Pulse Waveform Changes in Macular Choroidal Hemodynamics With Regression of Acute Central Serous Chorioretinopathy

Abstract

Purpose: To quantitatively evaluate the pulse waveform changes in macular choroidal blood flow by using laser speckle flowgraphy (LSFG) with regression of acute central serous chorioretinopathy (CSC).

Methods: This retrospective observational case series included 20 eyes of 20 patients with acute CSC. Laser speckle flowgraphy was performed at baseline and after 6 months. On the LSFG monochrome map, automatically divided 5 × 5 grid segments within the macula were classified into predominantly delayed filling (PDF) or minimally or no delayed filling (MDF) areas according to the degree of choroidal filling delay on early-phase indocyanine green angiography. The average mean blur rate (MBR) and the pulse waveform parameters, including the skew and blowout time (BOT), were compared between the total PDF and MDF areas during follow-up.

Results: The average MBR significantly decreased in both PDF (P = 0.005) and MDF (P < 0.001) areas during follow-up; in both areas, the skew decreased (P < 0.001 and P = 0.006, respectively) and BOT increased (P < 0.001 for each), showing significant reduction in vascular resistance at 6 months. The degree of the changes in the skew and BOT was significantly larger (P = 0.02 and P < 0.001, respectively) in the PDF area than in the MDF area.

Conclusions: Changes in the skew and BOT, indices for vascular resistance, confirmed the involvement of circulatory disturbance at the acute stage of CSC. The present findings suggested that the pathogenesis of CSC stems from imbalanced distribution of choroidal blood flow due to augmented vascular resistance.

Acute central serous chorioretinopathy (CSC) is a common disorder in middle-aged patients, characterized by serous retinal detachment in the macular region. Acute CSC typically regresses spontaneously after several months but recurs in approximately 30% to 50% of patients within several years after onset.¹ Precipitating factors for acute CSC include increased sympathetic activity,² a stress-prone personality,³ systemic corticosteroid use,⁴ and hypertension.³ Indocyanine green angiography (ICGA) demonstrates a localized delay in choroidal arterial filling along with choroidal venous dilatation⁵ and hyperpermeability observed in wider area than the initial dye leakage on fluorescein angiography (FA).⁶ Recent studies using enhanced depth imaging (EDI) optical coherence tomography (OCT) have shown a significant increase in choroidal thickness in affected eyes.^7,8 Swept-source OCT has further demonstrated an association between increased choroidal thickness and choroidal hyperpermeabiliy.⁹ These results support the mechanistic explanation of choroidal hyperperfusion or increased hydrostatic pressure leading to elevation of the retinal pigment epithelium (RPE)¹⁰ and subsequent development of RPE micro-tears corresponding to the leaking spots.¹¹ However, the mechanism underlying choroidal hyperpermeability remains incompletely understood. 

Laser speckle flowgraphy (LSFG) can noninvasively measure ocular circulation within 4 seconds, using the laser speckle phenomenon,^12–14 with the advantage of providing reproducible¹⁵ data throughout the course of diseases.¹³ The mean blur rate (MBR) is a quantitative index of relative blood flow velocity. The macular MBR, largely derived from the choroid,¹⁶ has been used to evaluate choroidal hemodynamics in various diseases.^17–26 Our recent study on CSC demonstrates that the macular MBR increases at the acute stage,¹⁹ reinforcing the suspected role of choroidal hyperperfusion in the development of acute CSC.¹⁰ 

With recent advances of LSFG, various parameters can be automatically calculated to analyze the pulse waveform of MBR averaged in a single heartbeat,^14,27–31 including the skew, blowout time (BOT), and acceleration time index (ATI), all of which are expressed as absolute values. These new indices have been reported to reflect the augmentation of blood flow resistance basically due to arterial sclerosis as well as aging (i.e., the skew increases and BOT decreases),^27–30 showing significant correlations with branchial-ankle pulse wave velocity²⁷ and the stage of glaucoma.³¹ The current study demonstrates functional data to further complement the rationale of circulatory abnormalities as the pathogenesis of CSC, showing sequential changes in these pulse waveform parameters on LSFG. 

