Switch to generic formulation of temozolomide results in statistically significant increase in grade 3 and 4 bone marrow toxicity in glioma patients in the province of Alberta

Abstract

Background

Temozolomide (TMZ) is an oral, systemic chemotherapy used chiefly for treating high-grade glioma. Due to the rising costs of systemic chemotherapy, many jurisdictions have replaced brand name with generic formulations. The aim of this study was to determine whether or not there was difference in the incidence of grade 3 or 4 bone marrow toxicity and median overall survival in patients treated with brand name versus generic TMZ in the province of Alberta, Canada. The province suspended the use of generic TMZ based on preliminary data pointing to excess toxicity.

Methods

This multicenter, retrospective study included data from patients with newly diagnosed high-grade glioma that received treatment with TMZ in Alberta. Multivariate logistic regression analysis was performed to determine the association between grade 3 or 4 toxicity to generic versus brand name TMZ exposure, ECOG score, and age. Kaplan–Meier survival estimates and log-rank testing were used to determine differences in overall survival between the brand name and generic TMZ cohorts, as well as the cytopenic versus non-cytopenic patients. Furthermore, a screening analysis for grade 3 or 4 bone marrow toxicity was conducted on all de novo glioma patients treated with brand name TMZ after Alberta preemptively stopped generic TMZ.

Results

Grade 3 or 4 neutropenia and thrombocytopenia were observed in 15% and 19% of patients treated with generic TMZ (n = 156) as compared to 3% and 5% of patients (n = 100) treated with brand name TMZ-treated patients; P= .003 and .001. A trend toward increased median overall survival in glioblastoma patients treated with generic TMZ (13.7 months) versus brand name (15.8 months, P = .178.) was also observed through meeting statistical significance. Based on these results, the province stopped the use of generic TMZ and reverted to the Merck TMZ. An initial review of all new glioma patients (n = 89) treated with Merck TMZ since the province stopped the generic drug demonstrated 3.4% and 10.1% grade 3 or 4 neutropenia, respectively.

Conclusions

The statistically significant difference in toxicity profile has prompted the province of Alberta to replace generic TMZ with brand name TMZ in high-grade glioma patients pending more detailed analysis. Our study provides evidence supporting the importance of conducting prospective studies on long-term safety for generic chemotherapies.

Temozolomide (TMZ) is an oral alkylating agent that has been the standard systemic first-line systemic agent for glioblastoma (GBM) since 2005, having also been used in the recurrent glioblastoma setting since 2001.^1–3 TMZ is a prodrug that is converted to [3-methyl-(triazen-1-yl)imidazole-4-carboximide MTIC and is able to penetrate the blood–brain barrier making it an effective treatment for GBM.⁴ A commonly used treatment protocol includes concurrent treatment with both TMZ and radiation therapy for 6 weeks at 75 mg/m² followed by 6–12 months of adjuvant TMZ with a dose of 150 mg/m² for the first cycle and 200 mg/m² for the remaining cycles. While TMZ is typically used for GBM, it is also commonly used for grade 2 and 3 gliomas as well as other malignancies including neuroendocrine tumors, metastatic melanoma, Ewing sarcoma, and cutaneous T-cell lymphoma.⁵

Given the rapidly rising costs for systemic cancer agents, many jurisdictions have moved to generic formulations of various drugs over brand name. In Alberta, Canada, brand name TMZ was used from 2005 to 2016 (Schering-Plough, subsequently Merck), but due to cost consideration, a switch to generic TMZ (Accord) was made from 2016 onward. Generic drugs are cheaper because of the reduced cost required to develop and test the drugs. Health Canada provides guidelines for the efficacy of a generic drug and deems a drug acceptable if its efficacy falls between −20% and +20% of the brand-name compound. The risk with switching to generic drugs is the absence of rigorous clinical trial data and long-term follow-up. Although comparable efficacy and toxicity are assumed between the brand name and generic drug, in the absence of rigorous follow-up, some generic anticancer agents may be more toxic and/or not as effective as the original drug.

