2.1.

Blood Preparation

Fresh human plasma concentrates were centrifuged to remove remaining blood cells (leukocytes $<1\times {10}^6∕\mathrm{unit}$ ; PLTs $<5\times {10}^{10}$ /L, and RBCs $<6\times {10}^9$ /L). The first sample (plasma 1) was centrifuged at 62,000 $g$ for $1\mathrm{h}$ (Allegra 64R, Beckman Coulter) and the second sample (plasma 2) at 1600 $g$ for $30\min$ (Labofuge A, Heraeus Christ). After centrifugation, no cells could be detected in plasma 1 or were close to the detection limit, as in plasma 2. The cell concentration was measured using a clinical blood analyzer (Micros 60 OT 18, ABX Diagnostics, Montpellier, France). The linear range of the blood analyzer is 0.5 to $80\times {10}^3∕\mathrm{\mu }$ L for leukocytes, 0.2 to $7.5\times {10}^6∕\mathrm{\mu }$ L for RBCs, and 10 to $1000\times {10}^3∕\mathrm{\mu }$ L for PLTs. Highly concentrated suspensions were diluted and PLT concentrations lower than $1\times {10}^3∕\mathrm{\mu }$ L could not be detected.

A fresh leukocyte-depleted platelet concentrate from a human donor was used. The PLTs were suspended in plasma with a concentration of $1075\times {10}^3∕\mathrm{\mu }$ L, and were diluted with plasma 1 to different concentrations 5 to $100\times {10}^3∕\mathrm{\mu }$ L. To investigate the influence of the surrounding media, the PLT sample $1075\times {10}^3∕\mathrm{\mu }$ L was carefully centrifuged $(10\min$ at $1800\mathrm{rpm})$ and the plasma was substituted for 0.9% saline solution. The PLT concentration was reduced by this procedure to $674\times {10}^3∕\mathrm{\mu }$ L possibly due to residual PLT in the supernatant which was removed from the sample.

The leukocytes were also suspended in plasma. The leukocyte sample included 4500/ $\mathrm{\mu }$ L leukocytes and $465\times {10}^3∕\mathrm{\mu }$ L PLTs. The measurements of the leukocytes sample were compared to a PLT sample of the same concentration without leukocytes. PLTs were investigated within $48\mathrm{h}$ because of their short lifetime, and all measurements on leukocytes were carried out within $24\mathrm{h}$ after the cells were received from the blood bank.

The influence of the surrounding media of erythrocytes was investigated after taking RBCs from cell concentrates taken from human healthy donors and washing them three times with 0.9% saline solution. Two samples were prepared with hematocrit 8% (RBC $1.0\times {10}^6∕\mathrm{\mu }$ L) and 40% (RBC $4.8\times {10}^6∕\mathrm{\mu }$ L), respectively. Half of the sample volume was centrifuged again $(10\min ∕1600g)$ and the buffer was substituted for plasma 2 without changing the hematocrit.

2.2.

Experimental Setup

The double integrating sphere technique combined with inverse Monte Carlo simulation (iMCS) has already been shown to be useful for determination of the optical parameters of blood.

The diffuse reflectance $R_d^M$ , the total transmission $T_t^M$ , and the diffuse transmission $T_d^M$ of thin samples were measured (350 to $1100\mathrm{nm}$ ) using an integrating sphere spectrometer (Lambda 900, PerkinElmer, Rodgau-Jügesheim, Germany), which was described in detail by Meinke ¹² A miniaturized extracorporal circulation system was set up to enable the blood to flow under controlled conditions. A special flow-through turbulence-free cuvette was used providing laminar flow up to shear rates of $3000{\mathrm{s}}^{-1}$ , enabling the use of a homogeneous rectangular measurement spot of $7\times 11.5\mathrm{mm}$ . The sample thickness was adjusted using a spacer with the required thickness. For the measurement of plasma and PLTs, the sample thickness was set to $2.6\mathrm{mm}$ . Leukocytes were measured using a sample thickness of $0.52\mathrm{mm}$ and because of the high absorption of hemoglobin, a value of $0.116\mathrm{mm}$ was selected for the RBCs. The flow was adapted to the thickness to provide a constant wall shear rate of $600{\mathrm{s}}^{-1}$ at the cuvette glass surface. The blood temperature was kept constant at 20°C and the pH at 7.4.

