2025 Volume 145 Issue 1 Pages 71-78
The relationship between the concomitant use of immune checkpoint inhibitors (ICIs) and elevated prothrombin time-to-international standard ratio (PT-INR) in patients receiving warfarin remains unclear. In the present study, 26 patients treated with ICIs during warfarin therapy were examined for increases in PT-INR within 60 d of ICI administration. Of these patients, 13 developed Grade 2 or higher PT-INR elevations, 5 of which required the immediate administration of vitamin K. The increased risk of bleeding and the impact on the continuation of cancer drug therapy are significant burdens for patients. Immune-related adverse events caused by ICIs have been suggested as one of the reasons for increases in PT-INR, and patients taking warfarin and ICIs need to be managed in consideration of the risk of elevated PT-INR by frequently checking the blood coagulation capacity.
The incidence of thromboembolism in cancer patients is increasing as advances in cancer treatment and new anticancer agents prolong the prognosis of cancer patients.1) In the acute phase of the onset of venous thromboembolism, treatment may be initiated with heparins, while the vitamin K antagonist oral coagulant warfarin may be used in the maintenance and chronic phases of treatment. Since sensitivity to warfarin not only varies among individuals, but also depends on environmental factors, such as pathophysiology and diet, and sometimes causes serious side effects, including bleeding, strict dosage control is required to ensure safe and appropriate anticoagulation therapy based on an evaluation by the prothrombin time-to-international standard ratio (PT-INR).2) Drug interactions are one of the factors contributing to elevated PT-INR and have been reported for some anticancer agents.3) However, limited information is available on the causal relationship with immune checkpoint inhibitors (ICIs), which have been used with increasing frequency in recent years and their effects on PT-INR are unknown. ICIs are therapeutic drugs with a mechanism of action that reactivates the function of immune cells, mainly T lymphocytes, by releasing the brake caused by cancer cells, enabling them to attack cancer cells.4) ICIs are considered to be less toxic than cytotoxic anticancer agents; however, immune-related adverse events (irAEs) may occur.5) irAEs attributed to excessive autoimmune reactions include a wide range of skin disorders, lung disorders, liver, bile, and pancreatic disorders, gastrointestinal disorders, and endocrine disorders, while blood disorders include thrombocytopenia and anemia. The relationship between elevated PT-INR and concomitant ICIs in patients receiving warfarin remains unclear. Therefore, the present study investigated whether PT-INR was elevated in patients receiving concomitant ICIs while being treated with warfarin and attempted to identify the factors contributing to this variation.
This was a retrospective single-arm cohort study conducted between April 2013 and March 2023 at the Medical Research Institute KITANO HOSPITAL, Yodogawa Christian Hospital, Sumitomo Hospital, Osaka International Cancer Center, and Osaka Red Cross Hospital. Thirty-four cancer patients were extracted who were taking warfarin at the start of the first dose of ICIs (pembrolizumab, nivolumab, atezolizumab, and ipilimumab) and had no significant PT-INR changes in the month before starting the combination. Among them, 26 patients were included after excluding those with inadequate PT-INR extraction before or during concomitant therapy and those whose warfarin continuation status or dosage was unknown (Fig. 1). None of the patients in the target group were taking concomitant medications that were contraindicated to potentiate the anticoagulant effects of warfarin. Note that all of the cancer patients in this study were administered ICIs for active cancer.
Sex, age, ICI type, number of ICIs administered, line of therapy, cancer type, purpose of warfarin use, warfarin dose at the start of concomitant therapy, PT-INR, aspartate aminotransferase (AST), alanine aminotransferase (ALT), serum albumin (ALB), total protein (TP), serum creatinine (sCr), the estimated glomerular filtration rate (eGFR), platelets (PL), tetraiodothyronine/thyroxine (T4), and thyroid stimulating hormone (TSH) were retrospectively collected from electronic medical records. The PT-INR/warfarin dose ratio (PT-INR/dose), corrected for changes in PT-INR before and after the combination and the warfarin dose, was also investigated. PT-INR elevations were evaluated based on the Common Terminology Criteria for Adverse Events (CTCAE) ver 5.0. Grade 1 was defined as a 1.2- to 1.5-fold increase, Grade 2 as a 1.5- to 2.5-fold increase, and Grade 3 as a 2.5-fold or greater increase. Warfarin dosages and PT-INR were observed for up to 60 d after the start of ICI administration, and values for each laboratory test were taken immediately before the initiation of ICIs. In post-ICI laboratory tests, values were taken at the time of the highest increase in PT-INR after ICI administration.
