New experimental work assessing tenability (how toxic gases and heat prevent escape and kill) using a full-scale room has found that any occupants would be at high risk of death if a smoke alarm wasn’t fitted.
Carried out by French authors, this is the first such published work for a decade. They measured temperature (from 0.8m to ceiling), heat flux, opacity and a range of toxic fire gases (from 0.6m to ceiling) in five test fires set up in a furnished domestic bedroom and adjacent corridor. 
The room, which measured 3m x 3m x 2.5m high, had non-flammable walls and ceiling and was fitted with PVC flooring and electrical ducting, contained a pine bed, mattress and bedding, a wardrobe, desk and table (all made of wood/fibreboard), a rug, clothes, towels, books, CD cases, papers and food packets. A smoke alarm was fitted on the ceiling. 
In the first three tests, different fire sources were placed on the bedding and the door remained closed, simulating a fire starting while a person is asleep in bed (under-ventilated). 
In the first two, a smouldering cigarette and a match were placed on the bedding (cotton-polyester fabric, polyester lining) and in both cases, it did not catch fire and did not significantly smoulder (domestic bedding in France is subject to legal fire safety regulations requiring resistance to ignition by a burning cigarette). 
In the third test, a small fire (equivalent to a burning ball of paper) was placed on the bedding. No flames were observed for 10 minutes, but carbon monoxide levels started to rise significantly after two, at which time the smoke alarm activated. The smoke reached such a density that escape would be compromised after about three minutes and concentrations of oxygen and carbon monoxide/hydrogen cyanide were at levels which would compromise tenability (occupant survival) after about four and six minutes respectively.
In the final two tests, fires of different sizes were started in a metal wastepaper bin under the desk with the door open, simulating a person escaping rapidly. In one, the bin contained 500g of crumpled paper and the fire burned out in the bin. In the other, the fire was lit in one bin situated adjacent to a second bin which also contained 500g of crumpled paper. 
In this, the fire spread to the second bin and the smoke alarm activated at 3 minutes 15 seconds. Gas toxicity did not reach untenable levels until around six minutes and smoke density did not compromise escape until the five minute mark, although temperatures threatened survival from about four minutes.
The authors concluded that the smoke alarm activated in time to make escape possible for a healthy individual. However, in both cases an individual with inhibited movement (invalid, elderly, or a child) would be at significant risk because there was only one or two minutes between the smoke alarm activating and both smoke density levels impeding escape and conditions becoming untenable for occupant survival (because of toxic gases in the under-ventilated room and temperature in the ventilated room). 
The authors also noted that while the results are highly specific to fire conditions and materials, and that risk is highly dependent on a person’s reaction and mobility, in all cases occupants would be at high risk of death if smoke alarms were not operational.
Commercial buildings, non-domestic and multi-occupancy premises in England and Wales are already forced to undertake a 'suitable and sufficient' fire risk assessment carried out under the Regulatory Reform (Fire Safety) Order 2005.
While the overwhelming majority of premises do this, if the assessment is thought to have been carried out to an insufficient extent, the Responsible Person can face an unlimited fine or up to two years in prison.