The burden of diabetes mellitus (DM) is growing globally, with a worldwide estimated prevalence of 4.1% in 2010 expected to increase to 7.7% by 2030 (1, 2). In China, up to 11.6% of the population aged 18 years and older has diabetes (3). Diabetic microvascular complications such as retinopathy (DR) and macular edema (DME) are the major causes of blindness in DM patients in China (4). Diabetic kidney disease (DKD) is another common microvascular complication that leads to kidney failure in approximately 40% of diabetic patients (5). Our group found that the elevation of urine albumin-to-creatinine ratio (UACR) or the worsening in UACR stages, a signature change for renal function in DKD, was correlated with DME (6) and macular thickening (7). Clinical evidence also supports this connection. DME patients with better renal function had greater chances of better prognoses, especially in refractory DME (8, 9). DKD and DR share similar pathogenesis (10). Hyperglycemia-induced inflammation and oxidative stress lead to loss of endothelial cells, pericytes, and disruption of cell junctions, which contribute to the breakdown of vascular barrier in both the kidney and retina (10). For DME, hyperglycemia-linked pathways also contribute to dysfunction of retinal pigment epithelium (RPE) cells and glial cells such as Müller cells, which normally drain the fluid in the retina to the systemic circulation and keep the retina dehydrated (11, 12). Increased entry of fluid through the damaged blood-retinal barrier (BRB) and decreased drainage result in accumulation of fluid, especially in the macula (13).

Interestingly, while DKD has been proven associated with the elevation in extracellular fluid volume (14), a recent study discovered that fluid overload also appeared in DME patients (15). Imbalance of hydrostatic and oncotic pressure explained by starling equation could account for the phenomenon. The fluid status was monitored by body composition measurement (BCM) objectively and non-invasively, which is widely assumed credible assessing the fluid status of patients with chronic kidney diseases in clinical practice (16, 17). DME characterized by exudative fluid accumulation in the macula was speculated as a focal manifestation of extracellular fluid overload. The accessibility of extracellular water to total body water (ECW/TBW) ratio makes it the perfect systemic fluid marker for DME studies and a potential biomarker for DME treatments (18). Still, the confounding effect between DKD and fluid overload has not been discussed deeply regarding the relationship with DME. There lacked evidence of whether the fluid expansion came up through DME pathophysiological alterations along or shared pathogenesis between DME and DKD.

In this paper, the fluid statuses among different DM patients, including DR patients with or without DME and diabetic patients with or without DR, were investigated. By further exploiting associated changes of fluid status and renal function parameters in DME, this study proposed a novel hypothesis for the role of body fluid status in DME mechanisms. It provided significant clinical evidence to guide future DME management and treatments.

This was a cross-sectional observational study. Diabetic patients diagnosed with or without DR and DME by a multi-disciplinary team at Guangdong Provincial People's Hospital were prospectively included from May 2020 to February 2021. Medical records were reviewed by a chief-resident doctor from the department of ophthalmology and a chief-resident doctor from the department of endocrine of Guangdong Provincial People's Hospital. All patients included were informed of the study contents and signed written consent. This study was conducted following the Declaration of Helsinki and under the supervision of the Research Ethics Committee of Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences (registration number: GDREC2018380H).

One hundred thirty-two patients (73 male and 59 female) with 132 eyes were enrolled in the study. Table 1 shows the details of general characteristics, systemic factors, ocular parameters and BCM parameters among the DM group (n = 35), DR group (n = 31), and DME group (n = 66). As for general characteristics, there was no statistically significant difference in sex, age, hypertension, blood pressure, weight among the three groups (all P > 0.05), and there were significant differences in DM duration (P < 0.001) and BMI (P = 0.013). BCVA, CST and MV were significantly different among three groups (all P < 0.001) while IOP was not (P = 0.942).

This study provided new evidence to demonstrate the role of fluid overload in DME. As previously described, the relationship between barrier dysfunction and macular edema, impaired renal function and fluid overload have been investigated, respectively (10, 15). However, the complicated interaction between barrier dysfunction and fluid overload has not been elucidated yet, which make it intricate to explain whether the fluid overload is the mediator between barrier dysfunction and macular edema, or it can lead to DME without the impact from BRB breakdown, as fluid overload also had other influencing factors (27–29). Our study corroborated that fluid overload existed in patients with DME. Furthermore, we were the first to discover that the relationship between fluid retention and DME is independent of renal functions, especially UACR and eGFR. ECW/TBW was proven to be a risk factor for DME, suggesting that fluid overload is a contributing factor to DME apart from BRB disruption.

Tsai et al. (15) had found that the level of fluid overload was correlated with CST in patients with DR, and volume overload was the influence factor of DME. Our study also found a significant difference in fluid status between DME patients and DM patients without DME, either with or without DR. These results suggested that volume expansion might contribute to DME development. Meanwhile, we did not find significant differences in fluid overload status between the DR and DM groups, indicating that fluid overload could be a risk factor for DME but not DR.

A recent study found that fluid overload was correlated with the stages of UACR in T2DM patients with DKD (14). DKD was regarded as a sign of endothelial dysfunction due to oxidative stress and inflammation (30). Also, multiple studies confirmed that UACR levels were significantly associated with DME (6, 7, 31). Hsieh et al. (32) found patients with higher baseline UACR levels to have higher risks for developing DME during the follow-up period than those with lower UACR levels. Nephropathy and retinopathy co-existed in diabetic patients (33). Hemodynamic and structural changes due to inflammatory cytokines and oxidative stress under high blood glucose have been speculated as the cause (31). The increased UACR level suggests albumin leakage in the glomerulus and indicates the vascular barrier breakdown systemically. Even though the BRB was formed predominantly with tight junctions different to peripheral capillaries (34), in our study, the UACR stages were associated with ECW/TBW after adjustment, indicating that UACR is a significant influence factor for fluid overload. Therefore, it was of great importance to consider the role of DKD presented by albuminuria in the process of fluid retention in DME.

