Response to De Bruyn H, Collaert B, Linden U, Johansson C, Albrektsson T: Clinical outcome of Screw Vent implant. Clin Oral Impl Res 1999;10:139-148.
July 29, 1999
Prof. Dr. Dr. H.C. Niklaus P. Lang, MS
University of Berne
School of Dental Medicine
Department of Periodontology
and Fixed Prosthodontics
Dear Dr. Lang:
SUBJECT: Invalidity of the study by De Bruyn, Albrektsson et al., entitled "Clinical outcome of Screw-Vent implants– A 7-year prospective follow-up study" 1999 Clinical Oral Implant Research 1: 139-148, ("De Bruyn, Albrektsson 1999 study").
HISTORY: This is a follow-up study to a 1-year study published in the same journal by De Bruyn et al. 1992 Clinical Oral Implants Research 3:32-41 ("De Bruyn 1992 study") that reported on a comparison of "the clinical success of 85 Screw-Vent and 107 Branemark implants, consecutively installed in a private periodontal clinic under the same conditions and by the same operator." Unlike the 1999 study, the original one included evaluation of Branemark implants placed by the same periodontist.
1999 STUDY CRITERIA: claims to use a bone loss criteria of success first proposed by Albrektsson et al (JOMI 1986) and later adopted by the European Academy of Periodontology (Albrektsson & Isidor 1994).
The validity of the De Bruyn, Albrektsson 1999 study deserves close scrutiny for these reasons:
added 0.2mm for each of the remaining 6 years of follow-up to reach what it terms the
"critical bone loss value" of "2.7mm below abutment implant junction, which
corresponds by coincidence to the first implant thread" of the Screw-Vent implant.
This Technology Report is being prepared to evaluate the scientific objectivity of this article, and to critically analyze the validity of using the Albrektsson criteria of success that, as will be demonstrated, has undergone repeated modifications since first proposed. (Albrektsson et al., JOMI 1986).
The core issue raised by the present De Bruyn, Albrektsson et al. article (COIR 1999:10) is the legitimacy of applying the Albrektsson et al. bone loss criteria to evaluate any implant system. De Bruyn, Albrektsson et al. essentially assert that the Screw-Vent implants, which were successfully osseointegrated and in function for 7 years, must be called failures if bone recession occurs to the first thread. By not applying the same criterion to the Brånemark implants placed in the same study, the authors eschew the fact that the same degree of crestal bone loss is routinely observed with Brånemark implants (Adell et al., Int J Oral Surg 1981; Lindquist et al., JPD 1988).
The De Bruyn 1999 article cites a number of references to support criticism of the Screw-Vent implant’s design, surface and clinical efficacy, but a review of these studies reveals an alarming discrepancy between what the referenced articles actually state. Most readers rely on the honesty of the authors and the thoroughness of the reviewers. The following comparisons of what the authors claim the reference reported, and what they actually state, should serve as a shocking example of how researchers or clinicians, affiliated with a company as a researcher and/or lecturer, can disguise marketing rhetoric in the form of
pseudoscience. Albrektsson, with financial ties to Nobel Biocare, has demonstrated this type of intellectual dishonesty in many of his previously published articles critical of Nobel Biocare’s competitors. He has done it again in this article:
INACCURATE REFERENCE #1: De Bruyn 1999 Article States:
"The short-term clinical survival of Screw-Vent implants (Core-Vent company Encino, USA) was first reported by De Bruyn et al. (1992). They reported 1-year failures of 11% in the maxilla with Screw-Vent versus 6% with Branemark fixtures."
FACT: De Bruyn 1992 Article States:
The characterization of the difference in maxillary success rates of the 1-year data from the De Bruyn 1992 article is misleading by not including the information that 37 of 62 (60%) Screw-Vent implants placed in the maxilla were of the 7-10mm length (7mm = 6; 10mm = 31) in contrast to only 13 of 56 (23%) Branemark implants (7mm = 0; 10mm = 13).
The De Bruyn 1992 study acknowledged more short implants as the cause of higher failures:
"We believe that the failure rate is more influenced by the implant length… the majority of all Screw-Vent implants in the upper jaw are 10mm or less...it is tempting to conclude that the large proportion of short Screw-Vent implants versus the large number of long Branemark fixtures highly influenced the absolute failure rates of the present study."
