The Sherlock 3CG Tip Confirmation System for placement of peripherally inserted central catheters

Full details of all clinical outcomes considered by the Committee are available in the assessment report overview.

The key clinical outcomes for the Sherlock 3CG Tip Confirmation System (TCS) presented in the decision problem were:

• accuracy of catheter tip placement

• incidence of catheter malposition

• need for catheter repositioning

• impact of malposition‑related complications such as infection or thrombosis

• treatment delay following catheter placement

• reduced staff time

• reduced hospital stay

• need for confirmatory chest X‑ray

• need for fluoroscopy to place the peripherally inserted central catheter (PICC) tip correctly

• time taken to insert PICC

• PICC failure and reinsertion rates

• patient experience measures

• quality of life

• device‑related adverse events.

The External Assessment Centre considered that 5 of the 14 outcomes in the decision problem were reported in the published evidence. These were: accuracy of catheter tip placement; incidence of catheter malposition; treatment delay following catheter placement; change in staff time, and need for confirmatory chest X‑ray. The External Assessment Centre considered that some additional outcomes had also been partially addressed. Outcomes for which no evidence was presented included reduced hospital stay and treatment delay following catheter placement. The company stated that it was unable to report on device‑related adverse events because of a lack of reported evidence.

The company identified 13 studies from its literature search but it excluded 9 and presented 4 published abstracts (Adams et al. 2013; Barton, 2014; Parikh, 2012; Stewart, 2013). It also presented responses from a questionnaire sent to 6 NHS hospitals, as supporting clinical evidence. The External Assessment Centre considered that the studies presented by the company were in keeping with the scope and were appropriate for inclusion. It also identified 1 study published after the company submission and that it considered suitable for assessment (Johnston et al. 2014). To ensure that all relevant evidence was identified, the External Assessment Centre carried out a further literature search with a wider scope which included any previous model of the device that had both magnetic tracking and electrocardiogram (ECG) tip confirmation components. One additional presentation was identified (Symington et al. 2013).

Johnston et al. (2014) reported a retrospective case‑series review of the first 250 patients to have PICCs inserted using the Sherlock 3CG TCS following its introduction to a UK NHS hospital. The population comprised patients in the intensive care unit (ICU). The vascular access team placed PICCs at the bedside, and used a portable chest X‑ray to confirm the tip location. Two independent reviewers examined the X‑rays. From the first 250 patients, 11 were excluded because of: failed insertion (n=2); no chest X‑ray being taken after the procedure (n=2); a failure to identify the tip position on the chest X‑ray (n=2); a failure to interpret the ECG criteria (n=4); and the catheter being too short (n=1). Tip location was reported for the 239 PICC placements where ECG was used for tip confirmation. Although there was no direct comparator for the intervention in this study, the same authors published a retrospective service evaluation a year before the Sherlock 3CG TCS was introduced, reviewing records for both ICU patients (n=246) and non‑ICU patients (n=233, Johnston 2013). The External Assessment Centre used this as a form of comparator to assess the impact of the Sherlock 3CG TCS on malposition rates. Both Johnston studies reported results using 2 different definitions of malposition, 1 from the USA and the other from Europe. The definition of appropriate placement typically used in US guidelines is the low superior vena cava or cavoatrial junction (National Association of Vascular Access Networks 1998, Infusion Nurses Society 2006, Funaki 2002). A European guideline uses a broader definition, stating that appropriate placement is in the mid or lower superior vena cava, cavoatrial junction, or high right atrium (Pittiruti 2009). Using the definition as per US guidelines, 56.1% (95% confidence interval [CI] 50% to 62%, n=134) of ICU patients had a malpositioned PICC using the Sherlock 3CG TCS compared with 76% (n=187) who had blind bedside insertion. Using the definition as per European guidelines, 20.5% (95% CI 16% to 26%, n=49) of ICU patients had a malpositioned PICC using the Sherlock 3CG TCS compared with 50.8% (n=125) who had blind bedside insertion. The malposition rate using the Sherlock 3CG TCS was significantly lower than blind placement using both sets of criteria (p<0.0001). However, it was also substantially higher than that reported in other studies. The authors suggest several reasons for this. They noted that it may be difficult to determine the exact point of a maximum or biphasic P wave for patients in intensive care, who may have ECG artefacts due to comorbidities. The authors also noted that tip position is not static, and that the catheter tip may move (due to, for example, arm movement, because the PICC is placed with the arm drawn away from the body and the chest X‑ray is taken with the arm drawn towards the body). The authors concluded that, if the European guideline definition of an adequate tip position is considered to be acceptable, the Sherlock 3CG TCS can be used for tip confirmation without chest X‑ray. If a more precise tip position of low superior vena cava or cavoatrial junction is used, as in the US guidelines, a chest X‑ray may be necessary.

