Optimising acid-base balance in patients with advanced chronic kidney disease: a quality improvement initiative

Abstract

Advanced chronic kidney disease (CKD) is commonly associated with disturbances in acid-base balance. Studies have shown that correcting metabolic acidosis in CKD offers several clinical benefits, including slower disease progression and reduced mortality. Acidosis can be addressed with oral sodium bicarbonate, an easy-to-administer and relatively inexpensive treatment compared with novel drugs such as sodium-glucose co-transporter-2 -(SGLT-2) inhibitors. Although sodium bicarbonate is frequently prescribed in CKD, less attention is paid as to whether serum bicarbonate levels are adequately optimised. This project aimed to increase the proportion of advanced CKD patients with serum bicarbonate levels within the normal range.

Retrospective data collection was conducted at our tertiary renal centre to establish baseline bicarbonate levels and assess prescribing practices for sodium bicarbonate. Quality improvement (QI) methodology was then applied to implement a series of interventions designed to increase awareness of acidosis and to promote the appropriate use of oral sodium bicarbonate.

Results from baseline data collection showed that 94% of patients had a serum bicarbonate level checked within the previous 6 months. However, only 46% of patients had levels within the optimal range. Although initial improvement was noted following our first intervention, this was not sustained through subsequent QI cycles, and the proportion of patients with optimal bicarbonate levels remained largely unchanged by the end of the project.

Optimising serum bicarbonate is an important component of the management of advanced CKD. While this project led to increased awareness and short-term gains, further work is required to achieve lasting improvements and embed change into routine practice.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Metabolic acidosis is common in advanced chronic kidney disease, and effective treatments are available; however, it remains inadequately managed in many patients. Guidance for renal centres on ways to improve the optimisation of this parameter at a departmental level is limited.

WHAT THIS STUDY ADDS

• Improving the management of metabolic acidosis at a departmental level requires a multipronged approach. Achieving consistent monitoring is an important first step, which must then be followed by effective interventions to promote action when a suboptimal serum bicarbonate measurement is detected.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings of this project emphasise the importance of auditing clinical practice relating to sodium bicarbonate prescribing in renal units and demonstrate the challenges associated with trying to change the status quo. The interventions described may act as a starting point for other units to begin to implement similar quality improvement initiatives, and we hope that they will feel encouraged to share both their successes and learning points.

Problem

Metabolic acidosis is a well-recognised complication of advanced chronic kidney disease (CKD), with an affordable treatment widely available in the form of sodium bicarbonate tablets. Despite this, many patients seen in advanced kidney care (AKC) or ‘low clearance’ clinics continue to have suboptimal serum bicarbonate levels—typically below the normal reference range. This issue was identified locally by team members working in our tertiary renal unit in the urban centre of Birmingham, which serves a large and diverse population. The AKC service manages approximately 430 adult patients, with weekly consultant-led clinics supported by a team of specialist renal nurses. These clinics provide holistic face-to-face reviews focusing on the management of complications of CKD stages 4 and 5, alongside preparing patients for potential renal replacement therapy (RRT) in the form of dialysis or transplant. Our patient population faces significant clinical and social challenges. Multimorbidity is common, with high rates of diabetes, hypertension and cardiovascular disease. Polypharmacy is the norm, and medication adherence can be challenging. Additionally, the community has high levels of economic deprivation, low health literacy and substantial language barriers, all of which influence disease understanding, communication and engagement with care plans.

The clinical team observed that a subset of patients in the AKC clinic had persistently low serum bicarbonate levels on routine clinic blood tests despite regular monitoring. On discussing this issue, the team recognised that there were multiple contributing factors, including patient-related factors (eg, adherence, health literacy and side effects) and system-level issues (eg, prescribing practices, follow-up processes and variation in clinical decision-making). It was clear that there would not be a ‘one size fits all’ solution to improving this metric and that a multi-faceted strategy would be required.

