- Email: [email protected]
Success of Multiple Endosseous Dental Implant Designs to Second-Stage Surgery Across Study Sites
J Oral Maxllla/ae Surg 55:76-82, 1997,Suppl5
Success of Multiple Endosseous Dental Implant Designs to Second-Stage Surgery Across Study Sites HAROLD F. MORRIS, DDS, MS,· MICHAEL C. MANZ, DDS, MPH,t AND JANET H. TAROLLI, RN, BSN:t: A multicenter clinical study of dental implants is being conducted by the Dental Implant Clinical Research Group to investigate the influence of implant design, application, and site on clinical performance and crestal bone. This article reports on the percentage of success up to implant uncovering for different implant designs and the distribution of failures across study sites. Data from 2,847 implants placed at 32 study sites were analyzed. Percentages of success up to implant uncovering were calculated for study implants overall, by implant design, by implant design within study strata, and according to individual study sites and quartile groupings of sites based on success. Comparisons were made, with chi-square and exact tests employed where appropriate. Differences were found between the different implant designs for the study overall, and between implant designs within the different study strata. Although some implant designs were found to have generally high success across study sites, some study sites designated as having surgeons with less experience tended to have higher failure levels, and one implant design failed at higher rates in a subset of study sites. The percentage and distribution of implant failures varied across study sites and by implant design. These differences appeared to be in part related to the level of experience of the surgeons. Further investigation should focus on identification of factors that contribute to higher success in implant placement with different implant designs. Highly favorable results from long-term prospective and retrospective clinical studies of various dental implant designs have been published.!" Efficacy of an implant design can be established in such studies performed under ideal conditions with a carefully selected patient population and highly experienced surgeons
who have attained proficiency with the particular design under study. Though the development of clinician proficiency with "perfected" implants has been reported to require the experience gained from inserting about 50 implants.l" this requirement may vary for different implant designs. The learning curves associated with the development of proficiency with different implant systems has not been discussed or formally considered in the evaluation of prospective studies required for the ADA Acceptance Program.?:" Effectiveness of an implant design can only be evaluated in clinical studies using surgeons of varying abilities and experience and with a broad patient base in conditions that resemble those in typical private practices in which implant treatment will be provided to the general public.') Valuable information can be gained from the effectiveness type of prospective study conducted with investigators at various points on the learning curve with an implant system. Specifically, surgeons or study centers can be sorted according to
... Codirector, Dental Clinical Research Center; Project Codirector, Dental Implant Clinical Research Group, Department of Veterans Affairs Medical Center, Dental Research, Ann Arbor, MI. t Research Fellow, Department of Veterans Affairs Medical Center, Ann Arbor, MI; Department of Periodontics. Prevention, and Geriatrics, University of Michigan School of Dentistry, Ann Arbor, MI; Program in Dental Public Health, University of Michigan School of Public Health, Ann Arbor, MI. :j: Dental Editor, Department of Veterans Affairs Medical Center. Dental Research. Ann Arbor. MI. Address correspondence and reprint requests to Dr Morris: Departmem of Veterans Affairs Medical Center, Dental Research (IS4). 2215 Fuller Rd, Ann Arbor, MI 4!lIOS.
This is a US govemment work. There are no restrictions on its use.
0278·2391197/5512-5010$0.00/0
76
77
MORRIS. MANZ. AND TAROLLI
Table 1.
Implants and Prostheses In the DICRG Study Strata
Stratum
Prosthetic Application
Number of Implants
Implant Design
Implant Material
Implant
Basket Screw Cylinder Basket Cylinder Grooved Screw Screw Grooved Cylinder Grooved
Ti6A14V Ti6Al4V HA-coated Ti6A14V HA-coated HA-coated HA-coated cpTi HA-coated HA-coated HA-coated
Core-Vent Screw-Vent Bio-Vent Core-Vent Bio-Vent Micro-Vent Screw-Vent Screw-Vent Micro-Vent Bio-Vent Micro-Vent
Lower completely edentulous (LCE)
Full denture
5 or 6
Lower posterior (LP)
Posterior bridge
2or3
Upper completely edentulous (UCE)
Bar overdenture
5 or 6
Upper posterior (UP)
Posterior bridge
2or3
Upper single tooth (UST)
Single crown
Name"
NOTE. cpTi, commercially pure titanium; HA. hydroxyapatite. * Micro-Vent, Screw-Vent. Bio-Vent, and Core-Vent and Spectra-System products (Core-Vent Corporation. Las Vegas, NV).
rates of success and then evaluated to see if past experience, operator skill, or other factors might explain the difference in success rates. This article reports on the distribution of implant failures across study sites and by implant design in an attempt to address these ideas on the relation of clinical factors to implant success. The source of data for this report was the ongoing prospective, multicenter clinical studies of the Dental Implant Clinical Research Group (DlCRG) that were started in 1991 to investigate the influence of implant design, application, and site on clinical performance and crestal bone." Because the report only spans the time from implant placement through uncovering, the trends seen in the results may not be predictive of the long-term survival of the study implants. Materials and Methods
The DlCRG study design includes five treatment categories (research strata) of approximately equal size (Table I). These are the lower completely edentulous stratum (LCE), the lower posterior partially edentulous stratum (LP), the upper completely edentulous stratum (UCE), the upper posterior partially edentulous stratum (UP), and the upper single tooth stratum (UST). Three types of implant were assigned and randomized for the LCE and UCE cases. Two types of implant were assigned and randomized for the LP and UP cases. One type of implant was used in the UST stratum cases. Thirty Department of Veterans Affairs medical centers and two universities are participating in the studies. They are administered as two study groups of approximately equal size, each using the same study design and yielding similar numbers of implants placed in the five study strata so that data may be compared. Clinicians received training in procedures related to the implant system and data collection.
