SLEEP, Vol. 29, No. 8, 2006 1083

NEONATE SLEEP

INTRODUCTION

CONTINUOUS MEASUREMENT OF BLOOD PRESSURE (BP) IS READILY AVAILABLE IN CLINICAL SETTINGS SUCH AS INTENSIVE-CARE MONITORING OF CRITICALLY ill infants and adults, where it is usually obtained invasively via arterial catheterization. However, continuous measurement of BP is also desirable for the investigation of the cardiovascular system and its autonomic control during sleep, in which arterial catheter-ization is problematic. In particular, there has been growing inter-est in studies related to obstructive sleep apnea syndrome1,2 and the sudden infant death syndrome,3-5 in which altered autonomic BP control during sleep may play a crucial role in the pathophysi-ology of such syndromes. Autonomic control of the cardiovascular system relies on the baroreflex mechanism for the short-term control of BP variations. Baroreflex sensitivity (BRS) is commonly assessed by power spectral analysis6 or spontaneous sequence analysis,7-9 with both approaches being dependent on accurate, beat-to-beat BP mea-surement. Because of ethical constraints on arterial catheteriza-tion in infants for research purposes, previous assessments of

BRS during sleep have relied on noninvasive techniques, such as oscillometry, that allow only intermittent measurements.10-12

A new noninvasive device, the FinapresTM (TNO, The Nether-lands) and its portable version, the PortapresTM (FMS, Finapres Medical Systems BV, The Netherlands), have become available for the continuous measurement of BP in adults. The FinapresTM operates via a photoplethysmographic cuff designed for the adult finger and utilizes the volume clamp method of Penaz13 for BP determination. Several studies have assessed the use of the Fi-napresTM in preterm infants, with application of the adult-sized finger cuff to the infant’s wrist,14-16 but, as yet, validation of this approach is not complete. Of these studies, 2 reported a good agreement between intraarterial catheters and FinapresTM/Porta-presTM in measurements of systolic (SAP) and diastolic (DAP) arterial pressure.14,15 However, in the most recent study, Andries-sen et al16 concluded that the FinapresTM provided absolute BP measurements of limited accuracy, though the device accurately measured beat-to-beat changes in BP. In addition to disagreement concerns relating to absolute accuracy, these previous studies have been limited to a single measurement in each infant, so that the reproducibility (precision) of repeated measurements within a subject is unknown. Furthermore, although the FinapresTM has been used to determine BRS,17 no comparison with arterial-cath-eter BRS measurements and no assessment of sleep state has been made. Finally, there is currently no information on the mea-surement accuracy of the FinapresTM for mean arterial pressure (MAP), the value most commonly used in the clinical setting. Recently, the FinometerTM (FMS, Finapres Medical Systems, The Netherlands), the successor of the FinapresTM, has become commercially available. The FinometerTM offers potential utility for recording beat-to-beat BP and contains an algorithm to recon-struct brachial artery pressure waveforms, but, as yet, this has not

Validation of a New Noninvasive Method to Measure Blood Pressure and Assess Baroreflex Sensitivity in Preterm Infants During SleepStephanie R. Yiallourou, BSc (Hons); Adrian M. Walker, PhD; Rosemary S.C. Horne, PhD

Ritchie Centre for Baby Health Research, Monash Institute of Medical Research, Monash University, Melbourne, Victoria, Australia

Disclosure StatementThis was not an industry supported study. Drs. Horne, Yiallourou, and Walker have indicated no financial conflicts of interest.

Submitted for publication January 2006Accepted for publication March 2006Address correspondence to: Rosemary SC Horne, PhD, Ritchie Centre for Baby Health Research, Level 5, Monash Medical Centre, 246 Clayton Rd, Clayton, Victoria, Australia 3168; Tel: 61 3 9594 5100; Fax: 61 3 9594 6811; E-mail: rosemary.horne@med.monash.edu.au

Study Objectives: Accuracy and precision of a noninvasive device for continuously measuring blood pressure (BP) (FinometerTM, FMS, The Netherlands) during sleep was assessed in preterm infants. Design: Absolute BP beat-to-beat values, interbeat changes, measure-ment precision, and baroreflex sensitivity were compared with BP mea-surements from intraarterial catheters. Participants: Ten preterm infants (gestational age 27-36 weeks; birth weight 964-2620 gm) were studied in the neonatal intensive care unit.Measurements and Results: The 2 modes of BP measurement were compared in 2-minute epochs (n = 10-12/infant). Mean arterial pressure, systolic arterial pressure, and diastolic arterial pressure were analyzed beat to beat, and baroreflex sensitivity was assessed using spontaneous sequence analysis. Mean differences for absolute BP (mm Hg) were as follows: mean arterial pressure, 3 (limits of agreement, -1 to 8); systolic arterial pressure, -4 (-8 to 1); and diastolic arterial pressure, 7 (4 to 10). Mean differences and limits of agreement for interbeat changes were es-sentially 0 for mean arterial pressure, systolic arterial pressure, and dia-

stolic arterial pressure. Precision (± 95% confidence intervals, mm Hg) for the FinometerTM were mean arterial pressure ± 7, systolic arterial pres-sure ± 8, and diastolic arterial pressure ± 6. Precision was greater for the arterial catheter (mean arterial pressure ± 3, systolic arterial pressure ± 4, and diastolic arterial pressure ± 4). Baroreflex sensitivity calculated from the FinometerTM BP was (mean ± SEM, ms/mm Hg) 1.74 ± 0.23 and, from the catheter system, BP was 1.56 ± 0.21 (p value NS). Conclusions: The FinometerTM provides accurate measurements of beat-to-beat BP and baroreflex sensitivity. The ability to continuously measure BP and baroreflex sensitivity during sleep in infants may provide vital clues into pathologic conditions associated with impaired autonomic control during sleep. Keywords: Blood pressure, baroreflex, sleep, cardiovascular control Citation: Yiallourou SR; Walker AM; Horne RSC. Validation of a new non-invasive method to measure blood pressure and assess baroreflex sensi-tivity in preterm infants during sleep. SLEEP 2006;29(8):1083-1088.

