Gastric cancer incidence is decreasing globally as food hygiene improves and Helicobacter pylori prevalence declines (1). However, according to cancer statistics compiled by Japan’s National Cancer Center, gastric cancer is the third most common type of cancer, after colorectal cancer and lung cancer (2).

According to The Japanese Gastric Cancer Treatment Guidelines 2025 (7th edition), Ramucirumab (RAM), used as a second-line treatment for gastric cancer, is a monoclonal antibody that targets vascular endothelial growth factor receptor 2 (VEGFR-2) (3). RAM inhibits endothelial cell proliferation, migration, and survival, thereby inhibiting tumor angiogenesis (4). Therefore, it is used to treat not only unresectable advanced or recurrent gastric cancer, but also colorectal cancer, non-small cell lung cancer and more recently, liver cancer, making it one of the treatment options for these types of cancer. Ramucirumab is effective, but caution is required regarding proteinuria, especially in East Asian populations (5). Proteinuria is a prevalent side effect of the VEGFR-2 inhibitory action of RAM (4). One cause of proteinuria is the inability of vascular endothelial growth factor to bind to VEGFR-2, which impairs signal transduction. This disruption also affects the structure of the glomerulus and its filtration function (6).

RAM administration can be started if proteinuria is 1+ or less; however, if proteinuria is 2+ or higher and a quantitative test shows more than 2 g, RAM must be stopped (7,8). Therefore, progression to proteinuria of 2+ or higher results in the discontinuation of RAM administration and a decrease in dose intensity. Thus, if such an event can be predicted based on evidence and preventative measures are implemented, a more effective treatment course can be achieved.

Several studies have investigated the relationship between RAM and proteinuria. According to Sakuma et al., the presence of trace proteinuria, i.e. proteinuria (+/−), at the initiation of RAM therapy, along with the concurrent use of paclitaxel (PTX), are risk factors for proteinuria onset (9). In addition, Ruiqi et al. also stated that the concomitant use of PTX is a risk factor for the development of proteinuria especially in East Asian populations (5). In another study, Horoiwa et al. found that taking non-steroidal anti-inflammatory drugs (NSAIDs) concurrently is a risk factor (10). Furthermore, Kanbayashi et al. identified systolic blood pressure (≥135 mmHg), treatment number (≥13), and concurrent use of calcium antagonists as risk factors, while Kimura et al. identified lung cancer as a risk reducing factor and proteinuria (+/−) at the start of RAM as risk factors (11,12). It is also believed that the presence of proteinuria at the time of initiation of various antiangiogenic drugs may be a risk factor for worsening proteinuria (13). All these factors are important at the start of RAM and after administration begins. Furthermore, Tanaka et al. and Dote et al. identified a history of bevacizumab (BV) use as a risk factor (14,15). According to reports, proteinuria trend and level over the prior year represent an adverse prognostic factor and good predictors of subsequent end stage renal disease in high-risk patients with IgA nephropathy (16-18). Based on these findings, we hypothesized that not only the presence of proteinuria at the start, but also fluctuations in proteinuria before RAM initiation may influence abnormal proteinuria after RAM administration. Therefore, to reduce the previously reported risk factors, this study limited participants to those with unresectable advanced or recurrent gastric cancer who were administered PTX, effectively excluding patients with lung cancer. Furthermore, we examined the impact of fluctuations in proteinuria within one year, six months or three months before initiating RAM-based treatment.

Data source and study design. This retrospective cohort study was authorised by the ethics boards of Gifu Prefectural General Medical Center, which was carried out in accordance with the approved guidelines. All study participants gave informed consent. Patients who received RAM-based therapy for gastric cancer and had previously used platinum-based chemotherapy at Gifu Prefectural General Medical Center between August 2015 and December 2021 were eligible.