Generally, TMZ is a well-tolerated treatment with significant toxicity occurring in a relative minority of cases. The product monograph for Merck TMZ reports rates of 13% and 8.8% for combined grade 3 and 4 toxicities in thrombocytopenia and neutropenia, respectively, during the combined concurrent and adjuvant phases of treatment.⁶ A summation of studies reporting grade 3 or 4 bone marrow toxicities for TMZ finds a range between 5% and 10% for incidence.⁷

Anecdotally, we clinically observed increased rates of thrombocytopenia after the switch to generic TMZ. A review of the literature showed a retrospective French study of 147 patients suggesting that generic TMZ was associated with an increased rate of thrombocytopenia: 19.6% (95% CI: 8.7–30.5) versus 3.1% in brand-name drug (95% CI: 0–6.6), P = .001.⁸ It also demonstrated a statistically significant decrease in overall survival from 20 months with brand-name TMZ to 17 months with generic TMZ (HR 1.83; 95% CI: 1.21–2.8; P = .004).

To further examine our ongoing clinical concern, we conducted a retrospective review of the efficacy and toxicity of brand name versus generic TMZ treatment of de novo or recurrent glioma. The primary objective was to assess the impact of brand name (Merck) versus generic (Accord) TMZ on the incidence of grade 3 or 4 thrombocytopenia and neutropenia. Secondary objectives were to compare the overall survival for GBM patients treated with generic and brand-name TMZ and compare the overall survival of de novo GBM patients who developed cytopenia with those who did not.

Methods

Data Sources

This multicenter, retrospective study included data from patients with newly diagnosed malignant gliomas (grade 2 to grade 4) who received treatment with TMZ at the Cross Cancer Institute (Edmonton, Alberta, Canada) and the Tom Baker Cancer Centre (Calgary, Alberta, Canada) covering northern and southern Alberta as well as the Northwest Territories, covering a population of 4.5 million people. The first cohort consisted of patients treated with generic TMZ from February 2020 to May 2021 (n = 157). Our second cohort consisted of patients treated with Merck TMZ from April 2015 to April 2016 (n = 100). This timing would be right before the switch from Merck to generic TMZ (June 2016 in northern Alberta and October 2016 in southern Alberta). Any patient treated with TMZ during any point in their cancer treatment was included in this study. Specifically, we assessed all patients receiving the standard first-line glioma TMZ treatment used in Alberta consisting of concurrent TMZ chemoradiation, (continuously for 3–6 weeks) and adjuvant TMZ (5 out of every 28 days for 6–12 months). We also included all patients treated for recurrent glioma using TMZ such as those treated with adjuvant dosing (5-day schedule) TMZ or daily dose-dense TMZ until progression. Data were gathered from our electronic medical records and our provincial database. Age was defined as a continuous factor in the analyses. Rates of grade 3 or 4 thrombocytopenia (platelets <50 × 10⁹/L and <25 × 10⁹/L, respectively) and grade 3 or 4 neutropenia (neutrophils <1.0 × 10⁹/L and <0.5 × 10⁹/L, respectively) were recorded. Following the initial retrospective review comparing the Accord cohort and Merck cohort as above, a preliminary screen was conducted with the same methods on 89 Merck-treated de novo glioma patients from December 2021 to December 2022 after the province of Alberta suspended the use of Accord TMZ because of initial concerns of excess toxicity.

Ethics

This review was given ethics approval.

Statistics Plan

Pearson chi-squared tests were performed to analyze the differences between groups. Patient characteristics were collected including age, pathology, sex, extent of resection, and ECOG score. Multivariate logistic regression analysis was performed to determine the association between grade 3 or 4 toxicity to TMZ exposure, ECOG score, and age. Kaplan–Meier survival estimates and log-rank testing were used to determine differences in overall survival between the Merck and generic TMZ cohorts for de novo glioblastoma only as there were too few cases of other pathologies to provide meaningful results, as well as the cytopenic versus non-cytopenic patients. All statistical analyses were conducted by SPSS software (version 23.0 IBM, NY) and SAS software (SAS Institute Inc 2013. SAS/ACCESS® 9.4 Interface to ADABAS).

