Of all neonates born alive in the United States in 1993, 7.2% weighed less than 2,500 g (1). This figure represents the highest incidence of low birth weight since 1976. The incidence of very low birth weight (births of less than 1,500 g) has also remained unchanged since 1991 with an incidence of 1.3% (1). Over the past 30 years, the incidence of neonates weighing less than 2,500 g has declined only slightly and the rate of neonates weighing less than 1,500 g has not changed substantially (2, 3).
In 1992, more than 34,000 children in the United States died in their first year of life. These deaths result in an infant mortality rate of 8.5 deaths per 1,000 live births (4). Diseases related to prematurity are a major cause. Although this is the lowest rate ever recorded in the United States, it compares unfavorably with most other Western industrialized countries. A variety of primary and secondary preventive strategies have been suggested to reduce the incidence of preterm delivery. In spite of these efforts, overall reduction in the incidence of preterm births over time has been slight.
Overview
Home uterine activity monitoring (HUAM) is a system of care to detect preterm labor. It uses a combination of the recording of uterine contractions with a tocodynamometer and daily telephone calls from a health care provider to offer patient support and advice. A recording of uterine contractions is transferred by telephone to the health care provider for rapid evaluation. The premise of HUAM is that women will have an identifiable increase in uterine contractions before the onset of preterm labor and that these prodromal uterine contractions otherwise may not be recognized by the patient. Advocates of the system propose that early identification of preterm uterine contractions will permit earlier administration of tocolytic therapy that, in turn, may be more effective than later therapy in preventing preterm births.
The ability to identify women at risk for preterm birth based on their history or risk factors is poor (5). Because many risk-scoring systems depend heavily on a history of preterm birth, the positive predictive value of risk scoring for nulliparous women is even lower than that for multiparous women. Thus, with widespread availability, HUAM may be used for many women who will not have premature labor, and some women who experience preterm labor will not receive HUAM. Widespread use could also lead to inappropriate treatment with tocolytic agents. This treatment can result in not only adverse maternal or fetal side effects but also increased cost of obstetric care.
At least 11 randomized controlled trials examining the efficacy of HUAM have published results (Table 1; print version only). Two of these studies include data from the same patients (6, 7). In addition, there are several prospective cohort studies (8- 10). The studies vary in design, criteria for inclusion of patients, and measurements of endpoints and outcomes. These differences make comparisons difficult. Furthermore, many of these studies have limitations with regard to research design, including sample size (power) or numbers of patients, which preclude reaching conclusions about the utility of HUAM. In an assessment of the work of Hill, Knuppel, and Watson, Keirse noted bias and errors sufficient to warrant dismissing their results (11). Both Grimes and Sachs have critically reviewed the literature regarding HUAM and have cautioned against widespread clinical application (12, 13). The U.S. Preventive Services Task Force performed an independent review and reached a similar conclusion (14).
Thus, the role of HUAM in the prevention of prematurity is controversial. In addition to difficulty in identifying patients who will deliver prematurely, a number of other issues should be considered: 1) accuracy and reliability of the technology, 2) the endpoint used to measure and evaluate effectiveness of HUAM, 3) effectiveness of treatment, 4) safety, and 5) a cost-benefit analysis.
Accuracy and Reliability
Criteria for the objective diagnosis of preterm labor vary among investigators and clinical settings. This variation in turn limits the precision with which preterm labor is detected by HUAM in these studies. Several investigators have attempted to measure the reliability of the home tocodynamometer (15, 16). They have reported only limited degrees of intraobserver and interobserver reliability in reading home tocodynamometry strips.
A possible benefit of HUAM might result from identification of women who will subsequently develop preterm labor. One small observational study reported that the frequency of uterine contractions increased in the 24-hour period before preterm labor developed (17). Most of the patients who developed preterm labor had more than four contractions per hour (17). No population-based studies have determined how many women have episodes of increased uterine contractions and do not go into premature labor.
Endpoints Used to Measure and Evaluate Effectiveness
The impact of HUAM on the prevention of preterm delivery is the important endpoint for determining the clinical utility and benefit of the technology. This impact could be established by analysis of both the entire population of patients in whom the technology is studied and controls. Unfortunately, many of the HUAM studies evaluate intermediate endpoints (eg, cervical dilatation on admission) in the subgroup of patients and controls who experience preterm labor. The results may reach statistical significance, but analyzing only the subgroup of women who experience preterm labor is a methodologic flaw and limits the value of these studies. In some of these same studies, analysis of results in the entire population does not show significant differences in the outcome of prematurity prevention.
