Effectiveness and safety of gemcitabine plus nab-paclitaxel in elderly patients with advanced pancreatic cancer: a single-center retrospective cohort study

Purpose. This study aimed to evaluate the effectiveness and safety of gemcitabine (GEM) plus nab-paclitaxel (GnP) in patients aged ≥ 75 years with advanced pancreatic cancer and compare it with monotherapy (GEM or S-1). Methods. We retrospectively reviewed the data of consecutive patients with advanced pancreatic cancer aged ≥ 75 years who received either GnP or monotherapy (GEM or S-1) between January 2014 and May 2020. The primary efficacy outcome was overall survival (OS). Results. A total of 96 patients were included in this study; 51 were treated with GnP and 45 with monotherapy (31 with GEM and 14 with S-1). The median OS and progression-free survival were 10.8 and 6.7 months in the GnP group and 10.7 and 4.3 months in the monotherapy group, respectively. The treatment effect on OS was consistently favorable in the GnP group across most subgroups, particularly in patients with locally advanced cancer, modified Glasgow prognostic score of 0 or 1, and neutrophil/lymphocyte ratio < 3.1. The disease control rates were 76% and 48% in the GnP and monotherapy groups, respectively, and grade 3 or 4 neutropenia occurred in 23 (45%) and 11 (24%) patients of the GnP and monotherapy groups, respectively. Conclusions. This study demonstrated that GnP was not superior to monotherapy with regard to OS. However, multivariate analysis showed that GnP treatment positively affected the OS and could be considered as a treatment option, even for elderly patients.


Introduction
Pancreatic cancer is a worldwide health problem with an increasing incidence and poor prognosis. Although surgical resection is the only curative treatment for pancreatic cancer, most patients have unresectable disease at diagnosis [1]. Distant lesions are observed in 40-45% of patients with pancreatic cancer and locally advanced disease in 15-20% [2]. Unfortunately, the postoperative recurrence rate in patients undergoing surgical resection is high, with relapse in approximately 75-92% of patients [3,4].
Approximately 40,900 patients are diagnosed with pancreatic cancer annually in Japan, making it the fourth and third leading cause of cancer deaths in men and women, respectively [5]. Advanced age is a high-risk factor for pancreatic cancer; in one report, at least 80% of all patients were aged > 65 years, and more than 60% were older than 75 years in 2018 [5]. Therefore, opportunities to provide chemotherapy to elderly patients with pancreatic cancer are increasing in Japan.
Regimens containing combinations of drugs, such as folinic acid (leucovorin), fluorouracil, irinotecan, and oxaliplatin (FOLFIRINOX) or gemcitabine (GEM) plus nabpaclitaxel (GnP) demonstrated their superiority over GEM Sae Ohwada and Akiko Todaka these authors contributed equally to this study. alone in two phase III studies [6,7] and have been used as a primary treatment in patients with locally advanced and metastatic pancreatic cancer. However, elderly patients are rarely treated with FOLFIRINOX because there are no data on its tolerability in those older than 75 years. The GnP trial enrolled patients under 80 years of age, with only 10% of them being aged over 75 years. Thus, neither regimen has sufficient data on its effectiveness and safety as a treatment in elderly patients with pancreatic cancer.
Conversely, it has been reported that monotherapy (GEM or S-1) was useful in elderly patients with fewer adverse events compared with combination therapy. Some reports have evaluated the effectiveness and safety of GEM as a single-agent therapy in elderly patients. Yamagishi et al. [8] and Katakura et al. [9] evaluated the therapeutic effects of GEM in patients with unresectable pancreatic cancer. They concluded that GEM monotherapy was as safe and effective in elderly patients as in younger ones. Additionally, Imaoka et al. [10] conducted a subgroup analysis in the GEST (Gemcitabine and S-1 Trial) study [11], showing the non-inferiority of S-1 to GEM and lower frequency of hematological toxicity in patients with advanced pancreatic cancer, thereby elucidating the efficacy and safety of S-1 as a first-line treatment in elderly patients. They concluded that S-1 could be a reasonable first-line treatment option in elderly patients with unresectable pancreatic cancer.
A meta-analysis [12] that included patients older than 65 years showed the superiority of GnP over GEM alone in progression-free survival (PFS) but not in overall survival (OS). A retrospective study [9] did not find GnP effective in improving PFS or OS in patients older than 75 years of age. There have been no reports comparing S-1 alone with combination therapy in elderly patients with pancreatic cancer. Hence, the therapeutic effectiveness of GnP, S-1, and GEM in the elderly has not been fully evaluated. In clinical practice, elderly patients often receive a single agent (GEM or S-1 alone) at the discretion of the attending physician. However, the choice of drug(s) for chemotherapy in the elderly is an important clinical issue.
This retrospective study aimed to evaluate the effectiveness and safety of GnP in comparison with monotherapy in patients aged ≥ 75 years with unresectable or recurrent pancreatic cancer.

