ABSTRACT
Background
Hemoglobin affinity for oxygen is modulated by ambient oxygen tension, acid/base status, 2,3 diphosphoglycerate (2,3DPG) concentrations and other factors, facilitating tissue oxygenation under changing conditions. 2,3DPG is a key regulator of oxygen affinity within red blood cells and its levels are affected by blood phosphate. P50, the partial pressure of oxygen at which 50% of its hemoglobin binding sites are occupied, is a marker of oxygen delivery to tissues. We measured P50 during hemodialysis and explored its relationship with mineral metabolites and left ventricular strain as a marker of cardiac function.
Methods
Venous blood gas and other laboratory parameters were measured in 20 prevalent patients pre- and post-hemodialysis. To avoid arterio-venous mixing, we selected patients dialyzing through tunneled dialysis catheters. Associations of P50 with demographics, laboratory parameters and echocardiographic measurements were examined using linear regression models.
Results
P50 levels decreased from 27.1 ± 0.9 mmHg to 26.2 ± 0.7 mmHg during hemodialysis (P < .001). Among 18 predictors evaluated, older age, and greater reductions in phosphate during hemodialysis were the strongest predictors of P50 changes in multivariate models. There was acute worsening in left ventricular global longitudinal strain (LVGLS) during hemodialysis (reduction of 1.4 ± 3.9%; P = .03). Greater reductions in P50 during hemodialysis and older age were significantly associated with greater reductions in LVGLS.
Conclusions
Hemodialysis consistently reduces P50. The magnitude of P50 change was strongly associated with concurrent phosphate changes. P50 reductions correlated with acute lowering of LVGLS. These observations illuminate a potential cause of systemic tissue hypoxia and potential cardiac dysfunction during hemodialysis.
Keywords: cardiac function, hemodialysis, oxygen delivery, P50, phosphate dynamics
Graphical Abstract
Graphical Abstract.
KEY LEARNING POINTS.
What was known:
Cardiovascular disease is prevalent in end-stage kidney disease (ESKD) patients, with traditional risk factors insufficient to explain the increased incidence.
Chronic hyperphosphatemia is linked to vascular calcification and cardiovascular disease risk, but acute phosphate changes during hemodialysis are less understood.
The effects of hemodialysis on P50, a marker of oxygen delivery, are unknown.
This study adds:
Hemodialysis consistently reduces P50 suggesting reduced oxygen delivery to tissues.
The magnitude of P50 change is strongly associated with concurrent phosphate changes.
P50 reductions during hemodialysis correlated with acute worsening of left ventricle global longitudinal strain.
Potential impact:
Understanding mechanisms of P50 alterations in hemodialysis patients may inform strategies to mitigate tissue hypoxia and cardiac dysfunction during dialysis.
Tailored phosphate management may improve cardiac outcomes in ESKD patients.
INTRODUCTION
Cardiovascular disease (CVD) remains the leading cause of death among patients with end-stage kidney disease (ESKD) [1]. The increased incidence of CVD in ESKD is not explained by traditional risk factors such as hypertension and diabetes. Unique non-traditional factors such as repetitive hemodynamic instability during hemodialysis sessions and abnormal mineral metabolism may confer additional risks.
Serum phosphate concentrations are often deranged in ESKD [2], and hyperphosphatemia is nearly ubiquitous in ESKD patients. The adverse vascular consequences of chronic hyperphosphatemia, including vascular calcification and CVD risk, are well appreciated [3], and efforts from pharmaceutical interventions to federal regulations have focused heavily on lowering phosphate concentrations in ESKD patients. These concentrations are typically measured at the start of the hemodialysis procedure. The consequences of acute phosphate reductions during the hemodialysis procedure have received much less investigation or attention. It is known that phosphate concentrations are typically lowered to a half to one-third of their pre-dialysis levels during the hemodialysis procedure, and then quickly rebound. The magnitude of intradialytic reductions in phosphate are larger in patients with higher pre-dialysis phosphate concentrations. Recent studies have begun to shed light on the complex dynamics of phosphate regulation during hemodialysis, suggesting that the muscle may serve as an endogenous source of phosphate during the procedure, and may account for the rapid rebound in serum phosphate concentrations after the procedure is completed [4–6]. The rapid and marked reductions of phosphate that occur early in the hemodialysis procedure may trigger important responses in the cellular machinery, including those affecting oxygen delivery.
Erythrocyte hemoglobin affinity for oxygen binding is modulated by ambient oxygen tension, acid/base status, intra-cellular 2,3 diphosphoglycerate (2,3DPG) concentrations and other factors. We previously demonstrated that 2,3DPG concentrations are lowered by systemic phosphate removal in patients receiving continuous dialysis therapy in the intensive care unit setting [4]. P50 is a marker of oxygen availability in tissues- a P50 reduction indicates increased hemoglobin:oxygen (Hgb:O2) affinity and decreased oxygen release, thereby marking risk for worsening systemic tissue hypoxia. To our knowledge, the effect of hemodialysis on P50 has not previously been evaluated.
