thereby reducing the rate of pre term birth. Newman et al? reported, however, that self detection of uterine contractions correlated poorly with tocodynamometer recording of uterine activity and suggested that self-detection may not be a reliable early method of detection of preterm labor. Ambulatory to codynamometry, with the Term Guard uterine activity monitor, has been reported by Katz et al. 3· 5 in a series of articles to be a safe and reliable method of uterine activity assessment that ha~ shown promise in the early identification of women in preterm labor. Because women using such a monitor are necessarily in daily contact with a nurse whose advice and support may lead to earlier identification of preterm labor, we
Home uterine activity monitoring
Volume 157 Number 3
designed a prospective random trial comparing preg nancy complications and outcomes in women at in creased risk of premature delivery. One group used the Term Guard uterine activity monitor (Tokos Medical Corporation, Santa Ana, California) and a control group received instruction in self-palpation and a sim ilar level of nursing education and support.
639
Table I. Symptoms of preterm labor Uterine contractions ;;.4/hr, with or without pain Low backache Increased vaginal discharge Menstrual-like cramps Pelvic pressure Abdominal cramping Spotting
Material and methods
One hundred fifty-seven pregnant women at m creased risk of preterm labor were recruited from area physicians (n = 136) and The Ohio State University prenatal clinic (n = 21). In all, there were 42 physicians or practice groups who contributed subjects to the study. None of the subjects had had preterm labor in the current pregnancy before enrollment, and all were between 20 and 34 weeks' gestation at entry.. In formed consent was obtained, and subjects were then randomized in a ratio of 1:2 to one of two groups, E or EM, according to a protocol approved by The Ohio State University Human Subjects Review Committee. GroupE patients received an initial educational session, usually lasting 60 to 75 minutes, about the symptoms and signs of preterm labor and were taught self palpation of uterine activity. They were asked to note the presence of symptoms of preterm labor, listed in Table I, and to palpate uterine activity for 1 hour twice daily. Each group E subject was contacted by a nurse at the Columbus Term Gu;.rd Monitoring Center daily on weekdays and as needed on weekends. Women in group EM were similarly educated at entry and in structed as well in the use of the Term Guard home uterine activity monitor. They were asked to note symp toms and to record contractions for 1 hour twice daily. Group EM patients were contacted by telephone daily including weekends to transmit and interpret monitor strips and symptoms. Women in both groups were in structed to call the monitoring center if symptoms and/or increased contractions (more than four per hour) were noted and persisted for 1 hour despite lat eral bed rest and oral hydration. Women with persistent symptoms and/or contractions in excess of their estab lished baseline contraction frequency were sent to the doctor's office or to the labor and delivery suite for a cervical examination, unless the doctor specifically or dered otherwise, for example, requested instead an ad ditional hour of home monitoring and hydration or, for those patients receiving oral tocolysis, an additional dose of medication. Participating physicians were asked to follow broad standard guidelines of care for preterm labor, using intravenous ritodrine or magnesium sulfate or subcu taneous terbutaline in standard doses when preterm labor was diagnosed. The definition of preterm labor used in the March of Dimes Multicenter Preterm Birth
Prevention Project was suggested, with parenteral to colysis recommended for persistent uterine contrac tions numbering eight or more per hour if the ce~vix was already 2 em dilated or 50% effaced but requiring cervical change if the cervix was <2 em dilated and less than 50% effaced. Prophylactic tocolysis was discour aged unless the patient had a history of preterm deliv ery. For purposes of data analysis, preterm labor was defined as admission to the hospital and administration of either an intravenous tocolytic or multiple doses of a subcutaneous wcolytic. Once parenteral tocolysis was begun, oral tocolytics were continued in all cases until ;;.35 weeks' gestation. Participating physicians were vis ited at least once during the study by a Tokos Medical Corporation representative and/or a project nurse who reinforced the study protocols. Medical and nursing in service educational programs were provided to the larger referring hospitals in the area. All nurses pro viding daily contact and monitor interpretation were employed at the Columbus Term Guard Monitoring Center and received a standard orientation and edu cation program provided by the Tokos Medical Cor poration. A system of "primary nursing" was adopted shortly after the study started, wherein each patient in both groups had one nurse who served as her principal contact. All nurses had subjects in both groups. A nurse was available to all patients in both groups 24 hours a day, 7 days a week. The monitoring center staff was aware of the crude preterm birth rates for both groups as the study progressed. Each birth before 37 completed weeks of gestation was reviewed by the principal investigator with the at tending physician and the Term Guard nursing staff to confirm gestational age and compliance with pro tocol guidelines. Results
One hundred fifty-seven women had been delivered of their infants as of mid-November, 1986 (107 in group EM and 50 in group E). There were nine women (8.4%) in group EM and six women (12%) in groupE who dropped out of the study. This difference is not significant. The most common reason for dropout in both groups was that the patient and/or the doctor found daily telephone contact or contraction data to be
640
lams et al.
