Calcium Channel Blocker Versus Renin–Angiotensin System Inhibitor in Risk of Kidney Cancer Among Patients With Hypertension: A Propensity Score‐Matched Cohort Study Authors: Minji Jung, Shufeng Li, Zhengyi Deng, Jinhui Li, Mingyi Li, Satvir Basran, Marvin E Langston, Benjamin I Chung Published: 2024 ABSTRACT ABSTRACTBackgroundUse of antihypertensive medications could be associated with an increased risk of kidney cancer. Despite their various mechanisms of action, whether this association differs between different classes of medications remains unclear.ObjectiveThe objective of this study is to compare the risk of kidney cancer between first‐line treatment options of antihypertensive medications in a hypertensive population.MethodIn this retrospective cohort study, we used the MarketScan Databases (2007–2021). We included individuals older than 30 years of age with a diagnosis of hypertension who received first‐line medications for hypertension, which included three classes: angiotensin‐converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), and dihydropyridine calcium channel blockers (dCCB). We applied a propensity score matching method and created three separate cohorts: (1) ARB versus ACEI, (2) dCCB versus ACEI, and (3) dCCB versus ACEI. For non‐dCCB, we repeated the analyses. The primary outcome was kidney cancer incidence. To assess kidney cancer risk, we applied multivariable conditional Cox proportional hazards models.ResultsIn the first cohort, ARB use was associated with an increased risk of kidney cancer compared to ACEI use (hazard ratio [HR] 1.10, 95% confidence interval [CI] 1.02–1.18). In the second cohort, dCCB use was associated with an increased risk of kidney cancer compared to ACEI use (HR 1.29, 95% CI 1.18–1.40). In the third cohort, dCCB use was associated with a higher risk of kidney cancer compared to ARB use (HR 1.17, 95% CI 1.08–1.28). Null association was shown when comparing non‐dCCB with ACEI or ARB use.ConclusionUse of dCCB showed a higher risk of kidney cancer compared to ACEI or ARB use in patients with hypertension.Keywords: antihypertensive drugs, calcium channel blocker, hypertension, kidney cancer, renin–angiotensin system inhibitor Introduction Kidney cancer is the sixth most common cancer in men and ninth most common cancer in women [1], and its incidence has been increasing [2, 3]. Due to a lack of nationwide screening and specific signs or symptoms, kidney cancer cases are likely to be detected at a later stage [4], which leads to poorer outcomes with a 5‐year relative survival below 10% [5, 6]. This illustrates the importance of identifying high risk groups and potential preventive strategies for kidney cancer. Hypertension is an established risk factor for kidney cancer development [3, 4, 7]. In the United States, it is estimated that ~50% of adults (119.9 million) have hypertension in 2023. The hypertensive population is on the rise, driven by an aging population, along with the rising prevalence of obesity [8]. Of those, 80% (94.9 million) are recommended to take antihypertensive medications to control their blood pressure and prevent cardiovascular diseases [8]. Due to the chronic nature of hypertension, it is common for individuals to take antihypertensive medications throughout their entire lives. Although antihypertensive medications are necessary to control blood pressure, evidence from preclinical experimental studies supporting a higher risk of kidney cancer with antihypertensive medication use has been growing. They have demonstrated that various classes of antihypertensive medications could potentially contribute to developing kidney cancer through different carcinogenic mechanisms, such as facilitating tumor cell proliferation, angiogenesis, or migration [9, 10, 11]. Considering their distinct mechanisms of actions in controlling blood pressure [12], it is reasonable to hypothesize that each class of these medications may have varying effects on kidney cancer development. To accurately evaluate these differential influences, it is necessary to conduct a direct comparison between different classes. To date, however, research in this area has been limited to case–control studies, which compared patients who took mediations to those who did not take them [13, 14, 15, 16]. It is also important to emphasize that the medications are prescribed not only for hypertension but also for other conditions such as heart failure and arrhythmia, which was not considered in previous studies. These significant limitations are more likely to introduce several biases (e.g., indication bias, selection bias, prevalent user bias). Given the widespread use of antihypertensive medications and their potentially distinct roles in cancer, we aimed to directly compare the risk of kidney cancer among first‐line antihypertensive medication options–angiotensin‐converting enzyme inhibitor (ACEI), angiotensin receptor blocker (ARB), dihydropyridine calcium channel blocker (dCCB), and non‐dCCB–in a hypertensive population. We conducted five head‐to‐head comparison analyses using