Lower extremity lymphatic flow is associated with diuretic response in acute heart failure

Abstract

Aims

The dysfunctional lymphatic system appears to play an important role in the development and progression of congestion in heart failure. We hypothesized that in acute heart failure (AHF), diuretic efficacy is associated with peripheral lymph flow.

Methods and results

We prospectively studied AHF patients who received protocolized diuretics followed by assessment of lower limb flow using indocyanine green lymphangiography (a validated method for visualization of lymphatic drainage). The lymph flow was defined as ‘present’ when it reached the ankle and ‘significant’ when it reached >10 cm within 10 min of dye injection, respectively. Based on diuretic efficacy (defined as cumulative diuresis and weight loss), patients were classified as diuretic responders vs. non‐responders. Overall, 65 patients (mean age: 67 ± 15 years, median [Q1–Q3] N‐terminal pro‐B‐type natriuretic peptide: 6901 [4478–12 723] pg/ml) were examined. There were 43 (66%) diuretic responders and 22 (34%) non‐responders who did not differ in baseline clinical/laboratory characteristics. Among the responders, there were more patients with lymph flow either ‘present’ or ‘significant’ (95% vs. 73% and 88% vs. 45% vs non‐responders, respectively, p < 0.01). Responders had significantly longer median distance lymph reached within 10 min (50 [24–75] vs. 10 [3–38] cm; p < 0.0005). There was a significant association between lymph flow and 6‐h diuresis with β (standard error) of 0.45 (0.13) (p < 0.005). In the multivariable analyses, lymph flow distance (odds ratio 1.48, 95% confidence interval 1.08–2.03) independently predicted diuretic efficacy (p < 0.05).

Conclusion

For the first time, faster lower limb lymph flow was linked with better diuretic efficacy in AHF. Studies are needed to determine whether the lymphatic system can become a therapeutic target for decongestion.

Introduction

The lymphatic system plays a crucial role in fluid homeostasis by maintaining the fluid balance between the interstitial and intravascular compartments. ¹ In health, the lymphatic system collects and removes filtered fluid from the interstitial space in the peripheral tissues at a rate that matches fluid accumulation and transports it to the central venous system. Any systemic process leading to lymphatic dysfunction may result in fluid accumulation in the interstitial tissue, which clinically manifests as peripheral congestion.

Congestion is a principal feature characterizing the clinical syndrome of heart failure with complex and multifactorial underlying mechanisms. Only recently has the potential role of the dysfunctional lymphatic system in the development and progression of congestion in HF been recognized. ² , ³ Additionally, it appears that targeting the lymphatic system to improve its function may facilitate interstitial fluid removal and result in more efficient decongestion, which is of particular clinical relevance in the settings of acute heart failure (AHF). ² , ³ On the other hand, as kidneys filter and excrete volume only directly from the plasma, an effective, high‐volume diuresis mandatory in patients with AHF, may require adequate lymphatic drainage to compensate for a rapid plasma volume loss and to prevent intravascular depletion and diuretic resistance. ⁴

Unfortunately, our current understanding of the interplay between congestion/decongestion and lymphatic function in AHF is not only limited but also poorly characterized. ² , ³ In this study, we sought to characterize the lymphatic flow in the lower extremities in patients with AHF and uncover the interaction between the features of the lymph flow and early decongestive potential.

Methods

This is a single‐centre, prospective study conducted at the Institute of Heart Diseases, Wroclaw Medical University, Poland. The study was approved by the local ethics committee and conducted in accordance with the guidelines of the Declaration of Helsinki principles. Informed consent was obtained from all participants.

Study population

The main inclusion criterion was a primary diagnosis of AHF in accordance with the 2021 ESC guidelines, ⁵ with the level of N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) >2000 pg/ml and the presence of peripheral oedema of the lower extremities and clinical indication to receive intravenous loop diuretic. Patients with known lymphatic dysfunction or other known non‐heart failure causes of peripheral oedema were excluded from the study. Patients permanently immobilized, with venous thrombosis, or with lower extremity ulcers were also excluded.

For a more objective assessment of the patient's peripheral oedema, at baseline the circumferences of the lower limbs (at the levels of the midfoot, ankle, and calf) were measured. Each patient's peripheral oedema was also assessed by a standard, more subjective scale, with zero allocated for patients with no oedema and +++ for those with severe peripheral oedema above the knees.

