Elastomeric pumps are widely used to deliver chemotherapy, antibiotics and analgesics through IV, epidural, subcutaneous or perineural catheters.1 Continuous perineural infusion of local anesthetic provides superior postoperative analgesia to opioids [lower mean and maximal visual analog scale (VAS) scores until the third postoperative day] and reduces both opioid consumption and side effects (nausea, vomiting, pruritus, sedation).2 Improved postoperative rehabilitation may result.3
Elastomeric pumps have several advantages over electronic pumps including portability, ease of use, and fewer technical problems such as undesirable alarm triggering.4–7 Nevertheless, the absence of an alarm renders abnormal drug delivery very difficult to detect. Manufacturers and in vitro studies have reported flow rates within ± 15% of their set rates during most of the infusion duration.8–11 Furthermore, a single clinical series, involving 25 Infusor™ devices (from 2 to 9 mL/h) weighed at the beginning and at the termination of infusion, has been published. The investigators reported flow rates between their nominal set rates ± 10%.12 However, other authors reported abnormal infusion times, without formal investigation.1,13,14
To detect potential infusion abnormality, elastomeric pumps are routinely weighed by nurses on our unit, several times a day, as part of standard postoperative care. This procedure also allows the calculation of the real local anesthetic flow rate delivered through the catheters. Due to these observations, we suspected that in vivo accuracy of elastomeric pumps was lower than claimed by manufacturers and in vitro studies. As a result, we retrospectively evaluated the in vivo accuracy and consistency of elastomeric pumps infusion rates in regional analgesia after orthopedic surgery.
The study was approved by the Ethics Committee of La Pitie Salpetriere Hospital (Paris, France) which waived the need for written informed consent, since data of this observational study were collected as part of a safety and quality control program of a standard therapy. After hospital discharge, each enrolled patient received a written informational document about the study. None of them expressed concern about their participation. During a 10-mo period beginning August, 2006, we examined computerized medical records (Actipidos Nursepad 4.2.7, Medicares, Pessac, France) of all consecutive adult patients who received an elastomeric pump with ropivacaine after orthopedic surgery. The patient list was confirmed by the pharmacy computerized file of drug prescriptions. In our center, general anesthesia followed by regional analgesia via an elastomeric pump is the standard care for patients scheduled for knee, foot and ankle surgery, elbow arthrolysis and for femoral shaft fracture repair. Contraindications for regional analgesia were epilepsy, local anesthetic allergy, severe peripheral neuropathy, local infection, major anticoagulation, or patient refusal. Neither electronic pumps nor patient-controlled perineural analgesia is used in our center.
All patients gave informed consent for the anesthetic procedure. Patients were premedicated with hydroxizin or alprazolam 1 h before surgery.
In the preoperative holding area, popliteal, femoral or infraclavicular nerve blocks were performed using a nerve stimulator (Multistim sensor, Pajunk, Geisingen, Germany) to elicit the required motor response at 0.5 mA. Twenty milliliters of ropivacaine 0.475% was first injected through the 19-gauge insulated needle. A 20-gauge “non-stimulating” catheter (Plexolong, Pajunk, Geisingen, Germany) was then advanced 4 ± 1 cm (popliteal or infraclavicular block) or 6 ± 1 cm (femoral block) beyond the needle tip. After needle removal, the catheter was connected to a filter and injected with 4 mL of 1% lidocaine with 1/200,000 epinephrine. The catheter was secured, and its filter (occluded by a luer cap) was fastened to the skin on the ipsilateral anterior part of the thigh or of the hemithorax or of the iliac crest (respectively for popliteal, infraclavicular or femoral catheters), using adhesive tape or the specific device provided in the catheter package.