To administer the RAM + PTX combination therapy, 8 mg/kg RAM (1 h intravenous infusion) was given on days 1 and 15, followed by 80 mg/m² PTX (1 h intravenous infusion) on days one, eight, and 15. RAM + PTX therapies were administered thrice in four weeks. To administer the RAM + nab-PTX combination therapy, 8 mg/kg RAM (1 h intravenous infusion) was given on days one and 15, followed by 100 mg/m² nab-PTX (1 h intravenous infusion) on days one, eight, and 15. RAM + PTX therapies were carried out three times in 4 weeks. All study participants were recruited. Patients were excluded from the study if they received primary care at another hospital, had a qualitative proteinuria level higher than 1+ before starting RAM-based treatment or lacked a qualitative proteinuria report before treatment initiation.

Data collection. Patient information was gathered from electronic patient files stored in the databases of the Gifu Prefectural General Medical Center. At the beginning of treatment, the following information was gathered: age, sex, performance status (PS) assessment by using Eastern Cooperative Oncology Group, stage of cancer, planned relative dose intensity (RDI) (%) of RAM, type of paclitaxel (PTX or nab-PTX), detailed treatment history (past chemotherapy: presence of platinum or BV, and the days that passed from the end of platinum-based chemotherapy if used), medical history (diabetes and hypertension), oral medication (antihypertensive agent, calcium antagonist and NSAIDs), qualitative analysis of proteinuria or creatinine clearance prior RAM-based therapy, and the beginning of aggravating level of proteinuria or creatinine clearance for the year prior RAM. Furthermore, data on aggravating levels of proteinuria within six or three months before initiating RAM-based treatment were collected. Information on the following variables was extracted throughout the treatment period: body height, body weight, body surface area (BSA), numbers of the RAM-based therapy; RAM dosage, PTX dosage, nab-PTX dosage, cumulative dose of RAM, and blood biochemical parameters such as leukocyte count (×10⁴/μl), neutrocyte count ratio (%), hemoglobin level (g/dl), platelet count (×10⁴/μl), serum aspartate aminotransferase level (IU/l), serum alanine aminotransferase level (IU/l), serum total bilirubin level (mg/dl); serum creatinine (Scr) (mg/dl); urea nitrogen (mg/dl); serum albumin (Alb) (g/dl); blood sugar (mg/dl); systolic blood pressure (mmHg); diastolic blood pressure (mmHg); urine parameters including qualitative analysis of proteinuria. Qualitative analysis of proteinuria was recorded as one of the following categories based on severity: (4+), (3+), (2+), (+), (+/−) or (−) depending on the patient's condition. Creatinine clearance (Ccr) was determined using the Cockcroft-Gault formula: Ccr (ml/min): [140 − age (years)] × weight (kg) × 0.85 (in women)/(Scr × 72).

Assessment. The severe group included patients with a qualitative proteinuria score of more than 2+ following RAM-based therapy. The others were assigned to the control group. Diabetes was defined as having blood sugar level greater than 200 mg/dl, HbA1c of 6.5%, or the disease name listed in a medical record. Hypertension was defined as having systolic blood pressure of 140 mmHg or higher, diastolic blood pressure of 90 mmHg, or having the disease name listed in a medical record. The aggravating group consisted of patients who had experienced an increase in the levels of qualitative proteinuria for the previous six or three months before starting RAM-based-treatment.

In terms of risk factor extraction, the factors that showed a significant difference between the severe and control groups in this study, as well as those reported in previous reports, were considered. The factors that showed significant differences between the severe and control groups in this study were as follows: borderline proteinuria levels, +/−; diastolic blood pressure with a threshold of 90 mmHg as per Japanese hypertension guidelines; the use of calcium antagonists and the onset of worsening qualitative proteinuria within one year or six months before starting RAM-based treatment. The factors reported in the previous report were grouped as follows: the use of NSAIDs, systolic blood pressure (135 mmHg with a border), and numbers of RAM (9 as the median of the population). Additionally, a decrease in Ccr levels exceeding 13 ml/min, based on the median decline within one year before RAM-based treatment for the entire population as the associated factor of the onset of worsening qualitative proteinuria included in the examination item.

