Peripheral Muscle Near-Infrared Spectroscopy in Neonates: Ready for Clinical Use? A Systematic Qualitative Review of the Literature

Near-infrared spectroscopy (NIRS) is an increasingly used non-invasive method to measure tissue oxygenation and perfusion in regions of interest, e.g. the brain, kidney or peripheral muscle in neonates. In the past few years efforts have been made to improve this method. The technique is improving and quality criteria to increase reproducibility in cerebral and peripheral muscle NIRS measurements have been introduced [1,2]. Parameters potentially influencing the peripheral oxygenation and perfusion have also been detected [3].

In critically ill neonates peripheral muscle NIRS measurements have a great potential, especially in recognizing early states of (compensated) shock, while other vital parameters, such as arterial oxygen saturation measured by pulse oximetry or blood pressure, still remain in the normal range [4,5]. Many studies have been published trying to demonstrate the potential usefulness of peripheral muscle NIRS measurements in the diagnosis and treatment of neonates using either the arterial [6,7] or the venous occlusion method [1,3,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22] or both [23,24,25]. The aim of this work was to perform a systematic qualitative review on peripheral muscle NIRS measurements in the clinical care of term and preterm neonates.

Articles were identified using the stepwise approach specified in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Statement [26].

A systematic search was performed of PubMed and Ovid Embase to identify articles in the English language published between January 1973 and March 2015. Studies had to address single, combined or multi-site measurements of peripheral muscle oxygenation and perfusion in neonates. Search terms included: neonate, neonates, newborn, newborns, infant, infants, near-infrared spectroscopy, NIRS, oxygenation, perfusion, oxygen extraction, peripheral, tissue, muscle, calf, forearm and thigh (Appendix). Additional published reports were identified through a manual search of references in the retrieved articles and in review articles. Only human studies were included.

The articles identified following the literature review were evaluated independently by two authors (N.H., G.P.) for inclusion using the titles and abstracts. Where uncertainty remained regarding eligibility for inclusion, the full texts were reviewed. Two reviewers independently identified relevant abstracts, critically appraised the full text of identified articles and assessed the methodological quality of the included studies. Data were analysed qualitatively, and data extraction included the characterization of study type, patient demographics, methods and results.

Our initial search identified 5,225 articles. After the removal of duplicates and rejection (e.g. peripheral muscle NIRS measurements of body parts other than extremities in human neonates), 41 studies fulfilled our inclusion criteria (fig. 1). All 41 studies performed peripheral muscle NIRS measurements in human neonates. Twenty-one studies were identified to use peripheral muscle NIRS measurements as a single method [1,3,6,7,8,9,10,11,12,13,14,15,16,17,18,19,23,24,25,27,28] (table 1), 17 studies combined cerebral and peripheral muscle NIRS measurements [20,21,22,29,30,31,32,33,34,35,36,37,38,39,40,41,42] (table 2) and 1 study used multi-site NIRS measurements [43] (table 2) in human neonates. Two randomized studies were identified [12,36]. Two additional publications were included because they provided important general information about peripheral NIRS measurements [44,45]. Tables 1 and 2 give an overview of the basic data of all included studies, while tables 3 and 4 compare the measured NIRS parameters, where available.

Over the past decades NIRS has become an increasingly widespread non-invasive method for measuring peripheral muscle oxygenation and perfusion in term and preterm neonates.

Peripheral muscle NIRS measurements can be combined with the venous occlusion [1,3,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22] and arterial occlusion methods [6,7]. With venous occlusion it is possible (due to the undisturbed arterial inflow to the extremity and the interrupted venous outflow) to measure blood flow and venous oxygen saturation (SvO₂). Furthermore, by taking into account arterial oxygen saturation, oxygen delivery (DO₂), oxygen consumption (VO₂) and fractional oxygen extraction (FOE) can be calculated. With arterial occlusion only VO₂ can be measured. Since arterial inflow and venous outflow are interrupted during arterial occlusion, changes in oxygenated and deoxygenated haemoglobin are due to the VO₂.

Hassan et al. [23] were the first to compare arterial occlusion with the venous occlusion method in healthy neonates for the measurement of VO₂. They found a weak correlation between these two methods. Only in another two studies [24,25] were both occlusion methods used. Differences between the measurement results after venous occlusion are due to the fact that there are different approaches in analysing changes during venous occlusion [1]. In several studies the first few seconds after occlusion were used for analysis [9,10,16,20,21,24,25]. Pichler et al. [1] showed that analysing the first 15 s after the beginning of venous occlusion showed a lower test-retest variability.