September 1987 Am J Obstet Gynecol
Table II. Demographic description of study subjects GroupE
Group EM
n
No. Parity (mean ± SD) Previous preterm delivery Twins Second-trimester loss Cerclage Uterine anomaly or diethylstilbestrol Bleeding after 14 wk
44 27 15 5 3 3
n
% 1.57 ± 1.2
I
%
98 61.4 34. I Il.4 6.8 6.8 2.3
60 25 I7 II 9 3
1.47 ± l.l
61.2 25.2 I7.3 11.2 9.2 3.I
Table III. Outcome of pregnancy GroupE
%
n
n Delivery < 35 wk <37 wk Gestational age* (mean ± SD) Birth weight (mean ± SD)
Group EM
44 9 23
I
%
ri
98
35.8 ± 2.7 27I8 ± 745
20.4 52 .3
23 42
35.7 ± 2.8 2722 ± 738
23.5 42.9
*All births ;.37 weeks were averaged as 37 weeks to avoid skew caused by postdate pregnancies.
annoying or confusing. None of these 15 women was delivered within 7 days of discontinuing the protocol, and their inclusion would not affect the results of the study. There were 98 patients in group EM and 44 in group E who continued to use the service until delivery or until37 weeks' gestation; these 142 women constitute the population used to obtain the results. Risk factors for preterm delivery and other demo graphic variables were similar in each group, as dis played in Table II. Some women had more than one risk factor. Previous preterril. birth and twins were the most common reasons for enrollment in both groups . There were no subjects with triplet gestations in either group. The incidence of prematurity, defined at 35 and 37 weeks, and the mean gestational age and birth weight are shown in Table III. Thirty-five weeks was selected for analysis because all participating physicians were aggressive in the use of parenteral tocolytics and were fully compliant with protocol guidelines through 35 completed weeks. Some did allow preterm labor after 35 weeks to progress without treatment in several in stances, evenly distributed between groups E and EM . The differences in preterm birth rates are not statis tically significant; analysis at 34 and 36 weeks also re veals no significant difference. Births before 37 completed weeks were extensively reviewed and categorized, as displayed in Table IV, according to the diagnostic category that directly led to
the preterm delivery : preterm labor, preterm prema turely ruptured membranes, or a medical or obstetric indication unrelated to preterm labor or preterm pre mature rupture of the membranes. Those preterm births that followed preterm labor were further cate gorized as (I) following a failed tocolytic attempt, (2) allowed by a physician because of a gestational age >35 completed weeks, or (3) occurring because of advanced cervical dilation >4 em on arrival. None of the differ ences in Table IV is significant. In no case was preterm labor before 35 weeks allowed to progress without treat ment unless the cervix was already > 4 em dilated on arrival. The incidence of preterm labor was 36% in groupE and 39% in group EM, as shown in Table V. The num ber of episodes of preterm labor eligible for treatment, the overall use of prophylactic and therapeutic toco lytics, and the gestational age at onset of preterm labor are also not statistically different. Weeks gained from study entry until delivery or until 37 weeks averaged 7.35 ± 4.2 for the 40 patients in group E and 9.0 ± 4.4 for the 94 in group EM who did not have an indicated preterm birth. Considering only those women in each group who developed pre term labor, the number of days gained from first ap pearance of preterm labor until delivery or until 37 weeks' gestation was 18.6 ± 21.0 in group E and 10.9 ± 12.9 in group EM. These differences are not significant.
Home uterine activity monitoring
Volume 157 Number 3
641
Table IV. Analysis of preterm births (<37 weeks' gestation) GroupE
n No. delivered <37 wk Cause of preterm delivery Preterm labor Failed tocolysis ;;;.35 wk Cervical dilation > 4 em Preterm premature rupture of membranes Indicated
Group EM
n
%
23
%
42
I4 5
60.9
9 0 5 4
21.7 I7.4
64.3
27 IO I5 2 II 4
26.2 9.5
Table V. Preterm labor GroupE n n
Subjects with preterm labor Episodes of preterm labor Treated Not treated 35 wk Cervical dilation >4 em Distress Prophylactic tocolytic used Gestational age at first preterm labor (mean ± SD) Days gained from first preterm labor to delivery (mean± SD)
During the course of the study, several physicians expressed concern that participation would lead to an excessive number of unscheduled office visits for cer vical examination because of the frequency and inten sity of contact with subjects in both groups. We there fore looked also at the frequency of unscheduled visits, defined as an office or labor and delivery suite visit initiated by the Term Guard nurse. There were 60 unscheduled visits in 44 groupE patients (1.36 ± 1.28 per subject) and 146 unscheduled visits in 98 group EM patients (1.49 ± 1.37 per subject). Seventy-three percent of group E and 74% of group EM subjects made at least one unscheduled visit. Sample size is always an issue in clinical studies. Be fore beginning this study, we based estimates of sample size on rates of preterm birth seen in our preterm labor clinic population, defining risk according to the Creasy scoring system. 12 The rate ofpreterm birth (<37 weeks' gestation) was approximately 30%. We postulated a 30% reduction in deliveries before 37 weeks with the use of the Term Guard monitor, which would require 230 subjects in each group to achieve a one-tailed p of 0.05 and a power of 0.8. Using a 2: 1 randomization schema, we planned to enroll 170 women in group E and 340 in group EM to achieve a harmonic mean of
44 I6 23 I3 IO 9
Group EM
% 36.3 56.5
I2
31.5 ± 3.8 18.6 ± 21
98 38 46 28 18 I5
% 38.8 60.9
2
0 I
n
27.3
I 26
32.I ± 3.3 I0.9 ± 12.9
26.5
230. However, entry criteria for this study proved to be substantially more stringent than those of the Creasy system, yielding a study population with a higher in herent risk of preterm birth. A revised power analysis at this point in the study reveals that a harmonic mean of 90 to 100 subjects (E = 75, EM = 150) will be nec essary to detect a drop in the preterm birth rate from 50% to 30%. A sufficient number of undelivered sub jects have been enrolled to reach that goal. Analysis of the data reported here by the method reported by Young et a!." reveals that the sample size is sufficient to have detected a 50% fall in the preterm delivery rate and sufficient as well to have detected a 10-day gain from the first episode of preterm labor with a power of 0.8 and a p of 0.05.