Statistical analysis. Normally distributed variables are expressed as mean±standard deviation, while non-normally distributed variables are expressed as median (interquartile range). Parametric and nonparametric pairwise data were compared using the unpaired t-test and Mann-Whitney U-test, respectively. Ratios were compared using the chi-squared and Fisher’s exact tests. The onset period of detection of proteinuria levels exceeding 2+ was determined using the Kaplan-Meier method, and the log-rank test was used to compare two groups. All statistical analyses were carried out using the Statistical Package for the Social Sciences (version 24.0; IBM Corporation, Armonk, NY, USA). Differences and associations with p-values less than 0.05 were deemed statistically significant.

Ethics. For all procedures involving human participants, this study followed institutional and national ethical standards, as well as the 1964 Declaration of Helsinki and its amendments or comparable ethical standards. The study protocol was authorized by the Epidemiology and Clinical Research Ethics Review Board of Gifu Prefectural General Medical Center (ethical approval number: 729, date: June 21, 2022).

Patient characteristics. In this study, out of 88 patients, 62 were included. Furthermore, 26 patients were excluded for the following reasons: One patient received the combination therapy of RAM + PTX at another hospital, 10 patients underwent a qualitative proteinuria analysis higher than 1+ before initiating RAM-based treatment, i.e., RAM + PTX or RAM + nanoparticle-albumin-bound paclitaxel (nab-PTX) and 15 patients did not have the qualitative proteinuria analysis before the initiation of RAM-based treatment. The participants were aged 71 (65-76) years, with 79.0% being men. Stage IV disease was noted in 96.8% of the participants (Table I). The factors showing a significant difference at baseline between the two groups were proteinuria, diastolic blood pressure and the use of a calcium antagonist (Table I). Additionally, the onset of aggravating levels of proteinuria within 1 year and 6 months before receiving RAM-based treatment showed a significant difference between the two groups (Table I). Similarly, there was no difference in the onset of aggravating levels of Ccr within one year before receiving RAM-based treatment between the two groups (Data not shown). Furthermore, no difference was found in the amount of chemotherapy received between the two groups (Table II).

Influence of expression days on qualitative proteinuria analysis higher than 2+ following RAM treatment. The risk factors associated with an incidence of proteinuria higher than 2+ in qualitative analysis after receiving RAM-based treatment, as identified in Table I, are proteinuria (+/−), higher diastolic blood pressure and use of a calcium antagonist before chemotherapy. The influences of these factors are depicted in each Kaplan-Meier curve. A significant difference was observed between +/− group and − group of baseline qualitative proteinuria (p=0.034) (Figure 1A). No difference was found between the two groups in baseline diastolic blood pressure (Figure 1B). A significant difference was observed between the administered and non-administered groups of calcium antagonist use (p=0.021) (Figure 1C).

The Kaplan-Meier curves illustrate the factors influencing the incidence of proteinuria levels higher than 2+ after receiving RAM-based treatment, highlighting the aggravation of qualitative proteinuria levels before RAM treatment at various time points. A significant difference was found between the aggravated group and the normal group within one year (p=0.007) (Figure 2A) and six months (p=0.007) (Figure 2B) before initiating RAM-based treatment. A marginally significant difference was found between the two groups within three months before initiating RAM-based treatment (p=0.081) (Figure 2C). However, no difference was found between the two groups regarding the onset of a ≥13 ml/min Ccr level decrement within one year before initiating RAM-based treatment (Figure 3).

The effects of factors reported in previous studies, such as NSAIDs use, higher systolic blood pressure value, and the number of RAM-based treatment numbers, are depicted in Kaplan-Meier curves (Figure 4A-C). Figure 4A-C all show no significant differences.

Progression of proteinuria prior to RAM-based therapy in patients who developed grade ≥2 proteinuria post-treatment. Table III shows the list of patients with ≥2+ proteinuria after receiving RAM-based treatment. The horizontal column indicates whether or not a worsening level of proteinuria occurred within six or three months before the start of RAM-based therapy and the baseline proteinuria level. The three of six patients (No. 2, 4, 5, 6, 9 and 10) that had the onset of aggravating level of proteinuria during the past six months had subsequently achieved negative level at the baseline and developed proteinuria (2+) after starting RAM-based therapy again.