Cervical Dilatation on Admission
Evidence supports the hypothesis that use of HUAM in pregnancies at high risk of preterm labor results in less cervical dilatation at the time of admission to the hospital for preterm labor. Morrison found that high-risk women who were monitored and who experienced preterm labor had less cervical dilatation and less effacement at the time of diagnosis of preterm labor than did controls (18). In other studies of twin pregnancies and high-risk pregnancies (19, 20), a higher percentage of monitored women presented with less advanced cervical dilatation than did women who were not monitored with HUAM. Hill's study may have included some or all of Knuppel's patients, thus limiting the conclusions that can be drawn from these findings (19, 20). Dyson was able to demonstrate a benefit from HUAM in terms of patients "favorable for suppression" of preterm labor at the time of their presentation (21), but this difference was only observed for the subgroup of women with twins. In addition, Dyson's analysis of twin data used infants rather than pregnancies as the unit of analysis; thus, Dyson artificially increased the statistical power for the twin result. Mou and Nagey also noted less cervical dilatation in high-risk pregnancies at the time of admission in preterm labor, but the latter study did not reach statistical significance (22, 23). Wapner, in a randomized control trial of HUAM without daily nursing contact in patients with a previous preterm birth, found significantly less cervical dilatation on admission in the patients who had preterm labor (24).
The Food and Drug Administration has approved the device for monitoring uterine activity and accepted cervical dilatation alone as an appropriate endpoint for the approval of this technology. It did not require proof that the device prevents premature birth and the associated neonatal morbidity and mortality, which are the most important outcomes for clinicians and patients (25). Some might argue that cervical dilatation and effacement is a subjective endpoint.
Success of Treatment of Preterm Labor
Investigators have attempted to determine the benefit of HUAM by studying its impact on the success of treatment for preterm labor. The hypothesis is that an improvement in success of treatment might be based on earlier initiation of treatment when cervical dilatation is less or before preterm labor becomes refractory to drug therapy. This approach is also limited to subgroup analysis of patients and controls with a diagnosis of preterm labor, rather than the entire study population.
Data from several of these studies suggest that treatment of preterm labor is more successful with the use of HUAM. Examining the subgroup of patients who experienced preterm labor, Morrison found that the percentage of monitored women in whom tocolysis was not effective was lower than that of women not monitored, although no statistical analysis was provided (18). Hill found a significant benefit of HUAM in preventing delivery within 48 hours in women who presented in preterm labor (19). Knuppel and Hill observed that significantly more controls failed tocolysis and delivered compared with those women who were monitored and experienced preterm labor (19, 20). In Watson's study, all of the women had preterm labor in the index pregnancy (26). In the subgroup of women who experienced recurrent preterm labor, significantly more of the controls delivered despite tocolytic treatment. Dyson also found this benefit in preterm labor with twins, although not with singleton pregnancies (21). As previously noted, the use of infants rather than pregnancies as the unit of analysis artificially increases the statistical power for the twin result. Wapner found a significant prolongation in the duration of pregnancy following treatment of preterm labor in monitored patients, and delivery was delayed by more than 48 hours in a significantly greater number of patients (24). In contrast, Iams examined the subgroup of patients who experienced preterm labor (7). He found no difference between those who were monitored and those who were not with regard to their suitability for tocolysis or the percentage of women who failed treatment.
Prevention of Preterm Delivery
A number of studies using different patient risk factors and different definitions of preterm delivery demonstrate a benefit in reducing preterm delivery but only in monitored women who had preterm labor compared with controls who had preterm labor and were not monitored (19-22, 24, 26). Reanalysis of the data from one study (19) found no significant difference in the incidence of preterm delivery when all enrolled patients are analyzed (13, 26). When the entire study population was analyzed, Wapner's study did not show a statistically significant improvement in pregnancy outcome as measured by gestational age, weight at delivery, and delivery of neonates weighing less than 2,000 g (24).
Only one of the studies that support the use of HUAM in preventing premature delivery does not depend on the analysis of subgroups of patients. In a study comparing HUAM to education and nursing contact, Morrison reported that the monitored women had a lower incidence of delivery before 37 weeks of gestation (18). Morrison also found significantly greater time elapsed from diagnosis of preterm labor to delivery in the monitored group than in controls (8.2-2.7 weeks versus 4.2-2.9 weeks). This study was not blinded and provided limited nursing contact to the control group.
