Abstract
Background/Aim: Capecitabine plus oxaliplatin (CapeOX) therapy is used as an adjuvant chemotherapy regimen for patients with colorectal cancer (CRC). Although oxaliplatin induces thrombocytopenia, the risk factors for thrombocytopenia in oxaliplatin-treated patients with CRC are not well established. We aimed to investigate the risk factors for thrombocytopenia in CapeOX-treated patients with CRC. In addition, we evaluated platelet counts and non-invasive liver fibrosis indices, specifically the aspartate aminotransferase-to-platelet ratio index (APRI) and the fibrosis-4 index (FIB-4), during CapeOX therapy in these patients. Patients and Methods: Between July 2017 and June 2020, we enrolled CapeOX-treated patients with high-risk stage II or stage III CRC at seven hospitals collaborating with the Division of Oncology, Aichi Prefectural Society of Hospital Pharmacists (Aichi prefecture, Japan). In this retrospective study, we investigated patients’ backgrounds, laboratory data, concomitant medications, number of cycles of CapeOX and oxaliplatin, cumulative dose of oxaliplatin, and administration period. The cut-off values were calculated using receiver operating characteristic analysis of platelet counts and APRI and FIB-4 scores. Results: Fifty-five patients without thrombocytopenia and 44 patients with thrombocytopenia were enrolled. During CapeOX therapy, the thrombocytopenia group showed a significant decrease in platelet count and a significant increase in APRI and FIB-4 scores compared to the non-thrombocytopenia group. Baseline albumin level ≤3.5 g/dl and platelet count ≤238×103/μl were independently associated with ≥grade 2 thrombocytopenia in CapeOX-treated patients. Conclusion: Baseline albumin level and platelet count may be useful for predicting thrombocytopenia in CapeOX-treated patients with high-risk stage II or stage III CRC.
Oxaliplatin-based chemotherapy is widely used as an established treatment for colorectal cancer (CRC) (1, 2). Capecitabine plus oxaliplatin (CapeOX) therapy is a standard adjuvant chemotherapy regimen for patients with high-risk stage II or III CRC after radical resection (3). However, oxaliplatin occasionally induces thrombocytopenia. In Japan, oxaliplatin can be administered to patients with a platelet count ≥75×103/μl (4). Therefore, oxaliplatin-induced thrombocytopenia may result in dose reduction or treatment postponement, causing exacerbation of the disease and potential treatment discontinuation. Thus, identifying risk factors for thrombocytopenia will be useful to ensure the safe and continuous administration of CapeOX adjuvant therapy.
One of the causes of oxaliplatin-induced thrombocytopenia is splenomegaly caused by the development of hepatic sinusoidal obstruction syndrome (SOS) (5). Splenomegaly was reported in the majority of oxaliplatin-treated patients, and increased spleen size correlated with the cumulative dose of oxaliplatin (6). Up to 78% of patients treated with oxaliplatin develop SOS (7), and thrombocytopenia resulting from SOS-induced splenomegaly has been shown to prolong thrombocytopenia resolution (6).
Recent noninvasive liver fibrosis indices, such as the aspartate aminotransferase (AST)-to-platelet ratio index (APRI) (8) and the fibrosis-4 index (FIB-4) (9), can be quickly and easily calculated in patients with chronic liver diseases. Therefore, such indices are expected to be applied to chemotherapy-associated liver diseases such as SOS. In oxaliplatin-treated patients, the APRI score is an independent predictive factor for splenomegaly induced by severe SOS (10). The change in FIB-4 score has also been shown to correlate with the increase in spleen size in patients undergoing oxaliplatin-based chemotherapy (11).
The purpose of this study was to identify risk factors for CapeOX-induced thrombocytopenia in patients with high-risk stage II or stage III CRC after radical resection. We also evaluated APRI and FIB-4 scores, along with platelet count, during CapeOX therapy in this population. Toward this end, we retrospectively investigated patients’ backgrounds, laboratory data, concomitant medications, number of cycles of CapeOX and oxaliplatin, cumulative dose of oxaliplatin, and administration period.
