Summary of MedSun Reports Describing Adverse Events: Large Volume Infusion Pumps and Issues Resulting In Over-infusion
MedSun: Newsletter #33, February 2009
Over the past year, MedSun has received 20 adverse event reports of large volume infusion pumps and issues related to over-infusion. These reports are associated with five manufacturers: Cardinal Health (11), Baxter Healthcare Corporation (5), Hospira, Inc/Abbott Laboratories (2), Smiths Medical (1), and Sigma International (1). The reports were submitted by 16 user facilities between January 1, 2008 and January 1, 2009. The most frequently reported device problems were:
One report involved a patient death, 1 report of injury, 12 reports indicate a malfunction, and 6 reports do not indicate the event report type. The following patient problems were reported:
•Breathing difficulties (1)
•Low blood glucose (1)
•Electrolyte imbalance (1)
•Hospitalization required (1)
•Identified diagnosis (1)
Of the reports that listed patient age, 6 had a patient age listed as less than 21 years and 10 had a patient age listed as greater than 21 years. Of the reports that listed patient gender, a total of 4 reports involved female patients and a total of 13 reports involved male patients.
These MedSun reports contributed to FDA awareness of the device problems. Reported events involving specific infusion pumps may, or may not, be involved in the recall(s) listed. Please note the lists of recalls below include recalls that could be related to issues with over-infusion from 2007 to present, and this report reviews Medsun reports from the last year.
•Cardinal Health Alaris Infusion Pump Module (formerly Medley Pump Module), Model 8100, Class I recall on October 29, 2007: During the manufacturing or servicing of the mechanism assembly, the occluder springs were misassembled (overlapping [nested], missing, bent or broken). If a spring is misassembled, there is a potential for inaccurate flow rate which may lead to a patient’s harm due to over-infusion.
•Baxter Healthcare Corp. COLLEAGUE and FLO-GARD Volumetric Infusion Pumps, Class I recall on August 15, 2007(updated recall information from June 21, 2007): A field service technician did not perform all of the hardware Recall: upgrades required in accordance with Baxter’s corrective action procedures related to eight open Colleague recalls. The service technician certified in the service documentation that the upgrades were done on the affected pumps even though the upgrades were not done. These incomplete upgrades were performed from May 22, 2007, through August 7, 2007.
•Baxter Healthcare Corp. COLLEAGUE and FLO-GARD Volumetric Infusion Pumps, Class I recall on June 21, 2007: On June 21, 2007, Baxter contacted their customers informing them that the company discovered, during ongoing quality control processes, that repair, test and inspection data sheets, including electrical safety data, were falsified. As a result, it may be possible that pumps sent to be serviced, repaired, or corrected were returned without service being performed on them. This may result in over infusion, under infusion, failure to detect an upstream or downstream occlusion, electrical shock hazard, failure to detect air in line and malfunctions where the pump will stop infusing and result in an interruption of therapy that can result in death. The company requested their customers to locate the affected pumps and take them out of service as soon as possible, and send them back to Baxter for repeat inspections and servicing. Baxter will provide loaner pumps to customers that require them, free of charge.