INACCURATE REFERENCE #2: De Bruyn 1999 Article States:
"Cune & De Putter (1994) have shown that tissues adjacent to Screw-Vent implants were less healthier than tissues surrounding other implant systems," and speculate that this was due to "possible aluminum leakage" from the abutments. Since this reference is false (see below), the speculation as to its cause is irrelevant (1994 JOMI Vol 5; 548-555).
FACT - Cune and DePutter Article States:
"With regard to implant survival and oral hygiene, no major differences could be observed, both among implant systems and between the suprastructure types… It was concluded that for the three outcome measures ‘implant’ survival, ‘health of the peri-implant tissues’, and ‘oral hygiene’, differences between implant systems...were small and probably not clinically relevant."
INACCURATE REFERENCE #3: De Bruyn 1999 Article States:
"Olefjord & Hansson 1993 could trace down contaminants of hydrofloric acid used in the cleaning process of Screw-Vent implants. They suggested that inorganic comtaminants should be avoided because these species can possibly provoke dissoluton of the titanium."
FACT: Olefjord and Hansson, in making the statement that "inorganic contaminants ...can possibly provoke dissolution of the titanium" were not referring to fluoride residue form acid etching. The article states:
"It is not know whether the Ca, Zn, and Si found on the surface in this study have the effect of promoting stress-corrosion cracking of the implant." Ca and Si were found on the Brånemark and Screw-Vent implants, and are common elements in water, perhaps deposited in the final rinsing process. Zn was also found on the Brånemark implants, but not on the Screw-Vent implant. Pre-1997 Screw-Vent implants were acid etched with HFl to remove loose titanium particles and other machining contaminants. Fluoride ions on the surface of an implant have been shown to enhance bone attachment.
Inaccurate Reference #4: De Bruyn 1999 Article States:
"Rosenberg et al. (1991) closely monitored 75 patients with various implant systems with respect to traumatic and infectious failure during 4 years. They reported 56% of failing Screw-Vent implants versus 18% of failing Brånemark implants."
FACT: "Rosenberg et al. (1991 COIR 2: 134-144) is not a comparative study of implant success, but a study of what kind of microflora are associated with implants that are already failing. Rosenberg does not report on the number or percentage of failed Screw-Vent or Brånemark implants in relationship to the number of implants placed, so no comparisons can be made as to success or failure of the two systems.
"75 patients with 335 osseointegrated implants were examined every 2 months for up to 4 years...A total of 11 patients exhibited failing implants and were entered into present study...Only 82 of the 335 implants in the study were evaluated for possible failures from trauma or infection."
Inaccurate Reference #5: De Bruyn 1999 Articles References Clelland et al. ( 1991 Finite Element Stress Analysis of the Screw-Vent Design):
"A stress analysis around the Screw-Vent implant, furthermore, revealed that maximum compressive stresses are concentrated within the cylindrical collar and upper one fifth of the implant body. This stress is transferred to the surrounding bone and may lead to bone loss. Further down the implant , the stress is minimized. This finding can be a possible explanation for implant fracture or ongoing bone loss, irrespective of implant length."
FACT: Clelland et al. 1991 Actually Reported:
"Based on the finite element model of the implant, the metal will not fatigue under normal occlusal forces." [No implant fractures were reported by De Bruyn 1999, so one wonders why the authors are speculating what might cause fractures.]
"Sufficient data concerning osseous healing and adaptation to dental implants on the market are unavailable. It remains uncertain how much force can be applied to a dental implant before the surrounding bone is jeopardized."
FACT: De Bruyn 1999 article’s selective use of studies to suggest that the design of the Screw-Vent is the cause of excessive bone loss, is apparent upon further review of stress analysis studies [Deines D. et al., Photoelastic stress. Int. J of Periodont Rest Dent. Vol 13: 6, 1993] reveal the following:
This study demonstrates that root form implants concentrate stress within the upper portion of the implant with minimum stresses more apically on the implant. "There appeared to be no difference in the stress distribution or intensity of the Nobelpharma and Screw-Vent designs. "
The requirement that implants not exhibit more than 0.2mm of annual crestal bone loss after the first year of placement as a criterion of success was proposed in 1986 (Albrektsson,
Zarb et al. JOMI 1986), by co-author, Dr. Tomas Albrektsson, a Nobel Biocare consultant and spokesman. It has undergone a number of modifications in published articles since that time although, to date, not a single published study on the Brånemark implant that has ever utilized the Albrektsson/Zarb criterion of crestal bone loss for determining success or failure of a single implant. For example, Adell et al.’s 15-year (Int J Oral Surg 1981) and 20-year (JOMI 1990) reports, and Zarb’s 9-year report (Int J Perio Rest Dent 1991) did not utilize bone loss as a criterion for categorizing Brånemark implants as failures. Albrektsson et al. (Int J Prosthodont 1993) admit that the 20-year report on the Brånemark implant by Adell et al. (JOMI 1990;5:347) "did not, for practical reasons, make individual radiograms, nor did they perform individual stability controls in every long-term follow-up patient.