Patients and Methods
Study design and enrolled patients. This multi-center retrospective cohort study analyzed patients with CRC at seven hospitals, including the Nagoya University Hospital (Nagoya, Japan), Konan Kosei Hospital (Konan, Japan), Japanese Red Cross Aichi Medical Center Nagoya Daiichi Hospital (Nagoya, Japan), Japan Community Health care Organization Chukyo Hospital (Nagoya, Japan), Nishichita General Hospital (Tokai, Japan), Tokoname Municipal Hospital (Tokoname, Japan), and Hekinan Municipal Hospital (Hekinan, Japan), all of which collaborate with the Division of Oncology, Aichi Prefectural Society of Hospital Pharmacists (Aichi Prefecture, Japan). Our study enrolled patients hospitalized between July 2017 and June 2020 who had high-risk stage II or stage III CRC and who underwent radical resection and received CapeOX as adjuvant chemotherapy. Patients were classified as having high-risk stage II or stage III CRC according to the judgment of each attending physician, with reference to the Japanese Classification of Colorectal Carcinoma 8th or 9th edition for the treatment of CRC (12, 13). The CapeOX regimen consisted of intravenous administration of 130 mg/m2 oxaliplatin on day 1 and oral administration of 1,000 mg/m2 capecitabine twice daily on days 1-14 every 3 weeks. If hematologic or nonhematologic toxicities occurred, oxaliplatin dose reduction or discontinuation was allowed. The treatments were continued until progressive disease or unacceptable toxicity occurred, or until the patient or doctor decided upon discontinuation. We excluded patients who met any of the following criteria: platelet count outside the normal range (150-400×103/μl), history of cirrhosis, prior radiation therapy, history of chemotherapy before the initiation of CapeOX adjuvant therapy, and fewer than five cycles of CapeOX adjuvant chemotherapy.
Data collection and assessment. This study was conducted in accordance with the principles of the Declaration of Helsinki and Ethical Guidelines for Medical and Health Research Involving Human Subjects. The study protocol was approved by the Ethics Committee of Nagoya University Hospital (Approval No. 2020-0623). The following data, which were recorded in the patients’ electronic medical records prior to administering chemotherapy, were retrospectively collected: age, sex, body surface area (BSA) calculated using the formula of DuBois and DuBois, body mass index (BMI) calculated from body height and weight, tumor site, CRC stage, Eastern Cooperative Oncology Group performance status, history of diabetes mellitus and hepatitis B and C infection, smoking and alcohol consumption status, as well as concomitant medications. The number of cycles of CapeOX and oxaliplatin, the dose of oxaliplatin per course, and the cumulative dose of oxaliplatin were calculated. The oxaliplatin relative dosing intensity (RDI) was calculated as [actual dose (mg/m2)/dosing period (weeks)/planned dose (mg/m2)/dosing period (weeks)]×100 (%)].
Various laboratory data were obtained before each CapeOX treatment cycle. APRI and FIB-4 scores were calculated based on AST, alanine aminotransferase (ALT), and platelet count. Based on the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 5.0, thrombocytopenia was defined as platelet count <75×103/μl, which is ≥grade 2 (14).
Calculation of noninvasive liver fibrosis indices. The two noninvasive liver fibrosis indices were calculated as follows: APRI=[AST (IU/l)/AST upper normal limit (IU/l)/platelet (×103/μl)]×100 and FIB-4=[(age (years)×AST (IU/l))/platelet (×103/μl)×ALT (IU/l)0.5].
Statistical analysis. Quantitative variables were compared between the non-thrombocytopenia and thrombocytopenia groups using the Student’s t-test and Mann-Whitney U-test as appropriate. Qualitative variables were compared using Fisher’s exact test. Receiver operating characteristic (ROC) curves for discrete variables were calculated to determine the cut-off values of platelet count and APRI and FIB-4 scores. The cut-off values of discrete variables that maximized the sum of sensitivity and specificity were identified. Univariate analysis of risk factors associated with the incidence of thrombocytopenia was conducted using logistic regression analysis. Factors exhibiting p<0.20 were included in the multivariate logistic regression analysis. All analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) (15), which is a graphical user interface for R version 4.1.2 (The R Foundation for Statistical Computing, Vienna, Austria). EZR is a modified version of R Commander (version 1.55, designed to add statistical functions frequently used in biostatistics). Statistical significance was set at p<0.05.
Results
Participant selection. During the study period, 231 patients with high-risk stage II or III CRC received CapeOX as adjuvant chemotherapy. Of these patients, we excluded 43 with platelet counts outside the normal range of each institution, 20 with a history of chemotherapy before the initiation of CapeOX adjuvant therapy, 2 with prior radiation therapy, and 67 who underwent fewer than five CapeOX cycles. The remaining 99 patients were eligible for this study and were divided into two groups based on whether they experienced thrombocytopenia (thrombocytopenia group, 44 patients) or not (non-thrombocytopenia group, 55 patients). A flow diagram of the eligibility criteria is shown in Figure 1.