[Note: The reports have been edited for clarity]
|Device||Device Identifiers||Event Description|
|Cardinal Health Alaris Infusion Pump||Models 8100, 8015||A patient was in the ED for chest pain and was on a heparin drip. The pump settings were checked by the nurse at the beginning of the morning shift which showed settings of heparin infusing at 17ml/hr, 18mcg/kg/hr, which was the appropriate dose. Approximately 2.5 hours later, the patient's 6 hour repeat PTT results were >150, which was up from the initial value of 28. The pump settings indicated no changes at that time. The doctor was notified of the lab results, and there was no order to change the dose or the infusion rate of the heparin. Approximately 1.5 hours after the PTT results were received, the heparin bag appeared to be abnormally empty to the nurse. The pump was checked by the nurse, and there were no changes in the settings. However, when the nurse checked the total volume infused, the pump stated 216.57ml. The total volume infused indicated an infusion rate of approximately 24ml/hr. A second nurse checked the pump display and verified the calculation discrepancy. The pump was removed from service. The patient was scheduled for a procedure later that afternoon. The doctor was notified of the issue and there was no change to the patient's treatment plan. There was no patient harm.|
|Cardinal Health Alaris Infusion Pump||Model 8015||Patient receiving TPN at 80 ml/hr. Infusion stopped at 0430 to draw labs. Nurse restarted infusion and thought that rate was set at 80 ml/hr with approximately 2000 ml still in bag. Approximately 2 hours later, pump alarming and TPN bag empty. Pump taken out of service. When event/error log downloaded revealed that rate had been set at 880 ml/hr. Concern that there may be keypad bounce that is not immediately evident to user. Pt developed a-fib and was transferred to ICU had no long term harm.|
|Cardinal Health Alaris Infusion Pump||Model 8100|| Programmed IVPB at 100 ml/hr for a total volume of 250 ml at 0900. The infusion was completed at 1100; however the total volume on the screen show 45 ml left. Main IVF supposed to be 75 ml/hr but patient was receiving 100ml/hr. BioMed assessment: The pump operation and accuracy were verified in biomed and all tests were within specifications. The pump module will normally continue to pump for the duration of the calculated time as programmed prior to switching back to the main IV rate; however, this was not the problem. The device was returned to the manufacturer:|
multiple tests were conducted and the pump operated according to specifications.
The conclusions: The pump appeared to have operated within specifications. No malfunction is believed to have occurred. The root cause of the reported problem is unknown. Based on this, the educators will share this report and review the loading and infusion methodologies for the pump. No patient harm.
|Cardinal Health Alaris Infusion Pump||Model 8015||55 ml Fentanyl syringe delivering 4 ml/hr (200mcg/hr) syringe was started and approximately 2.5 hours later, the syringe was found to be almost empty. Patient stable post event. Biomedical Assessment: Problem: Reported problem says: 60ml Fentanyl syringe set to deliver at 4ml/hr started and approximately 2.5 hours later the syringe was empty. The drug infused too quickly. It should have taken 15 hrs to deliver 60ml's. Corrective Actions/Recommendation: The event log from the PCA module shows that it was set to deliver 20 ml/hr instead of the 4 ml/hr as indicated by the nurse. There were no errors in the error logs for either the PCU or the PCA module. The pump functioned properly, but was set at the wrong rate.|
|Cardinal Health Alaris Infusion Pump||Model 8100|| Pump programmed to run at 125 ml for two hours for a total infusion of 250 ml. The IV infused completely in approximately 1 hour.|
Biomedical Assessment: Testing in Biomedical Engineering found the LVP module functioning properly. There were no errors in the error log. The operation log shows a number of rate and volume to be infused changes during the course of the day, but there is nothing conclusive as to what actually happened.
|Cardinal Alaris Infusion Pump||Models 8100, 8000|| Infant in NICU developed acute deterioration in respiratory status, requiring increased oxygen and ventilator support. It became apparent that the infant was over-infused with D13 Total Parenteral Nutrition (TPN) through an Alaris Large Volume Pump (LVP) connected to a Point of Care Unit (PCU). The Alaris Infusion Pump involved was accidentally programmed for 68 mL/hr instead of the ordered rate of 6.8 mL/hr. The pump infused for approximately 2 hours, infusing approximately 150mL of fluid during that time. The infant became extremely hyperglycemic and hyperkalemic, and continued to deteriorate despite aggressive medical and ventilatory management. The patient expired the next day. It is believed that, although the pump did not malfunction, inherent design flaws in the infusion pump may have contributed to this event. |
1. PCU Model #8000
2. Large Volume Pump (LVP) Model #8100
3. Syringe Module (attached, but not directly involved in incident)
Root Cause Analysis:
1. Preventative Maintenance was up to date at the time of the incident
2. Preventative Maintenance was performed after the incident and no issues were found
3. The historical data logs of the pump were downloaded and reviewed. It was verified that at the pump was programmed for a "basic infusion" at 68 mL/hr and a Volume To Be Infused (VTBI) of 150mL. The pump ran from approximately 2 hours, with a 2 minute pause. The total contents left in the IV bag of D13 TPN was 75mL, measured using a graduated cylinder. The bag was originally filled with 238.2 mL, therefore approximately 163 mL were infused in total.