propensity score matching: (1) ACEI versus ARB, (2) dCCB versus ACEI, (3) dCCB versus ARB, (4) non‐dCCB versus ACEI, and (5) non‐dCCB versus ARB. This approach addresses previous research limitations by considering the distinct mechanisms of these drugs and minimizing potential biases, leading to more accurate results. Materials and Methods Data Source and Study Design This retrospective cohort study was conducted using the Merative MarketScan Commercial and Medicare Databases (2007–2021, version 3.0) [17]. The Databases include more than 190 million commercially insured individuals with employer‐sponsored health insurance or individuals over the age of 65. These include de‐identified individuals' longitudinal health insurance claim information including socio‐demographics, insurance enrollment, medical institution visits, diagnostic codes for inpatient and outpatient settings, and records of prescription claims for outpatient setting. The diagnoses were recorded according to the International Classification of Diseases (ICD), 9th and 10th revision. The prescription claims were recorded according to generic drug names with the National Drug Codes. This study using the de‐identified data, hosted by the Stanford Center for Population Health Sciences (PHS), is covered under the PHS umbrella IRB 40974. Study Population We included patients with a diagnostic record of hypertension who initiated first‐line treatment options of antihypertensive medications [12]. Diagnosis of hypertension was identified using ICD‐9 and ICD‐10 codes (Table S1). To reduce potential indication bias, the first‐line treatment options for hypertension, which are the most commonly used three classes, included: renin–angiotensin–aldosterone system inhibitors including ACEI and ARB, and dCCB [12]. We repeated the main analyses for non‐dCCB. Due to the diverse indications of beta‐blockers (BB) and diuretics (DU), we considered them as covariates in this study. We defined the index date as the first prescription date of the medications. We excluded the following patients: (1) those under the age of 30 at the index date [18]; (2) those with a history of any cancers except for non‐melanoma skin cancer prior to the index date; (3) those with a history of Von Hippel–Lindau, tuberous sclerosis, or polycystic kidney disease prior to the index date; (4) those with missing data of age and sex at the index date; and (5) those without insurance enrollment within 1 year before and after the index date. To perform the head‐to‐head comparison analyses, we assembled three main separate study cohorts: (1) ARB versus ACEI, (2) dCCB versus ACEI, and (3) dCCB versus ARB. When comparing ACEI and ARB, we selected ACEI as a comparator because ACEI has been used more commonly than ARB [12]. When comparing dCCB with ACEI or ARB, we selected ACEI or ARB as an active comparator because ACEI and ARB have similar pathophysiological mechanisms and clinical effects by inhibiting renin–angiotensin–aldosterone system [12]. Index Antihypertensive Medications The index antihypertensive medication, the main exposure, was defined as an initiation of ACEI, ARB, or dCCB (Table S2) [12]. To ascertain their initiation, we included individuals with no previous prescription record of the index medication prior to the index date. To ensure their long‐term use, we included individuals who continued longer than 180 consecutive days with at least 80% adherence using the proportion of days covered (PDC) [19]. The PDC was calculated by dividing the total number of days the medication was supplied by the total prescribed duration, from the first to the end date of the last prescription. To assess whether the associations were different according to medication exposure, we identified a cumulative dose of defined daily dose (cDDD) and duration of antihypertensive medication use. The DDD is the assumed average maintenance dose per day for a medication for its main indication in adults [20]. The cDDD was calculated as a ratio of “total exposure amount” to the DDD value of the medication. The “total exposure amount” was calculated by multiplying the cumulative dose of the medication by the number of days' supply covered by the medication. As this provides a fixed unit of measure independent of strength or medication type, we were able to compare medication exposure between different classes. We categorized the cDDD as four groups: < 500, 500–1500, 1500–3000, and more than 3000 cDDD. We classified the duration as six groups: < 1, 1–3, 3–5, 5–7, 7–10, and longer than 10 years. Other Antihypertensive Medications In addition to the index medications, we identified the other antihypertensive medications, which included three classes: BB, DU, and non‐dCCB [12]. BB and DU are more likely to be added to the index medication or used for other compelling indications besides hypertension, such as atrial fibrillation or heart failure [12]. We identified the use of these medications during 1 year prior to the index date. For non‐dCCB, we repeated the analyses to assess the risk of kidney cancer. Non‐dCCB