You can respond to this article at http://www.anaesthesiacorrespondence.com
Closed-loop double-vasopressor automated system to treat hypotension during spinal anaesthesia for caesarean section: a preliminary study
Version of Record online: 28 SEP 2012
Anaesthesia © 2012 The Association of Anaesthetists of Great Britain and Ireland
Volume 67, Issue 12, pages 1348–1355, December 2012
How to Cite
Sia, A. T. H., Tan, H. S. and Sng, B. L. (2012), Closed-loop double-vasopressor automated system to treat hypotension during spinal anaesthesia for caesarean section: a preliminary study. Anaesthesia, 67: 1348–1355. doi: 10.1111/anae.12000
- Issue online: 7 NOV 2012
- Version of Record online: 28 SEP 2012
- Accepted: 19 August 2012
Hypotension occurs in most caesarean sections under spinal anaesthesia, necessitating vasopressor administration. However, the optimal dosing regimen remains unclear. We have developed a novel vasopressor delivery system that automatically administers phenylephrine or ephedrine based on continuous non-invasive blood pressure monitoring. This prospective cohort study recruited 55 healthy women under standardised spinal anaesthesia with 2.2 ml hyperbaric bupivacaine 0.5%, 15 μg fentanyl and 100 μg morphine. A 50-μg phenylephrine bolus was given at 1-min intervals when systolic blood pressure fell below 90% of baseline, and a 4-mg ephedrine bolus was given when hypotension developed with bradycardia (heart rate <60 beats.min−1). Systolic blood pressure was within 20% of baseline in 88% of all measurements. Six patients (11%) had one or more measurements above 120% of baseline (1% of all measurements), whereas 36 (65%) had at least one reading below 80% of baseline (11% of total measurements). The system maintained systolic blood pressure at a mean (SD) of −9.1 (7.0)% below baseline, with 5.4 (2.5)% fluctuation. Two patients (4%) experienced pre-delivery nausea. All 5-min Apgar scores were 9.
Peri-operative hypotension and fluctuation of arterial blood pressure (BP) during spinal anaesthesia are common. Recent data support the use of α1-agonists, such as phenylephrine, to maintain maternal systolic BP during spinal anaesthesia for caesarean section [1, 2]. However, use of ephedrine may be advocated in patients with hypotension associated with bradycardia after spinal anaesthesia. The optimal regimen for the administration of vasopressors is yet to be determined.
Phenylephrine is a potent, fast-acting vasopressor with a short duration of action and recent studies suggest that it has no detrimental effect on fetal lactate or acid-base status . Phenylephrine has been associated with reflex bradycardia that may require treatment. In contrast, the use of ephedrine is somewhat limited and high doses have been associated in a dose-dependent manner with fetal acidosis, possibly due to direct stimulation of fetal metabolism . Furthermore, ephedrine has a slow onset and long duration of action, making dose-titration difficult.
Optimal timing and dosing of vasopressors is dependent on accurate and timely BP monitoring. Non-invasive oscillometric BP monitoring is limited by its inability to cycle at less than 1-min intervals in clinical practice, and is therefore inadequate in detecting rapid BP changes and providing timely data to guide vasopressor dosing. An invasive arterial catheter would not be justified in healthy women undergoing elective caesarean section. In addition, manual dosing of vasopressor is labour-intensive and distracting. A continous non-invasive BP monitoring system (CNAP™; Draeger Medical, Lubeck, Germany) with a closed-loop feedback to control vasopressor administration would address these issues. The CNAP system was found in recent studies to be superior to intermittent oscillometric measurements in detecting both rapid BP changes and intra-operative hypotension during spinal anaesthesia for caesarean section , and its accuracy and precision have been compared favourably with invasive intra-arterial BP monitoring [6, 7].
Herein, we describe the use of a closed-loop, double-vasopressor, automated system that integrates beat-to-beat systolic BP data from a CNAP device using a customised algorithm that controls two infusion pumps, with phenylephrine and ephedrine administered according to the algorithm.