The real issue raised by the De Bruyn, Albrektsson article is not the bias of the authors, which is obvious from their inaccurate extrapolation from various references to portray the Screw-Vent implant in a poor light. It is the use of the Albrektsson bone loss criteria to evaluate any implant system. The following history of this criteria and the illogic of its application should provide a better understanding of how it’s application serves no valid purpose.
The De Bruyn, Albrektsson study established the benchmark level of bone loss they say is acceptable for the Screw-Vent by adding 1.5mm allowance for bone loss during the first year in function, and 0.2mm for each of the 6 years of follow-up, for a maximum of 2.7mm, which they acknowledge "corresponds by co-incidence with the first implant thread." The average bone loss was reported as being 2.92mm with "18 of 60 examined implants (30%) showed unacceptable radiological bone loss beyond the critical value of 2.7mm." Therefore, bone loss to or beyond the first thread was considered "unacceptable."
In contrast, bone loss with the Brånemark implant has been measured from the base of the countersink (Adell 1981), which is 1.5mm from the top of the implant. Brånemark, Albrektsson and Zarb (1985 Quintessence Textbook, Pg 181) reported an additional "1 to 1.5mm of marginal bone was lost during the first year after bridge connection, mainly as a response to the surgical trauma. This combined surgical removal by countersinking and subsequent bone loss totals 3mm and results in exposure of the first thread of the Branemark implant on a routine basis. Countersinking the crestal bone for the Branemark implant has the effect of widening the ridge by removing thin crestal bone that would eventually resorb. The Screw-Vent implants, with necks narrower than their outside threads, do not require countersinking, thereby preserving crestal bone initially and allowing the bone to resorb naturally.
Therefore, both the Branemark and Screw-Vent implants may lose bone to the first thread (3mm from top on the Branemark implant and 2.7mm from the top with the older Screw-Vent design used by De Bruyn) but according to the Albrektsson criteria, the Screw-Vent design would be considered a failure while the Branemark implant would be considered a success. The article states that the Albrektsson criteria has been accepted by the European Academy for Periodontology. The time has come for such organizations to review the history of, and modifications to, the Albrektsson criteria and question what purpose such criteria serve, other than for clinicians and researchers affiliated with Nobel Biocare to criticize competitors' products.
1986: The Albrektsson/Zarb Criteria of Success (Albrektsson T, Zarb G et al. JOMI 1986) included a requirement that crestal bone loss be "less than 0.2mm annually after the implant's first year of service." The 0.2mm bone loss after the first year in function was
based on an average bone measurements of implants splinted cross-arch in the symphysis of totally edentulous jaws (Branemark 1977 textbook).
1989: Following the reports of the University of Toronto replication study (Smith D, Zarb G et al., J Prosthet Dent 1989;62(5):567-572), Zarb modified this criterion by adding the proviso that: "Complications of an iatrogenic nature that are not attributable to a problem with material or design should be considered separately when computing the percentage of success."
Considering the long list of "complications of an iatrogenic nature" that can affect crestal bone loss, one should legitimately question how such a criterion could ever be applied to determine clinical success. Certainly this article provides no disclosure of any such complications, nor is it always possible to determine when such complications of an iatrogenic nature were the cause of the bone loss. A partial list of iatrogenic causes of crestal bone loss would include:
Number of implants, length of span of fixed prosthesis
Passive fit of framework, connection to natural teeth
1991: The report on the results of the University of Toronto Branemark Implant Study, (Chaytor D, Zarb G.: Int J Perio Rest Dent 1991; 11(2):113-125) tried to reconcile their observations with the Albrektsson/Zarb Bone Loss Criteria by evaluating average bone loss within a single patient:
"…considerable variations exist among subjects and from year to year for individual subjects." Annual means ranged from "as low as 0.04 to a high of 0.35…. Some of this observed variability may be related to the accuracy of the radiographic and measurement techniques used---or, in other words, to measurement error…. It is also likely that, at different years after treatment, patients do have increased or decreased bone levels at particular sites….Bone level measurements varied up and down at particular sites, either for biologic reasons or measurement errors."
