Brain penetrant calcium channel blockers do not reduce alcohol consumption: Converging results from two large independent cohort studies using electronic health records

Despite compelling genetic and preclinical evidence, two large, preregistered observational studies using electronic health records found that initiating brain-penetrant L-type calcium channel blockers does not reduce alcohol consumption compared to non-penetrant alternatives or unexposed controls, indicating these drugs are not effective for repurposing in alcohol use disorder treatment.

The Gist

The Big Question: Can Blood Pressure Pills Stop the Drinking?

Imagine you have a very popular, well-tested tool in your toolbox: a blood pressure pill called a Calcium Channel Blocker (specifically, the kind that can cross the "brain barrier"). Scientists had a hunch that this tool might also work as a "lock" for alcohol cravings.

Why? Because in the lab (on mice) and in our DNA, there were strong clues suggesting that these pills could calm down the brain's "reward center"—the part that screams, "I need another drink!" It was like finding a key that looked like it fit a different lock than the one it was originally made for. This is called drug repurposing: taking an old, safe medicine and trying to use it for a new problem.

The Experiment: Two Giant Test Kitchens

To see if this hunch was true, the researchers didn't just guess; they cooked up a massive experiment using real-world data from two giant "test kitchens":

1. The VA (Veterans Affairs): A massive healthcare system for US veterans.
2. KPNC (Kaiser Permanente): A huge healthcare system in Northern California.

They looked at millions of patient records (like checking millions of receipts) to see what happened when people started taking these specific brain-penetrating blood pressure pills.

The Setup:
They created three groups of people to compare:

• Group A: People who started taking the "Brain-Penetrating" blood pressure pills (the potential alcohol cure).
• Group B: People who started taking a different blood pressure pill that doesn't penetrate the brain (the control group).
• Group C: People who didn't take either pill (the baseline group).

They made sure these groups were "twins" in terms of age, health, and drinking habits before the study started, using a statistical magic trick called Propensity Score Matching. This ensured that if Group A drank less later, it was actually because of the pill, not because they were healthier to begin with.

The Results: The "Magic Key" Didn't Fit

The researchers waited to see if Group A (the brain-penetrating pill takers) would drink significantly less alcohol than Groups B and C.

The verdict?
Nothing happened.

It was like handing someone a "magic key" that everyone thought would open the "alcohol craving door," but when they tried it, the door didn't budge.

• People who took the brain-penetrating pills reduced their drinking, yes—but so did the people who took the non-brain-penetrating pills, and so did the people who took no pills at all.
• The difference between the groups was so tiny it was basically zero.

Why Did This Happen? (The "Brain Barrier" Mystery)

The paper suggests a few reasons why the "magic" didn't work in real life, even though it worked in the lab:

1. Lab vs. Reality: Mice in a cage are different from humans with complex lives, stress, and social habits. What works on a tiny brain in a petri dish doesn't always work on a human brain in the real world.
2. The "Brain Barrier" isn't a Wall, it's a Sieve: The researchers noted something interesting. The "control" pill (amlodipine), which was supposed to stay out of the brain, might actually sneak in a little bit more than we thought. If the "control" group is also getting a little bit of the brain effect, it makes it harder to see a difference between the groups.
3. Wrong Target: Maybe the specific "lock" (the calcium channel) isn't the main thing driving alcohol addiction in humans, even if the DNA clues suggested it was.

The Takeaway for Everyone

Think of this study as a very careful, high-stakes reality check.

For a long time, scientists were excited about using these blood pressure pills to treat alcohol use disorder. They had great genetic clues and animal studies to back it up. But this massive study, looking at hundreds of thousands of real people, said: "Stop the presses. This specific approach doesn't seem to work for reducing drinking."

What does this mean for the future?
It's not a failure; it's a success of science. It saves us from wasting time and money on a treatment that doesn't work. Now, researchers know they need to look for a different key to open the door to treating alcohol addiction. They have to go back to the drawing board and find new targets to help people struggling with alcohol use.

In short: The "brain-penetrating blood pressure pill" turned out to be a red herring for alcohol treatment. It's safe, it lowers blood pressure, but it doesn't stop the drinking.

Technical Summary

1. Problem Statement

Alcohol Use Disorder (AUD) represents a significant public health burden, with approximately 27.1 million adults in the US diagnosed with past-year AUD. Current pharmacotherapies (e.g., naltrexone, acamprosate) have limited efficacy, low utilization rates (<2.4%), and suboptimal adherence (<55%).

The Hypothesis:
L-type calcium channel blockers (LTCCBs) are a candidate class for drug repurposing in AUD treatment based on:

• Genetic Evidence: Genome-wide association studies (GWAS) link voltage-gated calcium channel subunit genes (CACNA1C, CACNA1E, CACNA2D3) to alcohol consumption and psychiatric disorders.
• Preclinical Evidence: Animal studies suggest LTCCBs (e.g., isradipine, nimodipine) reduce addiction-related behaviors (conditioned place preference, self-administration) by modulating dopamine activity in the mesolimbic and nigrostriatal pathways.
• Theoretical Mechanism: It was hypothesized that brain-penetrant (BP) LTCCBs (e.g., nifedipine, felodipine) would be more effective than non-brain-penetrant (non-BP) LTCCBs (e.g., amlodipine) in reducing alcohol consumption.