Comment The physiologic basis of contraction monitoring rests on a series of reports from multiple investigators cor relating increased uterine contraction frequency dur ing pregnancy and risk of preterm birth.4,8-IO Although normal and increased levels of uterine activity at each gestational age remain to be established, it appears from these reports that women destined to be delivered before term do, in fact, have more frequent contractile
642
lams et al.
activity throughout the latter half of pregnancy when compared with women who are delivered at term. The next question is how best to use this association to prevent unnecessary preterm births. Main eta!." ap pear to have been successful in using the association as a screening tool to identify a subset of their population at increased risk of preterm delivery, something which had proved to be difficult with more traditional his torical and risk scoring approaches in their population of indigent black women. 11 Herron et a!. 1 taught self-assessment of uterine ac tivity to women who had been already identified by a risk score as being at high risk for preterm labor. They demonstrated a substantial drop in the preterm deliv ery rate, and their methods were soon widely imitated with variable results. 1'· 12 The key features of preterm birth prevention programs patterned after the report by Herron et a!. are identification of women at in creased risk of preterm labor and intensive patient and staff education about the symptoms and signs of pre term labor. In this approach, the patient is a key par ticipant in the early identification of preterm labor. The patient must recognize the signs and/or symptoms of preterm labor and initiate contact with the health care system.· The report of Newman et a!." that high-risk women identified only 15% of their recorded <;ontrac tions cast doubt on the reliability of a patient-based "alarm system" for the early identification of preterm labor. The development of technology capable of re liably recording and transmitting uterine activity data iri.a timely fashion seemed a good way to avoid concerns about patients' perceptions of uterine activity, and the reports by Katz et aJ.'·' were promising. Perhaps over looked in the development and deployment of this ap proach is the increased frequency of patient contact with the health care system, in this case the Term Guard nursing staff. The study design reported here allows comparison of clinical outcomes in two groups of high risk women who all received frequent nursing contact. The groups differed primarily in the manner in which the uterine activity was assessed. The principal end point for this study is the number of women reaching 35 and 37 weeks at delivery. We believe this is the most appropriate outcome for analysis since it represents the answer to the first question posed by the patient and her doctor when considering use of such a system: Will this approach prevent the birth of a premature infant? Other end points such as eligibility for tocolysis may be potentially misleading: If the con traction detection system does detect contractions and if contractions are the basis for the diagnosis, then of course the system is always 100% "successful." The ad dition of cervical examination findings is helpful, but they may also be highly subjective and interpreted dif ferently according to the patient's previous obstetric
September 1987 Am J Obstet Gynecol
history, when used to determine the initiation of to colysis. All studies of pretenn labor are affected by these uncertainties, so it seems best to place greatest emphasis on the least subjective end points such as ges tational age and birth weight. The similarity of all outcomes in groups E and EM is striking: The incidence of preterm labor, preterm delivery, and successful tocolysis is similar. Women us ing the monitor were not sent in for examination or treated prophylactically more often than those using self-palpation. Physicians who enrolled their patients in the study often forgot which study group the patient was in, suggesting that they perceived similarly the care received by both groups. The nursing staff seemed to move from a primarily contraction frequency-based approach in the early weeks of the study to a more global, symptoms-plus-contraction frequency-based approach as their experience grew. The similarity in outcome for groups E and EM indicates that advantages previously attributed to the monitored contraction data rnay in fact be the result of frequent nursing contact and careful attention to symptoms and perceived con tractions. This in turn has implications for the prenatal care of women at increased risk of spontaneous pre term delivery: They should be educated about both preterm labor symptoms and contraction frequency as early markers of preterm labor and should be given ample opportunity to relate symptoms to a welcoming and supportive health care provider. We did not si multaneously enroll a third group of women who re ceived a standard level of nursing and physician contact and therefore cannot draw any conclusions about whether frequent contact, with or without a monitor, is superior to current practice in reducing the incidence of preterm birth in high-risk patients. Such a study is now underway. We doubt, however, that a system based principally on contraction detection without equal at tention to symptoms will be of use in reducing preterm births. Another question unanswered by this study is the identification of specific subsets of the population at risk for preterm labor who might benefit most or kast from tocodynamometer recording of uterine activity. Newman et al. 13 have reported increased levels of qter ine activity among women with twins, which might make them more or less susceptible to any benefit of home tocodynamometry. Similar questions remain un answered for women with uterine anomalies, incom petent cervices, and vaginal bleeding. Katz et al. 11 have also reported success with home monitoring in the man agement of oral tocolysis, with contraction frequency used as a guide to dosage adjustment, in 60 patients with preterm labor who used the Term Guard com pared with 60 non-randomly selected matched control subjects. The study reported here specifically excluded
Home uterine activity monitoring
Volume 157 Number 3
women treated for preterm labor before entry, but a similar prospective randomized trial of home monitor ing in these patients is also underway. In summary, this prospective random trial of inten sive patient education and nursing support with and without a home uterine contraction monitor indicates that increased attention should be paid to the role of frequent and supportive patient contact in preterm birth prevention; the role of ambulatory contraction monitoring has yet to be defined. We wish to express thanks to the physicians of central and southeast Ohio who enrolled their patients in this study, and to the Tokos Medical Corporation. REFERENCES l. Herron M, Katz M, Creasy R. Evaluation of a preterm
2. 3. 4. 5. 6.
