Tissue ischemia worsens during hemodialysis in end-stage renal disease patients

Introduction

Cognitive deficit is a common problem in chronic kidney disease patients. This problem occurs even in its milder stages (estimated glomerular filtration rate [eGFR] <30 mL/min per 1.73 m²), which were associated with a 47% increased odds of cognitive impairment relative to those with an eGFR 45 to 59 mL/min per 1.73 m² in the CRIC trial (1). As hemodialysis is widely available treatment of end-stage renal disease (ESRD) in developed countries, the hemodialysis population is getting older and sicker. As a consequence, up to 70% of hemodialysis patients aged ≥55 years have moderate-to-severe chronic cognitive impairment, yet it is largely undiagnosed (2, 3). Various studies investigated the reasons for this socio-economically important issue. Cognitive impairment is without doubt related to age, but also to education level and other factors. Interestingly, the global cognitive performance was also indirectly and significantly related to the level of glycated hemoglobin A_1C and of fibrinogen (3). Clinically, the term “uremic encephalopathy” is used. It accents the impact of metabolic changes associated with ESRD. However, ESRD patients treated by hemodialysis also suffer from cyclic hemodynamic changes. They include repeated slower water accumulation followed by fast fluid removal by ultrafiltration; these fluid changes are also accompanied by the changes of blood pressure. During hemodialysis session, fluid removal leads to blood pressure decrease. In some patients this decrease is more pronounced – they suffer from intradialytic hypotension and this phenomenon has been linked to neurological consequences (4) and also to worse prognosis (5).

Leukoaraiosis is the main magnetic resonance finding in ESRD patients. It is a form of subcortical injury, which occurs in a watershed area of the brain, where episodic intradialytic perfusion would be expected to have its maximal effect (6, 7). However, there is lack of specific data documenting this hypothesis. Recently, near-infrared spectroscopy (NIRS) has been used for monitoring the regional saturation of oxygen (SrO₂) of the frontal lobes of the brain. Dialysis patients had lower baseline brain SrO₂ than healthy controls; diabetic patients had even lower SrO₂ than non-diabetics (8, 9). The monitoring of brain SrO₂ is established in anesthesia and in critical status monitoring. Cerebral oxygen desaturation during cardiac surgery predicted cognitive decline after discharge and longer hospital stay (10). Similar relation was found after total knee replacement (11), but in ESRD population such data are still lacking. It is known that SrO₂ at the end of dialysis does not differ from the value prior to dialysis (12), but the time course of SrO₂ has not yet been published.

Hand ischemia is a frequent complication of dialysis vascular access creation. It occurs in up to 8% of dialysis patients (13). Several mechanisms are involved, including the steal phenomenon in the distal part of the radial artery at the forearm access and flow competition between the naturally high-resistant tissue arterial bed and low-resistant shunt. Some of the affected patients feel hand pain during dialysis, probably as a consequence of blood pressure and cardiac output decline (13, 14). Hand pain is a manifestation of ischemia and the hand is easily accessible by NIRS. High vascular access flow volume could have systemic hemodynamic impact including heart failure, characterized by inadequate organ blood supply (15).

Altogether, the changes of tissue oxygen saturations are probable both in the brain and in the dialysis access hand during hemodialysis. It was for this reason we started a pilot study of SrO₂ monitoring during a routine hemodialysis session and here we present the initial data.

Materials and methods

We asked patients treated by chronic hemodialysis in the General University Hospital, Prague, to take part in this study. Only those, who agreed and signed the Informed Consent, were included. Furthermore, the inclusion criteria included clinically stable state, lack of overt dementia, stroke or hand ischemia. However, no special intervention was made prior to the inclusion to improve the patient’s habits regarding their compliance to the drinking restriction or medication and the date of the study hemodialysis session was kept secret. We also did not change the setting of the dry weight for the purpose of this study. First we examined brain SrO₂ in a group of 10 ESRD patients at rest. After recognizing low brain SrO₂ values in ESRD patients, we examined another group of 17 subjects throughout hemodialysis.

Control group included 17 healthy individuals of the same age and gender distribution. It was assembled to take a measurement of cerebral and hand tissue SrO₂ at rest conditions (at least 5 minutes) for comparison with the baseline values of the study group.

This study was approved by the Institutional Ethical committee and conforms with the Helsinki Declaration of 1975, as revised in 2000. The principles of the study were explained to all patients and only those who signed the Informed consent were included.

