Impact of GFR on Mortality Risk After Biliopancreatic Diversion: Challenges and Pitfalls for the Clinician

Obesity is a major global health issue, significantly contributing to increased cardiovascular (CV) [1] and overall mortality risk [2]. The prevalence of obesity has reached epidemic proportions, leading to a rise in type 2 diabetes (T2D), hypertension, and chronic kidney disease (CKD), which together elevate the risk of both CV disease and mortality [3]. The impact on kidney function is particularly concerning, as obesity accelerates CKD progression and can lead to end-stage renal disease (ESRD), further compounding health risks [4]. In a global database including over 5.4 million healthy individuals, the risk of decline in GFR progressively and independently increased with rising of the BMI [5].

Bariatric surgery has emerged as an effective intervention for severe obesity, offering substantial and sustained weight loss, which translates into a wide range of health benefits [6, 7]. In addition to improving metabolic conditions, bariatric surgery has shown positive effects on kidney health, reducing proteinuria, and potentially slowing CKD progression and even reversing early renal damage [7, 8]. Procedures such as biliopancreatic diversion (BPD) and gastric bypass not only improve body mass index (BMI) and reduce obesity-related complications but also address hyperfiltration and other renal stressors associated with obesity. By reducing weight and improving metabolic profiles, bariatric surgery lowers overall and cardiovascular mortality risks, enhancing quality of life and long-term outcomes. At the same time, patients undergoing bariatric surgery may develop a risk of malnutrition, particularly those with CKD [9, 10]. In bariatric surgery, an accurate assessment of kidney health plays a crucial role in evaluating surgical risk, guiding postoperative recovery and determining long-term renal outcomes [11]. Accurately assessing GFR in patients with severe obesity, especially in the context of bariatric surgery, presents several challenges [12]. First, widely used equations for estimating GFR, like the CKD Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease (MDRD) equations, were developed based on older datasets from populations in which severe obesity was far less prevalent [13]. Additionally, large weight loss itself influences creatinine levels, regardless of actual kidney function, complicating the accuracy of these measurements [14].

A key issue in assessing GFR after bariatric surgery is the indexing of GFR to body surface area (BSA), as BSA changes significantly with major weight loss. Typically, both measured and estimated GFR (eGFR) are adjusted to a standard BSA of 1.73 m², based on the assumption that GFR is linked to kidney mass, which correlates with body mass in mammals [15]. This adjustment reduces variability in kidney function measurements among healthy people. However, in individuals with severe obesity, BSA indexing results in notably lower GFR values, raising questions about its appropriateness for post-bariatric surgery patients [16, 17].

In this study, the primary objective was to examine the performance of indexed and non-indexed eGFR using creatinine levels, comparing the different available formulas. The second aim was to assess the impact of GFR changes during the follow-up study period on the outcomes of patients who underwent biliopancreatic diversion.

Our cohort had a mean follow-up of 16.2 ± 9.1 years. The overall mean age was 36.1 ± 10.9 years, with 73.5% of the patients being women. Regarding comorbidities, 60% had a diagnosis of hypertension, and 43% had diabetes. The mean pre-surgery BMI, creatinine, and eGFR (estimated by unindexed CKD-EPI formula) were 47.0 ± 9.3, 0.87 ± 0.21 mg/dL, and 75.3 ± 15.9 mL/min/m², respectively.

No substantial differences were observed in the clinical and laboratory characteristics of the study population divided into baseline GFR tertiles (Table 1). Along with the growth in tertiles, patients were shown to be younger and to have a less severe grade of obesity. Hypertension was more prevalent in the first tertile, whereas diabetes was more frequent in the second tertile. A higher BMI (49.5 ± 10.0) was associated with a lower average eGFR (p = 0.0077).

We analyzed estimated GFR using different equations available over the three decades of follow-up. We identified five follow-up points: the first year, 3rd, 7th, 15th, and 25th years (Fig. 1). Supplemental Table 1 specifies the formulas used and the methods for unindexing for standard BSA and indexing for real BSA (Supplemental Table 1).

Bild vergrößern

Patients with a ΔeGFR greater than 20% after the first year were those with lower eGFR at baseline who showed a mean higher eGFR one year after the intervention and who benefited from a greater reduction in BMI in the first year and as annual ΔBMI (p < 0.001 and = 0.0744, respectively), as shown in Table 2.

