3.1 Case Series
Twenty-five cases of ONJ induced by TKIs and four cases of ONJ induced by ICIs (Table 2) were identified in the study. For the TKI-related ONJ cases, the median patient age was 63 years, the median time to onset was 12.8 months, 68.2% of patients were male (missing data were excluded), and the most common malignancy involved (65.64%) was renal cell carcinoma. Furthermore, sunitinib was implicated in 21 (84.0%) patients, lenvatinib was implicated in 2 (8.0%) patients, and cabozantinib and axitinib were each implicated in 1 (4.0%) patient. For the ICI-related ONJ cases, the median patient age was 53 years, the median time to onset was 10.2 months, two patients were male and two were female, and all cases had advanced melanoma. Ipilimumab (CTLA4 inhibitor) was implicated for two (50.0%) of the patients, nivolumab and pembrolizumab were implicated for one (25.0%) patient each, as shown in Table 2.
Table 2. Summary of case reports of immune checkpoint inhibitor-induced pure red cell anaemia reported in the literature.
|
Age
|
Sex
|
Bisphosphonate
|
Malignancy
|
Targeted Agent
|
Duration (month)
|
1[24]
|
65
|
M
|
Yes
|
RCC
|
Sunitinib
|
13
|
2[24]
|
67
|
M
|
Yes
|
RCC
|
Sunitinib
|
13
|
3[25]
|
59
|
M
|
No
|
RCC
|
Sunitinib
|
12
|
4[13]
|
58
|
F
|
No
|
RCC
|
Sunitinib
|
12
|
5[26]
|
55
|
-
|
Yes
|
RCC
|
Sunitinib
|
9
|
6[26]
|
56
|
-
|
Yes
|
RCC
|
Sunitinib
|
4
|
7[26]
|
62
|
-
|
Yes
|
RCC
|
Sunitinib
|
23
|
8[27]
|
59
|
M
|
Yes
|
RCC
|
Sunitinib
|
4
|
9[28]
|
19
|
F
|
No
|
RCC
|
Sunitinib
|
-
|
10[28]
|
64
|
F
|
No
|
RCC
|
Sunitinib
|
48
|
11[29]
|
51
|
M
|
Yes
|
RCC
|
Sunitinib
|
22
|
12[30]
|
62
|
M
|
Yes
|
RCC
|
Sunitinib
|
18
|
13[31]
|
42
|
F
|
Yes
|
Gastric Cancer
|
Sunitinib
|
-
|
14[32]
|
53
|
M
|
Yes
|
RCC
|
Sunitinib
|
5
|
15[33]
|
51
|
F
|
No
|
MTC
|
Cabozantinib
|
3
|
16[34]
|
70
|
M
|
No
|
RCC
|
Sunitinib
|
9
|
17[35]
|
58
|
M
|
No
|
Thyroid Cancer
|
Lenvatinib
|
12
|
18[36]
|
61
|
F
|
No
|
Thyroid Cancer
|
Lenvatinib
|
-
|
19[37]
|
55
|
F
|
Yes
|
RCC
|
Sunitinib
|
16
|
20[38]
|
67
|
M
|
No
|
RCC
|
Axitinib
|
2
|
21[39]
|
48
|
M
|
Yes
|
RCC
|
Sunitinib
|
3
|
22[40]
|
48
|
M
|
Yes
|
RCC
|
Sunitinib
|
14
|
23[40]
|
57
|
M
|
Yes
|
RCC
|
Sunitinib
|
7
|
24[40]
|
29
|
M
|
Yes
|
RCC
|
Sunitinib
|
4
|
25[41]
|
63
|
M
|
Yes
|
RCC
|
Sunitinib
|
30
|
26[42]
|
58
|
F
|
No
|
Melanoma Cells
|
Ipilimumab
|
36
|
27[43]
|
52
|
M
|
No
|
Melanoma Cells
|
Ipilimumab
|
2.2
|
28[44]
|
75
|
M
|
No
|
Melanoma Cells
|
Nivolumab
|
1.5
|
29[45]
|
28
|
F
|
No
|
Melanoma Cells
|
Pembrolizumab
|
1
|
3.2 Demographic Characteristics
From January 2004 to December 2022, the FAERS database contained 19,494,698 reports, of which 16,134,686 reports were retained after duplication records were excluded. Overall, there were 90,324 ICSRs (sunitinib: 21,840; cabozantinib: 4,324; lenvatinib: 13,802; axitinib: 12,697, pazopanib: 23,265; sorafenib: 14396) suspected to be related to TKIs and 10,3123 ICSRs (pembrolizumab: 32,010; nivolumab: 57,291; avelumab: 1,560; ipilimumab: 12,262) suspected to be related to ICIs. Notably, 240 ICSRs were related to ONJ AEs. Specifically, 77 (32.08%) ICSRs were linked to sunitinib, 40 (16.67%) to lenvatinib, 31 (12.92%) to pazopanib, 30 (12.50%) to nivolumab, 24 (10.00%) to axitinib, 13 (5.42%) to sorafenib, 12 (5.00%) to pembrolizumab, 10 (4.17%) to cabozantinib, and 3 (1.25%) to ipilimumab.
