Observational study finds eGFR below 25th percentile linked to higher dialysis risk — Evidence Review

Small changes in kidney function—even within the normal range—can signal higher risk for chronic kidney disease progression, according to a large Swedish study. Related research broadly supports the importance of early detection and individualized risk assessment for kidney disease, with most studies agreeing that earlier identification may improve patient management (1, 4, 5, 14).

• The new study aligns with prior findings that leveraging population-based data and novel risk tools can enhance early recognition of chronic kidney disease risk, especially when traditional single cutoff values may miss at-risk individuals (4, 14).
• Related research highlights the promise of machine learning and clinical prediction models in identifying early-stage chronic kidney disease, suggesting that integrating such tools into clinical practice could refine risk stratification and preventive care (1, 2, 3).
• While novel biomarkers and screening approaches are being explored, consensus remains that estimated glomerular filtration rate (eGFR) and albuminuria are practical and widely used markers for early detection, though gaps in screening and follow-up testing persist (5, 11, 12, 14).

Study Overview and Key Findings

Chronic kidney disease (CKD) is a growing global health concern, often diagnosed only after significant kidney damage has occurred. The new study from Karolinska Institutet addresses this gap by proposing a shift from traditional, fixed eGFR cut-off values to age- and sex-adjusted percentiles for more nuanced risk assessment. By analyzing over one million adults in Stockholm, the research highlights how even "normal" kidney function values can conceal elevated future risk, and introduces a freely available web-based calculator to support early intervention.

Property	Value
Organization	Karolinska Institutet
Journal Name	Kidney International
Authors	Yuanhang Yang, Antoine Creon, Juan Jesús Carrero
Population	Adults in the Stockholm region of Sweden
Sample Size	1.1 million adults
Methods	Observational Study
Outcome	Kidney function levels, risk of kidney failure, mortality
Results	eGFR below 25th percentile linked to higher risk of dialysis

Literature Review: Related Studies

To place these findings in context, we searched the Consensus paper database, which includes over 200 million research papers. The following search queries were used to identify relevant studies:

1. kidney disease early detection methods
2. eGFR dialysis risk association
3. chronic kidney disease screening guidelines

Topic	Key Findings
How effective are early detection methods for chronic kidney disease?	- Machine learning algorithms, including random forest and deep neural networks, can accurately detect early-stage CKD and outperform traditional methods (1, 2, 3). - Novel biomarkers and point-of-care screening strategies show promise for early detection, but current clinical practice still relies primarily on eGFR and albuminuria (4, 5).
What is the association between eGFR at dialysis initiation and patient outcomes?	- Higher eGFR at dialysis initiation is paradoxically associated with higher mortality, possibly due to confounding factors such as frailty and comorbidity (6, 7, 8, 9). - Both low and high eGFR values at dialysis start are linked to increased risk, supporting a U-shaped relationship between eGFR and mortality (8, 9).
How should chronic kidney disease screening be implemented in the general population?	- Targeted screening based on risk factors (age, hypertension, diabetes) is effective, but universal screening in asymptomatic adults lacks strong evidence for benefit (11, 12, 13, 15). - International guidelines and expert consensus recommend prioritizing high-risk individuals for screening and early intervention, especially in primary or community care settings (12, 13, 14).
Are there gaps in follow-up testing and early intervention for at-risk individuals?	- Many patients with early signs of CKD do not receive adequate follow-up testing, such as urinary albumin measurement, leading to missed opportunities for intervention (11, 14). - Socioeconomic disparities and lack of standardized protocols contribute to variation in CKD care and outcomes (12, 14).

How effective are early detection methods for chronic kidney disease?

Multiple studies demonstrate that advanced data-driven approaches, such as machine learning algorithms and deep learning models, can improve the early detection of CKD compared to conventional clinical assessment. However, while emerging biomarkers are being investigated, routine practice still depends heavily on eGFR and albuminuria as initial screening tools (1, 2, 3, 5). The new Karolinska study supports this trend by introducing population-based eGFR percentiles to refine early risk identification.

