The Association Between Medroxyprogesterone Acetate Exposure and Cerebral Meningioma Among a Medicaid Population

Simple Summary

Prior research has reported an increased association between depot medroxyprogesterone acetate, an injectable contraceptive, and cerebral meningioma, a cancer of the protective layers of the brain. Prior research has been limited to meningioma cases treated by surgery or by use of an administrative database of employer-provided commercial insurance; in addition, neither compared medroxyprogesterone acetate use to an active comparator. The current study utilized a population of women aged 18–55 with Medicaid, a type of public insurance in the United States, reporting associations between depot medroxyprogesterone acetate and cerebral meningioma that were similar in strength to prior research; associations for oral and injection medroxyprogesterone acetate were compared to a non-active (i.e., no exposure) and active (i.e., levonorgestrel or norethindrone exposure) comparator. The current study supports previously reported associations with the addition of a significantly increased association when compared to an active comparator.

Abstract

Background/Objectives: Medroxyprogesterone acetate (MPA) is a synthetic contraceptive that can be used orally or as a once-every-three-month injection (i.e., depot MPA [dMPA]). Prior research has reported an increased association between dMPA and cerebral meningioma but has been limited in generalizability to meningioma cases treated with surgery or cases derived from an administrative database of commercial insurance enrollees. The current study builds upon prior research by examining the association among public insurance enrollees utilizing both a non-active and active comparator. Methods: Utilizing Alabama Medicaid data, cases of cerebral meningioma were matched to up to ten controls based on age and year of Medicaid enrollment. A conditional logistic regression estimated odds ratios (ORs) and 95% confidence intervals (CIs) for the association between MPA and dMPA exposure and cerebral meningioma were compared to both an active and non-active comparator. Results: Among 469 cases and 4690 matched controls, there was no association between oral MPA and cerebral meningioma. Associations for dMPA exposure were similar when using a non-active (OR 1.87, 95% CI 1.16–3.00) or active comparator (OR 1.93, 95% CI 01.01–3.69). These associations were strongest for prolonged exposure compared to a non-active (OR 3.80, 95% CI 1.88–7.68) and active comparator (OR 3.67, 95% CI 1.09–12.29). Conclusion: The current results are consistent with the prior literature that dMPA exposure is associated with an increased likelihood of meningioma for prolonged use. More research is needed to examine whether the association is limited to a certain histology or grade of meningioma. Clinicians should consider discussing with patients these reported associations prior to using dMPA.

1. Introduction

Medroxyprogesterone acetate (MPA) is a synthetic progestin that is primarily prescribed for contraception and amenorrhea due to hormonal imbalances; however, it is also indicated in women who experience abnormal uterine bleeding to prevent endometrial hyperplasia as palliative treatment in those with inoperable endometrial carcinomas, and in women experiencing pain from endometriosis [1,2]. MPA binds to progesterone receptors in the hypothalamus, the female reproductive tract, and the pituitary gland, which results in the inhibition of the release of gonadotropin-releasing hormone [1,2]. The less frequent gonadotropin-releasing hormone release reduces the luteinizing hormone surge that occurs mid-cycle, preventing follicular maturation and ovulation [1]. Additionally, it can change the endometrium from proliferative to secretive, which makes implantation difficult.

MPA can be administered orally in tablet form at doses of 2.5, 5, or 10 milligrams; subcutaneously at 104 mg/0.65 mL; or as an intramuscular injection at a 150 mg/mL dose [1]. The intramuscular injection version of MPA (known as depot medroxyprogesterone acetate [dMPA]) is administered once every three months and has a typical use failure rate of 4% [3]. Recent estimates from the United States suggest that 3% of women who reported using contraception between 2017 and 2019 used the MPA injection, while 25% of women aged 15–49 who have had intercourse reported using the MPA injection; however, these percentages varied based on certain sociodemographic characteristics [3,4]. The MPA injection was more commonly used by young women, black women, lower-income, and publicly insured women [3,4]. Notably, four times as many women who were living below 100% of the federal poverty line reported using the injection compared to women living 400% above the poverty line, and the rate of usage among publicly insured patients was three times the rate of privately insured patients [3].

