Critical Reviews in Oncology/Hematology 94 (2015) 136–145
Pancreatitis with vascular endothelial growth factor receptor tyrosine kinase inhibitors Pooja Ghatalia a,1 , Charity J. Morgan b,1 , Toni K. Choueiri c , Pedro Rocha d , Gurudatta Naik e , Guru Sonpavde e,∗ a
e
Department of Internal Medicine, University of Alabama at Birmingham (UAB), Birmingham, AL, USA b Department of Biostatistics, UAB School of Public Health, Birmingham, AL, USA c Dana Farber Cancer Institute and Harvard Medical School, Boston, MA, USA d Department of Medicine, Hospital Del Mar, Barcelona, Spain Department of Internal Medicine, Section of Medical Oncology, UAB Medical Center, Birmingham, AL, USA Received 10 April 2014; received in revised form 31 October 2014; accepted 26 November 2014
Contents 1. 2.
3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Selection of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Data extraction and clinical end points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Statistical analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Search results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Trial quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Population characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Relative risk of all-grade pancreatitis events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Relative risk of high-grade pancreatitis events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Subset analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of interest statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of funding source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relevant disclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Author contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reviewers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix A. Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
137 137 137 137 137 140 140 140 141 141 141 141 141 143 143 143 144 144 144 144 145
Abstract A trial-level meta-analysis was conducted to determine the relative risk (RR) of pancreatitis associated with multi-targeted vascular endothelial growth factor receptor (VEGFR) tyrosine kinase inhibitors (TKI). Eligible studies included randomized phase 2 and 3 trials comparing arms with and without an FDA-approved VEGFR TKI (sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib,
∗ 1
Corresponding author. Tel.: +1 205 975 3742; fax: +1 205 934 9511. E-mail address: [email protected] (G. Sonpavde). Authors have equal contribution in this manuscript.
http://dx.doi.org/10.1016/j.critrevonc.2014.11.008 1040-8428/© 2014 Elsevier Ireland Ltd. All rights reserved.
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
137
regorafenib). Statistical analyses calculated the RR and 95% confidence intervals (CI). A total of 10,578 patients from 16 phase III trials and 6 phase II trials were selected. The RR for all grade and high-grade pancreatitis for the TKI vs. no TKI- arms was 1.95 (p = 0.042, 95% CI: 1.02 to 3.70) and 1.89 (p = 0.069, 95% CI: 0.95 to 373), respectively. No differential impact of malignancy type or specific TKI agent was seen on RR of all grade of high grade pancreatitis. Better patient selection and monitoring may mitigate the risk of severe pancreatitis. © 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Vascular endothelial growth factor receptor; Tyrosine kinase inhibitors; Approved; Pancreatitis; Meta-analysis
1. Introduction Several multi-targeted vascular endothelial growth factor receptor (VEGFR) tyrosine-kinase inhibitors (TKI) agents have been approved by the U.S. FDA including sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib and regorafenib. These agents have improved clinical outcomes in a wide range of malignancies including renal cell carcinoma (RCC), hepatocellular cancer (HCC), gastro-intestinal stromal tumor (GIST), medullary thyroid cancer (MTC), colorectal cancer, chronic myeloid leukemia and pancreatic neuroendocrine tumors. However, VEGFR TKIs are also associated with rare but severe life threatening toxicities, especially cardiovascular events and hemorrhage [1–3]. Several case reports have reported acute pancreatitis as an adverse event associated with sorafenib, and a phase II trial has suggested a high incidence with ponatinib [4–9]. However, the association of VEGFR TKIs with pancreatitis has not been addressed in a systematic manner. In order to determine the risk of pancreatitis associated with all of the currently approved VEGFR TKIs, we performed a meta-analysis of randomized clinical trials (RCT) published or presented in major oncology conferences.
ranking system [10]. Trials that did not list pancreatitis as an adverse event in any arm were excluded. 2.2. Data extraction and clinical end points Data abstraction was conducted independently by 3 investigators (PG, CM, GS) according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [11]. The variables extracted are shown in Table 1. High grade pancreatitis events require inpatient hospitalization for pain, vomiting or nutritional support (grade 3), or are life-threatening (grade 4) or result in persistent or significant disability, and may lead to death (grade 5). According to the Common Terminology Criteria for Adverse Events (CTCAE) v. 4.0, grade 2 pancreatitis consists of radiographic changes of pancreatitis or enzyme elevations without symptoms. However, for our meta-analysis, we excluded events recorded as solitary enzyme elevations, limiting our data to cases recorded as pancreatitis in order to avoid capturing duplicate entries (i.e. recording lipase/amylase elevation and pancreatitis as two separate events for the same patient) leading to over-estimation of pancreatitis. Additionally, enzyme elevations can be non-specific and may not represent clinically relevant pancreatitis. 2.3. Statistical analysis
