Needlestick injuries (NSI) are a major occupational hazard for health care workers. Current estimates indicate that between 600,000 and 800,000 NSI, or 30 NSI per 100 hospital beds, occur annually in the United States, and these injuries have the potential to cause serious bloodborne infections (1–3). At our institution in 2002, there were 212 reported NSI. Although treatment strategies for NSI have improved in recent years, the risk for contracting a bloodborne infection remains significant, and efforts to develop new devices that reduce exposure risk are continuing (2–4).
Central venous catheters traditionally have been secured to the skin by using sutures. The Arrow kits provide 3-0 silk suture on a straight Keith needle (Arrow Silk 000; K-04200-001; Arrow International, Inc., Redding, PA). In our hospital, however, this simple task of securing catheters contributes significantly to NSI. In our department, 4 (28%) of the 14 NSI in the first half of 2003 were related to securing central lines. In many areas of the hospital, the straight needle for this use is prohibited, and providers must secure the catheters by using alternative methods. The most widely used alternative, but one that does not completely eliminate the risk of NSI, involves using a needle driver and a curved cutting needle.
Recently, a stapler (NPX8506-A; Arrow International) was developed for attaching central venous catheters to the skin. Each stapler is preloaded with one staple that is easily deployed by squeezing the finger-held device. Although staples have proven to effectively close surgical wounds (5), little work has been performed to compare sutures and staples with respect to their ability to prevent dislodgement. Determining the anchoring strength of staples used for this purpose is important because there may be substantial risks involved in reinserting inadvertently dislodged catheters.
This study was designed to test the hypothesis that the axial force and torque necessary to detach a 16-gauge catheter hub from human skin is related to the mechanism of attachment. Specifically, we compared the peak axial force and torque required to remove a catheter hub attached to human skin with 3-0 silk suture, 0.022-in.-diameter staples (K-15703-007; Arrow International), and 0.025-in.-diameter staples (NPX-8599-s; Arrow International) (Fig. 1).
After IRB approval, we used discarded extremities after below-knee amputations to serve as a surrogate of the tissue that would be used in patients for catheter stabilization. The need for obtaining informed consent was waived by the IRB because the tissue was otherwise discarded. The most proximal portion of amputated specimens typically contains some amount of healthy tissue to improve the quality of the surgical closure of the remaining extremity stump by leaving the wound free of necrotic tissue. After amputation, the specimens were double-wrapped in plastic bags and stored at 2°C to 8°C in the surgical pathology refrigerator. During the dislodgment trials, catheter hubs were secured to areas of grossly normal-appearing tissue on the proximal end of the discarded limbs within 24 h of amputation.
For each dislodgement, an 8-mm-diameter stainless-steel rod was used to tunnel a 7F triple-lumen central venous catheter (MC-15703; Arrow International) under the skin from the site of hub attachment to simulate placement of the catheter beneath the skin and into a central vessel. The catheter hub was then secured to the skin by using 3-0 silk suture or staples. Securing the hub with suture was accomplished with 3-0 silk suture on a straight needle. For each of two eyelets, the needle was first passed through approximately 1 cm of skin and then through the eyelet. Five interlocking square knots were used to anchor the eyelet to the skin. The orientation of the stitch was perpendicular to the axis of the catheter (Fig. 2).
When securing the hub with staples, one end of the staple was passed through the hub eyelet and positioned against the skin. The other end of the staple was then positioned in the desired orientation, and the stapler was squeezed to close the staple. The four staples were oriented at 45°, 135°, 225°, and 315° with respect to the catheter axis (Fig. 2). A slight pinch of the skin was required to stabilize the area of skin to be stapled.
A handheld force transducer (Shimpo Digital Force Gauge, FGV-10; Nidec-Shimpo America Corp., Itasca, IL) was clamped to the distal end of the catheter hub by using a specially machined attachment adapter (Fig. 3). The transducer was connected to a Dell laptop computer (Inspiron; Dell, Austin, TX) for data acquisition via the serial port. The DASY Lab software package (Version 5.5; Dasytec USA, Amherst, NH) recorded force in pounds as a function of time. The frequency of data sampling was 10 Hz.
The catheter hub was covered with an opaque paper drape before the application of force. A brisk pull was then initiated manually, and the applied force was increased at a mean rate of 11 ± 2 N/s until the hub was dislodged from the skin, defined by the failure of the sutures or staples at all attachment points. The peak dislodgement force was subsequently recorded. The paper drape was then removed, and the mechanism of failure was identified. The process of inserting a new catheter into another suitable area was then repeated at least 3 cm from any previous site.
A trial consisted of three individual dislodgements—one for each attachment method. All dislodgements for each trial were performed on the same specimen, and multiple trials were repeated on the same specimen if enough normal-appearing tissue was present. Specifically, 2 specimens were used once, 4 specimens were used twice, and 2 specimens were used 3 times, for a total of 16 trials conducted on 8 specimens.