Flow diagram of patient selection. CapeOX: Capecitabine+oxaliplatin.
There were no significant differences in the baseline characteristics of the two groups (Table I). It is also noteworthy that there were no significant differences in the typical concomitant medications, median number of oxaliplatin cycles, mean oxaliplatin RDI, or median cumulative dose of oxaliplatin.
Baseline characteristics of the thrombocytopenia and non-thrombocytopenia groups.
Changes in platelet count and noninvasive liver fibrosis indices in the thrombocytopenia group. First, we confirmed whether the mean platelet count after CapeOX treatment differed significantly between the thrombocytopenia and non-thrombocytopenia groups. In this analysis, 26 patients in the thrombocytopenia group and 32 in the non-thrombocytopenia group completed all eight planned cycles of CapeOX adjuvant therapy. As expected, the mean platelet count was significantly lower in the thrombocytopenia group than in the non-thrombocytopenia group in the first through eighth cycles of CapeOX administration (Figure 2A and Table II). Previous studies demonstrated that APRI and FIB-4 scores are useful in predicting splenomegaly in patients with oxaliplatin-induced SOS (10, 11). Thus, to determine whether SOS-related splenomegaly was associated with CapeOX-induced thrombocytopenia, we calculated the longitudinal changes in the mean APRI and FIB-4 scores (Figure 2B, C and Table II). The mean APRI and FIB-4 scores were significantly higher in the thrombocytopenia group than in the non-thrombocytopenia group, raising the possibility that CapeOX-induced thrombocytopenia is not only caused by myelosuppression, but also by splenomegaly.
Longitudinal changes in platelet count and noninvasive liver indices in 58 patients (non-thrombocytopenia: n=32, thrombocytopenia: n=26) who completed eight cycles of CapeOX chemotherapy. A) changes in platelet count; B) changes in APRI score; C) changes in FIB-4 score. Data are expressed as mean±SD. APRI, Aspartate aminotransferase-to-platelet ratio index; FIB-4, fibrosis-4 index.
Platelet counts and APRI and FIB-4 scores during each cycle in 58 patients (non-thrombocytopenia: n=32, thrombocytopenia: n=26) who completed eight cycles of CapeOX chemotherapy.
Exploration of risk factors for thrombocytopenia. Next, we compared baseline laboratory data between the thrombocytopenia and non-thrombocytopenia groups. There was a significant difference in mean platelet count at baseline (p<0.001), but no significant difference in the median APRI or FIB-4 score (Table III). The threshold values of platelet count and APRI and FIB-4 scores at baseline for ≥grade 2 thrombocytopenia were subsequently estimated by ROC curve analysis (Figure 3 and Table IV). This analysis showed that the platelet count most accurately detected thrombocytopenia [area under the curve (AUC) of 0.719, 95% confidence interval (CI)=0.617-0.820]. The cut-off value of the baseline platelet count was estimated to be 238×103/μl. On the other hand, APRI and FIB-4 scores had AUCs of 0.587 (95% CI=0.473-0.701) and 0.561 (95% CI=0.445-0.676), respectively, suggesting that these cut-off values were unreliable. At baseline, therefore, a low PLT count ≤238×103/μl, but not a high APRI or FIB-4 score, might be a risk factor for ≥grade 2 thrombocytopenia induced by CapeOX therapy.
Baseline laboratory data of the thrombocytopenia and non-thrombocytopenia groups.
Receiver operating characteristic (ROC) curve analysis to calculate cut-off values for baseline platelet count, APRI and FIB-4 scores for ≥grade 2 thrombocytopenia in enrolled patients. APRI, Aspartate aminotransferase-to-platelet ratio index; FIB-4, fibrosis-4 index. Solid line represents the ROC curve for baseline platelet count; dashed-dotted line represents the ROC curve for baseline APRI; dashed line represents the ROC curve for baseline FIB-4.
Calculation of the cut-off values for ≥grade 2 thrombocytopenia.