1993: Chaytor provides a common sense approach to the validity of applying annual bone loss as a criterion of success (Chaytor D: Int J Prosthodont 1993;6(2):95-105.):
"Revisiting the first 9 years of data from the Toronto study of edentulous patients demonstrates, by descriptive statistics, a difference between research criteria and clinical criteria for success…No implants were removed because of excessive horizontal bone loss in the absence of mobility. It appears that clinical judgment outweighed the research-based rules for success."
"A decision to remove an implant on the basis of applying the research criteria of annual bone loss to a single implant in a single year would have deprived most patients of at least 2 additional years of service. Given that only 3 of the 63 implants in this group of 15 patients ever failed within the period of the study, the patients as a group would have been deprived of many years of satisfactory service form their implants."
Even Albrektsson and Zarb acknowledged the invalidity of applying their criteria to individual implants to determine clinical success (Albrektsson T, Zarb G: International Journal of Prosthodontics 6:95 106)
"Anyone who has worked with oral implants is aware that there are several implants that are difficult to categorize…the implant that shows an unacceptable, progressive resorption 2 and 3 years after placement. This implant will then meet the failure criteria. However, there are clinical records in which a reconstructed prosthesis will result in cessation of the loss in bone height and hence the same implant may turn into a successful one."
"Another problem with bone height measurements is related to the uncertainty in the radiographic evaluation. In the individual situation, this uncertainty is substantial." (emphasis added)
The VA study measured thickness of labial plate at time of implant placement and determined correlation between bone thickness and reduced crestal resorption (in manuscript). One analysis of data from the VA study (Manz M. Radiographic assessment of peri-implant vertical bone loss: DICRG. J Oral Maxillofac Surg 1997;55(Suppl 5):62-71) confirmed that bone loss was related to implant surface (less with HA coating 1/2mm from top), quality of bone (less in dense bone) and location in the jaw (greater in thin bone in anterior maxilla). Given all the iatrogenic causes of bone loss, many caused by inexperience of the surgeon, and its relationship with surface roughness, bone quality and ridge thickness, to judge an implant as meeting success criteria based on bone loss measurements is ludicrous, especially when reported by relatively inexperienced clinicians from one center with an affiliation to a competing company. Such reports should not pass peer review, or for that matter, the smell test!
De Bruyn 1992: "With the Screw-Vent implants, there is no need for countersinking as with the Brånemark system…The bone level is directed towards the first implant thread, about 2.7mm below the implant-abutment interface... Bone remodeling up to the first fixture thread (for the Branemark implants) about 1.5mm below the abutment-fixture interface, is observed...Obviously, bone apposition on unthreaded and smooth titanium surfaces is difficult to achieve."
The Branemark Implant is inserted with its 0.7mm external hex projection level with the bone following countersinking procedure which widens the crestal opening, removing thin bone and reducing the height of the ridge in the process of widening it. The first thread would therefore be 2.2mm below the surgically established crestal height which is even greater than if measured from the original crest of bone.
The Screw-Vent, used in the 1992 De Bruyn study, had a 2mm long, narrow neck that did not require countersinking. Therefore, when bone receded to the first thread (2.7mm below the implant top), there was, in reality, no greater bone loss than with the Brånemark implant. In 1995, the Screw-Vent design changed to shorten the neck by 0.5mm and bring up the first thread an additional 0.5mm, thereby equalizing the position of the first thread with that of the Brånemark implant. The length of the neck of the Screw-Vent is necessary to accommodate the 1.5mm deep internal hex.
Inaccurate Characterization of Screw-Vent Surface: De Bruyn 1999
"The design of the Screw-Vent implant, with a highly polished titanium collar of 2.7mm, was initially believed as predominant reason for marginal bone resorption (De Bruyn 1992)."
Wennerberg demonstrated that the acid etched surface roughness of the Screw-Vent implant (mean maximum peak-to-valley height = 10.21 µm) was slightly greater than that of the Branemark (mean maximum peak-to-valley height = 9.69 µm) machined surface. Screw-Vent is not "polished."
GERALD A. NIZNICK, DMD, MSD
President, Paragon Implant Company,
Core-Vent Bio-Engineering and