Patients
We retrospectively reviewed the medical records of consecutive patients with locally advanced or metastatic pancreatic cancer who were aged ≥ 75 years and received either monotherapy (GEM or S-1) or GnP as the first-line treatment between January 2014 and May 2020 at Shizuoka Cancer Center (Shizuoka, Japan). All patients were histologically or radiologically diagnosed with pancreatic cancer. Patients who relapsed within 6 months at the end of the adjuvant chemotherapy were excluded.
This research was conducted in accordance with the Declaration of Helsinki (as revised in Fortaleza, Brazil in October 2013) and approved by the Institutional Review Board of the Shizuoka Cancer Center (IRB No. J2020-128-2020-1-3). Patients were not required to give informed consent to the study because the analysis used anonymous clinical data that were obtained after each patient agreed to treatment by written consent.

Treatment
The GnP group received nab-paclitaxel (nab-PTX) at a dose of 125 mg/m 2 and concomitant GEM at a dose of 1,000 mg/ m 2 intravenously every 4 weeks on days 1, 8, and 15. The monotherapy group received either intravenous GEM at the dose and schedule as above or oral S-1 twice daily on days 1 − 28 of a 42-day cycle. The S-1 dose was calculated according to the body surface area: 80 mg/day for 1.25 m 2 , 100 mg/day for 1.25-1.50 m 2 , or 120 mg/day for 1.50 m 2 . The dose and/or treatment schedule of these regimens were modified at the attending physician's discretion or patient's discretion. The patients continued their treatment until disease progression, the occurrence of intolerable adverse events, or patient refusal.

Outcomes
The primary effectiveness outcome was OS, defined as the duration between the start date of first-line treatment and the date of death, the date of the last follow-up, or the cut-off date of this study (November 30, 2020). The secondary effectiveness outcomes were PFS, objective response rate (ORR), disease control rate (DCR), time-to-treatment-failure (TTF), time to deterioration by the Eastern Cooperative Oncology Group performance status (ECOG PS) score, and safety. PFS was defined as the duration between the start date of first-line treatment and the date of documented disease progression or death from any cause. The investigators assessed the response according to the Response Evaluation Criteria in Solid Tumors guidelines, version 1.1, using computed tomography and/or magnetic resonance imaging. TTF was defined as the duration between the start date of treatment and the date of its discontinuation for any reason; in the GnP group, treatment failure was defined as the discontinuation of both GEM and nab-PTX. Time to deterioration by the ECOG PS was defined as the duration between the start of firstline treatment and the date when worsening of the ECOG PS was observed. Safety was assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0.
Factors included in the univariate analyses for OS were age, sex, body mass index (BMI), ECOG PS, disease status, agents, and pretreatment laboratory tests, including neutrophil/lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), modified Glasgow prognostic score (mGPS), and carbohydrate antigen 19-9 (CA 19-9). The mGPS was defined as follows: patients with an elevated C-reactive protein level (> 1.0 g/L) and hypoalbuminemia (< 3.5 g/dL) were assigned a score of 2; patients with only one of these biochemical abnormalities were assigned a score of 1, and those with neither of these abnormalities were assigned a score of 0. The NLR was the ratio between the neutrophil and lymphocyte counts, while the PLR was the ratio between the platelet and lymphocyte counts.

Statistical analysis
The treatment groups were compared for the baseline characteristics, adverse events, and response rates using the χ 2 test, Fisher's exact test, or unpaired t-test. OS and PFS were analyzed using the Kaplan-Meier curves, and differences in survival rates between the groups were analyzed using the log-rank test. The Cox proportional hazards models were used to estimate hazard ratios (HRs) of factors affecting the OS; models were adjusted for age, sex, ECOG PS, mGPS, and disease status. The cut-off values for BMI, CA-19-9, NLR, and PLR were determined based on the median value. All statistical analyses were performed using Bell Curve for Excel, version 2.21 (Social Survey Research Information Co., Ltd. Tokyo, Japan). Statistical significance was set at P < 0.05.

