An update on progress on the KRUK DIAMONDS project which has come with us to the University of Cumbria……

Stoneygate Research Project

Development and validation of an algorithm for real-time prediction of
intradialytic hypotension using continuous non-invasive blood pressure
monitoring (The dialysis monitoring for decision support study – DIAMONDS)

Grant start date: 01/08/2021  Duration: 24 months

Research Team:

Clinical Problem: Intradialytic hypotension (IDH) occurs in 20% of haemodialysis treatments, leading to end-organ ischaemia, increased morbidity and mortality, and contributing to poor quality of life for patients. Brachial blood pressure (BP) is recorded only intermittently during haemodialysis making early detection and prediction of hypotension impossible. Currently, treating IDH is therefore reactive. Non-invasive continuous BP monitoring would allow earlier detection of IDH and support development of methods to predict and thus prevent hypotension.
Solution: We have developed dedicated hardware and novel analysis methods for continuously estimating BP using pressure sensors in the extra-corporeal dialysis circuit and have published this work in peer reviewed journals (1, 2). This approach allows BP to be estimated accurately from the arterial pressure in the arteriovenous fistula and can be achieved without any extraneous monitoring equipment attached to the patient.
We have validated this against concurrent BP measurement. We now propose to use this technology to develop, train and validate methods and algorithms to be used in combination with continuous non-invasive BP monitoring to predict IDH in real time with sufficient notice to facilitate interventions to prevent it.

Expected Impact

The output from this research will be a non-invasive, low-cost technology for continuous monitoring of intradialytic blood pressure and prediction of intradialytic hypotension (IDH) in real time to provide decision support to renal unit staff to initiate actions to prevent IDH, thereby delivering personalised treatment to improve patient experience as well as reducing morbidity and possibly also mortality

Research Proposal: The research is divided into three work packages (WP)

WP1: We will study a cohort of 50 participants (including a subgroup identified as being prone to IDH)
throughout three consecutive dialysis sessions each using continuous non-invasive BP monitoring to generate a large dataset. This will add both breadth and depth to our current data repository of intradialytic monitoring. The repository consists of (typically) 4-hour monitoring of physiological measurements such as blood pressure, ECG, arterial and venous dialysis line pressure, blood pump flow, treatment changes/interventions. The complete repository will be of sufficient size to be split into training and validation sets for the development of robust predictive algorithms in WP2.

WP2: We will use data from WP1 to develop an algorithm for real-time prediction of IDH using a hybrid of statistical signal processing and artificial intelligence techniques. The WP1 repository will contain a set of BP responses to treatment which are diverse enough to develop and train algorithms and computer models which in turn can predict the onset of IDH with sufficient accuracy and timeliness to inform staff, allowing interventions to be implemented to avoid the occurrence of hypotension. The output from this WP will be a laptop-connected monitoring device which displays historical and current BP concurrently with a window displaying predicted BP up to a defined time horizon.

WP3: We will validate the predictive algorithms in a further cohort of 20 hypotension-prone participants. This will demonstrate not only the feasibility of the approach, but also the expected prediction time horizon which can reasonably be expected. The output of WP3 will be prospective assessment and validation of the accuracy of the WP2 outputs during hypotensive episodes. Additionally, it will enable analysis of the relationship between accuracy and how far into the future BP predictions can reliable be made. Ultimately, this work aims to improve tolerability of haemodialysis, reducing patient symptoms and protecting from ischaemic end-organ dysfunction.