We investigated the risk factors associated with the onset of proteinuria higher than 2+ in qualitative analysis caused by RAM-based treatment. This study indicates that the presence of proteinuria (+/−) before initiating RAM-based treatment, concurrent use of a calcium antagonist and progression of proteinuria within six months before RAM-based treatment may affect its discontinuation.

Sakuma et al. reported that risk factors for developing proteinuria of 2+ or higher include proteinuria (+/−) at the start of RAM-based treatment and concurrent use of PTX (9). Some previous reports have also reported that the presence of proteinuria at the start of RAM is a risk factor (12,13). Ruiqi et al. (5) reported that the onset of severe adverse reaction including proteinuria caused by the combined treatment with RAM and PTX was higher than RAM monotherapy. The findings of the present study are consistent with the results for proteinuria (+/−) at the start of RAM-based treatment on previous studies. Patients with proteinuria (+/−) at the start of RAM-based treatment developed microalbuminuria, suggesting that RAM administration may have caused a breakdown in the glomerular structure and filtration function, leading to the onset of proteinuria. Regarding concurrent use of PTX, nab-PTX is considered nearly identical to PTX, as it is a PTX preparation with a 130 mm diameter, where PTX is bound to human serum albumin. In this study, as all patients received either PTX or nab-PTX, there may have been no significant difference between the two groups. This study found no clear association with the concomitant use of NSAIDs. Horoiwa et al. studied patients with lung cancer, a history of BV administration and proteinuria of 1+ or higher at the start of RAM-based treatment (10). These factors have been considered risk factors in some previous reports (9,12-15). Therefore, they were excluded from our study, which may account for differences in the findings compared to the previous studies. Systolic blood pressure, frequency of administration, and concurrent use of a calcium antagonist were identified as risk factors by Kanbayashi et al. (11). The current study yielded similar results, identifying only the concurrent use of a calcium antagonist as a risk factor. However, no discernible relationship was found between systolic blood pressure, dose frequency and the development of proteinuria higher than 2+. Lung cancer was associated with a reduced risk of proteinuria (12); however, we did not assess patients with lung cancer because the participants of this study were patients with gastric cancer. Platinum, which is commonly used as an initial treatment for gastric cancer, is well known for causing a decrease in renal function. Specifically, cisplatin is known to cause acute kidney injury via proximal tubular damage, oxidative stress, inflammation and renal vascular damage (19,20). Therefore, platinum used in the pre-treatment may have decreased the renal function, aggravating proteinuria during the subsequent treatment with RAM. However, no obvious correlation was found in this study. Kimura et al. reported that renal dysfunction during cisplatin use could be prevented with hydration and that this did not result in proteinuria (12). The findings of our study may be attributed to a similar background. We did not identify the history of BV administration as a risk factor as suggested by Tanaka et al. (14,15), as the study participants were limited to gastric cancer patients who had no indications for BV administration in the present study.

