Evaluation of Prediabetes in Cisplatin-induced Nephrotoxicity in the Short Hydration Method: A Subgroup Analysis

Abstract

Background/Aim

Cisplatin-induced nephrotoxicity (CIN) is one of the most attention-requiring adverse effects. We have reported that diabetes mellitus significantly increases the incidence of CIN in a short hydration method in real-world lung cancer treatment. However, the effect of prediabetes on CIN development remains unclear. This study investigated whether patients with prediabetes exhibit CIN at a greater rate during real-world cisplatin-including treatments as a subgroup analysis.

Patients and Methods

This retrospective observational study enrolled patients with lung cancer receiving cisplatin treatment (≥75 mg/m²) from May 2014 to January 2021 (n=169). Patients were divided into a prediabetes group (baseline HbA1c 5.7-6.4%) and a control group (baseline HbA1c <5.7%). The primary endpoint of this study was the incidence of CIN in all treatment cycles between the two groups. We also assessed variations in serum creatinine (SCr) levels and creatinine clearance (CCr).

Results

CIN occurred in 4.7% of controls and 8.3% of patients with prediabetes in all cycles, with no significant difference (p=0.37). In contrast, variation of SCr levels and CCr was significantly worse in the prediabetes group [median variation level (range) 0.11 mg/dl (–0.11-0.46 mg/dl) and 0.12 mg/dl (–0.02-1.08 mg/d) in controls and prediabetes, p=0.04 for SCr; –12.9 ml/min (–54.1-4.9 ml/min) and –16.3 ml/min (–49.4-3.0 ml/min), p=0.02 for CCr, respectively]. These results were also confirmed during the first cycle of treatment.

Conclusion

Patients with prediabetes did not develop problematic CIN, although they exhibited significant increases in SCr and decreases in CCr.

Cisplatin is a cytotoxic chemotherapeutic agent used in treating lung, ovarian, head and neck, urological, and esophageal malignancies (1-3). However, it has potent adverse effects, including gastrointestinal symptoms, neurotoxicity, ototoxicity, and nephrotoxicity (3). Cisplatin-induced nephrotoxicity (CIN) is one of the most important adverse effects of cisplatin treatment and affects ongoing and future treatment strategies. Because cisplatin is administered in a variety of treatment settings, including metastasis, perioperative treatment, and chemoradiotherapy, its management is important for effective and less onerous treatment. CIN is dose-dependent, cumulative, usually reversible (2,3), and is the dose-limiting toxicity of cisplatin (3–5). CIN used to occur in 30-40% of patients (3). However, supportive care, including magnesium and appropriate diuretic administration with quality antiemetic treatment, has decreased its incidence to 0-10% (6).

Many studies have evaluated the factors associated with CIN incidence, although cisplatin administration methods vary (3,6-17). We previously reported that diabetes mellitus (DM) significantly increases CIN incidence in a short hydration method, the most advanced recent cisplatin administration method, in real-world lung cancer treatment (6). The rationale remains unclear; we hypothesized that kidney aging and/or decreased autophagy at the proximal tubule, where CIN appears the most (4,5), induced by type 2 DM, promotes CIN degradation, as previously reported (6,18-24). However, we have not assessed whether CIN is affected by prediabetes, defined by glycemic variables greater than normal but less than DM thresholds, and a high-risk population for DM development (6,25). Prediabetes is increasing worldwide, similar to DM (26). In addition, cancer treatments, including corticosteroid use, can elevate glycemia and induce patients with prediabetes to DM. Consequently, determining whether patients with prediabetes exhibit CIN at a greater rate than those without prediabetic symptoms is important.

This study aimed to assess the effect of prediabetes on the development of CIN through a real-world subgroup analysis of cisplatin-including lung cancer treatment.

Patients and Methods

Patients. This retrospective observational subgroup study enrolled patients with lung cancer who received cisplatin treatment (≥75 mg/m²) between May 2014 and January 2021. The evaluated cisplatin-including regimens included cisplatin (80 mg/m², day 1)+ vinorelbine (20-25 mg/m², days 1, 8)±radiation, cisplatin (80 mg/m², day 1)+etoposide (100 mg/m², days 1-3)±radiation, and cisplatin (75 mg/m², day 1)+pemetrexed (500 mg/m², day 1)±bevacizumab (15 mg/kg, day 1). The study’s subgroup of patients was selected from our previous study (6). All patients met the following baseline criteria: 1) age ≥20 years; 2) Eastern Cooperative Oncology Group performance status (ECOG-PS) of 0 to 2; 3) HbA1c of <6.5%; 4) detailed patient information available from medical records; 5) and sufficient renal or liver function for cisplatin-including treatment induction. The study excluded patients previously administered cisplatin, co-administered immune checkpoint inhibitors, administered DM pharmacotherapy at baseline, could not complete the first cycle, transferred to another hospital during chemotherapy, or received dose reduction from treatment initiation. The patients were divided into the prediabetes group, which included patients whose baseline HbA1c levels were 5.7-6.4% (25) between May 2014 and January 2021, and the control group, whose baseline HbA1c levels were <5.7% between June 2014 and January 2021.

