Calcium Channel Blockers and Risk of Lymphedema among Breast Cancer Patients: Nested Case–Control Study

Arm lymphedema, the accumulation of lymphatic fluid in the interstitial tissues, is a major complication after breast cancer treatment. Although dependent on the type of therapy received, as many as 41% of patients treated for breast cancer develop lymphedema (1). Notably, there is no pharmacologic therapy to ameliorate arm lymphedema; rather, therapy involves compressive bandages, pneumatic pumps, massage, or surgical interventions (2). Once established, lymphedema is irreversible, and patients who develop lymphedema suffer from its disfiguring effects and endure serious complications that include lymphangitis, cellulitis, ulcers, and the rare development of malignant lymphangiosarcomas (2–8). This emphasizes the need to develop preventative strategies to reduce the risk of arm lymphedema development.

The development of arm lymphedema appears to be a multihit process where multiple variables, in combination, increase the risk (1, 9, 10). Factors associated with increased risk of lymphedema include BMI ≥ 30, advanced-stage cancer, more invasive surgery, radiation, certain chemotherapy regimens, infection, and hypertension (3, 5–7, 11–13). Strategies to prevent arm lymphedema development have focused mainly on the modification of surgical practices and cancer treatment regimens with positive results (2, 14–16). However, one of the primary goals of cancer treatment is preventing cancer recurrence, and whether changes to cancer treatment necessary to reduce the risk of lymphedema negatively affect cancer recurrence rates long-term remains to be determined (16, 17). This highlights the importance of identifying modifiable risk factors that are distinct from cancer treatment that may contribute to lymphedema development.

One potentially modifiable risk factor is the choice of medication class used to treat hypertension. Although hypertension has been associated with an increased risk of lymphedema, it remains unclear whether it is hypertension alone or hypertension treatment that increases the risk of lymphedema (12, 13). A common mechanism of action for antihypertensive medications is the regulation of calcium signaling and smooth muscle contraction, both of which are key elements in normal lymphatic function (18–21). Normally, the lymphatic system maintains fluid homeostasis through rhythmic contractions in lymph vessels to propel fluid from the interstitial tissues back into venous circulation. Calcium channel blockers (CCB) have been shown to inhibit these contractions in lymph vessels isolated from cows, guinea pigs, rats, and humans (18, 20–22); however, whether the inhibitory effects of CCBs in isolated lymph vessels translate to reductions in lymph flow and lymphedema development in patients has yet to be established. It is possible that CCBs, a widely used antihypertensive therapy (23), inhibit lymphatic transport in patients and therefore act as a “second-hit” toward arm lymphedema development in patients already exposed to other cancer treatment–related risk factors.

To our knowledge, no study has directly examined the impact of CCB use on lymphedema development. Therefore, we conducted a nested case–control study to determine the associated risk of arm lymphedema and CCB use in female breast cancer patients treated with antihypertensive medications. This study provides initial evidence to suggest CCB use is a modifiable risk factor to prevent lymphedema in these patients.

The data for the current study were from the IQVIA Health Plan Claims Data (formerly known as Parametric Plus), an administrative claims database that comprises adjudicated claims from 2006 until June 2015 (24). This database carries information on 70 insurance and managed care plans and over 10 million enrollees across the United States since 2006. The data are representative of the U.S. commercially insured population (25).

Information on medical claims (inpatient, outpatient, professional service claims along with their date and place of service), pharmacy claims, and enrollment is available (25). Disease diagnosis was extracted using International Classification of Diseases, version 9 Clinical Modification (ICD-9-CM) codes and procedures were extracted using Current Procedural Terminology (CPT). The enrollment file contains demographic information such as age, sex, geographic region, state, payer, and plan type. The pharmacy claims file includes information on Generic Product Identifier (GPI), National Drug Code, date of dispensed drug, days of supply, quantity dispensed, paid amount, copay, deductible, and allowed amount. The study was determined exempt by the local institutional review board. The study conducted secondary data analysis, and it was deemed as not human subjects research by the local institutional review board (IRB # 228444).