2.3.

Evaluation of Optical Properties

The optical parameters ${\mu }_a$ , ${\mu }_s$ , and $g$ were calculated^{6, 13} by iMCS. The iMCS uses forward Monte Carlo simulations iteratively to calculate the optical parameters ${\mu }_a$ , ${\mu }_s$ , and $g$ on the basis of a given phase function and the experimentally measured values for reflection and transmission ( $R_d^M$ , $T_t^M$ , and $T_d^M$ ). The iMCS uses an initial set of ${\mu }_a$ , ${\mu }_s$ , and $g$ to calculate the resulting simulated reflectance and transmission values $R_d^S$ , $T_t^S$ , and $T_d^S$ . These are then compared to the $R_d^M$ , $T_t^M$ , and $T_d^M$ values that have been measured experimentally. By systematic variation of ${\mu }_a$ , ${\mu }_s$ , and $g$ , the deviation of the simulated $R_d^S$ , $T_t^S$ , and $T_d^S$ values from the measured ones are minimized until a set of optical parameters is found where the deviations are within an error threshold of 0.10%.

For realistic Monte Carlo simulations of photon transport in turbid media, the selection of an appropriate effective scattering phase function that is suited to the investigated medium is very important. The Reynolds-McCormick (RM) phase function^{6, 8} and the Henyey-Greenstein (HG) phase function^{14, 15} have already been discussed for the scattering of RBCs. As described by Friebel, ⁶ the best effective phase function can be evaluated using a special iMCS. When using the RM phase function, an optimal value for the alpha factor can be derived that can be used to vary the phase function. In this investigation the RM phase function was tested with $\alpha =0.5$ (corresponds to HG) to $\alpha =3$ .

The refractive index of the surrounding medium must be known for calculation of the radiation transport within the Monte Carlo simulation as well as for comparing Mie calculations. Therefore, measurements of the real part of the refractive index of plasma at 400, 500, 600, and $700\mathrm{nm}$ were carried out using an Abbe Refractometer (Carl Zeiss, Oberkochen, Germany).

3.1.

Optical Parameters of Plasma

The refractive indices of plasma were calculated to be 1.3577 at $400\mathrm{nm}$ , 1.3506 at $500\mathrm{nm}$ , 1.3473 at $600\mathrm{nm}$ , and 1.3438 at $700\mathrm{nm}$ . The Sellmeier approximation with two parameters was used to fit the refractive index of plasma from 350 to $1100\mathrm{nm}$ . The calculated refractive indices of plasma were used within the iMCS for the refractive indices of the blood samples containing plasma as medium. For the samples containing saline solution, the refractive index of water was used for the iMCS as the refractive index of the samples.

The RM phase function, where $\alpha$ is between 1 and 2, was estimated to be the most appropriate effective scattering phase function for human plasma in the range of 350 to $1100\mathrm{nm}$ . The HG phase function led to 59% of erroneous simulations with an error larger than 0.05%. Using the RM phase function, with $\alpha =1.5$ , there is an error occurrence of only 6%. Figure 1 shows the optical parameters, calculated using the iMCS, with the evaluated effective phase function using $\alpha =1.5$ after measurement of $R_d$ , $T_t$ , and $T_d$ of two plasma samples. Plasma 1 was colorless with no detectable cell concentration. Plasma 2 had a PLT concentration of about 1000/ $\mathrm{\mu }$ L and a slight orange color indicative of a different absorption spectrum. In some cases, the absorption and scattering properties of the plasma samples were too low for the wavelength ranges and this resulted in unsuccessful simulations. These data have been removed from the figures. The error bars are standard deviations (SDs) for selected wavelengths resulting from preparation, measurement, and simulation errors. For ${\mu }_a$ , the SD is $0.005{\mathrm{mm}}^{-1}$ . The SD for ${\mu }_s$ is between 0.001 and $0.003{\mathrm{mm}}^{-1}$ , the SD for $g$ ranges from 0.02 to 0.05, and for ${\mu }_s^{\prime }$ the SD is estimated to be $0.007{\mathrm{mm}}^{-1}$ .