Methods

Patients and Diagnosis

This retrospective, observational case series included 20 eyes of 20 patients (12 men, 8 women) who received follow-up examinations for more than 6 months. These patients were part of a larger population of 26 consecutive patients with treatment-naïve acute CSC who visited the Medical Retina and Macula Clinic at Hokkaido University Hospital, Sapporo, Japan, between May 2010 and October 2013. Nine of these 20 patients have been also included in our recently published case series.¹⁹ The patients' age ranged from 38 to 71 years (mean ± SD, 52.8 ± 8.0 years), and the follow-up duration ranged from 6 to 29 months (14.4 ± 7.2 months). All investigations adhered to the tenets of the Declaration of Helsinki, and the study was approved by the institutional review board and ethics committee at Hokkaido University Hospital. Informed consent was obtained from all subjects after the nature and possible consequences of the study were explained. 

Acute CSC was diagnosed according to the following criteria: detachment of the neurosensory retina at the macula, one or more leakage spots at the RPE, subsequent expansion of the leaks on FA, and hyperfluorescence of choroidal hyperpermeability regions during the middle phase of ICGA. Eyes with diffuse RPE atrophy or diffuse leakage from undetermined sources on FA were excluded from this study. Eyes with other diseases such as epiretinal membrane and age-related macular degeneration were also excluded. None of the eyes underwent any medical treatment, laser photocoagulation, or photodynamic therapy during the follow-up period. 

All patients underwent routine ophthalmic examinations including the decimal best-corrected visual acuity (BCVA), FA, ICGA (TRC-50LX; Topcon, Tokyo, Japan, or F10 Digital Ophthalmoscope; NIDEK, Gamagori, Japan), spectral domain OCT (RS-3000 or RS-3000 Advance; NIDEK), and LSFG-NAVI (Softcare, Fukuoka, Japan). The BCVA was measured with a Japanese standard Landolt visual acuity chart, and results were converted to the logarithm of the minimal angle of resolution (logMAR) for statistical analyses. Enhanced depth imaging–OCT (RS-3000 Advance) was routinely applied to eight eyes of eight patients whose first visits to our clinic were later than October 2012, and central choroidal thickness (CCT) data were collected at the first visit and 6 months later. 

Mean Blur Rate Measurements on LSFG

LSFG-NAVI was used to evaluate choroidal circulation hemodynamics. The rationale for LSFG has been described elsewhere.¹⁶ Briefly, LSFG mainly targets moving red blood cells within the choroid, using a diode laser (wavelength, 830 nm). Light reflected from the tissue produces a speckle pattern on the plane where the area sensor is focused. The faster movement of erythrocytes causes more blurring within the speckle pattern.¹⁶ The MBR is calculated from variations in the blurring and used as a quantitative index of relative blood flow velocity. 

In the present study, LSFG measurements were performed three consecutive times at the initial visit and 6 months later. The pupils of each subject were dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride 20 minutes before performing LSFG. The original MBR values were continuously recorded at 118 frames within 4 seconds (Supplementary Fig. S1A, left), followed by averaging from the entire data set to synthesize a still image corresponding to the duration of one heartbeat (Supplementary Fig. S1A, right). The average MBR was determined as the mean value of the synthesized MBR histogram during one heartbeat (Supplementary Fig. S1A; right, dotted line). To evaluate changes in average MBR, the changing rate of average MBR against the initial baseline value was used, as previously described,^17,19–26 since the MBR is an index of “relative” blood flow velocity. 

Pulse Waveform Analyses on LSFG

Pulse waveform parameters were automatically calculated by the LSFG-analyzing software (LSFG Analyzer, version 3.1.34.0; Softcare), based on the synthesized MBR image during a single heartbeat. 