Patients who experienced an increase in eGFR greater than 20% during the first year after BPD had a better outcome, independently of basal BMI (Log Rank test p = 0.0045), as shown in Fig. 2.

Bild vergrößern

Considering the entire follow-up period, the mean BMI showed an annual change of −3.7 ± 5% per year, while the unadjusted eGFR demonstrated an annual increase of + 3.1 ± 7.8 mL/min/m². Patients with a lower ΔeGFR had lower preoperative weight, BMI, and BSA. Comorbidities did not significantly affect kidney function variation (Supplemental Table 2).

Kaplan–Meier survival curves based on ΔeGFR tertiles (indexed CKD-EPI) and adjusted for pre-surgery BMI show that patients in the second tertile (ΔeGFR between + 0.5 and + 2.2 ml/min/year) have a significantly higher survival rate (log-rank p = 0.034) compared to those in the other tertiles (Fig. 3).

Bild vergrößern

Both univariate and multivariate analyses indicate that baseline eGFR, BMI, and serum albumin do not have a significant impact on mortality. However,an elevation in eGFR over the very long term appears to be associated with significantly increased mortality (Table 3).

Indeed, the improvement in GFR during the first year after BPD appears to serve as a protective factor against mortality (HR 0.96 [95%CI 0.93–0.98], p = 0.002). In contrast, a higher annual ΔGFR beyond the first year of follow-up is associated with an increased risk of mortality (HR 1.31 [95%CI 1.15–1.48], p < 0.001). Furthermore, the annual decrease in BMI following surgery is linked to an independent increase in mortality risk (HR 2.48 [95%CI 1.01–6.07], p = 0.047).

This retrospective observational study highlights the intricate and time-dependent relationship between changes in eGFR and mortality risk following BPD. Notably, the timing of eGFR assessment emerged as a critical factor in understanding its prognostic implications. The lack of an early increase in eGFR and a long-term increase in eGFR are two unfavorable risk factors. While the former may indicate suboptimal outcomes of bariatric surgery, the latter may indicate malnutrition resulting from excessive muscle mass loss.

For people with very high BMI, BPD with duodenal switch resulted in greater weight loss than other procedures, resulting in greater improvement in weight loss outcomes and weight-associated comorbidities compared with non-surgical interventions [23, 24]. While sustained weight reduction after bariatric surgery leads to improvement in blood pressure and glycemic control, resulting in reduction in cardiovascular risk [25], the mean permanent reduction of about 75% of the initial excess weight following BPD is frequently associated with late complications including anemia, bone demineralization, and protein malnutrition [26]. Moreover, BPD is known to be associated with a decrease in serum creatinine and an early increase in eGFR, the prognostic effects in the long term remain to be investigated [27, 28].

Accurately evaluating GFR in patients with severe obesity, particularly following bariatric surgery, presents several challenges. First, commonly used equations for estimating GFR, such as the CKD-EPI and MDRD equations, were developed using older datasets from populations where severe obesity was far less prevalent. Second, both obesity and cachexia independently influence creatinine, regardless of GFR. Creatinine levels are directly linked to muscle mass, which decreases by 20–25% after BPD, opening the possibility for the development of malnutrition. Despite the advent of new specific formulas (i.e. EKCF, Salazar Corcoran) and attempts at indexing for real BSA [29, 30], the determination of GFR in these patients remains inconsistent, and no gold standard has been established yet. In this study, we found significant differences between GFR values calculated using different formulas and standard indexing compared to indexing with actual BSA. However, the trends remain similar in both the short- and long-term follow-up, providing consistency to the results (Fig. 1).

As previously reported in the literature, we found that the better GFR, the lower the prevalence of comorbidities, such as hypertension, diabetes, and more severe obesity (Table 1). In our cohort, renal hyperfiltration indicated by pathologically elevated eGFR values seems to be underrepresented, suggesting that severe obesity has already had time to cause long-term glomerular damage.

One of the key findings of our study is the dual impact of GFR changes on mortality risk.