2. Methods 2.1. Selection of studies An independent review of citations in the English language from PubMed/Medline from January 1966 to December 2013 was conducted. We searched individual VEGFR TKIs: sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib and regorafenib and narrowed the search to RCTs. Abstracts and virtual meeting presentations from major conferences (ASCO, ESMO, AACR), the most recent reports, updated manufacturer’s package inserts and clinicaltrials.gov were also searched. In clinicaltrials.gov, we used the advanced search option to search individual VEGFR TKI, and narrowed the search by entering ‘randomized’ in the interventions option. Phase 2 and 3 RCTs comparing arms with and without a VEGFR TKI were selected. We excluded trials that contained a VEGF inhibitor in all arms. Study quality was assessed by using the seven-point Jadad
Statistical analyses were performed by using R statistical software, version 3.0 [12,13]. Trials were considered evaluable for one or both categories of all grades or high-grade (grade ≥ 3) of pancreatitis based on reporting in the safety profiles of trials. Sub-analyses were performed for risk of pancreatitis based on malignancy type. The proportion of patients with pancreatitis and the 95% confidence intervals (CIs) were derived for each arm of each study and used to calculate the relative risk (RR) of pancreatitis. For studies reporting zero events in an arm, the classic half-integer correction was applied to calculate the RR and variance. For the meta-analysis, both the fixed-effects model and the random-effects model were considered. The latter was calculated with the method of DerSimonian and Laird, which considers both within-study and between-study variation [14]. Statistical heterogeneity among studies included in the meta-analysis was assessed using the Cochrane’s Q statistic, and inconsistency was quantified with the I2 (I-squared) statistic, which is used to describe the percentage
138
Table 1 Characteristics of randomized trials included in the final analysis of risk of pancreatitis with FDA approved and non-approved VEGF receptor TKIs. Author, year
Sunitinib Crown, 2013 [17]
Histology
No. of pts
Treatment arms
Evalu-able pts per arm
Median age (range) years
Median survival (95% CI) Months (unless stated)
Median progression-free survival (95% CI) Months (unless stated)
Median duration of therapy with range
No. of all grade pancreatitis
No. of grade ≥3 pancreatitis
Events
Jadad score
3
Breast cancer
442
Sunitinib 37.5 mg PO daily + Capecitabine
221
52 (27–79)
16.5 (14.5–19.6)
5.4 (4.4–5.8)
1 (0.46%)
0
Pancreatitis
4
Capecitabine
221
54 (31–77)
17.2 (15.5–9.3)
5.5 (4.3–6.8)
0
0
Sunitinib 37.5 mg PO daily + Docetaxel
295
54.3 (31–84)
26.0 (22.5–32.1)
8.2 (7.3–8.6)
1 (0.34%)
1 (0.34%)
Pancreatitis
4
Docetaxel Sunitinib 37.5 mg PO daily + Prednisone Placebo and Prednisone Sunitinib 37.5 mg PO daily Placebo
293 584
55.1 (28–78) 69 (39–90)
28.9 (25.5–33.4) 13.1 (12.0–14.1)
6.9 (6.5–7.3) 24.1 (23.4–28.1)
Sunitinib: 114 days (87–132), Capecitabine: 121 days (98–136) 143 days (120–169) Sunitinib: 26 wk (23–29) Docetaxel: 18 wk (17–21) 18 wk (16–19) 98 days (1–783)
0 1 (0.71%)
0 1 (0.71%)
Acute pancreatitis
5
289
68 (47–86)
11.8 (10.8–14.2)
17.9 (15.8–24.1)
97 days (6–661)
0
0
83
56 (25–84)
20.6 (20.6–NA)
11.4 (7.4–19.8)
1 (1.2%)
1 (1.2%)
57 (26–78)
1 (1.22%)
1 (1.22%)
Pancreatitis, Acute pancreatitis
5
82 238
53 (25–80)
15.3 (12.0–24.7)
4.6 months (0.4–17.5) 3.7 months (0.03–20.2) 61 days (1–485)
1 (0.42%)
1 (0.42%)
Acute pancreatitis
4
240 386
53 (23–80) 59 (25–83)
16.9 (14.5–26.0) 87.9 (75.2–NA)
4.2 (3.8–5.5) 33.6 (30.8–36.6)
61 days (4–540) NR
0 1 (0.26%)
0 1 (0.26%)
Acute pancreatitis
5
382
58 (25–82)
85.9 (81.1–NA)
36.6 (32.9–40)
NR
0
0
113
52 (32–81)
9.4 (5.8–11.2)
1.7 (1.5–2.6)
54 days (8–594)
1 (0.91%)
1 (0.91%)
Pancreatitis
4
104
52 (31–81)
10.5 (8.5–13.8)
1.7 (1.5–2.6)
0
0
375
62 (27–87)
Capecitabine: 56 days (1–294) Gemcitabine: 4 days (1–28) Vinorelbine: 8 days (2–13) Docetaxel: 4 days (2–5) Paclitaxel: 6 days (3–8) 6 months (1–15)
5 (1%)
5 (1%)
Pancreatitis
4
375 228
59 (34–85) 58 (23–84)
94.9 wk (77.7–117) 72.7 wk (61.3–83.0)
4 months (1–13) 56 days (1–236)
1 0
1 0
Acute Pancreaitits
5
114
55 (23–81)
64.9 wk (45.7–96)
29.5 days (2–168)
1 (0.88%)
1 (0.88%)
Bergh, 2012 [18]
3
Breast cancer
Michaelson, 2013 [19]
3
Prostate cancer 873
Raymond, 2011 [20]
3
593
Pancreatic 165 neuroendocrine tumors
Barrios, 2010 [21]
3
Breast cancer
478
Carrato, 2013 [22]
3
Colorectal cancer
768
Curigliano, 2013 [23]
2
Breast cancer
217
Motzer, 2007 [24]
3
Renal Cell carcinoma
750
Demetri, 2006 [25]
3
GIST
342
Sunitinib 37.5 mg PO daily Capecitabine Sunitinib 37.5 mg PO daily + FOLFIRI Placebo + FOLFIRI Sunitinib 37.5 mg PO dialy Standard of care
Sunitinib 37.5 mg PO daily Interferon-alfa Sunitinib 50 mg PO daily Placebo
NR
5.5 (3.6–7.4) 2.8 (2.4–4)