The Japanese Gastric Cancer Treatment Guidelines 2025 (7th edition) recommend RAM-based therapy as the primary option for second-line treatment. Reports have indicated that proteinuria is associated with a poor prognosis (21,22). Therefore, it is crucial to manage proteinuria while continuing treatment to improve the prognosis of patients with unresectable advanced or recurrent gastric cancer. Regarding RAM-based treatment administration, it is stipulated that treatment should be discontinued if proteinuria reaches more than three g per day (7,8). Furthermore, a quantitative test to determine proteinuria levels should be performed when the qualitative test yields a result of 2+ or higher. Hence, it is critical to prevent the onset of proteinuria of 2+ or higher at the start of RAM-based treatment. Therefore, identifying risk factors that may cause proteinuria of 2+ or higher in advance is considered extremely beneficial for the long-term continuous administration of RAM. Furthermore, the incidence of proteinuria is higher when RAM is administered in combination with other agents, compared to when administered alone (5,7-9,23-27). Caution should be exercised, particularly when RAM is administered concurrently with either PTX or nab-PTX. Although the combination therapy of RAM and PTX is generally well tolerated in patients aged 70 years or older, attention must be given to the increased risk of adverse events in this population (28). In this study, 6 out of 14 patients who developed ≥2+ proteinuria were aged 70 or older. However, no statistically significant association was observed between age and the onset of ≥2+ proteinuria. These findings suggest that elderly patients may still tolerate the development of ≥2+ proteinuria during RAM-based therapy. Goto et al. (29) reported the safe use of RAM without antihistamines in Japanese patients, including older adults for whom antihistamine administration may be challenging. The incidence of grade ≥2 proteinuria in their study was comparable to that observed in our cohort. Although all patients in the present study received antihistamines, these findings suggest that the necessity of antihistamine premedication may be evaluated based on individual symptoms.

Study limitations. One limitation of this study is its retrospective design and its exclusive focus on gastric cancer, which was intended to narrow the scope for identifying risk factors. Furthermore, RAM-based treatment was administered in the second or later stages of therapy. Thus, with many other options available, we were able to conduct only a limited analysis of 62 cases. Furthermore, since quantitative tests are rarely performed consistently in daily clinical practice, qualitative tests, which are more commonly used in routine clinical practice, were employed as evaluation items. Concurrent use of PTX, a previously reported risk factor (5,9), was excluded because all cases involved the administration of either PTX or nab-PTX. Because this study was limited to patients with gastric cancer, but not lung cancer (12), the influence of BV (14,15) did not have to be considered. Due to the small sample size, a Cox-Hazard analysis could not be performed. However, collecting a sufficient number of cases in future studies may provide a better understanding of the predictive factors. Future reanalysis will include a larger sample size.

The results of this study indicate that not only the presence of proteinuria at the initiation of RAM-based treatment, but also fluctuations in proteinuria within 6 months prior to RAM initiation, influence both the onset of ≥2+ proteinuria and the time to onset. This suggests that transitioning to the next treatment should only occur after carefully considering whether proteinuria was exacerbated during the pre-treatment period. We might predict the future exacerbation of proteinuria caused by RAM-based treatment if there is a change in proteinuria within six months, even if the qualitative test was negative at the start of the treatment. Therefore, it is recommended that qualitative proteinuria tests be performed regularly when patients with gastric cancer begin chemotherapy, in preparation for the transition to second-line treatment. This suggests that qualitative tests should be performed regularly when initiating primary treatment for patients with gastric cancer.

For patients receiving RAM-based therapy who have aggravated proteinuria within six months before receiving RAM-based treatment, proteinuria testing should be conducted throughout the treatment period. RAM-based treatment administration should be discontinued immediately if abnormalities are detected.

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

Conceptualization: T.H. (Takahiro Hayashi), T.H. (Tomoyuki Hirashita), and S.M.; methodology: T.H. (Takahiro Hayashi), T.H. (Tomoyuki Hirashita), S.M., N.N. and M.K.; validation: T.H. (Takahiro Hayashi), T.H. (Tomoyuki Hirashita) and S.M.; formal analysis: T.H. (Takahiro Hayashi) and S.M.; data curation: S.M. and H.O.; investigation: H,N, R.I. and S.M.; resources: T.H. (Tomoyuki Hirashita); supervision: T.H. (Takahiro Hayashi); visualization: T.H. (Tomoyuki Hirashita), S.M., H.O. and K.I.; writing - original draft preparation: T.H. (Takahiro Hayashi), T.H. (Tomoyuki Hirashita) and S.M.; writing - review and editing: T.H. (Takahiro Hayashi), T.H. (Tomoyuki Hirashita) and S.M. All Authors have read and agreed to the published version of the manuscript.

The Authors wish to thank the patients and their families for providing information, agreeing to the analysis, and consenting to the acknowledgement.

This research received no external funding.

No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.