Patients are from various socioeconomic levels and range in age from 20 to 80 years and older (mean, 62.9). They are predominately white male veterans (77.8% white; 15.0% African American, 4.5% Latin American, 1.4% Asian, 0.5% Native American, 0.8% missing data; 93.5% male, 6.4% female, 0.1 % missing data). Initial screening was performed to ensure that endosseous dental implant treatment was appropriate, and extensive dental and medical history data were collected. ASA health status categories represented are "healthy," "mild systemic disease," and "severe systemic disease." ASA status was previously reported to be associated with implant survival at uncovering. II The implants used are part of the Spectra-System (Core-Vent Corporation, Las Vegas, NV), which includes screw, cylinder, basket, and grooved press-fit designs, all with an internal hexagon-and-thread connection. Implant length and diameter were chosen by the treatment team. Implants were placed according to standard implant surgical protocol, and sutures were removed 5 to 7 days postoperatively. Use of preoperative antibiotics was optional, and use of an antimicrobial rinse postoperatively for 2 weeks was recommended. Fifty percent of the clinical centers were randomized to continue use of an antimicrobial rinse for the duration of the study. Standardized forms were used to collect data. Data were entered by the staff of the Data Management Center in Ann Arbor, Michigan, and analyzed by the study biostatistician at the Data Management Center. The analysis covers the period up to and including the end point of surgical uncovering, with results reported primarily as the proportion of implants surviving to the defined end point for this analysis. This report does not reflect the success of study implants after uncovering or long-term clinical results after implant loading. For this analysis, an implant was considered as a failure if it was removed because of mobility or chronic
78
pain, discomfort, or infection. "Stage I" was defined as the period between surgical placement and uncovering. "Stage II" was the time of surgical uncovering. The "time to uncovering" varied with the decision of the investigator and scheduling realities, but was treated here as a specific end point. For those implants removed before scheduled uncovering, that is, stage I failures, this time depended on the failure point. For the rest, a minimum of 4 months for the mandible and 6 months for the maxilla was specified. The database for this analysis includes 2,847 endosseous root form implants, which was the total number placed by the two study groups by May 1995 (2,910) less the implants replacing a failed implant (second surgery sites) (n = 63). The database includes 206 implants that had not yet been uncovered at the time of the analysis. Because certain implants were placed in strata other than those assigned, the database for the strata-related calculations includes only 2,817 implants. The clinical centers were ranked by implant success through stage II surgery. Then they were grouped into the upper quartile success class (eight centers), middle two quartiles success class (16 centers), and the lower quartile success class (eight centers). For each class, implant survival through uncovering was computed for the six study implant designs (hydroxyapatite [HA] cylinder, HA grooved, HA screw, Ti alloy screw, cpTi [commercially pure titanium] screw, alloy basket), which in general represent the designs and materials that are commercially available.P:" Within study strata, analyses were performed to compare implant survival for HA cylinders, Ti alloy baskets, and Ti alloy screws in the LCE stratum; HA cylinders and Ti alloy baskets in the LP stratum; HA grooved implants, cpTi screws, and HA screws in the UCE stratum; and HA cylinders and HA grooved implants in the UP stratum. Only one implant type was used (HA grooved) in the UST stratum. Chi-square tests (or exact tests when necessary) for differences in survival among implants and among implants within study strata were performed. These tests provided an indication of the differences in survival but were not intended to test differences in a statistically rigorous sense because they were not adjusted for the multiple comparisons aspect of this type of analysis. The distribution of failures among study sites was evaluated to determine if problems with specific implant designs appeared only in a subset of the study sites. Differences across study sites were then evaluated in relation to surgeon experience in the placement of dental implants before the study. The clinical centers were classified by prior experience with implants consistent with previous published findings.V' Those centers whose principal surgeon had placed 50 or more implants before participation in the study were called "more experienced" clinical centers and those whose
SUCCESS OF MULTIPLE IMPLANT DESIGNS
Implant Survival to Uncovering 100%
99.7%
80% 60% 40% 20% 0%
n=864 n=708 n=188 n=290 n=197 n=600 HA-cyl
FIGURE I.
HA-grv
HA-scr
A1loy-scr cpTi-scr Alloy-bskt
Implant survival comparison by implant type (N
=
2,847).
principal surgeon had placed fewer than 50 implants before participation were called "less experienced" clinical centers. The distribution of study sites by this prior experience designation was then evaluated in terms of the ranking of study sites by success.
Results There were 69 failed implants (N = 2,847), for a total stage II success rate of 97.6%. Failures per total placed and success rates by implant type were as follows: HA cylinder-3 of 864 (99.7%); HA grooved6 of 708 (99.2%); HA screw-5 of 188 (97.::Wn); alloy screw-8 of 290 (97.2%); cpTi screw-II of 197 (94.4%); and alloy basket-36 of 600 (94.0%). The success rates by implant type are displayed in Figure I. Statistically significant differences in survival to uncovering were found among the six implant types (P < .00 I), indicating a very low probability that the differences in success among the six implant designs were attributable to chance. RESULTS FOR IMPLANT TYPES ACROSS SUCCESS GROUPINGS
In the upper quartile success class (eight clinical centers), the six implant design survival percentages ranged from 100% to 96.9% (Fig 2), with only 2 of 733 implants failed (0.3%). In the middle quartilcs success class (16 clinical centers), implant survival percentages for the six designs ranged from 1000/r to 94.5% (Fig 3), with 30 of 1,491 implants failed (2.0%). In the lower quartile success class (eight clinical centers), implant survival percentages for the six designs ranged from 99.0% to 84.7% (Fig 4), with 37 of 6D implants failed (5.9%). Failure percentages across the quartile groupings were low and had a small range for the HA cylinder
79
MORRIS. MANZ. AND TAROLLI
Implant Survival to Uncovering 100% 80%
80%
60%
60%
40%
40%
20%
20% n=251 n=208 HA-cyl
HA-grv
n=27
n=73
HA-ser
A1loy-ser
n=32
n=142
epTI-ser A1loy-bskt
0%
n=41
n=69
HA-ser
A1loy-ser
n=206 n=125 HA-eyl
HA·grv
n=38
n=144
cpTi-ser A11oy-bskt
FIGURE 2. Implant survival comparison by implant type for the upper quartile success class (n = 733).
FIGURE 4. Implant survival comparison by implant type for the lower quartile success class (n = 623).
(0% to 1.0%) and HA grooved implants (0% to 1.6%). The failure percentages ranged higher for the HA screw (0% to 7.3%), alloy screw (0% to 7.2%), and cpTi screw (3.1% to 7.9%) implants. The alloy basket showed the widest range of failure percentages across the three success groupings (0% to 15.3%).
LP stratum, there were no failures among 113 HA cylinder and 72 alloy basket implants. In the UST stratum (not shown), there were no failures among the 63 HA grooved implants.
COMPARISON OF IMPLANTS WITHIN STUDY STRATA BY SUCCESS GROUPINGS
Top 25% In the upper quartile success class (Fig 5), only two implant failures occurred. Within the UCE stratum, I of 32 cpTi screws failed, and there were no HA grooved (n = 59) or HA screw (n = 27) failures. In the UP stratum, I of 53 HA cylinders failed, with no failures among 83 HA grooved implants. In the LCE stratum, there were no failures among 85 HA cylinder, 73 alloy screw, and 70 alloy basket implants. In the
Implant Survival to Uncovering 100% 100%
Middle 50% In the middle quartiles success class (Fig 6), there were 30 implant failures: 14 alloy basket, seven cpTi screw, four HA grooved, three alloy screw, and two HA screw implants. In the UCE stratum, 7 of 123 cpTi screws failed (5.7%), compared with 2 of 131 HA grooved (1.5%) and 2 of 118 HA screw (1.7%) implants. In the UP stratum, there were no failures among 74 HA cylinder and 110 HA grooved implants. In the
~cm n..tlg
HA-g'OO'f'ed
HA·1CnW
n-32
60"
-
..% 20"
-
0"
...3
cpTl-ecrew
HA1JfOOYM
r~_~_L] :rDa
80% 60% 40%
n' 85 HA..cyllnd.r
20% 0%
n~27
UP --
lmr*"tSU'''' 100"
n=407 n=375 n=120 n=148 n=127 n=314 HA-eyl
HA-grv
HA-ser
Alloy-ser
epTi-ser A1loy-bskt
FIGURE 3. Implant survival comparison by implant type for the middle quartiles success class (n = 1,491).
f'\~73
AlloV«'"
n"-70
Alloy-b.....
n~ 11 J Hlw:ybnd"
n.-72 A1toyftlk..
FIGURE 5. Implant survival comparisons within study strata for the upper quartile success class (n = 730). UCE. upper completely edentulous stratum (n = 118); UP. upper posterior partially edentulous stratum (n = 136); LCE. lower completely edentulous stratum (n = 228); LP. lower posterior partially edentulous stratum (n = 185). UST stratum (not shown) (upper single tooth stratum) (n = 63 HA grooved implants; IOOth- survival).