The heart has one of the highest oxygen demands in the body and may be especially susceptible to the effects of impaired oxygen delivery. Studies have shown that the hemodialysis procedure can induce transient cardiac regional wall motion abnormalities in some patients [7]. Left ventricular (LV) global longitudinal strain (GLS) from 2-D speckle tracking echocardiography (STE) is a sensitive and reproducible marker of LV systolic function compared with conventional LV ejection fraction (LVEF) by echocardiography [8], and may therefore provide an enhanced ability to evaluate LV function.
The goals of the current study were several fold. We aimed to determine the intradialytic changes in oxygen availability measured by P50 during the hemodialysis procedure. Second, we aimed to identify variables that associated with the magnitude of P50 changes during hemodialysis. Finally, we aimed to determine whether any P50 changes observed during hemodialysis are associated with cardiac function as measured by LVGLS in adults with kidney failure receiving hemodialysis.
MATERIALS AND METHODS
Study design and setting
This study was approved by the Greater Los Angeles Veterans Affairs Institutional Review Board (approval number 1615978), and all participants provided written informed consent. Using a cross-sectional study design in 20 prevalent hemodialysis patients we analyzed the association between P50 and laboratory studies before and at end of the hemodialysis session. In a subset of patients, we obtained echocardiographic parameters both before (pre) and at the midpoint of the dialysis sessions. Echocardiography was not done after the hemodialysis procedure due to feasibility constraints and the need to free the dialysis chair for other patients requiring hemodialysis treatments. We selected patients who were dialyzed through tunneled catheters to avoid arterio-venous mixing of sampled blood, which can occur when sampling from a arteriovenous fistula or graft. Inclusion criteria included adult patients with ESKD receiving maintenance dialysis thrice weekly for a minimum of 6 months. Exclusion criteria were: (i) sickle cell anemia (influences P50) and unexplained pre–chronic kidney disease anemias, (ii) inadequate echo windows, (iii) patients with LVEF <30% at baseline or baseline cardiac regional wall akinesis (which introduce bias in echocardiographic findings), (iv) chronic obstructive pulmonary disease patients dependent on home oxygen, or (v) patients with hemoglobin concentrations <8 g/dL.
Clinical and laboratory evaluation
Demographic parameters (age, sex, self-reported race, body weight) and clinical variables including primary cause of kidney disease and history of diabetes were collected by history and chart review. Hemodialysis parameters such as vintage, current prescription and urea reduction ratio were recorded.
Laboratory covariates
A complete blood count including hemoglobin was measured in whole-blood samples collected in EDTA tubes, using the Sysmex 9000 automated hematology analyzer. Serum phosphate, other electrolytes, brain natriuretic peptide (BNP) and high sensitivity troponin were measured at three time points during hemodialysis: pre-dialysis, the mid-point of the dialysis procedure and immediately post-dialysis. All post-hemodialysis blood samples were drawn within 5 min of treatment completion using the standardized slow-flow method to ensure consistency across patients. We collected venous blood in ice-cooled heparinized tubes for blood gas measurement using a Rapid Point 405 analyzer with an integral co-oximeter (Siemens Diagnostics). This instrument calculates P50 using measured oxygen saturation and partial pressure of oxygen.
Echocardiographic assessment
All patients were imaged pre-dialysis and mid-dialysis using a commercially available echocardiographic system (Philips Medical System, Andover, MA, USA) with the following probes 3V2c, 3V2c-S or S5-1. Mean frame rate for the study participants was 56 ± 4.8 Hz. Standard 2D echocardiographic views included apical long axis images. Electrocardiogram was simultaneously recorded during image acquisition.
In a subset of patients (N = 13), 2D speckle-tracking echocardiography (2D-STE) was conducted using vendor-independent software (2D Cardiac Performance Analysis, Ultrasound Workspace). End-diastolic and end-systolic volumes were used to calculate EF from the apical 4-, 3- and 2-chamber views. The endocardial borders were traced in the end-systolic frame of the 2D images from the three apical views and LVGLS strain curves were generated. Reported LVGLS reflects the average peak systolic longitudinal strain of the three apical views.