A retrospective nested case–control design was used to examine claims dated January 1, 2007, to June 30, 2015. One-year time before diagnosis of cancer was the baseline period. Breast cancer patients were identified by an ICD-9-CM diagnosis code for breast cancer (174.0–174.9, 175.0, and 233.0; ref. 26) during the study period. In order to improve specificity in identifying breast cancer patients, at least one inpatient claim or at least two outpatient claims separated by 30 days or more for breast cancer diagnosis was required. The date of first diagnosis of breast cancer was termed as “nest entry date.” All patients were required to have continuous enrollment and pharmacy benefits eligibility for 12 months prior to nest entry date. Male breast cancer patients and patients less than 18 years of age on the date of their first breast cancer diagnosis were excluded from the study. Patients with diagnosis of breast cancer within 180 days prior to nest entry date and who received any cancer-related procedure in the form of radiotherapy, chemotherapy, or surgery in the baseline period were excluded from the study. This ensured that only incident breast cancer patients were included in the study. Because the comparison of interest is among hypertensive patients, patients who were not exposed to antihypertensive agents during 90 days prior to nest entry date were excluded from the study (see Supplementary Table S1). Antihypertensive agent classes examined were CCB, beta blocker, thiazide diuretics, angiotensin converting enzyme inhibitor (ACE), angiotensin II receptor blockers (ARB), direct renin inhibitor (DRI), vasodilators, and alpha blockers. The use of loop diuretics and potassium sparing diuretics was excluded as these diuretics are not typically used in the management of hypertension alone (27).

Patients were selected as cases from the nest if they had one or more medical claims with a primary or secondary diagnosis of lymphedema (ICD-9-CM 757.0, 457.1, 457.0, and 729.81) or had one or more procedural claims related to lymphedema (CPT 97016, 97124, and 97140; ref. 28). The date of first lymphedema diagnosis after the nest entry date was defined as the “event date.” Median time from nest entry date to event date was 10 months. Patients without diagnosis of lymphedema were defined as controls. Up to 5 controls were selected for each case, matched on age (within 5 years), nest entry date (within 180 days), type of insurance, thiazide exposure, number of antihypertensive classes used and Charlson Comorbidity Index (CCI) in the baseline period. Insurance type was chosen as a matching variable based on previous literature that looked at adherence to CCB and statin use and risk of developing cardiovascular events (29). Thiazide exposure was used as a matching variable because it is a first-line agent in the management of hypertension and has been shown to decrease the risk of lymphedema (12, 27). Event date was imputed for controls as the difference between the matched case event date and nest entry date to have consistent follow-up for cases and controls (Fig. 1).

The exposure of interest was use of a CCB. Patients were considered exposed if they had at least one prescription for CCB in the 180 days prior to event date. Important confounding variables were collected in the 180 days before event date based on previous studies (7). Surgical interventions (mastectomy, axial node dissection/lymphadenectomy/lumpectomy, breast reconstruction surgery), exposure to radiotherapy, chemotherapy (anthracyclines, taxanes), hydrochlorothiazide, and CCI (23, 30, 31) were identified in the exposure period. Comorbidities were identified from inpatient and outpatient claims using ICD-9 codes and GPI codes.

Unadjusted bivariate analyses and conditional logistic regression were used to assess the effect of exposure among cases and controls for lymphedema development. Adjustment of covariates was carried out in the regression analysis. In order to assess robustness of the study assumptions, sensitivity analysis was carried out by adjusting the observation period from 180 to 90 days. Subgroup analyses were also used to determine the effect of dihydropyridine CCB as compared with patients exposed to nondihydropyridine CCB. All statistical analysis was carried out in SAS, version 9.3.

The sample consisted of 51,050 patients with a diagnosis of breast cancer between January 2007 and June 2015. After applying inclusion and exclusion criteria (Fig. 2), the nest comprised 4,051 female breast cancer patients exposed to antihypertensive agents in the 180 days prior to nest entry date (Table 1). Within the nest, 717 (30%) cases of lymphedema were identified and were matched to 1,681 controls (Table 2). In the final matched sample, 203 cases (28%) and 391 controls (23%) were exposed to CCB (P = 0.0035). The cases were more likely to have undergone mastectomy (8% vs. 5%, P = 0.0039), received radiation treatment (27% vs. 22%, P = 0.0046), and chemotherapy (taxanes: 12% vs. 7%, P = 0.0007; anthracyclines: 6% vs. 4%, P = 0.0073). Cases were more likely to have a CCI of 4 or more than controls (20% vs. 13%, P ≤ 0.0001).

In the unadjusted analysis of the matched group, the odds of exposure to CCB among cases compared with controls was 1.179 (95% CI, 0.939–1.481). Other patient characteristics such as mastectomy (OR: 1.727; 95% CI, 1.189–2.509), node dissection (OR: 1.389; 95% CI, 1.064–1.812), and radiation treatment (OR: 1.570; 95% CI, 1.240–1.986), anthracyclines-based chemotherapy (OR: 1.816; 95% CI, 1.168–2.825), taxane-based chemotherapy (OR: 1.867; 95% CI, 1.346–2.591) were associated with the development of lymphedema.