We devised a closed-loop dual-pump automated system by exporting haemodynamic data (systolic BP and heart rate) measured using two alternating finger cuffs (Infinity® CNAP SmartPod® (Draeger Medical)) to a laptop computer (via a custom-built RS232 interface), where it is integrated every 15 s using a customised program (C# programming language, Microsoft.NET v3.5 framework, running on the Windows XP operating system (Microsoft Corporation, Redmond, WA, USA)) created by our in-house computer engineer. The system provides systolic BP measurements continuously in real time with an accuracy of 5 mmHg, and activates either one of the syringe pumps (B. Braun, Melsungen, Germany) containing phenylephrine or ephedrine when hypotension occurs (defined as below 90% of the baseline systolic BP). Our algorithm administered 50 μg phenylephrine if the systolic BP integrated over 15 s fell below 90% of the baseline. However, should hypotension occur contemporaneously with bradycardia (heart rate <60 beats.min−1), 4 mg ephedrine would be administered instead. Vasopressor administration required 15 s, followed by a 30 s lockout period to permit the vasopressors to take effect. A schematic of the system and the algorithm is shown in Fig. 1. Being automated, this system would ameliorate the need for labour-intensive manual dosing by the anaesthetist.
After research ethics committee approval and written informed consent from all subjects, we enrolled 55 women of ASA physical status 1–2, aged between 21 and 45 years, of weight 40–99 kg and height 145–170 cm, with singleton full-term pregnancies presenting for elective caesarean section under spinal anaesthesia. We did not study women in labour and those with allergy to study drugs, contraindications to spinal anaesthesia, obstetric complications, such as pre-eclampsia, multiple gestations or placenta praevia.
All patients received ranitidine and sodium citrate prophylaxis pre-operatively. After resting for 15 min in a quiet environment with left uterine displacement, baseline BP and heart rate measurements were recorded at 1-min intervals and the mean of three readings obtained and used as baseline. An intravenous 18-G cannula was inserted in the forearm and electrocardiogram and pulse oximeter applied, followed by spinal anaesthesia, which was performed in the sitting position. After ensuring free flow of cerebrospinal fluid, a mixture of 2.2 ml hyperbaric buvivacaine 0.5%, 15 μg fentanyl and 100μg morphine was injected intrathecally over 15–20 s through a 27-G pencil-point needle at the L2-3 or L3-4 vertebral interspace. The patient was then placed supine, intravenous co-loading commenced with 1500 ml Ringer’s lactate solution, and our system was initiated and monitored by one of the investigators who could manually administer atropine and/or additional phenylephrine or ephedrine if there were more than three readings of systolic BP less than 70% of the baseline.
The sensory block height was measured 5 min after spinal anaesthesia by assessing the loss of sensation to ice. The timings of spinal anaesthesia, surgical incision and delivery were recorded. We also recorded the presence of nausea or vomiting, and the total amount of intravenous fluids given up to delivery. The duration of the study was from the administration of spinal anaesthesia to delivery. The attending midwife or neonatologist assessed the Apgar scores at 1 and 5 min after delivery.
Our primary outcome was the incidence of reactive hypertension, defined as systolic BP greater than 20% above the baseline value. Reactive hypertension may occur in about 47% of women due to phenylephrine infusion given in high doses during spinal anaesthesia . We utilised the pooled-data approach to assess the performance of our system as previously described .
Percentage performance error
The percentage performance error (PE) was defined as the percentage difference between each measured value of systolic BP from the baseline. The PE for the ith patient at the jth minute was calculated as follows:
Median absolute performance error
The median absolute performance error (MDAPE) is a measure of inaccuracy and indicates the absolute magnitudes of the differences between measured and baseline BPs. For each patient, it was defined as the median of the absolute values of PE (|PE|) and was calculated as follows:
where Ni is the number of values of |PE| for the ith patient and M is the number of patients in the study.