Statistical analysis
For each participant, we calculated the difference in P50 from beginning to end of hemodialysis. We analyzed continuous variables graphically and summarized baseline characteristics. Data are tabulated as mean ± standard deviation (SD) or median and interquartile range (IQR). Comparisons were made using analysis of variance (ANOVA) or ANOVA on ranks, as appropriate. Categorical variables were examined by frequency distribution, recorded as proportions and compared using the chi-square test. We examined correlations of variables with the change in P50 during hemodialysis. Evaluated variables included pre-dialysis serum phosphate, reduction in serum phosphate, pre-dialysis serum potassium, reduction of serum potassium, ionized calcium, BNP, lactate and troponin using Pearson's correlation coefficients. We first conducted univariate linear regression to evaluate factors associated with the reduction of P50 during hemodialysis. Our predictors of interest included hemodialysis parameters such as vintage, ultrafiltration, urea reduction ratio, demographic factors and laboratory parameters such as pre-dialysis phosphate, reduction of phosphate, pH and ionized calcium during hemodialysis. The factors that were significantly associated with reduction in P50 from the univariate analysis were then added to the multivariate analysis. Additionally, we used linear mixed models with random intercepts for patient ID to account for intra-subject correlations. We used multivariate linear regression to evaluate factors associated with change in LVEF and LVGLS. Our covariates of interest included reduction in P50, age, race and sex. We performed sensitivity analyses by further adjusting for potential confounders individually, including hemoglobin, pCO₂, systolic blood pressure, ultrafiltration and potassium.
Statistical analyses were performed with STATA Version 17 (StataCorp LLC, College Station, TX, USA). Two-sided P-values <.05 were considered statistically significant.
RESULTS
Demographic and clinical characteristics of patients treated by hemodialysis along with their hemodialysis parameters are detailed in Table 1. The mean (±SD) age of the 20 hemodialysis patients included in this study was 72 (±13) years, and two (10%) were female, reflecting that most patients were recruited from a Veterans Affairs center. Twelve patients reported African American descent (60%) and 13 patients reported history of diabetes (65%). All patients had oxygen saturation of 92% or greater as measured by finger pulse oximetry and none of the patients received supplemental oxygen. Median hemodialysis vintage was 2.5 (IQR 1.0, 3.7) years, and the mean urea reduction ratio was 76% (±5%). Mean values of pre-dialysis hemoglobin, phosphate, potassium and magnesium were 8.5 (±1.4) g/dL, 4.5 (± 1.2) mg/dL, 4.5 (±0.5) mmol/L and 2.0 (±0.3) mEq/L, respectively. A summary of pre- and post-hemodialysis laboratory and clinical parameters is provided in Supplementary data, Table S1. Finally, mean heart rate was 73 (±10) bpm and mean pre-dialysis LVEF was 41.7 (±10.7) %.
Table 1:
Characteristics of study participants (N = 20).
| Age (years), mean ± SD | 72 ± 13a |
| Male, n (%) | 18 (90) |
| Race, n (%) | |
| African American | 12 (60) |
| Hispanic | 1 (5) |
| Caucasian | 7 (35) |
| Diabetes, n (%) | 13 (65) |
| Dialysis vintage (years), mean ± SD | 2.5 (1.0, 3.7)a |
| Pre-dialysis SBP (mmHg), mean ± SD | 141.8 ± 23.6a |
| Pre-dialysis DBP (mmHg), mean ± SD | 81.3 ± 14.4a |
| Dialysis urea reduction ratio | 76% ± 0.05a |
| Dialysis ultrafiltration (mL) | 1900 (1250, 2750) |
| Dialysis bicarbonate (mEq/L) | 33.5 (32, 35) |
| Dialysis calcium (g/dL) | 2.5 ± 0.1a |
| Pre-dialysis hemoglobin (g/dL) | 8.5 ± 1.4a |
| Post-dialysis hemoglobin (g/dL) | 9.2 ± 1.5a |
| Pre-dialysis ferritin (ng/mL) | 740.7 (343.2, 1276.3) |
| Pre-dialysis transferrin saturation (%) | 31.3 (22.8, 40.8) |
| Pre-dialysis phosphate (mg/dL) | 4.5 ± 1.2a |
| Pre-dialysis potassium (mmol/L) | 4.5 ± 0.5a |
| Pre-dialysis magnesium (mEq/L) | 2.0 ± 0.3a |
| Pre-dialysis lactate (mmol/L) | 0.8 ± 0.5a |
| Pre-dialysis BNP (pg/mL) | 160.5 (80.5, 607.5) |
| Pre-dialysis Hs-Troponin (pg/mL) | 20.5 (10.5, 41.5) |
| Pre-dialysis LVEF (%) | 41.7 ± 10.7a |
| Pre-dialysis LVGLS (%) | –9.8 ± 0.8a |
| Pre-dialysis P50 (mmHg) | 26.7 (26.4, 27.5) |
Data are represented by mean ± SD or median (IQR).
DBP, diastolic blood pressure; Hs, high sensitivity; SBP, systolic blood pressure.
We found that P50 levels decreased during hemodialysis from 27.1 ± 0.9 mmHg to 26.2 ± 0.7 mmHg (P < .001, Fig. 1).
Figure 1:
Acute changes in P50 during the hemodialysis procedure among individual ESKD patients receiving hemodialysis. HD, hemodialysis. P50 is the partial pressure of oxygen where 50% of the oxygen-binding sites on hemoglobin are occupied.