• Machine learning models, such as random forest and XGBoost, achieve high accuracy in CKD prediction and can assist clinicians in early diagnosis (1, 2).
• Deep neural networks may outperform other machine learning approaches in identifying early CKD, suggesting the potential for integrating AI in risk prediction (3).
• Novel biomarkers (e.g., KIM-1, NGAL) are under investigation, but none have yet replaced traditional markers like eGFR and albuminuria for early detection (5).
• Point-of-care and population-based screening strategies, especially in resource-limited settings, can help identify high-risk individuals who warrant further testing (4).

What is the association between eGFR at dialysis initiation and patient outcomes?

Several large observational studies have reported a counterintuitive association: initiating dialysis at a higher eGFR is linked to higher mortality. This paradox is thought to be influenced by patient factors such as frailty and comorbidity, which often prompt earlier dialysis initiation. The new study's finding of a U-shaped relationship between eGFR percentiles and mortality risk aligns with these prior observations and underscores the need for nuanced risk stratification (6, 7, 8, 9).

• Higher eGFR at dialysis initiation is often associated with increased mortality, likely due in part to confounding by frailty and illness severity (6, 7).
• Both very low and very high eGFR values at dialysis start are linked to worse outcomes, supporting a U-shaped risk curve (8, 9).
• Patient age, comorbidities, and clinical presentation strongly influence the timing of dialysis initiation and subsequent outcomes (7, 8, 9).
• These findings highlight the complexity of interpreting eGFR in isolation and the value of age- and sex-adjusted norms, as proposed in the new study.

How should chronic kidney disease screening be implemented in the general population?

There is ongoing debate about the best approach to CKD screening. Most guidelines recommend targeted screening for high-risk groups rather than universal population screening, citing a lack of evidence for benefit in asymptomatic individuals without risk factors. The new study's approach—using individualized risk charts and web-based tools—may help clinicians better identify at-risk patients who fall outside traditional risk categories (11, 12, 13, 14, 15).

• Screening strategies that focus on individuals with hypertension, diabetes, or older age are efficient and capture most cases of CKD (11, 12).
• Universal screening in asymptomatic adults is not widely supported due to insufficient evidence of benefit and potential harms (13, 15).
• Recent consensus statements advocate for early identification and intervention in high-risk groups, especially in primary care and community settings (14).
• The new study's population-based eGFR percentiles may enable more personalized risk stratification within these targeted screening frameworks.

Are there gaps in follow-up testing and early intervention for at-risk individuals?

Despite advances in detection methods and risk stratification, gaps remain in the follow-up and management of patients with early signs of kidney dysfunction. Many individuals with abnormal eGFR or other risk factors do not receive recommended confirmatory tests or timely intervention, potentially leading to missed opportunities for prevention (11, 12, 14).

• Only a minority of patients with reduced eGFR receive follow-up urinary albumin testing, a key step in diagnosing early kidney damage (11).
• Socioeconomic disparities and lack of standardized care pathways contribute to variation in CKD outcomes (12, 14).
• Early intervention, including lifestyle modification and management of comorbidities, may be underutilized in at-risk populations (14).
• The new study's web-based tool could help address some of these gaps by providing clinicians with accessible, individualized risk information to prompt earlier action.

Future Research Questions

While the new study advances early risk identification in CKD, several important questions remain. Future research is needed to evaluate the clinical impact of implementing population-based eGFR percentiles, optimize integration with digital decision tools, and address persisting disparities in CKD diagnosis and management.

Research Question	Relevance
Does use of age- and sex-adjusted eGFR percentiles improve clinical outcomes in CKD patients?	Assessing whether individualized eGFR risk charts lead to earlier intervention and better outcomes will help determine their value in clinical practice (14).
How can machine learning tools be integrated into routine CKD screening and risk assessment?	Understanding the practical integration of AI-driven risk models may enhance early detection and support clinician decision-making (1, 2, 3).
What are the barriers to follow-up testing and early intervention in patients with abnormal eGFR?	Identifying and addressing gaps in confirmatory testing and management could reduce missed opportunities for prevention (11, 14).
Are novel biomarkers more effective than eGFR for early detection of CKD?	Research into new biomarkers may eventually provide more sensitive or specific tools for early CKD detection, given current limitations of eGFR (5).
How do socioeconomic factors influence CKD screening, diagnosis, and outcomes?	Addressing disparities in CKD care is critical for improving equity and ensuring that advances in risk prediction benefit all patient populations (12, 14).