birth prevention program: preliminary report. Obstet Gy necol 1982;59:452-6. Newman RB, Gill P, Wittreich P, Katz M. Maternal per ception of prelabor uterine activity. Obstet Gynecol 1986;68:765-9. Katz M, Gill P. Initial evaluation of an ambulatory system for home monitoring and transmission of uterine activity data. Obstet Gynecol 1985;66:273-7. Katz M, Newman RB , Gill P. Assessment of uterine activity in ambulatory patients at high risk of preterm labor and delivery. AMJ 0BSTET GYNECOL 1986;154:44-7. Katz M, Gill P, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity: a prelimi nary report. Obstet Gynecol 1986;68:773-8. Main DM, Gabbe SG, Richardson D, Strong S. Can pre-
7. 8. 9.
10.
11.
12.
13.
14.
643
term deliveries be prevented? AM J OBSTET GYNECOL 1985 ;151:892-8. Young MJ, Bresnitz EA, Strom BL. Sample size nomo grams for interpreting negative clinical studies. Ann In tern Med 1983;99:248-51. Aubry R, Pennington ]. Identification and evaluation of high risk pregnancy: the perinatal concept. Clin Obstet Gynecol 1973;16:3-27. Main D, Katz M, Chin G, et al. Intermittent weekly con traction monitoring to predict preterm labor in low risk women: a blinded study [Abstract No. 85]. In: Proceed ings of the seventh annual meeting of The Society of Perinatal Obstetricians, February 5-7, 1987, Lake Buena Vista, Florida. Nageotte MP, Dorchester W, Porto M, et al. Quantitation of uterine activity preceding preterm, term and post term labor [Abstract No. 129]. In: Proceedings of the seventh annual meeting of The Society of Perinatal Obstetricians, February 5-7, 1987, Lake Buena Vista, Florida. Main D, Richardson D, Gabbe S, et al. Prospective eval uation ,of a risk scoring system for predicting preterm delivery in black inner city women. Obstet Gynecol 1987 ;69:61-6. lams J, Johnson F, Creasy R. Preliminary report of a pro spective trial of a preterm birth prevention program [Ab stract No. 81]. In: Proceedings of the fifth annual meeting of The Society of Perinatal Obstetricians, January 31 February 2, 1985, Las Vegas, Nevada. Newman RB, Gill PJ , Katz M. Uterine activity during pregnancy in ambulatory patients: comparison of single ton and twin gestation. AM J 0BSTET GYNECOL 1986; 154:530-l. Katz M, Gill PJ, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity for the man agement of oral tocolysis. AM J 0BSTET GY1\ECOL 1986; 154:1253-6.
4. Goldstein I, Reece EA, Pilu G, Bovicelli L, Hobbins JC. Cerebellar measurements with ultrasonography in the evaluation of fetal growth and development. AMJ 0BSTET GYMCOL I987;I56:I065-9. 5. Usher R, McClean F. Intrauterine growth of live born Caucasian infants at sea level: standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr I969;74:90 I. 6. Battaglia FC, Lubchenco LO. A practical classification of newborn infants by weight and gestational age. J Pediatr I967;7I:I59. 7. Behrman RE, Lees MW, Peterson ED, Lannoy CS, Seeds AE. Distribution of the circulation in the normal and as phyxiated fetal primate. AM j 0BSTET GYNECO!. I970; I08:956. 8. Campbell AGM, Dawes GS, Fishman AP, eta!. Regional distribution of blood flow in mature fetal lamb. Circ Res I967;2I:229. 9. Rudolph AM, Haymann MA. The circulation of the fetus in utero: methods for studying distribution of blood flow, cardiac output and organ blood flow. Circ Res I967; 2I: I63.
September 1987 Am ] Obstet Gynecol
IO. Lockwood C, Weiner S. Assessment of fetal growth. Clin Perinatal I986;I3:335. II. Giezsson RT, Persson PH. Diagnosis of intrauterine growth retardation using ultrasound. Clin Obstet Gynae col I982; II :457. I2. Crane JP, Kopla MM. Prediction of intrauterine growth retardation via ultrasonically measured head/abdominal circumference ratios. Obstet Gynecol I979;54:597. 13. Vintzileos AM, Lodeiro JG, Feinstein SJ, Campbell WA, Weinbaum PJ, Nochimson DJ. Value of fetal ponderal index in predicting growth retardation. Obstet Gynecol I986;67:584. I4. Gohari P, Berkowitz RL, Hobbins JC. Prediction of intra uterine growth retardation by determination of intra uterine volume. AM J 0BSTET Gvr.;ECOL I977; I27:255. 15. Philipson EH, Sokol RJ, Williams T. Oligohydramnios: clinical associations and predictive value for intrauterine growth retardation. AMJ 0BSTET GYNECOL I983; I46:271. I6. Jeanty P, Romero R. Obstetrical ultrasound. New York: McGraw-Hill, I984.