To address this issue, we initiated a quality improvement (QI) project with the goal of improving management of metabolic acidosis in AKC clinics. Our SMART aim was to improve the proportion of patients in the AKC clinic with serum bicarbonate levels in the optimal range (22–29 mmol/L as defined by our local laboratory) by 50% compared with current practice. As this metric had not been recently audited within our unit, we first undertook a baseline data collection exercise to assess current performance and guide future interventions.

Background

Although the clinical relevance of serum bicarbonate levels in patients with CKD has recently been debated,¹ correction of metabolic acidosis still features in both national and international CKD guidelines as an area for consideration by nephrologists.^{2 3} Multiple intervention studies, though limited by small sample sizes, have consistently demonstrated that maintaining serum bicarbonate levels within the normal physiological range is associated with slower decline in kidney function and reduced likelihood of progression to end-stage renal failure.^{4 5} Furthermore, correction of metabolic acidosis contributes to the prevention of other CKD complications, including hyperkalaemia⁴ and mineral bone disorders.⁶ A recent retrospective observational study of over 24 000 patients with CKD stages 3–5 in the USA again demonstrated that increasing serum bicarbonate with medication was associated with a statistically significant reduction in the rate of CKD progression.⁷ Given the size and scope of this study, it has added further weight to the rationale for targeting metabolic acidosis as a modifiable risk factor in CKD management.

Sodium bicarbonate remains the most commonly used pharmacological agent for this purpose. It is affordable, widely available and has a good safety profile. Studies have shown no significant increase in gastrointestinal disorders compared with placebo in controlled settings.⁸ However, in clinical practice, some patients do report gastrointestinal side effects, particularly bloating, abdominal pain and nausea.⁹ A gastro-resistant formulation of sodium bicarbonate is now available in the UK, which offers an alternative for patients who are unable to tolerate standard preparations.

The UK Renal Registry collects national-level data on serum bicarbonate levels for patients on all forms of dialysis and with a kidney transplant.¹⁰ However, to our knowledge, there is no national or Midlands region-level data collection for CKD patients not on RRT. A brief review of the published literature identified no recently published audits or QI projects on optimising serum bicarbonate levels in adults not yet on RRT. It is possible that some units may be collecting this data locally but have not disseminated outcomes or formalised efforts to improve practice. Consequently, there is limited shared knowledge on effective strategies for addressing metabolic acidosis in this patient group. Given this knowledge gap and the existing clinical evidence supporting the correction of serum bicarbonate levels in advanced CKD, we felt it was both timely and necessary to explore potential interventions in our centre.

Measurement

To establish a baseline, on ‘day 0’, we collated a list of the last 100 patients seen in AKC clinics in our renal centre. This included patients from several clinics over the preceding 3-week period. For each patient, we collected retrospective data from the electronic patient record, including demographics, the most recent serum bicarbonate level and when this level was measured. We reviewed their recent clinic letters and electronic GP records for sodium bicarbonate prescriptions and for indications that the patient had reported problems with compliance or side effects. These data points were entered into a spreadsheet and subsequently analysed.

The results of baseline data collection showed that a majority of patients were having serum bicarbonate levels monitored regularly in the clinic (94% had a measurement within the preceding 6 months). The mean serum bicarbonate level in this cohort was 21.5 mmol/L, falling slightly below the optimal range of 22–29 mmol/L. When the distribution of serum bicarbonate levels was examined, only 46% of patients had a level within the optimal range. Although the mean value may appear only slightly below the reference range, this masked a clinically important distribution: more than half of patients had suboptimal bicarbonate levels (<22 mmol/L). This provided the key rationale for undertaking the project, as even modest and sustained deviations from the normal range are associated with adverse outcomes in CKD.

Of the patients with suboptimal bicarbonate levels, 24% of patients did not have sodium bicarbonate tablets prescribed. For two of these patients, there was documentation of previous side effects with sodium bicarbonate, and a further two patients had documented poor compliance.

