Main

During the perinatal period, the brain undergoes anatomic, functional, and metabolic changes. The anatomic changes include neuronal proliferation, migration, organization, and myelination, and the metabolic changes match the process of initial overproduction and subsequent elimination of excessive neurons, synapses, and dendritic spines known to occur in the developing brain. Noninvasive assessment of cerebral anatomic changes and of oxygen delivery and utilization is useful for evaluating the effectiveness of therapy and for preventing oxygen toxicity in seriously ill neonates.

Near-infrared spectroscopy (NIRS) has been used in the clinical field with various measuring devices using several wavelengths. A method using continuous-wave NIRS has been developed and reported to be suitable for clinical use in infants (1–7). However, current commercially available NIRS systems can detect only changes in cerebral Hb. Because NIRS is based on the modified Beer-Lambert law, a change in hematocrit and blood volume as well as developmental and pathophysiologic changes in brain tissue affect the pathlength of near-infrared light. In a few recent studies, absolute values of cerebral Hb oxygen saturation (Sco2) and cerebral blood volume (CBV) in infants were measured without inducing Hb concentration changes by using full-spectral near-infrared spectroscopy (8–11) and spatially resolved spectroscopy (12). However, these devices can measure only light absorption coefficient (μa), which represents the physiologic state, particularly the Hb concentration and oxygen saturation.

A recently developed time-resolved spectroscopy (TRS) device enables simultaneously quantitative analysis of μa and light-reduced scattering coefficient (μ's) in tissue by using the photon diffusion theory. μ's is thought to be a new parameter for assessment of structural changes in the brain, such as brain edema and myelination. Although TRS has been used in neonates, there have only been a few reports on its use in neonates, and measurements in neonates at the bedside have not been possible because of the size and the cost of typical laboratory equipment needed for these measurements. However, a new TRS device that is portable and has a high data acquisition rate was used clinically recently. This TRS system can be used 1) for continuous absolute quantification of hemodynamic variables and 2) for better estimation of light-scattering properties by measurement of μ's and differential pathlength factor (DPF). The aim of this study was to measure the values of μ's, μa, Sco2, CBV, and DPF in neonates using TRS and to determine the relationships between postconceptional age and μ's, Sco2, CBV, and DPF.

METHODS

Patient population.

Measurements were carried out in 27 neonates who were undergoing neonatal intensive care at the Maternal and Perinatal Center of Kagawa University Hospital. Written informed consent was obtained from the parents of each neonate. The study was also approved by a local ethics committee. Data from five neonates were excluded from the analysis because the measurements were affected by movement artifacts. Successful measurements were performed in 22 neonates. Their gestational age (mean ± SD) was 36.8 ± 3.1 wk, and birth weight was 2365 ± 791 g. The mean time after birth when measurements were carried out was 32.9 ± 21.1 h. The clinical diagnosis of each neonate is shown in Table 1. Mechanical ventilation was required in four neonates, and three neonates received catecholamines during this study. No neonates have problems in their prognosis at present.

Table 1 Clinical details
Full size table

Near-infrared time-resolved spectroscopy system and analysis.

We used a portable three-wavelength TRS system (TRS-10; Hamamatsu Photonics K.K., Hamamatsu, Japan) and attached a probe to the forehead of each neonate. In the TRS system, a time-correlated single-photon-counting technique is used for detection. The system is controlled by a computer through a digital I/O interface that consists of a three-wavelength (761, 795, and 835 nm) picosecond light pulser (PLP) as the pulse light source, a photon-counting head for single photon detection, and signal-processing circuits for time-resolved measurement. The PLP emits NIR light with a pulse duration of ∼100 ps and an average power of at least 150 μW at each wavelength at repetition of 5 MHz. The input light power to the patient was ∼20 μW.

