Regulatory Path

The Honest Version

This document describes what stands between SmartShot and a legal product. The answer is years of work, tens of millions of dollars, and a regulatory gauntlet that has broken companies with more resources than ours. Nothing here is sugarcoated. The FDA does not care about your pitch deck.

FDA Device Classification

SmartShot is almost certainly a Class III medical device. Class III is reserved for devices that sustain or support life, are implanted, or present a potential unreasonable risk of illness or injury. A wearable that autonomously injects medication into a human body without conscious patient action meets all three criteria.

The two available pathways for Class III:

Pre-Market Approval (PMA). The most rigorous FDA pathway. Requires valid scientific evidence of safety and effectiveness, typically from clinical trials. Average PMA review takes 180 days after submission, but the years of pre-submission work dwarf the review itself. PMA is the likely path for SmartShot.

De Novo Classification. Available when no predicate device exists and the device does not automatically default to Class III risk. De Novo creates a new regulatory classification and can establish SmartShot as the predicate for future competitors. Biolinq's Shine biosensor received De Novo classification in 2025 for its autonomous wearable glucose sensor, establishing a new device category. SmartShot could pursue De Novo for the detection-and-monitoring subsystem while the drug delivery function follows PMA.

The harder question is combination product designation. SmartShot pairs a drug (naloxone, epinephrine, nitroglycerin, or glucagon) with a device (wearable sensor and injection system). Under 21 CFR Part 3, the FDA's Office of Combination Products determines which center leads review based on the product's primary mode of action (PMOA).

If the drug's therapeutic action is deemed primary, CDER leads. If the device's delivery mechanism is deemed primary, CDRH leads. For SmartShot, the argument cuts both ways. The drug saves the life. The device decides when to deliver it.

The November 2025 Pre-RFD guidance update clarifies how sponsors can get preliminary feedback on classification and center assignment before formal submission. SmartShot should file a Pre-RFD request early. Getting the wrong center assignment can add years.

Precedent: Lasix ONYU (furosemide injection), approved October 2025, is a wearable drug-device combination for subcutaneous delivery in heart failure. It was reviewed as a combination product. SmartShot is more complex — ONYU is patient-activated, not autonomous — but the regulatory architecture applies.

Applicable Standards

Every standard listed here will require documented compliance before FDA submission. This is not optional.

ISO 13485:2016 — Quality Management Systems for Medical Devices. The foundation. As of February 2, 2026, the FDA's Quality Management System Regulation (QMSR) formally incorporates ISO 13485 by reference into 21 CFR Part 820. SmartShot's quality system must be ISO 13485-compliant from day one. There is no grandfathering.

ISO 14971:2019 — Application of Risk Management to Medical Devices. Requires a complete risk management file covering the entire product lifecycle. For SmartShot, this means documenting every failure mode of the detection algorithm, every scenario where a false positive injection could harm, every drug stability risk, every sensor degradation path. The risk management file will be the largest single document in the submission.

IEC 62304:2006+A1:2015 — Medical Device Software Lifecycle Processes. SmartShot's AI/ML detection algorithm is safety-critical software. IEC 62304 requires documented software development planning, architecture, detailed design, unit testing, integration testing, and system testing. The algorithm's classification under IEC 62304 will almost certainly be Class C (can cause death or serious injury), which triggers the most rigorous documentation requirements.

IEC 60601-1:2020 — Medical Electrical Equipment Safety. Covers electrical safety, electromagnetic compatibility, and essential performance. SmartShot must demonstrate that its electrical systems do not interfere with, or become interfered by, other medical devices. The wearable form factor adds specific requirements for patient-contact materials, battery containment, and thermal management.

ISO 10993 (series) — Biological Evaluation of Medical Devices. Any material that contacts the patient's skin or enters the body requires biocompatibility testing. For SmartShot, this includes the injection needle, the adhesive patch, any housing materials against skin, and the drug container closure system. Testing covers cytotoxicity, sensitization, irritation, and potentially implantation studies for the needle pathway.