A prospective random trial of home uterine activity monitoring in pregnancies at increased risk of preterm labor Jay D. lams, M.D., Francee F. Johnson, R.N., B.S.N., Richard W. O'Shaughnessy, M.D., and Lou C. West, R.N., B.S.N. Columbus, Ohio In a prospective trial we enrolled 157 women at increased risk of preterm birth, randomly assigning women in a ratio of 1 :2 to receive either frequent (;;.5 days/wk) nursing contact, education in preterm labor symptoms, and self-palpation of uterine activity (group E, n = 50), or daily nursing contact, preterm labor education, and the Term Guard home uterine activity monitor (group EM, n = 107). Comparison of the rate of preterm birth, the incidence of preterm labor and successful tocolysis, and the mean birth weight and gestational age revealed no significant differences and suggested that beneficial effects previously attributed to monitored contraction data may in fact be the result of frequent (five or more times per week) nursing contact and careful attention to preterm labor symptoms and perceived contractions. (AM J OesTET GYNECOL 1987;157:638-43.)
Key words: Term Guard uterine activity monitor, preterm birth, preterm labor Early detection of increased uterine activity may lead to more effective tocolysis in women at risk of preterm labor. Self-palpation of uterine activity as part of an educational program for women at risk of preterm la bor has been reported by Herron et al.' to increase the percentage of patients with preterm labor who are canFrom the Department of Obstetrics and Gynecology, The Ohio State University. Supported by a grant from the Tokos Medical Corporation and by March of Dimes Birth Defects Foundation Grant No. 2-187/C 185. Presented at the Seventh Annual Meeting of The Society of Perinatal Obstetricians, Lake Buena Vista, Florida, February 5-7, 1987. Reprints not available.
638
didates for tocolysis, thereby reducing the rate of pre term birth. Newman et al? reported, however, that self detection of uterine contractions correlated poorly with tocodynamometer recording of uterine activity and suggested that self-detection may not be a reliable early method of detection of preterm labor. Ambulatory to codynamometry, with the Term Guard uterine activity monitor, has been reported by Katz et al. 3· 5 in a series of articles to be a safe and reliable method of uterine activity assessment that ha~ shown promise in the early identification of women in preterm labor. Because women using such a monitor are necessarily in daily contact with a nurse whose advice and support may lead to earlier identification of preterm labor, we
Home uterine activity monitoring
Volume 157 Number 3
designed a prospective random trial comparing preg nancy complications and outcomes in women at in creased risk of premature delivery. One group used the Term Guard uterine activity monitor (Tokos Medical Corporation, Santa Ana, California) and a control group received instruction in self-palpation and a sim ilar level of nursing education and support.
639
Table I. Symptoms of preterm labor Uterine contractions ;;.4/hr, with or without pain Low backache Increased vaginal discharge Menstrual-like cramps Pelvic pressure Abdominal cramping Spotting
Material and methods
One hundred fifty-seven pregnant women at m creased risk of preterm labor were recruited from area physicians (n = 136) and The Ohio State University prenatal clinic (n = 21). In all, there were 42 physicians or practice groups who contributed subjects to the study. None of the subjects had had preterm labor in the current pregnancy before enrollment, and all were between 20 and 34 weeks' gestation at entry.. In formed consent was obtained, and subjects were then randomized in a ratio of 1:2 to one of two groups, E or EM, according to a protocol approved by The Ohio State University Human Subjects Review Committee. GroupE patients received an initial educational session, usually lasting 60 to 75 minutes, about the symptoms and signs of preterm labor and were taught self palpation of uterine activity. They were asked to note the presence of symptoms of preterm labor, listed in Table I, and to palpate uterine activity for 1 hour twice daily. Each group E subject was contacted by a nurse at the Columbus Term Gu;.rd Monitoring Center daily on weekdays and as needed on weekends. Women in group EM were similarly educated at entry and in structed as well in the use of the Term Guard home uterine activity monitor. They were asked to note symp toms and to record contractions for 1 hour twice daily. Group EM patients were contacted by telephone daily including weekends to transmit and interpret monitor strips and symptoms. Women in both groups were in structed to call the monitoring center if symptoms and/or increased contractions (more than four per hour) were noted and persisted for 1 hour despite lat eral bed rest and oral hydration. Women with persistent symptoms and/or contractions in excess of their estab lished baseline contraction frequency were sent to the doctor's office or to the labor and delivery suite for a cervical examination, unless the doctor specifically or dered otherwise, for example, requested instead an ad ditional hour of home monitoring and hydration or, for those patients receiving oral tocolysis, an additional dose of medication. Participating physicians were asked to follow broad standard guidelines of care for preterm labor, using intravenous ritodrine or magnesium sulfate or subcu taneous terbutaline in standard doses when preterm labor was diagnosed. The definition of preterm labor used in the March of Dimes Multicenter Preterm Birth