We then used the three key questions from the ‘Model for Improvement’ approach¹¹ as a framework to structure how we approached the project:

1. What are we trying to accomplish?

A higher proportion of patients with advanced CKD maintaining optimal serum bicarbonate levels as described in our SMART aim.

• 2 1How will we know change is an improvement?

We chose to use serum bicarbonate levels as our primary measure, as it was apparent from baseline data that this measurement was consistently available and provided a validated objective value for comparison between intervention cycles. We could therefore demonstrate change, and hopefully improvement, by monitoring the number of patients in the AKC clinic with serum bicarbonate levels in the target range through repeated cycles of data collection after interventions were implemented.

• 3 1What changes can we make that will result in the improvements that we seek?

The project team brainstormed the following list of potential areas of change, which we felt could lead to a higher proportion of patients having serum bicarbonate levels within the optimal range:

• Education (for both patients and prescribers).

• Having a flowchart or process map available in AKC clinics.

• Prompts to review every patient’s serum bicarbonate level in the low clearance clinic.

• Improved access to an alternative preparation of sodium bicarbonate for patients experiencing side effects.

Design

The project was jointly led by the two authors, who collaboratively brainstormed initial thoughts on potential interventions before seeking wider input. Early engagement with the multidisciplinary team (MDT) was a key priority. We consulted with consultant colleagues and the team of specialist renal nurses who are integral to the running of AKC clinics.

To raise awareness and build initial momentum, we presented the findings of the baseline audit to colleagues at a weekly departmental educational meeting. This session was used to highlight the clinical importance of optimising serum bicarbonate levels in advanced CKD and to share the initial data. Several colleagues expressed surprise at the proportion of patients with suboptimal bicarbonate levels, reinforcing our view that the issue had been under-recognised. We used this opportunity to emphasise the potential for improvement and hypothesised that bringing the issue to the forefront of people’s minds could prompt more frequent consideration of bicarbonate levels in subsequent clinic reviews.

In parallel, we wanted to improve access to an alternative gastro-resistant preparation of sodium bicarbonate to benefit patients who had previously experienced side effects with traditional preparations. Achieving this required engagement with key stakeholders, including pharmacy teams and representatives from the integrated care board (ICB), to address prescribing pathways and cost considerations.

We then developed and implemented a simplified flowchart to support healthcare professionals when monitoring metabolic acidosis and prescribing sodium bicarbonate in advanced CKD. The tool was designed to act as both a visual prompt and a practical guide, consolidating relevant clinical information and decision-making criteria in an accessible format. The design allowed for easy updates in response to emerging evidence or changes in local policy, supporting the long-term sustainability of the intervention.

Strategy

PDSA cycle 1

We initially focused on the educational aspect of the project. The aim was to raise awareness of metabolic acidosis in CKD and to flag that our unit had a relatively low proportion of patients with serum bicarbonate levels in the optimal range at that point. We presented the findings of our baseline audit at local departmental meetings, alongside slides detailing the evidence for sodium bicarbonate in CKD and potential clinical consequences of metabolic acidosis. The meetings were well attended by members of the MDT, including consultants, resident doctors and specialist renal nurses, thereby reaching multiple professionals involved in CKD care. Several improvements were seen in the data following this intervention. All 100 patients in this cohort had a bicarbonate level checked within the preceding 6 months. 63% of patients now had serum bicarbonate levels in the optimal range, and 86% had sodium bicarbonate tablets prescribed. There was also increased documentation of associated side effects (10% of patients). Positive feedback was received regarding the informative nature of the sessions; however, we felt that the effect was unlikely to be sustained with increasing time after the sessions if there were no further cognitive prompts. We also listened to concerns about patients who had tried sodium bicarbonate tablets previously and experienced side effects, for whom the treatment options were currently limited within our centre.