Fig. 1

Absorption coefficient ${\mu }_a$ , scattering coefficient ${\mu }_s$ , anisotropy $g$ , and effective scattering coefficient ${\mu }_s^{\prime }$ of two different plasma samples.

The absorption coefficient ${\mu }_a$ of plasma depending on the wavelength in Fig. 1, upper left panel, is shown in direct comparison to the absorption coefficient of water. Plasma 1 shows an absorption peak at about $450\mathrm{nm}$ , with a maximum at $460\mathrm{nm}$ , reaching a value of $0.08{\mathrm{mm}}^{-1}$ . Plasma 2 shows an absorption peak at $415\mathrm{nm}$ of $0.37{\mathrm{mm}}^{-1}$ and additionally at $580\mathrm{nm}$ of $0.02{\mathrm{mm}}^{-1}$ . Above $750\mathrm{nm}$ , absorption of plasma generally follows that of water. The absorption peak of water at $990\mathrm{nm}$ is smaller and shifted to $975\mathrm{nm}$ .

The scattering behavior of plasma is illustrated in Fig. 1, upper right and the lower panels. For plasma 1 the scattering coefficient ${\mu }_s$ decreases continuously from $0.12{\mathrm{mm}}^{-1}$ at $350\mathrm{nm}$ to $0.013{\mathrm{mm}}^{-1}$ at $1100\mathrm{nm}$ . The anisotropy factor $g$ in the spectral range of 350 and $1100\mathrm{nm}$ is between 0.12 and 0.25. The curve of the effective scattering coefficient ${\mu }_s^{\prime }={\mu }_s\left(1\text{‐}g\right)$ behaves similarly to ${\mu }_s$ and decreases from $0.1{\mathrm{mm}}^{-1}$ at $350\mathrm{nm}$ to below $0.02{\mathrm{mm}}^{-1}$ above $600\mathrm{nm}$ .

The plasma sample 2 shows different optical behavior characteristics, where ${\mu }_s$ is significantly higher, about $0.05{\mathrm{mm}}^{-1}$ , over the whole spectrum. The $g$ factor is also higher with values between 0.5 and 0.7, resulting in an effective scattering coefficient ${\mu }_s^{\prime }$ comparable to that obtained for plasma 1.

3.2.

Optical Properties of Platelets in Plasma

In whole blood, as well as in PLT concentrates, provided by the Department of Transfusion Medicine, PLTs are always suspended in plasma. The RM phase function with $\alpha$ between 1 and 2 was evaluated as having the most effective phase function for the PLTs plasma suspension. The RM phase function was used for all PLT samples with $\alpha =1.5$ because only 3% simulation errors occurred in contrast to the HG phase function with a total of 30% simulation errors.