The light from the PLP is sent to a patient by a source fiber with a length of 3 m, and the photon re-emitted from the patient is collected simultaneously by a detector fiber bundle with a length of 3 m (13). The light source fiber used in this study was a graded-index-type single fiber with a numerical aperture of 0.25 and a core diameter of 200 μm, and the light detector fiber was a bundle fiber with a diameter of 3 mm and numerical aperture of 0.21. Finally, a set of histograms of photon flight time, which is called a re-emission profile, is recorded (14,15). One temporal re-emission profile includes 1024 time channels that span ∼10 ns with a time step of ∼10 ps. In this study, the emerging light was collected over a period of 1 s to exceed at least 1000 count of photon at the peak channel of the re-emission profiles. The instrumental response was measured with the input fiber placed opposite the receiving fiber through a neutral density filter. The instrumental response of the TRS system was ∼150 ps Full-Width Half-Maximum at each wavelength.

The re-emission profiles observed at each measurement point were fitted by the photon diffusion equation proposed by Patterson et al. (16) to calculate the values of μa and μ's of the head at wavelengths of 761, 795, and 835 nm. In the fitting procedure, a nonlinear least squares fitting method based on Levenberg-Marquardt's method was used. In each iterative calculation, the function from the photon diffusion equation in reflectance mode, which was convoluted with the instrumental response, was fitted to the observed re-emission profile. The calculation regions were determined to include the observed profile data, and data of 600 channels were included into the fit.

After determination of the values of μa at three wavelengths, the oxyHb and deoxyHb concentrations were calculated from the absorption coefficients of oxyHb and deoxyHb using the following equations with the assumption that background absorption is due only to 85% (by volume) water (8):

MATH

In these equations, ε λ nm χ is the extinction coefficient at λ nm, and [oxyHb] and [deoxyHb] are concentration of oxyHb and deoxyHb, respectively.

First, water absorption was subtracted from μa at each of the three wavelengths, and then the concentrations of oxyHb and deoxyHb were estimated using the least squares fitting method. The absorption coefficients for oxyHb, deoxyHb, and water shown in Table 2 were used.

Table 2 Absorption coefficients for oxyHb, deoxyHb, and water
Full size table

Cerebral total Hb (totalHb) concentration, Sco2, and CBV were calculated as follows:

MATH

where [ ] indicates Hb concentration (μM), MWHb is the molecular weight of Hb (64,500), tHb is venous Hb concentration (g/dL), and Dt is brain tissue density (1.05 g/mL).

The mean pathlength was calculated from the difference between the center of gravity of the measured reemission profile and that of the instrumental function. We assumed that the value for the refractive index of brain tissue is 1.4 (17). The ratio of optical pathlength to interoptode distance is defined as the DPF (18).

All of the neonates were in the supine position during measurements. Their condition was stable, and they were sleeping during the procedure at least 12 min. The optode positions were on the forehead of each neonate, and interoptode space was 26–32 mm. At the same time, oxygen saturation by pulse oximeter (Spo2) was monitored using a Nellcor N550 (Tyco, Tokyo, Japan).

The average values of μ's, Sco2, CBV, and DPF for each patient were calculated for a period of 5 min during the 12-min measurement period in a steady state not affected by movement artifacts. Variation, particularly in cerebral oxygen delivery, can occur over short periods, and long averaging measurement time is needed. Sco2 depends on cerebral oxygen delivery and extraction, both of which vary with postconceptional age and postnatal age. We previously reported postnatal changes in CBV and Sco2 in normal infants determined by full-spectrum NIRS. CBV and Sco2 changed within the first 15 min after birth (10,11), but remained constant from 12 h after birth until day 5 (9). In the present study, the average time at which measurements were carried out after birth was 33 ± 21 h, and the values of Sco2 and CBV in this study therefore were not affected by birth stress. In a previous study, the cerebral blood velocity in infants measured by Doppler ultrasound showed cyclical fluctuations with frequency ranging from 1.5 to 5 cycles/min (19). NIRS studies have also shown oscillations of the Hb oxygenation state with frequency ranging from 3 to 5 cycles/min (20). Therefore, the use of average values for a period of 5 min in each neonate in this study seems sufficient for estimation of cerebral Hb in a steady state.

Statistical analysis.

A StatView-J 5.0 package for the Macintosh computer was used for statistical analysis. The level of statistical significance was set at a probability of p < 0.05 for all tests. All measurement results are expressed as means ± SDs.