80
SUCCESS OF MULTIPLE IMPLANT DESIGNS
Imp"nl Survtv••
UP
ImplanlSurvt.t.l
100%
100%
S"'.I'
Wnplant 100"
'0,," :>o'~
0%
,,"U1
",,111
HA-oroaved
HA·.arew
....,23
n"1.
C1IT~ ...ew
HA-
Implant 1"'VIYII 100"'-
..0 HA11 roovl d
nzl10
HA-gFOOVM
rn I I
I
".34
HA·lCfaw
•
Implanl SUrvNII 100%
I
..36 cpT"1Ia1W
".33
HA-cylnd.r Impl.nt Survl ..1
Implant SUrvival 100%
100%
80% 60% 40"A,
(,~
'0%
'0% n"bl HA-cyhnd.,
n'14B
Alloy-tc,"
n"Hi"
Alloy-b..kll
-. LP
_.--
'0%
;,")11.
'''0 HA-grooved
0 2
116
HA-cyhndlr
n=160
0%
-
n~6S
HA-cyllndlr
Alloy-bllklt
""'62
AIIOy..-c,"
0-58
n=108
HA-cyllnd.,
A1loy·tlull.ll
-
.
n=85 AlIoy-b . . . .t
FIGURE 6. Implant survival comparisons within study strata for the middle quurtilcs success class (n = I,4X3). UCE. upper cornplctcly edentulous stratum (n = 372); UP, upper posterior partiully edentulous stratum (n = I X4); LCE. lower completely edentulous stratum (n ~ 459); L/'. lower posterior partially edentulous stratum (n '" .H6). UST stratum (not shown) (upper single tooth stratum) (n = U2 IIA grooved implants; 9X.5% survival).
FIGURE 7. Implant survival comparisons within study strata for the lower quartile success class (n = 6(4). UCE, upper completely edentulous stratum (n = 113); UP. upper posterior partially edentulous stratum (n = 73); LCE. lower completely edentulous stratum (n = 185); LP. lower posterior partially edentulous stratum (n ~ 193). UST stratum (not shown) (upper single tooth stratum) (n 40 HA grooved implants; 97.5% survival).
LCE stratum, there were no failures among 157 HA cylinders, but 7 of 154 alloy baskets failed (4.5%) and 3 of 14H alloy screws failed (2.0%). In the LP stratum, 7 of 160 alloy baskets failed (4.4%) compared with no failures among 176 HA cylinders. In the UST stratum (not shown), 2 of 132 HA grooved implants (1.5%) failed.
n == 233), HA screw (97.2%, n == 179). and cpTi screw (94.2%, n == 1(1) implants (P < .(14). In the lower completely edentulous stratum (LCE), survival percentages differed statistically among HA cylinder (99.7(f(, n == 307), Ti alloy screw (97.2%, n == 2H3), and Ti alloy basket (95.4%, n == 282) implants (P < .rXl4). In the lower posterior partially edentulous stratum (LP), survival percentages for HA cylinder (99.7%, n == 397) and Ti alloy basket (92.7%, n == 317) implants differed (P < .00I). In the upper posterior partially edentulous stratum
In the lower quartile success class (Fig 7), there were 37 implant failures: 22 alloy basket, five alloy screw. three cp'Ti screw, three HA screw, two HA grooved, and two HA cylinder implants. In the UCE stratum, no HA grooved implants failed (n == 43), but 3 of 34 HA screw (8.H%) and 3 of 36 cp'Ti screw (8.3%) implants failed. In the UP stratum, I of 40 HA grooved implants failed (2.5%), compared with no failures among 33 HA cylinder implants. In the LCE stratum, failures included 6 of 58 alloy basket (10.3%), 5 of 62 alloy screw (H.I%), and I of 65 HA cylinder (1.5%) implants. In the LP stratum, 16 of H5 alloy baskets failed (I H.H%), compared with I of IOH HA cylinders (0.9%). In the UST stratum (not shown), I of 40 HA grooved implants failed (2.5%).
(99.1l}lo•
Impllnl Sur., .....'
Implant BIIIII.....1
100%
100%
UP
40% ",,"1%
0%
n-233 HA-grooved
n-119
HA·ecr. .
Imp lint BunriVIII
Implanl Survl,," 100%
100%
LP
80% 60%
""
40%
20"
OVlmALL STUDY STRATA RESULTS
Statistically significant results were found for three of the four study strata in which comparisons were possible (N == 2,H47) (Fig H), In the upper completely edentulous stratum (UCE), a statistically significant difference was found among survival percentages for HA grooved
n;;183 Alto~·lICr.w
n=397 HA-
n':111
A1loy-b ••llel
FIGURE X. Implant survival comparison by study strata (N c. 2.XI7). Implants that were placed in strata not specified by the study design were not included in this analysis. UCE. upper completely edentulous stratum; UP. upper posterior partially edentulous stratum; LCE. lower completely edentulous stratum; LP, lower posterior partially edentulous stratum.
81
MORRIS, MANZ, AND TAROLLI Table 2. Prior Experience With Implants of Clinical Center Success Groups
Clinical Centers Ranked by Implant Success
Number of More Experienced Clinical Centers
Number of Less Experienced Clinical Centers
Top 25% Middle 50% Bottom 25%
4 II I
5
4
7
(UP), only one of the HA cylinder (n = 160) and one of the HA grooved (n = 233) implants failed (survival percentages of 99.4% and 99.6%, respectively). In the upper single tooth stratum (UST) (not shown), in which only HA grooved implants were used, there was a 98.7% survival percentage, with only three failed implant'> (n = 235). STUDY SITE RESULTS Implant survival to uncovering for all implant designs combined differed among the 32 study sites, ranging from 100% to 89.8% (mean, 97.6%; N = 2,847 implants). For each of the implant designs, at least some of the study sites experienced 100% success rates. The lowest success among study sites was 96.2% (25 of 26) for HA cylinders; 93.3% (14 of 15) for HA grooved implants; 75% (three of four) for HA screws; 84.6% (II of 13) for alloy screws; 80% (four of five, two stations; 8 of 10, one station) for cpTi screws; and 75% (12 of 16) for alloy baskets. The results for the various implant designs by study site further showed that no site had more than two failures of any specific implant design except for the alloy basket implants. Five sites had three to five failures with this particular design, whereas 13 sites had no alloy basket failures and another 10 sites had only one alloy basket failure. The five sites with more than two alloy basket failures, while placing only 18% of the total alloy baskets in the study, accounted for half (18 of 36) of the total alloy basket failures. COMPARISON BY PRIOR EXPERIENCE WITH IMPLANT PLACEMENT Sixteen principal surgeons had placed 50 or more implants (more experienced), and 16 had placed fewer than 50 implants (less experienced) before participation in the study. The distribution of surgeons by prior experience among the quartile success groupings is displayed in Table 2. Although the upper quartile was evenly divided between more and less experienced surgeons, a definite difference appeared in the middle and lower groupings. More experienced surgeons exceeded less experienced surgeons in the middle grouping, and
all but one of the surgeons in the lower quartile grouping was classified as having less experience. The percentage of implant failure also differed by surgeon experience. The failure percentage was nearly twice as high (3.22% vs 1.69%) for surgeons having less experience in the placement of dental implants before the DICRG study (44 failures of 1,367 placed) than for the more experienced surgeons (25 failures/ 1,480 placed). Furthermore, the lowest six study sites in percentage success were sites with surgeons classified as having less experience before the DICRG study. They experienced 48% (33 of 69) of the total study failures at uncovering, while placing only 18% (525 of 2,847) of the total study implants. For all implant designs, higher percentages of failure were seen with the less experienced surgeons. Percentages of failure for the different implant designs by prior experience were HA cylinder-more experienced 0% (0 of 401), less experienced 0.65% (3 of 463); HA grooved-more experienced 0.81% (3 of 370), less experienced 0.89% (3 of 338); HA screw-more experienced 0.81 % (I of 123), less experienced 6.15% (4 of 65); cpTi screw-more experienced 4.92% (6 of 122), less experienced 6.67% (5 of 75); alloy screwmore experienced 1.37% (2 of 146), less experienced 4.17% (6 of 144); alloy basket-more experienced 4.09% (13 of 318), less experienced 8.16% (23 of 282).