The skew represents the degree of deformation of the pulse image that depends on two factors: the peak and slope of a waveform (Supplementary Fig. S1B, left).^30,31 If the waveform is bilaterally symmetrical, the skew will be zero. When the peak of the waveform is shifted leftward, the skew will become higher. The skew increases as the slope of the waveform after the peak becomes more concave. The skew is calculated as follows^30,31:  

 

 

  .

 

Blowout time is an index representing the rate of time in which the MBR is greater than half the amplitude during one heartbeat (Supplementary Fig. S1B, middle).^27,28,30 Blowout time is calculated according to the following formula^27,28,30:  

  .

 

Acceleration time index represents the time to reach the peak of MBR³¹ and is calculated as follows (Supplementary Fig. S1B, right)³¹:  

  .

 

Segmentation of Macular Area

On early-phase ICGA, choroidal filling delay was detected at the macular area (Fig. 1A, yellow arrowheads) and then used for segmentation of the region according to the degree of this circulatory change (Fig. 1B, within a yellow enclosure). On the LSFG monochrome map, the macula was automatically divided into 25 (5 × 5) equal grid segments (Fig. 1B, small white squares), each of which corresponds to 60 × 60 pixels or 2° × 2° view angle, using the LSFG analyzing software (Rubberband EX Plugin; Softcare). Next, each square was manually classified into areas of predominantly delayed filling (PDF; Fig. 1B, highlighted in blue) or minimally or no delayed filling (MDF; Fig. 1B, highlighted in red), according to the area ratio of choroidal filling delay (Fig. 1B, within a yellow enclosure) to a square exceeding 50% or not, respectively. 

Figure 1

 

Segmentation of the macular area using ICGA (A) and LSFG (B). Choroidal filling delay at the macular area ([A], yellow arrowheads) on the early-phase ICGA image. The macular area automatically divided into 25 equal grid segments (small white squares) on the LSFG monochrome map, with each square classified into two areas: PDF ([B], highlighted in blue) area and MDF ([B], highlighted in red) area, according to the area ratio of choroidal filling delay ([B], a yellow enclosure) to a square exceeding 50% or not, respectively. Average MBR values (C) from each of the PDF (blue values) and MDF (red values) squares at the baseline (left) and regression (right) stages.

The average MBR (Fig. 1C), the skew, BOT, and ATI were automatically calculated in each segment, and the mean values from the total PDF (Fig. 1C, blue values) or MDF (Fig. 1C, red values) squares were processed for statistical analyses to compare parameters between the total PDF and MDF areas and between the baseline (Fig. 1C, left) and regression (Fig. 1C, right) phases. 

Evaluation of Ocular Hemodynamics

Within a certain range, the relationship between choroidal blood flow and ocular perfusion pressure (OPP) is bilinear in healthy subjects with normal eyes, as previously demonstrated.³² To exclude the possibility of such physiological responses from the present results, the blood pressure and intraocular pressure (IOP) of patients were measured to calculate OPP. Mean blood pressure (MBP) was calculated from systolic blood pressure (SBP) and diastolic blood pressure (DBP), according to the following equation: MBP = DBP + ⅓(SBP − DBP). Ocular perfusion pressure was calculated by using the following equation: OPP = ⅔ MBP − IOP.¹⁷ 

Statistical Analyses

All results are expressed as the mean ± SD. The Wilcoxon signed rank test was used to compare changes in the logMAR BCVA, CCT, the changing rate of average MBR, pulse waveform parameters, and OPP. For relationship between the MBR changing rate and OPP, multiple regression analysis was used. All statistical analyses were performed by using a publicly available software program (“R” version 2.15.3; The R Foundation for Statistical Computing). For all tests, P values less than 0.05 were considered significant. 

Results

Patients' Characteristics

Patients' clinical characteristics are summarized in Table 1. Among the 20 patients, well-controlled hypertension was present in four patients and the previous history of systemic corticosteroid use was seen in two patients with idiopathic thrombocytopenic purpura and rheumatoid arthritis. After developing CSC, corticosteroid therapy was discontinued in both of these patients. Serous retinal detachment spontaneously regressed in all eyes within 6 months after the initial visit with an average duration of 2.7 ± 1.2 months. The mean logMAR BCVA was 0.15 ± 0.19 at the initial visit and 0.05 ± 0.22 at 6 months. The logMAR BCVA significantly improved at 6 months compared to that at the initial visit (P = 0.02). Central choroidal thickness significantly decreased from 376.6 ± 125.4 μm to 322.1 ± 106.8 μm during the 6-month follow-up in eight eyes examined (P = 0.006). 