PDSA cycle 2

As we were now aware of a fairly high burden of side effects among patients taking the standard preparation of oral sodium bicarbonate, we next focused on making an alternative preparation available for this group. Nephrotrans (gastro-resistant sodium bicarbonate capsules) was proposed as an alternative for patients experiencing gastrointestinal intolerance and associated poor adherence with medication. A submission to the Trust Medicines Management Committee was supported by clinical evidence demonstrating the benefits of correcting metabolic acidosis in slowing CKD progression and improving musculoskeletal outcomes.4,6 The rationale included improved tolerability, reduced pill burden and greater flexibility in dosing, with the potential to enhance patient adherence and therapeutic efficacy. Following review and approval at the trust level, the proposal was escalated to the ICB Medicines Formulary Group. Nephrotrans was subsequently accepted and added as a formulary option for patients in our local area with CKD stages 3–5 requiring oral bicarbonate supplementation but unable to tolerate conventional preparations. This change resulted in 5% of patients in our AKC clinics being prescribed Nephrotrans, of whom none reported any further side effects.

PDSA cycle 3

To address our previous concerns regarding the need for an intervention that would create sustainable long-term change, we moved on to create a printed flowchart. This was designed to support the monitoring of, and prescribing for, metabolic acidosis in advanced CKD by taking healthcare professionals through a stepwise approach when reviewing patients and their blood tests in AKC clinics (see online supplemental material for flowsheet). We sought feedback on draft versions of the flowsheet from colleagues before settling on a final design, which was printed and displayed in clinic rooms and renal unit offices. We hypothesised that this would act as a visual prompt for healthcare professionals and would keep the issue of metabolic acidosis at the forefront of people’s minds long term. It could also be used as a reference point for new members of staff, for example, rotating junior doctors or specialist nurses new to the AKC role. Unfortunately, subsequent data collection showed that this intervention had not had the desired effect. 100 patients were again sampled, with only 66% found to have sodium bicarbonate (either standard preparation or Nephrotrans) prescribed. 44% of patients had a serum bicarbonate level within the optimal range, which had decreased from earlier cohorts.

Results

Our primary outcome measure was the percentage of patients in each 100-patient cohort with serum bicarbonate levels in the optimal range. Following our first intervention, we were pleased to have achieved a significant increase in this measure from a baseline of 46% to 63%. There was also an increase in the number of patients who were prescribed sodium bicarbonate tablets in clinic from a baseline of 74% to 86%. We were aware that some variation was to be expected, given the relatively small sample size, but felt this represented a stronger focus on metabolic acidosis among the AKC team following our teaching sessions.

These results were, however, not sustained, and over subsequent Plan, Do, Study, Act (PDSA) cycles, we saw a disappointing decline in our primary outcome measure. The percentage of patients with optimal serum bicarbonate levels had dropped to 44% when we assessed our progress after the third intervention, which was slightly below the original baseline. We had not expected a significant change following our second intervention (introduction of an alternative preparation of sodium bicarbonate), as this only affected a very small number of patients. However, we had hoped for an improvement after the introduction of the metabolic acidosis flowchart, particularly as we had received positive feedback from stakeholder colleagues. Mean serum bicarbonate had also fallen from a baseline of 21.52 mmol/L to 21.18 mmol/L after the third intervention. On further analysis of the data, the number of patients who were prescribed sodium bicarbonate had reduced from 86% in the previous cycle to 66%. This suggests that a failure of implementation in terms of sodium bicarbonate prescribing was likely to be a significant contributor to the disappointing results in the final cycle.

A positive indicator was that 100% of patients had again had a serum bicarbonate measurement taken within the preceding 6 months. This suggests that although we did not meet our targets relating to achieving optimal serum bicarbonate levels, the initial improvement in monitoring had been sustained.

Data collection to reassess progress after the third intervention was delayed due to logistical issues following the rotation of resident doctors, and we noted that there was a longer duration between intervention and reassessment in this cycle. There may therefore have been a transient improvement in our primary outcome measure not captured in the data, but even so, this would suggest that any positive effects from the interventions were not sustained aside from the increase in serum bicarbonate monitoring. There will again be some variation due to sample size, but this nonetheless represented an unexpected failure.