Discussion Initial studies conducted on dental implant systems usually tend to focus on an evaluation of efficacy (performance under ideal conditions with carefully selected patients and with skilled and experienced surgeons). The DICRG studies recruited study sites without specific inclusion criteria on the amount of experience with the placement of dental implants in general, experience with specific implant systems, or the level of skill of the implant treatment team as indicated by past success rates with dental implant treatment. Additionally, exclusion criteria for study patients were well defined but were not strict as to age, medical status, or available bone. Therefore, it could be argued that the DICRG studies are more representative of a study of effectiveness (performance under normal conditions with a variety of dentists and patients) than efficacy. By viewing study data on success rates related to factors such as experience, adherence to study protocol, and patient selection, an approximate evaluation of efficacy could be conducted. This would involve a restricted analysis of data from study sites in which ideal study conditions were more closely approximated, surgeons had higher levels of experience and skill, and patients were more carefully selected to maximize the probability of implant treatment success. Alternatively, evaluating results that include sites with varied study conditions, treatment providers, and
82 patients, with the potential of associated higher rates of implant failure, can contribute valuable information. Such an investigation may better indicate the effectiveness of implant treatment and be more generalizable to private practices with dentists of varying skills and experience, and patients with varying medical and oral conditions. Problems might be observed with implant treatment in general, or with specific implant designs, that might not be apparent in tightly controlled efficacy studies, which maximize the conditions for success. These ideas seem to be borne out by this analysis of data from the DlCRG database. By ranking and grouping DICRG study sites into quartiles based on the percentage of success with study implants to stage 11 uncovering, lower percentages of failure were seen for many sites, whereas higher percentages are seen for a few sites. The upper quartile of study sites only had two failed implants overall (0.3%), but failures rose to 30 (2.0%) in the middle two quartiles and 37 (5.9%) in the lower quartile. Failure rates by study site ranged from 0% to 10.2%, with a mean of 2.4%. The median failure percentage of 1.8% indicates that the mean failure percentage is being increased by a small number of study sites with higher numbers of failures. The results show generally high success with the HA cylinder and the HA grooved implant designs. The three varieties of screw implant also had generally high success percentages, although they dropped below 93% in the lower quartile of study sites. Results for the alloy basket showed that the percentage of success dropped from 100% in the upper quartile to 84.7% in the lower quartile of study sites. Differences in success of the implant designs used within study strata also were noted. These differences will continue to be monitored to see if they continue after implant loading. Such long-term results could be used in the treatment-planning process to determine implant designs with the highest probability of success for specific applications, and to guide the development of future implant designs. The results for the various implant designs by study site show that no site had more than two failures with any specific implant design except for the alloy basket. Although 13 sites had no alloy basket failures and another 10 sites had only one alloy basket failure, five sites had three to five failures with this particular design. Although failures seemed to be low and evenly distributed among study sites for some implant designs, failures with specific implant designs tended to cluster in a subset of study sites, as seen with the alloy basket design. Clustering of failures may indicate that a greater level of experience and skill is necessary for high success with certain implant designs, possibly related to the complexity of the implant procedure with these designs, or it may indicate a difference in patients treated. The results after classification of the sites by surgeon's experience with the place-
SUCCESS OF MULTIPLE IMPLANT DESIGNS
ment of dental implants before the study support the plausibility of these relationships. All but one of the study sites in the lower quartile success group were classified as having a surgeon with less experience before the study. Overall, the percentage of failure for the less experienced surgeons was nearly twice that seen with the more experienced surgeons. Low failure percentages were seen for both experience levels with the HA cylinder and the HA grooved implants. Percentages of failure increased with the three screw designs and with the alloy basket implants, and were higher with the less experienced surgeons than with the more experienced surgeons. The screw and alloy basket designs have more complex procedures associated with their placement. Some minimum level of experience and proficiency would appear to have importance in the successful placement of these designs. The results of this investigation show that high success rates can generally be achieved with some implant designs, whereas lower success rates might be experienced with some designs, in a subset of dentists, indicating sensitivity of certain designs to operator, patient, or other clinical conditions. The relationships of these clinical factors to success with different implant designs require further investigation.
