14 / 60
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 32, No 1, January/February 2021
12
AFRICA
retinal vessel calibre models included two dependent variable
models: retinal arteries and veins. Independent variables for
these six models included
a priori
covariates, the respective
retinal artery/vein diameter and stress hormone responsiveness
(1) difference over three years (∆3yrs) (%), (2) prior to FLIP,
and (3) upon provocation (∆ FLIP%). Multiple linear regression
analyses were repeated by controlling for HRV, physical activity
and the use of cortisone derivatives,
α
- and/or
β
-blockers at
follow up.
We recently developed a method of determining risk for
chronic stress and stroke (filed 31 July 2020, international patent
application no. PCT/IB2020/05726). We applied this score to
determine whether retinal vascular responses would predict
chronic stress and stroke risk. Logistic regression analyses
were computed and included the covariates: stress hormone
changes over three years or upon provocation, HRV, diastolic
ocular perfusion pressure, HDL-C and hypertensive/diabetic
retinopathy. The statistical significance level was set at
p
≤ 0.05
(two-tailed). The
F
to enter in regression models was fixed at 2.5.
Results
Tertile characteristics (Table 1) showed an increasing trend
across u-NE tertiles for central obesity and decreasing trends in
cortisol and HDL, particularly in u-NE tertile 1. Again in u-NE
tertile 1, consistent inflammation (CRP) and raised BP were
observed, whereas a decrease occurred in u-NE tertile 3.
Stress hormones: Fig. 2A and Table 2 (median ± 95% CI) show
u-NE increases in u-NE tertile 1 (111.6%) but decreases in
u-NE tertiles 2 and 3 over three years (
p
≤ 0.01). ACTH levels
did not change in u-NE tertiles 1 and 2, however the increase
in u-NE tertile 3 was higher compared to tertile 1 (
p
≤ 0.001).
In u-NE tertile 1 (Fig. 2B), saliva post-FLIP cortisol (%) was
lower compared to u-NE tertile 2 (
p
≤ 0.05). Vein widening (Fig.
2C) was apparent in u-NE tertile 1 (245.3 MU) compared to
u-NE tertile 3 (239.5 MU) (
p
≤ 0.05). In Table 2 , medians were
compared and a five-second faster arterial constriction (Fig. 3A)
was evident in u-NE tertile 1 compared to u-NE tertile 2 (
p
≤
0.05). In Fig. 3B, veins dilated significantly more in u-NE tertile
2 when compared to u-NE tertile 3 (
p
≤ 0.05). The venous post-
FLIP recovery response was delayed (
p
≤ 0.05) in u-NE tertile
1 compared to u-NE tertile 3. In Table 4, hypertensive/diabetic
retinopathy was higher in u-NE tertile 3 compared to tertile 1.
Stress hormones and retinal vasculature associations: multiple
stepwise linear regression associations between retinal vessel
calibres (Table 3) and retinal FLIP responses (Table 5), and
stress hormones of u-NE tertile 1 are presented. Reduced arterial
dilation, faster constriction, narrowing and hypo-perfusion
were associated with increased SAM activity. Delayed venous
dilation, recovery and widening were associated with cortisol
hypo-secretion and low HDL-C (
p
≤ 0.05).
In u-NE tertile 1 (Table 6), delayed vein recovery responses
predicted stress and stroke risk, having large clinical significance
[odds ratio 4.8 (1.2–19.6);
p
= 0.03]. Associations between the
retinal vasculature and cortisol secretion in u-NE tertiles 2 and 3
showed effective cortisol functioning but no relationship existed
with norepinephrine (Table 7). Controlling for HRV, physical
activity and the use of cortisone derivatives,
α
- and/or ß-blockers
at follow up did not change the outcome of our findings.
Discussion
We aimed to (1) assess the relationships between the retinal
vasculature, SAM and HPA activity over three years and upon
provocation, and (2) determine chronic stress and stroke risk.
Findings showed that in the presence of low norepinephrine,
a reflex increase in SAM activity occurred, enhancing arterial
vasoconstriction and hypo-perfusion. Concomitant HPA
dysregulation attenuated retinal vein vasoactivity and tone,
reflecting delayed vein recovery responses and non-adaptation to
stress. These constrained vein recovery responses demonstrated
increased chronic stress and stroke risk, having large clinical
significance.The main findings are presented below (Fig. 4).
Table 7. Forward stepwise regression analyses depicting associations
between retinal vessel and stress hormone responses prior to and post
flicker light-induced provocation (FLIP) in norepinephrine:creatinine (u-NE
nmol/l:mmol/l) tertiles 2 and 3
u-NE tertile 2 median (min–max): 21.23 (15.05–28.62) (
n
= 87)
Artery max
dilation (%)
Artery time
max constric-
tion (s)
Vein max
dilation (%)
Vein post-FLIP
recovery
(% of baseline)
∆3yr stress hormones (%)
Adjusted
R
2
0.25
β
(95% CI)
< 0.10
0.20
β
(95% CI)
< 0.10
β
(95% CI)
u-NE (%)
–
–
–
–
Serum cortisol (%)
0.21
(0.03, 0.39)*
–
–
–
Stress hormone levels prior to FLIP
Adjusted
R
2
0.25
β
(95% CI)
< 0.10
β
(95% CI)
< 0.10
β
(95% CI)
< 0.10
β
(95% CI)
Saliva cortisol (nmol/l)
–0.26
(–0.46, –0.06)*
–
–
–
∆FLIP stress hormones (%)
Adjusted
R
2
0.29
β
(95% CI)
< 0.10
β
(95% CI)
0.25
β
(95% CI)
0.15
β
(95% CI)
Saliva
α
-amylase (%)
–
–
–
–
Saliva cortisol (%)
–
–
–
–0.36
(–0.60, –0.13)*
u-NE tertile 3 median (min–max): 40.62 (28.69–113.63)
(
n
= 89)
Artery max
dilation (%)
Artery time
max constric-
tion (s)
Vein max dila-
tion (%)
Vein post-FLIP
recovery
(% of baseline)
∆3yr stress hormones (%)
Adjusted
R
2
0.22
β
(95% CI)
0.12
0.20
β
(95% CI)
< 0.10
β
(95% CI)
u-NE (%)
–
–
–
–
Serum cortisol (%)
–0.22
(–0.42, –0.02)*
–
–
Stress hormone levels prior to FLIP
Adjusted
R
2
< 0.10
β
(95% CI)
0.11
β
(95% CI)
< 0.10
β
(95% CI)
< 0.10
β
(95% CI)
Saliva cortisol (nmol/l)
–
– NS
–
–
∆FLIP stress hormones (%)
Adjusted
R
2
0.15
β
(95% CI)
0.17
β
(95% CI)
0.15
β
(95% CI)
< 0.10
β
(95% CI)
Saliva
α
-amylase (%)
–
–
–
–
Saliva cortisol (%)
–
–
–
–
∆3yr; three-year stress hormone changes (%); Prior to FLIP, saliva stress hormone
levels prior to FLIP; ∆FLIP, stress hormone changes (%) obtained directly after
FLIP. ∆, changes.
Additional covariates included age and log-normalised waist circumference, cotinine,
gamma-glutamyl transferase and glycated haemoglobin; hypertensive/diabetic reti-
nopathy, diastolic ocular perfusion pressure and the respective retinal arterial/vein
diameter.