Klein BEK, Klein R, Myers CE, Lee KE. Complete Blood Cell Count and Retinal Vessel Diameters. Arch Ophthalmol. 2011;129(4):490-497. doi:10.1001/archophthalmol.2011.57
To examine the cross-sectional associations of components of the complete blood cell count with retinal vessel diameters.
The data are from the baseline examination of the Beaver Dam Eye Study cohort (n = 4730) from March 1, 1988, to September 14, 1990. Blood pressure was measured, a medical history including questions on cigarette smoking was obtained, and fundus photographs centered on the optic disc were taken and digitized. Retinal arteriole and venule diameters were measured using computer-assisted software. The central retinal arteriole equivalent and central retinal venule equivalent were computed. A complete blood cell count was done.
In age- and sex-adjusted analyses, red blood cell count, hemoglobin level, hematocrit, and white blood cell count were all statistically significantly associated with central retinal venule equivalent and central retinal arteriole equivalent, while platelet count was associated only with central retinal venule equivalent. These relationships persisted in more fully adjusted models, except platelet count became statistically significantly associated with both central retinal arteriole equivalent and central retinal venule equivalent.
Blood components as measured in a complete blood cell count are significant correlates of retinal vessel diameters and should be considered in analyses where retinal blood vessel diameters are outcomes.
Retinal vessel diameters are associated with a variety of systemic traits and diseases. For example, narrower retinal arterioles are associated with higher blood pressure,1- 4 myocardial perfusion,5 left ventricular remodeling,6 and stroke.7 Ikram et al8 have found that wider retinal arterioles are associated with impaired fasting glucose and diabetes. Wider retinal venules are associated with severity of diabetic retinopathy,9 and relatively smaller retinal venular diameters were found in persons with diabetes compared with those of similar age without type 2 diabetes9 and with incident proteinuria in persons with type 2 diabetes.10 Larger retinal venule diameters were associated with a variety of atherosclerosis indicators, such as higher carotid plaque scores and greater aortic calcifications, and risk factors such as body mass index and lower high-density lipoprotein cholesterol levels in the Rotterdam Study.11,12
It has been reported that hematocrit is significantly associated with retinal venular diameter,13 although in the few studies that have investigated this, relationships are not consistent.14 Also, white blood cell (WBC) count has been found to be associated with retinal venular diameters in a large population-based study.15 High WBC count, as is characteristic of leukemia,16 can alter the microvessel diameters owing to the high concentration of cells. Furthermore, it has been demonstrated in an experimental setting that leukocytosis in the peripheral and retinal circulation is associated with dilation of retinal arterioles and retinal venules.17 When retinal arterioles are embolized with glass microspheres in a laboratory setting, the diameters of the embolized vessels are clearly dilated to accommodate the particulate emboli.18
Because retinal vessels (arterioles and venules) are parts of dynamic microvascular beds that may reflect many physiological factors as well as the substances within them, we hypothesize that these structures may be influenced by the cells measured in the complete blood cell count in addition to other physiological factors. We examine this hypothesis in population-based data collected as part of the Beaver Dam Eye Study.
There were 4926 persons aged 43 to 86 years who participated in the baseline examination of the Beaver Dam Eye Study from March 1, 1988, to September 14, 1990. Ninety-nine percent of the population was of European ancestry.19 Informed consent was obtained from participants, and all examinations followed a similar protocol that was approved by the institutional human subjects committee of the University of Wisconsin and conformed to the tenets of the Declaration of Helsinki.
A medical history was obtained during the study evaluation. This included questions about cigarette smoking.