Baseline characteristics
A total of 96 consecutive patients were included in the study. Fifty-one received GnP, and 45 received monotherapy (31 were treated with GEM and 14 with S-1). Of the 45 patients who received monotherapy, 21 (47%) did so of their own choice, and the remaining at the discretion of the on-site physicians. The reasons for the decision were age (n = 13, 29%), poor ECOG PS (n = 7, 16), and organ dysfunction (n = 4, 9%). There were fewer patients aged ≥ 80 years (n = 8, 16% vs. n = 17, 38%, P = 0.02) and more patients with metastasis (n = 36, 71% vs. n = 23, 51%, P = 0.02) in the GnP group than in the monotherapy group. The baseline characteristics of the treatment groups are summarized in Table 1.

Response
The ORR in patients with measurable disease (n = 81) was higher in the GnP group than the monotherapy group, although the difference was not significant (n = 9, 22% vs. n = 4, 10%; P = 0.23). The DCR was significantly higher in the GnP group than the monotherapy group (n = 31, 76% vs. n = 19, 48%; P = 0.01).

Subgroup analysis of OS
The treatment effect on OS was consistently favorable in the GnP group across most subgroups. Patients with locally advanced cancer, mGPS 0 or 1, and NLR < 3.1 tended to have favorable OS in the GnP group. Patients with ECOG PS 2 and mGPS value of 2 tended to have a favorable outcome when receiving monotherapy (Fig. 2). The median OS in patients with ECOG PS of 0-1 and 2 in the GnP group was 11.6 and 2.7 months, respectively (P = 0.04). The median OS in patients with mGPS of 0-1 and 2 in the GnP group was 13.4 and 7.4 months, respectively (P = 0.005).

Time to deterioration of ECOG PS
Treatment with GnP resulted in a significantly longer time to ECOG PS deterioration than monotherapy did. The median time to deterioration in the ECOG PS score was 5.1 months in the GnP group and 3.7 months in the monotherapy group, with an HR of 0.61 (95% CI, 0.39-0.96; P = 0.03; Online Resource 1a). When categorized by baseline ECOG PS, treatment with GnP tended to extend the time to ECOG PS deterioration in patients with a baseline ECOG PS of 0 or 1 (HR, 0.61; 95% CI, 0.37-1.00; P = 0.04). There was no difference in the time to deterioration of ECOG PS between the groups in patients with a baseline ECOG PS of 2 (Online Resource 1b).

Treatment exposure
The number of patients experiencing a reduction of the initial dose was 39 in the GnP group and 29 in the monotherapy group (76% vs. 64%, P = 0.27). Of the 51 patients in the GnP group, 38 (75%) had treatment delay at least once; the most common reasons were neutropenia (n = 27), thrombocytopenia (n = 11), and fatigue (n = 4), with some overlaps. The frequency of neutropenia was significantly higher in the GnP group than in the monotherapy group (53% vs. 22%, P = 0.003). Twenty-three patients in the GnP group and 16 in the monotherapy group required dose reduction (45% vs. 35%, P = 0.41). Neutropenia (31% vs. 11%, P = 0.02) and peripheral sensory neuropathy (10% vs. 0%, P = 0.06) as reasons for dose reduction were more frequent in the GnP group.
Subsequent therapy was performed in 25 patients (49%) in the GnP group and 23 (51%) in the monotherapy group. The subsequent therapies in the treatment groups are shown in Online Resource 2.