Figure 2 shows the optical parameters of various concentrations of PLTs in plasma calculated using the iMCS compared to pure plasma. The mean standard deviations of ${\mu }_a$ , ${\mu }_s$ , $g$ , and ${\mu }_s^{\prime }$ correspond to those of the plasma measurements. The PLTs in plasma have an absorption behavior very similar to that of pure plasma, indicating that additional cells do not change the absorption significantly. Only if the PLTs are highly concentrated is the maximum absorption shifted to $420\mathrm{nm}$ $\left(0.079{\mathrm{mm}}^{-1}\right)$ . The scattering coefficient decreases with wavelength and increases linearly with PLT concentration. The $g$ factor increases with wavelength up to $600\mathrm{nm}$ . Above $600\mathrm{nm}$ , $g$ shows a slight decrease. Furthermore, $g$ increases for all wavelengths with increasing PLT concentration, for example, at $500\mathrm{nm}$ from 0.2 for pure plasma over 0.34 for $5\times {10}^3\mathrm{PLT}∕\mathrm{\mu }$ L to 0.953 for $1075\times {10}^3\mathrm{PLT}∕\mathrm{\mu }$ L. The scattering coefficient decreases with wavelength. Starting from the ${\mu }_s^{\prime }$ of plasma, ${\mu }_s^{\prime }$ increases with increasing PLT concentration. Due to the error of measurement, the values for ${\mu }_s^{\prime }$ at low PLT concentrations do not show a significant difference to those of plasma. The ${\mu }_s^{\prime }$ values of the highest PLT concentration are twice as high compared to pure plasma and reach $0.121{\mathrm{mm}}^{-1}$ at $350\mathrm{nm}$ and $0.028{\mathrm{mm}}^{-1}$ at $1100\mathrm{nm}$ .

Fig. 2

Absorption coefficient ${\mu }_a$ , scattering coefficient ${\mu }_s$ , anisotropy $g$ , and effective scattering coefficient ${\mu }_s^{\prime }$ of different amounts of platelets (concentration in ${10}^3∕\mathrm{\mu }\mathrm{L}$ ) in plasma compared to the corresponding plasma.

3.3.

Optical Properties of Leukocytes in Plasma

The leukocytes, provided by the Department of Transfusion Medicine, were suspended in plasma containing PLTs of a concentration of $465\times {10}^3\mathrm{PLT}∕\mathrm{\mu }$ L.

With regard to the error of measurement, no significant differences in absorption or scattering properties appear to have been caused by adding a physiological amount of leukocytes ( $4.5\times {10}^3$ / $\mathrm{\mu }$ L) to the PLT suspension in plasma. Therefore no figures are presented for these data.

3.4.

Influence of the Surrounding Medium

Figure 3 shows the comparison of platelet suspended in 0.9% saline solution $(674\times {10}^3\mathrm{PLT}∕\mathrm{\mu }\mathrm{L})$ to PLT in plasma $(1075\times {10}^3\mathrm{PLT}∕\mathrm{\mu }\mathrm{L})$ . The mean standard deviations of ${\mu }_a$ , ${\mu }_s$ , $g$ , and ${\mu }_s^{\prime }$ correspond to those of the plasma measurements. The absorption of the saline solution suspension is significantly lower at the absorption peak at $450\mathrm{nm}$ with values of $0.01{\mathrm{mm}}^{-1}$ . Although the PLT concentration of the saline solution is lower by a factor of 1.59, the scattering coefficient is higher compared to the plasma sample. The ${\mu }_s$ of PLT in saline solution is $2.18{\mathrm{mm}}^{-1}$ at $350\mathrm{nm}$ , which is a factor of 1.3 times higher than the ${\mu }_s$ of the more concentrated plasma suspension. PLTs in saline solution show a higher $g$ factor over the whole spectrum than PLTs in plasma. The $g$ factor of PLTs in saline solution varies slightly over the wavelength range: 0.963 at $350\mathrm{nm}$ , 0.967 at $700\mathrm{nm}$ , and 0.953 at $1100\mathrm{nm}$ , whereas the $g$ factor for PLTs in plasma is lower by about 0.012 but decreases more drastically for short wavelengths going down to 0.926 at $350\mathrm{nm}$ . As a consequence, the ${\mu }_s^{\prime }$ of PLTs in plasma exceeds the one in saline solution from 350 to $500\mathrm{nm}$ . Above $550\mathrm{nm}$ , ${\mu }_s^{\prime }$ for PLTs in saline solution is approximately $0.008{\mathrm{mm}}^{-1}$ higher than that in plasma.