Progesterone has been shown in prior studies to play a role in the development and existence of meningiomas, which are a type of central nervous tumor [5,6,7]. Meningiomas are mostly benign and slow-growing tumors (i.e., World Health Organization grade 1), and they are the most common type of brain tumor, accounting for 36% of all intracranial tumors [8]. They occur more frequently in females, and the WHO grade 1 tumors have a recurrence rate up to 25% [9]. The hormone dependency of meningiomas has been well studied. After puberty, three females develop meningiomas for every one male [10]. Moreover, between 33 and 89% of meningiomas express progesterone receptors [5]; however, this percentage can vary depending on the pregnancy status of a woman and the phase of the female’s menstrual cycle [5,10,11]. All meningiomas that developed in pregnant or postpartum patients expressed progesterone receptors compared to 75.7% of meningiomas in non-pregnant female patients [5]. Additionally, increasing age was associated with decreasing expression of progesterone receptors in meningiomas according to a recent systematic review [5].

A number of studies have identified an association between the use of progestogens and subsequent development of intracranial meningiomas [12,13,14,15]. Literature on the link between the intramuscular injection form of MPA specifically and subsequent development of meningiomas is less common. Two prior population-based case–control studies have identified an association between dMPA and intracranial meningiomas. The first, a matched case–control study derived from the French National Health Data System Database, reported a near six-times higher odds of meningiomas compared to age and location matched controls [13]; however, the study was limited by low exposure to dMPA in this population with only 0.05% of cases and 0.01% of controls reporting use of dMPA [13]. The second study was a case–control study conducted from a large private insurance claims database based in the United States [14], reporting similar associations to the prior study. This case–control study was based on employees of large companies who had private health insurance and could be generally healthier than a publicly insured population. Thus, the results may not generalize to a publicly insured patient population as several prior studies have noted differences in health care utilization and outcomes by type of health insurance in American patient populations [16,17,18,19]. In addition, both prior studies did not utilize an active comparator medication in their analyses, resulting in associations that may be biased due to confounding by factors related to the initiation of medication use (i.e., confounding by indication).

To address these limitations, the objective of the current study was to assess whether the previously reported association between MPA and cerebral meningiomas is also observed in a population of publicly insured women in the United States while also advancing the research by utilizing an active comparator in addition to a non-active comparator.

2. Materials and Methods

The current study utilized Alabama Medicaid administrative data from 2010 to 2023. The Medicaid claims data contains information on enrollment, demographic, inpatient and outpatient encounter claims, and pharmaceutical claims. The claims records, enrollment records, and demographics records were combined utilizing encrypted identification keys provided for both claims and persons enrolled.

Using a matched case–control design, cases were defined as females aged from 18 to 55 years with an ICD-9 or ICD-10 diagnosis code for a cerebral meningioma that was benign (225.2, D32.0), malignant (192.1, C70.0), or of an unknown histological behavior (D42.0). For each case, the date earliest diagnosis was used as the case date. For each case, up to 10 female controls were matched on age ± 1 year and year of Medicaid enrollment; in addition, controls must have been enrolled for at least as long as the case.

For each case and matched controls, data was collected on ICD-9 and ICD-10 diagnoses for claims made within the three years prior to the case date and claimed medication dispensations at any point prior to the case date. For the former, Elixhauser comorbidities were determined for each study participant. The count of comorbidities was summed for each subject, and a categorical variable of unweighted Elixhauser comorbidity count was defined as zero or one comorbidity, two comorbidities, three comorbidities, or four or more comorbidities.

Exposure to MPA was determined separately for both oral MPA and dMPA. For both routes of administration, duration of exposure was determined—following the definition of Roland et al. [13]—as no exposure, short-term (i.e., exposure only for one year), and prolonged-term exposure (i.e., exposure for one year followed by at least one dispensation the following year). There were 12 subjects (0.2% of the total population and all of whom were controls) that had exposure to both MPA and dMPA; these subjects were included in both the MPA and dMPA analyses. The same methodology was conducted for the active comparator exposure, defined as a claimed dispensation of norethindrone (alone or in combination) or of levonorgestrel in combination with another medication (e.g., levonorgestrel and ethinyl estradiol). Levonorgestrel alone is used an emergency contraceptive rather than continued contraceptive use and was excluded from consideration as an active comparator. These two medications were chosen due to their use as contraceptives and because they have not been reported to be associated with meningioma risk in cohort and case–control studies [20,21]. Using these variables, subjects were categorized as having no exposure (i.e., non-active comparator), exposure to an active comparator of levonorgestrel or norethindrone, and exposure to MPA (for the oral MPA analysis) or dMPA (for the injection MPA analysis).