Assessment of the lymphatic flow

All patients who were included into the study underwent an assessment of the lymphatic flow according to the protocol established at our institution and detailed in a previous publication. ⁶ In brief, indocyanine green (ICG) is an amphiphilic tricarbocyanine iodide dye that is widely used as a fluorescent tracer. After being injected subcutaneously, ICG disperses into the interstitial space, from where it is transported by lymphatic vessels to lymph nodes. Exposure to wavelengths between 750 and 800 nm enables high‐contrast imaging up to 2–3 cm deep into tissues. This property allows for the tracing of lymphatic drainage in the lower extremities. The IC‐Flow™ Imaging System (Diagnostic Green, Munich, Germany) was utilized for real‐time detection of ICG fluorescence in lymphatic vessels, facilitating visualization of lymphatic drainage.

ICG lymphangiography protocol

The ICG was injected subcutaneously in three pre‐defined locations on the patient's foot. These locations correspond to the areas from which separate groups of lymphatic vessels of the foot collect lymph ⁷ , ⁸ (Figure  1 ). The ICG was prepared following the manufacturer's guidelines. Three 1‐ml syringes (30‐G needles) containing 0.5 ml of ICG were prepared, one for each injection site. To minimize muscle pump effects on lymphatic flow, patients rested in a supine position for at least an hour before injections. Anesthetic cream (EMLA Cream, 5% lidocaine/prilocaine) was applied to three pre‐defined injection sites on each foot and covered for 45 min. After skin disinfection, 0.5 ml of ICG was injected subcutaneously at each site with minimal delay in between. Lymph flow (propagation of ICG in the lymphatic vessels) was monitored for 10 min.

Figure 1.

Indocyanine green (ICG) lymphography in acute heart failure patients. (Created in BioRender)

The lymphography was performed 3 h after the administration of the diuretic bolus (see below). This timing was chosen to align with the peak urine output at 2–3 h post‐administration of diuretics to optimally capture possibly the highest (or at least present) lymphatic drainage. ⁹ We also assumed that the plasma refilling requirement would be the highest during the peak diuresis to compensate for plasma volume loss.

Lymph flow measures and definitions

Several metrics of lymph flow were assessed during the lymphography.

Definitions of lymph flow:

• Presence of the lymph flow was defined as flow above the ankle within ≤10 min.

• The significant lymph flow was defined as lymph reaching >10 cm within ≤10 min.

The lymph flow was also quantitively assessed as a precise distance (in cm) reached by the lymph within 10 min of the injection point.

Additionally the lymph flow was assessed by measuring whether the most distant lymph reached any of pre‐defined lower extremity level (ankle, mid‐calf, knee, mid‐thigh, groin) flow within a 10‐min period.

Protocol of furosemide administration

To more objectively compare the patients' diuretic response and lymph flow, all patients received a standardized dose of furosemide (calculated as 1 mg/kg body weight). The dose was then administered as a 40 mg bolus, followed by a 2‐h intravenous drip of the remaining calculated dose. This protocol ensured all patients were exposed to the comparable diuretic dose (diuretic challenge). This approach has already been implemented previously. ⁹ , ¹⁰

Diuretic response assessment

The cumulative diuresis was monitored for 6 h following furosemide administration. Patients with a cumulative diuresis ≥900 ml per 6 h were classified as the diuretic responder group. Patients who had diuresis between 600 and 900 ml were classified as diuretic responders if their weight loss was greater than 1 kg within 24 h after the diuretic administration. All remaining patients who did not meet the above pre‐specified criteria were classified as the diuretic non‐responder group. The diuresis cutoff value was derived from the position statement of the Heart Failure Association of the European Society of Cardiology (ESC) and ESC guidelines. ⁵ , ¹¹

Laboratory assessments

Laboratory parameters were assessed using standard methods at a local laboratory. These included measuring plasma NT‐proBNP using an immunoenzymatic technique (Siemens, Marburg, Germany), as well as cancer antigen 125, renin, and aldosterone levels. The serum level of vascular endothelial growth factor C (VEGF‐C) was determined using a Quantikine ELISA immunoassay kit from R&D Systems (Minneapolis, MN, USA), utilizing frozen samples.