Results

Patient Characteristics

In total, 257 patients were included in this study. From April 2015 to April 2016, 100 patients were treated with brand-name (Merck) TMZ as compared to 157 patients treated with generic (Accord) TMZ from February 2020 to May 2021. Their baseline characteristics are presented in Table 1. There were more males than females, in keeping with known demographic data for gliomas. GBM was the predominant histology. The median age in the Merck group was 56 years and 58 years in the Accord group. Most patients had a subtotal or gross total resection with a combined 84% in the Accord group and a combined 79% in the Merck group. Similarly, most patients were ECOG 0-1 at presentation with a combined 72% in the Merck group and a combined 82% in the Accord group. All patients were treated with trimethroprim-sulfamethoxazole (TMP-SMX) prophylaxis during the concurrent chemoradiation phase of treatment. For the brand-name cohort in this study, it was common to use the anti-epileptics valproic acid and phenytoin if required. These anti-epileptics influence hepatic drug metabolism and valproic acid is known to cause cytopenias. For the generic cohort in this study, it was common to use the anti-epileptics lamotrigine and levetiracetam if required.

Table 1.

Patient characteristics at baseline

	Merck, n = 100 (%)	Generic n = 157 (%)
Performance status (ECOG)	ECOG 0 29 ECOG 1 43 ECOG 2 24 ECOG 3 4	ECOG 0 31 ECOG 1 51 ECOG 2 14 ECOG 3 4
Male	64	61
Female	36	39
Median age in years	56 (16–76)	58 (14–88)
WHO Pathology	GBM 70 Gr 2/3 Astrocytoma 19 Gr 2/3 Oligodendroglioma 4 HGG NOS 7	GBM 77 Gr 2/3 Astrocytoma 15 Gr 2/3 Oligodendroglioma 7 HGG NOS 1
MGMT status	Methylated 52 Unmethylated 41 Unknown 7	Methylated 56 Unmethylated 41 Unknown 3
IDH status	Mutated 25 Not mutated 64 Unknown 11	Mutated 37 Not mutated 62 Unknown 1
Extent of resection	Biopsy 4 Subtotal 49 Gross total 35 Unknown 11	Biopsy 20 Subtotal 57 Gross total 22 Unknown 1

^*HGG NOS includes oligoastrocytomas, gliomas with no subtype, gliomas with no specific grade.

Abbreviations: ECOG = Eastern Cooperative Oncology Group; GBM = glioblastoma; HGG = high-grade glioma; NOS = not otherwise specified; WHO = World Health Organization.

Toxicity Associations

As shown in Table 2, generic TMZ was associated with an increase in the frequency of both grade 3 or 4 thrombocytopenia(19.1% vs 5.0%, P = .02) and neutropenia (14.6% vs 3.0%, P = .05) as compared to brand name. We then used multivariate logistic regression analysis to identify any association between these toxicities and other clinical factors. The multivariate logistic regression analysis identified a statistically significant association between grade 3 and 4 toxicity for both neutropenia and thrombocytopenia and the use of generic TMZ (Table 3).

Table 2.

Incidence of grade 3 or 4 bone marrow toxicity and overall survival between Merck and generic cohorts

Grade 3 or 4 toxicity	Merck—n = 100, (%)	Generic—n = 156, (%)	P-value
Neutropenia	3 (3%)	23 (15%)	.003
Thrombocytopenia	5 (5%)	30 (19%)	.001
	(de novo GBM cohort only) N = 70	(de novo GBM cohort only) N = 123
Median overall survival	15.8 mo	13.7 mo	.18

Abbreviation: ECOG = Eastern Cooperative Oncology Group.

Table 3.