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al

PROVIGIL® (modafini l) TABLETS [C-IV]BRIEF SUMMARY: Consult Package Insert for Complete PrescribingInformationCONTRAINDICATIONS: Known hypersensitivity to PROVIGIL or its inactiveingredients.WARNINGS: Patients with abnormal levels of sleepiness who take PROVIGILshould be advised that their level of wakefulness may not return to normal.Patients with excessive sleepiness, including those taking PROVIGIL, shouldbe frequently reassessed for their degree of sleepiness and, if appropriate,advised to avoid driving or any other potentially dangerous activity.Prescribers should also be aware that patients may not acknowledgesleepiness or drowsiness until directly questioned about drowsiness orsleepiness during specific activities.PRECAUTIONS: Diagnosis of Sleep Disorders: PROVIGIL should be usedonly in patients who have had a complete evaluation of their excessivesleepiness, and in whom a diagnosis of either narcolepsy, OSAHS, and/orSWSD has been made in accordance with ICSD or DSM diagnostic criteria.Such an evaluation usually consists of a complete history and physical exami-nation, and it may be supplemented with testing in a laboratory setting. CPAPUse in Patients with OSAHS: In OSAHS, PROVIGIL is indicated as an adjunctto standard treatment(s) for the underlying obstruction. If continuous positiveairway pressure (CPAP) is the treatment of choice for a patient, a maximal effortto treat with CPAP for an adequate period of time should be made prior to initiating PROVIGIL. If PROVIGIL is used adjunctively with CPAP, the encour-agement of and periodic assessment of CPAP compliance is necessary.General: Patients should be cautioned about operating an automobile or otherhazardous machinery until they are reasonably certain that PROVIGIL therapywill not adversely affect their ability to engage in such activities. PatientsUsing Contraceptives: The effectiveness of steroidal contraceptives may bereduced when used with PROVIGIL and for one month after discontinuation.Alternative or concomitant methods of contraception are recommendedduring and for one month after discontinuation of PROVIGIL. CardiovascularSystem: In clinical studies of PROVIGIL, signs and symptoms including chestpain, palpitations, dyspnea and transient ischemic T-wave changes on ECGwere observed in three subjects in association with mitral valve prolapse orleft ventricular hypertrophy. It is recommended that PROVIGIL tablets not beused in patients with a history of left ventricular hypertrophy or in patientswith mitral valve prolapse who have experienced the mitral valve prolapsesyndrome when previously receiving CNS stimulants. Such signs may includebut are not limited to ischemic ECG changes, chest pain, or arrhythmia.Patients with a recent history of MI or unstable angina should be treated withcaution. Blood pressure monitoring in short-term controlled trials showed noclinically significant changes in mean systolic and diastolic blood pressure inpatients receiving PROVIGIL as compared to placebo. However, a greaterproportion of patients on PROVIGIL required new or increased use of anti-hypertensive medications (2.4%) compared to patients on placebo (0.7%).The differential use was slightly larger when only studies in OSAHS wereincluded, with 3.4% of patients on PROVIGIL and 1.1% of patients on placeborequiring such alterations in the use of antihypertensive medication. Increasedmonitoring of blood pressure may be appropriate in patients on PROVIGIL.Central Nervous System: There have been reports of psychotic episodesassociated with PROVIGIL use. One healthy male volunteer developed ideas ofreference, paranoid delusions, and auditory hallucinations in association withmultiple daily 600 mg doses of PROVIGIL and sleep deprivation. There was noevidence of psychosis 36 hours after drug discontinuation. Caution should beexercised when PROVIGIL is given to patients with a history of psychosis.Patients with Severe Renal Impairment: Treatment with PROVIGIL resultedin much higher exposure to its inactive metabolite, modafinil acid, but notPROVIGIL itself. Patients with Severe Hepatic Impairment: PROVIGILshould be administered at a reduced dose because its clearance is decreased.Elderly Patients: Elderly patients may have diminished renal and/or hepaticfunction; therefore, dosage reduction should be considered. Information forPatients: Physicians are advised to discuss the following with patients takingPROVIGIL. PROVIGIL is indicated for patients who have abnormal levels ofsleepiness. PROVIGIL has been shown to improve, but not eliminate thisabnormal tendency to fall asleep. Therefore, patients should not alter theirprevious behavior with regard to potentially dangerous activities (eg, driving,operating machinery) or other activities requiring appropriate levels ofwakefulness, until and unless treatment with PROVIGIL has been shown toproduce levels of wakefulness that permit such activities. Patients should beadvised that PROVIGIL is not a replacement for sleep. Patients should beinformed that it may be critical that they continue to take their previouslyprescribed treatments (eg, patients with OSAHS receiving CPAP shouldcontinue to do so). Patients should be informed of the availability of a patientinformation leaflet, and they should be instructed to read the leaflet prior totaking PROVIGIL. Pregnancy: Patients should notify their physician if theybecome pregnant or intend to become pregnant during therapy. They shouldbe cautioned of the potential increased risk of pregnancy when using steroidalcontraceptives (including depot or implantable contraceptives) with PROVIGILand for one month after discontinuation of therapy. Nursing: Patients shouldnotify their physician if they are breast feeding. Concomitant Medication:Patients should inform their physician if they are taking or plan to take anyprescription or over-the-counter drugs, because of the potential for drug inter-actions. Alcohol: It is prudent to avoid alcohol while taking PROVIGIL.Allergic Reactions: Patients should notify their physician if they develop arash, hives, or a related allergic phenomenon. Drug Interactions: CNS ActiveDrugs: In a single-dose study, simultaneous administration of PROVIGIL 200 mg with methylphenidate 40 mg delayed the absorption of PROVIGIL byapproximately one hour. In a single-dose study, simultaneous administrationof PROVIGIL 200 mg with dextroamphetamine 10 mg delayed absorption ofPROVIGIL by approximately one hour. Coadministration of a single dose ofclomipramine 50 mg with PROVIGIL 200 mg/day did not affect the pharmaco-kinetics of either drug. One incident of increased levels of clomipramine andits active metabolite desmethylclomipramine has been reported. In the druginteraction study between PROVIGIL and ethinyl estradiol (EE2), on the samedays as those for the plasma sampling for EE2 pharmacokinetics, a single doseof triazolam 0.125 mg was also administered. Mean Cmax and AUC0-� oftriazolam were decreased by 42% and 59%, respectively, and its eliminationhalf-life was decreased by approximately an hour after the modafinil