Correlations of changes in mineral metabolites and cardiac biomarkers during hemodialysis
The magnitude of reduction in serum phosphate during dialysis was strongly correlated with the change in P50 (r = 0.49, P = 0.025). Higher pre-dialysis phosphate concentration was also strongly associated with the observed change in P50 (r = 0.56, P = 0.008). However, the pre-dialysis phosphate concentration and the change in serum phosphate across the dialysis session were highly correlated with one another (r = 0.98, Table 2), so these two correlates of change in P50 may reflect a similar biology.
Table 2:
Correlation matrix of phosphate changes with mineral metabolism and cardiac biomarkers in ESKD.
| P50 delta | Phos delta | Phos pre-HD | Potassium delta | Ionized Ca delta | BNP delta | Lactate delta | Troponin delta | |
|---|---|---|---|---|---|---|---|---|
| P50 delta | 1.00 | 0.49* | 0.56* | 0.38 | –0.12 | –0.24 | 0.31 | –0.22 |
| Phos delta | 1.00 | 0.98* | 0.46* | 0.08 | 0.03 | 0.15 | 0.26 | |
| Phos pre-HD | 1.00 | 0.41 | 0.11 | –0.05 | 0.11 | 0.33 | ||
| Potassium delta | 1.00 | 0.12 | 0.21 | 0.19 | 0.10 | |||
| Ionized Ca delta | 1.00 | –0.22 | 0.24 | –0.00 | ||||
| BNP delta | 1.00 | 0.06 | –0.06 | |||||
| Lactate delta | 1.00 | –0.08 | ||||||
| Troponin delta | 1.00 |
*Correlation is significant at the 0.05 level (2-tailed) and **at the 0.01 level (2-tailed).
Ca, Calcium; Delta, change during hemodialysis; Phos, phosphate; Pre-HD, prior to hemodialysis.
Associations of mineral metabolite changes with P50 changes during hemodialysis
In univariate models, among 18 variables analyzed, those associated with greater reductions in P50 during the hemodialysis procedure were greater intra-dialytic serum phosphate reduction, higher pre-dialysis serum magnesium and age of the participant (Table 3). Figure 2 shows the scatter plot and regression line depicting the association between serum phosphate and P50 changes among individual patients.
Table 3:
Univariate and multivariate predictors of changes in P50 during hemodialysis (N = 20).
| Predictor variables | Univariate | Multivariate |
|---|---|---|
| per 1 SD increase | β (95% CI) | β (95% CI) |
| Age, per year | –0.03 (–0.05, –0.01) a | –0.03 (–0.05, –0.01) a |
| Male sex | –0.24 (–1.44, 0.96) | |
| Caucasian race (reference = AA) | 0.13 (–0.24, 0.49) | |
| Diabetes mellitus (reference = no DM) | –0.26 (–0.97, 0.45) | |
| HD vintage, per year | –0.04 (–0.18, 0.09) | |
| HD time, per h | 0.17 (–0.68, 1.03) | |
| Urea reduction ratio, % | –3.45 (–10.23, 3.34) | |
| Ultrafiltration, mL | 0.01 (–0.01, 0.01) | |
| Hemoglobin change during HD (per g/dL higher) | –0.02 (–0.24, 0.21) | |
| Pre-dialysis magnesium (per mEq/L higher) | –1.11 (–2.21, 0.01)b | –0.83 (–1.82, 0.17) |
| Serum phosphate during HD (per mg/dL higher) | 0.26 (0.06, 0.45) a | 0.29 (0.21, 0.39) a |
| Serum potassium during HD (per mmol/L higher) | 0.31 (–0.02, 0.63) | |
| BNP during HD (per pg/L reduction) | –0.01 (–0.00, 0.01) | |
| Troponin change during HD (per pg/mL reduction) | 0.01 (–0.01, 0.01) | |
| Ionized calcium change during HD (per mmol/L reduction) | –0.16 (–0.59, 0.26) | |
| pH during HD (per pH unit reduction) | –1.87 (–5.34, 1.59) | –0.31 (–2.95, 2.33) |
| pCO2 during HD (per mmHg higher) | 0.01 (–0.01, 0.02) | |
| Dialysate Calcium (per g/dL higher) | 0.82 (–2.31, 3.94) |
Indicates P < .05.
Indicates P = .05.
Multivariate model includes serum phosphate change during hemodialysis, pre-dialysis magnesium, pH change during hemodialysis, age.
Bold values indicate statistically significant results.
AA, African American race; DM, diabetes mellitus; HD, hemodialysis.
Figure 2:
Association between changes in serum phosphate and P50 among individual ESKD patients receiving hemodialysis.
