General anaesthesia in children results in a significant decrease of arterial blood pressure.1 To detect and treat hypotension during anaesthesia in young children, accurate blood pressure measurement is of the utmost importance, as prolonged hypotension has a detrimental effect on brain function in young children.2
Although directly obtained blood pressure via an arterial cannula is considered the gold standard of blood pressure measurement, the risks of arterial cannulation preclude the widespread use of intra-arterial measurement in paediatric anaesthesia.3 In addition, cannulation of an infant's radial artery may be technically difficult and time-consuming. Therefore, in the vast majority of surgical procedures under general anaesthesia in neonates and small children, arterial blood pressure is monitored by an oscillometric technique via an upper arm cuff. Unlike the traditional manual blood pressure measurement using Korotkoff sounds, the automatic noninvasive oscillometric techniques also provide mean arterial pressure, and consequently, it has widely replaced the manual determination of blood pressure.4 Data on the measurement performance of oscillometric measurement in this age group during surgical procedures are rare and were conducted more than 30 years ago.5,6 Since then, the statistical approach for assessing agreement between two measurement techniques has changed and we routinely use the oscillometric technique without up-to-date studies on the accuracy of such blood pressure measurements in this vulnerable patient group. As the inaccuracy of blood pressure measurement, whether falsely low or high, may lead to incorrect therapeutic decisions and adverse outcomes, we aimed to observe prospectively the differences between invasive and noninvasive arterial blood pressure measurements in infants and young children undergoing general anaesthesia for neurosurgical procedures.
Materials and methods
The study was conducted at a 2000-bedded German university hospital from November 2015 until January 2018. The study protocol (number 491–15) was approved by the Ethics committee of the faculty (Ethikkommission bei der LMU, Pettenkoferstr. 8a, Munich, Germany: chairman W. Eisenmerger) on 13 October 2015.
We conducted a prospective observational study to assess the differences between two established methods of blood pressure measurement in paediatric patients during general anaesthesia. Blood pressure was measured simultaneously via an intra-arterial cannula (reference standard) and by an oscillometric upper arm cuff.
Children up to the age of 2 years who were to undergo a next-day neurosurgical procedure under general anaesthesia and required invasive blood pressure monitoring for reasons unrelated to the study were eligible for inclusion. The parents were asked for written informed consent before a child was included in the study. None of the patients suffered from congenital heart disease. Patients were enrolled consecutively. We did not enrol patients on weekends and holidays, when most of the investigating team was absent.
General anaesthesia was induced intravenously by administration of either thiopental or propofol. Sufentanil was used as an opioid for induction and, if muscle relaxation was necessary, cisatracurium was administered. After endotracheal intubation, anaesthesia was maintained with sevoflurane or intravenously by continuous infusions of propofol and remifentanil.
An upper arm cuff for intermittent blood pressure measurement (Dräger Infinity, Dräger Lubeck, Germany) was placed on the upper arm of the child. The correctly sized cuff was chosen according to the manufacturer's guidance after measuring upper arm circumference. The automatic measurement cycle was then set at 10 min. The reference method was intra-arterial blood pressure in the radial artery. A 24G catheter was inserted in the radial artery of the contralateral arm and connected via meticulously fluid-filled connecting tubes to the pressure transducer. Skilled staff frequently checked zeroing, damping and continuous flushing of the invasive measurement system, as well as ensuring the correct height of the pressure transducer and the upper arm cuff at the level of the right atrium of the heart. The monitors used were checked by technicians at a yearly interval according to standards set by German directives.
We recorded simultaneous measurement of invasive and noninvasive blood pressure every 10 min from the onset of invasive blood pressure monitoring.
All vital parameters were stored electronically and extracted immediately after the procedure. Measurement values obtained at the same time as blood withdrawals or flushing of the system were deleted.
