| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1996;62:624-626
© 1996 The Society of Thoracic Surgeons
Medical Incorporated, 9605 West Jefferson Trail, IInve Grove Heights, MN 55077
To the Editor:
There has been a lack of information in the medical literature about the evolution of the two distinct generations of Omniscience heart valves. A clarification of the development of the current Omniscience valve design has been published [1], and the differences between the two Omniscience generations have appeared in at least one textbook [2]. However, this information is still not well disseminated. Unfortunately, Dr Akins' recent review article [3] rekindles the confusion about this subject and obfuscates the data pertaining to the current model of the Omniscience heart valve. I would like to take this opportunity to clarify the differences between the early, investigational model (1978 to 1982) and the currently marketed Omniscience valve (1982 to present).
A few European centers had high thromboembolic rates with the earlier Omniscience model in the mitral position [4–6], whereas satisfactory results were obtained from other centers [7]. Examination of multicenter results revealed that other factors, in addition to valve design, also contributed to the poor results. A combination of oversizing (during implantation or latent due to myocardial recovery), posterior orientation of the disc motion, and a thin suture ring, which allowed inadequate clearance between the disc path and ventricular structures, contributed to the high incidence of mitral thromboembolism. Refinement of the Omniscience design was undertaken in 1982 to eliminate its sensitivity to such surgical and patient factors [1].
Three of the six studies cited in the review article [4–6] report data that pertain exclusively to the investigational Omniscience valve. Because this first-generation prosthesis has been obsolete since 1982, it is more correct to remove these studies from the composite data presented. Actually, by including two publications from the same patient group (Rábago and associates [5] and Corina and colleagues [6]), the review has further confounded the Omniscience valve data. Recalculation of the complications presented in this review with exclusion of these articles yields a linearized rate of 2.3%/patient-year for mitral valve thromboembolism (including valve thrombosis) and 0.9%/patient-year for the subset of valve thrombosis, compared with 4.4%/patient-year and 2.3%/patient-year, respectively, as reported in the review article.
The review of the Omniscience performance in the aortic position included a report by Kazui and associates [8]. In an update of this cohort's experience [9] (not included in the review), Kazui and associates report lower thromboembolism rates: 2.9%/patient-year (559 patient-years) versus 4.7%/patient-year (192 patient-years) in the original report [8]. However, they also state that 71% of their patients received the obsolete Omniscience valve model (moreover, 17% of the cohort had double valve replacement). Based on this high percentage of patients with the earlier model, it can be argued that this report also should be excluded from the review.
The review also offers a couple of opinions on which I feel compelled to comment. The opening characteristic of the Omniscience valve was judged to have the lowest rating among the valves studied. However, complete opening of the Omniscience valve-ie, opening to its maximum angle of 80 degrees-is not expected in patients who are at rest. Patients in sinus rhythm have been well-served by monoleaflet prostheses designed to open to a maximum opening angle of 60 degrees (eg, Björk-Shiley valves). The Omniscience valve was designed with reserve opening capacity to respond to higher cardiac output conditions. Exercise hemodynamic studies have proved that the Omniscience valve does indeed respond to higher cardiac output, as evidenced by significantly increased effective valve area [8, 10]. In summary, the Omniscience disc opening is flow-responsive. Characterizing opening as "incomplete" without considering the adequate extent of opening, or flow requirements, is ambiguous terminology that can be misleading.
The Omniscience valve regurgitation/leakage was described as equivalent to that of the Medtronic-Hall prosthesis in the review article, and transvalve gradients were subjectively rated as falling between those of the ball and disc prostheses. There is no reference to support this judgement. Clinically, it is almost impossible to differentiate small pressure gradient differences by Doppler echocardiography. This is further complicated by the phenomenon of pressure gradient overestimation due to localized high-velocity flow patterns. Several reports, therefore, indicate no statistically significant difference between various monoleaflet and bileaflet valves [11–13]. In laboratory measurements (accuracy approximately 0.5 mm Hg), the Omniscience valve shows transvalve gradients equivalent to those of other monoleaflet valves (eg, Medtronic-Hall, Monostrut), which are generally slightly higher than that of the bileaflet St. Jude Medical valve [14]. However, it is well accepted that there are three distinct cardiac energy losses attributable to a prosthetic heart valve's opening, closing, and closed phases-ie, during forward flow and through regurgitation backflow. These losses can be measured accurately only under controlled laboratory conditions. In vitro experiments by the Helmholtz Institute for Biomedical Engineering in Germany, an independent test facility, found that the Omniscience valve actually induced the least overall energy loss [14] among the mechanical valves discussed in this review.
In addition, I would like to correct the date that the Omniscience valve was approved (Table 1 of the review article [3]). The investigational Omniscience valve model was introduced in 1978, the modified (current) design was available in 1982, and the Food and Drug Administration granted premarket approval in 1985.
I hope that the above comments help clarify to the profession the contradictory information about the Omniscience cardiac prosthesis and assist in avoiding such confusion in future reviews.