Based on our previous report, we hypothesized that the incidence of CIN would be 3-5% in the control group and 20% in the prediabetes group (6). To achieve 80% power with an alpha error of 5%, the required sample size was calculated as 65-88 subjects in each group. Ultimately, the study analyzed 85 and 84 participants in the control and prediabetes groups, respectively.

This study was approved by the Ethical Review Board for Life Science and Medical Research of Hokkaido University Hospital (approval number: 023-0196) and was conducted in accordance with the Declaration of Helsinki and the STROBE statement. Because this study was retrospective, the Ethical Review Board for Life Sciences and Medical Research of Hokkaido University Hospital waived the requirement for written informed consent from the participants.

Administration methods. All participants received a short hydration method for cisplatin administration, including 8 mEq of magnesium sulfate premedication, as described previously (27). Furthermore, all treatment regimens included the same prophylactic antiemetic therapy in accordance with the current national guidelines (28): 0.75 mg of palonosetron infusion on day 1, oral aprepitant 125 mg on day 1 and 80 mg on days 2 and 3, and intravenous dexamethasone 9.9 mg on day 1 and oral 8 mg on days 2-4, respectively. Additional medication for CIN and/or chemotherapy-induced nausea and vomiting (CINV) was administered depending on the patient’s condition, according to the physician’s decision.

Evaluation of CIN and other adverse effects. The required information was obtained from the patients’ medical records. Toxicities in all treatment cycles were assessed by physicians or pharmacists according to the Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. Pharmacists confirmed appropriate cisplatin administration, including oral hydration, for each patient. Renal function was evaluated using variations in serum creatinine (SCr) levels measured using an enzymatic method and creatinine clearance (CCr) calculated using the Cockcroft-Gault formula. CIN in the present study was defined as a CTCAE grade 2 or greater SCr elevation, as in our previous reports (3,6,27). The primary endpoint of our study was the incidence of CIN during all treatment cycles between the two groups. Secondary endpoints were CIN incidence in the first cycle, SCr and CCr variations, and cisplatin-related adverse effects between the groups.

Statistical analysis. Differences in baseline clinical backgrounds between the control and prediabetic groups were evaluated using Fisher’s exact probability test for categorical outcomes and the Mann-Whitney U-test for continuous parameters. CIN incidence was compared using Fisher’s exact probability test, and variations in SCr levels and CCr between groups were assessed using the Mann-Whitney U-test. The number of treatment cycles was compared using the Mann-Whitney U-test. The adverse effects other than CIN were evaluated using Fisher’s exact probability test for incidence and the Mann-Whitney U-test for severity. All analyses were conducted using the JMP version 16.1 statistical software (SAS Institute Japan; Tokyo, Japan). p-Values <0.05 were considered statistically significant.

Results

Patient characteristics. This study enrolled 169 patients (Figure 1). Baseline patient characteristics are shown in Table I. No differences were found between the two groups for sex, age, ECOG-PS, body surface area, staging, adjuvant chemotherapy setting, radiation combination treatment, treatment line, serum albumin levels, hemoglobin levels, SCr, CCr, estimated glomerular filtration rate (eGFR), serum sodium and potassium levels, co-administration of non-steroidal anti-inflammatory drugs (NSAIDs) and proton pump inhibitors, or complications of cardiovascular diseases. The median HbA1c level in the prediabetes group was 5.9% (range; 5.7–6.4%), which was significantly greater than that in the controls (p<0.001). None of the patients received DM medication during treatment. The total number of treatment cycles in the control and prediabetes groups were 8.2% and 13.1% for one cycle, 18.8% and 16.7% for two cycles, 14.1% and 7.1% for three cycles, 56.5% and 60.7% for four cycles, and 2.4% for both groups for six cycles, respectively, with no significant differences (p=0.92).

Figure 1. Design of this study. DM: Diabetes mellitus.

Table I. Patient characteristics.

**p<0.01. ECOG: Eastern Cooperative Oncology Group; BSA: body surface area; CCr: creatinine clearance; eGFR: estimated glomerular filtration rate; NSAIDs: non-steroidal anti-inflammatory drugs; PPIs: proton pump inhibitors.

Comparison of CIN incidence and variations in SCr and CCr levels. CIN occurred in 4.7% of controls and 8.3% of patients with prediabetes in all treatment cycles, with no significant difference (p=0.37, Figure 2A). In contrast, variation of SCr levels and CCr (worst levels – baseline levels) in all treatment cycles was significantly worse in the prediabetes group [median variation level (range) 0.11 mg/dl (–0.11-0.46 mg/dl) and 0.12 mg/dl (–0.02-1.08 mg/dl) in controls and prediabetes for SCr, p=0.04; –12.9 ml/min (–54.1-4.9 ml/min) and -16.3 ml/min (–49.4-3.0 ml/min) for CCr, p=0.02, respectively, Figure 2A and B]. We also evaluated the incidence of CIN and variations in SCr and CCr in the first cycle since most CIN incidences were confirmed in the first cycle in our previous study (3), resulting in the same results (Figure 2C and D). The incidence of CIN in the first cycle in the prediabetic group tended to be greater, but the difference was not statistically significant.