In terms of the ONJ-related ICSRs for TKIs, the highest number of ICSRs occurred in 2022. Specifically, reports related to axitinib and cabozantinib were collected since 2014, and the highest number of ICSRs occurred in 2022. For pazopanib, sunitinib, and sorafenib, ONJ AE reports were collected since 2010, and the highest numbers of ICSRs occurred in 2015, 2013, and 2011, respectively. For lenvatinib, ONJ AEs were collected since 2016, and the highest number of ICSRs occurred in 2022. In general, most cases were reported from North America and Europe. The median age of patients from all the TKI-related ICSRs was 63 years (IQR, 56–70 years; n = 143). More than half of the ICSRs related to male patients (63.1%). It is interesting that renal cancer was the most common indication for all the TKIs (65.64%), followed by thyroid cancer (10.26%). For all TKI-related ONJ cases documented in ICSRs (excluding missing data), the most common adverse outcome was other serious (88/168, 52.38%), and the most serious outcome was death/life-threatening condition (22/168, 13.10%). The mortality rate was highest for sunitinib-related ONJ (12/65, 18.46%), excluding missing data (n = 12), followed by sorafenib-related ONJ (2/13, 15.38%).
In ONJ-related ICSRs for ICIs, the highest number of ICSRs occurred in 2018. ICSRs on nivolumab were collected since 2017, and the highest number of ICSRs occurred in 2018. For pembrolizumab, ONJ AE reports were collected since 2018, and the highest number of ICSRs occurred in 2018, 2020, and 2022. In general, more males (62.22%) suffered ONJ AEs, and most of the cases were from North America and Europe. The median age of patients for all the ICI-related ICSRs was 63 years (IQR, 60–89 years; n = 36). More than half of the ICSRs related to male patients (62.2 %). Unlike for TKIs, the most common indication for ICIs was lung cancer. The most common indication for nivolumab was renal cancer. For all the ICI-related ONJ documented in ICSRs (excluding missing data), the most common adverse outcome was other serious (17/39, 43.59%), and the most serious outcome was death/life-threatening condition (4/39, 10.26%), as shown in Tables 3–5.
Table 3. Demographic and clinical data of ONJ with interested TKIs.
|
Axitinib ( N=24 )
|
Cabozantinib ( N=10)
|
Lenvatinib ( N=40)
|
Pazopanib (N=31)
|
Sunitinib ( N=77)
|
Sorafenib ( N=13)
|
Total (N=195)
|
Age median (range)
|
66
|
50-79
|
68
|
56-78
|
47
|
38-88
|
61
|
52-78
|
61
|
23-81
|
62
|
40-81
|
63
|
23-88
|
|
N
|
|
N
|
|
N
|
|
N
|
|
N
|
|
N
|
|
N
|
|
Sex
|
|
|
|
|
|
|
Females
|
5
|
20.83
|
0
|
0.00
|
15
|
37.50
|
10
|
32.26
|
18
|
23.38
|
4
|
30.77
|
52
|
26.67
|
Males
|
17
|
70.83
|
9
|
90.00
|
22
|
55.00
|
15
|
48.39
|
52
|
67.53
|
8
|
61.54
|
123
|
63.08
|
Missing
|
2
|
8.33
|
1
|
10.00
|
3
|
7.50
|
6
|
19.35
|
7
|
9.09
|
1
|
7.69
|
20
|