References
Cardiac Surgical Unit, Massachusetts General Hospital, White 503, 32 Fruit St, Boston, MA 02114
To the Editor:
I appreciate Dr Mikhail's efforts to clarify any misconceptions that may have arisen concerning Omniscience heart valves, as reported in my recent review [1]. His comments on the design changes in the 1980s expand upon remarks made by me about that topic in my original review [2].
Doctor Mikhail's comments about the articles used in the review warrant further response. Doctor Mikhail correctly notes that the data from two reports came from the same patient group from the same center in Spain [3, 4]. However, in my review only the aortic valve data from the original report by Rábago and associates [3] and only the mitral data from the later report by Cortina and colleagues [4] were used to generate the composite linearized rates of complications for the two valve positions. The later report by Cortina and colleagues, which focused only on mitral prostheses, contained longer follow-up on those valves.
Surprisingly, the longer follow-up by Cortina and colleagues showed a higher rate of thromboembolism for mitral Omniscience valves (7.6%/patient-year) than the earlier report by Rábago and associates (6.8%/patient-year). Usually longer follow-up shows lower thromboembolic rates because of the higher hazard function for that complication early after valve implantation, as in the case of the recent update [5] of the original Omniscience aortic valve experience by Kazui and associates [6]. In that study the more recent report notes a lower linearized thromboembolic rate with longer follow-up.
However, as Dr Mikhail noted, Kazui and associates' series exemplifies several other difficulties in finding comparable patient populations in the literature. The original report states that 15 of the 100 aortic valve patients also received a mitral prosthesis; the later report says that 17 patients had an associated mitral valve replacement. As Dr Mikhail suggests, maybe neither report ought to be used. Unfortunately, the update of Kazui and associates' experience also does not provide either the number of incidents or linearized rate for anticoagulation-related bleeding. Thus, that report could not be used to generate the composite thromboembolism and bleeding index for that valve.
Incomplete opening of the Omniscience valve, noted to be even less than 60 degrees in 18% of 50 unselected patients in one study [6], is a documented clinical feature of the Omniscience valve, the advantage or disadvantage of which remains open to debate. Although proponents of the valve may suggest that this represents "reserve opening capacity," patients in the cited study with an opening angle of less than 60 degrees had a thromboembolic rate of 44.4% versus 12.2% for those with valves that opened to 60 degrees or greater [6].
The assessment of valve gradients and regurgitation noted in the review was intentionally described as "very subjective." Not all in vitro and in vivo studies report the same findings. An in vitro study by Scotten and associates [7] notes modestly higher transmitral gradients and lower valve areas for the Omniscience versus the St. Jude Medical, Björk-Shiley convexo-concave, and Hall-Kaster (currently the Medtronic-Hall) valves, but also notes smaller closed leak back volumes. If the occluder does not open fully in an important percentage of patients clinically, then higher transvalve gradients must result.
In summary, Mikhail has reemphasized some important issues. We all look forward to continuing reports on Omniscience valves so that we may be able to compare them with the published reports on the other available mechanical valves.
Further, I wish to correct any confusion about the "newness" of the composite thromboembolism and bleeding index included in my recent review of mechanical valves [1]. For many years the dual complications associated with anticoagulation, namely thromboembolism and anticoagulant-related bleeding, have been a topic of considerable interest and discussion. In the cardiac surgical literature this discussion has been addressed by many authors, including McGoon [8], who in 1984 decried the poor quality of published reports on the risk of thromboembolism after valvular operations and noted that focusing only on thromboembolism did not "address the closely related and also important problem of hemorrhagic complications associated with anticoagulant therapy."
Indeed, in 1987 Edmunds [9], in one of the initial published reviews of thrombotic and bleeding complications associated with prosthetic heart valves, actually provided a summation of embolus, thrombosis, and bleeding complications from some of the then-available valvular prosthesis studies. His tables of data were generally provided for individual reports, were not cummulated for experience with each different brand of prosthesis, and in numerous cases suffered from the fact that "not all thrombotic and bleeding complications are included in citation(s)."
The only truly new aspect to my composite thromboembolism and bleeding index arises from a concept from the original review [2] of the "composite" rate for complications, that is, the numeric summation of all incidents of the same complication for the same prosthesis published in the literature, with the denominator arising from the cumulative patient-years of follow-up from the cited reports. The goal of the process was to cumulate the available published data in an effort to provide a broader view of the performance of various prostheses, while acknowledging that great variations existed in the patient populations whose results were cumulated. In essence it is at best a first approximation of a metaanalysis without statistical comparisons.
The view of thromboembolism and thrombosis on the one hand and anticoagulant-related bleeding on the other as two sides of a coin is not new. The goal of attaining no prosthetic thrombosis at the lowest rates of bleeding and thromboembolism has been emphasized by various authors, including Butchart and colleagues [10], from whom the principle depicted in Figure 2 [1] is derived. The importance of this concept has been one of the driving forces getting hospitals in the United States to convert to monitoring warfarin anticoagulation with the international normalized ratio, not just prothrombin times.
In summary, the recently published review is meant not so much to definitively evaluate mechanical heart valves as it is to cumulate results against which to compare new data, to stimulate our continued evaluation of present prostheses, and to encourage surgeons to reassess the guidelines by which we perform those evaluations.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