RESULTS

The values of DPF, μa, μ's, Sco2, totalHb, and CBV are shown in Table 3. The values of μ's in the 22 neonates at 761, 795, and 835 nm were estimated to be (mean ± SD) 6.46 ± 1.21, 5.90 ± 1.15, and 6.40 ± 1.16/cm, respectively. As shown in Fig. 1, there was a significant positive relationship between postconceptional age and μ's at the three wavelengths.

Table 3 Results of the study
Full size table
Figure 1
figure 1

Relationships between postconceptional age and μ's at 761, 795, and 835 nm in 22 neonates. The relationships between postconceptional age and μ's at 761, 795, and 835 nm were y 761 nm = 0.293x − 4.40 (r = 0.742, p < 0.001), y 795 nm = 0.202x − 1.58 (r = 0.539, p = 0.010), and y 835 nm = 0.298x − 4.65 (r = 0.787, p < 0.001), respectively.

Full size image

Mean Sco2 was 70.0 ± 4.6% (range 60.8–78.8%), and mean Spo2 was 98.6% (range 92–100%). Postconceptional age and Sco2 showed a significant negative linear relationship as shown in Fig. 2.

Figure 2
figure 2

Relationships between postconceptional age and cerebral Hb oxygen saturation in 22 neonates. The relationship between postconceptional age and cerebral Hb oxygen saturation was y = −0.74x + 97.5 (r = −0.498, p = 0.018).

Full size image

There was a significant positive relationship between postconceptional age and cerebral totalHb. The relationship between postconceptional age and cerebral totalHb was y = 4.23x + 91.9 (r = 686, p < 0.001). The mean CBV was 2.31 ± 0.56 mL/100 g (range 1.42–3.40 mL/100 g). As shown in Fig. 3, there was a significant positive relationship between postconceptional age and CBV. There was no relationship between blood Hb concentration and CBV.

Figure 3
figure 3

Relationships between postconceptional age and CBV in 22 neonates. The relationship between postconceptional age and CBV was y = 0.126x − 2.352 (r = 0.696, p < 0.001).

Full size image

The DPF values at 761, 795, and 835 nm were estimated to be 4.58 ± 0.41, 4.64 ± 0.46, and 4.31 ± 0.42, respectively. There was no relationship between postconceptional age and DPF at any of the three wavelengths.

DISCUSSION

This is the first report on the relationship between postconceptional age and μ's in the neonatal brain. We previously reported that there were no significant differences between the values of μ's at each wavelength for inspired fractional O2 levels in the range of 4–100% in a piglet hypoxia model. These results are similar to those obtained in a study by Zhang et al. (21) showing that scattering changes detected by frequency-domain spectroscopy were associated only with asphyxia and death. Yamashita et al. (22) reported the results of a preliminary study on μ's in the piglet brain by using TRS. Their results showed a notable decrease after death. Tissue edema and structural changes occur during severe hypoxia, particularly at and after death, and values of μ's are thought to change only during structural changes in tissue as a result of cerebral energy failure. Furthermore, developmental changes in the brain, especially neuronal proliferation, migration, organization, and myelination, were thought to be related to the positive relationship between μ's and postconceptional age. However, at present, magnetic resonance imaging techniques enable a much better assessment of anatomic development in infants.

Values of Sco2 in infants that were obtained in previous studies using NIRS (8–12,23,24) are summarized in Table 4. Mean Sco2 in the 22 neonates in this study was 70.0 ± 4.6%, and the range of Sco2 values was small (60.8–78.8%). This range is similar to those previously reported (63–69%) in infants (8–12,23). Results of measurements using the TRS system also showed a decrease in Sco2 with increasing postconceptional age. Sco2 measured by NIRS represents a mixed vascular Hb oxygen saturation of capillaries, arteriae, and veins in that tissue field. Absolute measurements are made on the basis of the assumption of homogeneity of tissue. The reason for decrement of Sco2 with increasing postconceptional age is that cerebral Hb oxygen consumption increases with advance of postconceptional age, and this leads to a decrease in venous Hb oxygen saturation. The lower values of cerebral oxygen consumption in neonates than in older children is likely to be due to changes in the structural complexity and functional activity of the brain that occur across the range of gestational ages (25–27). Another reason is that venous structural change in the brain surface occurs in this period. The ratio of venous to arterial vessels may increase, and this would lead to an increase in cerebral content of deoxyHb and therefore to decrease in Sco2.

