Objective
The purpose of this study was to investigate quality-of-care improvements after the introduction of ST waveform analysis as an adjunct to standard cardiotocography (CTG).
Study Design
This was a prospective clinical study that was conducted over 7 years. Four yearly cohorts of 12,832 term pregnancies were part of a detailed analysis. Cord blood metabolic acidosis and neonatal outcome were main outcome measures.
Results
The STAN (S31 Fetal Heart Monitor; Neoventa Medical AB, Mölndal, Sweden) usage rate increased from 26 to 69%. The cord metabolic acidosis rate was reduced from 0.72 to 0.06%. This 91.7% improvement was associated with a significant reduction in the number of cases with a prolonged response time, calculated as the time from CTG + ST indications to intervene until delivery and an ability of the staff to identify and act on preterminal and unstable fetal heart rate patterns at the onset of a recording.
Conclusion
Our data indicate a paradigm shift in the outcome of delivery related to a high rate of CTG + ST usage and the application of structured CTG analysis.
The development and validation of ST analysis of the fetal electrocardiogram as an adjunct to standard cardiotocography (CTG) has included a long series of studies, which started with experimental studies in the 1970s to recent randomized controlled trials (RCTs) that led on to positive conclusions in the Cochrane library.
Although the basic scientific requirement for the validation of a new medical method is successfully conducted RCTs, clinically relevant information may be gained from detailed clinical audits. Furthermore, an important aspect of health care developments is to conduct quality-improvement research to allow the assessment of quality of care in a total population beyond what is possible with randomized clinical trials.
Previously, we have evaluated the introduction of the STAN method (S31 Fetal Heart Monitor; Neoventa Medical AB, Mölndal, Sweden) in the total population of term deliveries over 2 years in the city of Gothenburg. Increasing STAN usage provided parallel reduction in the rate of cord metabolic acidosis as a marker of fetal outcome without increasing operative interventions for fetal distress. The aim of the current study was to extend the investigation beyond the initial 2 years to a 7-year time frame with the use of applicable auditing techniques that would focus on 1 of the delivery units that participated in the initial 2-year analysis. Thus, the population consisted of all term deliveries in active labor in a district hospital setting of low- and medium-risk pregnancies.
Materials and Methods
The current analysis investigates the perinatal outcome over a 7-year period from January 1, 2001–December 31, 2007. The delivery ward at Mölndal Hospital is a large obstetric unit with >3000 deliveries managed by 20 doctors and 97 midwifes on a yearly basis. The STAN method was introduced at the hospital in September 2000 when 2 STAN S21 units were acquired as part of a European Union-sponsored project. In 2005, 3 additional units were bought; in 2006, 8 units were in use. The ward had 1 doctor (H.N.) and 1 midwife (A.C.) who had been responsible for continuing education, data collection, and case discussions since 2001. The training consisted of lectures, written information, and multimedia-based teaching. The hospital was not a tertiary care neonatal unit, and high-risk pregnancies were not delivered in the labor ward. Only depersonalized data were used; therefore, ethical approval was not required.
The general indications for the use of ST analysis as an adjunct to standard CTG analysis were 36 completed gestational weeks. At our unit, continuous monitoring during the second stage of labor is the norm; in 70% of the deliveries, a scalp electrode is applied by midwifes that enabled ST analysis. The specific indications for use of the STAN monitor were an abnormal CTG pattern, meconium-stained amniotic fluid, postmaturity, preeclampsia, diabetes mellitus, intrauterine growth restriction, and twins. Because of a restricted number of STAN units, not all patients with CTG abnormalities could be monitored with CTG + ST analysis. The device was also used for external monitoring, in which case no ST information is available.
STAN clinical guidelines give information about the definition of normal, intermediary, abnormal, and preterminal CTG patterns; ST analysis was used when the CTG pattern was intermediary or abnormal. In case of unfavorable fetal heart rate (FHR) patterns at the onset of a recording, intervention was made on the basis of FHR information only, and ST data were not required. An unfavorable FHR pattern at onset was defined as lack of variability and reactivity (preterminal); an unstable FHR was defined as undetermined baseline heart rate during the initial 20 minutes. Fetal blood sampling (FBS) was left to the discretion of the attending physician. Cord artery and vein samples were obtained routinely by immediate cord clamping, and acid base data were checked for accuracy according to arterio-venous PCO 2 difference of 1.0 kPa and pH ≥0.03 units. Metabolic acidosis was defined as cord artery pH<7.05 and base deficit >12.0 mmol/L 2 . Base deficit in the extracellular fluid compartment was calculated with the Siggaard-Andersen Acid Base Chart algorithm.
Routine cord sampling forms the basis for the current obstetric outcome analysis. A high sampling rate (>90%) may still leave a small number of cases in which intrapartum asphyxia was present. To reduce this potential error, we used neonatal data that were obtained during the initial 60 minutes as an additional marker, similar to what had been done previously; in the case of neonatal data indicating metabolic acidosis, the case would be included in the metabolic acidosis group.