Median performance error
The median performance error (MDPE) is a measure of bias and indicates whether the differences between measured BPs were systematically above or below baseline BPs. For each patient, it was defined as the median of PE and was calculated as follows:
Wobble measures how much PE fluctuates around the MDPE with time, for each patient (i.e. intrasubject PE variability). It was calculated as follows:
Divergence is defined as the slope obtained from linear regression of each patient’s |PE| against time, and describes the trend of changes in |PE| with time. It gives us an indication of whether the accuracy of our system improves (negative divergence) or decreases (positive divergence) with time. Divergence (per minute) was calculated as follows:
where tij is the time of each individual measurement, in minutes.
All calculations were performed using Microsoft Office Excel, 2008.
All 55 women had successful administration of spinal anaesthesia, and none needed supplemental analgesia or conversion to general anaesthesia. Their characteristics and baseline haemodynamic parameters are summarised in Table 1.
|Age; years||32.9 (4.8)|
|Weight; kg||69.8 (10.9)|
|Height; cm||158.2 (5.7)|
|Baseline systolic blood pressure; mmHg||122 (114–130 [102–140])|
|Baseline heart rate; beats.min−1||79 (74–88 [65–120])|
The majority of the systolic BP measurements (88%) were within 20% variation of the baseline systolic BP. Six (11%) patients had readings 20% above the baseline, accounting for 1% of the total readings. Of these six patients, three were hypertensive within the first 2 min of spinal anaesthesia, before any vasopressors were administered, and did not experience further episodes of hypertension for the rest of the study. Conversely, 36 (65%) patients had readings 20% below the baseline, which accounted for 11% of the total readings. Fifty patients (91%) required phenylephrine for the period between spinal anaesthesia and delivery, whereas 12 (22%) required ephedrine for the same period. Two patients (4%) experienced nausea before delivery, but there were no episodes of vomiting.
No additional phenylephrine, ephedrine or atropine was administered by the attending anaesthetist. In all cases of hypertension, the BP decreased spontaneously without intervention towards the baseline, within 1 min. The clinical outcomes of the patients are summarised in Table 2. There was one neonate with an Apgar score of 6 at 1 min, whereas the rest had Apgar scores of 9. All the neonates had Apgar scores of 9 at 5 min.
|Block height at 5 min||T3 (T2–T4 [T1–T6])|
|Spinal to incision time; min||11 (8–14 [5–42])|
|Spinal to delivery time; min||18 (15–24 [9–65])|
|Spinal to end of surgery time; min||44 (38–58 [28–75])|
|Dose of phenylephrine before delivery; μg||250 (150–450 [0–1200])|
|Dose of ephedrine before delivery; mg||0 (0–16 [0–48])|
|Total dose of phenyleprine; μg||900 (650–1000 [0–2400])|
|Total dose of ephedrine; mg||0 (0–16 [0–56])|
|Total fluids; ml||1500 (1000–2000 [1000–2000])|
|Maximum systolic blood pressure; mmHg||129 (125–138 [109–200])|
|Maximum heart rate; beats.min−1||108 (97–122 [82–153])|
|Minimum systolic blood pressure; mmHg||84 (77–90 [68–125])|
|Minimum heart rate; beats.min−1||59 (54–66 [47–85])|
|Apgar score at 1 min||9 (6–9 [6–9])|
|Apgar score at 5 min||9 (9–9 [9–9])|
|Birthweight; g||3135 (2944–3520 [2426–4642])|
|Median performance error; %||−9.10 (7.0)|
|Median absolute performance error; %||10.86 (4.6)|
|Wobble; %||5.43 (2.5)|
|Divergence; %.min−1||0.01 (0.5)|
This is the first preliminary study combining continuous non-invasive BP monitoring using an automated closed-loop feedback computer-controlled infusion of vasopressors for maintaining systolic BP during caesarean section under spinal anaesthesia. The system required minimal intervention from the attending anaesthetist and resulted in good clinical maternal and neonatal outcomes, with a low incidence of mild pre-delivery nausea (4%) and good Apgar scores at 5 min. This compares favourably with both clinician-administered phenylephrine boluses and infusion techniques using conventional BP monitoring [10, 11]. Allen et al. reported the incidence of intra-operative nausea in women undergoing caesarean section under spinal anaesthesia to be 32% with phenylephrine infusion (100 μg.min−1) and 35% with 100-μg phenylephrine boluses .