treatment. In the absence of interaction studies with monoamine oxidase(MOA) inhibitors, caution should be exercised. Other Drugs: No significantchanges in the pharmacokinetics of warfarin occurred in healthy subjectsgiven one dose of warfarin 5 mg following chronic administration ofPROVIGIL. However, more frequent monitoring of prothrombin times/INR isadvised when PROVIGIL is coadministered with warfarin. PROVIGIL oncedaily 200 mg/day for 7 days followed by 400 mg/day for 21 days decreasedethinyl estradiol Cmax and AUC0-24 by a mean 11% and 18% with no apparentchange in the elimination rate. One interaction between PROVIGIL andcyclosporine has been reported in a 41-year-old female. After one month ofPROVIGIL 200 mg/day, cyclosporine blood levels decreased by 50%. Dosageadjustment for cyclosporine may be needed. Potential Interactions withDrugs That Inhibit, Induce, or are Metabolized by Cytochrome P-450Isoenzymes and Other Hepatic Enzymes: In primary human hepatocytes,PROVIGIL slightly induced CYP1A2, CYP2B6 and CYP3A4 in a dose-dependent manner. In vitro experiments do not necessarily predict responsein vivo; caution should be exercised when PROVIGIL is coadministered withdrugs that are metabolized by enzymes. In human hepatocytes, PROVIGILproduced a dose-related suppression of CYP2C9 activity suggesting apotential for metabolic interaction between PROVIGIL and substrates of thisenzyme (eg, S-warfarin and phenytoin). In healthy volunteers, chronicPROVIGIL treatment had no significant effect on single-dose pharmaco-kinetics of warfarin vs placebo. In human liver microsomes, PROVIGIL andmodafinil sulfone reversibly inhibited CYP2C19. Both compounds combinedcould produce sustained partial enzyme inhibition. Drugs largely eliminatedvia CYP2C19 metabolism, such as diazepam, propranolol, phenytoin (also viaCYP2C9) or S-mephenytoin may have prolonged elimination with PROVIGILcoadministration and may require dose reduction and monitoring for toxicity.CYP2C19 provides ancillary metabolism of some tricyclic antidepressants (eg,clomipramine and desipramine) primarily metabolized by CYP2D6. In tricyclicusers deficient in CYP2D6, CYP2C19 metabolism may be substantiallyincreased. PROVIGIL may elevate tricyclics in this patient subset. A reductionin tricyclic dose may be needed. Due to partial involvement of CYP3A4 elimination of PROVIGIL, coadministration of potent inducers of CYP3A4 (eg, carbamazepine, phenobarbital, rifampin) or inhibitors of CYP3A4 (eg, ketoconazole, itraconazole) could alter modafinil plasma levels.Carcinogenesis, Mutagenesis, Impairment of Fertility: Carcinogenesis:The highest dose studied in carcinogenesis studies represent 1.5 times(mouse) or 3 times (rat) the maximum human daily dose of 200 mg on amg/m2 basis. There was no evidence of tumorigenesis associated withPROVIGIL administration in these studies, but because the mouse study usedan inadequate high dose below that representative of a maximum tolerated

dose, the carcinogenic potential in that species has not been fully evaluated.Mutagenesis: There was no evidence of mutagenic or clastogenic potential ofPROVIGIL. Impairment of Fertility: PROVIGIL was administered orally tomale and female rats prior to and throughout mating and gestation at up to 23 times the recommended human dose of 200 mg/day on a mg/m2 basis withno effect on fertility. Pregnancy: Pregnancy Category C: PROVIGIL adminis-tered orally to pregnant rats throughout the period of organogenesis caused,in the absence of maternal toxicity, an increase in resorptions and an increasedincidence of hydronephrosis and skeletal variations in the offspring at a doseof 200 mg/kg/day (10 times the recommended human dose of 200 mg/day ona mg/m2 basis) but not at 100 mg/kg/day. However, in a subsequent study ofup to 480 mg/kg/day (23 times the recommended human dose on a mg/m2

basis), which included maternally toxic doses, no adverse effects on embryo-fetal development were seen. PROVIGIL administered orally to pregnantrabbits throughout the period of organogenesis at doses up to 100 mg/kg/day(10 times the recommended human dose on a mg/m2 basis) had no effects onembryofetal development. However, in a subsequent study in pregnantrabbits, increased resorptions, and increased alterations in fetuses from asingle litter (open eye lids, fused digits, rotated limbs), were observed at 180 mg/kg/day (17 times the recommended human dose on a mg/m2 basis),a dose that was also maternally toxic. PROVIGIL administered orally to ratsthroughout gestation and lactation at doses up to 200 mg/kg/day (10 times therecommended human dose on a mg/m2 basis), had no effects on the postnataldevelopment of the offspring. There are no adequate and well-controlledstudies in pregnant women. PROVIGIL should be used during pregnancy onlyif the potential benefit justifies the potential risk to the fetus. Labor andDelivery: The effect of PROVIGIL on labor and delivery in humans has notbeen systematically investigated. Nursing Mothers: It is not known whetherPROVIGIL or its metabolites are excreted in human milk. Caution should beexercised when PROVIGIL is administered to a nursing woman.PEDIATRIC USE: Safety and effectiveness in individuals below 16 years of agehave not been established. Leukopenia has been reported in pediatric patientstaking PROVIGIL.GERIATRIC USE: Safety and effectiveness in individuals above 65 years of agehave not been established.ADVERSE REACTIONS: PROVIGIL has been evaluated for safety in over 3500patients, of whom more than 2000 patients with excessive sleepinessassociated with primary disorders of sleep and wakefulness were given atleast one dose of PROVIGIL. In clinical trials, PROVIGIL has been found to begenerally well tolerated and most adverse experiences were mild to moderate.The most commonly observed adverse events (�5%) associated with the useof PROVIGIL more frequently than placebo-treated patients in the placebo-controlled clinical studies in primary disorders of sleep and wakefulness wereheadache, nausea, nervousness, rhinitis, diarrhea, back pain, anxiety,insomnia, dizziness, and dyspepsia. In the placebo-controlled clinical trials,8% of the 934 patients who received PROVIGIL discontinued due to anadverse experience. The most frequent reasons for discontinuation thatoccurred at a higher rate for PROVIGIL than placebo patients were headache

