Antibiotics May Interfere with Nivolumab Efficacy in Patients with Head and Neck Squamous Cell Carcinoma

Immune checkpoint inhibitors (ICIs) have been approved for the treatment of several types of cancer and have demonstrated potential treatment efficacy [1‒5]. Head and neck squamous cell carcinoma (SCC) is a cancer that is diagnosed in over 700,000 patients worldwide, and the number of patients with head and neck SCC is increasing every year [6]. Early-stage head and neck SCC is generally treated with chemoradiation therapy, which can decrease the 5-year mortality of patients by 6.5% (hazard ration for mortality: 0.83) [7]. In particular, in the case of local lesions with head and neck SCC, there are many cases in which a complete response to chemoradiation therapy can be achieved [8]. However, head and neck SCC with distant metastases or recurrence after radiotherapy cannot be treated with re-irradiation and is generally treated with systemic chemotherapy [9]. Nivolumab is an ICI approved for patients with head and neck SCC. Patients with head and neck SCC treated with nivolumab were found to have a longer overall survival time (OS) than those treated with methotrexate, docetaxel, or cetuximab [10]. In addition, nivolumab had demonstrated a longer time to deterioration of the quality of life than single agents such as methotrexate, docetaxel, or cetuximab in patients with platinum-refractory recurrent or metastatic head and neck SCC [11]. Based on these previous reports, nivolumab is considered the standard of care for patients with platinum-refractory recurrent or metastatic head and neck SCC.

It was recently reported that antibiotics can reduce the treatment efficacy of nivolumab in several types of cancers (lung cancer, renal cancer, and melanoma) [12‒15]. Head and neck SCC patients are often treated with antibiotics for various reasons, such as infection, myelosuppression due to chemotherapy, prevention of acne-like eruptions with anti-epidermal growth factor receptor antibody, and so on [16, 17]. Therefore, in head and neck SCC patients, the treatment efficacy of ICI might be reduced by antibiotic treatment. However, few studies have evaluated the effects of antibiotics on the treatment efficacy of nivolumab in patients with head and neck SCC. In this study, we retrospectively assessed the influence of antibiotics on the treatment efficacy of nivolumab in patients with head and neck SCC.

We reviewed the medical records of patients with unresectable advanced or recurrent head and neck SCC treated with nivolumab at the Department of Medical Oncology, Tohoku University Hospital, between April 2017 to January 2021.

The inclusion criteria were as follows:

• 1.

  Patients who underwent histopathological diagnosis of head and neck SCC were included. Patients diagnosed with cancer of unknown primary were also included (SCC proven on histopathological examination cervical lymph node tissue). The primary site of SCC included oral cavity (tongue, gingiva, and floor of the mouth), nasal cavity, paranasal sinus, pharynx, larynx, and head and neck skin.

• 2.

  Receiving at least one course of systemic chemotherapy with nivolumab at the Department of Medical Oncology, Tohoku University Hospital.

The patients were classified into the following two groups based on history of antibiotic treatment.

Antibiotic-treated group: patients who received oral or intravenous antibiotics between a month before the day of nivolumab initiation and the day of the first computed tomography (CT) evaluation of nivolumab efficacy.

Antibiotic-untreated group: patients who did not meet the eligibility criteria for the antibiotic-treated group.

The ICI used to treat the included patients was nivolumab. It was administered at a dose of 240 mg/body/day every 2 weeks or 480 mg/body/day every 4 weeks. All antibiotics were treated to enrolled patients according to the medical package insert.

Treatment response was assessed using the Response Criteria in Solid Tumor version 1.1. Response rate (RR) was calculated as follows:

(Total number of patients exhibiting a complete response [CR] and partial response [PR])/(Total number of patients with CR, PR, stable disease (SD), and progressive disease [PD]).

Disease control rate (DCR) was calculated as follows:

(Total number of patients with CR, PR, and SD)/(Total number of patients with CR, PR, SD, and PD).