We assessed the performance of our system based on measures that were previously utilised to evaluated similar closed-loop systems [9, 12]. We utilised the more complicated pooled-data method that assigns a weighting to each BP measurement according to the total number of measurements obtained for that patient. Due to the huge variability in the systolic BP from spinal anaesthesia until delivery between patients, and consequently the total number of BP measurements obtained for each patient, simply averaging the measurements would result in over-emphasis on patients with fewer BP readings and under-emphasis on those with greater numbers of measurements. This method allowed us to assign more weight to measures from patients in whom our system’s performance is known with more accuracy (i.e. a greater number of BP measurements) and vice versa, thus producing more precise population estimates . Our negative MDPE indicated that our system maintained BP below baseline values by a mean (SD) of 9.1 (7.0)%, with the wobble indicating that the BP fluctuated around this MDPE by 5.4 (2.5)%. Our MDAPE revealed that our system resulted in systolic BP values that differed from baseline by a mean (SD) of 10.9 (4.6)%, and small divergence suggested that our system’s accuracy decreases by 0.01% per minute, but there is significant variability with this parameter. Nevertheless, these results enable us to modify our system to improve performance further. For instance, our current MDPE and wobble suggest that we can increase the threshold for vasopressor administration, which should improve both MDPE (bias) as well as MDAPE (inaccuracy). In an ideal scenario, adjusting MDPE such that it closely mirrors the baseline BP would result in an MDAPE close to wobble, which we measured in this study to be 5.4%.
Ngan Kee et al. have shown that pre-emptive phenylephrine administration combined with crystalloid co-loading drastically reduce the incidence of hypotension, but at the cost of increased phenylephrine doses and incidence of reactive hypertension [8, 9]. In this preliminary observational study, we set the threshold for vasopressor dosing at 90% of baseline systolic BP to enable reactive treatment of hypotension, while reducing the incidence of reactive hypertension. Although 36 (65%) patients were hypotensive at some point during the study, the system automatically administered additional vasopressors to achieve baseline values in the majority within 3 min, and the few hypertensive episodes that did occur did so before any vasopressors were administered, and resolved spontaneously without intervention. The impact of reactive hypertension with reflex bradycardia on the mother and neonate may not be innocuous, especially in high-risk patients, such as those with pre-eclampsia or intra-uterine growth retardation. Hence, we decided to use the incidence of reactive hypertension as our primary outcome.
Compared with traditional intermittent BP monitoring, our system avails us with the advantages of continuous BP measurements, while circumventing the drawbacks of invasive BP methods. This translates to two clinically important benefits. First, the system is able to detect and respond to hypotension at any point in time, and thus achieve faster restoration of normotension. Although we limited doses to a maximum rate of 1 min, our system is capable of administering multiple doses each minute by altering the dosing cycle duration. Second, as the system integrates BP measurements over 15-s intervals, artefacts or loss of single readings due to movement or shivering have much less impact, resulting in increased accuracy. Furthermore, our ability to record rapid fluctuations in BP as well as track its change after each vasopressor bolus is very different from the intermittent BP monitoring used in other similar studies, which does not detect brisk fluctuations. In the case of fast-acting vasopressors like phenylephrine, an intermittent measurement following a bolus may reflect the BP after it has reached a new plateau, thus underestimating the true magnitude of the hypotensive episode.