4. Mechanical testing was performed to determine the force required to press the buttons on the pump. Particular interest was paid to whether or not the force required to press the decimal point was higher than the force to press the other number digits. It was found in preliminary testing that the force to press the decimal point was greater than the force to press the "6" or the "8" buttons. An RN programming the pump could potentially press the "6" then the decimal point, but not have it register on the pump because not enough force was exerted due to human factors design issues.
5. The Alaris pumps do not make a "beeping" noise as the buttons are pressed for this patient population modality. Therefore, there is no audible confirmation to the RN programming the pump that he/she has pressed a button and that it has registered.
6. It is being investigated what level of visual acuity is required to read the pump with varying lighting levels and heights of the pump.
|Cardinal Health Alaris Infusion Pump||Model 8100||A bag of potassium (30 MEQ k in 100ccNS) was hung. The nurse programmed the pump for the medication to go in at 25 cc/hr. When the nurse went to release the roller clamp, it was noted that all of the fluid, except for approximately 30cc had infused. The nurse immediately rolled the clamp back down and called in the resource person to double check the pump, and notify her of what had happened. She verified the pump was set up correctly. The pt was having frequent PAC's and PVC's prior to this incident and continued to have them after. There were no other adverse effects all VSS. All of the IV tubing replaced and a new bag of potassium rehung.|
|Cardinal Health Alaris Infusion Pump||Model 8100||Patient was ordered a medication bolus followed by a drip (gtt) at 18.4cc/hr. A bolus was given and a gtt was started immediately afterward at 1145am. The pump module kept beeping "air in line". The tubing was taken out of module several times in an attempt to get air out; however, there were only very small air bubbles that would not normally cause the pump to alarm. Upon attempting to remove the air bubbles, the clamp (tubing) was opened and the medication ran wide open into patient. This caused the patient to receive approximately 4 hours worth of the medication in 1 hour. Once I realized how low the volume in the bottle was and what had happened I immediately shut off the gtt. Notified pharmacy and the Charge Nurse. Biomed was notified of the issues I was having with the module and picked up the pump for further evaluation. The physician was notified and orders were received to stop the medication gtt and restart in 4 hours. No patient harm.|
|Cardinal Health Alaris Infusion Pump||Model Medley (8100)||Initially Alaris pump screen showed Heparin infusion at 5 ml/hour with appropriate drip chamber flow. Nurse noted drip chamber started to free flow without pump beeping and screen still reading 5 ml/ hour. Patient received approximately 5,000 unit bolus of heparin (50 ml). Infusion stopped and pump removed from service. Patient without apparent injury. Follow-up indicates that the pump had passed the occluder testing just prior to the event, and subsequently failed the test when the pump was tested after the event. The pump has since been sent back to the company for corrective action. The hospital has a total of 1,114 Alaris Large Volume Pump modules which are managed and serviced for the hospital by a third party distributor. The pump distributor has completed occluder testing of 1046 of the LVP modules, as of this date, using the latest protocol provided by Alaris. Pumps which fail the occluder testing are returned the distributor's District Office where the failure is verified, and failed pumps are then returned to Alaris for X-ray inspection. Of the 1046 pumps which have been tested, 80 LVPs have failed the occluder test, or 7.65%.|
|Cardinal Health Alaris Infusion Pump||Model unknown||The Alaris pump was set to deliver 65 ml/hr of packed red blood cells for a total volume of 260 ml over 4 hours. Approximately 2 hours later, the pump was found to be alarming with notification for "air in line". Total volume had infused in 1 hour and 50 minutes instead of 4 hours.|
|Cardinal Health Alaris Infusion Pump||Model Gemini PC-2TX||Infusion pump was set to infuse at a rate of 2 ml/hr to deliver 48 ml in 24 hours. Infusion started at 12:30 pm and at 23:15, there was a fluid side occlusion, 100 ml bag of Intralipids was empty. Performance tested unit for rate and volume accuracy. Pump is operating according to the manufacture's specifications and was returned to service.|