In the operating room, regional analgesia was nearly always supplemented by general anesthesia. General anesthesia was induced with propofol 2 to 3 mg/kg, sufentanil 0.2 to 0.5 mg/kg and atracurium 0.5 mg/kg. After tracheal intubation, patients’ lungs were mechanically ventilated with a mixture of oxygen, nitrous oxide and sevoflurane. Further intraoperative atracurium or sufentanil boluses were administered if necessary. Multimodal postoperative analgesia was started perioperatively, and continued for at least 24 h. It included acetaminophen in all cases, ketamine 2 μg/kg/min in the case of arthroplasty, and ketoprofen in the absence of contraindications; nefopam was added depending on physician preference.
In the recovery room, the trachea was extubated and patients were asked to evaluate their pain on a VAS ranging from 0 (no pain) to 100 (worst imaginable pain). If necessary, IV morphine titration was performed until the VAS score was less than 30 mm. Depending on the type of surgery, further rescue analgesia was provided via oral or subcutaneous morphine, or via an IV mixture of morphine plus droperidol Patient Controlled Analgesia.
Before use, elastomeric pumps and ropivacaine were stored at room temperature in the air-conditioned recovery room (24 ± 1°C). During the study period, 2 kinds of devices set at 5 mL/h were randomly used, depending on pharmacy supply: Infusor LV5 (Baxter, France) or Easypump™ (Braun, Germany) (Fig. 1). Based on intermediary results of the survey, anesthesiologists requested that Easypump devices were no longer to be provided. This was effective in April, at the end of the follow-up. Just before connection to the catheter, elastomeric pumps were filled with 280 mL (Infusor) or 400 mL (Easypump) of ropivacaine 0.2%, weighed, and the tubing luer cap was removed (Infusor) or the clamp opened (Easypump) to exclude air from the tubing. After connection, the glass flow rate restrictor was taped to the patient’s skin (Easypump). The devices were then placed in the patient’s bed, near the flank of the operated limb. The batch number, the weight of each new filled elastomeric pump, and the exact hour of its connection to the catheter were collected on a specific register.
On the surgical ward, each time they visited the patient (until catheter removal), nurses weighed elastomeric pumps at the bedside with an electronic scale (Tefal™, Rumilly, France), without the tubing. They recorded on the electronic medical patient file the weight, the exact hour of weighing, and all elastomeric pump-related incidents (ropivacaine leakage, accidental catheter removal, catheter disconnection, additional boluses, skin inflammation at the catheter entry). Except for the first elastomeric pump weighing (performed with the recovery room electronic scale), all weighings were performed with the same electronic scale. Several consecutive elastomeric pumps could be sequentially connected to a single catheter, depending on the chosen duration of analgesia and on the limited duration of use of these single-use devices. Specialized nurses performed elastomeric pump exchange on the surgical wards.
Accuracy of Electronic Scale Measures
Sensitivity of the electronic scales was 1 gram. For each electronic scale (n = 5), intra and interobserver variabilities of weighing were computed as the percentage of the absolute difference between extreme measures, divided by the mean of 8 measures of 2 saline-filled Easypump devices (1 with 400 mL and 1 with 100 mL to mimic a full device and 1 at its removal). Mean intraobserver variabilities in 400 and 100 mL filled device weighings were, respectively, 1.2% and 3.5%. For each electronic scale, three different operators performed three different weighings. Mean inter-observer variabilities in 400 and 100 mL filled device weighings were, respectively, 0.7% and 2.0%.
Clinical Assessment of Elastomeric Pump Flow Rate
Ropivacaine volume assessment considered device weight variations and the specific density of 0.2% ropivacaine (1.002–1.005 at 25°C), rounded off to 1.15,16 An elastomeric pump was considered as inefficient when its weight remained unmodified for more than 4 h. When this device secondarily deflated correctly, the first weight retained for flow rate analysis was the last one of the no flow period. Global elastomeric pump flow rate was computed as the difference between its first weight on the surgical ward and the last one, divided by the time (in min) between these two measures. To exclude measures of empty devices (at the end of infusion) from analysis, we considered the last measure >90 g for Easypump, and >70 g for Infusor (empty devices weighed, respectively, 75 and 55 g). For each device, percentage of the flow rate error was calculated as the measured flow rate minus the nominal one (5 mL/h), divided by the nominal flow rate and multiplied by 100.