114.6 wk (100.1–142.9) 47.7 wk (46.3 to 58.1) 22.1 wk (16.7–27.4) 22.9 wk (10.9–28) 6.0 wk (4.4–9.7)
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
Phase
Sorafenib Scagliotti, 2010 [26]
3
NSCLC
926
Lencioni, 2012 [27]
2
HCC
307
Paz-Ares, 2012 [28]
3
NSCLC
901
Moehler, 2011 [30]
McDermott, 2008 [31]
3
2
2
HCC
Biliary tract cancer
Melanoma
599
97
101
Brose, 2011 [32]
3
Thyroid cancer
417
Escudier, 2009 [33]
3
Renal cell cancer
902
Pazopanib Scagliotti, 2013 [34]
Vandetanib Gridelli, 2012 [35]
2
2
NSCLC
NSCLC
103
124
4
0
Pancreatitis
4
2 (1.32%) 1 (0.22%)
2 (1.32%) 1 (0.22%)
Pancreatitis
5
0
0
a
0
Pancreatitis
5
Pancreatitis
4
Pancreatitis
5
62 (34–86)
324 days (277–423) 139 days (132–160) 16.6 wk (0.1–69.9)
1 (0.22%)
1 (0.22%)
462
63 (34–82)
322 days (293–366) 139 days (132–160) 17.9 wk (0.1–75)
0
0
154
NA
NA days (554–NA) NA
0
153 452
NA 60 (28–81)
Placebo + Gemcitabine/Cisplatin
449
59 (22–82)
Sorafenib 400 mg PO bid Placebo
297
64.9 (NR)
302
66.3 (NR)
Sorafenib 400 mg PO bid + Gemcitabine Placebo + Gemcitabine Sorafenib 400 mg PO bid + Dacarbazine Placebo + Dacarbazine Sorafenib 400 mg PO bid Placebo Sorafenib 400 mg PO bid Placebo
49
NR
NA days (562–NA) NA 6.3 months 376 days (333–416) 183 days (168–208) Sorafenib: 17 wk Gemcitabine: 10.3 wk Cisplatin: 9.4 wk 379 days (335–414) 168 days (156–174) Sorafenib: 8 wk Gemcitabine: 10.4 wk Cisplatin: 9.6 wk 10.7 (9.4–13.3) NR 5.3 months (0.2–16.1) 7.9 (6.8–9.1) NR 4.3 months (0.1–16.6) 255 days (166–384)91 days (54–218) 64 days (1–330)
48
NR
51
b
Pazopanib 600 mg daily + Pemetrexed b Pazopanib 800 mg + Pemetrexed Cisplatin + Pemetrexed Vandetanib 100 mg PO daily + Gemcitabine Placebo + Gemcitabine
4.8 months
1 (0.34%)
0
0
1 (2.04%)
1 (2.04%)
341 days (306–488) 148 days (105–233) 66 days (8–375)
0
0
60 (18–88)
319 days (245–492) 148 days (112–196) 19.1 wk (3.1–71.1)
1 (1.96%)
1 (1.96%)
50
55 (31–82)
359 days (285–504)82 days (43–125) 12.1 wk (3–52)
0
0
207
61.5
NR
1 (0.48%)
1 (0.48%)
Pancreatitis
4
210 451
62 58.0 (19–86)
NR 175 days (160–238) NR 542 days (500–598) 167 days (139–174) 23 wk
0 a 3 (0.66%)
0 a 2 (0.44%)
Pancreatitis
5
451
59 (29–84)
461 days (408–526)84 days (78–91)
12 wk
a
0
8
60.4
NR
NR
0
0
Pancreatitis
3
61
60.8
NR
25.0 wk (17.3–34.1) NR
1 (1.64%)
1 (1.64%)
34
61.8
NR
22.9 wk (18.4–27.7) NR
0
0
61
75.03
262 days (170–356) 183 days (116 to 214)NR
1 (1.64%)
1 (1.64%)
Pancreatitis
5
63
75.48
305 days (214–355) 169 (95 to 194)
0
0
329 days (278–393) NR
NR
NR
1 (0.22%)
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
LLovet, 2008 [29]
Pancreatitis
464
Sorafenib 400 mg PO bid + Carboplatin/Paclitaxel Placebo + Carboplatin/Paclitaxel Sorafenib 400 mg PO bid + TACE Placebo + TACE Sorafenib 200 mg PO bid + Gemcitabine/Cisplain
139
140
– – – – Total
10583 All tumors 2 or 3
VEGFR TKI No VEGFR TKI
5569 5014
– –
18.9 wk (16.2 to 22.6) 35.8 wk (30.0–44.8) 62 (35–89) 309
FOLFIRI: Irinotecan + levo-leucovorin + 5-FU; GIST: gastro-intestinal stromal tumor; HCC: hepatocellular carcinoma; NR: not reported; RCC: renal cell carcinoma; NSCLC: non-small cell lung cancer; Pts: patients; TACE: transarterial chemoembolization; wk: weeks. a See the package insert for updated events. b The two arms were combined for analysis.
3–5 22 6
1 (0.32%) 1 (0.32%)
25 7
0 0
2.8 months (0.03–11) 2.3 months (0.03–11.1) 19.1 wk (17.1 to 24.4) 36.9 wk (30.1–41.1) 61 (34–84) 304 613 Pancreatic cancer 3 Axitinib Kindler, 2011 [38]
331 3 Wells, 2012 [37]
Medullary thyroid cancer
Axitinib 5 mg PO bid + Gemcitabine Placebo + Gemcitabine
0 0 39.9 wk 19.2 NR 53.4 (26–84) 100
Any
Pancreatitis
5
5 Pancreatitis 1 (0.43%) 1 (0.43%) 90.1 wk NR 50.7 (18–83) 231
30.5
0 0 85 days 11.9 wk (11.4 to 15.9) 35–83 278
10.5 (8.9–12.3) 60 (28–82) 534 3 de Boer, 2011 [36]
NSCLC
Vandetanib 100 mg PO daily + Pemetrexed Placebo + Pemetrexed Vandetanib 300 mg PO daily Placebo
256
9.2 (7.1–12)
5 Pancreatitis 1 (0.38%) 102 days
1 (0.38%)
Jadad score No. of grade ≥3 pancreatitis No. of pts Histology Phase Author, year
Table 1 (Continued)
Treatment arms
Evalu-able pts per arm
Median age (range) years
Median survival (95% CI) Months (unless stated)
Median progression-free survival (95% CI) Months (unless stated) 17.6 wk (13.4 to 18.9)
Median duration of therapy with range
No. of all grade pancreatitis
Events
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
of total variation across studies that is due to heterogeneity rather than chance; a value of 0% indicates no observed heterogeneity, while larger values between 0% and 100% show increasing heterogeneity [15]. The assumption of homogeneity was considered invalid for p-values 0.20 using the Egger test) (Supplementary Fig. 1). 3.6. Subset analysis We conducted a meta regression analysis to examine whether RR of pancreatitis varied by malignancy type. We found that there was no significant effect of histology on the RR for either all grade or high grade pancreatitis (p > 0.2, for both). We conducted a subgroup analysis comparing trials that used VEGFR-TKI alone to those that combined VEGFRTKI with another therapy in the non-placebo arm. There was no significant effect of the addition of another therapy on the risk of either all-grade or high-grade pancreatitis (p > 0.20 for both). Additionally, there was no significant effect of specific TKI agents on the risk of either all grade or high grade pancreatitis (p > 0.20 for both).