A complete blood cell count (Coulter counter method), a standardized medical history, measurements of blood pressure by standard protocol,20 and ocular examination with retinal photographs were obtained.20 Photographs of Early Treatment Diabetic Retinopathy Study field 1 were digitized and retinal vessel diameters were measured from these images with a computer-assisted standardized grading protocol.21,22
Stereoscopic 30° color retinal photographs were converted to digital images by a high-resolution scanner (Nikon LS2000; Nikon Corp, Tokyo, Japan) using standard settings for all photographs. Images were displayed on 21-in monitors set to 1600 × 1200 pixels. The diameters of all arterioles and venules coursing through a standard area 0.5 to 1.0 disc diameter from the optic disc margin (zone B) were measured using a customized computer program (Retinal Analysis; University of Wisconsin, Madison) according to a standard protocol. Each vessel was identified as an arteriole or venule by a trained grader, using the original color photograph for reference. The grader selected a segment of the vessel in zone B for measurement and used various tools to determine the validity of the measurement. A measurement was considered invalid if the grader could not get an accurate measurement after 3 attempts. The entire image was rejected if more than 1 arteriole or venule larger than 40 μm in diameter could not be measured accurately. This cutoff was based on preliminary data showing that vessels smaller than this diameter had no substantial effect on the summary values.21
Cigarette smoking status was determined according to the following criteria: a never smoker had smoked fewer than 100 cigarettes in his or her lifetime; a past smoker had smoked 100 or more cigarettes in his or her lifetime but reported that he or she was no longer smoking; and a current smoker had smoked more than 100 cigarettes in his or her lifetime and reported that he or she was still smoking. Pack-years were calculated by first dividing the average number of cigarettes smoked per day by 20 and then multiplying by the number of years smoked.
Statistical analyses were conducted in SAS version 9 statistical software (SAS Institute, Inc, Cary, North Carolina). Means were compared for statistically significant differences by the t test or analysis of variance when 2 or more than 2 groups, respectively, were involved. Because the distributions of the blood elements were highly skewed and differed by sex, each blood element was analyzed using sex-specific quintiles. We computed the quintiles for blood elements for each sex separately but analyzed the data with both sexes together (ie, quintile 1 [Q1] for men and Q1 for women were grouped together as Q1 for the entire group). Central retinal arteriolar equivalents (CRAEs) and central retinal venule equivalents (CRVEs) were computed based on the diameters of the 6 largest of each vessel type in standardized digitized photographs of the optic discs.21 Univariate analyses between each variable and CRAE and CRVE were conducted using simple linear regression. In additional analyses, age, sex, smoking status, diabetes status, and the other vessel measurement were added into a model sequentially to determine the effect on model fit. A final model for each blood element was developed using stepwise model selection.
Of the 4926 individuals who were interviewed and examined at the baseline examination, individuals were excluded from the analyses if there was no measurement for CRAE or CRVE in either eye (n = 114), if data were missing for any of the laboratory values used in the analysis (n = 14), or if the individual had a condition or disease such as leukemia, thrombocythemia, or polycythemia (n = 68). This left 4730 individuals eligible for analysis.
Excluded individuals were more likely to be older, female, and hypertensive and to have lower diastolic blood pressure and higher levels of glycated hemoglobin and all other blood elements except platelet count than those included. There were no differences in systolic blood pressure, pack-years smoked, platelet count, or current smoking between the 2 groups (Table 1).
An increase of 1 sex-specific quintile of WBC count, red blood cell (RBC) count, hemoglobin level, and hematocrit was significantly associated with larger CRAE and CRVE (Table 2). An increase of 1 sex-specific quintile of platelet count was also significantly associated with larger CRVE. Additional adjustment for age did not change the significance of these relationships. When the elements were considered as continuous variables, the relationships were unchanged from when they were considered as quintiles.
To investigate the relative strengths of the relationships of the blood elements to the vessel diameters, we developed models in which we sequentially included other known correlates of those diameters (Table 3). For all models, each element except platelet count was significantly associated with CRAE. The models indicate that including more of the additional variables increases the informativeness of the model measured by the change in R2. For all models, each blood element was significantly associated with CRVE. Additional adjustments for additional variables increased the R2 of the model. The presence of diabetes has a small incremental effect on the fit of the models.
We next performed the modeling using a stepwise approach including as possible variables all those considered in model 4 of Table 3 (Table 4 and Table 5). For CRAE, systolic blood pressure had the highest partial R2, followed by smoking status, blood element, and diabetes for each blood element. In the model assessing the contribution of platelet count to the other variables associated with CRAE, platelet count no longer contributed a significant amount of information. For CRVE, smoking status had the greatest partial R2, followed by the blood element and then sex and systolic blood pressure. Because of the importance of relative diameter, we repeated the analyses in Table 4, this time including CRVE in the models for CRAE and including CRAE in models for CRVE (Table 5). These additions markedly improved the R2 of all models. In all models, the relative importance of smoking status was diminished. The blood elements are the third most informative variable for CRVE or the fourth most informative variable for CRAE when controlling for the other vessel measurement.