Continuous quantitative values (weights, duration, VAS, flow rate errors) were reported as means and standard deviations. Comparisons between groups were performed using the Student’s t-test for quantitative data and the χS2 or Fisher exact tests for categorical data.
During the study period, 434 consecutive elastomeric pumps were studied. Four were excluded because they were not weighed. The remaining 430 were connected to 280 perineural catheters (189 femoral, 86 popliteal, and 5 infraclavicular) in 164 female and 116 male patients. Patients were 57 ± 19 years and ASA physical status I–III. They were operated for knee arthroplasty (43%), other knee or thigh surgery (25%), foot and ankle surgery (30%) or elbow arthrolysis (2%). Six catheters were accidentally removed 12 to 65 h after insertion. No catheter was kinked at removal. Mean duration of ropivacaine infusion was 120 ± 112 h.
Elastomeric Pump Reliability
After connection to the catheter, 88 elastomeric pumps (20.5%) did not deflate correctly.
In one Infusor, deflation began correctly but stopped during the first postoperative night (between the 9th and the 19th h), and then spontaneously continued during the 53 following hours. In the remaining cases, elastomeric pump weight stayed strictly unchanged. Three cases had obvious technical problems: one Easypump did not deflate even after disconnection from its effective catheter, and two catheters were definitely obstructed. In 21 cases, catheters were removed 11 to 72 h later, without being tested, because the analgesia was considered ineffective by the attending anesthesiologist, and/or because perineural analgesia was no longer needed at the time the dysfunction was discovered. In 20 cases, the device correctly deflated after the catheters were injected (cleared) with 5 mL local anesthetic (delay: 5–39 h). In four cases, the device correctly deflated after being disconnected and reconnected to the catheter, without any injection through the catheter. In the remaining 40 devices, spontaneous deflation initiated 6 to 43 h after connection, without any particular intervention.
During the first postoperative night, patients with elastomeric pumps that did not deflate correctly had higher maximal VAS scores and needed more rescue analgesia than the ones who received an efficient device (Table 1).
These deflation abnormalities were more frequent with Easypump than with Infusor devices (27% vs 6%, P < 0.0001, Table 2). Five differently numbered batches of Easypump and 4 Infusor devices were studied. Deflation abnormalities concerned all batches of the Easypump and two of the Infusor.
To differentiate catheter from pump dysfunctions, we performed two subgroup analyses. First, among the 126 catheters connected to 2 to 4 elastomeric pumps, a deflation problem occurred on the first device in 12 cases, and on the second device in 27 cases. The incidence of deflation problems with the second device was not higher when a problem occurred on the first one (4 of 12, or 30%) than when the first elastomeric pump deflated correctly (25 of 114 or 22%, P = 0.47). Second, an additional bolus of 0.2% ropivacaine was injected through the catheter in 32 cases, just before an Easypump was connected. Easypump deflation problems occurred similarly whether these additional boluses were performed or not (11 of 32 vs 69 of 268, P = 0.3).
Flow Rate Accuracy of Elastomeric Pumps
Flow rates were calculated over a mean period of 54 ± 18 h (Easypump) and 49 ± 19 h (Infusor) (Fig. 1). The last device weights were, respectively, 187 ± 99 and 117 ± 37 g. Flow rates differed from 5 mL/h ± 15% (4.25–5.75 mL/h) in 47% of the Easypump and in 34% of the Infusor (P = 0.01). A reduced flow rate (<85% of the set rate = 4.25 mL/h) was more frequently observed in Infusor than in Easypump elastomeric pumps (28% vs 11%, respectively, P < 0.0001). On the contrary, an increased flow rate (>115% of the set rate = 5.75 mL/h) was more frequently observed in Easypump than in Infusor devices (36% vs 6% respectively, P < 0.0001). No tendency to either overflow or underflow was observed depending on the months or the season. No adverse effect linked to over-infusion was observed.
