Charted: Sterile Injectables Are 71% of Active US Drug Shortages

Hospital pharmacies do not run out of “drugs” in the abstract. They run out of vials, ampules, and premixed bags—the sterile injectable products that keep operating rooms, ICUs, oncology chairs, and emergency departments moving. When those lines thin, the clinical problem is immediate and the economic problem is structural: few plants, thin margins, and long restart clocks.

The United States Pharmacopeia’s 2026 Annual Drug Shortages Report, mapping FDA Center for Drug Evaluation and Research shortage listings through the USP Medicine Supply Map as of 31 December 2025, counts 75 active shortages. Of those, 71% are sterile injectables—about 53 products—versus 16% oral solids and 13% other forms. That injectable share is the headline of this post. The second question is how little manufacturing depth sits behind the tip: ASHP and the University of Utah Drug Information Service put the broader molecule-level shortage tally at 227 in Q2 2026, and report that 48% of new 2026 shortages are sole-source. One disclosed plant vignette still matters for scale: Pfizer’s Rocky Mount, North Carolina site has been described by the company and the FDA as supplying roughly 8% of sterile injectables used in US hospitals.

The interactive dashboard above toggles Dosage mix, Shortage path, Site depth, Makers, and Geography. Switch path metrics between ASHP counts and injectable share; flip site units between share and an index on the ASHP tally; change the geography lens among volume share, shortage-volume share, and the gap between them.

The injectable share is not a rounding error

Dosage form is the first cut that matters. Oral solids can often be multi-sourced across dozens of finished-dose plants. Sterile injectables cannot. Aseptic fill-finish requires specialized cleanrooms, validated lines, and quality systems that fail loudly when they fail. USP’s end-2025 map makes the imbalance numerical: injectables are more than four times the oral-solid share of active shortages.

That imbalance is durable, not a one-quarter spike. Desk anchors that carry USP’s injectable share back through the early 2020s sit in the mid-to-high sixties even as absolute shortage counts rose and then partially fell. The 2025 print—71%—is the highest clear dosage-form tip in the recent USP series. When shortages fall in count while injectables stay dominant in share, the residual list is more hospital-critical, not less.

Read the table as a form-factor claim, not a claim that every injectable molecule is unavailable. Presentation-level FDA registries can show hundreds of injectable NDC entries in short supply even when molecule-level USP or ASHP counts sit lower. The form tip survives the unit-of-analysis change: injectables dominate whichever lens you pick.

Three counters, one fragility story

Shortage statistics confuse desks that treat every published number as the same object. Three counters run in parallel:

• USP / FDA molecule-mapped actives75 at end-2025, down 23% from 98 in 2024 on USP’s series. Nearly all were carryovers; only four new shortages first appeared in 2025.
• ASHP / Utah molecule actives216 at end-2025, 223 in Q1 2026, 227 in Q2 2026, still well below the 323 peak in Q1 2024 but rising for a third straight quarter.
• FDA presentation-level registrythousands of NDC/status rows; independent desk summaries put active injectable presentations in the high hundreds, again ~70% injectable by form.

These are not contradictions. ASHP registers disruption when pharmacies cannot obtain product; FDA lists national medically necessary mismatches; USP’s annual map freezes a year-end FDA-linked set and overlays Medicine Supply Map geography. A hospital can feel ASHP’s 227 while USP’s 75 correctly describes a narrower national confirmation set. For this post, the 71% injectable share is the USP year-end form tip; the 227 is the operational load still sitting on health-system pharmacies.

How many sites stand between patients and empty shelves?

“How many manufacturing sites?” is the question buyers ask after the share chart. The honest answer is layered.

At the manufacturer layer, ASHP reports that 48% of new shortages so far in 2026 are sole-source—one company left standing. Dual- and multi-source products still fail, but sole-source failures convert a quality event or business exit into a national gap overnight. Analytical site-redundancy buckets in the dashboard put roughly half of new shortages on a single manufacturer, with the remainder split between dual- and multi-source—labels that are carried where agencies publish sole-source shares without a full ANDA census.

At the plant layer, concentration is even more visceral. Public FDA and Pfizer disclosures after the July 2023 Rocky Mount tornado put that single facility at about 8% of US hospital sterile-injectable volume, inside a Pfizer franchise that itself was described as roughly one-third of the hospital sterile-injectable market. USP Medicine Supply Map work around that event pointed to well over a hundred products associated with the site, many on essential-medicine lists. Redundancy at the molecule level does not erase the fact that a weather event at one address can reallocate an entire hospital class of SKUs.