We evaluated multivariate models to identify factors associated with changes in P50 independently of the others. In the multivariate models, older age, and greater reductions in phosphate (Table 3) remained associated with greater P50 reduction during hemodialysis, but the P-value associated with the relationship between higher pre-dialysis serum magnesium and P50 reduction was rendered no longer statistically significant. Sensitivity analyses adjusting for potential confounders individually (hemoglobin, pCO₂, SBP) did not meaningfully change the association between phosphate and P50 (Supplementary data, Table S2)
Associations of P50 changes with cardiac function during hemodialysis
There was a reduction in LVEF from pre-dialysis to mid-dialysis (mean reduction 4.17 ± 9.8%; P = .07). In the univariate model, greater reduction in P50 during dialysis was not associated with change in LVEF [β = –1.25; 95% confidence interval (CI) for slope –9.81, 7.31; Table 4]. This relationship was unaltered after adjusting for age, gender and race (β = –1.97; 95% CI for slope –11.36, 7.42).
Table 4:
Association of changes in P50 with LVGLS (N = 13) and LVEF (N = 20).
| Variable | Univariate | Multivariatea | ||
|---|---|---|---|---|
| Outcome | Beta (95% CI) | P-value | Beta (95% CI) | P-value |
| Change in LVGLSb | ||||
| Change in P50 | 3.15 (–1.17, 7.48) | 0.13 | 4.35 (0.35, 8.36) | 0.03 |
| Age, per 1 year increase | 1.27 (–0.79, 3.33) | 0.21 | 2.13 (0.23, 4.03) | 0.03 |
| Male sex (reference = female) | 1.41 (–7.89, 10.72) | 0.74 | 3.86 (–3.91, 11.63) | 0.28 |
| Race (reference = Caucasian) | –0.71 (–3.29, 1.88) | 0.56 | –1.22 (–3.45, 0.99) | 0.24 |
| Change in LVEFb | ||||
| Change in P50 | –1.25 (–9.81, 7.31) | 0.76 | –1.97 (–11.36, 7.42) | 0.66 |
| Age, per 1 year increase | 1.36 (–3.48, 6.21) | 0.56 | 1.52 (–3.69, 6.74) | 0.54 |
| Male sex (reference = female) | –7.24 (–22.74, 8.25) | 0.34 | –7.21 (–23.88, 9.44) | 0.37 |
| Race (reference = Caucasian) | 2.01 (–2.96, 6.97) | 0.41 | 2.26 (–3.21, 7.73) | 0.39 |
Multivariate is adjusted for age, gender, race.
Change in LVGLS and LVEF is from pre-dialysis to mid-dialysis.
Bold values indicate statistically significant results.
Illustrative LVGLS strain tracing and bullseye plot for hemodialysis patients are shown in Fig. 3. Mean pre-dialysis LVGLS was –9.8 (±0.8) % which is considered impaired compared with previously published values in healthy individuals ranging from –16% to –19% (negative value of LVGLS that are closer to 0 reflects a more impaired LVGLS) [9, 10]. There was a worsening in LVGLS during dialysis (–9.8% to –8.4%; mean increase towards zero (i.e worsening) of 1.4 ± 3.9%; P = .03). In the univariate model, greater reductions in P50 during hemodialysis was not associated with worsening in LVGLS (β = 3.15; 95% CI for slope –1.17, 7.48). However, in the multivariate model that accounted for age, sex and race, greater reduction in P50 during hemodialysis was significantly associated with worsening in LVGLS (β = 4.35; 95% CI for slope 0.35, 8.36). Older age was also found to significantly associate with worsening in LVGLS (β = 2.13; 95% CI for slope 0.23, 4.03). Sensitivity analyses adjusting for potential confounders individually (ultrafiltration, potassium changes, presence of intradialytic hypotension) did not meaningfully change the association between P50 and LVGLS (Supplementary data, Table S3).
Figure 3:
LVGLS in an ESKD patient. 2D-STE tracing of the LV endocardial borders in the (A) apical 4-chamber view, (B) apical 3-chamber view and (C) apical 2-chamber view. (D) Depicts a bulls-eye plot with regional (each segment) and global (average) endocardial strain using a 17-segment model. Nearly normal strain is represented in red, whereas impaired strain is represented in shades of red/pink, with areas of most impaired strain depicted in very light pink or blue. In this patient GLS is impaired at –10.5%. AP4, apical 4-chamber; AP3, apical 3-chamber; AP2, apical 2-chamber; Endo, endocardial.
DISCUSSION
In this study of 20 patients with prevalent ESKD receiving maintenance hemodialysis, we found that P50 values significantly decreased during the hemodialysis procedure. Among several predictors that were evaluated, higher pre-dialysis serum phosphate, and greater reductions in phosphate during the hemodialysis procedure were associated with the magnitude of P50 reduction in univariate models. These two variables were highly correlated with one another. Pre-dialysis serum phosphate concentrations remained significantly associated with P50 changes in a multivariable model. While point estimates were similar for the magnitude of phosphate change, this association was rendered no longer statistically significant in a multivariable model. On average, patients had a worsening in LVGLS during hemodialysis, and greater reductions in P50 levels were significantly associated with the magnitude of worsening in LVGLS. These data provide evidence for a new potential mechanism of tissue hypoxia and impaired LV performance in patients treated with maintenance hemodialysis.