We found that the lack of improvement in GFR (< + 20% compared to baseline) during the first year is a red flag for mortality risk in patients who underwent BPD (Fig. 2, Log Rank test p = 0.003). Several mechanisms, including weight loss, could account for the positive impact of surgery on renal function. While pre-surgical BMI was comparable between patients who experience improved kidney function and those who do not in the first year post-surgery, it was observed that patients with more than a 20% improvement in GFR lost a greater amount of body mass (−34% vs 30% of BMI in the first year, data not shown). Therefore, we can hypothesize that, in an early phase, greater body mass loss translates into benefits that include improved kidney function, consistent with previous studies [27, 31].

Interestingly, the results regarding eGFR changes during long-term diverged dramatically. Dividing the cohort into tertiles based on the annual change in eGFR, a greater increase in eGFR (> 2.2 mL/min/m²/year, 3° tertile) in the long term translates into an increased risk of mortality (Fig. 3). It is known that BPD requires a close follow-up and multiple vitamin and mineral supplements, especially regarding fat-soluble vitamins [32]. In some cases, nutritional consequences that may occur in bariatric patients can potentially undermine the clinical benefits of BPD [33]. In this context, an increase in GFR may not accurately represent true renal function; instead, it may be a result of decreased serum creatinine levels, which are directly associated with progressive muscle mass loss due to protein malnutrition. This paradoxical finding underscores the necessity of distinguishing true renal function improvement from potential artifacts by malnutrition-induced creatinine reductions. Our study has several limitations, alongside strengths that merit consideration. First, the retrospective observational design precludes causal inference, and only associations can be established. Second, the absence of directly measured GFR or alternative biomarkers such as cystatin C limits the precision of renal function assessment. Nevertheless, creatinine-based eGFR remains the most widely available and routinely used method in clinical practice, ensuring that our findings reflect real-world patient evaluation and follow-up, thereby maintaining high practical relevance.

Third, we acknowledge that changes in patient characteristics over time—including body weight, nutritional status, comorbidities, and medication use—may act as dynamic confounders affecting renal outcomes and survival. Due to the historical and retrospective nature of the cohort, with irregular follow-up intervals spanning several decades, fully addressing these time-varying confounders through advanced statistical modeling (e.g., marginal structural models or joint longitudinal-survival models) was not feasible. Consequently, our study did not implement a formal time-varying analysis, and proper adjustment for dynamic confounders was limited. Our analyses therefore relied primarily on baseline covariates and pre-specified eGFR thresholds, which may not capture the full impact of longitudinal changes.

Finally, the monocentric nature and relatively homogeneous population of the study limit generalizability to broader or more diverse patient groups.

Despite these limitations, the relatively large sample size, the exceptionally long follow-up period, and the inclusion of unindexed GFR calculated with the CKD-EPI formula make this study a unique and valuable contribution, potentially representing one of the first explorations of the long-term relationship between bariatric surgery and kidney function. Furthermore, BMI-adjusted data were analyzed to assess mortality risk in Kaplan-Meier analyses. A number of potentially confounding variables were considered, including age, gender, BMI, and albumin levels, strengthening the independent association between GFR changes and mortality. Therefore, the findings described herein provide critical insights into the nuanced relationship between bariatric surgery, renal function, and long-term mortality, paving the way for future research and improvement in clinical management.

Changes in eGFR have been identified as an independent predictor of mortality, with opposing effects depending on the timing of the measurement after surgery (Table 3).

In the short-term follow-up, GFR improvement within 1 year post BPD was associated with a reduced risk of mortality. However, in the long-term follow-up, for each annual ml/min improvement in eGFR we observed a 30% increased risk of mortality (Table 3).

The advent of GLP-1 receptor agonists (GLP-1 RAs) raises new questions about the future of bariatric surgery. These medications offer promising alternatives for sustained weight loss, potentially altering the risk-benefit profile of procedures like BPD. Moving forward, careful patient selection and close monitoring of risk factors—such as excessive weight loss or the absence of early eGFR improvement—will be essential in optimizing outcomes.

This study demonstrates that changes in eGFR following BPD have distinct prognostic implications depending on the timing of assessment. The lack of early eGFR improvement predicts higher mortality risk, while long-term eGFR increases may reflect malnutrition and muscle mass loss. These findings highlight the importance of personalized postoperative monitoring and nutritional management to optimize patient outcomes. Given the advent of new weight-loss therapies, future research should compare their long-term renal effects with surgical interventions. Accurate GFR assessment methods tailored to severe obesity and bariatric surgery remain a priority for advancing clinical care.