Advances in imaging technology and digital signal processing have enabled the examination of retinal vessels not only as static images of a branching vasculature with high definition and high morphological resolution, but also as dynamic structures with time varying calibre [1,2]. In association with measurement of retinal blood flow using techniques of scanning optical Doppler flowmetry , these advances have enabled quantification of haemodynamic parameters in the retinal vasculature [1,4]. Because of the close association between pressure and flow in the retinal vessels with both intraocular pressure and cerebral dynamics, measurements of changes in retinal vascular function is enabling noninvasive assessment of intracranial parameters, such as cerebrospinal fluid pressure [5,6]. Retinal haemodynamic parameters have also recently been used in studies of patients with specific blood abnormalities; spontaneous retinal venous pulsatility has been found to be reduced in patients with congenital heart disease and with cyanosis and polycythemia .
The determinants of retinal haemodynamics involve a combination of extravascular factors associated with intraocular, cerebrovascular and cerebrospinal fluid mechanics, as well as intravascular factors involved in the regulation of blood flow in microcirculatory beds and intrinsic vascular properties. Indeed, measurement of pulse transit times in arterial segments has also been shown to be possible , and pulse wave velocity in retinal vessels (which is of the order of a few mm/s) has been found to be related to age and elevated arterial pressure  in a similar manner as in large arteries (wherein it is three orders of magnitude higher, of the order of 5–10 m/s). Pulse wave velocity is conventionally used as a surrogate measure of vascular stiffness in large conduit arteries; however, it is not clear if this can also be translated to small calibre vessels. In contrast to haemodynamics of large conduit arteries, in which blood can be considered a Newtonian fluid with a constant viscosity, vessels of the calibre of those in the retinal vasculature are subject to the non-Newtonian properties of blood as well as the variable viscosity that occurs with red cells and plasma separation in small calibre tubes . These properties will also affect blood transport and gas diffusion characteristics. The interaction of these phenoma in microcirculatory beds is complex and highly nonlinear, such that it could lead to the presence of chaotic states with spontaneous oscillations in vessel calibre . An important regulator of calibre and tone of retinal blood flow has been found to be nitric oxide (NO) .
In common parlance, NO is sometimes described as the ‘third respiratory gas’ . This implies that NO binds to haemoglobin, as does oxygen and carbon dioxide. The binding occurs through the ubiquitous protein S-nitrosylation mechanism, which is being found to be a significant cardiovascular signalling pathway . Thus, NO that is produced by the endothelial cells is scavenged by haemoglobin, but can also be transported by the haemoglobin . However, because of the high reaction rate between haemoglobin and NO , other mechanisms related to degree of oxygenation of haemoglobin and the presence of an erythrocyte-free layer, due to laminar flow in vessels of microcirculatory beds, compensate for the high propensity for haemoglobin to scavenge NO, such that flow itself plays a significant role in potentiating the vasodilatory effects of NO [15,16]. Measurement of changes in retinal arterial and venous diameters with the NO synthase inhibitor NG-monomethyl-L-arginine (L-NMMA) confirmed that NO is a significant contributor to regulation of vascular tone in the retinal vasculature .
In this issue of the Journal, Ritt et al. investigated the role of haemoglobin in modulating retinal blood flow through NO-mediated effects. The study uses the technique of scanning laser Doppler flowmetry to measure vasodilatory changes in retinal capillary blood flow in response to exposure to flicker light. Studies are conducted in a cohort of 139 healthy and nondiabetic male participants in an age range 18–75 years. Measurements are made under control conditions and following the infusion of the NO synthase inhibitor L-NMMA. The range of haemoglobin content for the whole cohort was 13–17 g/dl. Results were analyzed by dividing the cohort into two groups, one (Group 1) with haemoglobin below the median of 15.5 g/dl (mean of 14.7 g/dl) and the other (Group 2) above the median (mean of 16.0 g/dl). That is, an average difference between the two groups of 1.3 g/dl (8.8%). The response to flicker light for Group 1 was a 9.52% increase in retinal blood flow and 2.83% increase in Group 2. The corresponding response to L-NMMA infusion was a reduction of 0.92 and 7.35%, respectively. Regression analysis shows low but significant correlations, in which changes in haemoglobin account for only 6.2% (r = −0.249; P = 0.004) of the variance in retinal blood flow due to flicker light and 4.0% (r = −0.20; P = 0.018) of the variance because of L-NMMA.
In an attempt to relate systemic haemodynamic variables to the findings of response of retinal blood flow, Ritt et al. considered two separate multiregression models, one including systolic pressure and the other diastolic pressure as separate parameters. Only systolic pressure is found to have a significant effect, accounting for 3.8% (r = −0.196; P = 0.037) of the variance in change of retinal blood flow in response to flicker light exposure, but not for L-NMMA infusion. It is not immediately obvious why increases in systolic pressure would be associated with a reduced vasodilatory response to flicker light exposure. One of the possible mechanisms might be related to the interaction of retinal haemodynamics with cerebrovascular and cerebrospinal fluid dynamics  as well as the complex particulate properties of blood in small calibre vessels being affected by intravascular flow conditions modulating erythrocyte and plasma separation [11,18].