ISO 11608-1:2022 — Needle-Based Injection Systems for Medical Use. Directly applicable. Covers requirements and test methods for needle-based injection systems including dose accuracy, activation force, injection time, needle guard lockout, and holding force. SmartShot's automated injection mechanism must meet every requirement in this standard despite having no manual activation step, which creates novel compliance questions.

ISO 11608-5:2022 — Automated Functions. This part of the standard specifically addresses automated functions in needle-based injection systems. It was written for autoinjectors with patient-initiated triggers. SmartShot's fully autonomous trigger is beyond current scope, which means additional special controls.

FDA QMSR Final Rule. Effective February 2, 2026. Replaces the old Quality System Regulation. Aligns U.S. requirements with ISO 13485. SmartShot benefits from this: one quality system serves both FDA and international regulatory submissions.

Section 524B — Ensuring Cybersecurity of Devices. Added to the FD&C Act through the Food and Drug Omnibus Reform Act. As of October 2025, the FDA may reject submissions for "cyber devices" that lack a Software Bill of Materials (SBOM). SmartShot is definitionally a cyber device: it contains software, connects wirelessly, and transmits patient data. Required elements include a Secure Product Development Framework (SPDF), SBOM, ongoing vulnerability monitoring, and a patch management plan.

FDA AI/ML Guidance — Artificial Intelligence-Enabled Device Software Functions. Draft guidance published January 6, 2025. Final guidance on Predetermined Change Control Plans (PCCPs) published August 2025. PCCPs allow pre-authorized algorithm modifications without new submissions. This is critical for SmartShot because the detection algorithm will need continuous improvement based on real-world performance data.

Clinical Trial Design

The FDA will require clinical evidence that SmartShot does two things safely: detects the emergency correctly and delivers the drug appropriately. Each failure mode is different.

Primary Endpoints

Detection accuracy: sensitivity (true positive rate) and specificity (true negative rate). For a life-sustaining device, the FDA will expect sensitivity above 95% and specificity above 99%. A false negative means a patient dies. A false positive means an unnecessary injection, which ranges from uncomfortable (naloxone) to dangerous (epinephrine in a patient without anaphylaxis).

Drug delivery: bioavailability compared to standard administration route. Lasix ONYU demonstrated 112% bioavailability versus IV administration. SmartShot will need equivalent pharmacokinetic studies for each drug.

Time to intervention: measured from physiological threshold crossing to drug delivery. This is SmartShot's core value proposition and must be quantified against standard-of-care response times.

Study Population

Each drug targets a different population. Naloxone trials will recruit people who use opioids, likely through harm reduction programs and medication-assisted treatment clinics. Epinephrine trials will recruit patients with documented anaphylaxis history. Nitroglycerin trials will recruit patients with established coronary artery disease. Glucagon trials will recruit patients with type 1 diabetes and hypoglycemia unawareness.

Sample sizes will depend on event rates. Opioid overdose in active users occurs frequently enough to power a trial in 12-18 months. Anaphylaxis events are rarer, requiring larger enrollment or longer observation.

Safety Monitoring

The critical safety question is false positive injection. An independent Data Safety Monitoring Board (DSMB) will be mandatory. Stopping rules must be pre-specified for:

• False positive injection rate exceeding a threshold (drug-specific)
• Injection-site complications (infection, tissue damage, hematoma)
• Device malfunction (sensor failure, needle deployment failure, drug degradation)
• Algorithm drift (detection accuracy declining over time)

Comparator Arms

Standard of care comparison is the only design the FDA will accept. For naloxone, the comparator is bystander-administered nasal naloxone. For epinephrine, the comparator is manual EpiPen. For nitroglycerin, the comparator is patient-initiated sublingual tablet. For glucagon, the comparator is bystander-administered glucagon injection.

The primary hypothesis for each trial: SmartShot delivers the drug faster than human-initiated administration, with non-inferior or superior clinical outcomes, and an acceptable false positive rate.

Each Drug = Separate Pathway

This is the part most people underestimate. SmartShot is not one product. It is four products wearing the same hardware. Each drug requires its own regulatory submission, its own clinical trial, its own safety data, and its own FDA review.