References I. Spiekermann H, Jansen VK. Richter E-J: A IO-year follow-up
study of IMZ and TPS implants in the edentulous mandible using bar-retained overdentures. Int J Oral Maxillofac Implants 10:23I, 1995 2. Stultz ER, Lofland R, Sendax VI, et al: A multicenter 5-year retrospective survival analysis of 6,200 Integraloo implants. Com pend Contin Educ Dent 14:478, 1993 3. Lekholm U, van Steenberghe 0, Herrmann I, et al: Osseointegrated implants in the treatment of partially edentulous jaws: A prospective 5-year multicenter study. Int J Oral Maxillofac Implants 9:627, 1994 4. Adell R, Eriksson B, Lekholm U, et al: A long-term follow-up study of osseointegrated implants in the treatment of totally edentulous jaws. Int J Oral Maxillofac Implants 5:347. 1990 5. Lambert P, Morris HF, Ochi S, et al: Relationship between implant surgical experience and second-stage failures: DICRG interim report no. 2. Implant Dent 3:97, 1994 6. Lambert PM, Morris HF, Ochi S, et al: Positive effect of surgical experience with implants on second-stage implant survival. J Oral MaxiIlofac Surg 55:12, [997 (suppl 5) 7. Stanford JW: Acceptance program for endosseous implants: A service benefit of ADA membership. Int J Oral Maxillofac Implants 6:15, [991 8. Donovan TE, Chee WWL: ADA acceptance program for endosseous implants. J Calif Dent Assoc 20:60, 1992 9. Last JM (ed): A Dictionary of Epidemiology. New York. NY. Oxford University Press, 1983, pp 31-32 10. Monis HF. Oehi S, Dental Implant Clinical Research Group (planning Committee): The influence of implant design. application. and site on clinical performance and crestal bone: A multicenter. multidisciplinary clinical study. Implant Dent [:49. 1992 I I. Oehi S, Monis HF. Winkler S, et al: Patient demographics and implant survival at uncovering:Dental Implant Clinical Research Group interim report no. 6. Implant Dent 3:247, [994 12. Ismail JYH: A comparison of current root form implants: Biomechanical design and prosthodontic applications. NY State Dent J 55:34, 1989 13. English C: An overview of implant hardware. J Am Dent Assoc 121:360, 1990
Success of Multiple Endosseous Dental Implant Designs to Second-Stage Surgery Across Study Sites HAROLD F. MORRIS, DDS, MS,· MICHAEL C. MANZ, DDS, MPH,t AND JANET H. TAROLLI, RN, BSN:t: A multicenter clinical study of dental implants is being conducted by the Dental Implant Clinical Research Group to investigate the influence of implant design, application, and site on clinical performance and crestal bone. This article reports on the percentage of success up to implant uncovering for different implant designs and the distribution of failures across study sites. Data from 2,847 implants placed at 32 study sites were analyzed. Percentages of success up to implant uncovering were calculated for study implants overall, by implant design, by implant design within study strata, and according to individual study sites and quartile groupings of sites based on success. Comparisons were made, with chi-square and exact tests employed where appropriate. Differences were found between the different implant designs for the study overall, and between implant designs within the different study strata. Although some implant designs were found to have generally high success across study sites, some study sites designated as having surgeons with less experience tended to have higher failure levels, and one implant design failed at higher rates in a subset of study sites. The percentage and distribution of implant failures varied across study sites and by implant design. These differences appeared to be in part related to the level of experience of the surgeons. Further investigation should focus on identification of factors that contribute to higher success in implant placement with different implant designs. Highly favorable results from long-term prospective and retrospective clinical studies of various dental implant designs have been published.!" Efficacy of an implant design can be established in such studies performed under ideal conditions with a carefully selected patient population and highly experienced surgeons
who have attained proficiency with the particular design under study. Though the development of clinician proficiency with "perfected" implants has been reported to require the experience gained from inserting about 50 implants.l" this requirement may vary for different implant designs. The learning curves associated with the development of proficiency with different implant systems has not been discussed or formally considered in the evaluation of prospective studies required for the ADA Acceptance Program.?:" Effectiveness of an implant design can only be evaluated in clinical studies using surgeons of varying abilities and experience and with a broad patient base in conditions that resemble those in typical private practices in which implant treatment will be provided to the general public.') Valuable information can be gained from the effectiveness type of prospective study conducted with investigators at various points on the learning curve with an implant system. Specifically, surgeons or study centers can be sorted according to
... Codirector, Dental Clinical Research Center; Project Codirector, Dental Implant Clinical Research Group, Department of Veterans Affairs Medical Center, Dental Research, Ann Arbor, MI. t Research Fellow, Department of Veterans Affairs Medical Center, Ann Arbor, MI; Department of Periodontics. Prevention, and Geriatrics, University of Michigan School of Dentistry, Ann Arbor, MI; Program in Dental Public Health, University of Michigan School of Public Health, Ann Arbor, MI. :j: Dental Editor, Department of Veterans Affairs Medical Center. Dental Research. Ann Arbor. MI. Address correspondence and reprint requests to Dr Morris: Departmem of Veterans Affairs Medical Center, Dental Research (IS4). 2215 Fuller Rd, Ann Arbor, MI 4!lIOS.
This is a US govemment work. There are no restrictions on its use.
0278·2391197/5512-5010$0.00/0
76
77
MORRIS. MANZ. AND TAROLLI
Table 1.
Implants and Prostheses In the DICRG Study Strata
Stratum
Prosthetic Application
Number of Implants
Implant Design
Implant Material
Implant
Basket Screw Cylinder Basket Cylinder Grooved Screw Screw Grooved Cylinder Grooved
Ti6A14V Ti6Al4V HA-coated Ti6A14V HA-coated HA-coated HA-coated cpTi HA-coated HA-coated HA-coated
Core-Vent Screw-Vent Bio-Vent Core-Vent Bio-Vent Micro-Vent Screw-Vent Screw-Vent Micro-Vent Bio-Vent Micro-Vent
Lower completely edentulous (LCE)
Full denture
5 or 6
Lower posterior (LP)
Posterior bridge
2or3
Upper completely edentulous (UCE)
Bar overdenture
5 or 6
Upper posterior (UP)
Posterior bridge
2or3
Upper single tooth (UST)
Single crown
Name"
NOTE. cpTi, commercially pure titanium; HA. hydroxyapatite. * Micro-Vent, Screw-Vent. Bio-Vent, and Core-Vent and Spectra-System products (Core-Vent Corporation. Las Vegas, NV).
rates of success and then evaluated to see if past experience, operator skill, or other factors might explain the difference in success rates. This article reports on the distribution of implant failures across study sites and by implant design in an attempt to address these ideas on the relation of clinical factors to implant success. The source of data for this report was the ongoing prospective, multicenter clinical studies of the Dental Implant Clinical Research Group (DlCRG) that were started in 1991 to investigate the influence of implant design, application, and site on clinical performance and crestal bone." Because the report only spans the time from implant placement through uncovering, the trends seen in the results may not be predictive of the long-term survival of the study implants. Materials and Methods
The DlCRG study design includes five treatment categories (research strata) of approximately equal size (Table I). These are the lower completely edentulous stratum (LCE), the lower posterior partially edentulous stratum (LP), the upper completely edentulous stratum (UCE), the upper posterior partially edentulous stratum (UP), and the upper single tooth stratum (UST). Three types of implant were assigned and randomized for the LCE and UCE cases. Two types of implant were assigned and randomized for the LP and UP cases. One type of implant was used in the UST stratum cases. Thirty Department of Veterans Affairs medical centers and two universities are participating in the studies. They are administered as two study groups of approximately equal size, each using the same study design and yielding similar numbers of implants placed in the five study strata so that data may be compared. Clinicians received training in procedures related to the implant system and data collection.