In recent studies the venous occlusion method has become the preferred method [1,3,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]. This is mainly due to the better feasibility of venous occlusions in neonates. Arterial occlusions cause more discomfort to the neonates and, as a result, measurements are more unreliable due to an increase in movement artefacts.

Yoxall and Weindling [8] as well as Bay-Hansen et al. [11] compared peripheral SvO₂ measured by NIRS combined with the venous occlusion method with co-oximetry of peripheral and central venous blood. Although Yoxall and Weindling [8] took the blood from a superficial forearm vein instead of a central venous catheter as Bay-Hansen et al. [11] did, both studies showed a significant correlation between NIRS measurements of peripheral SvO₂ and co-oximetry of peripheral and central venous blood. However, SvO₂ measured by these two methods should be compared with caution, since peripheral NIRS measures SvO₂ directly in the tissue underlying the sensor, whereas in co-oximetry blood of a larger vessel represents SvO₂ of a larger region.

Zaramella et al. [24] used the ‘foot perfusion index', measured with pulse oximetry, to evaluate limb perfusion. They compared the results of the foot perfusion index with blood flow, DO₂, VO₂ and FOE, measured by NIRS on the calf. The foot perfusion index correlated with blood flow and DO₂, but not with VO₂ or FOE. They concluded that measurements of the foot perfusion index seemed clinically more feasible than NIRS measurements.

Apart from different occlusion methods, peripheral muscle NIRS measurements can be performed on the forearm or calf. Pichler et al. [14] were the first to compare forearm and calf muscle tissue oxygenation in healthy term neonates measured with NIRS. They found similar results of FOE, tissue oxygenation index (TOI) and SvO₂, whereas DO₂ and VO₂ were significantly higher in calf tissue. The differences of DO₂ and VO₂ were interpreted to be due to the increasing limb size and not because of a different tissue composition of both limbs. Nevertheless, this study showed that comparing the results of studies using different extremities might be difficult. In summary, when comparing DO₂ and VO₂ of the forearm and calf, circumference and tissue composition of the measured extremity can influence results.

Apart from differences in arterial and venous occlusion [23] or the use of different extremities for NIRS measurements [14], patient populations, technical aspects of the used devices and the measuring procedure itself can cause difficulties in the application of peripheral muscle NIRS. Therefore, recommendations for a standardized approach were published in 2008 [44]. These recommendations contained clinical and methodical items, especially the application of optodes and analysis. Furthermore, in 2009, quality criteria were introduced to increase the reproducibility of peripheral NIRS measurements in term and preterm neonates [1]. To overcome the problem of missing comparability due to different units, Wickramasinghe and Spencer [45] described a method for expressing NIRS-derived peripheral VO₂ units in the usually used standard unit of ml O₂/kg/min.

Schmitz et al. [39] analysed the feasibility of long-term continuous measurements of cerebral regional oxygen saturation (rSO₂) and peripheral rSO₂ on the first day of life in term and preterm neonates. They found that most artefacts were due to missing values and therefore easy to recognize. Fujioka et al. [40] evaluated cerebral and peripheral oxygenation and blood volumes between term and preterm infants during the first 3 days of life. They found a significantly higher cerebral blood volume in term neonates, whereas the peripheral blood volume was higher in preterm neonates. There was no difference in cerebral and peripheral oxygenation. The contrasting results might be explained by differences in the maturation of the physiological mechanism to control cerebral and peripheral blood volume.

The most widely used devices for NIRS measurements in neonates are the NIRO (Hamamatsu Inc., Hamamatsu, Japan) and the INVOS (Somanetics, Troy, Mich., USA) systems. Both are based on ‘spatially resolved spectroscopy' [46], but they differ in the mode of light absorption and the number of wavelengths [47]. Pocivalnik et al. [22] compared the values of cerebral and peripheral TOI measured by the NIRO 300 (Hamamatsu Inc.) with the values of cerebral and peripheral rSO₂ measured by the INVOS 5100 (Somanetics). Cerebral and peripheral TOI values were significantly lower than rSO₂ values. Reproducibility for cerebral measurements was similar, while in peripheral muscle measurements reproducibility for peripheral TOI was higher than for peripheral rSO₂. Differences in reproducibility can be explained by technical and calibration differences. Grossauer et al. [30] were the first to compare simultaneously measured cerebral TOI (70.4 ± 6.7%) and peripheral TOI values (62.1 ± 5.7%) in term and preterm neonates and introduced a cerebral TOI/peripheral TOI ratio of 1.14 ± 0.14. In summary, factors like patient population, measurement protocol, technical devices and analysis of measurements play an important role in interpreting and comparing data of different studies.