Statistical analysis

Continuous variables with a normal distribution were presented as mean ± standard deviation, while variables with a skewed distribution were described using medians with (upper and lower quartiles), and quantitative variables were presented as numbers (percentage). To demonstrate differences between study groups, we employed the Student's t‐test for variables with normal distribution, the Mann–Whitney U test for skewed variables, and the χ ² test for categorical variables. The multiple and logistic regression models were built to assess the associations between diuretic response and lymph flow. The selection of variables for multivariable models was based on prior literature and pathophysiological plausibility.

Power calculation was based on our previously limited observations, as no other clinical data were available. ⁶ The anticipated percentage of patients with detectable lymph flow at rest was estimated to be approximately 75% in patients with good diuretic response. In comparison, the flow at rest in the poor responders was estimated to be approximately 40%. Based on these assumptions, with an alpha of 0.05 and power of 80%, the study population was calculated to be 60 patients.

Results

Baseline characteristics

A total of 65 AHF patients, with 14 (22%) females, were included in the study whose baseline clinical and laboratory characteristics, and biomarkers of congestion and hormonal activation are presented in Table  1 . The mean age of the study population was 69 ± 15 years, the mean left ventricular ejection fraction (LVEF) was 38 ± 16%, the median (interquartile range [IQR]) NT‐proBNP was 6902 (Q1–Q3 4477.7–12 723.0) pg/ml, and the mean serum creatinine was 1.30 ± 0.61 mg/dl.

Table 1.

Baseline characteristics of the study population and comparison of diuretic responder versus non‐responder group

	All patients (n = 65)	Responder group (n = 43, 66%)	Non‐responder group (n = 22, 34%)	p‐value
Age (years)	67 ± 15	68 ± 14	64 ± 19	0.398
Female sex, n (%)	14 (22)	10 (23)	4 (18)	0.634
Height (cm)	172.3 ± 9.4	171.8 ± 9.4	173.3 ± 9.6	0.557
Weight (kg)	88.9 ± 19.9	89.7 ± 20.1	87.5 ± 19.9	0.684
Furosemide dose (intravenous) (mg)	90 [80–100]	90 [80–100]	90 [80–100]	0.776
AHF de novo, n (%)	33 (51)	19 (44)	14 (67)	0.091
Ejection fraction (%)	37.7 ± 15.7	37.62 ± 15.57	37.82 ± 16.40	0.964
TAPSE (cm)	1.63 ± 0.32	1.68 ± 0.34	1.52 ± 0.26	0.088
Haemoglobin (mg/dl)	12.5 ± 2.4	12.45 ± 2.40	12.66 ± 2.37	0.740
Haematocrit (%)	38.14 ± 6.73	38.2 ± 6.3	38.0 ± 7.6	0.929
Albumin (mg/ml)	3.36 ± 0.56	3.33 ± 0.48	3.44 ± 0.69	0.460
Total protein (g/L)	6.11 ± 0.67	6.04 ± 0.70	6.24 ± 0.61	0.305
Na+ (mmol/L)	139.1 ± 4.7	139.7 ± 4.64	137.9 ± 4.70	0.138
K+ (mmol/L)	4.75 ± 4.60	5.06 ± 5.60	4.15 ± 0.63	0.459
Creatinine (mg/dl)	1.30 ± 0.61	1.26 ± 0.58	1.38 ± 0.68	0.471
eGFR (ml/min/1.73 m2)	60.6 ± 30.0	62.4 ± 31.1	57.0 ± 28.1	0.496
Urea (mg/dl)	62.9 ± 34.9	58.13 ± 32.77	72.14 ± 37.72	0.126
CRP (mg/L)	12.3 [5.75–33.27]	12.3 [5.75–21.95]	13.1 [5.56–42.1]	0.501
Aspartate aminotransferase (IU/L)	33.0 [24.0–52.0]	31 [23–48]	30 [22–62]	0.454
Alanine aminotransferase (IU/L)	33.0 [18.0–54.0]	33 [18–48]	35 [15–88]	0.515
Gamma‐glutamyltransferase (IU/L)	115.0 [48.0–147.5]	113 [39–162]	116 [55–141]	0.562
Pharmacological treatment, n (%)
Beta‐blockers	60 (92)	39 (91)	21 (95)	0.496
Aldosterone receptor antagonists	24 (37)	15 (35)	9 (41)	0.634
SGLT2 inhibitors	28 (43)	20 (47)	8 (36)	0.434
ACE inhibitors	45 (69)	31 (72)	14 (64)	0.485
Sacubitril/valsartan	3 (5)	1 (2)	2 (9)	0.219
Medical history, n (%)
History of malignancy	3 (5)	3 (7)	0 (0)	0.110
Lower limb vein surgery or other invasive interventions	1 (2)	0 (0)	1 (5)	0.159
Biomarkers of congestion and hormonal activation
NT‐proBNP (pg/ml)	6902.8 [4477.7–12 723.0]	6391.56 [4044.70–11 169.00]	7836.5 [5229.0–17 326.7]	0.221
Renin (μIU/ml)	22.2 [3.3–154.2]	8.7 [2.3–115.8]	127.1 [22.2–750.0]	0.003
Aldosterone (ng/dl)	8.9 [5.2–19.4]	6.2 [3.5–11.5]	19.6 [8.9–32.6]	<0.001
Cancer antigen 125 (U/ml)	189.7 [91.2–469.1]	193.8 [92.6–469.1]	170.1 [91.2–380.4]	0.930
VEGF‐C (pg/ml)	5403.0 [3938.5–6418.0]	5489.0 [4440.0–6445.5]	5326.7 [3697.0–6324.0]	0.340
Assessment of lower limb oedema
Lower limb ooedema, n (%)				0.588
0	0 (0)	0 (0)	0 (0)
+	23 (35)	16 (37)	7 (32)
++	24 (37)	14 (33)	10 (45)
+++	18 (28)	13 (30)	5 (25)
Circumferences of the lower limbs at the level of:
Midfoot (cm)	27.0 [25.0–28.5]	27.3 [25.8–29.0]	26.5 [24.0–28.0]	0.102
Ankle (cm)	28.5 [26.5–31.0]	28.8 [27.0–31.8]	28.3 [26.0–31.0]	0.360
Calf (cm)	39.0 [36.5–42.0]	39.0 [37.0–41.5]	38.8 [35.0–43.0]	0.730