10.26
|
Reporter country
|
|
|
|
|
|
|
North America
|
4
|
16.67
|
4
|
40.00
|
14
|
35.00
|
6
|
19.35
|
39
|
50.65
|
3
|
23.08
|
70
|
35.90
|
South America
|
2
|
8.33
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
3
|
3.90
|
0
|
0.00
|
5
|
2.56
|
Europe
|
9
|
37.50
|
4
|
40.00
|
9
|
22.50
|
17
|
54.84
|
26
|
33.77
|
6
|
46.15
|
71
|
36.41
|
Asia
|
8
|
33.33
|
1
|
10.00
|
15
|
37.50
|
5
|
16.13
|
9
|
11.69
|
4
|
30.77
|
42
|
21.54
|
Other
|
0
|
0.00
|
1
|
10.00
|
2
|
5.00
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
3
|
1.54
|
Missing
|
2
|
8.33
|
0
|
0.00
|
0
|
0.00
|
3
|
9.68
|
0
|
0.00
|
0
|
0.00
|
5
|
2.56
|
Outcome
|
|
|
|
|
|
|
Death/life-threat
|
2
|
8.33
|
0
|
0.00
|
5
|
12.50
|
1
|
3.23
|
12
|
15.58
|
2
|
15.38
|
22
|
11.28
|
Disability
|
1
|
4.17
|
0
|
0.00
|
0
|
0.00
|
1
|
3.23
|
6
|
7.79
|
0
|
0.00
|
8
|
4.10
|
Hospitalization
|
2
|
8.33
|
5
|
50.00
|
18
|
45.00
|
6
|
19.35
|
17
|
22.08
|
2
|
15.38
|
50
|
25.64
|
Other serious
|
12
|
50.00
|
3
|
30.00
|
17
|
42.50
|
17
|
54.84
|
30
|
38.96
|
9
|
69.23
|
88
|
45.13
|
Missing
|
7
|
29.17
|
2
|
20.00
|
0
|
0.00
|
6
|
19.35
|
12
|
15.58
|
0
|
0.00
|
27
|
13.85
|
Indication
|
|
|
|
|
|
|
Renal cancer
|
19
|
79.17
|
7
|
70.00
|
12
|
30.00
|
29
|
93.55
|
55
|
71.43
|
6
|
46.15
|
128
|
65.64
|
Hepatic cancer
|
0
|
0.00
|
0
|
0.00
|
5
|
12.50
|
0
|
0.00
|
0
|
0.00
|
1
|
7.69
|
6
|
3.08
|
Thyroid cancer
|
0
|
0.00
|
2
|
20.00
|
16
|
40.00
|
0
|
0.00
|
0
|
0.00
|
2
|
15.38
|
20
|
10.26
|
Other
|
5
|
20.83
|
1
|
10.00
|
7
|
17.50
|
2
|
6.45
|
22
|
28.57
|
4
|
30.77
|
41
|
21.03
|
Time to report
|
2010
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
1
|
3.23
|
5
|
6.49
|
2
|
15.38
|
8
|
4.10
|
2011
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
17
|
22.08
|
1
|
7.69
|
18
|
9.23
|
2012
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
3
|
3.90
|
1
|
7.69
|
4
|
2.05
|
2013
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
4
|
12.90
|
0
|
0.00
|
3
|
23.08
|
7
|
3.59
|
2014
|
2
|
8.33
|
1
|
10.00
|
0
|
0.00
|
2
|
6.45
|
4
|
5.19
|
2
|
15.38
|
11
|
5.64
|
2015
|
1
|
4.17
|
1
|
10.00
|
0
|
0.00
|
6
|
19.35
|
8
|
10.39
|
0
|
0.00
|
16
|
8.21
|
2016
|
3
|
12.50
|
0
|
0.00
|
1
|
2.50
|
4
|
12.90
|
11
|
14.29
|
0
|
0.00
|
19
|
9.74
|
2017
|
3
|
12.50
|
1
|
10.00
|
1
|
2.50
|
3
|
9.68
|
6
|
7.79
|
0
|
0.00
|
14
|
7.18
|
2018
|
3
|
12.50
|
0
|
0.00
|
2
|
5.00
|
3
|
9.68
|
9
|
11.69
|
2
|
15.38
|
19
|
9.74
|
2019
|
0
|
0.00
|
2
|
20.00
|
7
|
17.50
|
2
|
6.45
|
6
|
7.79
|
1
|
7.69
|
18
|
9.23
|
2020
|
3
|
12.50
|
1
|
10.00
|
6
|
15.00
|
0
|
0.00
|
4
|
5.19
|
1
|
7.69
|
15
|
7.69
|
2021
|
2
|
8.33
|
1
|
10.00
|
10
|
25.00
|
0
|
0.00
|
3
|
3.90
|
0
|
0.00
|
16
|
8.21
|
2022
|
7
|
29.17
|
3
|
30.00
|
13
|
32.50
|
6
|
19.35
|
1
|
1.30
|
0
|
0.00
|
30
|
15.38
|
Table 4. Demographic and clinical data of ONJ with interested ICIs.
