|
On 18-feb-2022, pentax medical was made aware of a report that an fg-1017- cryoballoon controller had a malfunction where on initial insertion of nitrous oxide cartridge, gas began to slowly leak through controller body from front and rear exhaust ports preventing the balloon from being able to ablate.The case was cancelled.Initially this was determined to be a non-reportable event as there was no report of injury to the patient or user.If this malfunction were to recur, the controller is designed to contain nitrous oxide and vent slowly, therefore posed no risk to patient or user.There was also a review of the current complaint data that showed that there had been no adverse events associated with nitrous oxide venting from the cartridge on this device or any similarly marketed device in the past 36 months.A voluntary report form mw5107665 was received from the fda via mail on 18-mar-2022.The event description from the voluntary report matched the events documented within this complaint, an out of box failure due to a nitrous oxide leak within the controller body and as a result the balloon did not fill as designed.Although the "controller leak issue investigation" identified an overall low patient risk and no action is needed for the units in the field, this event was identified for reporting on 18-mar-2022 in an over-abundance of caution.The returned controller is being evaluated as part of the investigation.This event meets the requirements for fda reportability; however submission of this report does not constitute an admission that medical personnel, user facility, importer, manufacturer or product caused or contributed to the event.
|
|
(b)(4).The controller was returned under rma number (b)(4) and the investigation was completed on 01-mar-2022.The investigation included assessing for leaking.If leaking was found, the investigation would attempt to locate the source of the leak and perform a root cause investigation.A review of the device's lot history record showed it passed all testing performed during the production process.Most notably, it passed the "simulation testing" portion of the production process, which utilizes the device with nitrous oxide cartridges to perform simulation ablations in a water bath.Other tests were also performed to confirm the absence of leaks in the entirety of the nitrous pathway.All test results were passing and unremarkable.A review of the device log files showed the device had no nitrous oxide leak issues during "simulation testing" during production.Logs from the day of the complaint confirmed the device leaked to a similar degree, with the device losing most of the cartridge's contents in 15-30 seconds.As would be expected in a such a situation, the device produced "low pressure" errors.With regards to the production of "low pressure errors", the device was working as designed and performing as intended.Visual inspection of the returned controller with cap was performed, and no obvious signs of damage were observed.The piercing point was visually inspected and found unremarkable.The cap was also found unremarkable.Two nitrous oxide cartridges were screwed into the controller, using the controller's paired cap.A leak of nitrous oxide was observed and confirmed.The cartridge lots were examined, and it was noted that the device experienced leaking with at least 3 different lots of cartridges.Thus, it is unlikely that the leak is related to a specific set of cartridges.Based on "simulation testing", the leaking complaint controller was only able to perform 8mm of a swipe before repeatedly producing a "low balloon pressure" error due to having insufficient nitrous pressure due to the leak.For the leak to be missed (by the user), the controller would have to perform at 30mm swipe at 0.7mm/s speed without issue.In this, the device was working as designed and performing as intended.Thus, it is certain that if the controller were leaking to the extent found during the investigation, it would have failed the "simulation testing" test performed during production.Multiple tests were performed and during the "low-flow bubble testing", bubbles were observed to be leaking from the tower attachment screw holes adjacent to the high pressure sensor.There should be no outlet for air to be able to escape.A further investigation found that the bottom of the screw holes perforated the manifold cavity in which the high pressure sensor is housed.This is an unexpected result, as the screw holes function only as attachment points for the tower component and were not intended to be connected to the gas delivery pathways.This leak was confirmed absent during production testing but present during initial use of the controller in the field.Further disassembly and testing of the complaint controller components revealed a perforation between the bottom of a screw hole, used for holding the tower component to the manifold, and the cavity for the high pressure sensor.This perforation could present a leak with equal magnitude and similar symptoms as the complaint