Naloxone — First to Market

Most permissive regulatory environment. Naloxone has OTC status. The FDA has explicitly prioritized opioid overdose interventions. SAMHSA funding supports harm reduction technology. The detection signal (respiratory depression via SpO2 and respiratory rate) is the most technically mature biosensor application.

Risk profile of false positives is low. An unnecessary naloxone injection in someone not experiencing opioid overdose causes minimal harm — brief withdrawal symptoms in opioid-dependent individuals, essentially no effect in opioid-naive individuals.

This is the beachhead.

Epinephrine — Established but Novel

Epinephrine autoinjectors are standard of care for anaphylaxis. The drug is well-understood. But every existing device requires manual activation by a conscious person. Autonomous delivery is new territory.

Detection complexity is higher than naloxone. Anaphylaxis presents with variable symptom profiles: some patients present with primarily respiratory symptoms, others with cardiovascular collapse, others with cutaneous signs only. The algorithm must handle all presentations.

False positive risk is moderate. Unnecessary epinephrine causes tachycardia, hypertension, and anxiety. Dangerous in patients with underlying cardiac conditions but rarely fatal in otherwise healthy individuals.

Nitroglycerin — No Precedent

No device has ever delivered cardiac medication autonomously to a non-hospitalized patient. The Zoll LifeVest, approved in 2001, proved that autonomous cardiac intervention from a wearable is regulatorily viable — but defibrillation is not drug delivery.

Detection requires accurate arrhythmia identification and ST-segment analysis from a wearable form factor. Hospital-grade ECG uses 12 leads. Wearable ECG typically uses 1-3 leads. The accuracy gap is real.

False positive risk is significant. Nitroglycerin causes hypotension. In a patient experiencing vasovagal syncope rather than a cardiac event, nitroglycerin could worsen the episode. The specificity bar will be the highest of any SmartShot application.

Glucagon — Lowest Priority

The closed-loop insulin delivery market is mature. Medtronic, Insulet, Tandem, and Sequel have FDA-cleared systems. SmartShot would not compete on insulin delivery.

Glucagon for severe hypoglycemia is a narrower application. The detection signal (glucose level via CGM integration) is well-established, but requires partnership with or licensing from a CGM manufacturer.

This indication is strategically valuable but not urgent.

Regulatory Precedents

These are the devices and timelines that tell us what to expect.

Medtronic MiniMed 670G — First hybrid closed-loop insulin delivery system. Approved September 28, 2016. Development spanned over a decade. Breakthrough T1D (formerly JDRF) stated their involvement cut five years off the approval timeline. Even with advocacy support and Breakthrough Device designation, the path from concept to clearance exceeded ten years.

Medtronic MiniMed 780G — Second-generation closed-loop system. Approved in the EU in 2020. U.S. FDA approval delayed until April 21, 2023, partly due to manufacturing quality issues at Medtronic's facility. Timeline from 670G approval to 780G approval: seven years. Second-generation devices are not automatic.

Omnipod 5 (Insulet) — First tubeless closed-loop insulin system. FDA clearance January 28, 2022. Submission to clearance took approximately one year, but the preceding development and clinical trial period spanned years. The 510(k) pathway was available because the 670G established a predicate. SmartShot will not have that luxury for its first indication.

Zoll LifeVest — Wearable cardioverter defibrillator. Original FDA approval in 2001 for autonomous cardiac defibrillation. Proves that a wearable device can autonomously deliver a life-saving intervention without conscious patient action. Now in its fifth generation (launched December 2025). The LifeVest is the closest conceptual precedent to SmartShot: sense a cardiac emergency, intervene automatically. The difference: defibrillation is an electrical intervention, not a drug.

Lasix ONYU (SQ Innovation) — FDA approved October 2025. Wearable subcutaneous furosemide delivery for heart failure edema. First wearable drug-device combination product for cardiac care. Demonstrated 112% bioavailability versus IV Lasix. Uses a reusable unit (48 treatments) and single-use disposable component. Patient-activated, not autonomous. But the form factor, the regulatory classification, and the combination product designation are directly relevant to SmartShot's path.