Fig. 1. Selection process for trials included in the meta-analysis.
4. Discussion generally adequate for each trial and included a period of approximately 2–4 weeks after end of therapy on trial. The trials were all high quality (Jadad score 3–5).
3.3. Population characteristics A total of 10,578 patients were available for the metaanalysis, 5569 in the TKI group and 5009 in the control group. Patients were generally required to have an Eastern Cooperative Oncology Group (ECOG) performance status 0–1, adequate organ function and no brain metastasis. All studies were eligible for all and high-grade pancreatitis analysis.
To the best of our knowledge, this is the first and largest meta-analysis evaluating the risk of pancreatitis associated with all U.S. FDA approved VEGFR TKIs as of December 2013. In this comprehensive analysis of 10,578 patients, 22 randomized phase II and III trials using approved VEGFR TKIs (sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib, regorafenib) were included. We did not include phase I trials in our meta-analysis since these studies are not randomized and include a wide range of different dosages of drugs. In addition, we only included events reporting pancreatitis and did not consider amylase and lipase elevations as significant events for this meta-analysis. We observed a significant 1.95-fold increase (p = 0.042, 95%
142
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
Fig. 2. VEGFR TKIs were associated with a significantly higher risk for all-grade pancreatitis compared to no TKIs (RR = 1.95, p = 0.042, 95% CI 1.02 to 3.70). There was no evidence of heterogeneity (Q = 7.76, p = 0.996, I2 = 0%).
CI 1.02 to 3.70) in the risk of all grades of pancreatitis with VEGFR TKIs compared to controls not receiving TKIs. However, the risk of severe pancreatitis (grade ≥3) demonstrated a trend that was not statistically significant (RR = 1.89, p = 0.069, 95% CI 0.95 to 373). Inhibition of the VEGF pathway has been shown to increase fatal adverse events in several studies [3,39]. Several case reports and a phase I trial [40] report pancreatitis as an adverse event with the use of VEGFR TKIs. In these, however, sorafenib was reported as a culprit. The results of our meta-analysis imply that the observed adverse event can be generalized to other TKIs. Additionally, given the variety of malignancies in our meta-analysis, the observed effect can likely be generalized across a variety of patients receiving VEGFR TKIs. VEGF may play a role in regulating the cell cycle of acinar cells and maintaining adult pancreatic fenestrated capillaries. The etiology of TKI-induced acute pancreatitis is not well characterized, but possible explanations discussed in the above case reports include inhibition of VEGF leading to acinar cell apoptosis and release of autodigestive enzymes. Another mechanism proposed is decrease in gastrointestinal motility causing reflux of duodenal contents into
the pancreatic duct, activating pancreatic enzymes [6]. Conversely, bevacizumab and aflibercept, monoclonal antibodies that target VEGF have not been clearly associated with pancreatitis, suggesting that pancreatitis may be an off-target toxicity of VEGFR TKIs. Axitinib is a TKI that is relatively specific for VEGFR. However, as we identified only one trial evaluating axitinib, we were unable to perform a subanalysis evaluating therapy with axitinib or other relatively non-specific TKIs. Despite the size of this meta-analysis, our study has some limitations. First, confounding variables at the patient level, such as comorbidities, younger age and previous chemotherapeutic exposure could not be incorporated into the analysis. Second, we did not include trials recording enzyme elevation alone, even though this is considered a grade 2 adverse event according to CTCAE v 4.0. Acute pancreatitis is diagnosed clinically with at least 2 of the following 3 features: symptoms of severe abdominal pain, imaging characteristics and enzyme elevation. Including cases with enzyme elevation alone would lead to an over-estimation of cases with acute pancreatitis as transient lipase elevation may not always be clinically significant. Third, we did control for time on trial therapy, although a longer time on the TKI
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
143
Fig. 3. Subjects in the VEGFR TKI group were at higher risk of severe grade pancreatitis than subjects in the non-TKI group, but this difference did not reach statistical significance (RR = 1.89, p = 0.069, 95% CI 0.95 to 373). There was no evidence of heterogeneity (Q = 7.84, p = 0.998, I2 = 0%).
arm may be hypothesized to slightly increase the risk of pancreatitis merely attributable to longer duration of recording events; however, we believe such an analysis will not alter the overall results Finally, despite our search for trials investigating 8 VEGFR TKIs, we were only able to locate pancreatitis reported in trials with 5 of these agents (sunitinib, sorafenib, pazopanib, axitinib and vandetanib). This may be due to relatively fewer studies with the newer TKIs, regorafenib, ponatinib and cabozantinib. However, clinical pancreatitis and/or amylase/lipase elevations are a reported adverse event in package inserts of these drugs. There were no evaluable randomized trials of ponatinib, and the cabozantinib and regorafenib trials did not list pancreatitis as adverse events. Nevertheless, meta-analyses are considered reasonable to study rare events that cannot be comprehensively studied in prospective trials. We did not have access to patient level data, but studies suggest that trial level and patient level meta-analyses yield similar results [41]. In conclusion, the use of small-molecule VEGFR TKIs is associated with an increase in the risk of developing all grades of pancreatitis events in a broad range of malignancies regardless of the spectrum of other kinases inhibited. Acute pancreatitis should be considered in the differential diagnosis of patients receiving these drugs who present with symptoms of abdominal pain. Closer monitoring may be prudent in
those at higher risk of pancreatitis, e.g. those with a history of alcoholism or biliary stones. For patients with several comorbidities, the risks may outweigh the benefits, should they develop pancreatitis. Indeed, the lack of pancreatitis reported in RCTs of the 3 newer approved VEGFR TKIs (regorafenib, ponatinib, cabozantinib) suggests that patient selection and monitoring may have already improved. Nevertheless, these approved drugs do improve overall clinical outcomes, the absolute incidence of this serious adverse event is low (
Pancreatitis with vascular endothelial growth factor receptor tyrosine kinase inhibitors Pooja Ghatalia a,1 , Charity J. Morgan b,1 , Toni K. Choueiri c , Pedro Rocha d , Gurudatta Naik e , Guru Sonpavde e,∗ a
e
Department of Internal Medicine, University of Alabama at Birmingham (UAB), Birmingham, AL, USA b Department of Biostatistics, UAB School of Public Health, Birmingham, AL, USA c Dana Farber Cancer Institute and Harvard Medical School, Boston, MA, USA d Department of Medicine, Hospital Del Mar, Barcelona, Spain Department of Internal Medicine, Section of Medical Oncology, UAB Medical Center, Birmingham, AL, USA Received 10 April 2014; received in revised form 31 October 2014; accepted 26 November 2014
Contents 1. 2.