Basic hemodialysis data were recorded and included total ultrafiltration (UF), ultrafiltration rate, blood pump rate (Qb), together with the routine non-invasive measurement of blood pressure on the no-dialysis access upper extremity and the heart rate 30 minutes after the beginning of dialysis and then every hour. Although systolic and diastolic pressure was obtained, we calculated the mean arterial pressure (MAP), which better correlates with the tissue perfusion pressure, by the following equation: MAP = diastolic pressure + (systolic minus diastolic pressure)/3. Moreover, the finger systolic blood pressure was measured prior to the hemodialysis using pulse oximetry monitor and a special finger cuff. The systolic finger pressures were estimated on 2^nd-4^th fingers and the mean systolic finger pressure (FP) and finger/brachial pressure (FBP) index were calculated as shown earlier (13). The blood flow rate through a vascular access (Qa; arteriovenous fistula or arteriovenous graft) and cardiac output (CO) was measured by ultrasound using Vivid Q (General Electric Corp., Connecticut, USA) operated by a skilled ultrasonographer (>300 examinations per year) immediately before the study dialysis session. Qa was calculated using the inner vessel diameter and time-averaged mean flow velocity (brachial artery in arteriovenous fistulas and ePTFE graft in arteriovenous grafts) as reported earlier (16). The mean of the three measurements was calculated. Similarly, cardiac output was measured echocardiographically by multiplying the cross-sectional area and velocity time integral of the left ventricular outflow tract (also recorded by Vivid Q). Cerebral tissue oxygen saturation (cerebral SrO₂) and hand tissue oxygen saturation (hand SrO₂) were continuously monitored during the hemodialysis session. We used the INVOS 5100C oximetry system (Covidien Somanetics, Essex, UK) with two sensors – one was attached to the patient’s forehead at the dominant hemisphere, the other to the dorsum of the dialysis access hand. These sensors possess light-emitting diodes transmitting near infrared light at two wavelengths (735 and 810 nm) and two silicon photodiodes as light detectors. The system recorded values in two channels every 6 seconds. Baseline values were obtained during 10 minutes before the start of the hemodialysis. Data were ­averaged in 5-minute intervals and visualized in a chart. ­According preliminary experience, we analyzed the curves to identify the baseline, minimal, maximal and end values and the decrease slope (from baseline to minimum) in each channel. Summary charts of cerebral and hand SrO₂ were composed of data from all patients. In addition to routine blood samples, a blood count was taken before the hemodialysis session.

The data are expressed as means ± SD after passing D’Agostino & Pearson omnibus normality test; p>0.05 was considered as significant for the normal Gaussian distribution. Differences of parameters with normal Gaussian distribution were analyzed by the paired and unpaired t-test as appropriate. Pearson correlation between UF rate, Qb, Qa, CO and SrO₂ data was provided. P values <0.05 (two-tailed) were considered statistically significant. Statistical analysis and graphs were performed using Prism 5.0 (GraphPad, USA).

Results

Characteristics of the study population

We performed resting SrO₂ measurement at 27 ESRD patients (1 African, 2 Asians, 24 Caucasians; of these,14 (52%) women, the mean age was 60.3 ± 15.6 years, range 27-85 years, dialysis vintage 71 ± 59 months, 6 (22%) patients suffered from diabetes mellitus.

Seventeen of 27 patients were examined throughout hemodialysis. Their characteristics were as follows: 8 (47%) women and 9 men (53%), the mean age was 59 ± 18 years, range 27-85 years. Incidences of main comorbidities in the study population were the following: coronary artery disease 4 (24%), congestive heart failure 2 (12%), diabetes mellitus 4 (24%), arterial hypertension 15 (88%), dyslipidemia 6 (35%), history of thromboembolism 5 (29%), smokers/former smokers 8 (47%). Fourteen patients had a native arteriovenous fistula (11 radio-cephalic, 1 Gracz, 2 brachio-cephalic) and three had an arteriovenous graft (1 radio-basilic, 2 brachio-basilic). Finger pre-dialysis systolic pressure was 124 ± 31 mmHg and FBP index 0.80 ± 0.16. The control group consisted of 17 subjects (8 females and 9 males) aged 60.0 ± 5.1 years, their blood hemoglobin concentration was significantly higher than in the patient group (131 ± 9 g/L vs. 97 ± 13, p<0.0001) and mean arterial pressure (MAP) was lower than in the patient group before hemodialysis (101 ± 6 vs. 120 ± 31 mmHg, p<0.001.