The pattern across all precedents: plan for a decade.

Realistic Timeline

Year 1-2: Foundation

Sensor validation studies. Algorithm development and bench testing. Quality management system buildout (ISO 13485). Risk management file initiation (ISO 14971). Pre-submission meeting request to FDA (Q-Sub). Combination product Pre-RFD filing. Cybersecurity framework development. SBOM creation.

The pre-submission meeting is the single most important milestone. The FDA will tell you what they expect. Listen.

Year 3-4: Engineering and Pre-Clinical

Functional prototype development. Biocompatibility testing (ISO 10993). Electrical safety testing (IEC 60601). Software verification and validation (IEC 62304). Injection system testing (ISO 11608). Drug stability studies in the cartridge system. Animal studies if required. IND filing (for the drug component) and IDE filing (for the device component). Breakthrough Device designation request.

Year 5-7: Clinical Trials (Naloxone First)

Pivotal clinical trial for naloxone indication. Expected enrollment: 200-500 patients with documented opioid use disorder. Expected duration: 12-24 months of active monitoring per patient. Interim safety analyses at 6-month intervals. Pharmacokinetic sub-study comparing subcutaneous naloxone delivery via SmartShot to intranasal naloxone (Narcan).

Year 7-9: PMA Review and Clearance (Naloxone)

PMA submission preparation and filing. FDA review (180-day target, often extended). Advisory committee meeting (probable for a first-in-class autonomous drug delivery wearable). Facility inspection. Labeling negotiations. Post-market surveillance plan.

If everything goes well, the first SmartShot indication reaches market nine years from today. If it goes the way most Class III devices go, add two to three years.

Each Subsequent Drug: Additional 2-3 Years

Epinephrine: Years 8-11. Can begin clinical trials before naloxone clearance.
Nitroglycerin: Years 10-13. Requires the most stringent detection accuracy.
Glucagon: Years 12-15. Lowest priority. May be overtaken by CGM-integrated insulin systems adding glucagon delivery.

State-Level Considerations

Automated Medication Delivery Laws

States regulate who can administer medication and under what authority. Autonomous drug delivery by a device — without a licensed provider physically present — does not fit cleanly into any state's current pharmacy practice act or medical practice act. Legislative or regulatory changes may be required in some states before SmartShot can be prescribed.

Prescribing and Standing Orders

Naloxone is available under standing orders or pharmacist authority in all 50 states, which reduces prescribing barriers. Epinephrine prescribing for anaphylaxis is well-established. Nitroglycerin and glucagon require physician prescriptions. SmartShot's model — physician prescribes the device and the pre-loaded drug cartridge — will need to fit within each state's prescribing framework.

Telehealth and the Control Hub

SmartShot's Control Hub model, where remote medical professionals monitor device alerts and can override or confirm autonomous actions, intersects with state telehealth laws. Telehealth prescribing across state lines requires licensure in the patient's state, not only the provider's state. The Interstate Medical Licensure Compact covers 42 states as of 2026 but does not eliminate all barriers. Each state's telehealth regulations must be reviewed before the Control Hub can operate nationally.

Emergency Use Provisions

Most states have Good Samaritan protections for naloxone administration. Whether those protections extend to an autonomous device — rather than a human bystander — is untested law. SmartShot's legal team will need to work with state legislatures to clarify liability protections before commercial launch.

What This Means

SmartShot is a decade-long project. The technology is the easy part. The regulation is the hard part. Every closed-loop system that has reached market — insulin pumps, wearable defibrillators, now wearable infusion devices — took longer than its founders expected.

The path is real. The precedents exist. The FDA has approved autonomous wearable intervention devices. It has approved wearable drug-device combination products. It has approved AI/ML-driven medical software. SmartShot asks the FDA to approve all three in one system.

That has never been done. It will take time. The question is not whether it is possible. The question is whether someone will do the work.