Figure 2. Comparison of (A) cisplatin-induced nephrotoxicity (CIN) incidence and serum creatinine (SCr) variation, (B) creatinine clearance (CCr) variation in all treatment cycles, (C) CIN incidence and SCr variation, and (D) CCr variation in the first cycle.

Incidence and severity of CINV. CINV was evaluated because it can affect oral hydration, which is important in the short hydration method (Table II). The incidence and severity of the adverse gastrointestinal effects did not differ between the two groups.

Table II. Incidence and severity of nausea, vomiting, and anorexia in all subsequent cycles.

Discussion

We previously reported that patients with DM are at a significantly greater risk of developing CIN (6). The number of patients with prediabetes and DM worldwide is increasing; therefore, evaluating whether patients with prediabetes have a greater risk of CIN is important. This study is the first to identify the effect of prediabetes on CIN development in a short hydration method.

The CIN incidence during all treatment cycles did not differ between the control and prediabetes groups. In contrast, patients with prediabetes showed statistically significant SCr elevation and CCr decrease compared with controls. In addition, patients with prediabetes tended to develop CIN at a greater rate than those without symptoms during the first cycle, although this difference was not statistically significant. This study is the first to show that prediabetic patients exhibit a significant decrease in renal function compared to those without symptoms, although the difference was not clinically problematic.

Nephropathy is a problematic DM complication that develops slowly over several years (29). The baseline eGFR and CCr were not significantly different between the two groups in this study. Therefore, we hypothesized that the actual baseline clinical renal function was similar between the two groups, although we did not evaluate the proteinuria levels because of data unavailability. In our previous study, we hypothesized that autophagy is a possible mechanism of CIN degradation in patients with DM (6). CIN worsens in autophagy-deficient mice compared to controls (21,22). We also found that celecoxib, a cyclooxygenase-2 selective NSAID, shows nephroprotective effects against CIN through autophagy activation and oxidative stress suppression (23). In addition, Sakai et al. showed that autophagic activity in the proximal tubule is suppressed in type 2 DM because of the activation of mammalian target of rapamycin (mTOR) signaling via high plasma insulin levels; high basal autophagic activity has been confirmed in type 1 DM (24). Furthermore, renal injury induced by cisplatin administration was attenuated in mice administered streptozotocin, recognized as a model for type 1 DM (30,31). Considering the results of previous and present studies, lower insulin levels or shorter exposure periods with high insulin levels in prediabetic patients than in patients with DM may have induced mild autophagy suppression, resulting in a small aggravation of SCr and CCr. However, because autophagy activity differs between type 1 and 2 DM, evaluating patients with type 1 DM is necessary for further comprehension. Therefore, additional studies are required to evaluate CIN degradation under diabetic conditions.

CINV management is important for sufficient oral hydration in the short hydration method; therefore, appropriate prevention and handling of breakthrough symptoms according to guidelines are important (28). In this study, the incidence of all severe CINV cases was within the allowable range, and intravenous rehydration was performed adequately in severe CINV cases, resulting in a CIN incidence of 6.5% in the total population. Consequently, we should be mindful of the importance of CINV management and patient education for appropriate control of CIN.

This study has some limitations. First, it was performed retrospectively in a relatively small patient population from a single institution. Second, we evaluated prediabetes using a single baseline HbA1c assessment, considering data availability and that many factors can easily affect blood glucose levels. In addition, the duration of prediabetes was not assessed in this study. Because its level is related to blood sugar levels within the last 1-2 months, its recent variation may have affected HbA1c levels. Third, because the type of DM may affect CIN, further evaluation of the effect of DM type on CIN is necessary. Finally, we did not assess the genetic background of the patients, such as single-nucleotide polymorphisms of organic transporter 2, which reportedly affects CIN (32). Therefore, the preliminary results of this study need to be validated in future studies.

In conclusion, our study found that patients with prediabetes did not develop problematic CIN, although they exhibited significant increases in SCr and decreases in CCr. Further evaluation of CIN in diabetic conditions is necessary for its management in high-risk populations.

Funding

This work was funded by JSPS KAKENHI (grant number 22K15310).

Conflicts of Interest

The Authors declare no conflicts of interest in relation to this study.

Authors’ Contributions

Y.S. and M.K. designed the study. Y.S. and T.S. performed the experiments. Y.S. analyzed the data. Y.S. contributed new methods and models. Y.S. wrote the paper. All the Authors have read and approved the manuscript.