Table 1

 

Clinical Characteristics of Patients With Acute CSC

 

Laser Speckle Flowgraphy Findings (MBR and Pulse Waveform Parameters)

Figure 2 shows typical changes in the MBR pulse waveform of a representative case (case 5). During the 6-month follow-up, the overall height of pulse waves became much lower, causing the substantial decline of average MBR values in whole, PDF, and MDF areas. Notably, the reduction of steepness (distance between peak and trough) at 6 months as vertical changes in the waveform appeared to affect the sequential changes in the skew and BOT values (decreases and increases, respectively) in all areas. In contrast, there was no apparent peak shift as horizontal changes in the waveform with all the peaks remaining at frame 6 or 7, in consistence with the tendency of unremarkable changes in ATI values over time. 

Figure 2

 

Typical changes in the pulse waveform of a representative case (case 5) from the baseline (left) to regression phase (right) in whole (top), PDF (middle), and MDF (bottom) areas.

Indeed, Table 2 shows statistical results from the 20 CSC eyes on temporal (0M to 6M) changes in the mean values of average MBR, the skew, BOT, and ATI, so as to confirm the morphologic impact on pulse waves (Fig. 2). Unlike other waveform parameters, ATI showed no significant changes in all areas during follow-up (Table 2). The changing rate of average MBR against baseline significantly decreased at 6 months in each area of the macular region: −16.0% ± 2.0% in the whole area, −11.1% ± 3.0% in the PDF area, and −17.5% ± 2.5% in the MDF area (Supplementary Fig. S2A). The decreasing rate of average MBR was significantly larger (P = 0.005) in the MDF area than in the PDF area (Fig. 3). The skew (Supplementary Fig. S2B) and BOT (Supplementary Fig. S2C) values significantly decreased and increased, respectively, after 6 months as compared with baseline in each area. Importantly, the skew and BOT levels at baseline were significantly higher (P = 0.01) and lower (P = 0.03), respectively, in the PDF area than in the MDF area, suggesting the role of increased choroidal vascular resistance in the pathogenesis of acute CSC. Accordingly, the degree of the changes in the skew and BOT during the 6-month follow-up was significantly larger (P = 0.02 and P < 0.001, respectively) in the PDF area than in the MDF area (Fig. 3). 

Table 2

 

Average MBR and Pulse Waveform Parameters on LSFG

 

Figure 3

 

Comparison of changing rates of parameters between PDF and MDF areas. The degree of changes in the skew and BOT during the 6-month follow-up was significantly larger in the PDF area than in the MDF area, whereas the decreasing rate of average MBR was significantly higher in the MDF area than in the PDF area.

Ocular Perfusion Pressure Data

The OPP values were obtained in 17 of 20 eyes during the 6-month follow-up. The mean OPP was 49.6 ± 11.1 mm Hg at baseline and 46.8 ± 12.1 mm Hg at 6 months. There was a significant difference in OPP between baseline and regression phases (P = 0.02); however, the multiple regression analysis (objective variable: MBR; explanatory variables: the follow-up period, patients, OPP) demonstrated no significant association between MBR and OPP (P = 0.3), although a correlation between MBR and the follow-up period was statistically significant (P < 0.05). 