While the previous analyses considered each blood element individually, we next developed models in which they were included together. We chose only 1 of the 3 measures of RBC status (RBC count, hemoglobin level, and hematocrit) for the purposes of these analyses. We modeled the outcomes as was done in Table 4 and Table 5. In models where CRAE was the outcome and CRVE was not included, each of the blood element components added significantly to the models, with the RBC variables being most informative and WBC and platelet counts adding less information. When CRVE was included, the WBC count was no longer significant in the models. In models where CRVE was the outcome and CRAE was not included, each of the blood elements added significantly to the models; again, the RBC variables were the most informative and WBC and platelet counts added less information. When we added CRAE, all of the 3 blood components remained significant in most models. The relative importance of most of the other variables that were used in the models in Table 4 and Table 5 remained unchanged.
We have found that hematocrit, hemoglobin level, RBC count, WBC count, and platelet count are significantly associated with retinal vessel diameters. In addition, in our population these factors outweigh the relative importance of age, sex, and diabetes status in all models for both CRAE and CRVE. The effects related to RBC count are greater than those for WBC count. There are relatively smaller effects of platelet count. The associations are direct, ie, higher values are associated with wider diameters. While these data may indicate that these effects are related to that actual burden imposed by more particulate matter in the blood, it may be more informative to consider the physiological factors related to the blood elements (counts). It is possible that relative systemic factors such as chronic hypoxia due to systemic diseases or exposures may be associated with relatively higher RBC count (as well as hemoglobin level and hematocrit). An obvious exposure that is compatible with this notion is smoking, which we found to be associated with both CRAE and CRVE in analyses not adjusting for the diameter of either. Smoking, aside from potential effects on systemic hypoxia, is also associated with systemic inflammation. This mechanism has been invoked to explain previously described associations of WBC count to CRVE. In the Atherosclerosis Risk in Communities Study, higher WBC count was also associated with larger CRAE.13 Diabetes status, while significant in its effect, had a relatively small contribution compared with cigarette smoking for both CRAE and CRVE in our study.
Microvascular diameters undergo structural adaptation related to the local metabolic conditions.23 Hematocrit influences local hemodynamics.23 Vessel diameters are thought to change shear stress on vessel walls, dependent in part on oxygen deficit and RBC signaling.23 In the Rotterdam Study, retinal venule diameter was greater in persons with decreased cerebral blood flow and with decreased arteriolar oxygen saturation.11,24 Oxygen stimulation is associated with decreased flow-mediated diameters in large vessels.25 While this may have implications for cardiovascular events, it is uncertain what the health effect of changes in the microvascular beds may be.
Another possible effect (or correlate) of hematocrit is an increase in blood viscosity.26 This increase in viscosity has been shown to be associated with risk factors for cardiovascular disease such as blood lipid levels and blood pressure27 and may also be a pathway whereby the blood factors we examined influence retinal vessel diameters.
Our study is limited in that it is a cross-sectional analysis and these findings may not be applicable to longitudinal data. Also, while we have controlled for the most important confounders, there may be others that we did not measure or that were not important in our analyses but may be important in other studies. Lastly, variability of our outcome (as well as our covariate) measurements may have influenced our findings, although we suspect that these would have led to our underestimate of the relationships we report.
In summary, we have found that blood cell counts (and related measures of hematocrit and hemoglobin level) are significant correlates of retinal vessel diameters with greater effects than have been reported for some other correlates of these measures. We suggest that in research where retinal vessel diameters are the outcomes of interest, blood cell counts should be included as important determinants or correlates.
Correspondence: Barbara E. K. Klein, MD, MPH, Department of Ophthalmology and Visual Sciences, School of Medicine and Public Health, University of Wisconsin, Madison, 610 N Walnut St, Fourth Floor WARF, Madison, WI 53726-2336 (firstname.lastname@example.org).
Submitted for Publication: May 13, 2010; final revision received August 26, 2010; accepted September 13, 2010.
Author Contributions: Dr B. E. K. Klein had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This research is supported by grant EY06594 from the National Institutes of Health (Drs B. E. K. Klein and R. Klein). The National Eye Institute provided funding for the entire study, including collection and analyses of data. Additional support was provided by Senior Scientific Investigator Awards from Research to Prevent Blindness (Drs B. E. K. Klein and R. Klein).
Disclaimer: The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Eye Institute or the National Institutes of Health.
Additional Contributions: Heidi Gutt, BA, and Mary Kay Aprison, BS, assisted in editing and preparing the manuscript.