Risk of cisplatin-associated acute kidney injury (CP-AKI)

Clinical background

Cisplatin is one of the most widely used cytotoxic agents in oncology, with indications that include urothelial cancer, non-small cell lung cancer, squamous cell carcinoma of the head and neck, mesothelioma and germ cell tumours. Nephrotoxicity is the principal dose-limiting adverse effect and can occur after the very first dose. Severe cisplatin-associated acute kidney injury (CP-AKI) increases the risk of extrarenal toxicity, may make continued cisplatin treatment impossible and is strongly associated with shortened survival [1].

Earlier risk models for CP-AKI have been limited by small cohorts, permissive AKI definitions (for example a creatinine rise of 0.3 mg/dL) and a lack of external validation [1]. The clinical need has been for a tool that identifies patients at high risk of severe AKI (KDIGO stage 2 or 3) before the first dose, so that monitoring and hydration can be individualised.

Calculating the CP-AKI risk score

The score is the sum of nine variables, all available from routine laboratory tests and the patient's history before the first dose of cisplatin:

$$\text{CP-AKI risk} = P_{\text{age}} + P_{\text{hypertension}} + P_{\text{diabetes}} + P_{\text{smoking}} + P_{\text{dose}} + P_{\text{Hb}} + P_{\text{WBC}} + P_{\text{albumin}} + P_{\text{Mg}}$$

where each $P$ is the score for the respective variable according to the following categories:

The total score ranges from 0 to 22.5. The point allocation was derived by dividing the odds ratios from a multivariable logistic regression model by the smallest odds ratio in the model, after which the continuous variables were categorised using clinically relevant cut-offs [1].

The derivation cohort consisted of 11,766 adult patients who received their first intravenous dose of cisplatin at Memorial Sloan Kettering Cancer Center between 2006 and 2022. The median age was 59 years (IQR 50–67). The outcome, CP-AKI, was defined as a doubling of plasma creatinine or the need for dialysis within 14 days of the first dose, corresponding to KDIGO stage 2 or 3. The incidence was 5.2% in the derivation cohort [1].

Interpretation in practice

The total score divides patients into four risk bands. The table states what each band means in terms of management.

In the validation cohort, patients in the highest risk category had a 17.87-fold (95% CI 10.56–29.60) higher odds of CP-AKI than those in the lowest category [1]. In the derivation cohort the corresponding figure was 24.00 (95% CI 13.49–42.78) [1].

The score should be interpreted as support for decisions on the intensity of monitoring and the hydration strategy, not as an automatic decision to give or withhold cisplatin. The choice between cisplatin and an alternative cytotoxic agent depends primarily on tumour type and treatment intent.

Validation and performance

The external validation cohort consisted of 12,951 patients from five other large American academic cancer centres (Massachusetts General Hospital, Dana-Farber Cancer Institute, MD Anderson Cancer Center, University of Colorado and Northwell Health), collected over the same period. The median age was 60 years (IQR 50–67). The incidence of CP-AKI was lower in the validation cohort, 3.3%, probably reflecting differences in patient population and hydration practice between centres [1].

The primary multivariable model (which included serum creatinine and platelets in addition to the nine variables of the simple score) had a C-statistic of 0.75 [1]. The simple risk score showed a monotonic increase in CP-AKI risk across the categories in both the derivation and the validation cohort, suggesting good calibration of the risk groups [1].

The model was compared with three previously published models (Bhat et al., de Jongh et al. and Motwani et al.), which had C-statistics between 0.60 and 0.68. The new model was superior to all of them (DeLong P < 0.001 for each comparison) [1]. An important explanation is that the earlier models used considerably more permissive AKI definitions. Motwani et al., for example, defined AKI as a creatinine rise of 0.3 mg/dL, which also captures mild and clinically less relevant kidney injury [2]. Bhat et al. studied 233 patients with squamous cell carcinoma of the head and neck and identified African American ethnicity as a risk factor, a variable not included in the present score [3].

Decision curve analysis showed that both the primary model and the simple risk score gave a net clinical benefit across a wide range of threshold values compared with treating all or none [1].

Limitations

The score applies to the first dose of intravenous cisplatin. The risk with repeated cycles may differ, and the tool has not been validated for later cycles.

All six centres in the study were large American academic cancer centres. The population may differ from patients treated in other care settings, particularly outside the USA. The performance of the model in non-American populations has not been studied.

The score does not include serum creatinine or estimated GFR, even though these were significant predictors in the primary model. This was a deliberate choice to keep the tool simple and clinically applicable, but it means that patients with impaired renal function at baseline may receive a falsely low score if their kidney injury is not captured by the other variables [1].

The cisplatin dose is given in absolute milligrams, not per square metre of body surface area. This may mean that a low dose in a patient with a large body surface area gives a lower score than is warranted, and vice versa.

The tool is not designed for paediatric patients. Patients with end-stage renal failure were excluded from the derivation cohort.

The severity of CP-AKI is strongly linked to survival: the adjusted hazard ratio for 90-day survival was 4.63 (95% CI 3.56–6.02) for KDIGO stage 3 CP-AKI compared with no CP-AKI [1]. This underlines the importance of identifying high-risk patients, but the score itself predicts kidney injury only, not mortality.

Place in current practice

No national guideline specifically recommends use of the CP-AKI risk score. Practice for hydration during cisplatin treatment varies between centres, but hydration with isotonic saline is recommended throughout, and magnesium supplementation is given routinely at many clinics. Two systematic reviews support short hydration (2–4 L over 4–5 hours) rather than prolonged hydration, with the addition of mannitol at higher doses and magnesium supplementation irrespective of risk level [4, 5]. The CP-AKI risk score can be used to identify patients who may need intensified hydration or closer monitoring beyond the standard regimen, but it does not replace local hydration practice.

References

1. Gupta S, Glezerman IG, Hirsch JS et al. Derivation and external validation of a simple risk score for predicting severe acute kidney injury after intravenous cisplatin: cohort study. BMJ 2024;384:e077169. PMID: 38538012
2. Motwani SS, McMahon GM, Humphreys BD et al. Development and Validation of a Risk Prediction Model for Acute Kidney Injury After the First Course of Cisplatin. J Clin Oncol 2018;36(7):682–688. PMID: 29320311
3. Bhat ZY, Cadnapaphornchai P, Ginsburg K et al. Understanding the Risk Factors and Long-Term Consequences of Cisplatin-Associated Acute Kidney Injury: An Observational Cohort Study. PLoS One 2015;10(11):e0142225. PMID: 26556481
4. Crona DJ, Faso A, Nishijima TF et al. A Systematic Review of Strategies to Prevent Cisplatin-Induced Nephrotoxicity. Oncologist 2017;22(5):609–619. PMID: 28438887
5. Sikking C, Niggebrugge-Mentink KL, van der Sman ASE et al. Hydration Methods for Cisplatin Containing Chemotherapy: A Systematic Review. Oncologist 2024;29(4):e173–e186. PMID: 37995306