(2%), nausea, anxiety, dizziness, insomnia, chest pain, and nervousness (each<1%). The incidence of adverse experiences that occurred at a rate of �1%and were more frequent in patients treated with PROVIGIL than in placebopatients in the principal trials are listed below. Consult full prescribing infor-mation on adverse events. Body as a Whole: Headache, back pain, flusyndrome, chest pain, chills, neck rigidity Cardiovascular: Hypertension,tachycardia, palpitation, vasodilatation Digestive: Nausea, diarrhea,dyspepsia, dry mouth, anorexia, constipation, abnormal liver function, flatu-lence, mouth ulceration, thirst Hemic/Lymphatic: Eosinophilia Metabolic/Nutritional: Edema Nervous: Nervousness, insomnia, anxiety, dizziness,depression, paresthesia, somnolence, hypertonia, dyskinesia, hyperkinesia,agitation, confusion, tremor, emotional lability, vertigo Respiratory: Rhinitis,pharyngitis, lung disorder, epistaxis, asthma Skin/Appendages: Sweating,herpes simplex Special Senses: Amblyopia, abnormal vision, tasteperversion, eye pain Urogenital: Urine abnormality, hematuria, pyuria DoseDependency: In the placebo-controlled clinical trials the only adverse eventsthat were clearly dose related were headache and anxiety. Vital SignChanges: While there was no consistent change in mean values of heart rateor systolic and diastolic blood pressure, the requirement for antihypertensivemedication was slightly greater in patients on PROVIGIL compared to placebo.Weight Changes: There were no clinically significant differences in bodyweight change in patients treated with PROVIGIL compared to placebo-treatedpatients. Laboratory Changes: Mean plasma levels of gamma glutamyltrans-ferase (GGT) and alkaline phosphatase (AP) were higher followingadministration of PROVIGIL, but not placebo. Few subjects, however, had GGTor AP elevations outside of the normal range. Shifts to higher, but not clinicallysignificantly abnormal, GGT and AP values appeared to increase with time onPROVIGIL. No differences were apparent in alanine aminotransferase,aspartate aminotransferase, total protein, albumin, or total bilirubin. ECGChanges: No treatment-emergent pattern of ECG abnormalities was found inplacebo-controlled clinical trials following administration of PROVIGIL.Postmarketing Reporting: The following adverse reactions have beenidentified during post-approval use of PROVIGIL. Because these reactions arereported voluntarily from a population of uncertain size, it is not alwayspossible to reliably estimate their frequency or establish a causal relationshipto drug exposure. Decisions to include these reactions in labeling are typicallybased on one or more of the following factors: (1) seriousness of the reaction,(2) frequency of the reporting, or (3) strength of causal connection toPROVIGIL. Central Nervous System: Symptoms of psychosis, symptoms ofmania Dermatologic: Rare reports of serious skin reactions (includingsuspected cases of both erythema multiforme and Stevens-Johnsonsyndrome) Hematologic: Agranulocytosis Hypersensitivity: Urticaria (hives),angioedemaDRUG ABUSE AND DEPENDENCE: Abuse Potential and Dependence: Inaddition to its wakefulness-promoting effect and increased locomotor activityin animals, in humans, PROVIGIL produces psychoactive and euphoriceffects, alterations in mood, perception, thinking and feelings typical of otherCNS stimulants. In vitro, PROVIGIL binds to the dopamine reuptake site andcauses an increase in extracellular dopamine, but no increase in dopaminerelease. PROVIGIL is reinforcing, as evidenced by its self-administration in monkeys previously trained to self-administer cocaine. In some studies,PROVIGIL was also partially discriminated as stimulant-like. Physiciansshould follow patients closely, especially those with a history of drug and/or stimulant (eg, methylphenidate, amphetamine, or cocaine) abuse. Inindividuals experienced with drugs of abuse, PROVIGIL producedpsychoactive and euphoric effects and feelings consistent with otherscheduled CNS stimulants (methylphenidate). Patients should be observed forsigns of misuse or abuse. Withdrawal: Following 9 weeks of PROVIGIL use inone US clinical trial, no specific symptoms of withdrawal were observedduring 14 days of observation, although sleepiness returned in narcolepticpatients.OVERDOSAGE: Human Experience: In clinical trials, a total of 151 protocol-specified doses �1000 mg/day (5 to 8 times the recommended daily dose of200 mg) have been administered to 32 subjects, including 13 subjects whoreceived doses of 1000 or 1200 mg/day for 7 to 21 consecutive days. Inaddition, several intentional acute overdoses occurred; the two largest being4500 mg and 4000 mg taken by two subjects participating in foreigndepression studies. None of these study subjects experienced any unexpectedor life-threatening effects. Adverse experiences that were reported at thesedoses included excitation or agitation, insomnia, and slight or moderate eleva-tions in hemodynamic parameters. Other observed high-dose effects inclinical studies have included anxiety, irritability, aggressiveness, confusion,nervousness, tremor, palpitations, sleep disturbances, nausea, diarrhea, anddecreased prothrombin time. From post-marketing experience, there havebeen no reports of fatal overdoses involving PROVIGIL alone (doses up to 12 grams). Overdoses involving multiple drugs, including PROVIGIL, haveresulted in fatal outcomes. Symptoms most often accompanying PROVIGILoverdose, alone or in combination with other drugs have included insomnia,restlessness, disorientation, confusion, excitation, hallucination, nausea,diarrhea, tachycardia, bradycardia, hypertension, and chest pain. Cases ofaccidental ingestion/overdose have been reported in children as young as 11 months of age. The highest reported accidental ingestion on a mg/kg basisoccurred in a three-year-old boy who ingested 800-1000 mg (50-63 mg/kg) ofPROVIGIL. The child remained stable. The symptoms associated withoverdose in children were similar to those observed in adults. OverdoseManagement: No specific antidote to the toxic effects of PROVIGIL overdosehas been identified. Overdoses should be managed with primarily supportivecare, including cardiovascular monitoring. Emesis or gastric lavage should beconsidered. There are no data to suggest the utility of dialysis or urinary acidi-fication or alkalinization in enhancing drug elimination. The physician shouldconsider contacting a poison-control center on the treatment of any overdose.

Manufactured for: Cephalon, Inc., West Chester, PA 19380

For more information about PROVIGIL, please call Cephalon ProfessionalServices at 1-800-896-5855 or visit our Website at www.PROVIGIL.com

© 2004 Cephalon, Inc.PRO544 12/2004

All rights reserved. Printed in USA.