Statistical AnalysisThe Wilcoxon signed-rank test was used to compare each survey item before and after concomitant ICIs. A t-test was used when there was normality in the distribution of values in each group, and the Mann–Whitney U test was employed to test for differences between the 2 groups for a 2-group comparison of each survey item between the PT-INR Grade 0–1 and Grade 2–3 groups. Fisher’s exact test was used to test for frequency bias. The level of significance was set at p<0.05, and statistical analyses were performed using EZR6) (Saitama Medical Center, Jichi Medical University, Japan), a graphical user interface of R (The R Foundation for Statistical Computing, Vienna, Austria, version 3.5.2). More precisely, it is an improvement of R commander (version 2.5-1), designed to add statistical functions frequently used in biostatistics.
Ethical ConsiderationsThe present study was conducted in compliance with the “Ethical Guidelines for Life Science and Medical Research Involving Human Subjects” and was approved by the Ethical Review Committee of our hospital (P230201200).
Patient backgrounds are shown in Table 1. The analysis included 21 males and 5 females. The overall mean±S.D. of age was 72.0±5.6 y, and the ICIs used were nivolumab monotherapy in 20 patients and nivolumab plus ipilimumab, pembrolizumab alone, and atezolizumab alone in 2 patients each. The median [range] number of ICIs administered during the observation period was 3 [2–5], and the median [range] dose of warfarin before starting the combination was 1.54 mg [1.6–3.0]. It is important to note that patients in the present study were not limited to any specific chemotherapy. Therefore, patients with a wide range of primary tumor organs and various types of ICIs were included. Furthermore, no significant differences were observed in patient backgrounds between Grade 0–1 group and Grade 2–3 group.
| Total | Grade 0–1 | Grade 2–3 | p Value | |
|---|---|---|---|---|
| Sex, Male/Female | 21/5 | 10/3 | 11/2 | |
| Age, years | 72.0±5.6 | 73.9±5.2 | 70.0±5.4 | 0.07 |
| ICI | ||||
| Nivolumab | 20 | 8 | 12 | 0.16 |
| Nivolumab+Ipilimumab | 2 | 1 | 1 | 1.00 |
| Pembrolizumab | 2 | 2 | 0 | 0.48 |
| Atezolizumab | 2 | 2 | 0 | 0.48 |
| Total number of ICI doses during the period | ||||
| Nivolumab | 3.0 [2–5] | 3.0 [2.8–4.0] | 3.5 [1.8–5.0] | 0.69 |
| Nivolumab+Ipilimumab | 2.0 | 2.0 | 2.0 | |
| Pembrolizumab | 2.5 [2.3–2.8] | 2.5 [2.3–2.8] | 0 | |
| Atezolizumab | 3.5 [3.3–3.8] | 3.5 [3.3–3.8] | 0 | |
| Treatment line | ||||
| 1st | 4 | 3 | 1 | 0.59 |
| 2nd | 8 | 5 | 3 | 0.67 |
| ≥3rd | 14 | 5 | 9 | 0.24 |
| Primary tumor location | ||||
| Lung | 16 | 8 | 8 | 1.00 |
| Stomach | 5 | 2 | 3 | 1.00 |
| Liver | 1 | 1 | 0 | 1.00 |
| Kidney | 1 | 0 | 1 | 1.00 |
| Urinary tract | 1 | 1 | 0 | 1.00 |
| Others | 2 | 1 | 1 | 1.00 |
| Indication for Warfarin therapy | ||||
| Atrial fibrillation | 7 | 4 | 3 | 1.00 |