Multivariate logistic regression analysis for grade 3 or 4 neutropenia

Variable	Odds ratio	P-value	95% CI
			Lower	Upper
Generic (Accord) vs Brand Name (Merck)	5.76	.007	1.63	20.40
Sex (Female vs Male)	4.49	.001	1.82	11.11
ECOG 1	1.73	.64	0.17	17.21
ECOG 2	1.74	.63	0.19	16.06
ECOG 3	1.91	.59	0.18	20.24
Age*	1.00	.85	0.97	1.04

Abbreviation: ECOG = Eastern Cooperative Oncology Group

^*Age is a continuous variable

Additionally, female gender (see Table 4) showed a statistically significant association with grade 3 or 4 neutropenia and thrombocytopenia with an OR: 4.5 (95% CI: 1.8–11.1; P = .001) and 2.8 (95% CI: 1.3–5.9, P = .009), respectively. The median time to grade 3 or 4 toxicity was similar for both neutropenia and thrombocytopenia at 45 days for brand-name TMZ and 48 days for generic TMZ for either first-line adjuvant or continuous dosing. No patients had grade 3 or 4 neutropenia beyond 240 days on treatment; only 1 patient had grade 3 or 4 thrombocytopenia after 240 days on TMZ treatment and this was attributed to the use of second-line chemotherapy. 240 days is the duration for first-line treatment from the beginning of chemoradiation with daily dosing to the end of the sixth cycle of adjuvant TMZ with 5-day dosing. Seventy-five percent of toxicities occurred during concurrent chemoradiation and 95% of toxicities occurred by the second cycle of adjuvant treatment. No toxicity was recorded on retreatment with TMZ. Exact dose reduction data and dose delay data were not readily available but all patients with grade 3 or 4 toxicity either had a dose delay, dose reduction, or complete treatment cessation per standard practice.

Table 4.

Multivariate logistic regression analysis for grade 3 or 4 thrombocytopenia

Variable	Odds ratio	P-value	95% CI
			Lower	Upper
Generic (Accord) TMZ	4.66	.003	1.69	12.85
Sex (F)	2.76	.009	1.29	5.92
ECOG 1	0.63	.63	.85	2.1
ECOG 2	0.95	.95	0.17	5.19
ECOG 3	0.77	.77	0.21	8.03
Age*	1.01	.46	.98	1.04

Abbreviation: ECOG = Eastern Cooperative Oncology Group

^*Age is a continuous variable.

For the 12-month period from December 2021 to December 2022, we have treated 89 glioma patients with brand-name TMZ during concurrent chemoradiation and adjuvant TMZ. There have been 9 episodes of grade 3 or 4 thrombocytopenia and 3 episodes of grade 3 or 4 neutropenia (10.1% and 3.4%, respectively). Eleven percent of patients had either grade 3 or 4 thrombocytopenia or neutropenia.

Of the 7 Merck-treated patients with grade 3 or 4 bone marrow toxicity, 3 of them needed an ER visit or hospitalization following the toxicity whereas 23 of the 36 Accord-treated patients needed an ER visit or hospitalization following the toxicity. Overall, only 4 of these visits were directly attributed to neutropenia or thrombocytopenia with all of these occurring in the Accord arm. There were 3 admissions for thrombocytopenia requiring platelet transfusion and 1 admission for aplastic anemia. There were 8 COVID-positive cases in the Accord cohort and none of these had bone marrow toxicity. Platelet transfusion and other bone marrow supportive therapies were seldom used in either group preventing statistical comparison between arms. Overall, there was 1 patient in the Merck arm that required a platelet transfusion. There were 6 patients in the Accord arm that required a platelet transfusion, 1 that required a blood transfusion and 1 that was treated with Neulasta (pegfilgrastim) for neutropenia. Notably, 1 patient decided to be comfort care only who would normally have received platelet transfusions. Another was a Jehovah’s witness and, therefore, refused all transfusions.