Regarding the time of NIRS measurements, transition from the foetus to the newborn with all its complex physiological processes is of high interest. Urlesberger et al. [31] evaluated rSO₂ of the brain and pre- and postductal peripheral muscle tissue during this period of life in term infants after elective Caesarean section. rSO₂ showed an increase during the first 7 min of life, during which the brain had the highest saturation levels. Furthermore, fractional tissue oxygen extraction of the brain reached a plateau earlier compared to peripheral tissue. In another study, Urlesberger et al. [37] showed that term neonates, after elective Caesarean section with a left-to-right shunt via the ductus arteriosus, had significantly higher cerebral rSO₂ values compared to infants without shunting. They found no significant difference between the two groups regarding pre- and postductal peripheral rSO₂ values. Comparing preterm neonates with and without non-invasive ventilation, Binder et al. [38] as well as Schwaberger et al. [41] found significantly lower cerebral and peripheral rSO₂ values in the respiratory support group during the first 15 min of life. Pocivalnik et al. [42] showed that immediate postnatal oropharyngeal suctioning did not compromise cerebral and peripheral muscle tissue oxygenation in term neonates.

Concerning postnatal age, Pichler et al. [13] analysed changes in peripheral oxygenation in healthy term neonates measured twice by the NIRS and the venous occlusion method within the first week of life. They found an increase in VO₂ and FOE as well as a decrease in TOI over time.

Hassan et al. [6] described a decrease in peripheral VO₂ during limb cooling in healthy term neonates. In another study, the same group [7] showed a significant rise in global and peripheral VO₂ after a cold stimulus, which means that peripheral VO₂ is influenced by global VO₂.

Wardle et al. [10] examined the influence of hypotension on peripheral oxygenation in ventilated preterm neonates. The study showed that peripheral VO₂ and DO₂, as well as median haemoglobin flow, were lower in hypotensive preterm neonates compared to a normotensive group. After the treatment of hypotension, DO₂ and VO₂ increased to values comparable with the normotensive group [10]. Kissack and Weindling [16] explored the relationship between mean arterial blood pressure and two parameters of perfusion and oxygenation in peripheral tissue, peripheral blood flow and FOE during the first 12 h after birth in sick and ventilated very-low-birth-weight infants. Their results indicate that blood pressure should not be used as a surrogate for peripheral blood flow and DO₂. Victor et al. [20] investigated the relationship between mean arterial blood pressure, cerebral electrical activity (EEG), cerebral FOE and peripheral blood flow in very-low-birth-weight infants. At mean arterial blood pressure levels between 23 and 33 mm Hg peripheral blood flow was abnormally low, while EEG patterns and cerebral FOE remained normal. They concluded that cerebral perfusion is probably maintained at mean arterial blood pressure levels above 23 mm Hg. On the other hand, no apparent relationship between cardiac output, peripheral blood flow, EEG and cerebral FOE could be found [21]. In infants with low cardiac output and normal mean arterial blood pressure, cerebral and peripheral perfusion was maintained [21].

The summarized studies show that there are many physiological parameters associated with peripheral muscle oxygenation. Different approaches to investigate peripheral oxygenation and circulation with NIRS were used in the studies, making comparability and interpretation difficult. However, recommendations on peripheral muscle NIRS measurements have been published [44]. Recently, a large observational study in term and preterm neonates, which took these recommendations into account, also showed multi-associations of physiological parameters on peripheral oxygenation and circulation [3] that have to be taken into account when comparing studies and interpreting data.

A prospective cohort study using NIRS and the venous occlusion method demonstrated that TOI is reduced and FOE increased in healthy term neonates whose mothers smoked during pregnancy compared to healthy term neonates in the non-smoking group on the first day of life [15]. Milan et al. [28] used peripheral perfusion and oxygenation measured by NIRS in eight neonates to provide indirect information on circulatory failure in limb arterial thromboembolic disease. They found lower values of TOI, oxygenated and deoxygenated haemoglobin and blood volume in the affected limb compared to the unaffected one. Nonetheless, the TOI values in the unaffected group were also low compared to other studies [1,3,13,14,15,17,18,19,27]. We can only speculate on this observation. Firstly, there might be some methodical differences in the application of optodes and duration of measurements. Secondly, there might be differences in study populations and, thirdly, a low number of patients might have influenced the results.

Regarding perinatal asphyxia, Tax et al. [18] demonstrated an increasingly compromised peripheral oxygenation and perfusion measured by the NIRS and venous occlusion method in relation to the degree of acidosis in the umbilical cord blood in term and preterm neonates. In neonates with C-reactive protein elevation, impaired peripheral oxygenation was already observed even when routine haemodynamic parameters, such as heart rate and blood pressure, were still normal [17]. Regarding increasing leucocyte counts, Binder et al. [19] found a decrease in peripheral tissue VO₂, while vascular resistance increased using the NIRS and venous occlusion method.