As ECW/TBW was reported to be affectable by multiple factors (14, 35, 36), we gradually adjusted for systemic factors, eGFR, and albuminuria, when analyzing fluid status among all participants. DME and ECW/TBW were still positively related after adjusting for confounders, suggesting that in DM patients, DME happens simultaneously with the presence of fluid overload, which was not affected by apparent inner BRB breakdown. Our results confirmed that the two parameters, ECW/TBW and UACR, are independent risk factors for DME. Hence, it could be speculated that the role of systemic fluid overload in DME development may collaborate with other mechanisms than deteriorated BRB function.

Throughout the body, fluid volume, hydrostatic and oncotic pressure, levels of albumin, and capillary permeability are the keys to maintaining the fluid balance between the intracellular and extracellular spaces, which is also true in the retina case. One of the popular theories for DME development was explicated by the starling equation (13). The hydrostatic pressure and oncotic pressure in different compartments govern the movement direction of fluid, which maintain homeostasis under normal circumstances (37). For example, the water transport of RPE cells was proved coupling with lactate, which provide a positive oncotic pressure to drive the water flow (38). It can be speculated that when hydrostatic pressure increases in the capillaries under volume expansion, fluid in the capillaries tends to move to the extravascular space (13). Therefore, high fluid volume can lead to peripheral edema and macular edema in the eyes. Recently, the revised starling equation has risen to make up for the shortcoming of the traditional one (37). The vascular barrier was composed of not only endothelial cells but also a layer which is known as the endothelial surface layer (ESL). ESL is the inner surface of the vascular wall, a semi-permeable layer composed of the glycocalyx and plasma protein such as albumin. Especially, albumin plays a vital role in its filtration function. ESL is usually impaired under systemic inflammation, including DM. The increased production of acute-phase protein suppresses albumin synthesis in the liver due to the limited synthesis capacity. Moreover, increased excretion of albumin from the kidney in the long run in patients with impaired kidney function further decreases the albumin level, leading to the dysfunction of ESL and leakage of fluid and protein. While on the retina, the damaged Müller cells no longer sustain neovascular coupling with DR progression, creating a focal high permeable environment for the albumin (39). Taking together, both fluid overload and impaired blood-retinal barrier can lead to leakage of fluid into the retinal interstitial space, resulting in macular edema.

As retinal thickness decreases with aging (40), we found the model consisting of ECW/TBW, UACR, and age achieved high accuracy in predicting DME risk. This result tally with the hypothesis that fluid overload could be a systemic fluid marker, parallel to the barrier dysfunction represented by albuminuria in the development of DME, in addition to focal VEGF elevation and inflammation.

In clinical practice, anti-VEGF agents are considered the first-line treatment for DME. However, non-responders exits at a considerable percentage (41). Some cases are even refractory to both anti-VEGF treatment and corticosteroid (42), demonstrating the need for additional therapies to treat this disease more comprehensively. Also, it is known that DME does not necessarily fit the regular course of DR progression. It may occur at any stage of DR (43). Therefore, beyond ischemia and inflammation, increased fluid volume may be another critical component in the pathophysiology of DME that has not been given sufficient attention. Several researchers also reported cases in which patients with DME improved, both central retinal thickness reduced and visual acuity improved, after diuretic therapy and restricting salt intake (44, 45), some of whom were even resistant to ophthalmic intervention anti-VEGF and grid laser treatment (46). Considering our findings in this study, we propose that, on the one hand, fluid overload can be one of the mechanisms that cause DME. In this point, treatment to ameliorate extracellular volume expansion such as diuretics and sodium-glucose cotransporter 2 inhibitors may have a role in treating DME patients with high ECW/TBW. On the other hand, in cases of DME who inadequately respond to anti-VEGF therapies and anti-inflammatory drugs, the underlying disease pathology may be mediated by fluid retention. Hence, further study with longitudinal design to observe the treatment response of DME patients with different fluid statuses would be pivotal to confirm the hypothesis.

There were some limitations of the study. The sample size was calculated with preliminary study data. Still, more substantial statistical power might reveal more significant systemic influence factors. Also, the assessment of the risk of DME was to be tested prospectively, and further cohort studies should be conducted to collaborate with our results.

In conclusion, systemic fluid status elevated significantly in patients with DME compared with diabetic patients without DME. Fluid overload and albuminuria were independent risk factors for DME. Fluid overload might have initiated the extravasation in macula edema in addition to barrier dysfunction.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Research Ethics Committee of the Guangdong General Hospital (GDREC2018380H). The patients/participants provided their written informed consent to participate in this study.

JY, QP, and YL contributed to the study conception and design. Material preparation and data collection were performed by AL, JX, XZ, RC, YC, and ZW. Data analysis was conducted and the first draft of the manuscript was written by JY and QP. DC and LZ contributed to the manuscript reviewing and editing and funding acquisition. All authors read and approved the final manuscript.

This work was funded by Guangzhou Municipal Science and Technology Bureau (grant 202102080008) and Bethune-Merck Diabetes Research Fund (G2018030).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.