Age, race, and unweighted Elixhauser comorbidities were compared between case and matched controls utilizing a conditional logistic regression. Odds ratios (ORs) and associated 95% confidence intervals (CIs) for the association between MPA/dMPA exposure and cerebral meningioma were estimated from a conditional logistic regression. Models were adjusted for age, race, and categorical number of Elixhauser comorbidities, and separate models were created for any exposure, short-term exposure, and prolonged-term exposure. In a sensitivity analysis, based on recommendations for dMPA not being used longer than two years [1], models were created after categorizing as non-exposed those subjects whose last exposure (both MPA/dMPA and active comparator) was more than one year prior to the case date and more than two years prior to the case date. SAS v9.4 was used for all analyses, and a two-tailed alpha of 0.05 was used for determination of statistical significance.

3. Results

3.1. Bivariate Analysis

The study population included 469 cases of meningiomas and 4690 matched controls. A majority (90%) of patients identified as black or white, and there was no statistical difference in race between cases and controls (p = 0.4874) (

Table 1. Comparison of age and comorbidities of meningioma cases and matched controls.

	Cases (n = 469)	Controls (n = 4690)	p-Value *
Age
18–39	156 (33.3)	1562 (33.3)	1.0000
40–55	312 (66.7)	3120 (66.7)
Race
Asian/Pacific Islander	1 (0.2)	13 (0.3)	0.4874
Black	196 (41.8)	2059 (43.9)
White	230 (49.0)	2100 (44.8)
Hispanic	6 (1.3)	76 (1.6)
American Indian/Alaskan Native	3 (0.6)	23 (0.5)
Other/Unknown	33 (7.0)	419 (8.9)
Elixhauser comorbidities
Unweighted count
0–1	66 (14.1)	2476 (52.8)	<0.0001
2	47 (10.0)	406 (8.7)
3	58 (12.4)	420 (8.9)
≥4	298 (63.5)	1388 (29.6)
AIDS	5 (1.1)	33 (0.7)	0.3829
Alcohol abuse	20 (4.3)	141 (3.0)	0.1338
Deficiency anemias	151 (32.2)	778 (16.6)	<0.0001
Autoimmune conditions	41 (8.7)	214 (4.6)	<0.0001
Chronic blood loss anemia	12 (2.6)	59 (1.3)	0.0232
Leukemia	5 (1.1)	6 (0.1)	0.0005
Lymphoma	4 (0.9)	10 (0.2)	0.0180
Metastatic cancer	24 (5.1)	25 (0.5)	<0.0001
Solid tumor without metastasis, in situ	9 (1.9)	42 (0.9)	0.0369
Solid tumor without metastasis, malignant	90 (19.2)	131 (2.8)	<0.0001
Cerebrovascular disease	83 (17.7)	175 (3.7)	<0.0001
Heart failure	36 (7.7)	230 (4.9)	0.0091
Coagulopathy	21 (4.5)	98 (2.1)	0.0012
Dementia	17 (3.6)	37 (0.8)	<0.0001
Depression	179 (38.2)	1066 (22.7)	<0.0001
Diabetes with chronic complications	74 (15.8)	464 (9.9)	<0.0001
Diabetes without chronic complications	125 (26.7)	749 (16.0)	<0.0001
Drug abuse	66 (14.1)	416 (8.9)	0.0002
Complicated hypertension	54 (11.5)	293 (6.2)	<0.0001
Uncomplicated hypertension	274 (58.4)	1663 (35.5)	<0.0001
Mild liver disease	55 (11.7)	259 (5.5)	<0.0001
Moderate/Severe liver disease	4 (0.9)	16 (0.3)	0.1012
Chronic pulmonary disease	142 (30.3)	918 (19.6)	<0.0001
Neurological disorders affecting movement	31 (6.6)	129 (2.8)	<0.0001
Other neurologic disorders	206 (43.9)	155 (3.3)	<0.0001
Seizures and epilepsy	148 (31.6)	302 (6.4)	<0.0001
Obesity	186 (39.7)	1021 (21.8)	<0.0001
Paralysis	26 (5.5)	88 (1.9)	<0.0001
Peripheral vascular disorders	34 (7.2)	200 (4.3)	0.0027
Psychoses	90 (19.2)	589 (12.6)	<0.0001
Pulmonary circulation disorders	10 (2.1)	70 (1.5)	0.2854
Moderate renal failure	32 (6.8)	128 (2.7)	<0.0001
Severe renal failure	12 (2.6)	56 (1.2)	0.0156
Hypothyroidism	78 (16.6)	417 (8.9)	<0.0001
Other thyroid disease	27 (5.8)	171 (3.6)	0.0232
Peptic ulcer with bleeding	32 (6.8)	88 (1.9)	<0.0001
Valvular disease	39 (8.3)	244 (5.2)	0.0045
Weight loss	50 (10.7)	261 (5.7)	<0.0001