Lastly, we applied univariate and multivariate analyses to identify the risk factors for thrombocytopenia in the enrolled patients (Table V). A previous study showed that comorbid diabetes mellitus was associated with ≥grade 2 thrombocytopenia induced by oxaliplatin-based therapy, although the study was performed in patients with advanced gastric cancer (16). In addition, the Guidelines for the Treatment of Cirrhosis 2020 defined an albumin (ALB) level ≤3.5 g/dl as protein–energy malnutrition, one of the criteria for nutritional therapy (17). Taking these reports into consideration, various exploratory factors were selected in the univariate analysis, which revealed that risk factors with p<0.20 were male sex, rectal cancer, baseline ALB level ≤3.5 g/dl, and platelet count ≤238×103/μl. In a multivariate analysis of these four factors, ALB level ≤3.5 g/dl and platelet count ≤238×103/μl were independent risk factors for ≥grade 2 thrombocytopenia.
Univariate and multivariate analyses of the risk factors for ≥grade 2 thrombocytopenia.
Discussion
CapeOX is an effective adjuvant chemotherapy regimen for CRC (3). In a previous report, CapeOX therapy was reported to cause grade 3/4 severe thrombocytopenia in 5% of eligible patients (18). Another previous study also demonstrated that CapeOX therapy induced grade ≥3 severe thrombocytopenia in 8.3% of Japanese patients with stage II/III colon cancer (19). Therefore, the development of CapeOX-induced thrombocytopenia may have a significant impact on treatment adherence. Understanding the risk factors associated with CapeOX therapy is important to achieve safe, continued treatment. Accordingly, in this retrospective study we focused on risk factors for thrombocytopenia induced by CapeOX therapy in patients with CRC. There were two major findings. Firstly, we identified lower baseline ALB level and platelet count as independent risk factors for CapeOX-induced thrombocytopenia. Secondly, we found that the baseline APRI and FIB-4 scores were not associated with CapeOX-induced thrombocytopenia.
In this study, multivariate analysis identified a significant correlation between CapeOX-induced thrombocytopenia and baseline ALB level ≤3.5 g/dl, although there was no significant difference in the baseline level of ALT or AST between the thrombocytopenia and non-thrombocytopenia groups. Hypoalbuminemia can be caused by a variety of factors, including liver disorders, nutritional deficiencies, cancer, increased transcapillary escape rate, and decreased lymphatic clearance due to major surgery (20). Because all patients in this study had stage II or stage III CRC and had undergone radical resection, hypoalbuminemia may have been due to postoperative or tumor-induced effects. In addition, oxaliplatin has an ALB binding rate of over 90%, and hypoalbuminemia could increase the distribution of free oxaliplatin, thus increasing the risk of adverse effects. A previous study reported a higher incidence of oxaliplatin-induced chronic peripheral neuropathy in patients with hypoalbuminemia than in those without it (21). In light of the results of the aforementioned studies, the present results suggest that a lower baseline ALB level is associated with the development of CapeOX-induced thrombocytopenia.
We identified a significant correlation between thrombocytopenia and baseline platelet count in both univariate and multivariate analyses. Moreover, the platelet count during the first cycle was significantly lower in the thrombocytopenia group than in the non-thrombocytopenia group. This suggests that CapeOX-induced thrombocytopenia in patients with CRC and low baseline platelet counts should be cautiously monitored beginning in the initial cycle, because these patients are likely to have low platelet-producing capacity in the bone marrow or a larger platelet reservoir and degradation rate in the spleen. Li et al. reported that five factors including ALB and baseline platelet count were independent risk factors for oxaliplatin chemotherapy-induced thrombocytopenia (22). This previous report supports our results.
