Who Dies in Confined Spaces: 1,030 Deaths, and Why So Many Are Rescuers

A confined space kills twice. First it takes the worker who entered — overcome by a gas that has no color and often no smell, or by an atmosphere with no oxygen left to breathe. Then, more than half the time, it takes whoever ran in to save them. The U.S. Bureau of Labor Statistics counted 1,030 confined-space worker deaths from 2011 to 2018. NIOSH's landmark finding is that more than 60% of confined-space fatalities are would-be rescuers. This is the data behind OSHA's most procedure-heavy standard — the causes, the gases, the rescuer multiplier — and the Subpart AA rules written to break the chain.

At a glance: BLS recorded 1,030 confined-space deaths (2011–2018), 88–166 per year. Leading events: trench/material collapse 168, falls 156, single-episode inhalation 126 (H₂S 38 · CO 23 · methane 10), oxygen depletion 39, drowning 21. NIOSH: 60%+ of deaths are rescuers. OSHA's answer is Subpart AA — test before entry (1926.1204(e)), an attendant who never enters (1926.1209), and non-entry retrieval (1926.1211). The U.S. Bureau of Labor Statistics counted one thousand thirty confined-space worker deaths from 2011 to 2018, and NIOSH finds that more than sixty percent of confined-space fatalities are would-be rescuers who entered to help without testing the atmosphere.

The record: 1,030 deaths, 2011–2018

Confined-space deaths are not a rounding error and they are not falling on their own. Across 2011–2018 the BLS Census of Fatal Occupational Injuries logged 1,030 — a low of 88 in 2012, a high of 166 in 2017, roughly 129 a year. They span construction, agriculture, oil and gas, water and waste treatment, and manufacturing; the spaces are storage bins, manholes, tanks, sewers, trenches, and grain silos.

What matters for prevention is how they die. These are the leading events inside that total (BLS, 2011–2018 — selected categories, not the full 1,030):

Two honest notes. Trench and material collapse (168) is the largest single event — but for OSHA compliance, trenches are regulated under Subpart P (excavation, 1926.651/1926.652), not the confined-space standard. And falls (156) are often a secondary mechanism inside a shaft or vault. The category the confined-space standard is built around — the one where the space itself is the weapon — is the atmospheric group: inhalation and oxygen depletion.

The gas that has no color: 126 inhalation deaths

Of the 1,030, 126 workers died inhaling a harmful substance in a single episode inside a confined space. The gases are the same short list that recurs in every incident report — and the same four an entry team is supposed to test for:

Hydrogen sulfide — the sewer-and-oilfield gas that smells of rotten eggs at low concentration and then deadens your sense of smell before it kills — leads at 38. Carbon monoxide, produced by every gas engine and generator run near an opening, follows at 23. These are not exotic. They are exactly the readings a calibrated four-gas meter (oxygen, combustible gas/LEL, CO, H₂S) is built to catch — which is why OSHA makes testing a pre-entry gate, not an afterthought.

The rescuer multiplier: why one death becomes two

Here is the finding that defines confined-space safety. NIOSH's 1986 Alert — Request for Assistance in Preventing Occupational Fatalities in Confined Spaces (Publication 86-110) — concluded that more than 60% of confined-space fatalities are would-be rescuers. Four decades later it is still the most-cited number in the field, and OSHA relied on it in the preamble to the permit-required confined-space rule.

The mechanism is always the same. The atmosphere is invisible. The first entrant collapses without warning — hydrogen sulfide at high concentration can drop a person in a single breath. A coworker sees them down and does the most human thing possible: goes in after them, with no meter, no air supply, no retrieval line. The same atmosphere that took the first worker takes the second. In multiple-fatality confined-space incidents, the dead are disproportionately the rescuers.

This is the reason the confined-space standard reads the way it does — heavy on procedure, adamant that no one enters to rescue without training and equipment. Every rule below exists because untrained rescue kept killing people.

From data to law: how Subpart AA answers each number

The BLS causes and the NIOSH rescuer finding map almost one-to-one onto the construction confined-space standard, 29 CFR 1926 Subpart AA. The CSB-style lesson here is regulatory: OSHA turned the fatality pattern into a permit.

The rest of Subpart AA closes the remaining gaps. A competent person identifies every confined space and permit space before work begins (1926.1203(a)); an employer whose workers will enter needs a written permit space program (1926.1203(d), 1926.1204); the entry permit names the hazards, the tests, and the rescue service (1926.1206); entrants, attendants, and entry supervisors each have defined duties (1926.1208, 1926.1209, 1926.1210); and training gives every employee not authorized for rescue an understanding of the dangers of attempting it (1926.1207(a)).

Read the standard this way and its procedure stops feeling like paperwork. The permit exists because 165 people died breathing a bad atmosphere no one measured. The non-entry retrieval line exists because the person who would have pulled them out went in and died instead. The attendant who "just watches" is the control that breaks the rescuer chain — the single most important job on a confined-space entry.

What the record means for your crew

If your work touches manholes, vaults, tanks, pits, sewers, silos, or any space a body can enter and air can go bad, the 1,030 are your risk profile. The controls that break the chain are exactly the ones OSHA enforces:

• Test before entry, every time — O₂, combustible gas (LEL), CO, and H₂S — and keep monitoring. The atmospheric deaths are the ones a meter prevents.
• Never send an untrained rescuer in. Rig non-entry retrieval before entry; know who your rescue service is and that they can actually respond.
• Post an attendant whose only job is to stay outside, watch, and call for help — not to enter.
• Write the permit and make it the gate: no test, no entry.

The standard, tools & related reading

• The hub: Safety Permits — confined space & hot work
• The rules by role: Confined space training — entrant, attendant, supervisor
• The permit: Confined space entry permit — the required-space guide
• Construction vs general industry: Subpart AA vs 1910.146
• The hot-work companion piece: Why hot work keeps killing — 15 CSB investigations
• Tools: Confined Space Permit Generator · Hot Work Permit Generator
• Glossary: Confined space · Competent person · 2026 OSHA penalties

Sources & verification (verified 2026-07-21): the 1,030 confined-space fatalities (2011–2018), the 88 (2012) / 166 (2017) range, and the event and gas breakdowns (trench/collapse 168, falls 156, inhalation 126 [H₂S 38, CO 23, methane 10, sewer gas 6, solvents 5], oxygen depletion 39, drowning 21) are from the U.S. Bureau of Labor Statistics Fatal occupational injuries involving confined spaces, 2011–19 fact sheet (Census of Fatal Occupational Injuries). The "more than 60% of confined-space fatalities are would-be rescuers" finding is from NIOSH Publication No. 86-110 (January 1986). OSHA requirements per 29 CFR 1926 Subpart AA — 1926.1204(e) (atmospheric testing), 1926.1209 (attendant duties), and 1926.1211(a)/(b)(4)/(c) (rescue and non-entry retrieval) — verified against osha.gov; trenches are regulated under Subpart P (1926.651/652). General guidance, not legal advice.*