* Estimated from conditional logistic regression.

). Cases had a higher prevalence of four or more comorbidities compared to controls (63.5% vs. 29.6%, p < 0.0001). There was no difference between cases and controls in regard to comorbidities of AIDS (p = 0.3829), alcohol abuse (p = 0.1338), moderate or severe liver disease (p = 0.1012), or pulmonary circulation disorders (p = 0.2854). The prevalence of all other conditions comprising the Elixhauser comorbidity index were significantly different (p < 0.05).

3.2. Non-Active Comparator

Oral exposure to MPA occurred for 3.0% and 2.9% of cases and controls, respectively (

Table 2. Odds ratios * (ORs) and associated 95% confidence intervals (CIs) for the association between medroxyprogesterone acetate (MPA) exposure (compared to non-active comparator) and cerebral meningioma.

Exposure Type	Cases (n = 469)	Controls (n = 4690)	Crude OR (95% CI)	Adjusted † OR (95% CI)
ORAL MPA
Non-active comparator (%)	436 (93.0)	4383 (93.5)	Referent	Referent
MPA (%)
Any	14 (3.0)	137 (2.9)	1.01 (0.59–1.75)	0.65 (0.36–1.16)
Short-term	10 (2.1)	114 (2.4)	0.77 (0.39–1.54)	0.48 (0.24–0.97)
Prolonged	4 (0.9)	23 (0.5)	1.45 (0.43–4.92)	0.97 (0.28–3.33)
INJECTION MPA
Non-active comparator (%)	421 (89.8)	4335 (92.4)	Referent	Referent
MPA (%)
Any	29 (6.2)	182 (3.9)	1.77 (1.14–2.73)	1.81 (1.14–2.89)
Short-term	14 (3.0)	130 (2.8)	1.12 (0.61–2.05)	1.05 (0.55–2.01)
Prolonged	15 (3.2)	52 (1.1)	3.10 (1.64–5.86)	3.80 (1.88–7.68)

* Estimated from conditional logistic regression compared to a non-active comparator (i.e., no exposure to levonorgestrel or norethindrone). † Adjusted for age, race, and number of Elixhauser comorbidities.

). There was no association between any oral MPA exposure and subsequent development of cerebral meningiomas in both the crude (OR 1.01, 95% CI 0.59–1.75) and adjusted (OR 0.65, 95% CI 0.36–1.16) analyses. This association was significantly decreased for short-term exposures (adjusted OR 0.48, 95% CI 0.24–0.97) but not associated with prolonged-term exposures (adjusted OR 0.97, 95% CI 0.28–3.33). Exposure to dMPA occurred among 6.2% of cases and 3.9% of controls; in the adjusted model, cases were nearly twice as likely to have dMPA exposure than controls (OR 1.81, 95% CI 1.14–2.89). This association was only present for prolonged-term exposures (OR 3.80, 95% CI 1.88–7.68) as a null effect was observed for short-term exposures (OR 1.05, 95% CI 0.55–2.01).