Progression-free survival time (PFS) was defined as the duration from the day of nivolumab initiation to the day of imaging confirmation of disease progression or clinical confirmation of PD. OS was defined as the duration from the day of nivolumab initiation to the day of patient death. Neutrophil to lymphocyte ratio (NLR) was defined as the value of the neutrophil count divided by the lymphocyte count. Adverse events were determined by reviewing patient’s medical records and were evaluated according to the Common Terminology Criteria for Adverse Events version 5.0.

The PFS and the OS were analyzed using the Kaplan-Meier methods. The p values for intergroup differences in RR and DCR were calculated through Fisher’s exact test, and those for intergroup differences in patient age were calculated through a t test. All statistical analyses (including univariate and multivariate analyses) were performed using JMP^® 11 (SAS Institute Inc., Cary, NC, USA), and all differences were regarded as statistically significant when p < 0.05.

Sixty-four patients were enrolled in this study. The patient background is presented in Table 1. Forty-five patients were assigned to the antibiotic-treated group and 19 to the antibiotic-untreated group. More than 80% of patients in both groups were male. The most common primary site of SCC in the present study was the oral cavity, followed by the hypopharynx, oropharynx, and larynx. Approximately 90% of patients enrolled in the present study had 0 or 1 of performance status. The most common antibiotics taken by patients in the antibiotic-treated group were tetracyclines, followed by cephems, penicillins, and quinolones. The reasons for the antibiotic use were treatment for infections in 21 patients (46.7% in the antibiotic-treated group), prevention of acne-like eruption during the treatment with anti-epidermal growth factor receptor antibody in 18 patients (40.0% in the antibiotic-treated group), and prevention of infections in 6 patients (13.3% in the antibiotic-treated group). The median duration of antibiotic treatment was 21.5 days (range 3–113 days). The number of patients treated with nivolumab as the first, second, or third- or later-line therapy was 4 (8.9%), 34 (75.6%), and 7 (15.6%), respectively, in the antibiotic-treated group and 8 (42.1%), 6 (31.6%), and 5 (26.3%), respectively, in the antibiotic-untreated group. The percentage of patients who received nivolumab as the first line therapy in the antibiotic-untreated group was significantly higher than that in the antibiotic-treated group. Conversely, the percentage of patients treated with nivolumab as the second-line therapy in the antibiotic-treated group was significantly higher than that in the antibiotic-untreated group.

The percentage of patients treated with paclitaxel after nivolumab therapy in the antibiotic-treated group (24.4%) was significantly higher than that in the antibiotic-untreated group (0.0%) (p = 0.0179). Approximately 70% of patients in both groups had a neutrophil-to-lymphocyte ratio >4.

The RR and DCR of the two groups are presented in Table 2. The RR in the antibiotic-treated group and antibiotic-untreated group was 23.7% and 42.1%, respectively. Thera was no significant difference between two groups (p = 0.1518). The DCR in the antibiotic-treated group and antibiotic-untreated group was 63.2% and 73.7%, respectively. There was no significant difference between two groups (p = 0.4268). The median PFS and median OS are presented in Figures 1a, b. The median PFS and median OS of the antibiotic-treated group were 3.2 months (95% confidential interval [CI]: 2.0–4.1) and 5.8 months (95% CI: 2.3–16.7), respectively. The median PFS and median OS of the antibiotic-untreated group were 8.4 months (95% CI: 5.3–15.1) and 18.4 months (95% CI: 6.2–23.1), respectively. The PFS rate in the antibiotic-untreated group was significantly higher than that in the antibiotic-treated group (p = 0.021, log-rank test). There was no significant difference in the OS rate between two groups (p = 0.36, log-rank test).

