From the initial group of 290 glaucoma suspects who fulfilled the inclusion criteria, 2 did not provide informed consent, 20 did not complete all of the required tests, and 6 were unable to perform at least one of the tests included in the study protocol. These 28 cases were excluded and finally 262 eyes of 262 patients of Caucasian origin were included in the statistical analysis. The images obtained with the HRT3 for all preselected individuals had at least “acceptable” quality. The Kolmogorov-Smirnov test confirmed that all variables analyzed in this study followed a normal distribution.
The COVs in the OA were 7.4%, 8.5%, 8.2%, and 1.3% for PSV, EDV, MV, and RI, respectively. In the CRA, the COVs for PSV, EDV, MV, and RI were 9.1%, 10.3%, 10.0%, and 1.8%, respectively. In the TPCA, the COVs for PSV, EDV, MV, and RI were 15.5%, 15.7%, 15.3%, and 2.5%, respectively. In the NPCA, the COVs for PSV, EDV, MV, and RI were 15.2%, 16.6%, 16.4%, and 2.9%, respectively.
The study sample included 122 males and 140 females. Mean age was 51.5 ± 11.0 years, mean IOP was 23.56 ± 2.4 mmHg, and mean deviation of SAP was −0.49 ± 1.4 decibels. As defined above, according to the changes in the MRA over the 48-month follow-up, the sample was divided into two groups: 36 MRA-converters (13.7%) and 226 nonconverters (86.3%,
Fig. 1).
Table 1 shows the clinical characteristics of each group included in the study at baseline. There were no differences in age, best-corrected visual acuity (BCVA), IOP, central corneal thickness, mean deviation of SAP, pattern standard deviation of SAP, systolic and diastolic blood pressure, OPP, plasma glucose, and lipid profile between MRA-converters and nonconverters. Nevertheless, MRA-converters had a higher vertical-cup-to-disc ratio in stereophotographs than the nonconverter group. No significant changes in IOP were detected in any of the groups during the follow-up period.
Forty-seven eyes (17.9%) showed a significant negative trend based on the mean deviation of SAP. Twenty MRA converters also showed visual field worsening (55.5%). The agreement for detecting conversion to glaucoma evaluated by structural and functional tests was fair, with a kappa statistic of 0.38 (standard error, 0.07) between MRA conversion and visual field deterioration.
Table 2 shows the color-coded parameters of the MRA classification of the population at baseline. Most sectors in most individuals were within normal limits, and the most frequent abnormal sector was the nasal inferior (11.1%). At the beginning of the study, very few MRA sectors were outside normal limits, but around 90% of participants had at least a borderline sector.
EDV, RI, PI, and S/D of the OA were significantly different at baseline (Student
t-test,
P < 0.05) between MRA converters and nonconverters (
Table 3). The OA was the vessel with the largest difference between groups, and all hemodynamic parameters of the OA, except PSV and MV, were significantly different. The MRA converters had reduced EDV of the OA and increased RI and PI of the OA compared with the nonconverters.
Because RI is one of the most reliable hemodynamic parameters
34–38 (in this study, intraclass correlation coefficient for three consecutive measurements of RI of the OA was 0.968) and the greatest number of significant differences between MRA converters and nonconverters was in the OA, study authors calculated the survival and HR of glaucoma suspects with an RI of the OA, choosing 0.75 as the RI cut-off point for conversion to glaucoma (which was the mean RI of the OA obtained in the group of nonconverters and the mean RI of the OA described by Galassi et al. in a healthy population).
33 Thus, the whole sample was classified into two subgroups, according to the RI of the OA: 98 individuals had an RI lower than 0.75 and the remaining 164 had an RI greater than 0.75. At the end of the 48-month follow-up period, 30 patients in the group with an RI greater than 0.75 and 6 patients in the group with an RI lower than 0.75 met the criteria for conversion to glaucoma. The group with RI values lower than 0.75 in the OA had a survival rate of 93.9% at the 48-month follow-up (
Fig. 2), while the probability of survival in the group with RI values greater than 0.75 decreased to 81.7% (
P = 0.007; log-rank test). RI values over 0.75 in the OA positively correlated with conversion to glaucoma (HR: 3.306; 95% CI = 1.448–7.547;
P = 0.002). Based on the results of Galassi et al.,
34 study authors also selected an additional cut-off point of 0.78 for the RI of the OA (150 participants had an RI of the OA < 0.78 and 112 had an RI of the OA ≥ 0.78). With this condition, the HR increased slightly up to 3.757 (95% CI = 1.812–7.794;
P < 0.001).
Table 4 shows the RBF velocities for both the 0.75 and 0.78 RI-defined subgroups. RI, PI, and S/D of the OA, CRA, and TPCA differed between each RI-defined subgroup. EDV and MV only differed in the OA, and RI of NPCA was only different for the 0.78 RI-defined subgroups.
Neither the clinical variables (age, triglycerides, HDL cholesterol, LDL cholesterol, glycemia, systolic blood pressure, diastolic blood pressure, and OPP) nor the ophthalmic parameters (IOP, central corneal thickness, mean deviation, and pattern deviation of SAP) were significant in the Cox regression model; only the vertical cup-to-disc ratio evaluated by stereophotographs was significant (HR 1.973 per 0.1 increase; 95% CI = 1.106–3.518; P = 0.021).
With respect to the RBF velocities, each unit increase in the PSV of the OA resulted in a 6.2% increase in the risk of developing glaucoma (HR: 1.062; 95% CI = 1.008–1.120; P = 0.024). By contrast, each unit increase in the EDV of the OA decreased the risk of conversion to glaucoma by 30% (HR: 0.699; 95% CI = 0.552–0.884; P = 0.003). With respect to the CRV, each centimeter per second decrease in V min was associated with a 66% greater increase in the risk of developing glaucoma (HR: 0.342; 95% CI = 0.152–0.767; P = 0.008).
Because the vertical cup-to-disc ratio in the stereophotographs differed at baseline between the MRA converter and nonconverter groups, further analysis was performed. The whole sample was stratified into two subgroups according to this parameter. A 0.5 cut-off point was selected because it was the mean cup-to-disc ratio of the nonconverter group and the median value of the range for this variable. The clinical characteristics of both subgroups are detailed in
Table 5. The vertical cup-to-disc ratio did not differ between MRA converters and nonconverters of each subgroup. The RI of the OA was higher in the MRA converters with a cup-to-disc ratio higher than 0.5 (
Table 6). In the subgroup with a cup-to-disc ratio ≤0.5, the Mann-Whitney
U test was used because the MRA converters group included only six cases. RBF velocities in the OA, CRA, TPCA, NPCA, and CRV were not significantly different in this subgroup.
All 36 MRA converters were treated with ocular hypotensive drugs and neither HRT3 nor SAP showed any significant worsening through the end of the follow-up period.