3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Selection of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Data extraction and clinical end points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Statistical analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Search results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Trial quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Population characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Relative risk of all-grade pancreatitis events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Relative risk of high-grade pancreatitis events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Subset analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conflict of interest statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Role of funding source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relevant disclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Author contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reviewers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix A. Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
137 137 137 137 137 140 140 140 141 141 141 141 141 143 143 143 144 144 144 144 145
Abstract A trial-level meta-analysis was conducted to determine the relative risk (RR) of pancreatitis associated with multi-targeted vascular endothelial growth factor receptor (VEGFR) tyrosine kinase inhibitors (TKI). Eligible studies included randomized phase 2 and 3 trials comparing arms with and without an FDA-approved VEGFR TKI (sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib,
∗ 1
Corresponding author. Tel.: +1 205 975 3742; fax: +1 205 934 9511. E-mail address: [email protected] (G. Sonpavde). Authors have equal contribution in this manuscript.
http://dx.doi.org/10.1016/j.critrevonc.2014.11.008 1040-8428/© 2014 Elsevier Ireland Ltd. All rights reserved.
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
137
regorafenib). Statistical analyses calculated the RR and 95% confidence intervals (CI). A total of 10,578 patients from 16 phase III trials and 6 phase II trials were selected. The RR for all grade and high-grade pancreatitis for the TKI vs. no TKI- arms was 1.95 (p = 0.042, 95% CI: 1.02 to 3.70) and 1.89 (p = 0.069, 95% CI: 0.95 to 373), respectively. No differential impact of malignancy type or specific TKI agent was seen on RR of all grade of high grade pancreatitis. Better patient selection and monitoring may mitigate the risk of severe pancreatitis. © 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Vascular endothelial growth factor receptor; Tyrosine kinase inhibitors; Approved; Pancreatitis; Meta-analysis
1. Introduction Several multi-targeted vascular endothelial growth factor receptor (VEGFR) tyrosine-kinase inhibitors (TKI) agents have been approved by the U.S. FDA including sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib and regorafenib. These agents have improved clinical outcomes in a wide range of malignancies including renal cell carcinoma (RCC), hepatocellular cancer (HCC), gastro-intestinal stromal tumor (GIST), medullary thyroid cancer (MTC), colorectal cancer, chronic myeloid leukemia and pancreatic neuroendocrine tumors. However, VEGFR TKIs are also associated with rare but severe life threatening toxicities, especially cardiovascular events and hemorrhage [1–3]. Several case reports have reported acute pancreatitis as an adverse event associated with sorafenib, and a phase II trial has suggested a high incidence with ponatinib [4–9]. However, the association of VEGFR TKIs with pancreatitis has not been addressed in a systematic manner. In order to determine the risk of pancreatitis associated with all of the currently approved VEGFR TKIs, we performed a meta-analysis of randomized clinical trials (RCT) published or presented in major oncology conferences.
ranking system [10]. Trials that did not list pancreatitis as an adverse event in any arm were excluded. 2.2. Data extraction and clinical end points Data abstraction was conducted independently by 3 investigators (PG, CM, GS) according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [11]. The variables extracted are shown in Table 1. High grade pancreatitis events require inpatient hospitalization for pain, vomiting or nutritional support (grade 3), or are life-threatening (grade 4) or result in persistent or significant disability, and may lead to death (grade 5). According to the Common Terminology Criteria for Adverse Events (CTCAE) v. 4.0, grade 2 pancreatitis consists of radiographic changes of pancreatitis or enzyme elevations without symptoms. However, for our meta-analysis, we excluded events recorded as solitary enzyme elevations, limiting our data to cases recorded as pancreatitis in order to avoid capturing duplicate entries (i.e. recording lipase/amylase elevation and pancreatitis as two separate events for the same patient) leading to over-estimation of pancreatitis. Additionally, enzyme elevations can be non-specific and may not represent clinically relevant pancreatitis. 2.3. Statistical analysis