2. Methodology

The study employed a rigorous, preregistered (OSF: huawv) observational cohort design using Electronic Health Records (EHR) from two distinct, large US healthcare systems:

1. US Department of Veterans Affairs (VA): ~9 million patients annually.
2. Kaiser Permanente Northern California (KPNC): ~4.6 million members.

Study Design & Population:

• Exposure Groups:
  1. BP-LTCCBs: New users of nifedipine or felodipine.
  2. Non-BP-LTCCBs: New users of amlodipine (active comparator with limited brain penetrance).
  3. Unexposed: Patients from the same clinics with no LTCCB exposure.
• Inclusion Criteria: Adults (≥18), ≥1 outpatient visit in the prior year, self-reported alcohol use in the prior 2 years, and a 180-day washout period to ensure "new user" status.
• Follow-up: Up to 2 years post-index date.
• Covariates: Extensive adjustment for demographics, comorbidities (Charlson Index, VACS Index), substance use history, and concurrent medications (including other AUD treatments).

Statistical Analysis:

• Propensity Score Matching (PSM): Conducted separately for three contrasts:
  • BP-LTCCB vs. Unexposed
  • Non-BP-LTCCB vs. Unexposed
  • BP-LTCCB vs. Non-BP-LTCCB
  • Matching variables: Sex, AUD diagnosis, and baseline alcohol use were exactly matched; other covariates were matched on the logit of the propensity score (caliper 0.20 SD).
• Primary Outcome: Change in drinks per week from the most recent pre-index screen to the end of follow-up.
• Analytical Model: Multivariable Difference-in-Differences (DiD) linear regression to estimate the differential change in drinking between groups, controlling for baseline levels and time trends.
• Subgroup Analyses: Stratified by AUD diagnosis, baseline drinking level (light, moderate, heavy), and sex.

3. Key Results

The study analyzed data from both cohorts after propensity matching and outcome restriction.

VA Cohort Results:

• BP-LTCCB vs. Unexposed: No significant difference in reduction of drinks per week (DiD: 0.00, 95% CI: -0.18 to 0.19, p=0.97).
• BP-LTCCB vs. Non-BP-LTCCB: No significant difference (DiD: 0.14, 95% CI: -0.06 to 0.34, p=0.16).
• Subgroups: No differential effects observed across AUD diagnosis, sex, or baseline drinking levels.

KPNC Cohort Results:

• BP-LTCCB vs. Unexposed: No significant difference (DiD: 0.16, 95% CI: -0.67 to 0.99, p=0.70).
• BP-LTCCB vs. Non-BP-LTCCB: No significant difference (DiD: 0.31, 95% CI: -0.48 to 1.11, p=0.44).
• Subgroups: Consistent null findings across all subgroups.

Notable Observation:
In the VA study, the non-BP-LTCCB (amlodipine) group showed slightly greater reductions in drinking compared to the unexposed group, though the effect size was clinically negligible (DiD < 0.5). The authors note that recent literature suggests amlodipine may have higher brain penetrance than previously assumed, complicating the "brain-penetrant" vs. "non-penetrant" dichotomy.

4. Key Contributions

• Convergent Evidence: This is the first large-scale study to test the repurposing of BP-LTCCBs for AUD using two independent, diverse healthcare systems (VA and KPNC), yielding consistent null results.
• Rigorous Design: The study utilized preregistration, active comparators (amlodipine), and unexposed controls, alongside robust propensity score matching to minimize confounding by indication.
• Quantitative Outcomes: Unlike many EHR studies relying on diagnostic codes, this study utilized validated, quantitative alcohol screening tools (AUDIT-C in VA; structured questions in KPNC) to measure actual changes in consumption volume.
• Clarification of Mechanism: The results challenge the translational validity of preclinical LTCCB findings for AUD, suggesting that genetic and animal model support does not necessarily translate to clinical efficacy in human populations.

5. Significance and Implications

• Clinical Impact: The study provides strong evidence against repurposing brain-penetrant LTCCBs (nifedipine, felodipine) for the treatment of Alcohol Use Disorder. Clinicians should not prescribe these medications for the primary purpose of reducing alcohol consumption.
• Research Direction: The findings highlight the "translational gap" between preclinical models and human clinical outcomes for addiction pharmacotherapy. Future research should prioritize alternative neurobiological targets or etiologically informed subgroups rather than broad repurposing of LTCCBs.
• Methodological Benchmark: The study demonstrates the power of large-scale, preregistered, multi-cohort EHR analyses to rapidly evaluate drug repurposing hypotheses, potentially saving resources on ineffective clinical trials.

Conclusion:
Despite compelling genetic and preclinical data, brain-penetrant L-type calcium channel blockers do not reduce alcohol consumption in real-world clinical populations compared to non-penetrant analogs or unexposed controls. The study concludes that these agents are not viable candidates for AUD pharmacotherapy.