Values are given as n (%), mean ± standard deviation, or median [interquartile range].

ACE, angiotensin‐converting enzyme; AHF, acute heart failure; CRP, C‐reactive protein; eGFR, estimated glomerular filtration rate; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; SGLT2, sodium–glucose cotransporter 2; TAPSE, tricuspid annular plane systolic excursion; VEGF‐C, vascular endothelial growth factor‐C.

Comparison of baseline characteristics between the diuretic responder versus non‐responder group

There were 43 (66%) patients in the diuretic responder group and 22 (34%) in the non‐responder group. Patients from both groups did not differ across baseline clinical and laboratory characteristics (Table  1 ), including mean LVEF (38 ± 16% vs. 38 ± 16%), median [IQR] NT‐proBNP (6392 [4044.7–11 169] vs. 7837 [5229.0–17 326.7] pg/ml), percentage of patients with AHF de novo (44% vs. 67%), median [IQR] total dose of furosemide they received (90 [80–100] vs. 90 [80–100] mg) (responders vs. non‐responders, respectively, all p > 0.05), and baseline pharmacological heart failure treatment. However, the responder group had significantly lower serum renin (median [IQR]: 8.7 [2.3–115.8] vs. 127.1 [22.2–750.0] μIU/ml, p = 0.003) and aldosterone levels (median [IQR]: 6.2 [3.5–11.5] vs. 19.6 [8.9–32.6] ng/dl, p < 0.001), than non‐responders.

Assessment of the lower limb lymph flow in the entire group

Clinical assessment revealed at least modest lower limb oedema in all 65 patients (Table  1 ). In the entire group, the lymph flow was present (defined as lymph flow beyond the ankle within ≤10 min) in 57 (88%) patients, while the significant lymph flow was recorded in 48 (74%) patients. The median [IQR] distance lymph reached at 10 min was 40 [10–55] cm. In 36 (55%) patients, the lymph reached knee level, while at the groin level, it was present in 22 (34%). The median [IQR] time lymph needed to reach the ankle was 120 [60–300] s; more details of the lymph flow are presented in Table  2 .

Table 2.

The lymph flow assessment at rest in acute heart failure

	All patients (n = 65)	Responders (n = 43, 66%)	Non‐responders (n = 22, 34%)	p‐value*
Lymph flow at rest, n (%) a	57 (88)	41 (95)	16 (73)	0.0086
Significant lymph flow at rest, n (%) b	48 (74)	38 (88)	10 (45)	0.0002
The median distance lymph reach at 10 min (cm)	40 [10–55]	50 [24–75]	10 [3–38]	0.00005
Number (%) of patients with lymph flow reaching:
Ankle	57 (88)	41 (95)	16 (73)	0.0086
Mid‐calf	45 (69)	35 (81)	10 (46)	0.003
Knee	36 (55)	30 (70)	6 (27)	0.001
Mid‐thigh	22 (34)	19 (44)	3 (14)	0.014
Groin	22 (34)	19 (44)	3 (14)	0.014

^a Defined as lymph flow above the ankle within 10 min or less.