The severe (≥ grade 3) adverse events observed among the included patients are presented in Table 3. We observed 2 cases of elevated aspartate aminotransferase or alanine aminotransferase levels and one case of hypercalcemia in the antibiotic-treated group. There was no patient in whom the sever elevation of aspartate aminotransferase or alanine aminotransferase levels in the antibiotics-untreated group. The incidence rate of sever elevation of aspartate aminotransferase or alanine aminotransferase levels was not statistically significant between two groups. Further, we observed one case of hypercalcemia in the antibiotics-treated group and 1 case of hyponatremia in the antibiotic-untreated group. The hypercalcemia was not observed in the antibiotics-untreated group and the hyponatremia was not observed in the antibiotics-treated group, respectively. There was no significant difference in the occurrence rate of sever hypercalcemia and hyponatremia between the two groups.

We performed univariate and the multivariate analyses of the PFS and OS. The results of the univariate and the multivariate analyses are presented in Table 4. An Eastern Cooperative Oncology Group performance status (ECOG) PS of 0 or 1 and an NLR ≤4 were associated with a longer PFS. Conversely, a history of antibiotic treatment was associated with a shorter PFS. An ECOG PS of 0 or 1 and NLR ≤4 were associated with a longer OS. A history of antibiotic treatment was not associated with the OS of the participants. Patient sex, age, and the ICI treatment line were not associated with the PFS or OS in the participants in this study.

We retrospectively assessed the influence of antibiotic treatment on the efficacy of an ICI in patients with head and neck SCC. The PFS rate of the antibiotic-treated group in the present study was significantly lower than that of the antibiotic-untreated group (Fig. 1a). Although there was no significant difference, the RR in the antibiotic-treated group was lower than that in the antibiotic-untreated group (p = 0.1518) (Table 2). These findings indicate that antibiotic intake reduces the treatment efficacy of nivolumab in patients with head and neck SCC.

The negative impact of antibiotic treatment on the efficacy of ICIs has been extensively reported in several types of cancers [12, 15, 18‒20]. These previous reports showed that antibiotic intake was associated with a shorter PFS or OS or worse RR in patients with non-small lung cancer, melanoma, and renal cell cancer undergoing ICI therapy. Of these studies, 1 study included patients with head and neck SCC [20]; however, these patients comprised only 2% of the study population [20]. Therefore, it remains unclear if antibiotics influence the treatment efficacy of ICIs in patients with head and neck SCC. The present findings indicate the negative effects of antibiotic intake on the treatment efficacy of nivolumab in patients with head and neck SCC.

It has been reported that the alteration of fecal microbiota may affect the efficacy of ICIs in cholangiocarcinoma patients [21]. Antibiotics are well known to induce intestinal dysbiosis [22, 23]. Furthermore, it has been reported that the alteration of intratumoral microbiota influences the treatment efficacy of gemcitabine, one of the cytotoxic anticancer drugs [24]. Several retrospective data had shown the effect of antibiotics on the treatment efficacy of gemcitabine [25, 26]. The detailed mechanisms on how the antibiotics influence the treatment efficacy of ICI or the anticancer drugs is not yet clear. However, from these previous reports, the alteration of gut microbiota or the intratumor microbiota by the treatment of antibiotics might be an explanation for the negative effect of antibiotics on the treatment efficacy of chemotherapy.

There was no significant difference in the OS of the two groups in the present study. This may be attributed to a significantly high percentage of patients who underwent chemotherapy with paclitaxel following treatment with nivolumab in the antibiotic-treated group as compared to that in the antibiotic-untreated group (11 patients [24.4%] in the antibiotic-treated group vs. 0 patients (0.0%) in the antibiotic-untreated group). A detailed analysis of these 11 cases in the antibiotic-treated group indicated that the RR and median OS following paclitaxel treatment were 63.6% and 10.9 months, respectively (data not shown). A previous observational study reported that administering paclitaxel after nivolumab demonstrated a RR of 70% [27]. Consistent with the findings of this previous study [27], we observed a high RR to treatment with paclitaxel after nivolumab in the present study. The median OS of these 11 patients may have prolonged the median OS of the antibiotic-treated group and reduced the difference in the median OS of both groups.