shares a similar mechanism of action with dCCB, by antagonizing calcium channels, but has different primary target tissues: dCCB primarily targets blood vessels, whereas non‐dCCB affects heart. We assembled two additional study cohorts: (4) non‐dCCB versus ACEI and (5) non‐dCCB versus ARB. Similar to the main analysis, ACEI and ARB was used as an active comparator, respectively. Study Outcomes The study outcome was kidney cancer incidence, identified with at least one hospitalization or discharge diagnosis code or two outpatient diagnosis codes (ICD‐9 codes, 189.0 and ICD‐10 codes, C64). To ensure an incident case, we excluded prevalent cases before the index date. Each hypertensive patient was followed from the index date to the study outcome occurrence, disqualification of the insurance enrollment, or the end date of the study period (December 31, 2021), whichever came first. Given the chronic use of antihypertensive medications, we used an intention‐to‐treat approach in the main analysis, implying that treatment termination was disregarded. We assumed that the index medication may affect the risk of cancer, which is a long‐term progressive disease with a long latency period, even after termination of treatment. In a sensitivity analysis, we used an as‐treated approach. This approach accounted for a termination of treatment such as discontinuation or switch, which was defined as no claim filed for the medication within 6 or more months after the due date of a refill prescription. In this case, individuals were followed up by the outcome occurrence, termination of treatment, disqualification of the insurance enrollment, or the end date of the study period, whichever came first. Potential Confounders We considered the following variables as potential confounders: (1) demographic variables including age, sex, geographical region of individuals' residence, and insurance type at the index date; (2) hypertensive complications or comorbidities including myocardial infarction, ischemic stroke/transient ischemic stroke, hemorrhagic stroke, diabetes mellitus (DM) with or without complications, heart failure, angina, atrial fibrillation, dyslipidemia, peripheral vascular disease, chronic kidney disease (CKD), renal failure, and chronic obstructive pulmonary disease (COPD) during 1 year prior to the index date (Table S1); and (3) other medication use including lipid‐lowering agents, anti‐diabetic agents, anti‐thrombotic agents, anti‐platelet agents, and anti‐analgesic agents during 1 year before the index date [12, 21]. Charlson comorbidity index (CCI) was calculated to estimate the burden of individual's health condition [22]. Statistical Analysis We used propensity score (PS) matching in a 1:1 ratio using caliper without replacement to minimize the imbalance of potential confounders. Multivariable logistic regression models were used to estimate PS conditionally on the abovementioned variables including demographics, comorbidities, and medications. The caliper was calculated as 0.1 of the standard deviation of the logit of PS. To assess the statistically significant differences between PS‐matched groups, we used absolute standardized mean differences with a cutoff value of > 0.1 [23]. We calculated cumulative incidence during the follow up and incidence rates per 100,000 person‐years. We examined the proportional hazards assumptions using the Schoenfeld residual method and conducted the conditional Cox proportional hazards regression. To reduce potential confounding effects, we further adjusted the model for CCI, which was not used in the matching model. To assess whether the associations differ according to the baseline characteristics, we conducted several subgroup analyses according to age, sex, comorbidities, and the previous use of other antihypertensive medications. To confirm the robustness of the main findings, we conducted six sensitivity analyses. First, to examine the impact of possibly residual indication bias, we redefined the use of index antihypertensive medications as an initiation with no prescribed record of any other antihypertensive medications, such as BB or DU, prior to the index date. Second, to assess the effects of long‐term use, we restricted the study patients who took antihypertensive medications more than 1000 cDDD [13]. Third, to allow for an induction period of kidney cancer related to antihypertensive medication exposure and to minimize reverse causation that hypertension caused by undiagnosed kidney cancer, we applied 2 years of lag time after the index date. Fourth, to reduce the potential remaining indication bias, we repeated the analyses by excluding patients who had a history of potential compelling indications such as myocardial infarction, ischemic stroke, hemorrhagic stroke, DM, heart failure, angina, atrial fibrillation, peripheral vascular disease, CKD, or renal failure [12]. Fifth, to account for discontinuation of the index medication or switch to another medication, we applied the as‐treated approach as mentioned above. In such case, we assumed that the