Discussion

Our present study is the first to show pulse waveform changes in macular choroidal hemodynamics with regression of acute CSC. The current data on the sequential reduction of average MBR, an index of choroidal blood flow velocity, confirmed the robust reproducibility of our recent report¹⁹ showing the association of the initial MBR elevation with poor visual prognosis, supporting the rationale of choroidal hyperperfusion as the pathogenesis of acute CSC. More importantly, the pulse waveform analyses revealed that the skew and BOT values significantly changed over time, suggesting the involvement of increased vascular resistance especially in the acute phase of CSC. Furthermore, these pulse waveform changes were more prominent in the PDF area than in the MDF area, suggesting a close link between angiographic filling delay and functional blood flow resistance as a mechanistic explanation. In stark contrast, the degree of changes in average MBR during follow-up was less prominent in the area with profound arterial filling delay, suggesting the regional restriction of choroidal hyperperfusion at the acute stage of CSC. Reasonably, vascular resistance was elevated more notably in the ill-perfused area, leading to the relative decline of elevated blood flow velocity at baseline; therefore, the discrepancy between the pulse waveform parameters and average MBR is seemingly contradictory but may in fact be complementary. 

A previous analysis using subtracted images on ICGA has demonstrated early choroidal dye-filling patterns in normal volunteers and CSC patients.³³ Initially, the dye propagation in CSC eyes showed multiple patches with a significant time delay, a pattern different from normal volunteers, suggesting that blood flow from choroidal arterioles to the choriocapillaris is distributed in an imbalanced way in the macula of CSC eyes.³³ In concert with this angiographic observation, the current data led us to hypothesize (Fig. 4) that the etiology of increased macular MBR in acute CSC eyes lies in local vasoconstriction of choroidal arterioles, possibly due to sympathetic α-adrenoceptor activation, subsequently disturbing a perfusion into the choriocapillaris, and finally leading to a secondary passive overflow into the surrounding large choroidal veins via alternative pathways such as adjacent branches of circulatory units of lobules. In parallel, a net increase in entire choroidal blood flow is theorized to result from cardiac output elevation, possibly due to sympathetic β-adrenoceptor activation, in consistence with our present data showing a temporal reduction in OPP along with regression of CSC. Comparably, a substantial elevation of OPP, achieved by isometric exercise-induced activation of the sympathetic nervous system, increases choroidal blood flow more remarkably in patients with a history of CSC than in healthy subjects,³⁴ supporting our rationale (Fig. 4) of choroidal vascular dysregulation in response to increased sympathetic activity in acute CSC patients. 

Figure 4

 

Schema showing a possible mechanism of acute CSC in response to increased sympathetic activity. Adrenoceptor-mediated vascular dysregulation leading to augmented and imbalanced distribution of choroidal blood flow via adjacent lobular units. PED, retinal pigment epithelial detachment; SRF, subretinal fluid.

Indeed, the combined application of LSFG and ICGA in the present study may clearly represent these choroidal circulatory abnormalities: the pulse waveform changes as a result of arteriole vasoconstriction, the angiographic filling delay as a result of disturbed capillary perfusion, and the baseline elevation of blood flow velocity as a result of large vessel overflow. Moreover, a recent EDI-OCT study on CSC eyes shows thinning of the inner choroidal layers and enlargement of the underlying hyporeflective lumens (i.e., choroidal middle or large vessels),⁸ supporting our hypothesis on a morphologic basis. 

This study had a few limitations. This was a retrospective study with a relatively small population. The pulse waveform parameters analyzed on LSFG have been validated only recently. There was no significant correlation between visual recovery and any of the pulse waveform parameters currently examined (data not shown). Further studies are needed to establish the functional significance of these parameters. 

In conclusion, the currently observed changes in the skew and BOT, new and absolute indices for vascular resistance, further confirmed the involvement of circulatory disturbance at the acute stage of CSC. Our findings on LSFG suggested that the pathogenesis of CSC stems from imbalanced distribution of choroidal blood flow due to an increase in vascular resistance, possibly related to sympathetic activation. These pulse waveform parameters, in concert with MBR values, may be useful to quantitatively follow the activity of acute CSC. 

Acknowledgments

The authors alone are responsible for the content and writing of the paper. 

Disclosure: M. Saito, None; W. Saito, None; K. Hirooka, None; Y. Hashimoto, None; S. Mori, None; K. Noda, None; S. Ishida, None