Lay Description of Research

Haemodialysis is a form of dialysis during which a person’s blood is filtered by pumping it through an artificial filter. This partly replaces the normal function of the kidneys and is life sustaining for those who have developed kidney failure. To remain healthy, most people require four hours of haemodialysis, three times per week. Unfortunately, haemodialysis also has several negative side-effects. One of the most important negative effects is that a person’s blood pressure may drop suddenly during the treatment, a condition called intradialytic hypotension, or “IDH”. Patients who experience IDH may have unpleasant symptoms including muscle cramps, light headedness, fainting, sweating, sickness and vomiting. In addition, our previous research has shown that IDH is associated with a decrease in the ability of the heart to pump blood to other important organs like the brain, kidneys and bowels. Moreover, multiple episodes of IDH over time are associated with a permanent decrease in heart function and reduced survival on haemodialysis. Several approaches have been tried to reduce IDH but most are effective only in some patients, so IDH remains a common and dangerous complication of haemodialysis.
One of the main factors making it difficult to prevent IDH is that blood pressure can only be measured at time intervals that do not allow early detection of a sudden drop. Typically blood pressure is measured every 30 minutes during dialysis using a standard blood pressure machine. This means that low blood pressure is usually only detected after symptoms develop and steps are taken to improve the blood pressure only after it has dropped. It would be far better if methods were available to prevent IDH before it starts.

In order to help achieve this goal, our research group has developed a new method for monitoring the blood pressure continuously during a haemodialysis treatment. This relies on the use of pressure sensors that are placed on the plastic tubing that is used to pump the blood out of and back into the patient. By using computer programmes, we have been able to estimate the patient’s blood pressure directly from the pressure readings in the tubing. This means that we can now monitor the blood pressure continuously during haemodialysis without the need for a blood pressure cuff or any other external equipment that attaches directly to the patient. In the proposed project, we plan to use this new technology to develop a method for predicting a drop in blood pressure during dialysis before it happens, so that steps can be taken to prevent IDH (rather than treating it). The proposed research will be divided into three work packages (WP).

In WP1 we will study fifty patients during three of their normal dialysis treatments using our new method of continuous blood pressure monitoring. This will provide us with a large amount of information about how the blood pressure changes during dialysis and when it drops.

In WP2 we will use the information obtained during WP1 to develop a computer programme (algorithm) that will predict IDH before it happens. This will require computer modelling techniques and the use of artificial intelligence. We will divide the information from WP1 into two groups; one will be used to develop the computer algorithm and the other will be used to test that it works.

In WP3 we will study the new computer algorithm combined with our method for continuous blood pressure monitoring in twenty further patients to test how accurately it predicts IDH, and how long before the IDH occurs the computer can predict it. If the project is successful, we will have developed and validated a new method for continuous monitoring of blood pressure during haemodialysis and predicting IDH. This will transform the way in which dialysis is provided by enabling dialysis nurses to take action to prevent IDH before it happens, thereby protecting patients from the unpleasant symptoms and the long-term negative effects on their health and survival. It will also make it possible to provide dialysis in a personalised way so that each patient will have treatment that is tailored to the way that
their body responds to dialysis, rather than treating all patients in the same way.

Research Strategies

The vision for Kidney Research UK’s strategy is to “free lives from the restrictions, fear anxiety and life-limiting nature of kidney disease”. The strategy document highlights that five people every week die while on the waiting list for a kidney transplant. By developing an innovative method to predict and therefore prevent intradialytic hypotension (IDH) this project will directly contribute to improving quality of life on haemodialysis and also likely to improve survival, so that more people are able to benefit from a kidney transplant. We propose to develop a completely novel, low-cost method of monitoring blood pressure during dialysis that has the potential to transform the way in which haemodialysis is provided. We believe that this project supports the concepts of “bigger, faster and bolder” expressed in the strategy.

Purpose: Transform treatments by making dialysis and transplantation more tolerable and effective, until better alternatives are available: Our proposal directly addresses this by developing a technology for predicting and preventing IDH. This can be expected to result in improved quality of life and better survival for people on dialysis.
Actions: Accelerate discoveries: This proposal represents a step-change in the approach to monitoring blood pressure during dialysis, in addition to developing technology to predict IDH in real-time. This will rapidly facilitate the development of interventions to prevent IDH to improve quality of life and survival on haemodialysis.
Impact: i) Better treatments. Technological advances in dialysis, to make it more tolerable and effective: Our proposal directly addresses this desired impact. Moreover, the technology can be used with any haemodialysis machine and can therefore be widely applied without the need for new dialysis equipment. ii) Kinder monitoring. Developing and applying continuous non-invasive BP monitoring during dialysis will improve the quality of monitoring and reduce the need for blood pressure cuffs or other monitoring equipment that patients find uncomfortable and inconvenient.