At the upstream layer, USP’s KSM module shows that 44% of drugs in shortage have at least one key starting material manufactured in a single country—typically China or India. On the 2025 Vulnerable Medicines List, 48% carry the same single-country KSM flag, and 63% of the hundred vulnerable medicines are injectables. Finished-dose diversity can be an illusion when every ANDA holder pulls the same starting material from one geography.

Manufacturers at the tip of the registry

Presentation-level shortage registries concentrate on a short list of sterile-injectable labels. Desk roll-ups of recent FDA registry summaries put Hospira (Pfizer) near 166 active entries, Fresenius Kabi near 160, and Hikma USA near 97—about 423 entries across the top three alone. Those figures are not a complete market census, and they are not interchangeable with USP’s 75-molecule map. They do show where hospital buyers already know the pain: the same firms that fill OR and ICU carts are the firms whose quality holds or line outages ripple nationally.

Concentration is not evidence of malice. It is evidence of economics. Sterile injectable generics are hard to make, expensive to keep compliant, and priced as if they were commodities. When a line stops, rivals cannot spin up aseptic capacity in a quarter. The manufacturer ladder in the dashboard is therefore a fragility roster, not a league table of villains.

Geography: domestic plants still own the shortage volume

USP’s geography print for injectables is easy to misread as an offshoring morality play. The United States produced about 47% of injectable finished-dose volume in the mapped set and 57% of the volume tied to shortages. India produced 16% of volume and 26% of shortage volume. The European Union produced 20% of volume and only 5% of shortage volume. Domestic production does not immunize against shortage; it can concentrate shortage when the domestic plants are the ones under quality and margin stress.

The geography panel’s gap lens (shortage-volume share minus volume share) makes that concrete: the US and India sit positive; the EU sits deeply negative. Policymakers who hear “reshoring” as a shortage cure need that split. Bringing more volume onshore without fixing the incentive to maintain redundant, high-quality aseptic lines simply moves the chokepoint—it does not remove it.

Duration, discontinuations, and the price floor

Share and sites describe who fails. Duration describes how long patients wait. USP puts average shortage duration at 5.3 years in 2025, up from 4.3 in 2024 and roughly two years in 2019. 64% of active shortages have lasted more than three years; 39% more than five. Injectables drive that stickiness: once a sterile line is down, requalification, remediation, and second-source tech transfer chew years.

Discontinuations compound the problem. USP counts 170 discontinuations in 2025, a 60% jump from 2024 and the highest print since 2019. Prices on discontinued medicines fell—consistent with a market that exits when the floor is too low to fund compliance. A shortage that resolves because the last manufacturer walks away is not a recovery; it is a deletion.

What the dashboard is for

Use Dosage mix for the form tip and duration bands. Use Shortage path to watch ASHP counts ease from the 2024 peak while injectable share drifts upward. Use Site depth for sole-source exposure and the Rocky Mount-scale capacity facts. Use Makers for the presentation-entry ladder. Use Geography for the volume-versus-shortage scatter and gap bars. Therapy-stress indexes and upstream KSM flags are analytical overlays—useful for prioritization conversations, not FDA severity grades.

Caveats

This is a desk synthesis, not a substitute for FDA, ASHP, or hospital pharmacy allocation systems. USP year-end molecule counts, ASHP rolling molecule counts, and FDA presentation-level rows measure different objects; do not average them into a fake “true” shortage total. Manufacturer entry tallies and dual/multi-source splits mix disclosed anchors with carried estimates and will move as registries refresh. Rocky Mount’s 8% hospital sterile-injectable figure is a 2023 disclosure snapshot, not a live telemetry feed. Geography shares describe mapped finished-dose volume, not every vial in every hospital cupboard. Stress indexes by therapy class are analytical, not regulatory. Nothing here is clinical advice, formulary guidance, or a claim about any single patient’s access.

The economic punchline survives the caveats. Seventy-one percent of the USP-mapped active shortage list is sterile injectable. Nearly half of new shortages in 2026 still lean on a single manufacturer. And a single disclosed plant can still represent on the order of eight percent of what US hospitals inject. Empty shelves are not a mystery of demand. They are a map of thin aseptic capacity priced like a commodity.