| Venous thrombosis | 6 | 2 | 4 | 0.65 |
| Cerebral thrombosis | 5 | 4 | 1 | 0.32 |
| Myocardial infarction | 2 | 0 | 2 | 0.48 |
| Aortic valve replacement | 2 | 1 | 1 | 1.00 |
| Pulmonary artery thrombosis | 1 | 0 | 1 | 1.00 |
| Others | 3 | 2 | 1 | 1.00 |
| Warfarin dose (mg/d) | 2.0 [1.6–3.0] | 2.0 [1.5–2.5] | 2.5 [2.0–3.0] | 0.25 |
| PT-INR | 1.54 [1.24–2.10] | 1.76 [1.54–2.12] | 1.24 [1.15–2.02] | 0.09 |
| PT-INR/dose | 0.78 [0.59–1.02] | 0.85 [0.71–1.03] | 0.62 [0.39–0.87] | 0.09 |
| AST (IU/L) | 23 [19–27] | 23 [18–25] | 23 [20–28] | 0.68 |
| ALT (IU/L) | 14 [10–20] | 13 [10–20] | 14 [13–18] | 0.50 |
| ALB (g/dL) | 3.5 [3.1–3.8] | 3.2 [3.1–3.8] | 3.5 [3.0–3.7] | 0.94 |
| TP (g/dL) | 6.8 [6.3–7.2] | 6.6 [6.3–7.2] | 6.9 [6.3–7.3] | 0.71 |
| sCr (mg/dL) | 1.06 [0.76–1.48] | 1.05 [0.79–1.32] | 1.07 [0.73–1.98] | 0.66 |
| eGFR (mL/min/1.73 m2) | 52.1 [33.1–75.8] | 54.6 [37.0–70.0] | 51.0 [27.0–80.2] | 0.70 |
| PLT (×104/µL) | 23.6 [16.9–30.2] | 20.9 [14.0–25.3] | 23.9 [22.7–31.0] | 0.15 |
| T4 (ng/dL) | 1.10 [0.85–1.20] | 1.00 [0.88–1.18] | 1.10 [0.88–1.19] | 0.97 |
| TSH (μIU/mL) | 2.42 [1.82–9.44] | 2.07 [1.30–2.70] | 2.58 [2.10–27.0] | 0.17 |
Date are represented as the mean±S.D. or median [range]. ICI: Immune checkpoint inhibitor, PT-INR: prothrombin time international normalized ratio, AST: aspartate aminotransferase, ALT: alanine aminotransferase, ALB: serum albumin, TP: total protein, sCr: serum creatinine, eGFR: estimated glomerular filtration rate, PLT: platelets, T4: tetraiodothyronine/thyroxine, TSH: thyroid-stimulating hormone.
The grade classification of PT-INR elevations in the present study with or without adjustments for warfarin and with or without the administration of vitamin K after PT-INR elevations are shown (Table 2). Of the 26 patients analyzed, 21 (81%) had elevated PT-INR after concomitant ICIs. Eight patients were classified as Grade 1, 5 as Grade 2, and 8 as Grade 3. Warfarin was reduced or discontinued in 12 patients (46%) and vitamin K was administered to 5 (19%). Laboratory parameters before and after the combination are shown (Table 3). There were no significant differences in liver function, renal function, Alb, or platelet counts; a comparison of PT-INR before and after concomitant ICIs showed a significant increase after ICIs (pre-combination [range] vs. post-combination [range]: 1.54 [1.24–2.10] vs. 2.49 [1.98–4.73], p<0.00001) (Fig. 2). PT-INR/dose also significantly increased after concomitant ICIs (0.78 [0.59–1.02] vs. 1.07 [0.85–2.32], p<0.00001) (Fig. 3).
| Outcomes | Patients (n=26) |
|---|---|
| Elevations in PT-INR, n (%) | 21 (81) |
| Grade 1, n (%) | 8 (31) |
| Grade 2, n (%) | 5 (19) |
| Grade 3, n (%) | 8 (31) |
| Discontinuation or dose reduction of warfarin, n (%) | 12 (46) |
| Vitamin K rescue, n (%) | 5 (19) |
ICI: Immune checkpoint inhibitor, PT-INR: prothrombin time international normalized ratio.