Our clinical evaluation of elastomeric pumps in postoperative perineural analgesia raises concerns about their in vivo reliability. First, 20.5% of them did not deflate correctly after connection to their catheter, with either a delay in deflation or no deflation at all. These dysfunctions were associated with insufficient analgesia during the first postoperative night. Second, in vivo flow rates of 40% of these devices were out of the precision range defined in vitro. Thus, an “elastomeric pump-related” technical problem must be considered in the case of early, insufficient postoperative perineural analgesia.
Absence (or abnormal delay) of elastomeric pump deflation could be due to catheter or pump dysfunction. Several points argue in favor of the second hypothesis: 1) only 2 catheter obstructions were documented among the 23 tested, and no catheter was kinked at removal; 2) an early catheter obstruction was unlikely since this problem persisted even when an additional bolus of ropivacaine was injected immediately before connecting the pump; 3) deflation defects occurred in elastomeric pumps connected soon after catheter insertion as often as those connected several days later; 4) in 126 catheters connected to 2 to 4 devices, problems occurred similarly whether the first pump deflated correctly or not; and 5) 91% of these problems were observed with the same brand of device (Easypump). The manufacturer was contacted, but did not have any explanation for this problem, provided that local anesthetics and all elastomeric pumps were stored at room temperature, at a standard altitude, and that the flow restrictor site was correctly taped to the patient’s skin.7,8,13 These factors can be excluded in our study.
Another possible explanation could be an abnormal prolonged adherence of the two inner plastic surfaces of elastomeric pump tubing at the clamp site (Fig. 2). A partially obstructed tubing, acting as a second flow restrictor, would allow air exclusion and a very low flow rate of ropivacaine. This would stop when the tubing is connected to a long catheter with very small diameter (which generates further important flow resistance). Clamping the tubing is necessary to fill the Easypump but not the Infusor devices. This process was already suspected to decrease elastomeric pump flow rate.13 Batch-to-batch variation of deflation problem incidence might be explained by slight differences in batch-to-batch plastic tubing properties. No kinked pump tubing was reported in our series, but we cannot exclude that kinks were hidden in the clamps themselves (Fig. 2).
Previous studies suggested that there are deflation defects with elastomeric pumps, with reports of abnormal prolonged times of infusion (up to 200% of nominal time).1,13 These were interpreted as poorly accurate flow rates but, in the absence of pump weighing, we cannot exclude a delayed device deflation followed by a subnormal flow rate, as in our study. An additional clue is that in 1 in vitro study, 2 devices of 40 were excluded because delivery stopped during the assay.9 Based on the literature and our data, we recommend regular weighing of elastomeric pumps. From in vitro trials, flow rate is known to increase significantly at the beginning and at the end of the infusion.7,9–11,17 The effect of this phenomenon on our results is anecdotal since we excluded measures at the beginning and the end of infusion.
Several additional factors could explain the observed inaccurate flow rates of elastomeric pumps. Patient movements and positioning may induce transient occlusions of the tubing or catheter. Due to their flow restrictor being calibrated at skin temperature, each modification of skin temperature (fever, chills, warm suits or blankets) will increase the pump flow rate.7,10,17 Greater variability of the Easypump flow rate, and its tendancy toward a higher flow rate than expected could also be explained. In vitro, the Infusor flow restrictor is far more consistent than the Easypump flow restrictor.18 Due to its soft external housing, an involuntary external compression of an Easypump device (by the patient leaning against it during the night for example) will increase its flow rate by 87% to 105%.18 This problem cannot occur with rigid Infusor devices. Discrepancies between in vitro and in vivo flow rates may be explained by these numerous clinical variables, that are controlled in vitro, but unpredictable in vivo in terms of frequency and intensity from one patient to another.
Our clinical study on elastomeric pumps shows that device dysfunction, mainly an absence of deflation, must be considered in case of early failure to achieve postoperative perineural analgesia. In vivo measurements of their infusion rate yielded greater variability than in vitro measurements. For these reasons, we recommend weighing such devices before connection to the catheter, and then regularly, especially during the first hours of infusion, to verify that their weights are decreasing and local anesthetic is being infused.