In contrast, several investigators found no difference in the incidence of preterm delivery or gestational age at delivery between monitored and unmonitored women (7, 23, 27). These investigators, who did not show a benefit of HUAM, tended to analyze all women who had been randomized rather than only the subgroup of women who experienced preterm labor.
Other neonatal outcomes have been examined, such as respiratory distress syndrome, neonatal intensive care unit admissions, total length of neonatal intensive care, and total nursery days. The results regarding these other endpoints also are not consistent.
Some investigators have attempted to isolate the benefit of the components of HUAM as a system of care, specifically the health provider contact and the tocodynamometer. Because of limitations and differences in design, consistent conclusions regarding these components cannot be made. The following conclusions can be made, however:
From a clinical point of view, use of an intermediate endpoint, such as cervical dilatation, is not justified when the endpoint--prevention of preterm delivery--can be measured.
The available data do not support the effectiveness of HUAM for the prevention of preterm birth.
Effectiveness of Treatment of Preterm Labor
A system of care that is designed to provide early detection of preterm labor is of benefit only if an effective treatment of preterm labor exists. There is evidence that tocolytic therapy is of benefit, but this may be limited to prolongation of pregnancy for 48 hours following the initiation of treatment. The benefits of corticosteroids for inducing fetal pulmonary maturation have recently been reviewed (28, 29). The studies of HUAM have not been designed to determine whether benefit is derived from early detection of preterm labor that permits corticosteroid treatment.
In clinical application, oral tocolytic agents are commonly used with HUAM. They may be administered either before or after an episode of acute treatment of premature labor. A recent metaanalysis of oral tocolytic therapy fails to support a role for oral tocolytic therapy following resolution of an acute episode of preterm labor (30). The following conclusions can be made about the effectiveness of treatment:
Any potential benefit of HUAM is limited by the effectiveness of treatment available for preterm labor.
A benefit of HUAM derived from allowing time for corticosteroid treatment has not been evaluated.
No apparent benefit is derived from HUAM in assisting the clinician in prescribing oral tocolytic therapy.
Risks
Use of HUAM is not invasive. Therefore, the risks are confined to interventions that might result from use of the system, such as tocolytics and corticosteroids (28, 29).
Conclusion
Well-designed, prospective, randomized clinical studies of sufficient power are still needed to establish the benefit, if any, of HUAM for the prevention of preterm delivery or for the prevention of associated adverse neonatal outcomes. Without an established benefit, cost-benefit analysis is not possible. Data are insufficient to support a benefit from HUAM in preventing preterm birth. Therefore, the American College of Obstetricians and Gynecologists does not recommend the use of this system of care. Other recently published reviews are available that do not change these conclusions (31, 32).
References
1. Ventura SJ, Martin JA, Taffel SM, Mathews TJ, Clarke SC. Advance report of final natality statistics, 1993. Monthly vital statistics report; vol 44, no. 3, suppl. Hyattsville, Maryland: National Center for Health Statistics, 1995
2. Institute of Medicine. Division of Health Promotion and Disease Prevention. Committee to Study the Prevention of Low Birthweight. Preventing low birthweight. Washington, DC: National Academy Press, 1985
3. March of Dimes Birth Defects Foundation. March of Dimes statbook: statistics for healthier mothers and babies. White Plains, New York: MODBDF, 1993
4. Kochanek DK, Hudson BL. Advance report of final mortality statistics, 1992. Monthly vital statistics report; vol 43, no. 6, suppl. Hyattsville, Maryland: National Center for Health Statistics, 1995
5. American College of Obstetricians and Gynecologists. Preterm labor. ACOG Technical Bulletin 206. Washington, DC: ACOG, 1995
6. Iams JD, Johnson FF, O'Shaughnessy RW. A prospective random trial of home uterine activity monitoring in pregnancies at increased risk of preterm labor. Am J Obstet Gynecol 1987;157:638-643