| Before | After | p Value | |
|---|---|---|---|
| AST (U/L) | 23 [19–27] | 22 [19–28] | 0.98 |
| ALT (U/L) | 14 [10–20] | 14 [10–18] | 0.96 |
| ALB (g/dL) | 3.5 [3.1–3.8] | 3.4 [3.0–3.9] | 0.21 |
| TP (g/dL) | 6.8 [6.3–7.2] | 6.7 [6.4–7.4] | 0.91 |
| sCr (mg/dL) | 1.06 [0.76–1.48] | 1.06 [0.82–1.54] | 0.06 |
| eGFR (mL/min/1.73 m2) | 52.1 [33.1–75.8] | 49.0 [35.5–69.6] | 0.10 |
| PLT (×104/µL) | 23.6 [16.9–30.2] | 23.6 [16.9–27.5] | 0.90 |
| PT-INR | 1.54 [1.24–2.10] | 2.49 [1.98–4.73] | 0.00001* |
Wilcoxon rank sum test. ICI: Immune checkpoint inhibitor, AST: aspartate aminotransferase, ALT: alanine aminotransferase, ALB: serum albumin, TP: total protein, sCr: serum creatinine, eGFR: estimated glomerular filtration rate, PLT: platelets, PT-INR: prothrombin time international normalized ratio, * p<0.00001.
PT-INR: Prothrombin time international normalized ratio, ICI: immune checkpoint inhibitor. *Wilcoxon signed-rank test, * p<0.00001.
PT-INR: Prothrombin time international normalized ratio, ICI: immune checkpoint inhibitor. *Wilcoxon signed-rank test, * p<0.00001.
Table 4 shows the concomitant medications of concern for drug–drug interactions in warfarin. 6 patients (55%) in the Grade 2–3 group used levothyroxine sodium hydrate, in contrast to none in the Grade 0–1 group (p<0.05). No significant differences were observed in the use of other drugs. The timing of initiation of drugs that may interact with warfarin and the date of onset of PT-INR elevation and T4/TSH variation were examined in detail for each case (Table 5). In 2 out of 6 cases (Nos. 16, 17) who received concomitant levothyroxine sodium had been receiving it prior to the start of ICI. In 4 out of 6 cases (Nos. 14, 20, 21, 23) who received concomitant levothyroxine sodium had it added after the ICI start date. In 3 out of 6 cases (Nos. 16, 17, 21) showed improvement in T4/TSH, while 3 out of 6 cases (Nos. 14, 20, 23) showed no improvement in T4/TSH.
| Total | Grade 0–1 | Grade 2–3 | p Value | |
|---|---|---|---|---|
| Concomitant medications | ||||
| Acetaminophen | 4 | 0 | 4 | 0.10 |
| Celecoxib | 1 | 0 | 1 | 1.00 |
| Naproxen | 2 | 0 | 2 | 0.48 |
| Amiodarone | 1 | 1 | 0 | 1.00 |
| Fluvastatin | 1 | 1 | 0 | 1.00 |
| Rosuvastatin | 2 | 1 | 1 | 1.00 |
| Levothyroxine sodium hydrate | 6 | 0 | 6 | 0.01* |
| Aspirin | 3 | 3 | 0 | 0.22 |
| Clarithromycin | 1 | 0 | 1 | 1.00 |
| Clopidogrel Sulfate | 1 | 0 | 1 | 1.00 |
| Levofloxacin | 1 | 0 | 1 | 1.00 |
| Other area-specific drugs | ||||
| Stomach medicine | 18 | 9 | 9 | 1.00 |
| Anti-hypertensive agent | 14 | 5 | 9 | 0.24 |
| Diuretic | 9 | 4 | 5 | 1.00 |
| Non-Steroidal Anti-Inflammatory Drugs | 7 | 1 | 6 | 0.07 |
| Dyslipidemia drugs | 7 | 3 | 4 | 1.00 |
| Hypoglycemic drug | 7 | 4 | 3 | 1.00 |
| Opioid analgesics | 4 | 0 | 4 | 0.10 |
| Antibiotic | 3 | 1 | 2 | 1.00 |
PT-INR: Prothrombin time international normalized ratio. *Fisher’s exact test, * p<0.05.