Mr. Stéphane Cornec for designing the device register in the recovery room, Mr Philippe Pilloy for organizing temperature follow up in the recovery room. Mrs Andrée Verrier and Stephanie Pybot, secretaries, for editing patients’ information files.
1. Skryabina EA, Dunn TS. Disposable infusion pumps. Am J Health Syst Pharm 2006;63:1260–8
2. Richman JM, Liu SS, Courpas G, Wong R, Rowlingson AJ, McGready J, Cohen SR, Wu CL. Does continuous peripheral nerve block provide superior pain control to opioids? A meta-analysis. Anesth Analg 2006;102:248–57
3. Capdevila X, Dadure C, Bringuier S, Bernard N, Biboulet P, Gaertner E, Macaire P. Effect of patient-controlled perineural analgesia on rehabilitation and pain after ambulatory orthopedic surgery: a multicenter randomized trial. Anesthesiology 2006;105:566–73
4. Zahnd D, Aebi S, Rusterholz S, Fey MF, Borner MM. A randomized crossover trial assessing patient preference for two different types of portable infusion-pump devices. Ann Oncol 1999;10:727–9
5. Capdevila X, Macaire P, Aknin P, Dadure C, Bernard N, Lopez S. Patient-controlled perineural analgesia after ambulatory orthopedic surgery: a comparison of electronic versus elastomeric pumps. Anesth Analg 2003;96:414–7
6. Schleis TG, Tice AD. Selecting infusion devices for use in ambulatory care. Am J Health Syst Pharm 1996;53:868–77
7. Ackermann M, Maier S, Ing H, Bonnabry P. Evaluation of the design and reliability of three elastomeric and one mechanical infusers. J Oncol Pharm Pract 2007;13:77–84
8. Mizuuchi M, Yamakage M, Iwasaki S, Kimura A, Namiki A. The infusion rate of most disposable, non-electric infusion pumps decreases under hypobaric conditions. Can J Anaesth 2003;50:657–62
9. Chung IS, Cho HS, Kim JA, Lee KH. The flow rate of the elastomeric balloon infusor is influenced by the internal pressure of the infusor. J Korean Med Sci 2001;16:702–6
10. Ilfeld BM, Morey TE, Enneking FK. Portable infusion pumps used for continuous regional analgesia: delivery rate accuracy and consistency. Reg Anesth Pain Med 2003;28:424–32
11. Ilfeld BM, Morey TE, Enneking FK. Delivery rate accuracy of portable, bolus-capable infusion pumps used for patient-controlled continuous regional analgesia. Reg Anesth Pain Med 2003;28:17–23
12. Dadure C, Raux O, Gaudard P, Sagintaah M, Troncin R, Rochette A, Capdevila X. Continuous psoas compartment blocks after major orthopedic surgery in children: a prospective computed tomographic scan and clinical studies. Anesth Analg 2004;98:623–8
13. Kaye T. Prolonged infusion times with disposable elastomeric infusion devices. Am J Hosp Pharm 1994;51:533–4
14. Ganapathy S, Amendola A, Lichfield R, Fowler PJ, Ling E. Elastomeric pumps for ambulatory patient controlled regional analgesia. Can J Anaesth 2000;47:897–902
15. McLeod GA. Density of spinal anaesthetic solutions of bupivacaine, levobupivacaine, and ropivacaine with and without dextrose. Br J Anaesth 2004;92:547–51
16. Trissel LA. Handbook on injectable drugs. 14th ed. Bethesda: Am Society of Health-System Pharmacists, 2007
17. Ilfeld BM, Morey TE, Enneking FK. The delivery rate accuracy of portable infusion pumps used for continuous regional analgesia. Anesth Analg 2002;95:1331–6
18. Thiveaud D, Demaziere V, Lafont J. Comparison of the performance of four elastomeric devices. Eur J Hosp Pharm Pract 2005;11:54–6