Prevention Project was suggested, with parenteral to colysis recommended for persistent uterine contrac tions numbering eight or more per hour if the ce~vix was already 2 em dilated or 50% effaced but requiring cervical change if the cervix was <2 em dilated and less than 50% effaced. Prophylactic tocolysis was discour aged unless the patient had a history of preterm deliv ery. For purposes of data analysis, preterm labor was defined as admission to the hospital and administration of either an intravenous tocolytic or multiple doses of a subcutaneous wcolytic. Once parenteral tocolysis was begun, oral tocolytics were continued in all cases until ;;.35 weeks' gestation. Participating physicians were vis ited at least once during the study by a Tokos Medical Corporation representative and/or a project nurse who reinforced the study protocols. Medical and nursing in service educational programs were provided to the larger referring hospitals in the area. All nurses pro viding daily contact and monitor interpretation were employed at the Columbus Term Guard Monitoring Center and received a standard orientation and edu cation program provided by the Tokos Medical Cor poration. A system of "primary nursing" was adopted shortly after the study started, wherein each patient in both groups had one nurse who served as her principal contact. All nurses had subjects in both groups. A nurse was available to all patients in both groups 24 hours a day, 7 days a week. The monitoring center staff was aware of the crude preterm birth rates for both groups as the study progressed. Each birth before 37 completed weeks of gestation was reviewed by the principal investigator with the at tending physician and the Term Guard nursing staff to confirm gestational age and compliance with pro tocol guidelines. Results
One hundred fifty-seven women had been delivered of their infants as of mid-November, 1986 (107 in group EM and 50 in group E). There were nine women (8.4%) in group EM and six women (12%) in groupE who dropped out of the study. This difference is not significant. The most common reason for dropout in both groups was that the patient and/or the doctor found daily telephone contact or contraction data to be
640
lams et al.
September 1987 Am J Obstet Gynecol
Table II. Demographic description of study subjects GroupE
Group EM
n
No. Parity (mean ± SD) Previous preterm delivery Twins Second-trimester loss Cerclage Uterine anomaly or diethylstilbestrol Bleeding after 14 wk
44 27 15 5 3 3
n
% 1.57 ± 1.2
I
%
98 61.4 34. I Il.4 6.8 6.8 2.3
60 25 I7 II 9 3
1.47 ± l.l
61.2 25.2 I7.3 11.2 9.2 3.I
Table III. Outcome of pregnancy GroupE
%
n
n Delivery < 35 wk <37 wk Gestational age* (mean ± SD) Birth weight (mean ± SD)
Group EM
44 9 23
I
%
ri
98
35.8 ± 2.7 27I8 ± 745
20.4 52 .3
23 42
35.7 ± 2.8 2722 ± 738
23.5 42.9
*All births ;.37 weeks were averaged as 37 weeks to avoid skew caused by postdate pregnancies.
annoying or confusing. None of these 15 women was delivered within 7 days of discontinuing the protocol, and their inclusion would not affect the results of the study. There were 98 patients in group EM and 44 in group E who continued to use the service until delivery or until37 weeks' gestation; these 142 women constitute the population used to obtain the results. Risk factors for preterm delivery and other demo graphic variables were similar in each group, as dis played in Table II. Some women had more than one risk factor. Previous preterril. birth and twins were the most common reasons for enrollment in both groups . There were no subjects with triplet gestations in either group. The incidence of prematurity, defined at 35 and 37 weeks, and the mean gestational age and birth weight are shown in Table III. Thirty-five weeks was selected for analysis because all participating physicians were aggressive in the use of parenteral tocolytics and were fully compliant with protocol guidelines through 35 completed weeks. Some did allow preterm labor after 35 weeks to progress without treatment in several in stances, evenly distributed between groups E and EM . The differences in preterm birth rates are not statis tically significant; analysis at 34 and 36 weeks also re veals no significant difference. Births before 37 completed weeks were extensively reviewed and categorized, as displayed in Table IV, according to the diagnostic category that directly led to
the preterm delivery : preterm labor, preterm prema turely ruptured membranes, or a medical or obstetric indication unrelated to preterm labor or preterm pre mature rupture of the membranes. Those preterm births that followed preterm labor were further cate gorized as (I) following a failed tocolytic attempt, (2) allowed by a physician because of a gestational age >35 completed weeks, or (3) occurring because of advanced cervical dilation >4 em on arrival. None of the differ ences in Table IV is significant. In no case was preterm labor before 35 weeks allowed to progress without treat ment unless the cervix was already > 4 em dilated on arrival. The incidence of preterm labor was 36% in groupE and 39% in group EM, as shown in Table V. The num ber of episodes of preterm labor eligible for treatment, the overall use of prophylactic and therapeutic toco lytics, and the gestational age at onset of preterm labor are also not statistically different. Weeks gained from study entry until delivery or until 37 weeks averaged 7.35 ± 4.2 for the 40 patients in group E and 9.0 ± 4.4 for the 94 in group EM who did not have an indicated preterm birth. Considering only those women in each group who developed pre term labor, the number of days gained from first ap pearance of preterm labor until delivery or until 37 weeks' gestation was 18.6 ± 21.0 in group E and 10.9 ± 12.9 in group EM. These differences are not significant.