After adjusting for demographics, comorbidities, and treatment characteristics, use of CCB was significantly associated with the development of lymphedema. Subjects using CCB had 32% higher odds of developing lymphedema as compared with those not exposed to CCB (OR: 1.320; 95% CI, 1.003–1.737). Factors such as mastectomy (OR: 1.630; 95% CI, 1.077–2.467), radiation treatment (OR: 1.443; 95% CI, 1.119–1859) and taxane-based chemotherapy (OR: 1.521; 95% CI, 1.033–2.239) were associated with risk of developing lymphedema after adjustment for covariates (Table 3). These results were robust to changes in the exposure period from 180 to 90 days (OR: 1.307; 95% CI, 1.024–1.667). No significant difference in effect was seen between dihydropyridine and nondihydropyridine CCB use (OR: 1.102; 95% CI, 0.583–2.085). Baseline characteristics were compared between patients who were included in the study and those who were excluded due to lack of continuous enrollment. There were no statistically significant differences between the groups.

Our results highlight the impact of concurrent CCB use on the risk of lymphedema development and raise questions for future studies. To our knowledge, no studies have assessed relative risk of lymphedema development associated with antihypertensive medication use by pharmacologic class and timing of treatment. However, a few studies have begun to explore this potential (12, 13). A prospective cohort study that followed female breast cancer patients for a median of 10.2 years after breast cancer treatment found that hypertension was associated with lymphedema development (HR = 1.49; 95% CI, 1.06–2.10), and after adjusting for individual antihypertensive agent classes, only diuretic use decreased risk of lymphedema development (12). It is difficult to draw conclusions from this study given the broad use of diuretics in hypertension management and because no information regarding duration of treatment, number of antihypertensive medications used, or proximity of antihypertensive medication use in relation to lymphedema development is included. However, given the difference seen with diuretic use compared with other antihypertensive classes, we chose to control for thiazide diuretic use in our study to limit confounding (12).

Another descriptive cross-sectional study was conducted by Ridner and Dietrich to compare self-reported comorbid conditions and medication usage between female breast cancer patients with and without lymphedema (13). In this study, the use of medications to treat cardiac conditions, such as antihypertensive drugs, antiarrhythmic drugs, and agents used in heart failure was associated with increased risk of lymphedema; however, these medications were classified into broad therapeutic categories rather than delineated by drug class or mechanism of action (13). This broad categorization without considering the timing of medication use in relation to cancer treatment and lymphedema development may not accurately reflect the risk of lymphedema associated with individual medication classes. Thus, our study was intentionally designed to not only include individual classes of antihypertensive medications but also account for timing of exposure to these medications in relation to cancer treatment and lymphedema development to assess the potential coinciding risk.

In particular, our study focused on the contribution of CCBs to the development of lymphedema since previous work defining mechanisms of lymphatic transport have identified activation of L-type calcium channels as the primary contributor to lymphatic contraction and thus the driving force of lymph flow (18, 19). Therefore, CCBs that target and block L-type calcium channels may act to impair lymphatic transport resulting in reduced lymph flow and subsequent lymphedema development. Here we provide initial evidence that CCB use increases the risk of lymphedema development in breast cancer patients, suggesting CCB use may play a role in impaired lymphatic transport associated with lymphedema and that avoiding CCB use in hypertension therapy represents a modifiable risk factor that could be addressed to reduce lymphedema development in these patients.

CCBs are traditionally associated with another condition of tissue fluid accumulation known as peripheral edema, which is thought mainly to arise from venous insufficiency but may be partially due to inhibitory effects on lymphatic transport (18, 19). Peripheral edema development with CCB use increases with duration of treatment is dependent on the type of CCB used. The incidence of peripheral edema after 90 days of treatment is approximately 10% and increases to 24% as duration of treatment reaches 180 days (32, 33). The estimated incidence of lymphedema also increases over time, with the vast majority of diagnoses occurring within 180 days of breast cancer treatment (7). Thus, we chose an initial exposure period of 180 days to capture all CCB use and lymphedema development and a 90-day exposure period for our sensitivity analysis; the results were robust to this change.