STAN recordings automatically are given a separate identification number at the time of the recording. These unique case numbers were associated with patient identification numbers that are used in the hospital-based patient database (Obstetrix; Siemens AB, Mölndal, Sweden). These data and the FBS data were available only from 2005 and onward. For 2001, mode of monitoring and FBS data had been collected independently of the hospital-based patient database. Clinical data were obtained from the information that had been collected as part of the Swedish National Perinatal Registry. Retrospective assessment of CTG + ST recordings was made in cases with cesarean deliveries for fetal distress (CSFD) and metabolic acidosis at delivery and in which the FBS had been nonreassuring (scalp pH <7.20; scalp lactate >4.7 mmol/L ). The aim was to assess to what extent the CTG + ST clinical guidelines had been followed and to calculate the response time (time from CTG + ST guideline indications to delivery).
The results were statistically evaluated with the Medcalc statistical software (v. 5; Medcalc, Mariakerke, Belgium). χ 2 or Fisher’s exact test were used for discrete variables, and the odds ratios (ORs) with 95% confidence intervals (CIs) were given. The Student t test was used for testing statistical significance for continuous variables. Probability values < .05 were considered significant.
Approximately 5000 deliveries are required to obtain statistical power to detect a change in metabolic acidosis rates by 0.4% (β = .20; α = .05).
Results
Table 1 provides annual obstetric data for the period 2001–2007. During the last 7 years, the number of term deliveries in active labor increased by 47%, from 2417 deliveries in 2001 to 3561 deliveries in 2007. Table 2 provides information on term deliveries in active labor, cord metabolic acidosis rate, CTG + ST usage rate (including those cases in which additionally a fetal blood sample was obtained for scalp pH or lactate analysis), overall CSFD, and the total operative delivery rate for fetal distress (ODFD). Detailed analysis was made for the years 2001 and 2005–2007 that covered 12,832 deliveries, of which 7663 deliveries had ST analysis as an adjunct to standard CTG.
| Year | Total deliveries, n | Birthweight, g a | Total cesarean delivery rate, % | Operative vaginal delivery rate, % | Total operative delivery rate, % | Cord artery pH <7.00, % |
|---|---|---|---|---|---|---|
| 2001 | 2677 | 3574 ± 507 | 17.9 | 4.6 | 22.5 | 1.54 |
| 2002 | 3110 | 3578 ± 490 | 14.6 b | 4.2 | 18.7 b | 0.51 b |
| 2003 | 3340 | 3559 ± 493 | 12.4 b | 6.3 c | 18.7 b | 0.74 d |
| 2004 | 3558 | 3568 ± 496 | 11.8 b | 5.9 d | 17.7 b | 0.45 b |
| 2005 | 3596 | 3555 ± 473 | 13.2 b | 4.4 | 17.6 b | 0.61 c |
| 2006 | 3786 | 3561 ± 471 | 13.8 b | 4.9 | 18.7 b | 0.46 b |
| 2007 | 3828 | 3535 ± 474 | 14.8 c | 5.1 | 19.9 d | 0.31 b |
a Data are given as average ± SD;
b P < .001; χ 2 was used with annual comparisons that were made with the use of 2001 data as reference;
| Year | Patients at term in active labor, n | Cord blood gas sampling rate, % | Metabolic acidosis rate, % | Cardiotocography + ST usage rate, % | Cardiotocography + ST + fetal blood sampling rate, % | Cesarean delivery for fetal distress, % | Operative delivery for fetal distress, % |
|---|---|---|---|---|---|---|---|
| 2001 | 2417 | 80.7 | 0.72 | 26 | 9.9 | 2.4 | 4.8 |
| 2002 | 2880 | 89.9 | 0.66 | NA | NA | 2.5 | 4.5 |
| 2003 | 3121 | 91.5 | 0.56 | NA | NA | 2.3 | 5.6 |
| 2004 | 3338 | 94.6 | 0.25 a | NA | NA | 2.3 | 5.2 |
| 2005 | 3321 | 95.7 | 0.31 a | 58 | 4.0 | 2.0 | 4.6 |
| 2006 | 3533 | 96.1 | 0.24 a | 68 | 5.0 | 2.7 | 5.3 |
| 2007 | 3561 | 96.3 | 0.06 b | 69 | 7.5 | 3.0 | 5.9 |
b P < .001, χ 2 was used with annual comparisons that were made with the use of 2001 data as reference.
The rate of CTG + ST usage increased significantly every year, as did the FBS rate between –05 and –07 (OR, 0.51; 95% CI, 0.39–0.68). CSFD and ODFD increased significantly during the same time period (OR, 0.65; 95% CI, 0.48–0.89 and OR, 0.77; 95% CI, 0.62–0.95, respectively).
Over the last 3 years of the study, there were 20 cases with cord metabolic acidosis. Table 3 provides information on observations made for each of these cases.