The availability of oxygen at the cellular level is critical for maintaining efficient energy production and normal organ function. P50 is the partial pressure of oxygen where 50% of the oxygen-binding sites on hemoglobin are occupied. This is an important parameter on the Hgb:O2 dissociation curve as factors that alter P50 and the shape of the Hgb:O2 dissociation curve determine tissue delivery [11]. These include the pH, temperature of the blood, and concentration of 2,3DPG in the erythrocytes. A reduced P50 indicates higher affinity of hemoglobin for oxygen, so that less oxygen is released at the tissue level, and cells may experience hypoxic stress. The normal physiological range for P50 is narrow so even minor numerical shifts in P50 can have significant relevance for oxygen delivery and tissue oxygenation. For comparison, the magnitude of the mean P50 change observed with the hemodialysis procedure in this study was within the range of changes observed among Mt Everest climbers [12] and critically ill COVID-19 patients in the intensive care unit [13].
At the cellular level in erythrocytes, phosphate may play a key role in oxygen transport and P50 by influencing the intracellular concentration of 2,3DPG. 2,3DPG has an inverse relationship with tissue hypoxia such that reductions in 2,3DPG concentrations lower P50, increasing affinity of hemoglobin for oxygen, inducing greater tissue hypoxia. Erythrocyte intracellular phosphate concentrations (including 2,3DPG) have been shown to be dependent and linked to extracellular phosphate through specific phosphate transporters in the erythrocyte cell membranes [14]. We previously reported that reductions of phosphate during dialysis are strongly associated with concurrent 2,3DPG reductions [4], further supporting this potential direct link. Other groups have reported that hypophosphatemia in hemodialysis patients is associated with depression of myocardial performance which is reversible with phosphate repletion [15]. Furthermore, higher pre-dialysis serum phosphate has been independently associated with complications such as intradialytic hypotension in ESKD [16]. Each hemodialysis session is associated with an acute decline in serum phosphate, and because of the diffusion dependent nature of this treatment, patients with higher pre-dialysis serum phosphate concentrations experience more pronounced reduction in phosphate levels. Indeed, we found that higher pre-dialysis serum phosphate, and greater reductions in phosphate during dialysis had a correlation near unity. Importantly, both were significant predictors of intra-dialytic P50 changes in univariate models, with similar findings in multivariable models although only pre-dialysis phosphate remained statistically significant. These observations suggest a complex interplay between phosphate kinetics and oxygen transport dynamics during hemodialysis in ESKD patients. This observation may be of clinical relevance. While the risks and vascular impacts of chronic hyperphosphatemia are well recognized, this study highlights potential biological pathways and risks that may occur because of acute reductions in phosphate during the hemodialysis procedure itself.
The effects of tissue hypoxia secondary to excessive Hgb:O2 affinity become particularly pronounced in organs with high oxygen demand. Even minor reductions in oxygen delivery can potentially lead to acute myocardial ischemia and may result in reduced contractility and overall cardiac performance. In ESKD patients, this vulnerability is further compounded by the rapid fluid and electrolyte shifts that occur during the hemodialysis procedure. Measurements of myocardial deformation by echocardiographic strain imaging allow for a more sensitive, non-invasive measurement of cardiac function through comparison of how much the cardiac muscle shortens during systole compared with its resting length during diastole. Among various strain parameters, LVGLS has emerged as the most robust clinical indicator of cardiac function. LVGLS is considered a predictor of mortality, independent of and incremental to LVEF [8, 17]. We chose to utilize LVGLS in our study as LVGLS is a more reproducible measure of LV function, is independent of preload, and can detect early signs of myocardial dysfunction in asymptomatic patients or those with preserved LVEF [18]. We found that greater reduction in P50 during hemodialysis and older age were significantly associated with acute worsening in LVGLS during the hemodialysis procedure, whereas no significant association was found between P50 changes and LVEF. Given the small sample size, and that LVGLS was only available in a subset of our participants, these findings require confirmation and should be considered as exploratory. Nonetheless, the findings suggest that acute changes in oxygen affinity during the hemodialysis procedure may be a critical determinant of myocardial function. This, in turn, may contribute to myocardial stunning, and intradialytic hypotension, which are both complications that are highly prevalent and strongly associated with adverse outcomes in hemodialysis patients [11, 19]. If these findings are confirmed, it will be important to explore interventions to mitigate the impact of P50 fluctuations and optimize oxygen delivery and cardiac function in the dialysis population. If our hypothesis proves true, preventing acute hypophosphatemia during the hemodialysis procedure should be tested as one such modifiable factor that may hold promise to ameliorate tissue hypoxia and improve cardiac function. Whether this can be accomplished without exacerbating chronic hyperphosphatemia, either through delivery of phosphate through the return line during dialysis, extending dialysis treatment time or other interventions all require future study.