SLEEP, Vol. 29, No. 8, 2006 1084

been validated in infants. Using arterial catheter measurements as the gold standard, the aims of this study were to measure BP during sleep using the FinometerTM and assess its capability in a number of key aspects: (1) accuracy of the absolute values and interbeat values of MAP, SAP, and DAP; (2) reproducibility of repeated measurements (precision) within a subject; and 3) ac-curacy in estimating BRS using spontaneous sequence analysis.

METHODS

The protocol for this study was approved by the Southern Health and Monash University Human Research Ethics Committees. Written parental consent was obtained for all subjects prior to commencement of the study. Ten preterm infants (6 girls/4 boys) born at 27 to 36 weeks gestational age (mean 31 ± 1 weeks), with birth weights between 964 and 2620 gm (mean 1624 ± 215 gm) and Apgar scores of 2 to 8 (median 5) at 1 minute and 5 to 9 (median 8) at 5 minutes were recruited from the Neonatal Intensive Care Unit, Monash Medical Centre. Infants were studied 1 to 4 days postnatally. All infants were implanted with arterial catheters for intensive care monitoring prior to enrollment. Only infants approximately 1000 gm or larger were included in the selection criteria, because the BP cuff could not be applied appropriately in smaller infants.

Noninvasive BP Measurement (FinometerTM)

BP was measured with the Finometer TM cuff placed around the infant’s wrist. In each subject, an appropriate-sized cuff (small, medium, or large) was selected, and 10 to 12 BP measurements were performed, each of 2 minutes duration, with a 2-minute rest period between successive measurements. The cyclic approach of 2 minutes (off-on) was chosen to avoid venous pooling in the infant’s hand. Following the initial start-up calibration, the auto-matic calibration (Physiocal) was switched off to ensure an un-interrupted recording. All BP measurements were referenced to heart level via the built-in height-correction system. Infant sleep states were scored as active sleep, quiet sleep, or indeterminate sleep using behavioral observations based on crite-ria described for preterm infants.18

Invasive BP Measurement (AgilentTM – Catheter system)

BP was also measured directly via a preexisting 3.5 Fr or 5 Fr intraarterial catheter inserted into the umbilical (n = 9) or ul-nar artery (n = 1). Decisions on inserting catheters, as well as the connection, zeroing, calibration, and maintenance were made independently by clinical staff for the care of the infant. To pre-vent clotting, the catheter was routinely flushed with heparinized saline. The catheter was connected to a pressure transducer and referenced to heart level with a fluid-filled manometer tube.

Physiologic Recordings

Physiologic variables measured for routine clinical monitoring of vital functions during intensive care were recorded and dis-played on a patient-monitoring device (AgilentTM monitor, Agilent Technologies, MA). Measurements included BP (3.5-5 Fr cathe-ter), electrocardiogram, and respiration (Kendall KittycatTM Foam Prewired Neonatal/Pediatric Monitoring Electrodes, The Ludlow Company, Chicopee, MA) and oxygen saturation (Dolphin TM, 2000 Oximetry Sensors, OSS Medical, Singapore). Signals, in-

cluding FinometerTM BP measurements, were digitized using a 16-channel Powerlab system (ADInstruments, Sydney, Australia) at a sampling frequency of 400 Hz and stored on a personal com-puter running a specialized program for data storage, analysis, and signal display (Chart 5.0, ADInstruments).

Data Analysis

Absolute BP Beat-to-Beat Values

Absolute beat-to-beat values of MAP, SAP, and DAP were ob-tained and matched by peak detection using Chart 5.0 software (ADInstruments) for both the FinometerTM (MAPf, SAPf, and DAPf) and catheter-system (MAPc, SAPc, and DAPc) BP mea-surements. Any sections within the 2-minute epoch containing movement artifact were excluded from further analysis. The dif-ference between the catheter-system and the FinometerTM MAP, SAP, and DAP measurements were calculated for each beat, and a mean value was calculated for each 2-minute recording for each subject. Mean values were calculated for each subject; data was tested for normality using Kolmogorov-Smirnov test and com-pared using Bland-Altman analysis.19

Interbeat Differences

Inter-beat differences between the FinometerTM and catheter system were compared by calculating the change in BP from one beat to the next and determining the difference between both sys-tems for each beat. Values were compared using Bland-Altman analysis19 for each subject.

Measurement Precision

Precision (reproducibility) was estimated using the method of Youden20 as the 95% confidence interval (CI) of a single esti-mate: CI = 2S And,

2 (n-1) S = Σd2

Where, S equals the standard deviation of single estimate, d equals the difference of extremes of replicate measurements, and n is the number of replicates. The number of replicate measurements (n) for the FinometerTM required to obtain an estimate with precision equal to that of the catheter system was calculated as:

Sc n = Sf

Where Sf is the standard deviation of a single estimate for the FinometerTM and Sc is the standard deviation of a single estimate for the catheter system.

BRS Assessment

BRS was assessed using spontaneous sequence analysis.7-9 Paired MAP and heart period (obtained from the electrocardiogram signal) sequences characterized by an increase or decrease of at

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al

SLEEP, Vol. 29, No. 8, 2006 1085

differed by -4 mm Hg, with limits of agreement of -8 to 1 mm Hg (Figure 2b). DAP differed by 7 mm Hg, with limits of agreement of 4 to 10 mm Hg (Figure 2c).

Interbeat BP

Mean differences and limits of agreement for the interbeat changes recorded by the 2 systems were essentially 0 (< 0 .01 mm Hg) for each of MAP, SAP, and DAP.

least 1 mm Hg during 3 or more consecutive beats were identified during each 2-minute epoch. BRS was calculated as the slope of the linear regression of MAP and the proceeding heart period for each sequence and averaged for each 2-minute epoch. Baroreflex sequences were defined by changes in MAP and heart period proceeding in the same direction, ie, sequences having positive slopes. A mean BRS was calculated for each infant for both FinometerTM and catheter-system MAP measurements. A paired Student's t test was used to compare BRS estimates both within each infant and for pooled subject means.