On the basis of the intraoperative observation of similar simultaneous measurements in three children before the study, a mean bias of 5 (4) mmHg was observed during the procedures. A difference of 5 mmHg is clinically relevant in young children. Assuming an α of 0.05, a β of 0.2, a sample size of n = 21 was calculated. To account for a drop-out rate of 15%, we planned to include 25 patients.
Data analysis was performed using SPSS 23 (IBM, Armonk, New York, USA) and R 3.2.3 (The R project for statistical computing, Vienna, Austria).
Normally distributed demographic parameters are expressed either as total number n (%) or as mean (SD). Demographic data that were found to be not normally distributed are expressed as median [IQR].
Although it is not the method of choice for the comparison of two methods, we included the calculation of the correlation coefficient, because all previous studies with comparable data were published before the Bland–Altman method became established. The Bland–Altman analysis took account of repeated measurements7 and the percentage error was calculated as suggested by Critchley and Critchley.8
In addition, measurement pairs that included a low invasive mean arterial pressure (≤ 45 mmHg) were analysed separately, as high accuracy is especially important in such hypotensive cases.
Twenty-five children were included in the study. The data from four were excluded from analysis: in one case, the arterial catheter was placed in the posterior tibial artery; one patient had the blood pressure cuff placed on the thigh; in two cases, an error in data transmission occurred.
The median age of the children was 6 [5 to 11] months. Height and weight were 70 (11) cm and 8 (3) kg, respectively. The mean ratio of cuff width to arm circumference was 0.52 (0.04).
In four patients, a brain tumour was removed, seven patients underwent a correction of craniosynostosis, four patients had surgery for spina bifida, three patients had an intracranial cyst and three had other procedures. No patient suffered from a heart disease. Duration of anaesthesia and surgery were 6.2 (2.7) and 3.5 (1.9) hours, respectively. Anaesthesia time included insertion of peripheral intravenous cannulae, application of intra-operative neuromonitoring, positioning for surgery and time until extubation.
The median blood loss was 100 [40 to 350] ml.
After checking the data for obvious artefacts due to blood sampling or clotting, 820 pairs of simultaneous invasive and noninvasive blood pressure measurement were analysed from 21 patients. This equates to 28 [21 to 49] pairs of measurements per patient.
Descriptive blood pressure data for noninvasive and invasive arterial pressure values are presented in Table 1.
The correlation coefficient, bias, SD, 95% limits of agreement and percentage error of systolic, mean and diastolic arterial pressure comparison are summarised in Table 2.
The Bland–Altman plot of systolic and mean arterial pressures is shown in Fig. 1.
In 14 patients, 90 pairs that included a hypotensive invasive mean arterial pressure (≤ 45 mmHg) were obtained. The analysis of these data revealed a mean bias of −9 (5) mmHg, with 95% limits of agreement of –18 to 0 mmHg (percentage error 20%). The bias of systolic arterial pressure in these data indicated that the systolic measurement from the upper arm cuff was 13 mmHg higher (percentage error 41%) than the intra-arterial pressure.
In this study, we compared oscillometric with direct intra-arterial measurement of blood pressure in infants and small children during scheduled general anaesthesia for neurosurgical procedures. Although the overall performance of the noninvasive device was acceptable, in cases of low invasive blood pressure, there was a clinically relevant bias between the methods.
Surprisingly, the data regarding the accuracy of automated oscillometric devices in this age group in healthy children are sparse. Many studies compare the two measurement techniques in the paediatric ICU.5 Studies comparable with our work are more than 30 years old,6,9 implying that data analysis was based on linear regression analysis. Today, Bland–Altman analysis is the recognised approach for testing the interchangeability of two methods for cardiovascular monitoring.7
In their 1987 study, Cullen et al.6 compared an oscillometric device with intra-arterial blood pressure in 16 neonates and infants, weighing up to 6 kg and calculated the correlation coefficient as 0.73. However, only a very small minority of invasive measurements of mean arterial pressure were lower than 45 mmHg (approximately six out of 260).6
In a similar study from 1981, Friesen and Lichtor9 analysed blood pressure in the same age group. Twenty of the studied patients had direct arterial pressure measurement. Unfortunately, the authors do not report mean arterial pressure. The oscillometric device in both studies was the Dinamap monitor (GE Healthcare, Chicago, Illinois, USA).
