Letter to the Editor - Letter to the editor regarding triglyceride-glucose index and acute kidney injury in diabetic ketoacidosis

Sep 30, 2026

Uploaded by Aditya

Letter to the editor responding to a retrospective cohort study evaluating whether the triglyceride-glucose (TyG) index at admission is associated with subsequent acute kidney injury (AKI) in adults with diabetic ketoacidosis (DKA). The correspondence raises methodological issues, focusing on the data-driven derivation of a TyG cut-off (~10.92), incomplete reporting of AKI definition and baseline creatinine assessment, limited confounder control, and missing AKI stage distributions. The authors urge clearer baseline creatinine ascertainment, expanded model adjustment, causal versus predictive analyses, and independent validation.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages2

Reading Time5 min

AccessFree to download

Letter to the editor regarding association between TyG index and AKI in DKA

Concerns about TyG cut-off derivation

Concerns about AKI definition, baseline creatinine, and confounding control

Requested revisions and validation

Conclusion

Author contributions

Funding and data availability

References

AnnAls of Medicine 2026, Vol. 58, no. 1, 2604878 [https://doi.org/10.1080/07853890.2025.2604878](https://doi.org/10.1080/07853890.2025.2604878) LETTER TO THE EDITOR Letter to the editor regarding: ‘association between triglycerideglucose index upon admission and the subsequent occurrence of acute kidney injury in adult patients with diabetic ketoacidosis: a single center retrospective cohort study’ To the editor We read with great interest the retrospective cohort study by Jin et al. assessing the association between the triglyceride–glucose (TyG) index at admission and acute kidney injury (AKI) in adults with diabetic ketoacidosis (DKA) [1] . The authors reported a nonlinear relationship between TyG and subsequent AKI and proposed a threshold around 10.92 as a practical cut-off for risk stratification. This is an important and clinically relevant topic, given the high burden of renal complications in DKA. However, we would like to raise several methodological concerns that may limit the robustness and generalisability of their conclusions. First, the derivation and use of the TyG cut-off at 10.92 appear to be strongly data driven. In this single cohort, TyG was divided into quartiles, with the lower bound of the highest quartile coinciding with 10.92, and restricted cubic spline analyses suggested a similar turning point. The same value was then adopted asa clinical cut-off and used as the exposure in propensity score matching. A threshold determined entirely within one dataset is prone to overfitting and optimism bias, so the observed risk gradient may be larger than would be seen in routine clinical practice. Without external validation, the ‘optimal’ cut-off is likely to shift across populations, and building algorithms or management bundles around this value risks conveying a level of validation that has not yet been achieved [2] . Second, the reporting of AKI definition, baseline renal function and confounder control seems incomplete. Because urine output criteria were unavailable, AKI was defined solely by serum creatinine, yet the manuscript does not clearly specify how baseline creatinine was obtained (previous outpatient measurement, admission value or a back-calculated estimate), even though this choice directly affects AKI incidence and KDIGO staging [3] . Moreover, the distribution of AKI stages is not presented, so it is unclear whether TyG is particularly associated with moderate-to-severe AKI. In the multivariable models, baseline kidney function, indices of DKA severity (such as pH, bicarbonate or anion gap) and key treatment-related factors appear tobe only partially accounted for, leaving room for substantial residual confounding [4] . We therefore encourage the authors to provide a clearer description of baseline creatinine ascertainment, present AKI stage distribution and extend their models to incorporate baseline renal function, DKA severity indices and relevant treatment factors. Separate models aimed at causal interpretation and at risk prediction could help clarify whether TyG should be viewed primarily as a potential causal factor or as a low-cost risk marker. Ideally, the proposed TyG threshold should also be examined in an independent cohort, with appropriate internal or external validation [5] . In conclusion, Jin et al. contribute valuable data suggesting a nonlinear association between admission TyG and AKI in adults with DKA and highlighting a possible high-risk range. With more transparent definitions, more comprehensive adjustment for confounding and independent validation of the data-derived cut-off, this line of research may provide a stronger basis for renal risk stratification and early management in this vulnerable population. Author contributions CRediT: Yi-Fan Guo: Conceptualization, Writing – original draft; Mao-Sheng Xu: Investigation, Writing – review & editing. Disclosure statement No potential conflict of interest was reported by the author(s) . 2

FAQs

What study is being discussed in this letter?

The letter discusses a retrospective cohort study assessing the association between the triglyceride-glucose (TyG) index at admission and subsequent acute kidney injury (AKI) in adults with diabetic ketoacidosis (DKA).

Why does the letter question the TyG cut-off value of 10.92?

The cut-off appears data-driven within a single dataset, which can lead to overfitting and optimism bias. Without external validation, the optimal threshold may vary across populations.

What methodological reporting concerns are raised about AKI assessment?

The letter notes incomplete AKI definition details, especially how baseline creatinine was obtained. It also highlights that AKI stage distribution is not presented and that confounder control may be only partial.

Sep 30, 2026

Uploaded by Aditya