Patients are from various socioeconomic levels and range in age from 20 to 80 years and older (mean, 62.9). They are predominately white male veterans (77.8% white; 15.0% African American, 4.5% Latin American, 1.4% Asian, 0.5% Native American, 0.8% missing data; 93.5% male, 6.4% female, 0.1 % missing data). Initial screening was performed to ensure that endosseous dental implant treatment was appropriate, and extensive dental and medical history data were collected. ASA health status categories represented are "healthy," "mild systemic disease," and "severe systemic disease." ASA status was previously reported to be associated with implant survival at uncovering. II The implants used are part of the Spectra-System (Core-Vent Corporation, Las Vegas, NV), which includes screw, cylinder, basket, and grooved press-fit designs, all with an internal hexagon-and-thread connection. Implant length and diameter were chosen by the treatment team. Implants were placed according to standard implant surgical protocol, and sutures were removed 5 to 7 days postoperatively. Use of preoperative antibiotics was optional, and use of an antimicrobial rinse postoperatively for 2 weeks was recommended. Fifty percent of the clinical centers were randomized to continue use of an antimicrobial rinse for the duration of the study. Standardized forms were used to collect data. Data were entered by the staff of the Data Management Center in Ann Arbor, Michigan, and analyzed by the study biostatistician at the Data Management Center. The analysis covers the period up to and including the end point of surgical uncovering, with results reported primarily as the proportion of implants surviving to the defined end point for this analysis. This report does not reflect the success of study implants after uncovering or long-term clinical results after implant loading. For this analysis, an implant was considered as a failure if it was removed because of mobility or chronic
78
pain, discomfort, or infection. "Stage I" was defined as the period between surgical placement and uncovering. "Stage II" was the time of surgical uncovering. The "time to uncovering" varied with the decision of the investigator and scheduling realities, but was treated here as a specific end point. For those implants removed before scheduled uncovering, that is, stage I failures, this time depended on the failure point. For the rest, a minimum of 4 months for the mandible and 6 months for the maxilla was specified. The database for this analysis includes 2,847 endosseous root form implants, which was the total number placed by the two study groups by May 1995 (2,910) less the implants replacing a failed implant (second surgery sites) (n = 63). The database includes 206 implants that had not yet been uncovered at the time of the analysis. Because certain implants were placed in strata other than those assigned, the database for the strata-related calculations includes only 2,817 implants. The clinical centers were ranked by implant success through stage II surgery. Then they were grouped into the upper quartile success class (eight centers), middle two quartiles success class (16 centers), and the lower quartile success class (eight centers). For each class, implant survival through uncovering was computed for the six study implant designs (hydroxyapatite [HA] cylinder, HA grooved, HA screw, Ti alloy screw, cpTi [commercially pure titanium] screw, alloy basket), which in general represent the designs and materials that are commercially available.P:" Within study strata, analyses were performed to compare implant survival for HA cylinders, Ti alloy baskets, and Ti alloy screws in the LCE stratum; HA cylinders and Ti alloy baskets in the LP stratum; HA grooved implants, cpTi screws, and HA screws in the UCE stratum; and HA cylinders and HA grooved implants in the UP stratum. Only one implant type was used (HA grooved) in the UST stratum. Chi-square tests (or exact tests when necessary) for differences in survival among implants and among implants within study strata were performed. These tests provided an indication of the differences in survival but were not intended to test differences in a statistically rigorous sense because they were not adjusted for the multiple comparisons aspect of this type of analysis. The distribution of failures among study sites was evaluated to determine if problems with specific implant designs appeared only in a subset of the study sites. Differences across study sites were then evaluated in relation to surgeon experience in the placement of dental implants before the study. The clinical centers were classified by prior experience with implants consistent with previous published findings.V' Those centers whose principal surgeon had placed 50 or more implants before participation in the study were called "more experienced" clinical centers and those whose
SUCCESS OF MULTIPLE IMPLANT DESIGNS
Implant Survival to Uncovering 100%
99.7%
80% 60% 40% 20% 0%
n=864 n=708 n=188 n=290 n=197 n=600 HA-cyl
FIGURE I.
HA-grv
HA-scr
A1loy-scr cpTi-scr Alloy-bskt
Implant survival comparison by implant type (N
=
2,847).
principal surgeon had placed fewer than 50 implants before participation were called "less experienced" clinical centers. The distribution of study sites by this prior experience designation was then evaluated in terms of the ranking of study sites by success.
Results There were 69 failed implants (N = 2,847), for a total stage II success rate of 97.6%. Failures per total placed and success rates by implant type were as follows: HA cylinder-3 of 864 (99.7%); HA grooved6 of 708 (99.2%); HA screw-5 of 188 (97.::Wn); alloy screw-8 of 290 (97.2%); cpTi screw-II of 197 (94.4%); and alloy basket-36 of 600 (94.0%). The success rates by implant type are displayed in Figure I. Statistically significant differences in survival to uncovering were found among the six implant types (P < .00 I), indicating a very low probability that the differences in success among the six implant designs were attributable to chance. RESULTS FOR IMPLANT TYPES ACROSS SUCCESS GROUPINGS
In the upper quartile success class (eight clinical centers), the six implant design survival percentages ranged from 100% to 96.9% (Fig 2), with only 2 of 733 implants failed (0.3%). In the middle quartilcs success class (16 clinical centers), implant survival percentages for the six designs ranged from 1000/r to 94.5% (Fig 3), with 30 of 1,491 implants failed (2.0%). In the lower quartile success class (eight clinical centers), implant survival percentages for the six designs ranged from 99.0% to 84.7% (Fig 4), with 37 of 6D implants failed (5.9%). Failure percentages across the quartile groupings were low and had a small range for the HA cylinder
79
MORRIS. MANZ. AND TAROLLI
Implant Survival to Uncovering 100% 80%
80%
60%
60%
40%
40%
20%
20% n=251 n=208 HA-cyl
HA-grv
n=27
n=73
HA-ser
A1loy-ser
n=32
n=142
epTI-ser A1loy-bskt
0%
n=41
n=69
HA-ser
A1loy-ser
n=206 n=125 HA-eyl
HA·grv
n=38
n=144
cpTi-ser A11oy-bskt
FIGURE 2. Implant survival comparison by implant type for the upper quartile success class (n = 733).
FIGURE 4. Implant survival comparison by implant type for the lower quartile success class (n = 623).
(0% to 1.0%) and HA grooved implants (0% to 1.6%). The failure percentages ranged higher for the HA screw (0% to 7.3%), alloy screw (0% to 7.2%), and cpTi screw (3.1% to 7.9%) implants. The alloy basket showed the widest range of failure percentages across the three success groupings (0% to 15.3%).
LP stratum, there were no failures among 113 HA cylinder and 72 alloy basket implants. In the UST stratum (not shown), there were no failures among the 63 HA grooved implants.
COMPARISON OF IMPLANTS WITHIN STUDY STRATA BY SUCCESS GROUPINGS
Top 25% In the upper quartile success class (Fig 5), only two implant failures occurred. Within the UCE stratum, I of 32 cpTi screws failed, and there were no HA grooved (n = 59) or HA screw (n = 27) failures. In the UP stratum, I of 53 HA cylinders failed, with no failures among 83 HA grooved implants. In the LCE stratum, there were no failures among 85 HA cylinder, 73 alloy screw, and 70 alloy basket implants. In the
Implant Survival to Uncovering 100% 100%
Middle 50% In the middle quartiles success class (Fig 6), there were 30 implant failures: 14 alloy basket, seven cpTi screw, four HA grooved, three alloy screw, and two HA screw implants. In the UCE stratum, 7 of 123 cpTi screws failed (5.7%), compared with 2 of 131 HA grooved (1.5%) and 2 of 118 HA screw (1.7%) implants. In the UP stratum, there were no failures among 74 HA cylinder and 110 HA grooved implants. In the
~cm n..tlg
HA-g'OO'f'ed
HA·1CnW
n-32
60"
-
..% 20"
-
0"
...3
cpTl-ecrew
HA1JfOOYM
r~_~_L] :rDa
80% 60% 40%
n' 85 HA..cyllnd.r
20% 0%
n~27
UP --
lmr*"tSU'''' 100"
n=407 n=375 n=120 n=148 n=127 n=314 HA-eyl
HA-grv
HA-ser
Alloy-ser
epTi-ser A1loy-bskt
FIGURE 3. Implant survival comparison by implant type for the middle quartiles success class (n = 1,491).
f'\~73
AlloV«'"
n"-70
Alloy-b.....
n~ 11 J Hlw:ybnd"
n.-72 A1toyftlk..
FIGURE 5. Implant survival comparisons within study strata for the upper quartile success class (n = 730). UCE. upper completely edentulous stratum (n = 118); UP. upper posterior partially edentulous stratum (n = 136); LCE. lower completely edentulous stratum (n = 228); LP. lower posterior partially edentulous stratum (n = 185). UST stratum (not shown) (upper single tooth stratum) (n = 63 HA grooved implants; IOOth- survival).