Devices utilising an oscillometric technique inflate the upper arm cuff until the pressure is well above the estimated systolic arterial pressure; then, subsequently, the cuff is deflated. During deflation, oscillations from the brachial artery that transmit to the cuff are detected. A patient's mean arterial pressure correlates well with the maximum amplitude of the pressure oscillations in the cuff.10 It is important to understand that oscillometric devices do not measure directly the systolic and diastolic pressure, but calculate these based on a proprietary formula.4 Hence, the performance of devices from different companies is not interchangeable.11 In our study, we used the oscillometric technology embedded within the Dräger monitor. To our knowledge, there are no published data in infants using this device during general anaesthesia. For three other devices, Dannevig et al.12 presented results comparable to our findings in critically ill neonates.
As a primary endpoint, we chose mean arterial pressure accuracy and precision rather than systolic pressure. Despite intra-arterial pressure monitoring being regarded as the gold standard of blood pressure monitoring, there are still sources of measurement errors such as damping or systolic overshoot.13 In this regard, mean intra-arterial arterial pressure is more reliable than systolic or diastolic intra-arterial pressures. In addition, as explained above, devices using an oscillometric technology measure the mean arterial pressure. Therefore, we compared mean pressures, as these are the most accurate values for both measurement techniques in our study.
The results of our study raise an important question. Commonly, the blood pressure of children undergoing general anaesthesia is monitored by noninvasive means. Recent studies suggest that blood pressure stability is of the utmost importance during anaesthesia in neonates and infants. In severe hypotension, cerebral autoregulation is impaired.14 In infants, the effect of low mean arterial pressure, as measured noninvasively on cerebral desaturation, was demonstrated by Michelet et al.15 Impaired cerebral perfusion may contribute to brain damage.2,16 In accordance with these results, a prospective animal study by Ringer et al.17 showed metabolic disturbances due to moderate hypotension in piglet brains. Whether a mean arterial pressure of 40 mmHg derived by an oscillometric device in an infant during general anaesthesia should be regarded as sufficient is debatable, yet many accept this limit.18,19 Also, one should take into account the fact that normal blood pressure in children from birth to 2 years of age varies considerably: a mean arterial pressure of 40 mmHg is probably acceptable in a one-month-old baby but indicates profound hypotension at 18 months of age. In reality, for these children, the lowest acceptable systolic or mean arterial pressure under general anaesthesia is unknown. Furthermore, the technique of measurement is critical. Our data show that a mean arterial pressure of 40 mmHg measured by an upper arm cuff may correspond to only 31 mmHg within the artery. This relative overestimation of blood pressure by oscillometry would lead to a physician regarding a patient's blood pressure as being within well tolerated boundaries. In addition, factors such as the type of the monitor, the size of the cuff and the measurement site (e.g. upper arm, calf or thigh) all have an impact on measurement accuracy.
Consequently, invasive blood pressure measurement should be established in neonates and infants undergoing procedures associated with a significant blood loss and/or the risk of hypotension. Apart from its ability to measure blood pressure accurately, a main advantage of the intra-arterial catheter is the possibility of beat to beat measurement. In conditions such as acute bleeding, embolism or elevation of intracranial pressure, it is critical to monitor blood pressure continuously. Establishing arterial access with ultrasound guidance probably reduces the risk of failure or haematoma.
There are some limitations to the results of the study; first, the data come from one centre, and second, only one device was studied. Further research with a multicentre approach and different devices is essential in order to obtain reliable data to allow the conduct of well tolerated anaesthesia in this most vulnerable patient group.20 Furthermore, we did not measure blood pressure in both arms to detect possible differences prior to study inclusion. Data on inter-arm differences of blood pressure in young children are unavailable.