3.3. Active Comparator

Similarly to the non-active comparator analysis, there was no association between meningioma and oral MPA exposure overall (adjusted OR 0.70, 95% CI 0.33–1.50). There was no significant association in short term (adjusted OR 0.50, 95% CI 0.20–1.25) or prolonged term (adjusted OR 1.34, 95% CI 0.25–7.00) exposures (

Table 3. Odds ratios * (ORs) and associated 95% confidence intervals (CIs) for the association between medroxyprogesterone acetate (MPA) exposure (compared to active comparator) and cerebral meningioma.

Exposure Type	Cases (n = 469)	Controls (n = 4690)	Crude OR (95% CI)	Adjusted † OR (95% CI)
ORAL MPA
Any (%)
MPA	14 (3.2)	137 (2.9)	0.91 (0.44–1.89)	0.70 (0.33–1.50)
Oral levonorgestrel/norethindrone	19 (4.1)	168 (3.6)	Referent	Referent
Short-term (%)
MPA	10 (2.1)	114 (2.4)	0.70 (0.29–1.69)	0.50 (0.20–1.25)
Oral levonorgestrel/norethindrone	14 (3.0)	133 (2.8)	Referent	Referent
Prolonged (%)
MPA	4 (0.9)	23 (0.5)	1.37 (0.28–6.80)	1.34 (0.25–7.00)
Oral levonorgestrel/norethindrone	5 (1.1)	35 (0.7)	Referent	Referent
INJECTION MPA
Any (%)
MPA	29 (6.2)	182 (3.9)	1.46 (0.79–2.71)	1.93 (1.01–3.69)
Oral levonorgestrel/norethindrone	19 (4.1)	173 (3.7)	Referent	Referent
Short-term (%)
MPA	14 (3.0)	130 (2.8)	0.94 (0.42–2.10)	1.10 (0.47–2.56)
Oral levonorgestrel/norethindrone	14 (3.0)	137 (2.9)	Referent	Referent
Prolonged (%)
MPA	15 (3.2)	52 (1.1)	2.12 (0.69–6.51)	3.67 (1.09–12.29)
Oral levonorgestrel/norethindrone	5 (1.1)	36 (0.8)	Referent	Referent

* Estimated from conditional logistic regression. † Adjusted for age, race, and number of Elixhauser comorbidities.

). A significant association was observed for dMPA exposure overall (adjusted OR 1.93, 95% CI 1.01–2.56) with the association being significantly increased for prolonged-term exposure to dMPA (adjusted OR 3.67, 95% CI 1.09–12.29).

3.4. Sensitivity Analysis

Associations for oral MPA exposure were more decreased than the main analysis when excluding exposures then ended over a year prior to the case date, but the associations were similar to the main analysis when excluding exposures ending more than two years prior to the case date (

Table 4. Odds ratios * † (ORs) and associated 95% confidence intervals (CIs) for the association between medroxyprogesterone acetate (MPA) exposure and cerebral meningioma excluding last exposures one year and two years prior to case date.

			Adjusted OR (95% CI)
Exposure Type	Cases MPA, N (%)	Controls MPA, N (%)	Vs Non-Active Comparator	Vs Active ‡ Comparator
ORAL MPA
Excluding > 1 year prior
Any	3 (0.6)	43 (0.9)	0.37 (0.11–1.25)	0.75 (0.15–3.69)
Short-term	2 (0.4)	32 (0.7)	0.14 (0.02–1.04)	0.41 (0.03–5.08)
Prolonged	1 (0.2)	11 (0.2)	0.49 (0.06–3.88)	0.60 (0.05–7.66)
Excluding > 2 years prior
Any	5 (1.1)	66 (1.4)	0.48 (0.19–1.22)	0.59 (0.18–1.99)
Short-term	3 (0.6)	53 (1.1)	0.20 (0.05–0.85)	0.26 (0.05–1.48)
Prolonged	2 (0.4)	13 (0.3)	0.92 (0.20–4.22)	1.05 (0.15–7.50)
INJECTION MPA
Excluding > 1 year prior
Any	15 (3.2)	55 (1.2)	2.74 (1.45–5.19)	4.16 (1.35–12.79)
Short-term	6 (1.3)	33 (0.7)	1.85 (0.71–4.82)	5.03 (0.88–28.75)
Prolonged	9 (1.9)	22 (0.5)	3.82 (1.63–8.98)	3.20 (0.71–14.45)
Excluding > 2 years prior
Any	19 (1.7)	83 (1.8)	2.55 (1.44–4.51)	2.69 (1.09–6.63)
Short-term	8 (1.7)	51 (1.1)	1.74 0.75–4.01)	2.03 (0.57–7.33)
Prolonged	11 (2.3)	32 (0.7)	3.88 (1.79–8.40)	3.25 (0.88–12.03)