Fig. 3

Optical parameters ${\mu }_a$ , ${\mu }_s$ , $g$ , and ${\mu }_s^{\prime }$ of $1075\times {10}^3∕\mathrm{\mu }\mathrm{L}$ platelets in plasma and $674\times {10}^3∕\mathrm{\mu }\mathrm{L}$ platelets in 0.9% sodium chlorine solution.

In Fig. 4 the optical parameters are shown for RBCs suspended in plasma and saline solution. The cell concentrations amount to $1.0\times {10}^6∕\mathrm{\mu }\mathrm{L}$ for both samples. This corresponds to a hematocrit (Hct) of 8%. For the ${\mu }_a$ of RBCs in saline solution, the mean relative SD is 2.2%. The mean relative SDs for ${\mu }_s$ and $g$ are 1.6 and 0.056%, respectively, and for ${\mu }_s^{\prime }$ they are in the range of 1.8 to 7%. The SD for the optical properties of the plasma samples is slightly higher. The Soret absorption at $415\mathrm{nm}$ and also the two absorption maxima at 542 and $577\mathrm{nm}$ of the RBCs in saline solution exceed that of RBCs in plasma. The ${\mu }_s$ of RBCs in saline solution is about $4{\mathrm{mm}}^{-1}$ on average higher than of RBCs in plasma over the whole wavelength range. Up to $490\mathrm{nm}$ , the $g$ factor of the cells in saline solution is higher than that in plasma, whereas above this wavelength, $g$ is, on average, 0.004 lower. The effective scattering coefficient of RBCs in saline solution is between 0.2 to $0.4{\mathrm{mm}}^{-1}$ higher compared to RBCs in plasma. Investigations of RBCs at a concentration of $4.8\times {10}^6∕\mathrm{\mu }\mathrm{L}$ (Hct 40%)—not shown here—show similar results but the differences between RBCs in the two media decrease relatively.

Fig. 4

Optical parameters ${\mu }_a$ , ${\mu }_s$ , $g$ , and ${\mu }_s^{\prime }$ of $1\times {10}^6∕\mathrm{\mu }\mathrm{L}$ red blood cells in plasma and in 0.9% sodium chlorine solution, corresponding to a Hct of 8%.

4.1.

Optical Properties of Plasma

The absorption of plasma in the near IR is dominated by the water absorption. In the UV to visible range, plasma shows a significant absorption due to the chromophores found in proteins and other molecules. The plasma absorption exhibits a wide individual variability brought about by various protein concentrations, nutritive compounds, or pharmaceuticals, for example, contraceptives. Even the visual assessment of plasma samples from different donors shows a variety of colors ranging from yellow to green to orange to brown and their transitions. The absorption peak of plasma 1 at about $450\mathrm{nm}$ can be associated with the bilirubin absorption. Often the most prominent absorption peak is at $415\mathrm{nm}$ , which is due to the remaining RBCs or free hemoglobin from the plasma production. Furthermore, scattering particles, such as lipids, can result in turbid plasma samples that are usually excluded in routine plasma quality control. This means that the results shown here are exemplary.

The scattering properties of pure plasma are described by Rayleigh scattering of protein molecules, resulting in a scattering cross section that decreases with increasing wavelength $(\sim {\lambda }^{-4})$ and isotropic scattering corresponding to a $g$ factor of zero. For plasma 1, the $g$ factor is about 0.2 in the investigated wavelength region, indicating a slightly forward scattering, and ${\mu }_s$ is much higher than expected by Rayleigh scattering, calculated for a 10% albumin solution. On the basis of calculations according to Mie theory, it is estimated that particles with a size of approx. $100\mathrm{nm}$ and a total concentration of 10% must be present to explain the deviation from Rayleigh scattering. In plasma 1, these particles might be protein aggregates and lipids or cell components which could not be removed by the given centrifugation procedure.