In this study of neonates and young children, blood pressure derived by the oscillometric device showed acceptable agreement provided the invasive blood pressure was within the normal range. However, in hypotension (mean invasive arterial pressure ≤ 45 mmHg), there was a considerable and clinically relevant overestimation of blood pressure measured by the oscillometric upper arm cuff. Thus, arterial cannulation should be established in this group of vulnerable patients when undergoing major surgical procedures associated with a significant risk for hypotension.
Acknowledgements relating to this article
Assistance with the study: none
Financial support and sponsorship: the study was funded by institutional sources.
Conflict of interest: none.
Presentation: preliminary data were presented in 2017 at a national congress (Deutscher Anästhesie Congress) in Nuremberg.
1. Sottas CE, Cumin D, Anderson BJ. Blood pressure and heart rates in neonates and preschool children: an analysis from 10 years of electronic recording. Paediatr Anaesth
2. McCann ME, Schouten AN, Dobija N, et al. Infantile postoperative encephalopathy: perioperative factors as a cause for concern. Pediatrics
3. Frezza EE, Mezghebe H. Indications and complications of arterial catheter use in surgical or medical intensive care units: analysis of 4932 patients. Am Surg
4. Butani L, Morgenstern BZ. Are pitfalls of oscillometric blood pressure measurements preventable in children? Pediatr Nephrol
5. Park MK, Menard SM. Accuracy of blood pressure measurement by the Dinamap monitor in infants and children. Pediatrics
6. Cullen PM, Dye J, Hughes DG. Clinical assessment of the neonatal Dinamap 847 during anesthesia in neonates and infants. J Clin Monit
7. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res
8. Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput
9. Friesen RH, Lichtor JL. Indirect measurement of blood pressure in neonates and infants utilizing an automatic noninvasive oscillometric monitor. Anesth Analg
10. Alpert BS, Quinn D, Gallick D. Oscillometric blood pressure: a review for clinicians. J Am Soc Hypertens
11. Kaufmann MA, Pargger H, Drop LJ. Oscillometric blood pressure measurements by different devices are not interchangeable. Anesth Analg
12. Dannevig I, Dale HC, Liestol K, Lindemann R. Blood pressure in the neonate: three noninvasive oscillometric pressure monitors compared with invasively measured blood pressure. Acta Paediatr
13. Romagnoli S, Ricci Z, Quattrone D, et al. Accuracy of invasive arterial pressure monitoring in cardiovascular patients: an observational study. Critical care
14. Vavilala MS, Lee LA, Lam AM. The lower limit of cerebral autoregulation in children during sevoflurane anesthesia. J Neurosurg Anesthesiol
15. Michelet D, Arslan O, Hilly J, et al. Intraoperative changes in blood pressure associated with cerebral desaturation in infants. Paediatr Anaesth
16. McCann ME, Schouten AN. Beyond survival; influences of blood pressure, cerebral perfusion and anesthesia on neurodevelopment. Paediatr Anaesth
17. Ringer SK, Ohlerth S, Carrera I, et al. Effects of hypotension and/or hypocapnia during sevoflurane anesthesia on perfusion and metabolites in the developing brain of piglets: a blinded randomized study. Paediatr Anaesth
18. de Graaff JC, Pasma W, van Buuren S, et al. Reference values for noninvasive blood pressure in children during anesthesia: a multicentered retrospective observational cohort study. Anesthesiology
19. Weber F, Honing GH, Scoones GP. Arterial blood pressure in anesthetized neonates and infants: a retrospective analysis of 1091 cases. Paediatr Anaesth
20. Weiss M, Bissonnette B, Engelhardt T, Soriano S. Anesthetists rather than anesthetics are the threat to baby brains. Paediatr Anaesth