2. Methods 2.1. Selection of studies An independent review of citations in the English language from PubMed/Medline from January 1966 to December 2013 was conducted. We searched individual VEGFR TKIs: sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib and regorafenib and narrowed the search to RCTs. Abstracts and virtual meeting presentations from major conferences (ASCO, ESMO, AACR), the most recent reports, updated manufacturer’s package inserts and clinicaltrials.gov were also searched. In clinicaltrials.gov, we used the advanced search option to search individual VEGFR TKI, and narrowed the search by entering ‘randomized’ in the interventions option. Phase 2 and 3 RCTs comparing arms with and without a VEGFR TKI were selected. We excluded trials that contained a VEGF inhibitor in all arms. Study quality was assessed by using the seven-point Jadad
Statistical analyses were performed by using R statistical software, version 3.0 [12,13]. Trials were considered evaluable for one or both categories of all grades or high-grade (grade ≥ 3) of pancreatitis based on reporting in the safety profiles of trials. Sub-analyses were performed for risk of pancreatitis based on malignancy type. The proportion of patients with pancreatitis and the 95% confidence intervals (CIs) were derived for each arm of each study and used to calculate the relative risk (RR) of pancreatitis. For studies reporting zero events in an arm, the classic half-integer correction was applied to calculate the RR and variance. For the meta-analysis, both the fixed-effects model and the random-effects model were considered. The latter was calculated with the method of DerSimonian and Laird, which considers both within-study and between-study variation [14]. Statistical heterogeneity among studies included in the meta-analysis was assessed using the Cochrane’s Q statistic, and inconsistency was quantified with the I2 (I-squared) statistic, which is used to describe the percentage
138
Table 1 Characteristics of randomized trials included in the final analysis of risk of pancreatitis with FDA approved and non-approved VEGF receptor TKIs. Author, year
Sunitinib Crown, 2013 [17]
Histology
No. of pts
Treatment arms
Evalu-able pts per arm
Median age (range) years
Median survival (95% CI) Months (unless stated)
Median progression-free survival (95% CI) Months (unless stated)
Median duration of therapy with range
No. of all grade pancreatitis
No. of grade ≥3 pancreatitis
Events
Jadad score
3
Breast cancer
442
Sunitinib 37.5 mg PO daily + Capecitabine
221
52 (27–79)
16.5 (14.5–19.6)
5.4 (4.4–5.8)
1 (0.46%)
0
Pancreatitis
4
Capecitabine
221
54 (31–77)
17.2 (15.5–9.3)
5.5 (4.3–6.8)
0
0
Sunitinib 37.5 mg PO daily + Docetaxel
295
54.3 (31–84)
26.0 (22.5–32.1)
8.2 (7.3–8.6)
1 (0.34%)
1 (0.34%)
Pancreatitis
4
Docetaxel Sunitinib 37.5 mg PO daily + Prednisone Placebo and Prednisone Sunitinib 37.5 mg PO daily Placebo
293 584
55.1 (28–78) 69 (39–90)
28.9 (25.5–33.4) 13.1 (12.0–14.1)
6.9 (6.5–7.3) 24.1 (23.4–28.1)
Sunitinib: 114 days (87–132), Capecitabine: 121 days (98–136) 143 days (120–169) Sunitinib: 26 wk (23–29) Docetaxel: 18 wk (17–21) 18 wk (16–19) 98 days (1–783)
0 1 (0.71%)
0 1 (0.71%)
Acute pancreatitis
5
289
68 (47–86)
11.8 (10.8–14.2)
17.9 (15.8–24.1)
97 days (6–661)
0
0
83
56 (25–84)
20.6 (20.6–NA)
11.4 (7.4–19.8)
1 (1.2%)
1 (1.2%)
57 (26–78)
1 (1.22%)
1 (1.22%)
Pancreatitis, Acute pancreatitis
5
82 238
53 (25–80)
15.3 (12.0–24.7)
4.6 months (0.4–17.5) 3.7 months (0.03–20.2) 61 days (1–485)
1 (0.42%)
1 (0.42%)
Acute pancreatitis
4
240 386
53 (23–80) 59 (25–83)
16.9 (14.5–26.0) 87.9 (75.2–NA)
4.2 (3.8–5.5) 33.6 (30.8–36.6)
61 days (4–540) NR
0 1 (0.26%)
0 1 (0.26%)
Acute pancreatitis
5
382
58 (25–82)
85.9 (81.1–NA)
36.6 (32.9–40)
NR
0
0
113
52 (32–81)
9.4 (5.8–11.2)
1.7 (1.5–2.6)
54 days (8–594)
1 (0.91%)
1 (0.91%)
Pancreatitis
4
104
52 (31–81)
10.5 (8.5–13.8)
1.7 (1.5–2.6)
0
0
375
62 (27–87)
Capecitabine: 56 days (1–294) Gemcitabine: 4 days (1–28) Vinorelbine: 8 days (2–13) Docetaxel: 4 days (2–5) Paclitaxel: 6 days (3–8) 6 months (1–15)
5 (1%)
5 (1%)
Pancreatitis
4
375 228
59 (34–85) 58 (23–84)
94.9 wk (77.7–117) 72.7 wk (61.3–83.0)
4 months (1–13) 56 days (1–236)
1 0
1 0
Acute Pancreaitits
5
114
55 (23–81)
64.9 wk (45.7–96)
29.5 days (2–168)
1 (0.88%)
1 (0.88%)
Bergh, 2012 [18]
3
Breast cancer
Michaelson, 2013 [19]
3
Prostate cancer 873
Raymond, 2011 [20]
3
593
Pancreatic 165 neuroendocrine tumors
Barrios, 2010 [21]
3
Breast cancer
478
Carrato, 2013 [22]
3
Colorectal cancer
768
Curigliano, 2013 [23]
2
Breast cancer
217
Motzer, 2007 [24]
3
Renal Cell carcinoma
750
Demetri, 2006 [25]
3
GIST
342
Sunitinib 37.5 mg PO daily Capecitabine Sunitinib 37.5 mg PO daily + FOLFIRI Placebo + FOLFIRI Sunitinib 37.5 mg PO dialy Standard of care
Sunitinib 37.5 mg PO daily Interferon-alfa Sunitinib 50 mg PO daily Placebo
NR
5.5 (3.6–7.4) 2.8 (2.4–4)
114.6 wk (100.1–142.9) 47.7 wk (46.3 to 58.1) 22.1 wk (16.7–27.4) 22.9 wk (10.9–28) 6.0 wk (4.4–9.7)
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
Phase
Sorafenib Scagliotti, 2010 [26]
3
NSCLC
926
Lencioni, 2012 [27]
2
HCC
307
Paz-Ares, 2012 [28]