In plasma 2, residual RBCs and PLTs would be a plausible explanation for the relatively high absorption at $415\mathrm{nm}$ and the higher $g$ value compared to plasma 1. This could lead to the assumption that the plasma samples are composed of different constituents such as molecules, aggregates of molecules, or cells of different sizes. The absorption or scattering of samples is given by the sum of the absorption coefficients ${\mu }_a^C$ or by the sum of scattering coefficients ${\mu }_s^C$ of each compound,

4.2.

Optical Properties of Platelets

As shown in Fig. 2, the addition of a physiological amount of PLTs does not change the absorption of plasma. The plasma mainly determines the absorption. This could be demonstrated by substitution of plasma by saline solution. The absorption between 350 and $500\mathrm{nm}$ is reduced by up to a factor of 10 even when there is residual plasma.

The linear increase of ${\mu }_s$ of PLTs in plasma with increasing PLT concentration suggests that measurement of the amount of PLTs would be possible by reflectance measurements. The composite $g$ factor of PLTs in plasma is lowered when the PLT concentration decreases because the influence of the low plasma $g$ factor increases. Development of devices for the determination of low PLT concentrations would be required to account for the changes in $g$ because the effective scattering coefficient does not change significantly in concentrations up to $100\times {10}^3\mathrm{PLTs}∕\mathrm{\mu }\mathrm{L}$ .

4.3.

Influence of the Medium

The scattering properties of particles suspended in a medium depend on the difference in their refractive indices and, therefore, on the refractive index of the medium. The refractive index of PLTs is higher than that of plasma. Mattley ¹⁶ determined the refractive index of PLTs to be 1.39 at $600\mathrm{nm}$ and that of plasma to be $n_{\mathrm{plas}}=1.35$ . The latter is in good agreement with the Abbe refractive index measurement of plasma determined in this study to be $n_{\mathrm{plas}}=1.347$ at $600\mathrm{nm}$ . This is significantly higher than that of water^{17, 18} $(n_{{\mathrm{H}}_2\mathrm{O}}=1.333)$ , which is not supposed to differ from the index of a 0.9% saline solution. The refractive index of RBCs is higher than that of PLTs: $n_{\mathrm{Hb}}=1.417$ at $600\mathrm{nm}$ for an intracellular hemoglobin concentration¹⁹ of $32\mathrm{g}∕\mathrm{dL}$ .

Therefore, the change in the scattering properties of PLTs or RBCs, when plasma is replaced by saline solution, is a result of an increase in the difference between the refractive indices of the cells and the medium. A calculation with Mie theory for spheres of the same volume and refractive index as PLTs or RBCs shows an increase of the scattering cross section and a decrease in the anisotropy factor when replacing plasma by saline solution. If the medium has absorption and scattering properties itself, as is the case for plasma, the absorption coefficient ${\mu }_a^C$ , the scattering coefficient ${\mu }_s^C$ and the $g$ factor are made up as shown in Eqs. 1, 2, 3.

As a consequence, changes in refractive indices, as well as in the composition of the optical parameters, are two different effects and both should be considered when comparing cells in saline solution with cells in plasma. One effect can dominate the other, dependent on the quantities of the optical parameters.

The scattering coefficients of PLTs and RBCs in saline solution are higher than those of the cells in plasma, indicating a predominating effect of the refractive index change.

The $g$ factor of PLTs in saline solution is higher than that in plasma, indicating that the very low plasma $g$ factor has led to a significant decrease in the composite $g$ factor of PLTs in plasma. This is because the scattering coefficients of plasma are nonnegligible compared to those of PLTs. This effect is more dominant for short wavelengths where the plasma ${\mu }_s$ has more influence compared to the PLT ${\mu }_s$ , due to the predominating exponential increase $\left({\lambda }^{-4}\right)$ with decreasing wavelength. As a result, ${\mu }_s^{\prime }$ of PLTs in plasma for short wavelengths exceeds the one in saline solution, but over $500\mathrm{nm}$ the effective scattering of PLTs in saline solution is higher.