Importantly, a single randomized, double-blinded, placebo-controlled cross-over study previously investigated the role of lymphatic dysfunction in CCB-associated peripheral edema (34). Although no apparent evidence of lymphedema development was observed in response to CCB use in this study (34), it is important to note the small sample size of this study (n = 6) was not sufficient to detect the expected 2% to 5% incidence of peripheral edema for the exposure period. Therefore, it is difficult to draw conclusions regarding the role of lymphatic transport to peripheral edema development (32). Another distinction between this clinical trial and the current study are the baseline characteristics of the patient population. This clinical trial was conducted in normal healthy male volunteers, as compared with our study in female breast cancer patients whose lymphatics may additionally be compromised by cancer and cancer therapy and thus CCBs may serve as an additional variable to increase the risk of lymphedema development in specific patient populations.

Furthermore, CCBs can be divided into two classes, dihydropyridines and nondihydropyridines, with dihydropyridines being associated with higher incidences of peripheral edema compared with nondihydropyridines (32, 33). However, our subgroup analysis revealed no significant difference between CCB subgroups and lymphedema development, suggesting the increased risk of lymphedema development may be a class-wide association. This interpretation should be taken cautiously because the study was not powered to examine the difference between these subgroups.

Some limitations of our study should be acknowledged. First, the breast cancer diagnosis code date was used as a proxy for diagnosis of cancer. Second, administrative claims do not allow for control of all known risk factors for lymphedema development. For example, BMI greater than 30 has been shown to increase the risk of lymphedema (1, 6, 10, 35); however, BMI was not available in our data set. Additionally, tumor size and nodal involvement, also known risk factors, could not be identified in the data (35). Thus, the analysis controlled for degree of surgical intervention and CCI as surrogate markers for disease severity. Third, previous reports have shown that the risk of lymphedema is increased in African-American women treated for breast cancer who were also diagnosed with hypertension (12). Importantly, African-American patients are three times more likely to be treated with a CCB for hypertension therapy, which may contribute to this increased risk (36). However, the distribution of CCB use by race in these data is unknown. Fourth, the limited sample size precluded further subgroup analysis between antihypertensive agent classes or regarding the combinatory effect of individual cancer therapies, such as chemotherapeutic agents or radiation, and antihypertensive agent classes. Fifth, cases were matched based on the number of antihypertensive agents used, but we did not account for concurrent use of antihypertensive agent classes throughout the exposure period due to limited sample sizes. Sixth, we did not distinguish CCB use for indications other than hypertension; thus, we cannot draw conclusions regarding individual comorbid conditions and risk of lymphedema development. Lastly, in this study we used prevalent use of CCB rather than incident CCB use; therefore, we did not account for duration of CCB use outside of the exposure period. Future studies should expand this work to determine risk of lymphedema across each antihypertensive class, racial/ethnic backgrounds, combination treatments, and other comorbid conditions to identify specific patient populations that may benefit from hypertension therapy management to reduce the risk of lymphedema after breast cancer treatment.

The increased risk of lymphedema associated with CCB use in breast cancer patients suggest that CCB use is a modifiable risk factor in lymphedema development. Thus, avoiding the use of CCBs in breast cancer patients may serve as a viable strategy to reduce the risk of lymphedema development. Our findings also raise the possibility that other medications that alter calcium signaling or mechanisms of muscle contraction may predispose patients to lymphedema and should be avoided in patients at risk. Collectively, comprehensive studies assessing the relative risk of antihypertensive medication classes on developing breast cancer-related lymphedema have the potential to identify hypertension therapy management as an important lymphedema prevention strategy that could be rapidly implemented in the clinical setting.

J.T. Painter is a health research scientist at the Department of Veterans Affairs. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.J. Stolarz, J.T. Painter

Development of methodology: A.J. Stolarz, J.T. Painter

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.J. Stolarz, J.T. Painter

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.J. Stolarz, M. Lakkad, V.S. Klimberg, J.T. Painter

Writing, review, and/or revision of the manuscript: A.J. Stolarz, M. Lakkad, V.S. Klimberg, J.T. Painter

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.J. Stolarz, J.T. Painter

Study supervision: J.T. Painter

The statements, findings, conclusions, views, and opinions contained and expressed in this article are based in part on data obtained under license from IQVIA. Source: IQVIA Health Plan Claims Data (formerly known as PharMetrics Plus) 2007–2015, IQVIA. All rights reserved. The statements, findings, conclusions, views, and opinions contained and expressed herein are not necessarily those of IQVIA or any of its affiliated or subsidiary entities. This research was supported by the UAMS College of Pharmacy Seed Grant (to A.J. Stolarz).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.