Lessons and limitations

We began the project by setting a specific, measurable aim which was of clinical relevance to our patient cohort and had been identified as a potential area for improvement in our renal unit. We valued input from our MDT and involved staff members as important stakeholders throughout the project. Although it was not our primary outcome measure, the improvement seen in regular monitoring of serum bicarbonate levels to 100% compliance is a positive point to take forward.

We took a broad view of potential interventions which could achieve the primary outcome of the project. While the intentions of this were good, it may have been an unexpected downfall, as we could have seen more success by taking one type of intervention (for example, the flowchart) and developing this fully rather than branching out into other areas in a short time frame.

When designing and implementing the flowchart, we made the assumption that a reference aid visible in clinical areas would act as a prompt to those prescribing sodium bicarbonate and that this would translate into a greater number of patients having optimal serum bicarbonate levels. This did not materialise in our data, and we are planning to seek further feedback from colleagues on their experiences of using the flowchart to better understand why this was the case. There may have been additional contributing factors to the failure to achieve our primary outcome measure following the third cycle, for example, patient adherence with medications, which was beyond the scope of this project. However, given that the reduction in the number of patients meeting the serum bicarbonate target corresponded with a reduction in the prescription of sodium bicarbonate tablets, we feel that this would logically be the best initial area to focus on in subsequent interventions.

Our sample size of 100 patients per data collection round initially felt substantial, but with hindsight, this may not be the case, given the size of our AKC patient population. We did not perform statistical measurements to look at chance, which may have contributed both to the initial significant improvement seen in cycle one and the disappointing results in later cycles. For logistical reasons, we stopped collecting data about side effects and patient compliance in the third cycle, so it is difficult to evaluate whether these factors affected the results. Overall, the collection of a greater number of data points would have been beneficial throughout the project.

It is important to emphasise that the baseline deviation in mean bicarbonate level, while numerically modest, is not clinically trivial. Persistent mild metabolic acidosis in CKD has been linked to faster progression of kidney disease and increased risk of hyperkalaemia and bone and muscle complications, as previously discussed. Therefore, the finding that less than half of our patients were within the optimal range highlighted a significant care gap and underscored the need for focused interventions.

We have not yet carried out any directly patient-facing interventions relating to this project, which is an area we could explore further in the future. This would allow us to look at patient-specific factors, for example, concerns about adherence, which are relevant to our primary outcome measure.

Conclusion

This project used a multifaceted approach to attempt to enhance the management of metabolic acidosis in patients with advanced CKD, aligning local practice with national and international guidelines. Baseline data collection confirmed that serum bicarbonate levels were already being measured reliably in AKC clinics, but the results were not consistently acted on. This insight allowed us to focus our efforts on improving the appropriate and effective use of sodium bicarbonate.

Some successful outcomes were achieved during the course of the project. Most notably, we improved access to a gastro-resistant preparation of sodium bicarbonate for patients who had previously struggled with side effects with standard formulations—an intervention which was well received by both clinicians and patients. Although a formal cost-effectiveness analysis was not undertaken, improved tolerability and treatment adherence have the potential to delay progression to end-stage renal disease, which may yield downstream economic benefits. This represents a potential area for future evaluation, although the small number of eligible patients in our centre would present challenges in demonstrating a significant impact.

Although the project did not meet its aim of increasing the proportion of patients with serum bicarbonate levels in the optimal range, important lessons were learnt. A key challenge was the sustainability of interventions over time. The experience has highlighted the importance of incorporating detailed qualitative and quantitative data collection at multiple time points to better evaluate the effectiveness of change and to guide ongoing improvement efforts. We are committed to taking the project forward with this learning with a renewed focus on sustainability and measurement. Future work will explore more robust, data-driven strategies to embed positive change into routine practice and optimise the management of metabolic acidosis for our advanced CKD patient population.

Data availability statement

Data are available upon reasonable request.