Brain SrO₂ – baseline values

The studied ESRD group had significantly lower brain SrO₂ prior to the dialysis than the healthy controls (51.5 ± 10.9 vs. 68 ± 7%, p<0.0001).

Hemodialysis and hemodynamics

During hemodialysis, there were significant changes in the mean arterial pressure (MAP) and in the heart rate (HR); minimal values of MAP and minimal/maximal values of HR differed significantly from baseline. Detailed data are shown in Table I. Comparing the baseline and end-dialysis values, MAP dropped significantly from 120 ± 25 to 108 ± 28 mmHg (p = 0.02), but HR was even the same (81 ± 15; 77 ± 16 bpm, p = NS).

TABLE I -

Basic hemodynamic data and tissue oxygenation changes during hemodialysis

Hemodialysis and hemodynamics	Result and comparison to baseline
* p<0.05.
** p<0.01.
*** p<0.001; ns = non-significant; comparison with the baseline values.
Dry weight (kg)	79.6 ± 30.1
Total ultrafiltration (mL)	3027 ± 1501
Ultrafiltration rate (mL/h)	692 ± 279
Blood pump rate (mL/min)	270 ± 10
Vascular access flow (mL/min)	1249 ± 694
Cardiac output (mL/min)	6035 ± 2011
Blood hemoglobin level (g/L)	97 ± 13
MAP baseline (mmHg)	120 ± 31
MAP end of dialysis (mmHg)	108 ± 28*
HR baseline (beats/min)	81 ± 15
HR end of dialysis (beats/min)	77 ± 16ns
Blood hemoglobin baseline (g/L)	96.7 ± 12.2
Blood hemoglobin end of dialysis (g/L)	109.6 ± 16.2***
Brain SrO2
Baseline (%)	52 ± 8
Minimal value (%)	47 ± 8***
Maximal value (%)	56 ± 7***
End of dialysis (%)	53 ± 7ns
Slope to minimal value (%/h)	-4 ± 3
Hand SrO2
Baseline (%)	55 ± 16
Minimal value (%)	45 ± 14***
Maximal value (%)	63 ± 15***
End of dialysis (%)	53 ± 14***
Slope to minimal value (%/h)	-8 ± 4

Brain SrO₂ during hemodialysis

The averaged time course of cerebral SrO₂ during the first 3 hours of hemodialysis is visualized in Figure 1. There is an obvious drop in the minimum value during the first 35 minutes (from basal 52 ± 8 to 47 ± 8%, p = 0.0002) and consequent slow increase to the maximum (56 ± 7%, p<0.0001). However, the end-dialysis value did not differ from baseline values (53 ± 7%, p = NS). The faster ultrafiltration rate was linked to the steeper SrO₂ decrease according to the correlation analysis (r = -0.72, p<0.001). The blood pressure decrease during the first 30 minutes of hemodialysis was not significantly correlated to the SrO₂. No further significant links to brain SrO₂ was observed.

Fig. 1 -

Hand tissue SrO₂

Resting hand tissue SrO₂ values were significantly lower in the study population in comparison to healthy controls (55 ± 16% vs. 66 ± 8%, p = 0.03). Baseline hand SrO₂ levels were significantly related to finger pressure and FBP index (r = 0.59, p = 0.03 and r = 0.67, p = 0.03, respectively). Similar links were found for the minimal and maximal hand SrO₂ values. On the contrary, the blood hemoglobin was inversely related to the hand SrO₂ (r = -0.75, p = 0.001). The averaged time course of hand tissue SrO₂ is shown in Figure 2. In conformity with cerebral SrO₂ there was a drop to the minimal values during the first 35 minutes (from basal 55 ± 16% to 45 ± 14%, p<0.0001) and then the SrO₂ value cyclically rose to the maximum (63 ± 15 %, p = 0.0003). The end-dialysis value was similar to baseline (53 ± 14%, p = NS).

Fig. 2 -

Discussion

Our study has shown that the regional oxygen saturation both of the frontal lobe and of the dialysis access hand dorsum are lower in ESRD patients and further decline after the initiation of hemodialysis.