80
SUCCESS OF MULTIPLE IMPLANT DESIGNS
Imp"nl Survtv••
UP
ImplanlSurvt.t.l
100%
100%
S"'.I'
Wnplant 100"
'0,," :>o'~
0%
,,"U1
",,111
HA-oroaved
HA·.arew
....,23
n"1.
C1IT~ ...ew
HA-
Implant 1"'VIYII 100"'-
..0 HA11 roovl d
nzl10
HA-gFOOVM
rn I I
I
".34
HA·lCfaw
•
Implanl SUrvNII 100%
I
..36 cpT"1Ia1W
".33
HA-cylnd.r Impl.nt Survl ..1
Implant SUrvival 100%
100%
80% 60% 40"A,
(,~
'0%
'0% n"bl HA-cyhnd.,
n'14B
Alloy-tc,"
n"Hi"
Alloy-b..kll
-. LP
_.--
'0%
;,")11.
'''0 HA-grooved
0 2
116
HA-cyhndlr
n=160
0%
-
n~6S
HA-cyllndlr
Alloy-bllklt
""'62
AIIOy..-c,"
0-58
n=108
HA-cyllnd.,
A1loy·tlull.ll
-
.
n=85 AlIoy-b . . . .t
FIGURE 6. Implant survival comparisons within study strata for the middle quurtilcs success class (n = I,4X3). UCE. upper cornplctcly edentulous stratum (n = 372); UP, upper posterior partiully edentulous stratum (n = I X4); LCE. lower completely edentulous stratum (n ~ 459); L/'. lower posterior partially edentulous stratum (n '" .H6). UST stratum (not shown) (upper single tooth stratum) (n = U2 IIA grooved implants; 9X.5% survival).
FIGURE 7. Implant survival comparisons within study strata for the lower quartile success class (n = 6(4). UCE, upper completely edentulous stratum (n = 113); UP. upper posterior partially edentulous stratum (n = 73); LCE. lower completely edentulous stratum (n = 185); LP. lower posterior partially edentulous stratum (n ~ 193). UST stratum (not shown) (upper single tooth stratum) (n 40 HA grooved implants; 97.5% survival).
LCE stratum, there were no failures among 157 HA cylinders, but 7 of 154 alloy baskets failed (4.5%) and 3 of 14H alloy screws failed (2.0%). In the LP stratum, 7 of 160 alloy baskets failed (4.4%) compared with no failures among 176 HA cylinders. In the UST stratum (not shown), 2 of 132 HA grooved implants (1.5%) failed.
n == 233), HA screw (97.2%, n == 179). and cpTi screw (94.2%, n == 1(1) implants (P < .(14). In the lower completely edentulous stratum (LCE), survival percentages differed statistically among HA cylinder (99.7(f(, n == 307), Ti alloy screw (97.2%, n == 2H3), and Ti alloy basket (95.4%, n == 282) implants (P < .rXl4). In the lower posterior partially edentulous stratum (LP), survival percentages for HA cylinder (99.7%, n == 397) and Ti alloy basket (92.7%, n == 317) implants differed (P < .00I). In the upper posterior partially edentulous stratum
In the lower quartile success class (Fig 7), there were 37 implant failures: 22 alloy basket, five alloy screw. three cp'Ti screw, three HA screw, two HA grooved, and two HA cylinder implants. In the UCE stratum, no HA grooved implants failed (n == 43), but 3 of 34 HA screw (8.H%) and 3 of 36 cp'Ti screw (8.3%) implants failed. In the UP stratum, I of 40 HA grooved implants failed (2.5%), compared with no failures among 33 HA cylinder implants. In the LCE stratum, failures included 6 of 58 alloy basket (10.3%), 5 of 62 alloy screw (H.I%), and I of 65 HA cylinder (1.5%) implants. In the LP stratum, 16 of H5 alloy baskets failed (I H.H%), compared with I of IOH HA cylinders (0.9%). In the UST stratum (not shown), I of 40 HA grooved implants failed (2.5%).
(99.1l}lo•
Impllnl Sur., .....'
Implant BIIIII.....1
100%
100%
UP
40% ",,"1%
0%
n-233 HA-grooved
n-119
HA·ecr. .
Imp lint BunriVIII
Implanl Survl,," 100%
100%
LP
80% 60%
""
40%
20"
OVlmALL STUDY STRATA RESULTS
Statistically significant results were found for three of the four study strata in which comparisons were possible (N == 2,H47) (Fig H), In the upper completely edentulous stratum (UCE), a statistically significant difference was found among survival percentages for HA grooved
n;;183 Alto~·lICr.w
n=397 HA-
n':111
A1loy-b ••llel
FIGURE X. Implant survival comparison by study strata (N c. 2.XI7). Implants that were placed in strata not specified by the study design were not included in this analysis. UCE. upper completely edentulous stratum; UP. upper posterior partially edentulous stratum; LCE. lower completely edentulous stratum; LP, lower posterior partially edentulous stratum.
81
MORRIS, MANZ, AND TAROLLI Table 2. Prior Experience With Implants of Clinical Center Success Groups
Clinical Centers Ranked by Implant Success
Number of More Experienced Clinical Centers
Number of Less Experienced Clinical Centers
Top 25% Middle 50% Bottom 25%
4 II I
5
4
7
(UP), only one of the HA cylinder (n = 160) and one of the HA grooved (n = 233) implants failed (survival percentages of 99.4% and 99.6%, respectively). In the upper single tooth stratum (UST) (not shown), in which only HA grooved implants were used, there was a 98.7% survival percentage, with only three failed implant'> (n = 235). STUDY SITE RESULTS Implant survival to uncovering for all implant designs combined differed among the 32 study sites, ranging from 100% to 89.8% (mean, 97.6%; N = 2,847 implants). For each of the implant designs, at least some of the study sites experienced 100% success rates. The lowest success among study sites was 96.2% (25 of 26) for HA cylinders; 93.3% (14 of 15) for HA grooved implants; 75% (three of four) for HA screws; 84.6% (II of 13) for alloy screws; 80% (four of five, two stations; 8 of 10, one station) for cpTi screws; and 75% (12 of 16) for alloy baskets. The results for the various implant designs by study site further showed that no site had more than two failures of any specific implant design except for the alloy basket implants. Five sites had three to five failures with this particular design, whereas 13 sites had no alloy basket failures and another 10 sites had only one alloy basket failure. The five sites with more than two alloy basket failures, while placing only 18% of the total alloy baskets in the study, accounted for half (18 of 36) of the total alloy basket failures. COMPARISON BY PRIOR EXPERIENCE WITH IMPLANT PLACEMENT Sixteen principal surgeons had placed 50 or more implants (more experienced), and 16 had placed fewer than 50 implants (less experienced) before participation in the study. The distribution of surgeons by prior experience among the quartile success groupings is displayed in Table 2. Although the upper quartile was evenly divided between more and less experienced surgeons, a definite difference appeared in the middle and lower groupings. More experienced surgeons exceeded less experienced surgeons in the middle grouping, and
all but one of the surgeons in the lower quartile grouping was classified as having less experience. The percentage of implant failure also differed by surgeon experience. The failure percentage was nearly twice as high (3.22% vs 1.69%) for surgeons having less experience in the placement of dental implants before the DICRG study (44 failures of 1,367 placed) than for the more experienced surgeons (25 failures/ 1,480 placed). Furthermore, the lowest six study sites in percentage success were sites with surgeons classified as having less experience before the DICRG study. They experienced 48% (33 of 69) of the total study failures at uncovering, while placing only 18% (525 of 2,847) of the total study implants. For all implant designs, higher percentages of failure were seen with the less experienced surgeons. Percentages of failure for the different implant designs by prior experience were HA cylinder-more experienced 0% (0 of 401), less experienced 0.65% (3 of 463); HA grooved-more experienced 0.81% (3 of 370), less experienced 0.89% (3 of 338); HA screw-more experienced 0.81 % (I of 123), less experienced 6.15% (4 of 65); cpTi screw-more experienced 4.92% (6 of 122), less experienced 6.67% (5 of 75); alloy screwmore experienced 1.37% (2 of 146), less experienced 4.17% (6 of 144); alloy basket-more experienced 4.09% (13 of 318), less experienced 8.16% (23 of 282).