medication may not affect the cancer risk after termination of treatment. Lastly, we repeated the analyses including the entire population before PS matching to examine the associations independently with unknown probabilities of receiving different antihypertensive medications. For sensitivity analysis, to account for potential imbalance between two groups, the Cox regression model was adjusted for all confounders, which were used in the matching model in the main study. We set up statistical significance as a two‐tailed p‐value < 0.05. All data analyses were conducted using SAS Version 9.4 and R Version 4.0. Results Baseline Characteristics After applying our inclusion and exclusion criteria, the first cohort included patients who received ACEI (n = 640,671) and ARB (n = 640,671); the second cohort included those with ACEI (n = 373,601) and dCCB (n = 373,601); and the third cohort included those with ARB (n = 408,491) and dCCB (n = 408,491) (Figure S1). Overall, baseline characteristics were well‐balanced between the groups (Table 1). Across the three cohorts, ~50% were over 60 years old, women, and lived in the Midwest. In this hypertensive population, dyslipidemia accounted for the highest proportion of comorbidities (< 60%), followed by DM with or without complications (< 30%). TABLE 1Baseline characteristics of the three study cohorts.Study cohort 1Study cohort 2Study cohort 3ACEIARBACEIdCCBARBdCCB N = 640,671 N = 640,671 N = 373,601 N = 373,601 N = 408,491 N = 408,491 N % N % N % N % N % N %Age (median, IQR)58 (50–64)58 (51–64)59 (51–65)58 (51–65)59 (51–65)59 (51–65)< 50 years144,10122.5135,78621.281,95521.983,00922.285,96321.090,33322.150–60 years217,28333.9216,30233.8119,26631.9118,96031.8131,20632.1128,99731.660–70 years180,86828.2186,66629.1105,32628.2104,62528.0117,93428.9114,67728.1≥ 70 years98,41915.4101,91715.967,05417.967,00717.973,38818.074,48418.2SexMen323,86650.6319,89849.9190,21350.9192,23951.5206,49950.6210,21551.5Women316,80549.4320,77350.1183,38849.1181,36248.5201,99249.4198,27648.5Geographical regionNorth90,22114.1109,97417.256,31915.162,48216.771,18217.466,66316.3Northeast168,93426.4134,07120.999,89926.789,74624.090,17322.1100,37524.6Midwest278,35343.4291,74645.5156,45641.9164,26344.0176,60243.2179,59444.0South91,45114.394,24614.755,68614.952,98114.265,63216.157,01014.0West11,7121.810,6341.752411.441291.149021.248491.2Insurance typeHMO79,14112.474,21111.645,56612.250,34613.549,72612.254,32713.3PPO356,14655.6354,17155.3203,94954.6196,81652.7220,38654.0217,07053.1Others205,38432.1212,28933.1124,08633.2126,43933.8138,37933.9137,09433.6Duration of hypertension diagnosis to medication initiation, days (median, IQR) a 548 (367–1313)570 (325–1532)630 (430–1559)655 (480–1862)620 (420–1875)643 (510–1785)Previous use of other AHTNBB84,65113.281,24312.755,18814.852,33014.055,48413.658,08414.2DU103,78616.299,68215.663,08716.959,15715.872,54417.867,54416.5Thiazide DU63,5879.960,7999.538,24510.239,52410.644,45710.945,20711.1Loop DU25,6194.025,2983.914,9364.011,3013.017,8304.413,8373.4K+ sparing DU22,3543.521,1923.314,0433.812,3363.315,2073.713,3043.3ComorbiditiesDyslipidemia374,56258.5378,29459.0210,26556.3209,86656.2238,64958.4233,82557.2DM w complications44,6107.047,8017.527,2257.327,8497.532,5338.033,1118.1DM w/o complications157,13524.5162,35225.382,38622.183,07722.293,72422.995,58623.4PVD42,1676.643,6226.828,1747.528,2087.632,2437.932,3707.9Ischemic stroke/TIA41,1426.441,5086.528,9737.828,5427.630,1177.431,5497.7Hemorrhagic stroke28590.427170.426040.727390.720660.525320.6Myocardial infarction18,8722.918,8532.911,4243.111,5093.113,1493.213,0223.2Angina61561.061811.049101.352081.458831.459951.5Heart failure31,1094.931,0574.817,4554.716,7364.519,8324.919,5654.8Atrial fibrillation31,8365.032,1685.016,9344.515,8494.218,6114.618,0594.4Chronic kidney disease28,7034.531,1814.920,9965.623,0246.227,9016.829,1097.1Renal failure62411.066411.045611.251761.457201.463351.6COPD38,6966.039,0746.124,2366.523,1976.226,4856.526,4136.5CCI (median, IQR)2 (1–3)2 (1–3)2 (1–3)2 (1–3)2 (1–3)2 (1–3)Other medication useLipid‐lowering agents116,55518.2113,13817.761,43016.461,42216.471,32517.571,17117.4Anti‐diabetic agents51,2618.050,3217.919,7005.320,4125.524,7116.025,4816.2Anti‐thrombotic agents14,8102.314,7072.380292.176352.087412.187402.1Anti‐platelet agents19,5943.119,3763.012,2623.312,4243.313,7733.414,0633.4Anti‐analgesic agents56,9998.956,3018.834,2079.233,1338.935,8318.835,7508.8Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; AHTN, antihypertensive drugs; ARB, angiotensin receptor blocker; BB, beta‐blocker; CCI, Charlson comorbidity index; COPD, chronic obstructive pulmonary disease; dCCB, dihydropyridine calcium channel blocker; DM, diabetes mellitus; DU, diuretic; HMO, Health Maintenance Organization; IQR, interquartile range; PPO, preferred provider organization; PVD, peripheral vascular disease; TIA, transient ischemic attack. a Duration of hypertension was calculated as time period (days) from the first diagnosis of hypertension to the first prescription date of index medication. Baseline characteristics of the three study cohorts. Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; AHTN, antihypertensive drugs; ARB, angiotensin receptor blocker; BB, beta‐blocker; CCI, Charlson comorbidity index; COPD, chronic obstructive pulmonary disease; dCCB, dihydropyridine calcium channel blocker; DM, diabetes mellitus; DU, diuretic; HMO, Health Maintenance Organization; IQR, interquartile range; PPO, preferred provider organization; PVD, peripheral vascular disease; TIA, transient ischemic attack. Duration of hypertension was calculated as time period (days) from the first diagnosis of hypertension to the first prescription date of index medication. Main Analyses Table 2 summarizes the results of the main analysis. In the first cohort, when comparing ARB with ACEI use, 1551 (0.24%) and 1578 (0.25%) kidney cancer events occurred in ACEI and ARB groups, respectively. ARB use was associated with an increased risk of kidney cancer compared with ACEI use (adjusted hazard ratio [HR] 1.10, 95% confidence interval [CI] 1.02–1.18). In the second cohort, when comparing dCCB with ACEI use, 894 (0.24%) and 1091 (0.29%) kidney cancer events occurred in ACEI and dCCB groups, respectively. The use of dCCB showed a higher risk of kidney cancer compared with ACEI use (HR 1.29, 95% CI 1.18–1.40). In the third cohort, when comparing dCCB with ARB use, 964 (0.24%) and 1166 (0.29%) kidney cancer events occurred in ARB and dCCB groups, respectively. The use of dCCB had an elevated risk of kidney cancer compared with ARB use (HR 1.17, 95% CI 1.08–1.28). TABLE 2Risks of kidney cancer associated with the use of ACEI, ARB, and dCCB in the three cohorts.Study cohort 1ACEI (N = 640,671)ARB (N = 640,671)ARB vs. ACEIEvent, N %PYIR per 100,000 PYEvent, N %PYIR per 100,000 PYUnadjusted HR (95% CI)Adjusted HR a (95% CI)15510.242,062,283.475.21,5780.252,070,866.176.21.10 (1.03–1.18)1.10 (1.02–1.18)Study cohort 2ACEI (N = 373,601)dCCB (N = 373,601)dCCB vs. ACEIEvent, N %PYIR per 100,000 PYEvent, N %PYIR per 100,000 PYUnadjusted HR (95% CI)Adjusted HR a (95% CI)8940.24879,996.0101.61,0910.29913,761.4119.41.30 (1.19–1.42)1.29 (1.18–1.40)Study cohort 3ARB (N = 408,491)dCCB (N = 408,491)dCCB vs. ARBEvent, N %PYIR per 100,000 PYEvent, N %PYIR per 100,000 PYUnadjusted HR (95% CI)Adjusted HR a (95% CI)9640.24957,562.9100.71,1660.29964,669.1120.91.18 (1.08–1.28)1.17 (1.08–1.28)Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CI, confidence interval; dCCB, dihydropyridine calcium channel blocker; HR, hazard ratio; IR, incidence rate; PY, person‐years. a The model was adjusted for Charlson comorbidity index. Risks of kidney cancer associated with the use of ACEI, ARB, and dCCB in the three cohorts. Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CI, confidence interval; dCCB, dihydropyridine calcium channel blocker; HR, hazard ratio; IR, incidence rate; PY, person‐years. The model was adjusted for Charlson comorbidity index. Subgroup and Sensitivity Analyses Across the three cohorts, results from the subgroup analyses supported the main findings (Figure 1 and Tables S3–S6). When it comes to renal failure, when comparing dCCB to ACEI use, a greater risk was observed among patients without renal failure than those with renal failure (p for difference = 0.04). When comparing dCCB to ARB use, similar results were observed among patients without and with renal failure (p for difference = 0.02). Similarly, compared to ACEI or ARB use, the use of dCCB showed a higher risk of kidney cancer among patients without COPD. Among those without COPD, null associations were observed when compared dCCB use to ACEI or ARB use. According to the medication exposure, consistent results supporting the main findings were shown with a trend toward a higher cDDD associated with an increased risk of kidney cancer (Figure 2 and Table S7). The results from the sensitivity analyses showed robust results, regardless of constructing different study populations, applying a lag time, or considering termination of treatment (Table 3). Forest plots of subgroup analyses for kidney cancer risk associated with ACEI, ARB, and dCCB. Across the three cohorts, results from the subgroup analyses generally supported the main findings. In the second and third cohorts, a greater risk was observed among patients without renal failure/COPD than those with renal failure/COPD. The model was adjusted for Charlson comorbidity index. ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CI, confidence interval; COPD, chronic obstructive pulmonary disease; dCCB, dihydropyridine calcium channel blocker; DM, diabetes mellitus; HR, hazard ratio; PVD, peripheral vascular disease; TIA, transient ischemic attack. Plots of exposure to antihypertensive medications with the risk of kidney cancer. According to the medication exposure, consistent trends supporting the main finding were shown. Stronger hazard ratios were observed in the highest exposure group across the three study cohorts. The hazard ratios were adjusted for Charlson comorbidity index. ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; cDDD, cumulative defined daily dose; CI, confidence interval; dCCB, dihydropyridine calcium channel blocker; HR, hazard ratio. TABLE 3Sensitivity analyses for kidney cancer risk associated with the use of ACEI, ARB, and dCCB.Study cohort 1ACEIARBARB vs. ACEI N Event, N % N Event, N %Adjusted HR (95% CI) a Main analysis640,67115510.24640,67115780.251.10 (1.02–1.18)Redefinition of new AHTN use479,63510950.23484,20511610.241.11 (1.02–1.20)Long‐term use (1000+ cDDD)185,8986630.36186,8087070.381.12 (1.02–1.25)Lag time (2 years)640,6718700.14640,6718530.131.11 (1.01–1.22)Without compelling indications402,4877420.18395,2187960.201.14 (1.03–1.26)As‐treated approach640,67110190.16640,67111370.181.03 (0.96–1.11)Without propensity score matching816,30118820.23726,04518960.261.11 (1.04–1.20)Study cohort 2ACEIdCCBdCCB vs. ACEI N Event, N % N Event, N %Adjusted HR (95% CI) a Main analysis373,6018940.24373,60110910.291.29 (1.18–1.40)Redefinition of new AHTN use255,6625350.21256,9087290.281.35 (1.21–1.52)Long‐term use (1000+ cDDD)98,5323600.37101,8304800.471.30 (1.15–1.47)Lag time (2 years)373,2054980.13373,0175070.141.20 (1.06–1.28)Without compelling indications235,3644150.18235,7535600.241.38 (1.22–1.64)As‐treated approach373,6015630.15373,6018500.231.16 (1.07–1.26)Without propensity score matching749,99016160.22470,34813850.291.30 (1.20–1.41)Study cohort 3ACEIdCCBdCCB vs. ARB N Event, N % N Event, N %Adjusted HR (95% CI) a Main analysis408,4919640.24408,49111660.291.17 (1.08–1.28)Redefinition of new AHTN use242,5425660.23245,7006830.281.17 (1.05–1.32)Long‐term use (1000+ cDDD)107,4844060.38105,9345110.481.22 (1.07–1.38)Lag time (2 years)408,0285010.12407,8485230.131.05 (0.99–1.12)Without compelling indications253,4484720.19252,3155800.231.20 (1.10–1.30)As‐treated approach408,4916860.17408,4919130.221.18 (1.08–1.28)Without propensity score matching727,06517830.25492,55814600.301.18 (1.10–1.27)Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; cDDD, cumulative defined daily dose; CI, confidence interval; dCCB, dihydropyridine calcium channel blocker; HR, hazard ratio. a For sensitivity analysis, the model was adjusted for all confounders, which were used in the matching model in the main study, to account for potential imbalance between two groups. Sensitivity analyses for kidney cancer risk associated with the use of ACEI, ARB, and dCCB. Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; cDDD, cumulative defined daily dose; CI, confidence interval; dCCB, dihydropyridine calcium channel blocker; HR, hazard ratio. For sensitivity analysis, the model was adjusted for all confounders, which were used in the matching model in the main study, to account for potential imbalance between two groups. Repeated Analysis: Comparing Non‐dCCB Use With ACEI or ARB Use The results of the additional analyses comparing non‐dCCB use with ACEI or ARB use were presented in Tables S8–S13. Compared non‐dCCB use to ACEI or ARB use, no significant differences in the risk of kidney cancer were observed. Discussion To our knowledge, this is the first long‐term, large PS‐matched retrospective cohort study that investigated the risk of kidney cancer across the three most commonly used antihypertensive medications in hypertensive population using head‐to‐head comparison design. We observed that dCCB use was associated with an increased risk of kidney cancer compared to ACEI or ARB use. In addition, ARB use was associated with a higher risk of kidney cancer compared to ACEI use. In subgroup analyses, similar associations were generally observed, whereas null associations were shown in those with renal failure or COPD. Robust results from the sensitivity analyses supported the main findings. In contrast, there was no significant association of non‐dCCB use with kidney cancer risk compared to ACEI or ARB use. Although no previous cohort study has directly compared the risk of kidney cancer between different classes of antihypertensive medications using PS matching, several case–control studies have supported our findings. In one case–control study, dCCB use was associated with a 30% higher risk of renal cell carcinoma (RCC) (odds ratio [OR] 1.32, 95% CI 1.05–1.67) compared to ARB use. Compared to ACEI use, dCCB use showed a trend of an increased risk of RCC (OR 1.21, 95% CI 0.95–1.53) [13]. In another case–control study conducted in a hypertensive population, CCB, but not differentiating dCCB and non‐dCCB, use was associated with a two‐fold higher risk of kidney cancer compared with non‐CCB, combining ACEI, BB, and DU [24]. Although they applied an active comparator design, two major limitations exist in these studies. First, due to the exposure information collected based on the outcome occurred in a case–control study design, there may be a lack of representative exposure distributions, and this leads to potential selection bias and lack of temporal associations. Second, since no information regarding the initiation of exposure to the medication was identified, indication bias and prevalent user bias could influence their results. When