^b Defined as lymph reaching >10 cm within 10 min or less.

^* P‐value for comparison between responders versus non‐responders.

Association between lymph flow and diuresis

There was a significant association between lymph flow and 6‐h diuresis (both analysed as continuous variables) in the study population. The linear regression model revealed a coefficient β (standard error [SE]) of 0.45 (0.13) (p < 0.005). Furthermore, the association remained significant even after adjusting for baseline weight (β [SE] = 0.45 [0.14], p < 0.005) and baseline estimated glomerular filtration rate (β [SE] = 0.49 [0.15], p < 0.005).

Predictors of the lower limb lymph flow

In the multivariable model, incorporating clinical, laboratory variables and biomarkers of congestion and hormonal activation (Table  3 ), only serum aldosterone levels (β [SE] = −0.30 [0.14], p < 0.05) and tricuspid annular plane systolic excursion (TAPSE) (β [SE] = 0.28 [0.14], p < 0.05) were independently associated with the lymph flow distance.

Table 3.

Predictors of the lower limb lymph flow distance within 10 min (multivariable regression)

Variable	Model's R = 0.53, p < 0.05
	β (SE)	p‐value
TAPSE (mm)	0.28 (0.14)	0.047
CRP (mg/g)	0.04 (0.13)	0.79
eGFR (ml/min/1.73 m2)	−0.03 (0.14)	0.84
Renin (μIU/ml)	−0.14 (0.13)	0.30
Aldosterone (ng/dl)	−0.30 (0.14)	0.04
NT‐proBNP (log) (pg/ml)	0.06 (0.14)	0.70

CRP, C‐reactive protein; eGFR, estimated glomerular filtration rate; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; SE, standard error; TAPSE, tricuspid annular plane systolic excursion.

Comparison of the lower limb lymph flow in the diuretic responder versus non‐responder group

In the diuretic responder group, lymph flow was present in 41 (95%) patients versus 16 (73%) patients in the non‐responder group (p = 0.009) (Figure  2A ). The significant lymph flow was more frequent in the responder versus non‐responder group (38 [88%] vs. 10 [45%], p < 0.001) (Figure  2B ). The distance the lymph reached at 10 min was significantly longer in the diuretic responder versus the non‐responder group (median [IQR]: 50 [24–75] vs. 10 [3–38] cm, p < 0.001) (Figure  3 ). The percentage of patients in whom the lymph reached each pre‐defined lower limb level was significantly higher in the diuretic responder group (all p < 0.05) (Table  2 , Figure  4 ).

Figure 2.

Comparison of the presence of lymph flow between the responder and non‐responder groups. (A) The lymph flow. (B) The significant lymph flow.

Figure 3.

Comparison of the median distance lymph reach at 10 min in both study groups.

Figure 4.

Comparison between the responder versus non‐responder groups in the percentage of patients with lymph flow reaching the pre‐defined points.

Determinants of diuretic response

In the multivariable model, incorporating baseline clinical and laboratory variables, biomarkers of congestion, and hormonal activation, it appeared that only significant lymph flow and serum aldosterone levels were independently associated with the allocation to the diuretic responder group. In a model in which the lymph flow was imputed as a continuous variable (per 10 cm change), the distance lymph reached at 10 min was an independent predictor of the allocation to the responder group (odds ratio [95% confidence interval]: 1.48 [1.08–2.03], p = 0.012) (Table  4 ).

Table 4.

Clinical and laboratory predictors of diuretic response (multivariable regression)

Variable	Model's χ 2 = 27.95, p < 0.0001
	Odds ratio (95% CI)	p‐value
Distance lymph reached at 10 min (per each 10 cm)	1.48 (1.08–2.03)	0.012
Aldosterone (ng/dl)	0.93 (0.88–0.99)	0.032
NT‐proBNP (log) (pg/ml)	0.53 (0.19–1.44)	0.200
eGFR (ml/min/1.73 m2)	0.99 (0.96–1.02)	0.667
Serum Na+ (mmol/L)	1.10 (0.92–1.30)	0.256

CI, confidence interval; eGFR, estimated glomerular filtration rate; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.

