Suture needles of different sizes are often used during a single case and is some cases hundreds of needles are used. Surgical needles are the most frequently miscounted surgical item and their management contributes to the complexity of the scrub person's duties. ORs should have systems in place to account for needles. The most common practice is manual counting of individual needles when the case begins and at the end of the case. Discrepancies are usually resolved by recounting or if they can't be resolved, xrays are taken to look for the needle in the patient. There are case reports of retained needles causing pain and requiring operation for removal but these reports have involved needles that were >17mm or in very special spaces (e.g. the eye). There is good agreement that micro needles and needles <10mm are not detectable or detected with low frequency on plain intraoperative xrays. Even if these needles could be seen it is unlikely that they could be found to remove them and there is no evidence that they cause harm in large body cavities. There are no reports of needles <17mm causing harm and the ability to see these needles on plain films varies. So the cutoff for what is to be considered a large needle is probably between 13-17mm. Needle sizes should be referred to in mm and not as a suture size since many different sized needles can be swedged onto the same size suture. It is up to each facility to decide on a policy. We outline below a 3 step practice for prevention of retention and safe needle management.
Second, use needle counter boxes and keep the number of needles "in play" at one time <40. We have found that one of the most frequently ordered needle counter boxes has slots for 80 needles and scrub persons often put two needles in one slot. That means counting 160 needles in one small box and overrides the safety design of the counter. The counters are designed to make it easy to see that each slot has only one needle and provide a visual aid to account for all the needles used. 
INCORRECT NEEDLE COUNT, PATIENT DISCLOSURE AND MRI RISK
If at the end of an
operation, a small (<15mm) needle is missing in a large body cavity (e.g.
chest) and a thorough search has been made but the needle has not been found,
the needle count in the operative record should be recorded as incorrect. It is
good practice to document in the operative report any and all actions taken in
the setting of the incorrect count. Another action is to disclose to the patient
the fact that a small needle was missing. Ethically this is sound because
missing a needle is not an anticipated outcome of the operation and actions
were taken in the OR to look for the needle but it is not known with certainty
where the missing needle is. If intraoperative Xrays identify the needle, it is
a clinical decision whether or not to remove it. It may be that removal may not
be possible or cause more harm than leaving the needle alone. If the needle is
“found” but not removed the final count is recorded as correct and it is good
practice to disclose to the patient that the needle is inside of them. It is
useful to show the patient what the missing needle looks like and discuss why
it is unlikely to cause harm. Surgical metal clips and staples of larger sizes
are used routinely and remain in patients. If there remains any question or
uncertainty about the needle’s whereabouts, a CT scan may be obtained. CT scans
can identify needles of any size. Disclosure around the time of the operation
is a good strategy because many patients have a CT scan sometime in their
future. It is better to hear about a retained needle from the surgeon rather
than months or years later if the patient has a CT scan for some other reason
and is told there is a needle inside of them of which they had no prior
knowledge.
With regards to leaving a small needle in a large space and the patient undergoing a MRI there should be no danger with these small needles. Concerns with metallic objects in MRI are related to the heat generation in the magnetic field and is a danger related to the length of the object. Another concern is the question of wobble or movement of the object but after objects have been in spaces for a time they develop a fibrous reaction which prevents them from moving (and in the case of guidewires in the heart this fibrous reaction can also hinder complete removal of the wire because it becomes adherent to the heart tissue). A small needle in a small or sensitive space, such as the eye might however be a cause of concern in MRI.