Table 4 Some values of Sco 2 in infants using NIRS
Full size table

In this study, we noninvasively estimated values of CBV in neonates using a TRS system. Brazy et al. (1,28) monitored changes in CBV in infants, but quantification of CBV has not been possible. Methods for calculating CBV using oxyHb as a tracer with continuous-wave NIRS have been reported (29,30). Moreover, without changing oxyHb, CBV has been measured using indocyanine green (ICG) with spatially resolved NIRS (31). However, to our knowledge, there have been no studies in which CBV in infants was estimated by using a TRS system without changes in oxyHb or without using ICG injection. The mean value obtained in the neonates in the present study, 2.31 ± 0.56 mL/100 g, is similar to the values estimated in infants by using the continuous-wave NIRS method with changes in arterial Hb saturation (2.2–3.0 mL/100 g) (29) and with changes in Pco2 (3.7 mL/100 g) (30). These values in infants all are lower than those estimated in human adults using singe photon emission computed tomography (4.8 ± 0.4 mL/100 g) (32) and using positron emission tomography (4.7 ± 1.1 mL/100 g) (33). The reason for smaller values of CBV in neonates than in adults is that regional CBV is smaller in white cerebral matter than in gray cerebral matter, and the relatively low mean CBV may reflect a relative preponderance of white matter compared with that in the adult brain (29). In this study, the values of CBV increased with advance of postconceptional age. This relationship between postconceptional age and CBV is based on the results of an anatomic study of cerebral blood vessels showing that the percentage of blood vessel area in gray matter and white matter increased as a function of gestational age (34). Indeed, the CBV value estimated by using spatially resolved NIRS with ICG (1.72 ± 0.76 mL/100 g) was smaller than that obtained in our study, because the gestational age of the subjects in that study (mean gestational age 28 wk; mean postnatal age 6 d) was less than that of our patients (31).

DPFs at 761, 795, and 835 nm were estimated to be 4.58 ± 0.41, 4.64 ± 0.46, and 4.31 ± 0.42, respectively. Various experimental techniques have been used to determine DPF in infants. These are based on measurements of absorption of light by time-of-flight spectroscopy (35,36), phase-resolved spectroscopy (26,37), or the water peak method (8). The mean value of DPFs estimated by using time-of-flight spectroscopy at 783 nm in postmortem infants has been reported to be 4.39 ± 0.28 (n = 6) (35) and 3.85 ± 0.57 (n = 10) (36). Duncan et al. (26,37) measured DPFs in a group of 35 infants using phase-resolved spectroscopy and calculated the mean values to be 5.11 ± 0.48 at 744 nm and 4.67 ± 0.65 at 832 nm. DPFs that were calculated by the water peak method at 730 nm and 830 nm were 4.66 ± 1.01 and 3.91 ± 0.75, respectively (8). The DPF values obtained in this study are similar to those obtained in the group of infants using phase-resolved spectroscopy.

The sensitivity and the reliability of our TRS method were previously assessed by using an in vitro model and a piglet hypoxia model. In the in vitro study, the use of intralipids and a blood phantom showed that qualitative measurements of Hb concentrations and of oxygen saturation could be made under the same conditions as those in an in vivo study. In the piglet hypoxia model, the mean Sco2 value at normoxia was calculated to be 62%, and the contributions of arterial blood and venous blood were estimated to be 41 and 59%, respectively (38). The ratio of the contribution of arterial blood to that of venous blood obtained in the study using a piglet hypoxia model is almost the same as the ratio reported by Brun et al. (39) and by Kusaka et al. (40). The results of the present study demonstrated that our TRS method can be used to monitor Sco2 and CBV. However, the number of such neonates in this study was too small for us to reach any definite conclusion, and further study is required.

In conclusion, the results of this study confirm that the new TRS is a practical method for measurements of μ's, Sco2, CBV, and DPF in neonates at the bedside in an intensive care unit.