3
NSCLC
901
Moehler, 2011 [30]
McDermott, 2008 [31]
3
2
2
HCC
Biliary tract cancer
Melanoma
599
97
101
Brose, 2011 [32]
3
Thyroid cancer
417
Escudier, 2009 [33]
3
Renal cell cancer
902
Pazopanib Scagliotti, 2013 [34]
Vandetanib Gridelli, 2012 [35]
2
2
NSCLC
NSCLC
103
124
4
0
Pancreatitis
4
2 (1.32%) 1 (0.22%)
2 (1.32%) 1 (0.22%)
Pancreatitis
5
0
0
a
0
Pancreatitis
5
Pancreatitis
4
Pancreatitis
5
62 (34–86)
324 days (277–423) 139 days (132–160) 16.6 wk (0.1–69.9)
1 (0.22%)
1 (0.22%)
462
63 (34–82)
322 days (293–366) 139 days (132–160) 17.9 wk (0.1–75)
0
0
154
NA
NA days (554–NA) NA
0
153 452
NA 60 (28–81)
Placebo + Gemcitabine/Cisplatin
449
59 (22–82)
Sorafenib 400 mg PO bid Placebo
297
64.9 (NR)
302
66.3 (NR)
Sorafenib 400 mg PO bid + Gemcitabine Placebo + Gemcitabine Sorafenib 400 mg PO bid + Dacarbazine Placebo + Dacarbazine Sorafenib 400 mg PO bid Placebo Sorafenib 400 mg PO bid Placebo
49
NR
NA days (562–NA) NA 6.3 months 376 days (333–416) 183 days (168–208) Sorafenib: 17 wk Gemcitabine: 10.3 wk Cisplatin: 9.4 wk 379 days (335–414) 168 days (156–174) Sorafenib: 8 wk Gemcitabine: 10.4 wk Cisplatin: 9.6 wk 10.7 (9.4–13.3) NR 5.3 months (0.2–16.1) 7.9 (6.8–9.1) NR 4.3 months (0.1–16.6) 255 days (166–384)91 days (54–218) 64 days (1–330)
48
NR
51
b
Pazopanib 600 mg daily + Pemetrexed b Pazopanib 800 mg + Pemetrexed Cisplatin + Pemetrexed Vandetanib 100 mg PO daily + Gemcitabine Placebo + Gemcitabine
4.8 months
1 (0.34%)
0
0
1 (2.04%)
1 (2.04%)
341 days (306–488) 148 days (105–233) 66 days (8–375)
0
0
60 (18–88)
319 days (245–492) 148 days (112–196) 19.1 wk (3.1–71.1)
1 (1.96%)
1 (1.96%)
50
55 (31–82)
359 days (285–504)82 days (43–125) 12.1 wk (3–52)
0
0
207
61.5
NR
1 (0.48%)
1 (0.48%)
Pancreatitis
4
210 451
62 58.0 (19–86)
NR 175 days (160–238) NR 542 days (500–598) 167 days (139–174) 23 wk
0 a 3 (0.66%)
0 a 2 (0.44%)
Pancreatitis
5
451
59 (29–84)
461 days (408–526)84 days (78–91)
12 wk
a
0
8
60.4
NR
NR
0
0
Pancreatitis
3
61
60.8
NR
25.0 wk (17.3–34.1) NR
1 (1.64%)
1 (1.64%)
34
61.8
NR
22.9 wk (18.4–27.7) NR
0
0
61
75.03
262 days (170–356) 183 days (116 to 214)NR
1 (1.64%)
1 (1.64%)
Pancreatitis
5
63
75.48
305 days (214–355) 169 (95 to 194)
0
0
329 days (278–393) NR
NR
NR
1 (0.22%)
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
LLovet, 2008 [29]
Pancreatitis
464
Sorafenib 400 mg PO bid + Carboplatin/Paclitaxel Placebo + Carboplatin/Paclitaxel Sorafenib 400 mg PO bid + TACE Placebo + TACE Sorafenib 200 mg PO bid + Gemcitabine/Cisplain
139
140
– – – – Total
10583 All tumors 2 or 3
VEGFR TKI No VEGFR TKI
5569 5014
– –
18.9 wk (16.2 to 22.6) 35.8 wk (30.0–44.8) 62 (35–89) 309
FOLFIRI: Irinotecan + levo-leucovorin + 5-FU; GIST: gastro-intestinal stromal tumor; HCC: hepatocellular carcinoma; NR: not reported; RCC: renal cell carcinoma; NSCLC: non-small cell lung cancer; Pts: patients; TACE: transarterial chemoembolization; wk: weeks. a See the package insert for updated events. b The two arms were combined for analysis.
3–5 22 6
1 (0.32%) 1 (0.32%)
25 7
0 0
2.8 months (0.03–11) 2.3 months (0.03–11.1) 19.1 wk (17.1 to 24.4) 36.9 wk (30.1–41.1) 61 (34–84) 304 613 Pancreatic cancer 3 Axitinib Kindler, 2011 [38]
331 3 Wells, 2012 [37]
Medullary thyroid cancer
Axitinib 5 mg PO bid + Gemcitabine Placebo + Gemcitabine
0 0 39.9 wk 19.2 NR 53.4 (26–84) 100
Any
Pancreatitis
5
5 Pancreatitis 1 (0.43%) 1 (0.43%) 90.1 wk NR 50.7 (18–83) 231
30.5
0 0 85 days 11.9 wk (11.4 to 15.9) 35–83 278
10.5 (8.9–12.3) 60 (28–82) 534 3 de Boer, 2011 [36]
NSCLC
Vandetanib 100 mg PO daily + Pemetrexed Placebo + Pemetrexed Vandetanib 300 mg PO daily Placebo
256
9.2 (7.1–12)
5 Pancreatitis 1 (0.38%) 102 days
1 (0.38%)
Jadad score No. of grade ≥3 pancreatitis No. of pts Histology Phase Author, year
Table 1 (Continued)
Treatment arms
Evalu-able pts per arm
Median age (range) years
Median survival (95% CI) Months (unless stated)
Median progression-free survival (95% CI) Months (unless stated) 17.6 wk (13.4 to 18.9)
Median duration of therapy with range
No. of all grade pancreatitis
Events
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
of total variation across studies that is due to heterogeneity rather than chance; a value of 0% indicates no observed heterogeneity, while larger values between 0% and 100% show increasing heterogeneity [15]. The assumption of homogeneity was considered invalid for p-values 0.20 using the Egger test) (Supplementary Fig. 1). 3.6. Subset analysis We conducted a meta regression analysis to examine whether RR of pancreatitis varied by malignancy type. We found that there was no significant effect of histology on the RR for either all grade or high grade pancreatitis (p > 0.2, for both). We conducted a subgroup analysis comparing trials that used VEGFR-TKI alone to those that combined VEGFRTKI with another therapy in the non-placebo arm. There was no significant effect of the addition of another therapy on the risk of either all-grade or high-grade pancreatitis (p > 0.20 for both). Additionally, there was no significant effect of specific TKI agents on the risk of either all grade or high grade pancreatitis (p > 0.20 for both).