* Estimated from conditional logistic regression. † Adjusted for age, race, and number of Elixhauser comorbidities. ‡ Dispensation of norethindrone (alone or in combination) or levonorgestrel in combination (e.g., levonorgestrel and ethinyl estradiol).

). For dMPA exposure, when excluding exposures that ended more than a year prior to the case date, any dMPA exposure was associated with a near three-fold increased odds of meningioma compared to a non-active comparator (adjusted OR 2.74, 95% CI 1.45–5.19) and an over four-fold increased odds compared to an active comparator (adjusted OR 4.16, 95% CI 12.79). Unlike the main analysis, the strongest association, though not significant, was observed for short-term exposures (adjusted OR 5.03, 95% CI 0.88–28.75); however, when excluding exposures ending more than two years prior to the case date, associations were similar between the non-active and active comparators.

4. Discussion

The results of this matched case–control study showed a strong and persistent association between dMPA exposure and development of cerebral meningiomas in a publicly insured patient population in the United States. This association was limited to exposures for at least two continuous years; additionally, in the sensitivity analyses, limiting exposures to those most recent to the case date increased the strength of associations for any exposure. No significant increased associations were observed for oral MPA exposure.

The associations reported in the current study are similar to associations previously reported [13,14,22] and support the findings of increased likelihood of meningioma for use of MPA in a disproportionality study utilizing data from the Food and Drug Administration’s Adverse Events Reporting System [23]. In a case–control study based on the

National Health Data System in France [13], though there was low overall dMPA exposure (0.05% of cases, 0.01% of controls), prolonged dMPA exposure (using the same definition as in the current study) was associated with 5.6 times the odds of developing meningioma compared to the controls, stronger than the association for prolonged dMPA in the current study though within the confidence limits of the current association. Short-term exposure could not be analyzed in the prior study due to only two participants having short-term exposure to dMPA; however, the results of the current study suggest no association is present within the first year of use. Another case–control study conducted from a large private insurance claims database based in the United States [14] reported that an exposure of two or more years to the dMPA was associated with over twice the odds of developing a cerebral meningioma compared to the controls, even after accounting for age and comorbidities [14]. The association for one year of use was similar (OR 1.35) to what the current study reports. Differences in the association with the current study are likely due to differences in the study population, with the current study including only those with public insurance and the previous study including only those with commercial, employer-provided insurance.

Other studies have reported meningioma occurrence among those with long-term dMPA use. In a retrospective study among women receiving treatment for meningiomas at the University of Pittsburgh Medical center between 2014 and 2021 [24], twenty-five women had a total of 49 meningiomas which all stained positively in progesterone immunohistochemical staining. The average time of dMPA use was 15.5 years among the 25 women. Additionally, of the ten women instructed to stop using dMPA, there was evident tumor shrinkage in half of them, further supporting evidence of a potential link between dMPA usage and development of cerebral meningiomas [24]. Two additional studies in Indonesia and Sweden looked at various types of contraceptive use and development of meningiomas or gliomas [15,23]. A case–control study in Indonesia found elevated odds of meningiomas with ten or more years of hormonal contraceptive use compared to those with no use (OR: 18.22); however, dMPA was combined with other types of contraception, and the isolated effect of dMPA could not be ascertained [23]. A Sweden-based study found that women who had used long-acting hormonal contraceptives (subdermal implants, injections, or hormonal intrauterine devices) for between five and nine years had 2.5 times the odds of meningioma, but there was no association with development of gliomas [15].