| No. | ICI treatment regimen | Number of ICI doses during the period | PT-INR increased Grade | Development day | Vitamin K rescue | Discontinuation or dose reduction of warfarin | Before T4/TSH | After T4/TSH | Drugs used concomitantly before the start of ICI | Drugs added after ICI start date |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Nivolumab | 2 | 0 | — | 1.1/1.82 | 1.0/5.26 | — | — | ||
| 2 | Nivolumab+Ipilimumab | 2 | 0 | — | 1.4/0.99 | 1.2/1.29 | Amiodarone, Aspirin | — | ||
| 3 | Nivolumab | 3 | 0 | — | 1.2/2.06 | 1.3/1.56 | — | — | ||
| 4 | Nivolumab | 3 | 0 | — | 0.95/4.64 | — | — | — | ||
| 5 | Atezolizumab | 3 | 0 | — | 0.84/2.49 | 0.91/2.85 | — | — | ||
| 6 | Nivolumab | 1 | 1 | 7 | 〇 | 1.0/2.08 | — | Aspirin | — | |
| 7 | Pembrolizumab | 3 | 1 | 18 | 0.81/12.3 | 0.88/6.24 | — | — | ||
| 8 | Atezolizumab | 4 | 1 | 42 | 1.2/1.13 | 1.1/3.48 | — | — | ||
| 9 | Nivolumab | 4 | 1 | 28 | 1.3/1.13 | 1.3/2.39 | Fluvastatin, Aspirin | — | ||
| 10 | Nivolumab | 3 | 1 | 42 | 1.0/1.11 | 1.1/1.19 | — | — | ||
| 11 | Nivolumab | 5 | 1 | 53 | 0.85/2.76 | 0.87/2.93 | — | — | ||
| 12 | Nivolumab | 4 | 1 | 4 | — | 1.1/4.33 | Rosuvastatin | — | ||
| 13 | Pembrolizumab | 2 | 1 | 45 | — | 0.69/0.98 | — | — | ||
| 14 | Nivolumab | 3 | 2 | 14 | 〇 | 0.92/11.0 | 1.1/12.3 | — | Levothyroxine sodium hydrate (day 1–) | |
| 15 | Nivolumab | 5 | 2 | 35 | 〇 | — | 1.0/1.00 | Acetaminophen | — | |
| 16 | Nivolumab | 5 | 2 | 16 | 1.1/2.16 | 0.9/0.64 | Celecoxib, Levothyroxine sodium hydrate, Clarithromycin | — | ||
| 17 | Nivolumab | 5 | 2 | 56 | 0.8/40.6 | 0.9/14.1 | Levothyroxine sodium hydrate | — | ||
| 18 | Nivolumab | 3 | 2 | 15 | 〇 | 1.3/9.44 | — | Clarithromycin | — | |
| 19 | Nivolumab | 1 | 3 | 8 | 〇 | 〇 | 0.82/0.33 | — | Acetaminophen, Naproxen | — |
| 20 | Nivolumab | 5 | 3 | 60 | 〇 | 1.17/23.6 | 1.1/36.0 | Acetaminophen | Levothyroxine sodium hydrate (day 2–), Levofloxacin (day 34–) | |
| 21 | Nivolumab | 1 | 3 | 15 | 〇 | 〇 | 0.83/30.4 | 1.0/10.7 | — | Levothyroxine sodium hydrate (day 1–) |
| 22 | Nivolumab | 5 | 3 | 29 | 〇 | 〇 | 1.45/2.42 | 1.51/2.56 | Naproxen | — |
| 23 | Nivolumab | 1 | 3 | 24 | 〇 | 0.68/1.24 | 0.51/0.38 | — | Levothyroxine sodium hydrate (day 17–) | |
| 24 | Nivolumab+Ipilimumab | 2 | 3 | 35 | 〇 | 1.1/2.58 | 1.0/4.46 | — | — | |
| 25 | Nivolumab | 2 | 3 | 21 | 〇 | 〇 | 1.1/0.54 | — | Acetaminophen | — |
| 26 | Nivolumab | 4 | 3 | 15 | 〇 | 〇 | 1.2/2.04 | 1.14/1.48 | Rosuvastatin | — |
ICI: Immune checkpoint inhibitor, PT-INR: prothrombin time international normalized ratio, T4: tetraiodothyronine/thyroxine, TSH: thyroid-stimulating hormone.
The present results suggest that the concomitant use of ICIs increased PT-INR in warfarin-treated patients with stable PT-INR control. In addition, 50% of patients had Grade 2–3 PT-INR elevations, approximately 46% required a reduction in or the discontinuation of warfarin, and 19% required the immediate administration of vitamin K. In patients with Grade 2 or higher PT-INR elevations, ICIs were more common with nivolumab monotherapy and in combination with ipilimumab, and these ICIs required more attention; however, the number of cases was small. Patients with elevated PT-INR in the present study did not have markedly prolonged PT-INR before each ICI administration; however, since it remains unclear whether a direct interaction occurs between ICIs and warfarin, this needs to be considered when they are used concomitantly. Limited information is available on increases in PT-INR with concomitant use; therefore, we herein discussed the factors involved, including hepatic dysfunction, a decreased serum ALB concentration, decreased thyroid function, vitamin K intake, and drug–drug interactions.