compared to non‐use, various classes of antihypertensive medication use were differently associated with 10%–50% increased risks of kidney cancer [25]. Although the results supported our findings that CCB had the higher risk of kidney cancer among different classes, it is important to note that comparing non‐use is susceptible to introduce indication bias. Contrary to these studies, the current PS‐matched cohort study provides more reliable relative risks with reduced bias. In this study, dCCB use was found to be associated with the highest risk of kidney cancer compared to ACEI and ARB use. This study contributes to the existing literature and clinical practice by providing evidence that medication selection may be a modifiable factor in minimizing a higher risk of kidney cancer in a hypertensive population, who need pharmacotherapy but are already at a high risk for kidney cancer, while simultaneously providing optimal hypertension care in accordance with the current guidelines. Although similar results were observed across the pre‐defined subgroups, the associations were different according to the presence of renal failure or COPD. In patients with renal failure or COPD, no differences in the risk of kidney cancer were observed when comparing dCCB use to ACEI or ARB use. Additional risk from selecting antihypertensive medication use may be likely small in hypertensive patients with renal failure or COPD, who are already susceptible to a high risk of kidney cancer. Renal failure has been associated with a three‐ to ten‐fold increased risk of kidney cancer compared to the general population [26, 27, 28]. COPD history was used as a proxy of smoking [29], which is a strong risk factor for kidney cancer [4, 30, 31, 32]. Potential mechanisms underpinning carcinogenesis of antihypertensive medications are available, while stronger evidence exists for CCB compared to ACEI or ARB. Several studies found that CCB may allow for an accumulation of damaged cells such as potentially malignant genetic variants and trigger calcium‐mediated cell differentiation and apoptosis [10, 33, 34]. On the other hand, ACEI or ARB has conflicting evidence of both carcinogenic and anti‐cancer effects. ACEI may contribute to developing kidney cancer via increasing bradykinin levels [35], whereas ARB may promote cellular proliferation or angiogenesis and lead to angiotensin‐mediated tumor progression [11]. In contrast, ACEI or ARB may down‐regulate vascular endothelial growth factor‐mediated angiogenesis through angiotensin II and angiotensin receptors, which could potentially have anti‐cancer effects for skin, urogenital, or breast cancers [36, 37, 38]. Since our study compared dCCB and ACEI or ARB directly, we were unable to rule out the possibility that the observed increase in kidney cancer risk with dCCB may partly result from ACEI or ARB reducing the risk, rather than solely from dCCB increasing it. Therefore, we suggest future studies to clarify the potential carcinogenic effects of CCB compared to the protective effects of ACEI or ARB on kidney cancer risk. In fact, the current study found differences in kidney cancer risk when comparing dCCB with ACEI or ARB, while we observed null associations for non‐dCCB. As limited experimental studies that distinguish between dCCB and non‐dCCB are available, further research is warranted to explore their potentially distinct carcinogenic mechanisms. For patients with hypertension, who are already at‐risk of developing kidney cancer but need to control their blood pressure using antihypertensive medications, every effort to provide evidence‐based prevention strategies for kidney cancer is necessary. Given their life‐long use and various mechanisms of action, the choice of antihypertensive medications might influence the risk of kidney cancer differently. Until now, potential carcinogenic mechanisms of dCCB have been less extensively investigated. Our findings emphasize the need for further investigation into potential biological links between dCCB use and kidney cancer development. Our study has several strengths. First, our study with a long‐term, large analytic cohort allowed us to reduce selection and recall bias and apply systematic approaches to assess the robustness using several subgroup and sensitivity analyses. Second, our PS‐matched cohort reduced indication bias by balancing characteristics between study groups, while the new user, active comparator design minimized prevalent user bias by identifying treatment initiators, unlike previous case–control studies. Third, comprehensive diagnosis and prescription data helped us estimate cumulative exposure to each medication over the study period. Also, we were able to account for hypertensive complications or other potential compelling indications of medications and adherence to the medications. We were further able to account for other medications, such as analgesics, anti‐diabetic agents, or lipid‐lowering agents, which might be associated with a risk of kidney cancer [12, 21]. Our findings suggest that given that antihypertensive