Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting

Published evidence

Five studies are summarised in this briefing, involving a total of 1,715 patients.

The studies summarised are 1 systematic review (which assesses 2 randomised controlled trials and 6 retrospective cohort studies), and 4 observational studies.

The clinical evidence and its strengths and limitations is summarised in the overall assessment of the evidence.

Overall assessment of the evidence

Many studies evaluate Cor‑Knot in valve surgery. The most relevant are included in this briefing and include outcomes such as cardiopulmonary bypass time and aortic cross-clamp time. These are relevant because of the company claims that Cor‑Knot may reduce these times. The systematic review by Salmasi et al. 2019 includes 8 studies, of which 2 are randomised controlled trials. The studies that were not randomised or controlled are generally of a low methodological quality.

Some safety issues have been identified with Cor‑Knot. Sagheer et al. 2019 reported a metallic embolus in the brain of a patient 7 months after valve surgery with Cor‑Knot. The embolus is thought to have originated from the metallic valve or the Cor‑Knot fastener. Garrett Jr 2017 report a delayed metallic embolisation 5 years after robotic mitral valve repair using Cor‑Knot, which caused a stroke in the patient. The company said that after it discussed the case with Dr Garrett, they concluded that the metallic embolisation was not caused by Cor‑Knot. The author highlights the need for long-term surveillance after using Cor‑Knot.

Navas-Blanco et al. 2018 reported a full thickness injury to the left coronary sinus during an aortic valve replacement after using Cor‑Knot. The company says this was a result of unusually thin tissue and operator error. Baciewicz Jr 2018, Biefer et al. 2018, Brescia et al. 2017 and Balan et al. 2017 reported patients developing post-operative valve regurgitation after Cor‑Knot damaged the valve leaflets. The company acknowledges that 9 in 438,000 cases (0.0021%) result in leaflet perforation. The authors highlight the need to orient the Cor‑Knot fastener away from the prosthesis and for meticulous technique. Biefer et al. 2018 also suggested that a rigid mitral ring may be a better valve prosthesis because they have not had any episodes of perforation since using this. The company acknowledges that 11 titanium fasteners out of more than 7,188,000 sold have been involved in reported adverse events, with a complaint rate of 0.00015%. The company says the events are primarily related to surgical technique and that appropriate fastener orientation prevents leaflet damage.

More high-quality controlled trials in the NHS are needed to compare patient outcomes with Cor‑Knot with hand-tied knots and endoscopic knot pushers. These studies should include long-term follow up to assess the longer-term efficacy and safety of Cor‑Knot.

Salmasi et al. (2019)

Study size, design and location

A systematic review and meta-analysis of 8 studies (6 retrospective cohort studies, 2 randomised controlled trials) comparing operative efficacies between Cor‑Knot and manual knot tying in patients undergoing aortic valve replacement (AVR) or mitral valve repair. Location: UK.

Intervention and comparator(s)

Intervention: Cor‑Knot.

Comparator: manual knot tying.

Key outcomes

Four studies looked at AVR (Loberman et al. 2018, Beute et al. 2018, Lee et al. 2018 and Plestis et al. 2018). In these studies aortic cross-clamp time was significantly shorter with Cor‑Knot than with manual tied knots. The weighted mean difference (WMD) was 86.62 (95% confidence interval [CI] 15.9 to 157.4; p=0.016). Cardiopulmonary bypass time was shorter in 2 studies: median 86 minutes (interquartile range [IQR] 40) for Cor‑Knot compared with 113.5 minutes (IQR 42), p=0.02; and mean 104 minutes (± 22.5 minutes) for Cor‑Knot compared with 118.3 minutes (± 30.1 minutes), p<0.001. However, there was no difference in another study, which reported a mean of 100.1 minutes (± 21.3 minutes) using Cor-Knot compared with 107.3 minutes (± 20.4 minutes), p=0.1. The fourth study did not record cardiopulmonary bypass time. A meta‐analysis of the 4 studies showed no significant difference for 30‑day mortality between Cor‐Knot and hand tying (odds ratio 1.73; 95% CI 0.50 to 6.02; p=0.678). Using Cor‑Knot did not increase the risk of permanent pacemaker implantation, paravalvular leak, or 30‐day mortality. Five studies reported length of intensive care stay, 4 reported no difference between the 2 groups. One study showed significantly reduced time in hospital with Cor‑Knot: a median of 6 days (range 4 days to 17 days) compared with 7 days (range 4 days to 79 days), p=0.002.

Four studies looked at mitral valve repair (Grapow et al. 2015, Etiwy et al. 2018, Sabik et al. 2018 and Perin et al. 2019). Meta-analysis of these studies showed significantly reduced cardiopulmonary bypass time in the Cor‐Knot group compared with the manual knot tying group (WMD: 110.0; 95% CI 12.3 to 207.7; p=0.027). There was also significantly shorter aortic cross-clamp time in the Cor‑Knot group (WMD: 79.0; 95% CI 10.4 to 147.5; p=0.024). There were no differences in length of intensive care or hospital stay.

Strengths and limitations

The systematic review with meta-analysis is high-quality evidence. Statistical analysis of the relevant outcomes helps to objectively assess efficacy of Cor‑Knot compared with standard care.

The authors excluded publications if cardiopulmonary bypass time or aortic cross-clamp time were not reported, limiting the number of studies reviewed. The mitral valve intervention papers included minimal access studies and combined tricuspid repair studies, which may have affected operative times and outcomes. Only 2 randomised controlled trials were included. The remaining 6 were retrospective cohort studies, which are lower quality evidence.

Morgant et al. (2019)

Study size, design and location

A single-centre prospective study involving 221 patients undergoing isolated AVR surgery with stented prosthesis between September 2009 to June 2018. Location: France.