|
Ipilimumab ( N=3 )
|
Nivolumab ( N=30)
|
Pembrolizumab ( N=12)
|
Total (N=45)
|
Age median (range)
|
66
|
50-79
|
68
|
56-78
|
47
|
38-88
|
63
|
23-88
|
|
N
|
%
|
N
|
%
|
N
|
%
|
N
|
%
|
Sex
|
Females
|
1
|
33.33
|
3
|
10.00
|
8
|
66.67
|
12
|
26.67
|
Males
|
0
|
0.00
|
25
|
83.33
|
3
|
25.00
|
28
|
62.22
|
Missing
|
2
|
66.67
|
2
|
6.67
|
1
|
8.33
|
5
|
11.11
|
Reporter country
|
North America
|
2
|
66.67
|
3
|
10.00
|
5
|
41.67
|
10
|
22.22
|
South America
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
Europe
|
1
|
33.33
|
17
|
56.67
|
3
|
25.00
|
21
|
46.67
|
Asia
|
0
|
0.00
|
10
|
33.33
|
4
|
33.33
|
14
|
31.11
|
Outcome
|
Death/life-threat
|
0
|
0.00
|
3
|
10.00
|
1
|
8.33
|
4
|
8.89
|
Disability
|
0
|
0.00
|
1
|
3.33
|
0
|
0.00
|
1
|
2.22
|
Hospitalization
|
0
|
0.00
|
12
|
40.00
|
2
|
16.67
|
14
|
31.11
|
Other serious
|
0
|
0.00
|
12
|
40.00
|
5
|
41.67
|
17
|
37.78
|
Missing
|
3
|
100.00
|
2
|
6.67
|
4
|
33.33
|
6
|
13.33
|
Indication
|
Renal cancer
|
0
|
0.00
|
12
|
40.00
|
2
|
16.67
|
14
|
31.11
|
Lung cancer
|
1
|
33.33
|
11
|
36.67
|
5
|
41.67
|
17
|
37.78
|
Malignant melanoma
|
0
|
0.00
|
4
|
13.33
|
1
|
8.33
|
5
|
11.11
|
Other
|
2
|
66.67
|
3
|
10.00
|
4
|
33.33
|
9
|
20.00
|
Time to report
|
2017
|
0
|
0.00
|
3
|
10.00
|
0
|
0.00
|
3
|
6.67
|
2018
|
0
|
0.00
|
8
|
26.67
|
3
|
25.00
|
11
|
24.44
|
2019
|
2
|
66.67
|
4
|
13.33
|
1
|
8.33
|
7
|
15.56
|
2020
|
1
|
33.33
|
4
|
13.33
|
3
|
25.00
|
8
|
17.78
|
2021
|
0
|
0.00
|
4
|
13.33
|
2
|
16.67
|
6
|
13.33
|
2022
|
0
|
0.00
|
7
|
23.33
|
3
|
25.00
|
10
|
22.22
|
Table 5. Demographic and clinical data of ONJ with combined drugs.