Adams et al. (2013) presented a poster reporting on the introduction of the Sherlock 3CG TCS to a healthcare centre in the US. Over a 9‑month period, 333 patients had PICC insertion using the Sherlock 3CG TCS, which was subsequently verified using chest X‑ray and confirmed by 2 radiologists. Accurate placement was defined as the catheter tip being in the distal superior vena cava or at the cavoatrial junction. The Sherlock 3CG TCS was used to confirm tip position in 83.5% of patients (278/333). In the remaining 16.5% (55/333), the ECG system could not be used either because of an abnormal P wave (12.9%) or because of technical factors such as loose connections and poor electrode placement (3.6%). When the Sherlock 3CG TCS was used to confirm tip position, 1 radiologist reported that 96.4% (268/278) of PICCs were placed accurately, and that 3.6% (10/278) were malpositioned; the other radiologist reported that 98.2% (273/278) of PICCs were placed accurately and 1.8% (5/278) were malpositioned. In 2011, the malposition rate using the predecessor device, the Sherlock Tip Location System (magnetic tracking only) was reported to be 14% based on a subsequent chest X‑ray. Adams et al. also reported that the PICC was ready for infusion 61.0 minutes earlier using the Sherlock 3CG TCS (39.5 minutes) than using a chest X‑ray and a radiologist report for tip position confirmation (101.0 minutes), although no information was reported on how this was measured. The researchers confirmed that chest X‑rays are no longer mandatory for PICCs placed using the Sherlock 3CG TCS at this centre.

The abstract by Barton (2014) described the introduction of the Sherlock 3CG TCS to a nurse‑led PICC service at a UK NHS hospital. In an initial trial, clinicians used the Sherlock 3CG TCS for PICC placement in 65 adults with no atrial fibrillation. They used chest X‑rays, reviewed by an independent physician, to confirm tip location. Following the initial trial, an application was made to amend local protocol and remove the need for a mandatory chest X‑ray following PICC placement. During the application process, clinicians placed another 160 PICCs using the Sherlock 3CG TCS, with position confirmed using a chest X‑ray. In total, data were reported on 225 patients. The definition of acceptable tip position was the lower third of the superior vena cava or the cavoatrial junction, as used in US guidelines. Chest X‑rays confirmed that tip position was acceptable in 100% of cases reported. Only success rates of tip positioning in patients for whom magnetic tip position and ECG tip confirmation could be used were reported. Cases where the Sherlock 3CG TCS was not suitable or where there was a failure of the ECG system were not included. The authors reported to the External Assessment Centre that, during the trial period, 2 patients were not suitable for the Sherlock 3CG TCS and had PICCs placed with fluoroscopy. Since the introduction of the Sherlock 3CG TCS, 11 patients needed chest X‑rays due to the failure of the ECG system to provide tip confirmation. Five of these cases were because of atrial fibrillation, and 6 because of a failure of the electrode connections. The hospital has since removed the need for chest X‑ray after PICC placement using the Sherlock 3CG TCS.

Parikh (2012) presented a poster reporting a prospective case series from October 2011 to April 2012 of 247 PICCs placed in 221 patients (mean age 62 years, range 15 to 100) in a US hospital. The Sherlock 3CG TCS was used for tip placement and confirmation, except in patients with atrial fibrillation, atrial flutter or no discernible P wave (15.4%, 38/247). Tip position was confirmed by chest X‑ray and evaluated by 2 independent observers. Successful tip placement was defined as the superior vena cava or cavoatrial junction, as in the US guidelines. The study was divided into 2 phases. Phase 1 was a voluntary training phase. Nurses who wished to be trained in the use of the Sherlock 3CG TCS (4 of 7 nurses) had training, which consisted of a PICC refresher course, an online course, a 1‑hour taught course and 1‑to‑1 training with a nurse trainer provided by the company. The nurses then placed 62 PICCs using the Sherlock 3CG TCS. As per the exclusion criteria, 3 patients were excluded. Successful tip placement was 83% (n=62) for those using the Sherlock 3CG TCS. For phase 2, the 3 other nurses who had not had training in the Sherlock 3CG TCS had phase 1 training. All 7 nurses then inserted 5 PICCs while being observed by a nurse trainer. All staff completed phase 2 training. Staff placed 147 PICCs using the Sherlock 3CG TCS, excluding 35 patients as per the criteria. Successful tip placement was 96% (n=147) for those using the Sherlock 3CG TCS. From November 2012 to May 2013, staff placed a further 567 PICCs, 437 using the Sherlock 3CG TCS. Of these, 24.9% (109/437) still needed a chest X‑ray for confirmation, for reasons that included unclear baseline rhythm, and complicated or uncertain PICC placement. It is unknown if the PICCs which did not use the Sherlock 3CG TCS needed a chest X‑ray to confirm, but this is probable.