Lower resting baseline SrO₂ values were already reported in dialysis patients and these were even more pronounced in diabetic patients (7, 9). However, our observation of significant decrease of frontal SrO₂ after the beginning of hemodialysis is new. The explanation for these findings is not easy. None of our patients suffered from a clinically significant blood pressure decrease (intradialytic hypotension) and the observed decrease of SrO₂ did not significantly correlate with blood pressure decrease, but it correlated with the ultrafiltration rate. The ultrafiltration rate was much higher in the 3-times-weekly hemodialysis schedule than in daily dialysis and reached almost 700 mL/hour in our study, which significantly exceeds the normal rate of fluid loss by diuresis. Fast ultrafiltration is known to be non-physiological and is associated with hypotension risk, mostly due to the preload decrease in other studies (17). At the beginning of the 1990s, Postiglione et al documented significant decrease of middle cerebral artery flow velocity after hemodialysis using ­ranscranial ­Doppler (18). Increased resistance and pulsatility indices after hemodialysis could have reflected the higher peripheral vascular resistance. It was speculated (18) that the decreased cerebral perfusion must not lead to reduced brain oxygen delivery thanks to the hemo-concentration caused by the hemodialysis. Our data document, unfortunately, that the cerebral autoregulation fails to cover adequate brain oxygenation during the beginning of hemodialysis.

Prohovnik et al reported lower cerebral blood flow in dialysis patients in comparison to age-matched healthy subjects (8). Thus, cerebral blood flow and also cerebral oxygenation are lower in ESRD patients and both these variables even decline during the beginning of hemodialysis. What are the mechanisms involved in these changes? Besides age, it was the arterial oxygen content, which mostly determined the middle cerebral artery flow velocity in one study (19). In the study by Ito et al (9), blood pH was independently linked to cerebral SrO₂.

Hemoglobin concentration was significantly lower in the study than in the control groups, as is typical for ESRD patients. This difference could be a reason for the observed lower baseline SrO₂ values in the study group. Interestingly, hemoglobin normalization led to worsened cerebral perfusion and to increased oxygen extraction in a study using positron emission tonometry (20). The authors hypothesized that it was due to decreased red blood cell deformability, but cardiac output also lowered in their study. We can only hypothesize that the increased cerebral blood transition time was also affected by higher blood viscosity. The influence of hemoglobin increase on cerebral oxygenation in adult subjects is not known. A study of the effect of blood transfusion on cerebral oxygenation in preterm newborns has been published (21) and increase in cerebral, splanchnic and renal ­oxygenation was documented. During the initiation of hemodialysis, the saline in the dialysis set is replaced by the patient´s blood, which leads to blood dilution in the patient´s vascular tree. As the ultrafiltration continues, the blood concentration process prevails. This phenomenon could be an alternative explanation for the observed changes in SrO₂ during hemodialysis. However, the fact that SrO₂ values at the end of dialysis did not differ from the values prior to dialysis despite approximately 10% increase of blood hemoglobin concentration, does not support this mechanism. Another alternative explanation for the observed decrease of SrO₂ is the faster metabolic change of serum urea, osmolarity, etc.

The hand represents a high resistant arterial bed with a minimal flow volume under resting conditions. The hand with a dialysis access is ischemia-prone and hand ischemia develops in up to 10% of ESRD patients having a dialysis vascular access and usually the hand pain occurs during hemodialysis. This was why we chose the hand SrO₂ for comparison with the brain. This study has shown that even symptom-free patients have resting SrO₂ lower than age-matched controls and that the tissue oxygenation even decreases after the beginning of hemodialysis similarly to the brain SrO₂. Moreover, hand SrO₂ was directly related to finger pressure, which advocates the usability of SrO₂ for the investigation of hand ischemia. The fact that the tissue oxygenation at both studied arterial beds decreased during the same time period further supports the robustness of our data. Peripheral circulation is regulated by other mechanisms than the brain circulation, so we cannot extrapolate the hand data to the brain oxygenation changes. Nevertheless, patients with lower finger pressure were more prone to ischemia (13).

The limitation of these preliminary data is the small number of included patients – it is a preliminary study and the inclusion of other patients continues. The results are statistically significant and the time course of SrO₂ of brain and hand are interrelated. Blood hemoglobin levels were lower than recommended in our patients. Another limitation is that detailed laboratory data explaining the mechanisms were not collected. Furthermore, the mental status analysis is lacking in this pilot study. We, therefore, cannot determine any “dangerous” baseline SrO₂ or even intradialytic SrO₂ decrease cut-off values. However, a link between SrO₂ decrease and worse neurological outcome and prognosis has been described in surgical patients (10, 11) as mentioned above.

The time course of brain SrO₂ since the initiation of hemodialysis is depicted – records of all patients were averaged into the graph with the sampling frequency 6 seconds. X-axis shows time and y-axis represents the values of SrO₂.