Intervention and comparator(s)

Intervention: Cor‑Knot (n=63).

Comparator: Hand-tied knots (n=158).

Key outcomes

Compared with the hand-tied knot group, the Cor‑Knot group had shorter aortic cross-clamp time (74 minutes [± 13.8 minutes] compared with 90.4 minutes [± 23.7 minutes], p<0.0001) and cardiopulmonary bypass times (100.8 minutes [± 20.6 minutes] compared with 117.6 minutes [± 33.1 minutes], p<0.0001). The differences in 30‑day mortality (1.2% and 0%, p=0.37) and stroke and transient ischaemic attack rates (2.5% and 1.6%, p=0.67) between the 2 groups were not statistically significant.

Strengths and limitations

Statistical analyses were done using a propensity score with 1:1 matching for automatically tied and hand-tied knots, helping comparability of outcomes between the 2 groups. Large sample size helps to improve reliability. The study was not done in the UK and so is not generalisable to the NHS.

Nifong et al. (2013)

Study size, design and location

A single-centre retrospective observational study involving 336 patients receiving Cor‑Knot or robot-assisted mitral valve repair between May 2000 and December 2012. Location: USA.

Intervention and comparator(s)

Intervention: Cor‑Knot (n=48).

Comparator: robotic suture tying (n=288).

Key outcomes

The mean time to place and secure sutures was significantly shorter for Cor‑Knot compared with robotic knot tying (107.4 seconds [± 50.4 seconds] compared with 151.8 seconds [± 71.4 seconds], p<0.02). Three other outcome times were also significantly shorter for Cor‑Knot than for robotic tying:

  • annuloplasty band placement (26.9 minutes [± 7.4 minutes] and 36.6 minutes [± 10.2 minutes], p<0.02)

  • cardiopulmonary bypass (144.9 minutes [± 30.1 minutes] and 160.3 minutes [± 40.1 minutes], p<0.02)

  • aortic cross clamp (94.7 minutes [± 31.1 minutes] compared with 123.0 minutes [± 33.3 minutes], p<0.02).

Strengths and limitations

The study is presented as a conference abstract, which is low-quality evidence, and with limited reporting of outcomes. It is not in the UK and so not generalisable to the NHS. The study period (from 2000 to 2012) is relatively old and therefore potentially less useful. There were fewer patients in the Cor‑Knot group than in the robotic suture tying group, with no information about the groups' baseline characteristics. This limits comparability of outcomes between the 2 groups.

Wang et al. (2017)

Intervention and comparator(s)

Intervention: MIAVR with adjuncts (Cor‑Knot, coronary sinus catheter; n=78).

Comparators: MIAVR without adjuncts (n=78), AVR through a median sternotomy (n=78).

Key outcomes

There was no difference in preoperative preparation time between the MIAVR with adjuncts group and the MIAVR without adjuncts group (83.1 minutes compared with 81.1 minutes, p=0.56, CI -4.7 to 8.6). MIAVR with adjuncts had shorter cross-clamp time (70.5 minutes compared with 108.1 minutes, p<0.0001, CI 31.1 to 44.0) and shorter cardiopulmonary bypass time (101.1 minutes compared with 166.1 minutes, p<0.0001, CI 54.0 to 76.1) compared with the without adjuncts group. Compared with patients who had AVR via sternotomy, MIAVR with adjuncts also had shorter cross-clamp time (70.5 minutes compared with 84.4 minutes, p<0.0001, CI 8.2 to 19.6) and cardiopulmonary bypass time (101.1 minutes compared with 127.7 minutes, p<0.0001, CI 19.3 to 33.9).

Strengths and limitations

Patients who had MIAVR with adjuncts were propensity matched against those who had MIAVR without adjuncts and AVR through median sternotomy for comorbidities such as age, gender and ejection fraction. This made baseline characteristics similar, allowing comparability. The procedures were performed by 12 surgeons, which helps to increase the reliability of the results. The type of prosthesis used varied between the groups, which may have affected outcomes. The MIAVR with adjuncts group included Cor‑Knot and coronary sinus catheter, so it's not possible to attribute outcomes solely to Cor‑Knot.

Liang et al. (2019)

Study size, design and location

A single-centre retrospective cohort study involving 114 patients who had MIAVR for isolated aortic valve disease between January 2011 and August 2018. Location: USA.

Intervention and comparator(s)

Intervention: Cor‑Knot, single-shot del nido cardioplegia, and rapid deployment valves.

Comparator: no facilitating technology.

Key outcomes

After Cor‑Knot was introduced, compared with no facilitating technology, aortic cross-clamp time reduced from a median of 96 minutes (IQR 84 to 103) to 84 minutes (IQR 73 to 90); p<0.001. Cardiopulmonary bypass time reduced from a median of 157 minutes (IQR 134 to 182) to 129 minutes (IQR 113 to 144); p<0.001.

Strengths and limitations

This is a single-centre study, and 98% of the procedures were done by the same surgeon, which limits the reliability of the results. Cor‑Knot was introduced in 2012. Only the outcomes for aortic cross-clamp time and cardiopulmonary bypass time were reported for Cor‑Knot. Other outcomes were reported after 2 other interventions were introduced, which makes results difficult to interpret or attribute to an individual intervention. Only 2 of the 114 people in the study were women.


The company has not submitted any sustainability claims.

Recent and ongoing studies

Multi-center register for the use of the Cor-Knot automated knotting system in heart valve surgery (RECORVA). identifier: NCT03681834. Status: not yet recruiting. Indication: vascular surgery. Last update posted September 24, 2018. France. Devices: Cor‑Knot.