7. Iams JD, Johnson FF, O'Shaughnessy RW. A prospective random trial of home uterine activity monitoring in pregnancies at increased risk of preterm labor. Part II. Am J Obstet Gynecol 1988;159:595-603
8. Katz M, Gill PJ, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity: a preliminary report. Obstet Gynecol 1986;68:773-778
9. Katz M, Gill PJ, Newman RB. Detection of preterm labor by ambulatory monitoring of uterine activity for the management of oral tocolysis. Am J Obstet Gynecol 1986;154:1253-1256
10. Smith JC, Floyd RC, Pittman KP, Grillo D, Roach H, Morrison JC. Reduction of preterm births in a low-income population utilizing home uterine activity monitoring. J Maternal-Fetal Investigation 1994;4:87-92
11. Keirse MJ, Van Hoven M. Reanalysis of a multireported trial on home uterine activity monitoring. Birth 1993; 20:117-122
12. Grimes DA, Schulz KF. Randomized controlled trials of home uterine activity monitoring: a review and critique. Obstet Gynecol 1992;79:137-142
13. Sachs BP, Hellerstein S, Freeman R, Frigoletto F, Hauth JC. Home monitoring of uterine activity. Does it prevent prematurity? N Engl J Med 1991;325:1374-1377
14. U.S. Preventive Services Task Force. Home uterine activity monitoring for preterm labor. Review article. JAMA 1993;270:371-376
15. Katz M, Newman RB, Gill PJ. Assessment of uterine activity in ambulatory patients at high risk of preterm labor and delivery. Am J Obstet Gynecol 1986;154:44-47
16. Scheerer LJ, Campion S, Katz M. Ambulatory tocodynamometry data interpretation: evaluating variability and reliability. Obstet Gynecol 1990;76S-79S
17. Katz M, Newman RB, Gill PJ. Assessment of uterine activity in ambulatory patients at high risk of preterm labor and delivery. Am J Obstet Gynecol 1986;154:44-47
18. Morrison JC, Martin JN Jr, Martin RW, Gookin KS, Wiser WL. Prevention of preterm birth by ambulatory assessment of uterine activity: a randomized study. Am J Obstet Gynecol 1987;156:536-543
19. Hill WC, Fleming AD, Martin RW, Hamer C, Knuppel RA, Lake MF, et al. Home uterine activity monitoring is associated with a reduction in preterm birth. Obstet Gynecol
20. Knuppel RA, Lake MF, Watson DL, Welch RA, Hill WC, Fleming AD, et al. Preventing preterm birth in twin gestation: home uterine activity monitoring and perinatal nursing support. Obstet Gynecol 1990;76(suppl):24S-27S 1990;76(suppl):13S-18S
21. Dyson DC, Crites YM, Ray DA, Armstrong MA. Prevention of preterm birth in high-risk patients: the role of education and provider contact versus home uterine monitoring. Am J Obstet Gynecol 1991;164:756-762
22. Mou SM, Sunderji SG, Gall S, How H, Patel V, Gray M, et al. Multicenter randomized clinical trial of home uterine activity monitoring for detection of preterm labor. Am J Obstet Gynecol 1991;165:858-866
23. Nagey DA, Bailey-Jones C, Herman AA. Randomized comparison of home uterine activity monitoring and routine care in patients discharged after treatment for preterm labor. Obstet Gynecol 1993;82:319-323
24. Wapner RJ, Cotton DB, Artal R, Librizzi RJ, Ross MG. A randomized multicenter trial assessing a home uterine activity monitoring device used in the absence of daily nursing contact. Am J Obstet Gynecol 1995;172:1026-1034
25. Yin L. Home monitoring for uterine activity. A response from the FDA. N Engl J Med 1991;325:1377
26. Watson DL, Welch RA, Mariona FG, Lake MF, Knuppel RA, Martin RW, et al. Management of preterm labor patients at home: does daily uterine activity monitoring and nursing support make a difference? Obstet Gynecol 1990;76(suppl):32S-35S
27. Blondel B, Breart G, Berthoux Y, Berland M, Mellier G, Rudigoz RC, et al. Home uterine activity monitoring in France; a randomized, controlled trial. Am J Obstet Gynecol 1992;167:424-429
28. American College of Obstetricians and Gynecologists. Antenatal corticosteroid therapy for fetal maturation. ACOG Committee Opinion 147. Washington, DC: ACOG, 1994
29. Effect of corticosteroids for fetal maturation on perinatal outcomes. NIH Consensus Development Panel on the Effect of Corticosteroids for Fetal Maturation of Perinatal Outcomes. JAMA 1995;273:413-418
30. Macones GA, Berlin M, Berlin JA. Efficacy of oral beta- agonist maintenance therapy in preterm labor: a meta- analysis. Obstet Gynecol 1995;85:313-317
31. The Collaborative Home Uterine Monitoring Study Group. A multicenter randomized controlled trial of home uterine monitoring: active versus sham device. Am J Obstet Gynecol 1995;173:1120-1127
32. Colton T, Kayne HL, Zhang Y, Heeren T. A metaanalysis of home uterine activity monitoring. Am J Obstet Gynecol 1995;173:1499-1505