Discussion Initial studies conducted on dental implant systems usually tend to focus on an evaluation of efficacy (performance under ideal conditions with carefully selected patients and with skilled and experienced surgeons). The DICRG studies recruited study sites without specific inclusion criteria on the amount of experience with the placement of dental implants in general, experience with specific implant systems, or the level of skill of the implant treatment team as indicated by past success rates with dental implant treatment. Additionally, exclusion criteria for study patients were well defined but were not strict as to age, medical status, or available bone. Therefore, it could be argued that the DICRG studies are more representative of a study of effectiveness (performance under normal conditions with a variety of dentists and patients) than efficacy. By viewing study data on success rates related to factors such as experience, adherence to study protocol, and patient selection, an approximate evaluation of efficacy could be conducted. This would involve a restricted analysis of data from study sites in which ideal study conditions were more closely approximated, surgeons had higher levels of experience and skill, and patients were more carefully selected to maximize the probability of implant treatment success. Alternatively, evaluating results that include sites with varied study conditions, treatment providers, and
82 patients, with the potential of associated higher rates of implant failure, can contribute valuable information. Such an investigation may better indicate the effectiveness of implant treatment and be more generalizable to private practices with dentists of varying skills and experience, and patients with varying medical and oral conditions. Problems might be observed with implant treatment in general, or with specific implant designs, that might not be apparent in tightly controlled efficacy studies, which maximize the conditions for success. These ideas seem to be borne out by this analysis of data from the DlCRG database. By ranking and grouping DICRG study sites into quartiles based on the percentage of success with study implants to stage 11 uncovering, lower percentages of failure were seen for many sites, whereas higher percentages are seen for a few sites. The upper quartile of study sites only had two failed implants overall (0.3%), but failures rose to 30 (2.0%) in the middle two quartiles and 37 (5.9%) in the lower quartile. Failure rates by study site ranged from 0% to 10.2%, with a mean of 2.4%. The median failure percentage of 1.8% indicates that the mean failure percentage is being increased by a small number of study sites with higher numbers of failures. The results show generally high success with the HA cylinder and the HA grooved implant designs. The three varieties of screw implant also had generally high success percentages, although they dropped below 93% in the lower quartile of study sites. Results for the alloy basket showed that the percentage of success dropped from 100% in the upper quartile to 84.7% in the lower quartile of study sites. Differences in success of the implant designs used within study strata also were noted. These differences will continue to be monitored to see if they continue after implant loading. Such long-term results could be used in the treatment-planning process to determine implant designs with the highest probability of success for specific applications, and to guide the development of future implant designs. The results for the various implant designs by study site show that no site had more than two failures with any specific implant design except for the alloy basket. Although 13 sites had no alloy basket failures and another 10 sites had only one alloy basket failure, five sites had three to five failures with this particular design. Although failures seemed to be low and evenly distributed among study sites for some implant designs, failures with specific implant designs tended to cluster in a subset of study sites, as seen with the alloy basket design. Clustering of failures may indicate that a greater level of experience and skill is necessary for high success with certain implant designs, possibly related to the complexity of the implant procedure with these designs, or it may indicate a difference in patients treated. The results after classification of the sites by surgeon's experience with the place-
SUCCESS OF MULTIPLE IMPLANT DESIGNS
ment of dental implants before the study support the plausibility of these relationships. All but one of the study sites in the lower quartile success group were classified as having a surgeon with less experience before the study. Overall, the percentage of failure for the less experienced surgeons was nearly twice that seen with the more experienced surgeons. Low failure percentages were seen for both experience levels with the HA cylinder and the HA grooved implants. Percentages of failure increased with the three screw designs and with the alloy basket implants, and were higher with the less experienced surgeons than with the more experienced surgeons. The screw and alloy basket designs have more complex procedures associated with their placement. Some minimum level of experience and proficiency would appear to have importance in the successful placement of these designs. The results of this investigation show that high success rates can generally be achieved with some implant designs, whereas lower success rates might be experienced with some designs, in a subset of dentists, indicating sensitivity of certain designs to operator, patient, or other clinical conditions. The relationships of these clinical factors to success with different implant designs require further investigation.
References I. Spiekermann H, Jansen VK. Richter E-J: A IO-year follow-up
study of IMZ and TPS implants in the edentulous mandible using bar-retained overdentures. Int J Oral Maxillofac Implants 10:23I, 1995 2. Stultz ER, Lofland R, Sendax VI, et al: A multicenter 5-year retrospective survival analysis of 6,200 Integraloo implants. Com pend Contin Educ Dent 14:478, 1993 3. Lekholm U, van Steenberghe 0, Herrmann I, et al: Osseointegrated implants in the treatment of partially edentulous jaws: A prospective 5-year multicenter study. Int J Oral Maxillofac Implants 9:627, 1994 4. Adell R, Eriksson B, Lekholm U, et al: A long-term follow-up study of osseointegrated implants in the treatment of totally edentulous jaws. Int J Oral Maxillofac Implants 5:347. 1990 5. Lambert P, Morris HF, Ochi S, et al: Relationship between implant surgical experience and second-stage failures: DICRG interim report no. 2. Implant Dent 3:97, 1994 6. Lambert PM, Morris HF, Ochi S, et al: Positive effect of surgical experience with implants on second-stage implant survival. J Oral MaxiIlofac Surg 55:12, [997 (suppl 5) 7. Stanford JW: Acceptance program for endosseous implants: A service benefit of ADA membership. Int J Oral Maxillofac Implants 6:15, [991 8. Donovan TE, Chee WWL: ADA acceptance program for endosseous implants. J Calif Dent Assoc 20:60, 1992 9. Last JM (ed): A Dictionary of Epidemiology. New York. NY. Oxford University Press, 1983, pp 31-32 10. Monis HF. Oehi S, Dental Implant Clinical Research Group (planning Committee): The influence of implant design. application. and site on clinical performance and crestal bone: A multicenter. multidisciplinary clinical study. Implant Dent [:49. 1992 I I. Oehi S, Monis HF. Winkler S, et al: Patient demographics and implant survival at uncovering:Dental Implant Clinical Research Group interim report no. 6. Implant Dent 3:247, [994 12. Ismail JYH: A comparison of current root form implants: Biomechanical design and prosthodontic applications. NY State Dent J 55:34, 1989 13. English C: An overview of implant hardware. J Am Dent Assoc 121:360, 1990