The 'twice the pipe diameter' dead-leg rule isn't a rule
Written by The BlueWave team · Published 9 September 2026 · 6 min read
General information, not legal or regulatory advice — your duties need your own competent advice.
BlueWave is Legionella compliance software for UK water hygiene contractors. BlueWave is not affiliated with or endorsed by the Legionella Control Association.
Somebody on site will tell you a dead leg is fine as long as it is shorter than twice the pipe diameter. It is one of the most repeated numbers in water hygiene, and it is not a rule. HSE's guidance does not specify a minimum length, or a minimum ratio of length to diameter, for what counts as a dead leg. The "twice the diameter" figure is a design rule of thumb that has hardened, through repetition, into a regulatory-sounding threshold it never was.
Start from what the guidance does say, because the definitions are precise where the numbers are not. HSG274 Part 2 defines a dead leg as "a length of water system pipework leading to a fitting through which water only passes infrequently when there is draw off from the fitting, providing the potential for stagnation". A blind end, sometimes called a dead end, is "a length of pipework that is closed at one end through which no water passes". Notice what those definitions turn on: whether water moves, and how often. Not how long the pipe is. A short dead leg on a never-used fitting is a worse risk than a longer one that gets drawn through daily.
That is the point the ratio misses. A dead leg is dangerous because water sits in it, warms to the temperature of the room around it, and settles in the band where legionella multiplies, with no flow to carry the growth away. Then someone finally uses the fitting and the leg empties its contents into the live system. Length matters only as a rough proxy for how much stagnant water you are holding. Whether the water moves is the thing that actually decides the risk.
Why the numbers don't agree with each other
If "twice the diameter" were a real threshold, the sources would agree on it. They don't. WRAS is where the two-diameter rule of thumb comes from, and some versions of it cite a volume instead, something under 500 ml. CIPHE is commonly quoted at six pipe diameters, three times the WRAS figure. Healthcare guidance HTM 04-01 takes a different approach again and says branch pipework should generally not exceed three metres. Three sources, three different answers, none of them a law.
That spread is the tell. A regulatory limit does not come in a two-diameter version, a six-diameter version and a three-metre version. These are design targets for people laying new pipework, useful when you are deciding where to put a tee, close to meaningless as a pass mark for pipework that already exists. Chasing a ratio on an installed system is measuring the wrong thing. The question is never "is this leg shorter than twice the diameter", it is "does water move through this pipe often enough to stop it stagnating, and if not, why is the pipe still here".
How you actually find them
Dead legs are found by physically tracing the pipework during the risk assessment, not by reading a drawing that stopped being accurate two refits ago. The common sources are all the same story, which is change that removed a use but left the pipe:
- Pipework left in place after equipment came out. An old sink, a dishwasher, a washing machine gone, the supply to them still there and now going nowhere.
- Supplies capped off for a future use that never arrived.
- Outlets isolated but not removed, valved shut and forgotten.
- Tee branches feeding a fitting used so rarely the branch barely sees water.
The physical trick that finds them is temperature. Run your hand along the pipe. A branch that is cold where the hot line around it is warm, or warm where the cold line is cold, is a length water has stopped moving through. Stagnant pipe drifts to room temperature, and room temperature on either service is a flag.
What to do when you find one
There is a hierarchy here, and the order is the whole point. Do the thing near the top if you possibly can, and treat the bottom of the list as the compromise it is.
- Remove it. Cut the redundant pipework back to the circulating pipe, including the tee itself, so there is no stub left to stagnate. This is the actual fix. Everything below it is management of a problem you could have deleted.
- Flush it, if the outlet has to stay. A retained low-use outlet can be brought onto a weekly flushing regime, recorded each time, which keeps water moving through the leg. This is a control, not a cure, and it works for exactly as long as the flushing actually happens.
- Valve it off, or fit a non-return, where an unused supply genuinely has to remain in place.
- Design to the two-diameter rule, as a last resort on new work, never as the default you reach for first.
Most sites over-use option 2 and under-use option 1, because flushing feels like progress and cutting pipe out feels like a job. But a flushing regime is a promise to do something every week forever, and a removed dead leg is a risk that is gone. Given the choice, delete the pipe.
A found dead leg is a job, not a note
The failure mode with dead legs is not finding them. It is finding them, writing them into the risk assessment, and then losing track of whether anything was ever done. A dead leg noted in a survey and left open for three years is exactly the kind of open remedial action that turns up in the evidence after an incident.
BlueWave treats a found dead leg as a remedial action on the record that stays open until it is closed honestly: either the pipe was removed, with the date and who did it, or a flushing regime is running against that outlet and each flush is logged. What matters is that it cannot fall off the list unnoticed. Cutting the pipe out or running the flush is still the engineer's work. The system's only job is to stop the action being forgotten before one of them happens.
If you keep the leg and manage it by flushing, the evidence bar is real, and it is per outlet and per week, the same standard as any little-used outlet. That flushing task then joins the rest of the monitoring calendar, and the dead leg becomes a live commitment your risk assessment now depends on, which is one more reason a change-triggered review beats a date on a certificate.
Forget the ratio. Walk the pipe, find the lengths water has stopped moving through, and get them cut back to the main. A dead leg you removed needs no flushing log, no annual argument about millimetres, and no line in next year's remedial actions.