Discussion
We found in this retrospective study that GnP was superior to monotherapy in terms of PFS, TTF, and DCR, even though it showed no improvement in OS. Subgroup analysis according to treatment regimens showed that the treatment effect on OS was consistently favorable in the GnP group across most subgroups, especially in patients with locally advanced cancer, mGPS 0 or 1, and NLR < 3.1. However, the GnP treatment effect was not observed in patients with an mGPS or ECOG PS of 2. GnP proved to be as safe as monotherapy, even in elderly patients. Furthermore, our data showed that GnP contributed to a delay in ECOG PS deterioration compared with monotherapy.
Elderly patients are heterogeneous in terms of comorbidities and physical capacity. Combination therapies, such as FOLFIRINOX and GnP, are generally recommended as firstline treatment for patients with advanced pancreatic cancer in whom physicians have determined that these therapies could be administered. However, GEM or S-1 monotherapy Fig. 2 Forest plot of treatment effects on overall survival in subgroup analyses PS, performance status; BMI, body mass index; mGPS, modified Glasgow prognostic score; NLR, neutrophil/lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; CA 19-9, carbohydrate antigen 19-9; GnP, gemcitabine (GEM) plus nab-paclitaxel; Monotherapy (GEM or S-1); HR, hazard ratio; CI, confidence interval is used for most elderly patients. A systematic review [12] assessed the efficacy of combination therapy in treating elderly patients with pancreatic cancer. The study showed that combination therapy, especially GnP, significantly improved OS compared to GEM alone. However, that study defined all patients aged ≥ 65 years as "elderly." Trials that specifically assess the role of combination therapy versus GEM in elderly patients with pancreatic cancer are lacking. Considering that the number of elderly patients with pancreatic cancer is increasing in Japan, an investigation of the effectiveness of anti-cancer therapy in such patients aged ≥ 75 years is required.
We found in this study that GnP was not superior to monotherapy with regard to OS, possibly due to the higher rate of patients with metastatic pancreatic cancer in the GnP group (71%) than in the monotherapy group (51%). In general, the prognosis in patients with metastatic cancer is worse than that in patients with locally advanced pancreatic cancer [13,14]. On-site physicians tend to select GnP to control disease progression if the patient's condition is stable. Randomized controlled trials might be necessary to exclude this bias and accurately assess the effects of the various treatments.
We found in the subgroup analysis that mGPS and ECOG PS were predictive factors for the difficulty of achieving therapeutic effects of GnP in patients with advanced pancreatic cancer. ECOG PS has been widely recognized as an important prognostic factor in multiple cancers [15]. It has also been shown to be associated with OS in pancreatic cancer patients [16][17][18]. Poor ECOG PS might be due to jaundice, anemia, and/or undernutrition, which might reflect disease progression. The poor condition and decreased tolerability to chemotherapy are possible factors associated with   a worse outcome. mGPS is also recognized as a prognostic factor in advanced pancreatic cancer [19,20]. Hwang et al. [19] reported a significant association between mGPS and OS in pancreatic cancer. Similar to previous reports [19], our study showed that patients with poor ECOG PS and mGPS had significantly worse OS. Therefore, we considered it difficult to determine the therapeutic effect of GnP in patients with an ECOG PS or mGPS of 2. The hematological toxicity in the GnP group was tolerable, and its frequency was similar to that in a GnP trial that included patients younger than those in our study [7]. Given their age, the higher proportion of patients receiving initial dose reduction in the GnP group could have influenced the frequency of hematological side effects. The pharmacodynamic and pharmacokinetic changes in elderly patients differ from those in younger patients. Elderly patients are more susceptible to the negative effects of chemotherapy on their organs. In this regard, appropriate dose reduction could contribute to tolerability. Our data demonstrated that GnP significantly slowed PS deterioration compared with monotherapy. However, neither GnP nor monotherapy prolonged the time to PS deterioration in patients with ECOG PS of 2, indicating that GnP could relieve symptoms associated with the disease better than monotherapy only for patients in good general condition at the start of treatment.
Our study has several limitations. First, this study might have been affected by a selection bias due to its retrospective nature. Indeed, the monotherapy group had a higher proportion of patients with ECOG PS of 2. Second, dose modification at the physician's discretion could affect the effectiveness and safety outcomes. Third, the data were obtained from a single institution, and the number of enrolled patients was relatively small.
In conclusion, this study showed that GnP treatment in elderly patients with unresectable pancreatic cancer led to a stronger improvement in PFS, TTF, and DCR than monotherapy did. Although the Kaplan-Meier curve analysis found no superiority of GnP with regard to OS, multivariate analysis showed that the treatment effect on OS was favorable in the GnP group across most subgroups. Additionally, GnP was as safe as monotherapy when administered at an appropriate dose. Therefore, GnP could be a treatment option for elderly patients. We believe that this study provides useful information on comparative effectiveness and safety between GnP and monotherapy in clinical practice. Further studies are required to determine which regimen should be selected for elderly patients with pancreatic cancer.