|Baxter Healthcare Corporation Infusion Pump||unknown||A patient's cardiac rhythm changed from a sinus rhythm in the 80's down to the 40's, and the blood pressure increased to 200's/100's. The patient was observed and all drips were checked; pump C on the IV pump had failed. The screen said pump failure and it had apparently gave the patient a bolus of Levophed. The tubing was pulled out and another pump was used. The patient's heart rate and blood pressure returned to normal after a few minutes.|
|Baxter Healthcare Corporation Infusion Pump||Colleague 3CX||A 250ml lipid infusion was to run over 12 hours. Seven hours after the infusion was started, the infusion pump detected air in the tubing. The bag was empty. The pump settings were correct - 21ml/hr for 250ml total. The pump still showed 119ml were left to be infused. A triple pump was used and the other channels were working correctly. The lipids were infusing through channel B.|
|Baxter Healthcare Corporation Infusion Pump||Syringe pump, AS50||Syringe pump programmed to run lipids at 0.2ml/hr. Syringe changed out and new lipids to run at the same rate over 24 hrs. Approximately 85 minutes later the pump alarm went off. At this time the syringe that had 4.2ml of lipids was now empty.|
|Baxter Healthcare Corporation Infusion Pump||Flogaurd 6301||TPN hung on pump at 2121 to infuse at 19 CC/HR. Bag empty (approx 500CC) at 0330. Biomed performed functional test per service manual. Found no problem.|
|Baxter Healthcare Corporation Infusion Pump||Colleague 2M8163||Patient was receiving TPN infusion at a rate of 45 ml/hr. A new infusion bag was hung and the IV pump's volume setting was changed accordingly. The rate was allegedly unchanged, but 45 minutes later, the patient became confused and tachypneic. The pump was found to be operating at 700 ml/hr. The pump was discontinued and sent for inspection. Insulin therapy was started and lab tests were sent on the patient to compensate for the over-infusion. The patient recovered.|
|Hospira, Inc/Abbott Laboratories Infusion Pump||Plum pump, Lifecare 5000||New bottle of integrellin started at 12cc/hr (100cc bottle) approximately three hours prior to the bottle being empty. Integrellin bottle was empty with pump still running. Pump removed from service immediately.|
|Hospira, Inc/Abbott Laboratories Infusion Pump||Plum A+||Plum "A+" pump. An Insulin drip had been hung at a rate of 2.5 cc/Hr. Approximately two hours later the bag was empty. The patient did not present as if he had received that much insulin. The pump showed that 90cc had infused. Concerned that the pump is malfunctioning.|
|Smiths Medical Infusion Pump||Medfusion 3010A||Male infant on ECMO . Infant received approximately 40ml of heparin via syringe pump, instead of 0.4ml. The medication eas heparin 100u/ml. Patient had elevated ACT levels, but no apparent bleeding and no change in cranial ultrasond. Heparin discontinued temporarily until ACT levels returned to normal.|
|Sigma International Infusion Pump||Spectrum|| Patient was to receive Septra at 500ml/hr for a 1036ml bag in which there was 36ml of Septra and 1000ml of D5. The primary IV (.9NS) fluids were set to run at 125ml/hour. When the secondary IV infusion rate of 500ml/hour was supposed to have switched back to the primary rate of 125ml/hour; it didn't, and the secondary rate of 500ml continued for an amount of time unknown to the nurse. The infusion of the Septra that was mixed in the D5 took 2hrs and 45minutes to run, when it was programmed to run over 2 hours. This was due to the fast rate and subsequent running in of some of the primary fluids at the same time as the secondary fluids (Septra/D5). Between the secondary rate of 500ml/hr of running the primary fluids (.9NS) for a time, and the primary fluid (.9NS) dripping in with the Septra/D5, the patient received approximately 240-345ml of NS at the rate of 500ml/hr. Staff unsure if flow rate was set incorrectly or if something happened with the pump. This could be prevented by staff checking the patient when the secondary fluids should be finished, and also confirming the pump set up is correct. The nurse should also be sure that the primary fluid bag is hung low enough to not flow in at the same time as the secondary infusion.|
Clinical Engineering evaluation: A review of the event log within the pump correlated to the event information. The staff, per order attempted to infuse a secondary infusion at 500ml/hr. Operational testing of the pump in the stated configuration showed that without the primary line clamped, siphoning would occur. Current pump software does not prompt the user to monitor and clamp, if necessary, the primary line when infusing a secondary infusion above 300ml/hr.
Staff to be advised to monitor any secondary infusions for siphoning of the primary bag as evidenced by primary line drip chamber activity specifically at rates nearing or above 300ml/hr.