RESULTS

A typical example of simultaneous BP signals recorded by the catheter and FinometerTM systems is presented in Figure 1. Of the 10 infants studied, 1 infant was excluded (infant 3) because the mean catheter-system BP measurements for this infant were more than 2 SDs above the pooled subject mean, suggesting a zeroing error in this system. A total of 101 2-minute epochs of simultaneous FinometerTM

and catheter-system BP measurements were made. Of these, 58 epochs were analyzed free of movement artifacts, 55 epochs were recorded in active sleep, 1 in quiet sleep, and 2 in indeterminate sleep. Measurements in different sleep states were combined, and a total 11,637 paired beats were analyzed. BRS was assessed in 8 of 10 infants because 1 infant (infant 5) was excluded from BRS assessment related to a poor electrocardiogram signal.

Absolute BP

The ranges of BP (mm Hg) measured by the catheter system (c) and the FinometerTM (f) were 34 to 46 (MAPc) versus 39 to 47 (MAPf), 45 to 62 (SAPc) versus 46 to 61 (SAPf), and 23 to 43 (DAPc) versus 31 to 45 (DAPf). Mean differences (± SEM) and limits of agreement for MAP, SAP, and DAP are presented in Tables 1, 2, and 3, respectively. MAP estimates corresponded most closely with a difference between measurements averaging 3 mm Hg and close limits of agreement (-1 to 8) (Figure 2a). SAP

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al23

Catheter

Finometer

Figure 1

20

40

60

80

1 2 3 4 5

Times (s)

Blo

od P

ress

ure

(mm

Hg)

Blo

od P

ress

ure

(mm

Hg)

80

60

40

20

Time (s)1 2 3 4 5

20

40

60

80

1 2 3 4 5

Times (s)

Blo

od P

ress

ure

(mm

Hg)

20

40

60

80

1 2 3 4 5

Times (s)

Blo

od P

ress

ure

(mm

Hg)

Blo

od P

ress

ure

(mm

Hg)

80

60

40

20

Time (s)1 2 3 4 5

Figure 1—Comparison of FinometerTM and catheter-system arterial blood pressure recordings in a preterm infant in active sleep. Note that the FinometerTM tracks the arterial waveform closely, with a small underestimation of pulse pressure compared with the catheter system. Separation of the signals by approximately a half beat delay is due to an intrinsic delay in the catheter measurement system.

Figure 2a

Figure 2c

-20-15-10-505101520

30 35 40 45 50 55 60Mean (MAPc + MAPf)

)gH

mm( cP

AM - fP

AM

-20-15-10-505101520

40 45 50 55 60 65 70

Mean (SAPc + SAPf)

gH

mm( cP

AS - fPA S

)

-20

-15

-10

-5

0

5

10

15

20

20 25 30 35 40 45 50Mean (DAPc + DAPf)

)gH

mm( cP

AD- fP

AD

Figure 2b

Figure 2—Bland-Altman plots displaying mean differences for indi-vidual subjects (n = 9) between FinometerTM (f) and catheter-system (c) blood pressure measurement calculated on a beat-to-beat basis for (2a) mean arterial pressure (MAP), (2b) systolic arterial pressure (SAP), and (2c) diastolic arterial pressure (DAP). The solid line rep-resents the mean difference for all subjects between both methods of measurements, and the dashed lines represents 95% limits of agree-ment (± 2 SD).

SLEEP, Vol. 29, No. 8, 2006 1086

Precision of BP Measurements

The precision (95% CI, mm Hg) for each of the measurements was MAPc (-1 to 4) versus MAPf (-3 to 10), SAPc (-2 to 6) versus SAPf (-4 to 11), and DAPc (-2 to 5) versus DAPf (-3 to 9). Numbers of replicates required for the FinometerTM to measure with a precision equal to that of the catheter were 5, 3, and 3 for MAP, SAP and DAP, respectively.

Baroreflex Sensitivity

BRS data are presented in Table 4. There was no difference between the mean BRS calculated for the FinometerTM (1.74 ± 0.23) and the catheter system (1.56 ± 0.21, p = 0.2). The number of ramps identified to calculate BRS estimates in the catheter system (n = 713) was not different than that in the FinometerTM (n = 679).

DISCUSSION

Our study has demonstrated that the FinometerTM is a useful tool that can measure BP accurately and with high precision in preterm infants during sleep using the adult finger cuff placed around the wrist of the infant. Our findings support those of earlier workers,14,15 who found close agreement between FinometerTM and catheter estimates of SAP and DAP. Our data extend these validations to include MAP estimates. In addition, we have demonstrated that the FinometerTM accurately reconstructs the beat-to-beat BP profile, allowing its application to assess interbeat changes in BP and BRS estimations in the time domain. This is also the first study to assess the precision of single measurement

versus repeated measurements of BP within an infant, and we have calculated the number of replicate measurements of BP to match the precision of the FinometerTM to that of the AgilentTM catheter system. The FinometerTM closely determined MAP, with the mean difference between arterial catheter and FinometerTM measurements being only 3 mm Hg, a difference not considered to be clinically significant in preterm infants with BP typically in the range of 25 to 45 mmHg. This is the first study to compare MAP values between a catheter system and the FinometerTM, which is important because, in long-term BP catheter recordings, MAP is least affected by loss of catheter patency and is more reliable than systolic and diastolic pressures.21 Absolute BP values of the FinometerTM underestimated SAP by a mean of -4 mm Hg and overestimated DAP by a mean of 7 mm Hg, indicating that the FinometerTM underestimates pulse pressure, although MAP accuracy is maintained. Our findings are similar to those of Harrington et al,15 who analysed 5-minute epochs of BP in term neonates in intensive care and also found that the PortapresTM underestimated SAP (-2 ± 4 mm Hg) and overestimated DAP (4 ± 4 mm Hg). In contrast with our finding and those of Drouin et al14 and Harrington et al,15 Andriessen et al16 concluded that the FinapresTM was of limited clinical value for estimating individual BP values because the difference from catheter values for SAP was as much as -6.5 or +17.6 mmHg and, for DAP, was -5.6 or +17.9 mm Hg in individual subjects. In this study,16 the widely disparate values were for just 1 infant for SAP and for a different infant for DAP; when the data are averaged for the group, the difference for SAP is 2.2 mm Hg and DAP is 2.6 mm Hg, values similar to ours and to those of previous studies.14,15 In our study,

Table 1—Mean MAP Values Measured by the Catheter System and the FinometerTM for 9 Subjects

Subject MAPc MAPf DMAP Limits of agreement 1 40 47 7 0 to 14 2 37 42 5 2 to 7 4 42 44 2 -1 to 4 5 39 45 6 3 to 9 6 42 47 5 -1 to 10 7 48 45 -3 -6 to 8 8 40 40 0 -1 to 2 9 34 40 6 1 to 10 10 36 39 3 -3 to 8 Mean ± SEM 40 ± 1 43 ± 1 3 ± 1 -1 to 8

Mean arterial pressure (MAP) values, in mm Hg, were measured by the catheter system (MAPc) and the FinometerTM (MAPf). DMAP represents the mean difference and limits of agreement between each method of measurement.