Survival Outcomes

The de novo GBM subpopulation represents the most prevalent pathology in either cohort, so survival was calculated just for these patients. Non-GBM glioma represented a minority of cases in both the generic and Merck cohorts, so analysis of these patients was not incorporated. Overall, the mOS for GBM patients showed a nonsignificant difference of 13.7 months for the 123 generic-treated patients (95% CI: 11.4–16.0 months) as compared to 15.8 months (95% CI: 13.2–18.3 months), P = .178 in the 70 brand-name-treated patients (Figure 1). The mOS for de novo GBM patients with a grade 3–4 cytopenia versus those without was a nonsignificant difference of 14.1 months (95% CI: 11.4–16.5 months) for the 43 patients with grade 3 or 4 cytopenia as compared to 17.6 months (95% CI: 14.05–20.7 months) for the 213 without grade 3–4 cytopenia (P = .173)

Figure 1.

Kaplan–Meier curve with overall survival in generic versus Merck-treated patients

Discussion

TMZ is the standard of care for GBM in the concurrent and adjuvant phases of treatment. It is also a commonly used regimen for the management of grade 2 and 3 gliomas as well as commonly used in the recurrent setting for all gliomas. The product monograph for the Merck formulation reports a relatively low rate of bone marrow toxicity with grade 3/4 neutropenia in 8.8.% of cases and grade 3/4 thrombocytopenia in 13% of cases.⁶ Some single-center studies report an even lower incidence of grade 3 and 4 toxicity at 2%,⁹ and data collected across multiple studies suggest a rate of 5%–10% for grade 3 and 4 combined thrombocytopenia and neutropenia, including Merck brand data.⁷ The product monograph of the generic references the same data.

To our knowledge, our study represents the largest evaluation of generic versus brand-name TMZ with regard to toxicity in the literature with at least 100 cases in each cohort. The Accord formulation of TMZ is the only generic formulation used in Alberta. We detected a fivefold increase in grade 3 or 4 bone marrow toxicity in our generic-treated patients compared to those treated with brand name. The vast majority of toxicities occurred during concurrent chemoradiation or early on after the start of adjuvant TMZ, as indicated by the median times to neutropenia and thrombocytopenia of 48 and 45 days, respectively.

Currently, TMZ formulations from 6 manufacturers are approved for use in Canada: Merck Canada Inc (brand name), Accord Healthcare Inc, Apotex Inc, Jamp Pharma Corporation, Taro Pharmaceuticals, and Teva Canada Limited. The process for access to TMZ across Canada’s health jurisdictions varies by province, ranging from coverage via a formulary to the use of a patient’s own private insurance at a retail pharmacy. The formulation from Accord Healthcare Inc. is the only generic formulation used in Alberta. Accord brand temozolomide is marketed in many other countries including the United States, Japan, and the European Union.

Generic drugs are approved without the same clinical trial testing because of their similarity in basic composition and properties with innovator formulations. However, the observation that generic drugs induce toxicity and survival differences as compared to brand name has been reported in the literature previously.¹⁰ For instance, it has been shown that grade 3 leukopenia was more frequent in patients treated with concurrent generic cisplatin chemoradiation for uterine cancer compared to brand-name cisplatin.¹¹ Another study found a higher proportion of ifosfamide-induced encephalopathy in the group of patients receiving generic formulation versus brand-name product at 10.2% versus 1.9%, respectively.¹² The exact mechanism for any survival difference is not known. Manufacturing details are proprietary and not available, and it is possible that differences in manufacturing process could contribute to any differences in survival and toxicity between generic and brand-name formulations. Additionally, the allowed ±20% difference in efficacy for generic drugs by Health Canada could be more significant in cytotoxic drugs like TMZ which have a much tighter therapeutic index than most drugs.

No significant difference was seen in GBM survival between the 2 patient groups although there was a slight trend toward lower survival with the generic brand. There was a slight difference in performance status between groups with 72% of Merck patients being ECOG 0-1 compared with 83% of patients treated with Accord TMZ. Additionally, there was a higher proportion of gross total resection in the Merck group (32%) compared to the Accord group (22%) across all pathologies. Notably, postoperative MRI became more common in recent years, during the Accord group, so gross total resections were likely downgraded to subtotal resections more commonly in that group. More biopsies were also done in the Accord group (13%) versus the Merck group (4%). Numbers were too small to meaningfully control for these subgroups in OS calculations. Due to these variables, we would not make any conclusions on OS in these groups.