drugs are used over long‐term periods, the risk of kidney cancer should be carefully considered when selecting the primary agent for hypertension treatment. Several limitations should also be noted. First, due to the nature of observational studies using claims data, we were unable to confirm whether patients took the prescribed medications. To mitigate this concern, we restricted the study to patients who had been prescribed antihypertensive medications for more than 6 months with an adherent use, indicating their regular use. In addition, we were unable to investigate the associations based on lifestyle factors including smoking status or body mass index, which are known risk factors for kidney cancer. Also, there could be unmeasured confounders related to severity of hypertension, potentially influencing the impacts of antihypertensive medications on kidney cancer. Indeed, it is impossible to separate the influence of high blood pressure or hypertension severity from the effects of antihypertensive medications on kidney cancer. More severe hypertension could be associated with higher cumulative dose and longer use of medications, as well as kidney cancer risk. However, in this PS‐matched study with an active comparator, this potential confounding effect is largely nondifferential across the study groups and usually biases the results toward null. Robust findings from dose–response relationship, along with different subgroup and sensitivity analyses including only longer users, suggest a possible association of different antihypertensive medication use with an elevated risk of kidney cancer, independent of hypertension. Second, we were unable to fully reflect on real‐world settings, which often involve more complex practices such as switching to another medication class or using combination therapies for patients with uncontrolled or severe hypertensive or for those with different indications. Although it may seem unrealistic to conduct observational studies to gather evidence on this matter, our findings suggest that dCCB have potential carcinogenic effects for kidney cancer that differ from those of other classes. Further studies are needed to explore the associations between different combination therapies and kidney cancer risk. Lastly, our study patients from the MarketScan Databases may remain relatively stable in their enrollment in health insurance in comparison to the general population [17]. Accordingly, the population may not fully represent individuals with a low socioeconomic status. In conclusion, we found that dCCB use was associated with an increased risk of kidney cancer compared to ACEI or ARB use in a hypertensive population. Given their long‐term use and the fact that patients with hypertension are already at‐risk of developing kidney cancer, the choice of antihypertensive medications may be one of modifiable factors in minimizing the potential risk of kidney cancer while also controlling blood pressure and preventing cardiovascular diseases. Author Contributions Minji Jung: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (equal), software (lead), visualization (lead), writing – original draft (lead), writing – review and editing (lead). Shufeng Li: methodology (supporting). Zhengyi Deng: writing – review and editing (supporting). Jinhui Li: writing – review and editing (supporting). Mingyi Li: writing – review and editing (supporting). Satvir Basran: writing – review and editing (supporting). Marvin E. Langston: conceptualization (equal), investigation (equal), methodology (equal), project administration (lead), supervision (lead), validation (equal), writing – review and editing (lead). Benjamin I. Chung: conceptualization (equal), project administration (lead), supervision (lead), writing – review and editing (supporting). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Data S1. Acknowledgments Data for this project were accessed using the Stanford Center for Population Health Sciences (PHS) Data Core. The PHS Data Core is supported by the National Institutes of Health (NIH) National Center for Advancing Translational Science Clinical and Translational Science Award (UL1TR003142) and from Internal Stanford funding. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Funding: The authors received no specific funding for this work. Marvin E. Langston and Benjamin I. Chung contributed equally to the manuscript as the senior authors. Data Availability Statement The study used the Merative MarketScan Commercial and Medicare Databases. The authors cannot legally distribute these data, but details on data access can be found here: https://www.merative.com/content/dam/merative/documents/brief/marketscan‐explainer‐general.pdf. References Associated Data Supplementary Materials Data S1. Data Availability Statement The study used the Merative MarketScan Commercial and Medicare Databases. The authors cannot legally distribute these data, but details on data access can be found here: https://www.merative.com/content/dam/merative/documents/brief/marketscan‐explainer‐general.pdf.