Home uterine activity monitoring
Volume 157 Number 3
641
Table IV. Analysis of preterm births (<37 weeks' gestation) GroupE
n No. delivered <37 wk Cause of preterm delivery Preterm labor Failed tocolysis ;;;.35 wk Cervical dilation > 4 em Preterm premature rupture of membranes Indicated
Group EM
n
%
23
%
42
I4 5
60.9
9 0 5 4
21.7 I7.4
64.3
27 IO I5 2 II 4
26.2 9.5
Table V. Preterm labor GroupE n n
Subjects with preterm labor Episodes of preterm labor Treated Not treated 35 wk Cervical dilation >4 em Distress Prophylactic tocolytic used Gestational age at first preterm labor (mean ± SD) Days gained from first preterm labor to delivery (mean± SD)
During the course of the study, several physicians expressed concern that participation would lead to an excessive number of unscheduled office visits for cer vical examination because of the frequency and inten sity of contact with subjects in both groups. We there fore looked also at the frequency of unscheduled visits, defined as an office or labor and delivery suite visit initiated by the Term Guard nurse. There were 60 unscheduled visits in 44 groupE patients (1.36 ± 1.28 per subject) and 146 unscheduled visits in 98 group EM patients (1.49 ± 1.37 per subject). Seventy-three percent of group E and 74% of group EM subjects made at least one unscheduled visit. Sample size is always an issue in clinical studies. Be fore beginning this study, we based estimates of sample size on rates of preterm birth seen in our preterm labor clinic population, defining risk according to the Creasy scoring system. 12 The rate ofpreterm birth (<37 weeks' gestation) was approximately 30%. We postulated a 30% reduction in deliveries before 37 weeks with the use of the Term Guard monitor, which would require 230 subjects in each group to achieve a one-tailed p of 0.05 and a power of 0.8. Using a 2: 1 randomization schema, we planned to enroll 170 women in group E and 340 in group EM to achieve a harmonic mean of
44 I6 23 I3 IO 9
Group EM
% 36.3 56.5
I2
31.5 ± 3.8 18.6 ± 21
98 38 46 28 18 I5
% 38.8 60.9
2
0 I
n
27.3
I 26
32.I ± 3.3 I0.9 ± 12.9
26.5
230. However, entry criteria for this study proved to be substantially more stringent than those of the Creasy system, yielding a study population with a higher in herent risk of preterm birth. A revised power analysis at this point in the study reveals that a harmonic mean of 90 to 100 subjects (E = 75, EM = 150) will be nec essary to detect a drop in the preterm birth rate from 50% to 30%. A sufficient number of undelivered sub jects have been enrolled to reach that goal. Analysis of the data reported here by the method reported by Young et a!." reveals that the sample size is sufficient to have detected a 50% fall in the preterm delivery rate and sufficient as well to have detected a 10-day gain from the first episode of preterm labor with a power of 0.8 and a p of 0.05.
Comment The physiologic basis of contraction monitoring rests on a series of reports from multiple investigators cor relating increased uterine contraction frequency dur ing pregnancy and risk of preterm birth.4,8-IO Although normal and increased levels of uterine activity at each gestational age remain to be established, it appears from these reports that women destined to be delivered before term do, in fact, have more frequent contractile
642
lams et al.
activity throughout the latter half of pregnancy when compared with women who are delivered at term. The next question is how best to use this association to prevent unnecessary preterm births. Main eta!." ap pear to have been successful in using the association as a screening tool to identify a subset of their population at increased risk of preterm delivery, something which had proved to be difficult with more traditional his torical and risk scoring approaches in their population of indigent black women. 11 Herron et a!. 1 taught self-assessment of uterine ac tivity to women who had been already identified by a risk score as being at high risk for preterm labor. They demonstrated a substantial drop in the preterm deliv ery rate, and their methods were soon widely imitated with variable results. 1'· 12 The key features of preterm birth prevention programs patterned after the report by Herron et a!. are identification of women at in creased risk of preterm labor and intensive patient and staff education about the symptoms and signs of pre term labor. In this approach, the patient is a key par ticipant in the early identification of preterm labor. The patient must recognize the signs and/or symptoms of preterm labor and initiate contact with the health care system.· The report of Newman et a!." that high-risk women identified only 15% of their recorded <;ontrac tions cast doubt on the reliability of a patient-based "alarm system" for the early identification of preterm labor. The development of technology capable of re liably recording and transmitting uterine activity data iri.a timely fashion seemed a good way to avoid concerns about patients' perceptions of uterine activity, and the reports by Katz et aJ.'·' were promising. Perhaps over looked in the development and deployment of this ap proach is the increased frequency of patient contact with the health care system, in this case the Term Guard nursing staff. The study design reported here allows comparison of clinical outcomes in two groups of high risk women who all received frequent nursing contact. The groups differed primarily in the manner in which the uterine activity was assessed. The principal end point for this study is the number of women reaching 35 and 37 weeks at delivery. We believe this is the most appropriate outcome for analysis since it represents the answer to the first question posed by the patient and her doctor when considering use of such a system: Will this approach prevent the birth of a premature infant? Other end points such as eligibility for tocolysis may be potentially misleading: If the con traction detection system does detect contractions and if contractions are the basis for the diagnosis, then of course the system is always 100% "successful." The ad dition of cervical examination findings is helpful, but they may also be highly subjective and interpreted dif ferently according to the patient's previous obstetric