leak.An additional investigation was performed for the "controller leak issue" for both gen ii version 2.1 (serial number (b)(4)) & gen ii version 2.2 (serial number (b)(4)) controllers.The additional investigation was completed on 22-mar-2022.The controllers were put through a series of examinations and diagnostic testing including documentation review (cartridge lot, controller log file, lhr), component examination (cap, piercing point, and interior components), various testing (simulation testing, tilt measurements, piercing point mechanics/seating effects, and bubble testing) finally resulting in disassembly to the component level careful examination and pressure testing identified the leak path as being a connection between the high-pressure sensor bore with two mounting screws for the tower assembly number r1090.A further investigation of the released manifold drawing as compared to the manifold solid model found a drawing error.The dimensions defining the depth of the #4 screw holes securing the tower to the manifold were incorrect.The image on the drawing, like the model, for the screw depth is correct.Altering the depth of the screw holes on the manifold model to the incorrect dimension on the drawing confirmed the breach between two of the #4 tower mounting holes and the high-pressure sensor cavity.Documentation research found that there have been no changes to tower mounting screws (prt 1757) or mounting screw holes in the manifold (pn 2184) since the initial release.A sample of 38 randomly selected manifolds from 3 lots (200915-0001, 210629-0007, and 211228-0001) in inventory confirmed that all selected units had a connection between the pressure sensor cavity and two of the tower mounting holes.The assembly protocol for the tower mounting confirmed the use of loctite 222ms (pn 1343) during the manufacturing process.Loctite 222ms is a thread locker and sealer which has been in use since the release of the manifold.A review of the process steps for the application of loctite 222ms found that the application instructions were ambiguous in describing the amount and method of applying the loctite 222ms.The manufacture's literature for loctite 222ms was reviewed noting the following: "loctite® 222ms¿ is designed for the locking and sealing of threaded fasteners".Data is presented for accelerated age testing of 222ms at a temperature of 120c for 5,000 hours (208.3 days) with no appreciable loss in strength.A review of the controller risk identified a risk item corresponding to the current condition resulting in an overall all low patient risk.The investigation conclusion was that the cause of the leak in the controller serial number (b)(4) was due to incorrect dimension on drawing p/n, and insufficient amount of loctite application.If an insufficient amount of loctite is applied, this results in a leak path which is detected during the first simulated procedure.The reason no other controller in the field has leaked is that a sufficient amount of loctite was applied to seal the gaps between the threads.Based on the analysis of the loctite used, 222ms, once the loctite cures, it acts as a bonding and sealing compound.If a sufficient amount of loctite is applied, there is sufficient sealing.Additionally, the strength of the bond does not decline over specified operating or storage temperatures.Because the screws are tightened with an eighth-to-quarter turn, they are under tension and hold the tower securely to the manifold with significant frictional contact and are not impacted by the installation of the cartridge, or by handling.Therefore, based on this analysis, this failure (leakage through screw holes) is not expected to occur in the filed for those controllers that have not experienced this issue as out-of-the-box failure.The current measured probability of leakage for controllers, excluding out of the box failures, in the field is minimum to none.Remediation included an additional final leak test per deviation (da 107), had been put into place for 100% of the controllers after they have been completed and are ready for packaging along with in-process leak testing.A verification test is in process on 30 manifold assemblies produced with the drawing error (mu0d00002421/eco2431) providing 95/90 statistical assurance of sealing by the 222ms.A drawing revision has been completed correcting the dimensional error and aligning the drawing to the model.An order had been placed for replacement manifolds per the updated released drawing.An update to mpi 1343 to better control the application of the 222ms which is aligned with the verification testing.No action is needed for the products in the field.The loctite 222ms has been demonstrated to provide a reliable seal for out of box leakage found.Patient harm is minimal for failure to treat.There is no difference between version 2.1 and 2.2 manifolds.Had the amount of loctite been insufficient it is highly likely the units would have already demonstrated a leak.We will continue to monitor leak failures including out-of-box failures.
|