Moreover, our system utilises two vasopressors: phenylephrine and ephedrine. Phenylephrine was required to maintain BP in 90% of our study cohort. Phenylephrine also conveyed the advantages of high efficacy, ease of titration and minimal fetal acidaemia; however, bradycardia often limits its use . This situation arose in about 22% of our study cohort, for whom ephedrine was administered to prevent further reflex bradycardia. Despite the slower onset and longer duration of action of ephedrine, which negates some of the advantages of our rapid-response system and making titration more difficult, we felt that this was acceptable because: (i) its use as a second-line vasopressor mirrors our institution’s practice; (ii) whereas inferior to phenylephrine, ephedrine has a long track record of use in obstetrics and its side-effect profile is well known; and (iii) we wanted an agent that would simultaneously treat both hypotension and bradycardia.
Hence, our novel system would be able to detect and respond automatically and instantaneously to hypotension in real time. In addition, by utilising both phenylephrine and ephedrine, the effects of hypertension, bradycardia and reduction of uteroplacental blood flow due to high doses of phenylephrine alone may be modulated. Finally, by integrating measurements over a period of time, our system markedly reduces the influence of artefacts. Our system may allow closer monitoring and less variation in systolic BP swings.
There are several limitations of our study. The incidence of adverse maternal and fetal outcomes used in this study were based on clinical markers, such as nausea, vomiting, headaches and Apgar scores. The sensitivity of these clinical markers may be inadequate to detect fluctuations in haemodynamics that may be clinically insignificant in healthy women; however, large changes would probably not be seen in low-risk elective cases. The addition of umbilical venous and arterial pH, lactate and blood gases may prove a more sensitive marker of adverse haemodynamics and this should be incorporated in future studies. The system administered vasopressors only if the average systolic BP over 15 s was below 90% of baseline. Although this reactive strategy of treating hypotension is consistent with our aim of reducing the incidence of overtreatment resulting in reactive hypertension, it may result in a higher incidence of hypotension than a pre-emptive method such as that described by Ngan Kee et al. [8, 9]. However, we hypothesise that refining our algorithm, namely, increasing the threshold to treatment of hypotension and increasing the dosing frequency from once per minute to once per 30 s may ameliorate the risk of hypotension further, whilst reducing reactive hypertension. Smiley et al. found that β-adrenoceptor polymorphism may lead to lower vasopressor requirements in spinal anaesthesia . Further research may also involve the trend of systolic BP response with adrenoceptor polymorphism using this algorithm. Finally, as the baseline BP was measured via oscillometry, and CNAP readings were obtained via volume-clamp photoplethysmography, it is possible that inherent inaccuracies in CNAP could have adversely affected the performance parameters of our system. However, as the accuracy and precision of CNAP has compared favourably with intra-arterial BP measurements in previous studies [6, 7], and because CNAP automatically recalibrates itself against oscillometric BP measurements every 10 min, we felt that this issue was unlikely.
In conclusion, our novel system combining continuous non-invasive BP monitoring with an automated closed-loop feedback computer-controlled system with phenylephrine, and ephedrine has been shown in this preliminary study to be clinically effective in maintaining BP during caesarean section under spinal anaesthesia, with minimal attention from the anaesthetist, a low incidence of nausea pre-delivery, and minimal maternal and fetal adverse effects. Future research involving a comparison between our system and manual dosing based on oscillometric systolic BP measurements is required.
The authors acknowledge Philip Cheong, computer engineer, for creating the program and for his technical expertise, and Agnes Teo, research nurse, for her assistance in data collection.
This study was funded by National Medical Research Council, Singapore, grant number NMRC/EDG/1003/2011. No competing interests declared.
- 5Continuous non-invasive blood pressure monitoring during spinal anesthesia for cesarean section. European Journal of Anaesthesiology2010; 27: 166–7., , , , .
- 9Closed-loop feedback computer-controlled infusion of phenylephrine for maintaining blood pressure during spinal anaesthesia for caesarean section: a preliminary descriptive study. Anaesthesia2007; 62: 1251–6., , , , , .