Discussion

To the best of our knowledge, for the first time, we have been able to characterize the lymph flow in the lower extremities in patients with AHF and peripheral oedema receiving intravenous loop diuretics. In these patients, faster and more distant lymph flow was linked with better diuretic efficacy and lower serum aldosterone levels (Graphical Abstract).

The lymphatic system has been a highly intriguing yet largely inaccessible aspect of AHF pathophysiology for decades. ² , ³ , ¹² This stems from its plausible role in the development and progression of peripheral oedema and consequently, decongestion due to its function in regulating the fluid balance of the interstitial space. On the other hand, the challenges in assessing the system's performance with a lack of simple diagnostic tools and the inherent complexity of visualization and access to the system itself have hindered progress in understanding its role. Yet, the early attempts to facilitate decongestion by direct cannulation and drainage of the thoracic duct have revealed that decompression of the thoracic duct in decompensated patients was associated with prompt and significant improvement in peripheral oedema. ¹³ , ¹⁴

In this study, for the first time, we have described that the lymph flow in the lower extremities was present in most patients with AHF and additionally was considered as significant in three‐quarters of them (defined as lymph reaching >10 cm within ≤10 min). This indicates that it appears to be a dynamic and active process in patients with AHF and peripheral oedema most likely associated/initiated with decongestion (as we assessed the flow after intravenous furosemide administration). It sharply contrasts with our previous findings in compensated heart failure patients studied in ambulatory settings, in whom lymph flow at rest was only minimal. ⁶

This study reports a clinically important (and new) association between more effective decongestion (as indicated by a more efficient diuretic response in an early phase of hospitalization) and improved lymphatic drainage. Although we believe this observation has a significant pathophysiological meaning, we cannot conclude any causal relationships or indicate any responsible mechanisms.

There are several putative causes underlying lymphatic dysfunction (with pathological accumulation of interstitial fluid) in AHF: increased capillary hydrostatic pressure, elevated central venous pressure leading to impaired drainage of the lymph into central venous circulation due to increased permeability of the lymphatic vasculature, and dysfunction of the renal lymphatic system. ² , ³ More effective decongestion seems to correct elevated central venous pressure and capillary hydrostatic pressure, which may facilitate lymphatic drainage. On the other hand, insufficient lymphatic drainage and, consequently, inadequate refilling of the intravascular space during diuresis may result in intravascular volume depletion that limits further diuresis to prevent the system collapse. Conversely, a poor diuretic response and the lack of sufficient volume excretion may lead to insufficient loss of intravascular volume, which would otherwise be effectively replenished by lymphatic drainage. Answering this fundamental question regarding which process influences the other will be very challenging because numerous concurrent adaptive mechanisms/factors play crucial roles in maintaining intravascular integrity, like baseline plasma volume, vascular tone (stressed blood volume), thirst, and neurohormonal control. ¹⁵ , ¹⁶ , ¹⁷

The lymphatic system plays an essential role in the renal interstitium, particularly within the periarterial region, which can also lead to perfusion abnormalities and further consequences. ¹² , ¹⁸ The experimental model demonstrated that furosemide may paradoxically induce vasodilatation and decrease lymphatic contractility in renal collecting vessels, leading to impaired drainage of the renal interstitium. ¹⁹ However, in our cohort, we did not observe this phenomenon on peripheral lymphatic vessels when clinically visible lymph flow was assessed.

The elevated aldosterone levels were independently associated with lower lymphatic drainage. This also needs to be interpreted with caution as the observation may result from the known impact of the hormone on the diuretic response, while we did not examine the direct impact of aldosterone on lymph flow. Remarkably, the lymphatic marker VEGF‐C was not associated with lymph drainage in our population. This is surprising as previous data have suggested the association between the marker and peripheral congestion. ²⁰ Moreover, lymphatic drainage was not higher in de novo AHF patients in our cohort, nor was it associated with the diuretic dose.