This study has several strengths. We measured laboratory parameters at several time points during hemodialysis, providing a dynamic view of changes. The focus on acute phosphate reductions, oxygen affinity, its determinants and consequences are novel parameters in hemodialysis. We also used 2D-STE, which is a sensitive technique offering more detailed cardiac function assessment than traditional methods and has been studied in detail [8, 17].
This study also has important limitations. The study had limited power because of the small number of patients, requirement for hemodialysis through a tunneled dialysis catheter, and restriction to a single-center outpatient hemodialysis unit. Given heavy recruitment from a Veterans Affairs program, most participants were male. This was an observational study without long-term follow-up. Therefore, the concurrent timing of measurements and cross-sectional nature of the study design preclude determining causation or temporality between concurrent phosphate (and other electrolyte) changes with P50, or with P50 and myocardial function. We acknowledge that venous P50 serves as a practical surrogate for assessing systemic oxygen affinity trends, and that we did not have direct tissue-level oxygen measurements. Our findings regarding P50 changes should be interpreted as reflecting hemoglobin oxygen affinity in the venous compartment, and caution should be used in extrapolating these results directly to tissue oxygenation. We did not have 2,3DPG data for this population. This population had a reduced LVEF, but the etiology of reduced LVEF (ischemic vs non-ischemic) is unknown.
In conclusion, we found that the hemodialysis procedure acutely reduces P50 indicative of increased hemoglobin oxygen affinity, decreased oxygen delivery to tissues and potentially greater tissue hypoxia. We found that higher pre-dialysis serum phosphate and greater reduction in serum phosphate during the hemodialysis procedure were associated with the magnitude of P50 reduction. Participants with the greatest acute reductions in P50 experienced the greatest deteriorations in LVGLS, on average. Future studies should explore mechanisms of P50 alterations in hemodialysis patients and examine effects of intra-dialytic phosphate changes on cardiac stunning, and intradialytic hypotension.
Supplementary Material
ACKNOWLEDGEMENTS
We extend our sincere gratitude to the laboratory staff, dialysis nurses at VA Greater Los Angeles Health Care System for their invaluable assistance in conducting this study. Their expertise and dedication were crucial to the successful completion of this research. We also wish to express our heartfelt thanks to the patients who participated in this study, without whom this work would not have been possible.
Contributor Information
Shilpa Sharma, Division of Nephrology, Department of Medicine, David Geffen School of Medicine at UCLA, and VA Greater Los Angeles Healthcare System, CA, , USA.
Kim-Lien Nguyen, Division of Cardiology, Department of Medicine, David Geffen School of Medicine at UCLA, and VA Greater Los Angeles Healthcare System.
Isidro B Salusky, Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, CA, , USA.
Tomas Ganz, Department of Medicine, David Geffen School of Medicine at UCLA.
Joachim H Ix, Division of Nephrology-Hypertension, University of California San Diego, and Veterans Affairs San Diego Healthcare System, San Diego, CA, , USA.
FUNDING
VA-IK2-CX002195 (SS), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) K24DK110427 (JHI), VA-MERIT I01-CX001901 (KLN), National Heart, Lung, and Blood Institute (NHLBI) R01HL148182 (KLN), NIDDK U2CDK129496, UO1DK122013 (IBS).
AUTHORS’ CONTRIBUTIONS
Study design: S.S., K.-L.N., I.B.S., T.G., J.H.I. Data analysis: S.S. and J.H.I. Drafting of manuscript: S.S. Revising of final version: S.S., K.-L.N., I.B.S., T.G., J.H.I. Approving final version and decision to submit for publication: S.S., K.-L.N., I.B.S., T.G., J.H.I.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are not publicly available due to patient privacy concerns but are available from the corresponding author upon reasonable request.
CONFLICT OF INTEREST STATEMENT
J.H.I. is principal investigator of an investigator-initiated research grant from the Breakthrough T1D foundation and has served on advisory boards for Alpha Young, AstraZeneca, Ardelyx Inc., Marius Pharmaceuticals and Jnana Inc. I.B.S. has served on advisory boards for Akebia, Inozyme and Ardelyx. T.G. has been a consultant for Ionis Pharma, Silence Therapeutics, Chugai and Disc Medicine, and is a scientific founder and shareholder of Intrinsic LifeSciences, LLC.
The remaining authors have declared no relevant conflicts of interest.