We also observed that women were more likely to experience clinically significant neutropenia using the generic drug. This gender difference has been reported in the literature for TMZ and chemotherapy in general previously and is not surprising to be seen here with the generic TMZ as well.^13,14 The amount of increased risk was not significantly different from that seen with the brand name although small numbers prevent a true comparison.

The impact of concurrent medications taken with TMZ was also considered. The use of TMP-SMX for PJP prophylaxis was the same between the generic and brand-name cohorts. Enzyme-inducing anti-epileptics could alter the metabolism of TMZ. However, our standard of care is to avoid the use of enzyme-inducing anti-epileptics with TMZ to avoid any potential effects on serum TMZ concentration that could result from altering hepatic metabolism. Moreover, our current preferred first-line anti-epileptics levetiracetam and lamotrigine do not significantly feature cytopenias as an adverse event, although it is rarely reported.^15,16 This is not the case with agents that were more commonly used in the brand-name era like phenytoin or valproic acid. If brand-name-era patients were treated with levetiracetam and lamotrigine, it is possible the incidence of cytopenia would be even lower. We have had a shift in radiation practice to greater use of the elderly protocol of 40 Gy in 15 fractions as well as a hypofractionation trial of 60 Gy over 20 fractions running in the later time period.¹⁷ If anything, the shorter course of radiation, and therefore TMZ, should lead to less bone marrow toxicities in this cohort. The amount of radiation to the bone marrow in the skull is not significantly different between treatment protocols. As well, the skull is seldom implicated in bone marrow toxicity because of the insignificant volume of bone marrow it contains.

We observed increased grade 3 and 4 bone marrow toxicity in patients treated with generic TMZ. Our data are consistent between the 2 provincial cancer centers that treat CNS tumors in the province and include all patients treated chronologically during the 2 time periods. Treatment for gliomas during these time periods was the same apart from the previously mentioned alterations We believe that these data may represent a significant patient safety concern for other patients being treated with Accord formulation TMZ and possibly likewise for other generic TMZ formulations being used elsewhere

To this end, the Alberta Provincial Health Authority has stopped the use of generic TMZ and adopted brand-name Merck TMZ in December 2021. A screening analysis of all de novo glioma patients (n = 89) treated with Merck TMZ since this switch has identified reduced rates of grade 3–4 neutropenia and grade 3–4 thrombocytopenia with 3.4% and 10.1%, respectively. For comparison, our Accord arm (n = 100) had a 19% incidence of grade 3 or 4 neutropenia and a 15% incidence of grade 3 or 4 thrombocytopenia.

Limitations of this study include its retrospective nature and the small sample size. In our study, the small sample size has likely affected the veracity of our survival calculations and hindered our ability to reach definitive conclusions regarding survival. Also, the difference in the timing of treatment (2015–2016 vs 2020–-2021) may have resulted in demographic or institutional changes that were not entirely accounted for.

We will be following up this report with a more comprehensive analysis of our entire patient population from 2010 to 2021 treated with both formulations of TMZ. This will provide us with more robust data detailing the impact on cytopenias and potentially provide the statistical power needed to detect an OS difference. These findings do not necessarily apply to other generic versions of the TMZ, but do prompt consideration for other clinicians to conduct similar reviews of other generic formulations and to ultimately conduct prospective, controlled studies to address the issue with more rigor.

Funding

None declared.

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest aside from OA who declares the following: Advisory board/ honoraria with Ipsen; Eisai; Amgen; Roche; Bayer, and Lilly.

Authorship statement

Conception and design of study: JE, KY, EO. Acquisition of data: EO and EB. Analysis and/or interpretation of data: EO, JE, OA, SP, KY. Drafting the manuscript: EO; revising the manuscript critically: JE, OA, SP, KY, EO, RC. Approval of the version of the manuscript to be published: EO, EB, OA, SP, JE, KY, RC.