|
Cabozantinib+Nivolumab ( N=6)
|
Pembrolizumab+Lenvatinib ( N=4)
|
Pembrolizumab+Axitinib ( N=1)
|
Age median (range)
|
58
|
43-65
|
69
|
57-81
|
57
|
57-57
|
|
N
|
%
|
N
|
%
|
N
|
%
|
Sex
|
Females
|
5
|
83.33
|
2
|
50.00
|
0
|
0.00
|
Males
|
1
|
16.67
|
2
|
50.00
|
1
|
100.00
|
Missing
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
Reporter country
|
North America
|
0
|
0
|
1
|
25.00
|
0
|
0.00
|
Europe
|
6
|
100.00
|
1
|
25.00
|
1
|
100.00
|
Asia
|
0
|
0.00
|
2
|
50.00
|
|
0.00
|
Outcome
|
Hospitalization
|
4
|
66.67
|
|
0.00
|
|
0.00
|
Other serious
|
1
|
16.67
|
4
|
100.00
|
1
|
100.00
|
Missing
|
1
|
16.67
|
|
0.00
|
|
0.00
|
Indication
|
Renal cancer
|
6
|
100.00
|
1
|
25.00
|
1
|
100.00
|
Endometrial cancer
|
0
|
0.00
|
2
|
50
|
0
|
0
|
Other
|
0
|
0.00
|
1
|
25
|
0
|
0.00
|
Time to report
|
2018
|
1
|
16.67
|
0
|
0.00
|
0
|
0.00
|
2019
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
2020
|
0
|
0.00
|
0
|
0.00
|
0
|
0.00
|
2021
|
1
|
16.67
|
0
|
0.00
|
0
|
0.00
|
2022
|
4
|
66.67
|
4
|
100.00
|
1
|
100.00
|
A total of 135 cases (ICIs: 20; TKIs: 115) were suitable for calculating the median time to onset value. We found that the median time to onset was 834 days for axitinib-related ICSRs, 853 days for cabozantinib-related ICSRs, 867 days for lenvatinib-related ICSRs, 877 days for pazopanib-related ICSRs, 921 days for sunitinib-related ICSRs, and 840 days for sorafenib-related ICSRs. Additionally, the overall median time to onset values for TKI-related ICSRs and ICI-related ICSRs were 840 days and 666 days, respectively.
3.3 Disproportionality Analysis
For monotherapy, ONJ ICSRs were reported for all six TKIs (axitinib, cabozantinib, lenvatinib, pazopanib, sunitinib, and sorafenib). In particular, the majority of TKI-associated ONJ events were reported for sunitinib (39.5%), followed by lenvatinib (20.5%). Concurrently, based on the ROR criterion, sunitinib and lenvatinib were significantly associated with over-reporting of ONJ (ROR: 1.84, 95% confidence interval [CI]: 1.47–2.31; and ROR: 1.86, 95% confidence interval [CI]: 1.36–2.54, respectively). Among the four ICIs (avelumab, ipilimumab, nivolumab and pembrolizumab), no avelumab-related ONJ ICSRs were found. The majority of ICI-associated ONJ events were reported for nivolumab (66.7%), followed by nivolumab (26.7%). However, the four ICIs showed no significant association with over-reporting of ONJ.
For the six combination therapies (lenvatinib + everolimus, nivolumab + ipilimumab, pembrolizumab + axitinib, avelumab + axitinib, nivolumab + cabozantinib, pembrolizumab + lenvatinib), only the combinations of cabozantinib + nivolumab, pembrolizumab + lenvatinib, and pembrolizumab + axitinib had related ICSRs. Moreover, the majority of combination therapy-related ICSRs were reported for cabozantinib + nivolumab (54.5%), followed by pembrolizumab + lenvatinib (36.4%). However, with the criteria of Ω shrinkage, only the combination therapy of cabozantinib + nivolumab was significantly associated with over-reporting of ONJ (Ω 1.94, Ω025 1.30), as shown in Table 6.
Table 6. Associations of different interested drugs with ONJ.
Drugs
|
N
|
ROR(95%CI)
|
Ω (Ω025)
|
TKIs
|
195
|
1.31(1.14-1.51)
|
|
Axitinib
|
24
|
1.41(0.94-2.10)
|
|
Cabozantinib
|
10
|
1.24(0.67-2.30)
|
|
Lenvatinib
|
40
|
1.86(1.36-2.54)
|
|
Pazopanib
|
31
|
1.00(0.70-1.42)
|
|
Sunitinib
|
77
|
1.84(1.47-2.31)
|
|
Sorafenib
|
13
|
0.43(0.25-0.75)
|
|
ICIs
|
45
|
0.36(0.27-0.48)
|
|
Nivolumab
|
30
|
0.43(0.30-0.62)
|
|
Ipilimumab
|
3
|
0.22(0.07-0.67)
|
|
Pembrolizumab
|
12
|
0.29(0.16-0.51)
|
|
Cabozantinib+Nivolumab
|
6
|
|
1.94(1.30)
|
Pembrolizumab+Lenvatinib
|
4
|
|
0.41(-0.23)
|
Pembrolizumab+Axitinib
|
1
|
|
-0.48(-1.12)
|