Specific KRUK research strategy themes addressed:

Emerging Technologies: This proposal will support the development of a new technology including the use of advanced mathematical modelling and artificial intelligence techniques to predict IDH in real time.

Translational and proof of concept: We will build on our previous basic research that used an experimental ex vivo model of the circulatory system during haemodialysis and seek to translate this into a clinically applicable technique for continuous non-invasive monitoring of intradialytic blood pressure and prediction of IDH.

Managing and treating complications of kidney diseases: Intradialytic hypotension is a common complication of haemodialysis and is associated with ischaemia of multiple vascular beds including in the heart, brain, kidneys and gastrointestinal tract which in turn contributes to multimorbidity. Predicting and preventing IDH will therefore contribute to reducing the adverse effects of haemodialysis.

Improving clinical care and quality of health/life: The technology is completely non-invasive and will have no negative impact on the patient experience of dialysis. Moreover, a reduction in IDH would likely have a beneficial impact on symptoms and post-dialysis recovery, thereby improving quality of life.

Moreover, this project will directly address several evidence gaps identified in the UK Renal Research Strategy, including:

Dialysis:

• Biological and psychosocial effects: Intradialytic hypotension represents a major biological adverse
  effect of haemodialysis that substantially impacts quality of life and survival. Addressing the lack of
  adequate measures to prevent IDH is therefore an important priority.
• Understanding and reducing cardiovascular ill-health and mortality: IDH directly impacts cardiovascular health by provoking myocardial stunning and chronic myocardial damage as well as ischaemia in multiple critical vascular beds in the brain, kidneys and gastrointestinal system. This project will enhance understanding of the mechanisms of IDH and importantly will develop the technology to predict IDH in real time with enough notice to enable preventative interventions.
• Technological innovations: Application of advanced mathematical modelling and artificial intelligence techniques has enabled our group to achieve non-invasive monitoring of intradialytic blood pressure using pressure sensors placed on the dialysis lines. A key innovation was the ability to filter out the impact of the dialysis machine peristaltic pumps to estimate arterial pressure transmitted to the fistula. We will build on this technology further by using artificial intelligence techniques to predict episodes of IDH in real time.

Patient experience, quality of care and quality of life:

• Making the leap into personalised medicine: This innovation represents a substantial step forward in developing personalised delivery of haemodialysis. Continuous non-invasive monitoring of blood pressure will enable dialysis to be tailored specifically to the haemodynamic responses of an individual patient to help prevent IDH, reduce intradialytic symptoms and improve patient experience.

Health Inequalities

People on dialysis represent an extremely vulnerable group due to an associated high symptom burden, largevnumber of comorbid conditions and poor long-term survival. In addition, the treatment burden associated with having to attend the dialysis unit three times per week for four hours in considerable. Moreover, many patients report feeling very tired or unwell after haemodialysis, often to the point of requiring them to rest for the remainder of the day. These post-dialysis symptoms likely result from intradialytic hypotension (IDH) and the resultant organ ischaemia.
It is important to note that only a minority of people on dialysis are suitable for a kidney transplant because many are excluded due to advanced age and/or comorbid conditions that place them at high risk of peri-operative and post-transplant complications. These people face the rest of their lives on dialysis and therefore, poor quality of life. The proposed project will seek to address this unfortunate health inequality by developing technology to predict and prevent IDH and the associated organ ischaemia. This affords the prospect of providing haemodialysis in a manner tailored to each patient’s physiological response to minimise unpleasant intradialytic symptoms and reduce comorbidity. Following this project we will seek to test several interventions to be used with the DIAMONDS technology to prevent IDH. This will lead to a funding application to conduct a large clinical trial to investigate the benefits of our approach on patient quality of life and survival.