Wardle et al. [12] performed a randomized controlled trial to demonstrate the use of FOE to guide the need for blood transfusions in preterm neonates. Fewer transfusions were given to neonates in the NIRS group than in the conventional group. However, FOE measurements failed to identify many neonates felt by clinicians to require blood transfusion. Baenziger et al. [27] used peripheral NIRS measurements in haemodynamically stable preterm neonates to detect changes of the FOE due to alterations in DO₂ induced by blood transfusions. They found a positive correlation between the decrease of FOE after transfusion and the number of transfused red cells and the increase in haematocrit [27].

Zaramella et al. [25] examined whether the cord clamping time affects limb perfusion measured by NIRS and heart haemodynamics using M-mode echocardiography in vaginally delivered healthy term neonates. Although late cord clamping (4 min after birth) showed higher haematocrit and haemoglobin values as well as a larger left ventricular end-diastolic diameter than early cord clamping (30 s after birth), no changes in peripheral perfusion (calf blood flow and DO₂) and oxygen metabolism (VO₂, SvO₂, FOE) were found.

In critically ill neonates with low cardiac output syndrome, levosimendan, a new inodilator that enhances myocardial contractility and additionally causes peripheral and coronary vasodilatation, had a beneficial effect on cerebral and systemic perfusion as well as oxygenation [35]. In a phase I, randomized study, Pellicer et al. [36] investigated the haemodynamic effects of the routinely used milrinone, a selective phosphodiesterase III inhibitor, that has inotropic and lusitropic effects on the myocardium and a relaxing effect on the vascular muscle, and the novel molecule levosimendan in neonates undergoing cardiovascular surgery using NIRS measurements to assess changes in cerebral and peripheral perfusion and oxygenation. The study provided data on the early intraoperative use of both inodilators, but also emphasized the persistent haemodynamic effects of levosimendan after the end of drug infusion. Regarding simultaneous NIRS measurements of cerebral and peripheral tissue oxygenation, Redlin et al. [33] used this method in combination with haemoglobin concentrations to detect regional malperfusion during paediatric cardiac surgery. These measurements were also used as a monitor for cerebral and lower body tissue oxygenation during cardiopulmonary bypass operations with a blood-sparing approach [32] and during deep hypothermic circulatory arrest with regional low-flow perfusion [29]. In a clinical observational study simultaneous cerebral and peripheral NIRS measurements were also used to investigate the treatment effects of partial exchange transfusions in newborns with polycythaemia [34]. Partial exchange transfusions led to an increase in cerebral oxygenation, whereas peripheral oxygenation remained unchanged.

This review shows that there are still some limitations of this method. First of all, when comparing studies with peripheral muscle NIRS measurements, one has to bear in mind the differences between the arterial and venous occlusion method as well as the differences between upper and lower extremities. Secondly, different NIRS devices are currently available and used for clinical and scientific purposes. Studies showed different results for the measured NIRS parameters and, therefore, caution is advised when comparing data due to many influencing physiological and pathological parameters. Furthermore, there are no standardized values for the differential path length factor or the interoptode distance in neonates (tables 1, 2). Citation of the used values in studies is recommended. Finally, different units are used for measured NIRS parameters (table 3). Limitations of peripheral muscle NIRS measurements have been published with recommendations as a consensus statement by Pichler et al. [44].

Peripheral muscle NIRS measurements are an increasingly used method to monitor peripheral oxygenation and perfusion in neonates. Over the past decades significant effort has been made to increase the reproducibility and accuracy of peripheral muscle NIRS measurements. Especially in the recognition of early states of centralization (compensated shock), peripheral muscle NIRS may provide a promising tool. At the moment there is one ongoing randomized controlled trial using peripheral muscle NIRS in combination with cerebral NIRS to detect the early signs of centralization in order to start an early intervention to avoid hypotension in preterm neonates (the AHIP trial; ClinicalTrials.gov identifier: NCT01910467). Especially in the care of critically ill neonates, peripheral muscle NIRS measurements alone or in combination with cerebral or multi-site NIRS measurements provide useful additional information about the state of health in this vulnerable group of patients in the future. However, before this method can be used in the clinical routine it has to be tested as monitoring to guide interventions in further studies.

No external funding was secured for this study. The authors have no financial relationships relevant to this article to disclose. The authors have no conflict of interest to disclose.