The persistence of a strong association in different populations between prolonged dMPA use and later development of cerebral meningiomas supports the notion that there may be a causal link between exposure to dMPA and meningiomas. Moreover, there is a biological explanation for the observed associations between dMPA use and meningiomas. Combined with prior research [14], multiple studies have reported a similar difference in association between oral MPA and dMPA exposure in which the increased association was observed only for dMPA. This difference could be due to the pharmacokinetics of the medication routes. Relative to oral MPA, dMPA is more slowly absorbed into the body, reaching a peak concentration in approximately 50 days compared to a few hours for oral MPA and has been reported to have higher maximum serum concentration [1,25]. This additionally results in a more sustained serum MPA concentration than through the oral MPA route, the absorption of which can be affected by adherence to the daily medication regimen and whether the medication was taken with food, which can increase absorption [26]. Thus, it is possible that the increased meningioma association being observed for dMPA could be due to the difference in exposure, particularly higher concentrations, longer sustained concentrations, and the potential for less variation in daily serum MPA levels for the depot MPA-exposed group.

The sustained serum concentration of MPA could result in greater possibility of interaction with progesterone receptors. A recent systematic review on hormone reception expression in meningiomas noted that slightly over two-thirds of meningiomas expressed progesterone receptors, but this percentage varied depending on the hormone treatment and pregnancy status of the women [5]. Progesterone levels are elevated during pregnancy, and all meningiomas that developed during pregnancy expressed progesterone receptors [5]. Further, MPA has one of the highest progesterone receptor affinities of the progestins [27]. Progesterone receptor status of meningiomas has been associated with location (with a higher prevalence in skull-base meningiomas [28,29]), recurrence [29], and WHO grade [29,30]. For the latter, higher progesterone receptor prevalence is associated with lower grade (WHO grade 1) meningiomas. These grade 1 meningiomas are benign, but treatment, when including surgical resection, can be costly (with a reported median cost of approximately USD 40,000 for the initial admission [31]) and has a reported complication rate as high as 41% with the most common complications including readmissions, seizures, dysrhythmia, intracranial hemorrhage, and cerebral artery occlusion [31].

In light of the above, the current study needs to be viewed in terms of its limitations. First, the study was based on data from medical and pharmaceutical claims of publicly insured women in the U.S., which has two limitations. The first limitation is that it is only known that a Medicaid beneficiary received a prescription for oral MPA; thus, it is possible that exposure misclassification occurred for oral MPA if the enrollee did not use the medication. In this case, the enrollee would be classified as exposed when in fact exposure did not occur. This potentially results in a bias towards the null and a reported association that is weaker than the true association. This misclassification is not present for dMPA as it is a one-time injection. Second, results may not be generalizable to all women. As another limitation, while the associations in this study were adjusted for age, race, and comorbidities, residual confounding may still be present; thus, it is possible, depending on the direction of the confounding (whether positive or negative confounding) that reported associations may lose significance if the effect of the confounding is strongly positive. Accordingly, the results of the current study should be interpreted with caution.

Regarding pathology, this study could not consider grade or pathology of the meningiomas, as only ICD diagnosis codes were available to determine meningioma status, and the pathology data needed to determine the grade was not available. Analyses, therefore, could not be performed to determine whether the reported associations were differential by WHO grade of meningiomas. Approximately 75% of meningiomas are benign and slow-growing tumors (i.e., WHO grade 1); however, even benign tumors can cause significant debilitating neurologic complications which warrants further investigation. Finally, there was a large discrepancy in prevalence of comorbidities between cases and controls; however, when limiting logistic models to those with three or more comorbidities and, separately, 0–2 comorbidities, associations were similar to reported associations.

5. Conclusions

The current study noted elevated odds of cerebral meningiomas with at least two years of exposure to dMPA, which is consistent with prior studies [13,14,15,23,24] and provides further evidence of the link between dMPA exposure and development of meningiomas. In particular, the strong association observed in the three previous case–control studies and the current study, the observation in the current and prior studies finding the association only among dMPA rather than oral MPA, and the progesterone receptor expression in meningiomas provides evidence of a potential causal link between dMPA use and subsequent development of cerebral meningioma.

Clinicians should consider, for patients who are considering dMPA as their contraceptive method, informing the patients of the risks of developing cerebral meningiomas. Future research needs to investigate whether dMPA exposure is associated with specific tumor grades and pathologies.