End-stage renal disease typically requires haemodialysis to help replace kidney function. However, changes in oxygen uptake during haemodialysis have been linked to increased all-cause mortality. This complication of haemodialysis is linked to decreasing fluid volume, compromising blood flow to tissue and leukostasis within pulmonary tissue. However, an alternative cause of reduced oxygen availability (hypoxia) during haemodialysis is acute alkalosis. Alkalosis during haemodialysis can cause hypoxia via dysregulated ventilation and impaired ability for tissue to extract oxygen. Despite strong rationale for these mechanisms, few studies have fully explored causes of hypoxia during haemodialysis. Greater understanding may help to mitigate the risk associated with this vital treatment option.

The study will comprise of end-stage renal disease patients who regularly undergo haemodialysis. Three blood samples will be attained before, during and after haemodialysis to assess arterial blood gases. In a small subset of patients, white blood cell (WBC) count and cardiac output will be assessed via a non-invasive cardiac output monitor during treatment. Regression analysis will be performed to help identify predictors of hypoxia during haemodialysis. Patient burden is negligible, with blood samples attained from the dialyser as part of routine treatment. In the patients who agree for cardiac output assessment, the patient will be required to have four small noninvasive sensor pads placed on the chest. Patients will be assessed over 3 consecutive treatments during a single week.

Alkalosis may explain multiple complications of HD treatment, most notably hypoxemia. Despite a strong rationale for these mechanisms, current consensus suggests ischaemia as the predominant catalyst for hypoxia with HD. However, few studies have characterised pH changes during HD with many using serum bicarbonate as a proxy of pH status. Nevertheless, these studies predominantly indicate acute alkalosis occurrence with HD. Despite these findings, few HD units individualise their bicarbonate prescription or assess pH prior to treatment. Of 554 dialysis facilities worldwide, 49% did not individualise bicarbonate prescriptions. Non-individualised bicarbonate prescription risks sub-therapeutic serum bicarbonate levels and sub-clinical hypoxemia during HD. Assumption of acidosis and consistent baseline pH prior to each HD session may further risk acute alkalosis with HD. Indeed, a normal pH was identified in 26 ESRD patients on commencing HD whilst prescribed dialysate bicarbonate of 32 mmol/L (7.38 ± 0.06) or 26 mmol/L (7.35 ± 0.06). However, pH increased throughout HD with acute alkalosis within the 32 mmol/L group (7.48 ± 0.05). These data, suggest individualised assessment of pH prior to HD could mitigate alkalosis and subsequent hypoxemia. However, small cohorts, limited studies of pH changes over consecutive HD treatments, and few studies of links between hypoxemia and pH limit current understanding. This study therefore aims to explore pH changes and occurrence of hypoxia with HD.

Adult patients (18+) will be recruited irrespective of age, gender, or ethnicity. Due to the observational nature of the study, no personal benefits will be expressed to the patient. Signed consent will be attained from a patient after a minimum of 24hrs has been provided on receipt of the participation information sheet. Any questions will be clearly answered prior to obtaining consent from the patient. Only patients who have full capacity to give consent and full understanding of the purpose and procedures of the study will be recruited.

There is no foreseen risk to the patient from the study. All blood samples are attained via arterial-venous catheter which is inserted prior to treatment as part of the patient's routine haemodialysis. In the small subset of patients who have their cardiac output assessed, the use of four small non-invasive sensors poses no additional discomfort, risk or restriction throughout routine haemodialysis treatment. In a previous project using the noninvasive cardiac output monitor (NICOM), no patient expressed any discomfort whilst using the device during haemodialysis.

All patients involved in the project will be anonymised with a study number for data collection purposes. No identifiable information will be used during post-hoc analysis. Signed consent forms or any other identifiable data will be securely logged within the study site file and secured within the renal research office. Only the chief investigator and renal research nurses will be aware of patient involvement in the study. Access to identifiable data will be strictly limited to the chief investigator.

There is no conflict of interest from the chief investigator or other members of the research team that may influence the study. After completion of the study, data will be compiled and drafted for scientific publication with all data securely archived within the study site file at the renal research department. Any patient expressing interest in the study findings will be provided with the appropriate information verbally or by a written summary of findings.

All blood samples will consist of ~1.5ml which will be analysed immediately after sampling. Samples will subsequently be disposed of immediately after analysis and not be used for any subsequent studies.

Inclusion Criteria:

• On haemodialysis for at least 3 months.
• Haemodialysis 3 times per week.
• Age 18 years or older.

Exclusion Criteria:

• Central venous catheter.
• Planned kidney transplant during the duration of study.

End-stage renal disease patients will be recruited from the University Hospitals Coventry & Warwickshire hemodialysis unit identified by the research team and renal medical professionals.