It has been thought that the development of oxaliplatin-induced thrombocytopenia is related to myelosuppression, immune-mediated reaction, and splenomegaly due to hepatic SOS (5). Myelosuppression is the main cause of thrombocytopenia in patients with CRC receiving oxaliplatin-based therapy. Thrombocytopenia induced by myelosuppression is mild or moderate in severity, and the platelet count recovers before the next chemotherapy cycle in most patients (5). In immune thrombocytopenia, the platelet count tends to decrease rapidly, as a result of hypersensitivity reactions (5, 23), although there were no patients with immune thrombocytopenia in the present study. Thrombocytopenia resulting from SOS-induced splenomegaly is mild, but recovery is slow, and the condition tends to persist for 2-3 years after treatment is discontinued (6, 24). A previous report has indicated that oxaliplatin induces perisinusoidal fibrosis and veno-occlusive lesions in the liver (25). These injuries induce portal hypertension followed by SOS, and their complications include splenomegaly with associated thrombocytopenia (26-28). In a previous study, an increase in spleen size was observed in 86% of patients receiving oxaliplatin-based adjuvant chemotherapy, and 24% of these patients had an increase in spleen size of 50% or more (6). Increased spleen size also predicted sinusoidal injury with a higher histologic grade and strongly correlated with thrombocytopenia and liver dysfunction (6, 24). A high APRI score has also been reported to predict the onset of SOS (10). In a previous study, mean platelet counts were lower in patients with splenomegaly than in those without it, and this difference was most pronounced between 12 and 24 weeks after the start of oxaliplatin-based chemotherapy (29). In patients who received FOLFOX (oxaliplatin, fluorouracil, and levofolinate combination therapy) or CapeOX treatment, higher FIB-4 scores correlated well with increased spleen volume (11). The present study showed that during each cycle of CapeOX treatment, APRI and FIB-4 scores were significantly higher in the thrombocytopenia group than in the non-thrombocytopenia group. Since the enrolled patients received at least five cycles of CapeOX treatment, thrombocytopenia due to CapeOX could have been caused not only by bone marrow suppression, but also by splenomegaly due to SOS. Furthermore, this study showed that the baseline APRI and FIB-4 scores were not related to the development of CapeOX-induced thrombocytopenia. Miura et al. reported that splenomegaly due to FOLFOX-associated hepatotoxicity was predicted by an APRI score of 0.17 or higher, before FOLFOX therapy (30). However, our results were inconsistent with this, possibly due to differences in the cancer stage and regimens of enrolled patients.
Another previous study demonstrated that a history of diabetes mellitus was a risk factor for thrombocytopenia induced by oxaliplatin-based therapy in patients with advanced gastric cancer (16). Our results showed that there was no significant difference between the thrombocytopenia and non-thrombocytopenia groups regarding the history of diabetes mellitus. These findings are inconsistent with those of the aforementioned report, probably due to different cancer types and regimens in the two studies. Many drugs are known to induce thrombocytopenia, including nonsteroidal anti-inflammatory drugs and proton pump inhibitors (31, 32). However, there were no significant differences in concomitant drugs between the thrombocytopenia and non-thrombocytopenia groups. In addition, there were no significant differences between the thrombocytopenia and non-thrombocytopenia groups in terms of the median number of oxaliplatin cycles, mean oxaliplatin RDI, and median cumulative dose of oxaliplatin. The oxaliplatin RDI of both groups was low [70.5±12.9% in the thrombocytopenia group vs. 73.0±19.1% in the non-thrombocytopenia group], suggesting that dose reduction or discontinuation in the non-thrombocytopenia group was the result of side effects other than thrombocytopenia (e.g., neutropenia, nausea, peripheral neuropathy).
There are several limitations to this study. Firstly, it used a non-randomized, retrospective cohort design. Secondly, despite the fact that this was a multi-center retrospective study, we analyzed only 99 patients. Thirdly, spleen volume and SOS could not be directly assessed using computed tomographic imaging and liver histopathological analysis. We hope that the results of this study will be validated in the future by a large prospective study with imaging confirmation of SOS-induced splenomegaly.
Conclusion
The baseline ALB level and platelet count may be useful for predicting thrombocytopenia in CapeOX-treated patients with high-risk stage II or stage III CRC. Accordingly, baseline ALB and platelet count should be cautiously monitored from the initial cycle of CapeOX therapy in these patients. Although the present study provides some new findings, further basic and clinical studies are needed to elucidate the detailed mechanisms of CapeOX-induced thrombocytopenia.
Acknowledgements
The authors thank Dr. Tatsuya Hisada (Department of Pharmacy, Toyota Memorial Hospital, Aichi, Japan) and Mr. Kosuke Kawai (Department of Pharmacy, Tokai Hospital, Nagoya, Japan) for their management in Division of Oncology, Aichi Prefectural Society of Hospital Pharmacists, and Dr. Masahiko Ando (Department of Advanced Medicine, Nagoya University Hospital, Nagoya, Japan) for his help in statistical analysis.
Footnotes
Authors’ Contributions
TN and HM designed the study. NK, TN, SKodama, SKoyama, SN, YW, HO, and HK performed data acquisition. NK and TN analyzed the data. NK, TN, YM, and MM performed data interpretation. NK and TN wrote the draft and edited it. All authors reviewed the draft. SN, MM, SY, and KY supervised the study. All authors read and approved the final manuscript.
Conflicts of Interest
The Authors declare no competing financial interests.
- Received December 22, 2023.
- Revision received February 16, 2024.
- Accepted February 19, 2024.
- Copyright © 2024, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.