The findings of the study by Ritt et al. are of interest in so far as they suggest that haemoglobin does indeed play an independent role in regulating retinal blood flow through the possible NO scavenging effects of haemoglobin [13,19]. The study is essentially observational and at this stage only establishes probable associations at best, without implying causality. However, the findings are consistent with the known mechanisms of NO in affecting vascular tone of microcirculatory beds and so regulating blood flow. In addition, small but significant differences are found in the normal range for haemoglobin. Indeed, the changes are all relatively small, in the realm of a few percent. The cohort is relatively large, so the study is sufficiently powered to detect relatively small changes. The small effects put into question the relevance to potential clinical applications. However, the statistically significant findings and the consistency with underlying mechanisms of NO-mediated effects in regulation of blood flow elucidate the complex and subtle interplay between haemoglobin, NO and fluid dynamics in the retinal vasculature. If causality can be established, this is of potential significance in investigations involving retinal haemodynamics, as it implies that blood haemoglobin content would need to be taken into account when comparing results related to retinal blood flow.
Conflicts of interest
There are no conflicts of interest.
1. Pournaras CJ, Riva CE. Retinal blood flow evaluation. Ophthalmologica
2. Vilser W, Nagel E, Lanzl I. Retinal vessel analysis: new possibilities. Biomed Tech (Berl)
2002; 47 (Suppl 1 Pt 2):682–685.
3. Michelson G, Schmauss B, Langhans MJ, Harazny J, Groh MJ. Principle, validity, and reliability of scanning laser Doppler flowmetry. J Glaucoma
4. Golzan SM, Avolio A, Graham SL. Hemodynamic interactions in the eye: a review. Ophthalmologica
5. Golzan SM, Graham SL, Leaney J, Avolio A. Dynamic association between intraocular pressure and spontaneous pulsations of retinal veins. Curr Eye Res
6. Golzan SM, Kim MO, Seddighi AS, Avolio A, Graham SL. Noninvasive estimation of cerebrospinal fluid pressure waveforms by means of retinal venous pulsatility and central aortic blood pressure. Ann Biomed Eng
7. Mojtaba Golzan S, Leaney J, Cordina R, Avolio A, Celermajer DS, Graham SL. Spontaneous retinal venous pulsatility in patients with cyanotic congenital heart disease. Heart Vessels
8. Kotliar KE, Baumann M, Vilser W, Lanzl IM. Pulse wave velocity in retinal arteries of healthy volunteers. Br J Ophthalmol
9. Kotliar KE, Lanzl IM, Hanssen H, Eberhardt K, Vilser W, Halle M, et al. Does increased blood pressure rather than aging influence retinal pulse wave velocity? Invest Ophthalmol Vis Sci
10. Fahraeus R, Lindqvist T. The viscosity of blood in narrow capillary tubes. Am J Physiol
11. Geddes JB, Carr RT, Wu F, Lao Y, Maher M. Blood flow in microvascular networks: a study in nonlinear biology. Chaos
2010; 20:045123 (1–16).
12. Dorner GT, Garhofer G, Kiss B, Polska E, Polak K, Riva CE, Schmetterer L. Nitric oxide regulates retinal vascular tone in humans. Am J Physiol Heart Circ Physiol
13. Lima B, Forrester MT, Hess DT, Stamler JS. S-nitrosylation in cardiovascular signaling. Circ Res
14. Allen BW, Stamler JS, Piantadosi CA. Hemoglobin, nitric oxide and molecular mechanisms of hypoxic vasodilation. Trends Mol Med
15. Liao JC, Hein TW, Vaughn MW, Huang KT, Kuo L. Intravascular flow decreases erythrocyte consumption of nitric oxide. Proc Natl Acad Sci U S A
16. Gladwin MT, Crawford JH, Patel RP. The biochemistry of nitric oxide, nitrite, and hemoglobin: role in blood flow regulation. Free Radic Biol Med
17. Ritt M, Harazny JM, Schmidt S, Raff U, Ott C, Michelson G, Schmieder RE. Haemoglobin and vascular function in the human retinal vascular bed. J Hypertens
18. Pries AR, Secomb TW, Gaehtgens P. Biophysical aspects of blood flow in the microvasculature. Cardiovasc Res
19. Kim-Shapiro DB, Schechter AN, Gladwin MT. Unraveling the reactions of nitric oxide, nitrite, and hemoglobin in physiology and therapeutics. Arterioscler Thromb Vasc Biol