ECOG PS ≥2 and high NLR have been reported as negative predictive factors for both PFS and OS, following nivolumab treatment for head and neck cancer [28, 29]. The univariate and the multivariate analyses performed in the present study (Table 4) revealed that an ECOG PS ≥2 and NLR >4 were independent negative predictive factors for both short PFS and short OS, which is consistent with previous findings.

As expected, antibiotic treatment was significantly correlated with a short PFS in the present study (Table 4; univariate analysis, p = 0.0247; multivariate analysis, p = 0.025), indicating that antibiotics reduced the PFS with nivolumab therapy. However, interestingly, antibiotic intake was not correlated with a short OS (Table 4; univariate analysis, p = 0.3616; multivariate analysis, p = 0.5455). Previous studies have reported that the use of antibiotics decreased the OS of ICI treatment in several types of cancer [30, 31]. On the other hand, there were also previous reports that the antibiotic use did not change the OS of ICI treatment in several types of cancer, such as the present study [32, 33]. Therefore, it is considered controversial whether antibiotics use worsens the OS by ICI treatment. Further studies are needed to clarify this point.

The present study had several limitations. First, it was a retrospective study with a small number of patients. Second, it did not include data on the combined positive score. A high combined positive score correlates with a high RR, long PFS, and long OS with ICI treatment [34]. This may be a reason why no significant difference in the RR or OS was observed between two groups in this study. Third, the patients in the antibiotic-treated group were treated with multiple types of antibiotics. The fluoroquinolone, one of the antibiotics, has been reported to worsen the prognosis of lung cancer or melanoma patients who were treated with ICI [30, 35]. Only 10 patients were treated with quinolone in this study. This number is too small for statistical analysis to assess the negative effect of quinolone on ICI treatment. Nevertheless, our study suggests that antibiotic use may significantly reduce the PFS associated with nivolumab therapy in patients with head and neck SCC. Multicenter studies with large sample sizes are needed to validate the current findings and to determine the influence of antibiotics on the efficacy of nivolumab by antibiotic type.

This study protocol was reviewed and approved by the Ethics Committee of Tohoku University Hospital, approval number: 2023-1-169. The Ethics Committee of Tohoku University Hospital waived the need for written informed consent of the participants for this retrospective study. Consent was obtained through an opt-out system. All data in the current study had no personal identifiers and were kept confidential.

Chikashi Ishioka received research funding from the Tokyo Cooperative Oncology Group. Chikashi Ishioka has also received contributions from Chugai Pharmaceutical; Novartis Pharma K.K.; Ono Pharmaceutical; MSD; Pfizer; AstraZeneca; Bristol-Myers Squibb; Kyowa Kirin Co., Ltd.; Janssen Pharmaceutical; Taiho Pharmaceutical; Daiichi Sankyo Company, Limited; Takeda Pharmaceutical; Merck Biopharma Co., Ltd.; Eli Lilly Japan K.K; Bayer Yakuhin, Ltd; and Incyte Biosciences Japan G.K. Chikashi Ishioka is a representative of the Tohoku Clinical Oncology Research and Education Society, a nonprofit organization. The authors report no other conflicts of interest in this work.

This study is not funding by any sponsor.

Reio Ueta and Hiroo Imai designed this retrospective study and drafted the manuscript. Chikashi Ishioka is the corresponding author and contributed to data analysis and interpretation and manuscript preparation. All other authors (Ken Saijo, Yoshifumi Kawamura, Shuto Kodera, Keigo Komine, Kota Ouchi, Yuki Kasahara, Sakura Taniguchi, Yuya Yoshida, Keiju Sasaki, Hidekazu Shirota, Masanobu Takahashi) contributed to data collection and interpretation and critical review of the manuscript. All authors (Reio Ueta, Hiroo Imai, Chikashi Ishioka, Ken Saijo, Yoshifumi Kawamura, Shuto Kodera, Keigo Komine, Kota Ouchi, Yuki Kasahara, Sakura Taniguchi, Yuya Yoshida, Keiju Sasaki, Hidekazu Shirota, Masanobu Takahashi) had served as attending physicians of enrolled patients, have read and approved the final version of the manuscript, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.