Table 2— Mean SAP Values Measured by the Catheter System and the FinometerTM for 9 Subjects Subject SAPc SAPf DSAP Limits of agreement 1 62 57 -5 -16 to 5 2 45 47 2 -1 to 5 4 49 49 0 -4 to 3 5 50 49 -1 5 to 3 6 61 51 -10 -16 to -3 7 64 48 -16 -23 to -9 8 60 58 -2 -4 to 0 9 56 60 4 -2 to 10 10 55 46 -9 -16 to -2 Mean ± SEM 55 ± 2 52 ± 2 -4 ± 2 -8 to 1

Systolic arterial pressure (SAP) values, in mm Hg, were measured by the catheter system (MAPc) and the FinometerTM (MAPf). DSAP represents the mean difference and limits of agreement between each method of measurement.

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al

SLEEP, Vol. 29, No. 8, 2006 1087

the maximum individual differences for SAP and DAP were similar to those of Andriessen et al,16 but the differences for MAP were much less (-1 to 7, Table 1). The FinometerTM tracked interbeat changes in MAP, SAP, and DAP with high agreement, compared with the arterial catheter, with mean differences of 0 and limits of agreement also 0. In support of our data, Andriessen et al16 achieved similarly accurate beat-to-beat changes for SAP and DAP. Importantly, that the FinometerTM

can accurately track beat-to-beat changes in BP supports the use of this new technology in detecting rapid changes in BP, which are essential when assessing autonomic cardiovascular control. Furthermore, our study has shown that BRS in the time domain can be accurately assessed from FinometerTM BP measurements, because there was no difference compared with BRS derived from arterial-catheter measurements. The BRS values we obtained from spontaneous sequence analysis (arterial catheter: 1.56 ± 0.21 ms/mm Hg; FinometerTM: 1.74 ± 0.23 ms/mm Hg) are similar to those previously reported for preterm infants at similar gestational ages (1.80 ± 0.14 ms/mm Hg) using the technique of spontaneous sequence analysis.22 A novel approach of our study was to estimate the precision of each method by calculating the 95% CI for a single estimate. The precision of the catheter system was somewhat greater than the FinometerTM because the 95% CIs were less; for example, for MAP, the 95% CI was -1 to 4 mm Hg versus -3 to 10 mm Hg, respectively. Importantly, using the precision calculation described by Youden,20 we have shown that, with 5 or fewer replicate measurements, the FinometerTM precision is equal to that of the catheter system. The source of greater precision of the catheter system is uncertain, but it may represent a greater intrinsic

stability of the measurement system or a lesser sensitivity to physiologic BP changes. Alternatively, the lesser precision of the FinometerTM could be due to the displacement of the cuff between replicated measurements. Because the majority of measurements were recorded during active sleep, the position of the cuff on the wrist could have altered due to the characteristic body movements of this state. Therefore, during a research study, this could be rectified by ensuring that the wrist with the BP cuff is kept still or is confined to 1 position as the infant sleeps. Despite the overall good agreement between methods of measurement, we did observe variation in mean differences and 95% limits of agreement between subjects, with the FinometerTM both underestimating and overestimating absolute BP in some infants. This variation may have been due to a poor-fitting cuff in these particular infants. The inflatable cuff operates to clamp the diameter of the artery constant by opposing pulsatile changes occurring during each heart beat. Any change is counteracted by a fast pressure servo controller that increases or decreases pressure accordingly in the inflatable bladder. If the BP cuff is too large, this could lead to an overestimation of BP caused by a pressure gradient over the air bladder due to overinflation. Conversely, if the cuff is too small or wrapped too tightly, this could result in an underestimation of BP. This may explain the overestimation of BP in infant 1, who had the lowest birth weight (964 gm) of all infants studied. Despite using the smallest cuff on this infant, the FinometerTM overestimated DAP by 13 mm Hg, which may suggest that the cuff was too large. Other authors have also reported considerable variation in bias between subjects and have questioned whether or not it is acceptable to pool individual BP

Table 4—Determinates of BRS Estimates for 8 Infants for the Catheter and FinometerTM Systems

Catheter FinometerTM Infant Gestational BRS ms/mm Hg Ramps, no. BRS ms/mm Hg Ramps, no. p ValueNo. age, wk 1 30 0.57 ± 0.70 54 0.80 ± 0.19 48 NS 2 36 1.64 ± 0.09 84 2.22 ± 0.68 70 NS 4 32 2.30 ± 0.75 8 2.70 ± 0.59 8 NS 6 27 1.63 ± 0.38 63 2.09 ± 0.47 65 NS 7 33 1.86 ± 0.72 35 0.91 ± 0.13 42 NS 8 28 2.28 ± 0.59 169 1.90 ± 0.51 154 NS 9 28 1.23 ± 0.15 207 1.70 ± 0.27 200 NS 10 28 1.01 ± 0.12 93 1.66 ± 0.27 92 NS Mean 30 ± 1 1.56 ± 0.21 713 1.74 ± 0.23 679 NS

The mean baroreflex sensitivity (BRS) ± SEM is presented in ms/mm Hg.

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al

Table 3—Mean DAP Values Measured by the Catheter System and the FinometerTM for 9 Subjects Subject DAPc DAPf DDAP Limits of agreement 1 29 42 13 7 to 20 2 33 39 6 4 to 8 4 39 42 3 1 to 5 5 33 43 10 7 to 13 6 33 45 12 7 to 17 7 40 44 4 -2 to 10 8 30 31 1 -1 to 3 9 23 32 9 6 to 11 10 27 35 8 4 to 13 Mean ± SEM 32 ± 2 39 ± 2 7 ± 1 4 to 10

Diastolic arterial pressure (DAP) values, in mm Hg, were measured by the catheter system (MAPc) and the FinometerTM (MAPf). DDAP represents the mean difference and limits of agreement between each method of measurement.