The catheter hubs used for the torque dislodgement trials were predrilled with a 1/16-in. drill bit to create 2 holes through which a U-shaped pin was inserted from the underside of the catheter hub. The pin was positioned to align the hub eyelets with the torque axis. A handheld torque meter (TMG-15; Imada, Inc., Northbrook, IL) was then attached to the pin on the catheter hub (Fig. 4). The transducer was connected to the laptop computer, and data were acquired in the same fashion as described for axial force dislodgement. A twist was then initiated manually: the applied torque was increased until the hub was completely dislodged from the skin, defined by the failure of the sutures or staples at all attachment points.
Statistical analysis was performed with the SPSS statistical analysis software package (SPSS Inc., Chicago, IL). Mean and sd for all attachment methods were calculated. One-way analysis of variance and post hoc comparisons with Dunnett’s test were conducted to compare the peak dislodgement force among the three groups by using the suture group as the control. P values <0.05 were considered significant.
The mean peak force required for axial dislodgement was greater for 3-0 silk suture (40.9 ± 10.7 N; n = 16) than for 4 0.022-in.-diameter staples (34.0 ± 7.2 N; n = 16; P = 0.04) but was not different from that required for 4 0.025-in.-diameter staples (40.4 ± 5.8 N; n = 16). With respect to the mechanism of failure, the suture group failed at the level of the skin in 2 of 32 trials (intact suture loops were pulled through the skin), at the suture in 29 of 32 trials (the suture broke), and at the level of the eyelet in 1 of 32 trials (intact suture loops were pulled through the eyelets). In all staple trials, including both 0.022-in.- and 0.025-in.-diameter staples, all attachment failures were the result of deformed staples pulling free from the skin, and none of the attachment failures resulted in tearing of the eyelets or skin.
During these trials, pure rotation around the catheter hub resulted in an elastic deformity of the hub and eyelet preventing enough torque to be translated to the anchoring points to cause failure. Consequently, we were unable to obtain peak torque measurements for this mechanism of dislodgement, and these trials were abandoned early in the study.
The major finding of our study was that attaching a central venous catheter hub to human skin with 3-0 silk suture prevents axial dislodgement better than 4 0.022-in.-diameter staples and is similar to 4 0.025-in.-diameter staples. Our results are different from those of the only published study that used staples to secure central lines in processed human cadavers (6). In that study, the dislodgement force required to remove a catheter hub secured with only 2 staples from the skin was similar to the force required to remove a catheter secured with 3-0 silk suture from the skin; the peak dislodgement force for catheters secured with 4 staples was significantly greater than that required to dislodge those secured with 2 staples or suture. Our results also differ from those of an unpublished study conducted by Arrow investigators, who found no difference in the axial force required to remove a central venous catheter hub secured to sheep skin with 4 of the smaller 0.022-in.-diameter staples or with 3-0 silk suture.
Because the methods of our axial dislodgement trials were patterned after the previously described studies, one explanation for the difference in our results is that fresh human skin exerts less frictional force against staples than does either processed cadaveric skin or dried sheep skin. Anecdotally, the authors found, during a few “practice” catheter dislodgement attempts, that the staples were much harder to deploy into the tougher and thicker sheep skin than into the supple human tissue; also, once deployed, they were more difficult to extract.
Our assumption that the normal-appearing skin from surgically removed legs has similar mechanical properties to the skin of the neck and chest of living humans is one of the primary limitations of this study. We chose a surrogate for the skin of a living patient because it would be unethical to subject patients to the risks of tissue injury and catheter dislodgement. Although we performed our tests within 24 hours of limb removal and the specimens appeared grossly normal, microscopic tissue necrosis may have been occurring during this period.
Another limitation of this study involves conducting multiple axial dislodgement trials on a single specimen. The rationale was to maximize our ability to collect data from a limited supply of specimens. Ideally, only 1 set of 3 individual trials (sutures, 0.022-in. staples, or 0.025-in. staples) would have been performed on each specimen.
In a pure rotational motion around the catheter hub’s axis, material properties appear to provide protection from dislodgement by allowing the hub and eyelets to deform and prevent the translation of rotation into force at the points of attachment to the skin. However, this observation applies only to these specific catheters with their relatively pliable hubs. Other catheters, particularly those with stiff plastic hubs, as typically found on introducer sheaths, and those with adjustable catheter clamps may be able to more effectively translate rotation into force at the attachment points. Further investigation with different catheters may be warranted in this area.
In conclusion, central venous catheters secured to human skin with 4 0.022-in.-diameter staples have less peak dislodgement force than either 4 0.025-in.-diameter staples or 3-0 silk suture. There appears to be no difference in peak dislodgement force between 4 0.025-in.-diameter staples and 3-0 silk suture. Future investigation may be needed to determine whether 4 0.025-in.-diameter staples and 3-0 silk sutures have similar securing strength in other mechanisms of dislodgement and whether safety devices, including central line staples, actually reduce the incidence of NSI among health care providers.