September 1987 Am J Obstet Gynecol
history, when used to determine the initiation of to colysis. All studies of pretenn labor are affected by these uncertainties, so it seems best to place greatest emphasis on the least subjective end points such as ges tational age and birth weight. The similarity of all outcomes in groups E and EM is striking: The incidence of preterm labor, preterm delivery, and successful tocolysis is similar. Women us ing the monitor were not sent in for examination or treated prophylactically more often than those using self-palpation. Physicians who enrolled their patients in the study often forgot which study group the patient was in, suggesting that they perceived similarly the care received by both groups. The nursing staff seemed to move from a primarily contraction frequency-based approach in the early weeks of the study to a more global, symptoms-plus-contraction frequency-based approach as their experience grew. The similarity in outcome for groups E and EM indicates that advantages previously attributed to the monitored contraction data rnay in fact be the result of frequent nursing contact and careful attention to symptoms and perceived con tractions. This in turn has implications for the prenatal care of women at increased risk of spontaneous pre term delivery: They should be educated about both preterm labor symptoms and contraction frequency as early markers of preterm labor and should be given ample opportunity to relate symptoms to a welcoming and supportive health care provider. We did not si multaneously enroll a third group of women who re ceived a standard level of nursing and physician contact and therefore cannot draw any conclusions about whether frequent contact, with or without a monitor, is superior to current practice in reducing the incidence of preterm birth in high-risk patients. Such a study is now underway. We doubt, however, that a system based principally on contraction detection without equal at tention to symptoms will be of use in reducing preterm births. Another question unanswered by this study is the identification of specific subsets of the population at risk for preterm labor who might benefit most or kast from tocodynamometer recording of uterine activity. Newman et al. 13 have reported increased levels of qter ine activity among women with twins, which might make them more or less susceptible to any benefit of home tocodynamometry. Similar questions remain un answered for women with uterine anomalies, incom petent cervices, and vaginal bleeding. Katz et al. 11 have also reported success with home monitoring in the man agement of oral tocolysis, with contraction frequency used as a guide to dosage adjustment, in 60 patients with preterm labor who used the Term Guard com pared with 60 non-randomly selected matched control subjects. The study reported here specifically excluded
Home uterine activity monitoring
Volume 157 Number 3
women treated for preterm labor before entry, but a similar prospective randomized trial of home monitor ing in these patients is also underway. In summary, this prospective random trial of inten sive patient education and nursing support with and without a home uterine contraction monitor indicates that increased attention should be paid to the role of frequent and supportive patient contact in preterm birth prevention; the role of ambulatory contraction monitoring has yet to be defined. We wish to express thanks to the physicians of central and southeast Ohio who enrolled their patients in this study, and to the Tokos Medical Corporation. REFERENCES l. Herron M, Katz M, Creasy R. Evaluation of a preterm
2. 3. 4. 5. 6.
birth prevention program: preliminary report. Obstet Gy necol 1982;59:452-6. Newman RB, Gill P, Wittreich P, Katz M. Maternal per ception of prelabor uterine activity. Obstet Gynecol 1986;68:765-9. Katz M, Gill P. Initial evaluation of an ambulatory system for home monitoring and transmission of uterine activity data. Obstet Gynecol 1985;66:273-7. Katz M, Newman RB , Gill P. Assessment of uterine activity in ambulatory patients at high risk of preterm labor and delivery. AMJ 0BSTET GYNECOL 1986;154:44-7. Katz M, Gill P, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity: a prelimi nary report. Obstet Gynecol 1986;68:773-8. Main DM, Gabbe SG, Richardson D, Strong S. Can pre-
7. 8. 9.
10.
11.
12.
13.
14.
643
term deliveries be prevented? AM J OBSTET GYNECOL 1985 ;151:892-8. Young MJ, Bresnitz EA, Strom BL. Sample size nomo grams for interpreting negative clinical studies. Ann In tern Med 1983;99:248-51. Aubry R, Pennington ]. Identification and evaluation of high risk pregnancy: the perinatal concept. Clin Obstet Gynecol 1973;16:3-27. Main D, Katz M, Chin G, et al. Intermittent weekly con traction monitoring to predict preterm labor in low risk women: a blinded study [Abstract No. 85]. In: Proceed ings of the seventh annual meeting of The Society of Perinatal Obstetricians, February 5-7, 1987, Lake Buena Vista, Florida. Nageotte MP, Dorchester W, Porto M, et al. Quantitation of uterine activity preceding preterm, term and post term labor [Abstract No. 129]. In: Proceedings of the seventh annual meeting of The Society of Perinatal Obstetricians, February 5-7, 1987, Lake Buena Vista, Florida. Main D, Richardson D, Gabbe S, et al. Prospective eval uation ,of a risk scoring system for predicting preterm delivery in black inner city women. Obstet Gynecol 1987 ;69:61-6. lams J, Johnson F, Creasy R. Preliminary report of a pro spective trial of a preterm birth prevention program [Ab stract No. 81]. In: Proceedings of the fifth annual meeting of The Society of Perinatal Obstetricians, January 31 February 2, 1985, Las Vegas, Nevada. Newman RB, Gill PJ , Katz M. Uterine activity during pregnancy in ambulatory patients: comparison of single ton and twin gestation. AM J 0BSTET GYNECOL 1986; 154:530-l. Katz M, Gill PJ, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity for the man agement of oral tocolysis. AM J 0BSTET GY1\ECOL 1986; 154:1253-6.