SLEEP, Vol. 29, No. 8, 2006 1088

values.16 However, as evident from our Bland-Altman analysis, in particularly for MAP (Figure 2a), the majority of individual subject differences are close to 0, and mean differences were normally distributed. Therefore, in a research setting with adequate repeated measurements in each subject and normally distributed data, we are confident that a pooled comparison will be representative of population estimates. A second limitation was that because preterm infants spend the majority of their sleeping time in active sleep, the majority of our measurements (55/58) were made in this state, and potential sleep-state effects could not be examined. However, because physiologic variability is less in quiet sleep, we anticipate that any differences in the measurement systems would be less in this state. In summary, our study has found that the FinometerTM is a useful tool to measure both absolute and beat-to-beat changes in BP in infants during sleep with the adult-sized finger cuff placed around the wrist. The continuous noninvasive measurement of BP will provide new opportunities to study the maturation of baroreflex control of BP in infants and may provide important insights into a number of infant sleep-related problems such as sudden infant death syndrome, which has been proposed to result from a failed baroreflex response to profound hypotension.23

ACKNOWLEDGMENTS

We would like to thank the medical and nursing staff at the Neonatal Intensive Care Unit, Monash Medical Centre, and all of the parents who volunteered their infants for participation in the study. This project was supported by the National Health and Medical Research Council of Australia.

REFERENCES

1. Amin RS, Carroll JL, Jeffries JL, et al. Twenty-four-hour ambulatory blood pressure in children with sleep-disordered breathing.[see comment]. Am J Respir Crit Care Med 2004;169:950-6.

2. Suzuki M, Guilleminault C, Otsuka K, Shiomi T. Blood pressure "dipping" and "non-dipping" in obstructive sleep apnea syndrome patients. Sleep 1996;19:382-7.

3. Schechtman VL, Harper RM, Kluge KA, Wilson AJ, Hoffman HJ, Southall DP. Cardiac and respiratory patterns in normal infants and victims of the sudden infant death syndrome. Sleep 1988;11:413-24.

4. Harper RM, Woo MA, Alger JR. Visualization of sleep influences on cerebellar and brainstem cardiac and respiratory control mechanisms. Brain Res Bull 2000;53:125-31.

5. Galland BC, Reeves G, Taylor BJ, Bolton DP. Sleep position, autonomic function, and arousal. Arch Dis Child Fetal Neonatal Ed 1998;78:F189-94.

6. Robbe HW, Mulder LJ, Ruddel H, Langewitz WA, Veldman JB, Mulder G. Assessment of baroreceptor reflex sensitivity by means of spectral analysis. Hypertension 1987;10:538-43.

7. Bertinieri G, Di Rienzo M, Cavallazzi A, Ferrari AU, Pedotti A, Mancia G. Evaluation of baroreceptor reflex by blood pressure monitoring in unanesthetized cats. Am J Physiol 1988;254:H377-83.

8. Blaber AP, Yamamoto Y, Hughson RL. Methodology of spontaneous baroreflex relationship assessed by surrogate data analysis. Am J Physiol 1995;268:H1682-7.

9. Zoccoli G, Andreoli E, Bojic T, et al. Central and baroreflex control of heart rate during the wake-sleep cycle in rat. Sleep 2001;24:753-8.

10. Chong A, Murphy N, Matthews T. Effect of prone sleeping on circulatory control in infants. Arch Dis Child 2000;82:253-6.

11. Thoresen M, Cowan F, Walloe L. Cardiovascular responses to tilting in healthy newborn babies. Early Hum Dev 1991;26:213-22.

12. Andrasyova D, Kellerova E. Blood pressure and heart rate response to head-up position in full-term newborns. Early Hum Dev 1996;44:169-78.

13. Penaz J VA, Teichmann W. Contribution to the continuous indirect blood pressure measurement. Z Gesamte Inn Med 1976;15:1030-3.

14. Drouin E, Gournay V, Calamel J, Mouzard A, Roze JC. Feasibility of using finger arterial pressure in neonates. Arch Dis Child Fetal Neonatal Ed 1997;77:F139-40.

15. Harrington C, Kirjavainen T, Teng A, Sullivan CE. Cardiovascular responses to three simple, provocative tests of autonomic activity in sleeping infants. J Appl Physiol 2001;91:561-8.

16. Andriessen P, Schoffelen RL, Berendsen RC, et al. Noninvasive assessment of blood pressure variability in preterm infants. Pediatr Res 2004;55:220-3.

17. Drouin E, Gournay V, Calamel J, Mouzard A, Roze JC. Assessment of spontaneous baroreflex sensitivity in neonates. Arch Dis Child Fetal Neonatal Ed 1997;76:F108-12.

18. Curzi-Dascalova L, Mirmiran M. Manual of methods for recording and analysing sleep-wakefulness states in preterm and full-term infants. Paris: Les Editions INSERM; 1996.

19. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346:1085-1087.

20. Youden WJ Statistical Methods for Chemists. New York: Wiley Publications in Statistics; 1951.

21. Oosting J, Struijker-Boudier HA, Janssen BJ. Validation of a continuous baroreceptor reflex sensitivity index calculated from spontaneous fluctuations of blood pressure and pulse interval in rats. J Hypertens 1997;15:391-9.

22. Gournay V, Drouin E, Roze JC. Development of baroreflex control of heart rate in preterm and full term infants. Arch Dis Child Fetal Neonatal Ed 2002;86:F151-4.

23. Harper RM. Sudden infant death syndrome: a failure of compensatory cerebellar mechanisms? Pediatr Res 2000;48:140-142.

Noninvasive Blood Pressure Measurement In Infants—Yiallourou et al

Toll Free: 1-888-447-6673Email: USinfo@somnomed.com

� Durable � Patient Friendly � Very Comfortable � Customer Made for Precision Fitting

Non-Compliant CPAP Patients?Mild to Moderate Sufferers?

Provide your patients an alternative to CPAP in accordance with the February 2006 AASM guidelines.Call now to join the sleep centres who have already addressed their CPAP non-compliant patients.Join a referral network in your area today.