The first factor was the possible effect of liver dysfunction caused by ICIs. The liver produces vitamin K-dependent coagulation factors II, VII, IX, and X.7) Warfarin is considered to exert its anticoagulant effects by indirectly decreasing the production of these coagulation factors through its inhibition of vitamin K metabolism. When hepatic dysfunction occurs, the production of coagulation factors is reduced, which may enhance the effects of warfarin by promoting bleeding. On the other hand, ICIs exert anti-tumor effects by enhancing the activity of the immune system; however, irAEs occur when the immune system is impaired.5) The ONO-4538-06 study showed that 1.8% of patients with unresectable advanced or recurrent non-small cell lung cancer who received nivolumab developed hepatobiliary system disorders (Opdivo® Intravenous Infusion Interview Form, Ono Pharmaceutical Co.). Among the study subjects, one patient had Grade 3 hepatic dysfunction diagnosed as irAE due to nivolumab, which may have resulted in an increase in PT-INR. In addition, another patient had abnormal liver enzymes at the time of the PT-INR elevation; however, it was unclear whether liver dysfunction was due to irAE. CYP3A4 and CYP2C9 are involved in the metabolism of warfarin, and the activities of these enzymes are decreased in cirrhosis and other conditions.8,9) Increases in PT-INR may be due to a decreased ability to metabolize warfarin due to ICI-related hepatic dysfunction. In addition, genetic variations in CYP2C9 are associated with warfarin dosages.9) In the present study, CYP2C9 was not identified in patients; therefore, its relevance is unknown. On the other hand, since elevated PT-INR was observed even in patients without hepatic dysfunction, other factors were also considered.
We propose that a decrease in serum ALB concentrations caused increases in PT-INR. The liver produces ALB, and serum ALB levels may be reduced in patients with hepatic dysfunction. Warfarin has a high plasma protein-binding rate, with 1–10% of warfarin in the blood being in a pharmacologically active free form and 90–99% being in a pharmacologically inactive bound form to ALB. The former antagonizes the effects of vitamin K and its anticoagulant effects are induced by inhibiting the production of vitamin K-dependent coagulation factors in hepatocytes; therefore, warfarin is more likely to exert anticoagulant effects when serum ALB levels are reduced. Furthermore, a previous study reported that low serum ALB levels increased the risk of bleeding.10) Since no significant differences were observed in serum ALB concentrations between pre- and post-combination blood tests performed in the present study, we considered the possibility of an increase in PT-INR due to a decrease in serum ALB concentrations to be low. On the other hand, serum ALB concentrations may decrease due to anorexia caused by anticancer drug treatment. Therefore, in patients with anorexia or hepatic dysfunction, serum ALB concentrations need to be monitored.
Thyroid hormones are involved in blood coagulation factors, such as prothrombin and factors VIII and X. Fluctuations in thyroid function have been shown to affect anticoagulants.11) As indicated above, warfarin exerts its anticoagulant effects by inhibiting the production of vitamin K-dependent coagulation factors in hepatocytes; however, under hypothyroid conditions, catabolic metabolism is considered to decrease along with the production of vitamin K-dependent coagulation factors, resulting in the weakened effects of warfarin. If PT-INR is controlled within an appropriate range in an hypothyroid state, the initiation of treatment with thyroid preparations and allowing thyroid function to return to normal may potentiate the effects of warfarin and result in excessive anticoagulant effects.12) This is listed as a concomitant use precaution in the Warfarin package insert and other documents (Warfarin® Tablets Attachment, Eisai Inc., revised July 2019). Improvement of T4/TSH was observed in 3 out of 6 cases (Nos. 16, 17, 21) treated with levothyroxine sodium in this study. This was suggested to have contributed to the increase in PT-INR due to the effect of normalization of thyroid function. On the other hand, in 3 out of 6 cases (Nos. 14, 20, 23) did not have normalization of thyroid function, resulting in an increase in PT-INR. Therefore, these cases were considered to be due to different factors. Previous studies reported that the interaction between levothyroxine sodium and warfarin did not increase the risk of bleeding and, thus, opinions are divided.13) Regarding variables of thyroid function, the regular monitoring of FT4 and TSH is necessary because of concerns about the impact of ICI administration on the development of irAEs.