The time course of hand tissue SrO₂ since the initiation of hemodialysis is depicted – records of all patients were averaged into the graph with the sampling frequency 6 seconds. X-axis shows time and y-axis represents the values of SrO₂.

• 1. Kurella Tamura M Xie D Yaffe K et al. Vascular risk factors and cognitive impairment in chronic kidney disease: the Chronic Renal Insufficiency Cohort (CRIC) study. Clin J Am Soc Nephrol 2011 6 2 248 256
• 2. Murray AM Cognitive impairment in the aging dialysis and chronic kidney disease populations: an occult burden. Adv Chronic Kidney Dis 2008 15 2 123 132
• 3. Seidel UK Gronewold J Volsek M et al. The prevalence, severity, and association with HbA1c and fibrinogen of cognitive impairment in chronic kidney disease. Kidney Int 2014 85 3 693 702
• 4. Eldehni MT McIntyre CW Are there neurological consequences of recurrent intradialytic hypotension? Semin Dial 2012 25 3 253 256
• 5. Assimon MM Flythe JE Intradialytic Blood Pressure Abnormalities: The Highs, The Lows and All That Lies Between. Am J Nephrol 2015 42 5 337 350
• 6. McIntyre CW Recurrent circulatory stress: the dark side of dialysis. Semin Dial 2010 23 5 449 451
• 7. Pereira AA Weiner DE Scott T et al. Subcortical cognitive impairment in dialysis patients. Hemodial Int 2007 11 3 309 314
• 8. Prohovnik I Post J Uribarri J Lee H Sandu O Langhoff E Cerebrovascular effects of hemodialysis in chronic kidney disease. J Cereb Blood Flow Metab 2007 27 11 1861 1869
• 9. Ito K Ookawara S Ueda Y et al. Factors affecting cerebral oxygenation in hemodialysis patients: cerebral oxygenation associates with pH, hemodialysis duration, serum albumin concentration, and diabetes mellitus. PLoS ONE 2015 10 2 e0117474.
• 10. Slater JP Guarino T Stack J et al. Cerebral oxygen desaturation predicts cognitive decline and longer hospital stay after cardiac surgery. Ann Thorac Surg 2009 87 1 36 44 discussion 44-45.
• 11. Salazar F Doñate M Boget T et al. Relationship between ­intraoperative regional cerebral oxygen saturation trends and cognitive decline after total knee replacement: a post-hoc analysis. BMC Anesthesiol 2014 14 1 58
• 12. Hoshino T Ookawara S Goto S et al. Evaluation of cerebral oxygenation in patients undergoing long-term hemodialysis. Nephron Clin Pract 2014 126 1 57 61
• 13. Malik J Tuka V Kasalova Z et al. Understanding the dialysis access steal syndrome. A review of the etiologies, diagnosis, prevention and treatment strategies. J Vasc Access 2008 9 3 155 166
• 14. Modaghegh MH Kazemzadeh G Pezeshki Rad M et al. Chronic hemodialysis access-induced distal ischemia (HAIDI): distinctive form of a major complication. J Vasc Access 2015 16 1 26 30
• 15. Vaes RH Tordoir JH Scheltinga MR Systemic effects of a high-flow arteriovenous fistula for hemodialysis. J Vasc Access 2014 15 3 163 168
• 16. Kudlicka J Kavan J Tuka V Malik J More precise diagnosis of access stenosis: ultrasonography versus angiography. J Vasc Access 2012 13 3 310 314
• 17. Stegmayr BG Ultrafiltration and dry weight-what are the cardiovascular effects? Artif Organs 2003 27 3 227 229
• 18. Postiglione A Faccenda F Gallotta G Rubba P Federico S Changes in middle cerebral artery blood velocity in uremic patients after hemodialysis. Stroke 1991 22 12 1508 1511
• 19. Macko RF Ameriso SF Akmal M et al. Arterial oxygen content and age are determinants of middle cerebral artery blood flow velocity. Stroke 1993 24 7 1025 1028
• 20. Metry G Wikström B Valind S et al. Effect of normalization of hematocrit on brain circulation and metabolism in hemodialysis patients. J Am Soc Nephrol 1999 10 4 854 863
• 21. Dani C Pratesi S Fontanelli G Barp J Bertini G Blood transfusions increase cerebral, splanchnic, and renal oxygenation in anemic preterm infants. Transfusion 2010 50 6 1220 1226