The two effects for the $g$ factor of RBCs, as already discussed, are in direct competition with one another. Below $600\mathrm{nm}$ , the $g$ factor of RBCs in saline solution exceeds that of RBCs in plasma, as found for the whole wavelength range for PLTs. This effect is due to the low $g$ factor and relatively high ${\mu }_s$ for plasma, and the low $g$ factor and relatively low ${\mu }_s$ for RBCs in this wavelength region resulting in a lower composite $g$ value for RBCs in plasma. For wavelengths over $600\mathrm{nm}$ , the smaller difference in refractive index of RBCs and plasma is the dominant effect, resulting in a higher composite $g$ factor of RBCs in plasma compared to RBCs in saline solution.

These changes in $g$ and ${\mu }_s$ result in an increase of ${\mu }_s^{\prime }$ of RBCs in saline solution compared to that of RBCs in plasma over the whole wavelength range. This significant change in the scattering behavior must be considered when using RBC concentrates instead of whole blood for optical investigations.

A general discussion of the optical properties of RBCs can be found in Friebel ⁶

4.4.

Influence of Blood Components on the Optical Behavior of Human Whole Blood

To compare the different blood components, examples of the optical parameters of RBCs, PLTs, and plasma are shown in Fig. 5 . The RBCs and the PLTs are at physiological concentration (Hct 40%, $100\times {10}^3\mathrm{PLTs}∕\mathrm{\mu }\mathrm{L}$ ) and plasma 1 contains almost no residual cells. Data relating to ${\mu }_a$ , ${\mu }_s$ , and ${\mu }_s^{\prime }$ are shown on a logarithmic scale. The absorption of the RBCs exceeds the others by far. The absorption and the scattering coefficient of the RBCs are two to three orders of magnitude larger than those of PLTs or plasma. This is due to the high refractive index of the RBC.

Fig. 5

Comparison of the optical parameters ${\mu }_a$ , ${\mu }_s$ , and $g$ of the various cell compounds of blood: $4.8\times {10}^6\mathrm{RBCs}∕\mathrm{\mu }\mathrm{L}$ in plasma corresponding to a Hct of 40%, $1075\times {10}^3$ $\mathrm{platelets}∕\mathrm{\mu }\mathrm{L}$ in plasma and cell-reduced pure plasma.

The $g$ values for RBCs are, on average, higher except for the high absorption of hemoglobin ranging from 300 to $450\mathrm{nm}$ . All blood cells show strong forward-scattering behavior. The resulting effective scattering coefficients represent two orders of magnitude difference between RBCs and the other blood components. Investigations of RBCs without PLTs and leukocytes, instead of whole blood, would give results with an error of less than 1% for the scattering and absorption properties. Plasma itself shows very small absorption and scattering coefficients compared to RBCs, but its influence as a medium is an important aspect. The differences in refractive index between saline solution and plasma change the scattering properties of the RBCs by a factor of up to two.

In addition to the direct optical effects, the replacement of plasma by saline solution also leads to physiological effects that can significantly influence the optical behavior of blood. Plasma proteins cause the aggregation of RBCs at low shear rates. Plasma has a higher viscosity than saline solutions, leading to different flow conditions which then lead to a change in flow-dependent phenomena such as cell deformation or cell alignment or contribution effects such as axial migration. For example, in the case of high hematocrit levels, the scattering anisotropy of RBCs is greatly influenced by the flow conditions^{20, 21, 22} particularly the shear rate-dependent aggregation and disaggregation phenomena.^{23, 24, 25}

Holding the shear rate above the level where aggregation can occur could minimize these physiological influences. Furthermore, the flow velocity of plasma suspensions was diminished to compensate the increase in shear forces induced by the increased viscosity.

Due to the wide variation in the optical properties of plasma, investigations of RBCs should be carried out in saline solutions, which result in a higher level of measurement reproducibility.