A relationship with vitamin K intake is also discussed. Since vitamin K is closely associated with the production of blood coagulation factors, warfarin drug efficacy is affected by changes in dietary vitamin K intake.14) However, due to the lack of studies on the effects of ICI administration on vitamin K biosynthesis, we assumed that it was not directly involved. Nevertheless, vitamin K intake needs to be considered in patients taking warfarin. The possibility that vitamin K deficiency may have contributed to the elevation of PT-INR cannot be ruled out because the present study did not assess dietary intake before and after the combination.
Other factors will also be discussed. Of the 13 cases with Grade 2–3 PT-INR elevations, all but 1 case (No. 24) were taking concomitant medications that interacted with warfarin other than levothyroxine sodium. In order to minimize the influence of warfarin and interacting drugs as much as possible in this study, 34 cancer patients were selected who had no significant PT-INR change in the month prior to the start of concomitant ICI. However, the effects of interactions cannot be completely ruled out. Therefore, it is necessary in the future to study the effects of warfarin in a larger population without concomitant use of drugs that may interact with warfarin. Other possible reasons for the increase in PT-INR were the effects of decreased renal function; however, no significant decrease was observed before and after concomitant use. Moreover, patients did not use over-the-counter drugs during the period of PT-INR elevation. There were also no cases of long-term antibiotic administration or gastrointestinal symptoms, such as diarrhea; however, the possible involvement of a disorder of the intestinal microflora cannot be ruled out. Although the potential for medication errors was inferred from interviews with patients and residual medication, no information to suggest an overdose due to medication errors was obtained. PT-INR/dose was significantly increased after concomitant ICIs in this study. On the other hand, we cannot rule out the possibility that the significant increase was influenced in some cases. For a more comprehensive analysis, a comparison of the incidence of increased PT-INR/dose before and after ICI administration should be considered in the future. In addition, the population selected for this study had a variety of diseases for which warfarin was intended to be administered, so individual PT-INR target ranges exist. Therefore, it was necessary to confirm whether or not there was any deviation from the target range with the use of ICIs. However, it was difficult to obtain information on the target range for each individual patient, and this study was limited in its ability to extract such information. In the future, we should also investigate the relationship between the target range of PT-INR and the timing of ICIs administration when the PT-INR deviates from the target range.
ICI-induced irAE include acquired hemophilia as well as thyroid dysfunction.15–17) It has been suggested that acquired coagulation factor V deficiency may be caused by irAE.18) This suggests that one of the possible mechanisms of PT-INR elevation is a deficiency of multiple coagulation factors due to irAE. In some cases, elevated PT-INR was observed as early as 8 d after the concomitant use of ICIs; therefore, it is necessary to monitor the onset of PT-INR elevations from a relatively early stage. On the other hand, there were cases in which PT-INR elevations were observed after a certain period of time following concomitant therapy. Basically, the uncertainty of when irAEs will occur is the same as their uncertainty. Therefore, the risks associated with concomitant use for more than 60 d need to continue to be considered. In a case report of a patient who intentionally overdosed on warfarin, the intravenous administration of vitamin K resulted in rebound even when a temporary improvement in PT-INR was achieved. When an increase in PT-INR is observed with the combination of ICIs and warfarin, changes in PT-INR need to be monitored continuously after a single intravenous infusion of vitamin K.
In conclusion, the present study suggests that the concomitant use of ICIs induces an increase in PT-INR in warfarin-treated patients with stable PT-INR control. PT-INR may have increased due to liver damage, the concomitant use of levothyroxine, and abnormal thyroid function caused by ICIs. Since elevated PT-INR may increase the risk of major bleeding or intracerebral hemorrhage, the concomitant use of ICIs during warfarin administration needs to be carefully implemented, and frequent checks of liver function, thyroid function, and coagulability as well as the concomitant use of levothyroxine need to be performed in advance to detect the risk of increased PT-INR and initiate appropriate measures. Therefore, it is necessary to detect and respond to the risk of PT-INR elevations at an early stage by frequently checking liver function, thyroid function, and the blood coagulation capacity and also by confirming in advance whether levothyroxine is used concomitantly.
The authors declare no conflict of interest.