Fig. 1. Selection process for trials included in the meta-analysis.
4. Discussion generally adequate for each trial and included a period of approximately 2–4 weeks after end of therapy on trial. The trials were all high quality (Jadad score 3–5).
3.3. Population characteristics A total of 10,578 patients were available for the metaanalysis, 5569 in the TKI group and 5009 in the control group. Patients were generally required to have an Eastern Cooperative Oncology Group (ECOG) performance status 0–1, adequate organ function and no brain metastasis. All studies were eligible for all and high-grade pancreatitis analysis.
To the best of our knowledge, this is the first and largest meta-analysis evaluating the risk of pancreatitis associated with all U.S. FDA approved VEGFR TKIs as of December 2013. In this comprehensive analysis of 10,578 patients, 22 randomized phase II and III trials using approved VEGFR TKIs (sunitinib, sorafenib, pazopanib, axitinib, vandetanib, cabozantinib, ponatinib, regorafenib) were included. We did not include phase I trials in our meta-analysis since these studies are not randomized and include a wide range of different dosages of drugs. In addition, we only included events reporting pancreatitis and did not consider amylase and lipase elevations as significant events for this meta-analysis. We observed a significant 1.95-fold increase (p = 0.042, 95%
142
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
Fig. 2. VEGFR TKIs were associated with a significantly higher risk for all-grade pancreatitis compared to no TKIs (RR = 1.95, p = 0.042, 95% CI 1.02 to 3.70). There was no evidence of heterogeneity (Q = 7.76, p = 0.996, I2 = 0%).
CI 1.02 to 3.70) in the risk of all grades of pancreatitis with VEGFR TKIs compared to controls not receiving TKIs. However, the risk of severe pancreatitis (grade ≥3) demonstrated a trend that was not statistically significant (RR = 1.89, p = 0.069, 95% CI 0.95 to 373). Inhibition of the VEGF pathway has been shown to increase fatal adverse events in several studies [3,39]. Several case reports and a phase I trial [40] report pancreatitis as an adverse event with the use of VEGFR TKIs. In these, however, sorafenib was reported as a culprit. The results of our meta-analysis imply that the observed adverse event can be generalized to other TKIs. Additionally, given the variety of malignancies in our meta-analysis, the observed effect can likely be generalized across a variety of patients receiving VEGFR TKIs. VEGF may play a role in regulating the cell cycle of acinar cells and maintaining adult pancreatic fenestrated capillaries. The etiology of TKI-induced acute pancreatitis is not well characterized, but possible explanations discussed in the above case reports include inhibition of VEGF leading to acinar cell apoptosis and release of autodigestive enzymes. Another mechanism proposed is decrease in gastrointestinal motility causing reflux of duodenal contents into
the pancreatic duct, activating pancreatic enzymes [6]. Conversely, bevacizumab and aflibercept, monoclonal antibodies that target VEGF have not been clearly associated with pancreatitis, suggesting that pancreatitis may be an off-target toxicity of VEGFR TKIs. Axitinib is a TKI that is relatively specific for VEGFR. However, as we identified only one trial evaluating axitinib, we were unable to perform a subanalysis evaluating therapy with axitinib or other relatively non-specific TKIs. Despite the size of this meta-analysis, our study has some limitations. First, confounding variables at the patient level, such as comorbidities, younger age and previous chemotherapeutic exposure could not be incorporated into the analysis. Second, we did not include trials recording enzyme elevation alone, even though this is considered a grade 2 adverse event according to CTCAE v 4.0. Acute pancreatitis is diagnosed clinically with at least 2 of the following 3 features: symptoms of severe abdominal pain, imaging characteristics and enzyme elevation. Including cases with enzyme elevation alone would lead to an over-estimation of cases with acute pancreatitis as transient lipase elevation may not always be clinically significant. Third, we did control for time on trial therapy, although a longer time on the TKI
P. Ghatalia et al. / Critical Reviews in Oncology/Hematology 94 (2015) 136–145
143
Fig. 3. Subjects in the VEGFR TKI group were at higher risk of severe grade pancreatitis than subjects in the non-TKI group, but this difference did not reach statistical significance (RR = 1.89, p = 0.069, 95% CI 0.95 to 373). There was no evidence of heterogeneity (Q = 7.84, p = 0.998, I2 = 0%).
arm may be hypothesized to slightly increase the risk of pancreatitis merely attributable to longer duration of recording events; however, we believe such an analysis will not alter the overall results Finally, despite our search for trials investigating 8 VEGFR TKIs, we were only able to locate pancreatitis reported in trials with 5 of these agents (sunitinib, sorafenib, pazopanib, axitinib and vandetanib). This may be due to relatively fewer studies with the newer TKIs, regorafenib, ponatinib and cabozantinib. However, clinical pancreatitis and/or amylase/lipase elevations are a reported adverse event in package inserts of these drugs. There were no evaluable randomized trials of ponatinib, and the cabozantinib and regorafenib trials did not list pancreatitis as adverse events. Nevertheless, meta-analyses are considered reasonable to study rare events that cannot be comprehensively studied in prospective trials. We did not have access to patient level data, but studies suggest that trial level and patient level meta-analyses yield similar results [41]. In conclusion, the use of small-molecule VEGFR TKIs is associated with an increase in the risk of developing all grades of pancreatitis events in a broad range of malignancies regardless of the spectrum of other kinases inhibited. Acute pancreatitis should be considered in the differential diagnosis of patients receiving these drugs who present with symptoms of abdominal pain. Closer monitoring may be prudent in
those at higher risk of pancreatitis, e.g. those with a history of alcoholism or biliary stones. For patients with several comorbidities, the risks may outweigh the benefits, should they develop pancreatitis. Indeed, the lack of pancreatitis reported in RCTs of the 3 newer approved VEGFR TKIs (regorafenib, ponatinib, cabozantinib) suggests that patient selection and monitoring may have already improved. Nevertheless, these approved drugs do improve overall clinical outcomes, the absolute incidence of this serious adverse event is low (