For this study, we divided the population into two groups based on diuretic response to compare lymph flow between them. However, there is no universal definition of a good diuretic response. Therefore, we followed the recent ESC guidelines to define both groups based on 6‐h diuresis. Additionally, we incorporated weight loss as an enrichment factor only in patients with borderline diuresis despite the known weak correlation between diuretic response and short‐term diuresis. ²¹ However, we assumed that those who lost weight must have experienced effective diuresis, mobilization of the fluid from the interstitial space, and, potentially, good lymph flow. Furthermore, in this first study, some arbitrary decisions were necessary, as when designing the study, we were uncertain whether this phenotype would be associated with lymph flow or if any phenotype would be associated with it at all. Therefore, we opted for a broader definition of a good diuretic response. Moreover, the definition of the non‐responder group is based on the patient's exposure to a standardized dose of diuretics, which is scientifically desirable, but it leads to limitations. As some patients may have been underdosed, and if they had been exposed to higher doses of diuretics, they might no longer meet the criteria for the non‐responder group.

The lymphatic system seems to play an important role during decongestion as kidneys filter and excrete volume from the plasma. Effective diuretic‐induced diuresis may require sufficient lymphatic drainage to offset plasma volume loss and sustain intravascular volume stability. If the role of lymphatics in managing congestion/decongestion in AHF is confirmed, it may become the therapeutic target for the therapy. The lymphatic system‐targeted interventions may potentially benefit most patients with residual congestion and those with refractory peripheral congestion. ²² , ²³ It might also be a supportive device to prevent intravascular volume depletion during effective decongestion. ⁴ However, this needs further development and a better understanding of lymphatic pathophysiology.

Interestingly, lymphatic drainage was also observed in the non‐responder group; however, the flow in this group was significantly less effective, as indicated by a slower flow and a shorter distance reached after 10 min. Based on this observation, we can hypothesize that sufficient lymph flow, rather than any lymph flow, is needed to meet the intravascular demand during decongestion for it to be effective. Thus, techniques for quantitative assessment of the flow are needed to enhance our understanding of the phenomenon. Importantly, the differences in lymphatic drainage between both groups did not result from differing degrees of peripheral oedema, as this remained consistent. The same degree of peripheral oedema was confirmed based on clinical scale and objective measurements (circumferences).

Limitations

As data on the lymph flow in AHF were either scarce or unavailable, in this study, we have adopted some arbitrary definitions (e.g. any lymph flow at rest or significant lymph flow) and assumptions, such as the hypothesis that lymph flow should be present/assessed 3 h after diuretic administration which corresponds to the peak diuretic response. While this assumption is biologically plausible, no scientific data currently exist to support or refute it. The ICG lymphography also has inherent limitations, the tissue penetration depth of 2–3 cm may limit visualization in patients with more subcutaneous tissue and obese in whom the oedema may have multifactorial aetiology. ²⁴ , ²⁵ However, the sensitivity analysis did not reveal the correlation between the patient's weight and lymph flow, and the profile of peripheral oedema was identical for both study groups. The major limitation of ICG lymphography is that the ICG remains in the lymphatic vessels after application, making subsequent examinations and applications pointless for an extended period. This prevents the possibility of performing serial assessments, which would otherwise provide highly informative insights. Although the patient population is based on power calculation, the study group is relatively small, which results in several general limitations and makes it underpowered to detect differences in some comparisons for which the study was not designed. Unfortunately, we cannot provide the natriuresis data, which might offer intriguing insight into the overall concept.

Conclusions

For the first time, we demonstrated that lymph flow in the lower extremities was present in the majority of patients with AHF 3 h after the diuretic administration. Moreover, faster and more distant lower limb lymph flow (more effective lymph drainage) was linked with better diuretic response. The lymph flow was inversely correlated with serum aldosterone levels and positively with TAPSE. Further studies are needed to determine if the lymphatic system can become a therapeutic target for better decongestion in AHF. ²⁶

Funding

This research was financed by the Polish‐German Science Prize ‘Nicolaus Copernicus’ (Copernicus Prize) COP/01/2018 awarded by the Deutsche Forschungsgemeinschaft (DFG) and the Foundation for Polish Science (FNP) for outstanding scientific achievements resulting from the German‐Polish cooperation. This research was also financed by a grant from Wroclaw Medical University IDUB.A46B.24.002. The presented research results were funded by the Development Strategy of the Wroclaw Medical University entitled ‘UMW in the Light of Scientific Excellence 2024–2026’.

Conflict of interest: J.B. has received honoraria from Bayer, Boehringer Ingelheim, AstraZeneca, Alleviant Medical, Reprieve Cardiovascular, and WhiteSwell. All other authors have nothing to disclose.