An abstract and poster by Stewart (2013) presented a study in an Australian hospital which recruited over 65 patients between November 2012 and March 2013. The exact number of patients and methodology were not reported. Clinicians placed PICCs using the Sherlock 3CG TCS and confirmed the tip position using a chest X‑ray. No information was given on what tip positions were considered to be acceptable or who reported on the chest X‑ray. The abstract reported that 100% of malpositions were corrected at time of placement. Of PICC placements using the Sherlock 3CG TCS, 96% were within the cavoatrial junction. The other 4% were reported in the right atrium. Discrepancies were noted between locations reported by ECG and X‑ray, which were resolved with clinical experience and collaboration. A time saving of 1 hour and 51 minutes was reported, being the average wait time between PICC insertion and X‑ray results. No information was given on how time savings were measured or if there was any resulting change in treatment time or outcomes.

Symington et al. (2011) was a conference presentation on a US centre using the Sapiens TCS in conjunction with the Sherlock TLS II. These devices are the predecessor devices to, and when used together have the same mode of action as, the Sherlock 3CG TCS. The author reported on a consecutive case series during April 2011 (n=63). No information was given on the patient population. The company provided training. Tip placement was verified with a chest X‑ray, reviewed by the author. The author reported a 5% technical failure rate, including difficulties with cannulation of the vein, advancement of the catheter and occluded veins. It was reported that technical failures were 'thrown out', although this was not explained in greater detail. The authors reported that 62 of 63 tip placements were appropriately positioned, although no specific criteria for appropriate placement were reported. They reported that, by July 2011, there had been 604 PICC placements using the Sapiens TCS in conjunction with the Sherlock TLS II. The lead author reported that he was formally requesting that his hospital remove the need for mandatory chest X‑ray from its procedural guidelines. He was also a paid presenter for Bard Access Systems (a division of CR Bard), which was clearly stated.

The company contacted 7 UK NHS hospitals currently using the Sherlock 3CG TCS and collected questionnaire responses from them. The initial clinical evidence submission did not include these data, but they were later provided to the External Assessment Centre. The External Assessment Centre judged that the structuring of the questions and format of the answers did not allow for the assessment of relevant evidence on outcomes as defined in the decision problem. The External Assessment Centre noted the variation in reported clinical practice for issues such as hospital policy for confirmation, typical levels of malposition, dealing with malpositions and PICC reinsertions. In general, all respondents reported fewer malpositions using the Sherlock 3CG TCS than before its introduction. The External Assessment Centre noted that there was a risk of bias because not all hospitals using the Sherlock 3CG TCS were asked to provide data to the company. To explore this, the External Assessment Centre contacted 7 of the other 8 hospitals (not included in the company's survey) currently using the device, and 1 hospital that had not responded to the company's initial questionnaire. The External Assessment Centre concluded that the hospital questionnaires provided no assessable data relevant to the scope.

The company did not identify any adverse events from the published literature, or from a search of the Medicines and Healthcare Products Regulatory Agency's website. The company retrieved 51 records from the US Food and Drug Administration's MAUDE database, but stated that they were not necessarily device‑related adverse events. The External Assessment Centre retrieved 100 records from the same database, using a wider search strategy. Adverse events submitted to MAUDE are not verified. No searches were carried out for adverse event reports from PICC insertion using comparator technologies (blind PICC insertion with chest X‑ray, or fluoroscopy). Reported adverse events included: broken or damaged wire tip or stylet (n=29); adverse patient reactions (such as shortness of breath; n=23); catheter malfunction (such as leaks or splits; n=18) and tip malposition (n=14). The External Assessment Centre sought clinical expert opinion, but could not rule out the possibility that the adverse events reported with the Sherlock 3CG TCS were common to all PICC insertion techniques.