REFERENCES
- 1. Foley RN, Parfrey PS, Sarnak MJ. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis 1998;32:S112–9. 10.1053/ajkd.1998.v32.pm9820470 [DOI] [PubMed] [Google Scholar]
- 2. Nadkarni GN, Uribarri J. Phosphorus and the kidney: what is known and what is needed. Adv Nutr 2014;5:98–103. 10.3945/an.113.004655 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Jankowski J, Floege J, Fliser D et al. Cardiovascular disease in chronic kidney disease: pathophysiological insights and therapeutic options. Circulation 2021;143:1157–72. 10.1161/CIRCULATIONAHA.120.050686 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Sharma S, Brugnara C, Betensky RA et al. Reductions in red blood cell 2,3-diphosphoglycerate concentration during continuous renal replacment therapy. Clin J Am Soc Nephrol 2015;10:74–79. 10.2215/CJN.02160214 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Chazot G, Lemoine S, Kocevar G et al. Intracellular phosphate and ATP depletion measured by magnetic resonance spectroscopy in patients receiving maintenance hemodialysis. J Am Soc Nephrol 2021;32:229–37. 10.1681/ASN.2020050716 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Ginsberg C, Miller LM, Ofsthun N et al. Differences in phosphate and parathyroid hormone concentrations over the day among patients on hemodialysis. J Am Soc Nephrol 2022;33:2087–93. 10.1681/ASN.2021111493 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Burton JO, Jefferies HJ, Selby NM et al. Hemodialysis-induced repetitive myocardial injury results in global and segmental reduction in systolic cardiac function. Clin J Am Soc Nephrol 2009;4:1925–31. 10.2215/CJN.04470709 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Stanton T, Leano R, Marwick TH. Prediction of all-cause mortality from global longitudinal speckle strain: comparison with ejection fraction and wall motion scoring. Circ Cardiovasc Imaging 2009;2:356–64. 10.1161/CIRCIMAGING.109.862334 [DOI] [PubMed] [Google Scholar]
- 9. Marwick TH, Leano RL, Brown J et al. Myocardial strain measurement with 2-dimensional speckle-tracking echocardiography: definition of normal range. JACC Cardiovasc Imaging 2009;2:80–4. 10.1016/j.jcmg.2007.12.007 [DOI] [PubMed] [Google Scholar]
- 10. Bussadori C, Moreo A, Di Donato M et al. A new 2D-based method for myocardial velocity strain and strain rate quantification in a normal adult and paediatric population: assessment of reference values. Cardiovasc Ultrasound 2009;7:8. 10.1186/1476-7120-7-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Stefansson BV, Brunelli SM, Cabrera C et al. Intradialytic hypotension and risk of cardiovascular disease. Clin J Am Soc Nephrol 2014;9:2124–32. 10.2215/CJN.02680314 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Winslow RM, Samaja M, West JB. Red cell function at extreme altitude on Mount Everest. J Appl Physiol Respir Environ Exerc Physiol 1984;56:109–16. [DOI] [PubMed] [Google Scholar]
- 13. Ceruti S, Minotti B, Glotta A et al. Temporal Changes in the oxyhemoglobin dissociation curve of critically ill COVID-19 patients. J Clin Med 2022; 11:788. 10.3390/jcm11030788 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Shoemaker DG, Bender CA, Gunn RB. Sodium-phosphate cotransport in human red blood cells. Kinetics and role in membrane metabolism. J Gen Physiol 1988;92:449–74. 10.1085/jgp.92.4.449 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Davis SV, Olichwier KK, Chakko SC. Reversible depression of myocardial performance in hypophosphatemia. Am J Med Sci 1988;295:183–7. 10.1097/00000441-198803000-00004 [DOI] [PubMed] [Google Scholar]
- 16. Ravi KS, Reeves PB, Correa S et al. Predialysis serum phosphate and intradialytic hypotension. Hemodial Int 2022;26:38–47. 10.1111/hdi.12971 [DOI] [PubMed] [Google Scholar]
- 17. Hung CL, Verma A, Uno H et al. Longitudinal and circumferential strain rate, left ventricular remodeling, and prognosis after myocardial infarction. J Am Coll Cardiol 2010;56:1812–22. 10.1016/j.jacc.2010.06.044 [DOI] [PubMed] [Google Scholar]
- 18. Thavendiranathan P, Poulin F, Lim KD et al. Use of myocardial strain imaging by echocardiography for the early detection of cardiotoxicity in patients during and after cancer chemotherapy: a systematic review. J Am Coll Cardiol 2014;63:2751–68. 10.1016/j.jacc.2014.01.073 [DOI] [PubMed] [Google Scholar]
- 19. Flythe JE, Xue H, Lynch KE et al. Association of mortality risk with various definitions of intradialytic hypotension. J Am Soc Nephrol 2015;26:724–34. 10.1681/ASN.2014020222 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are not publicly available due to patient privacy concerns but are available from the corresponding author upon reasonable request.