successes and failures with orthodontic mini-screws: a ... · with orthodontic mini-screws: ......

Address for correspondence:

A. COSTI,21, avenue Niel,75017 [email protected]

DOI: 10.1051/odfen/2010407 J Dentofacial Anom Orthod 2010;13:401-412� RODF / EDP Sciences

401

Article received: 06-2010Accepted for publication: 07-2010

C E C S M O R U B R I C

Successes and failureswith orthodontic mini-screws:a retrospective clinical studyof 95 mini-screw implantations

Arnaud COSTI

CECSMO Memoire directed by Doctor M.-H. DECOSSE

The use of mini-screws has increasedbroadly over the last ten years. Effectively,these implantable devices, greatly reducedin size and easy to set in place and toremove, have provided orthodontists with alow cost means of reinforcing anchorage fora pre-determined period of time. However,their success rate is far from satisfactorycompared to the well-established implantsused for tooth replacement or for othersources of skeletal anchorage. In a recentmeta-analysis Schatzle9 reported a failure

rate of 16.4% for mini-screws as opposed to10.5% for palatal implants and only 7.3% formini-plates.

The Faculty of Dental Surgery, Paris 7,determined, accordingly, that there wasneed for a retrospective study on the useof mini-screws. Beyond a simple calculationof the failure rate, we felt it was especiallyimportant for us to try to gain an under-standing of the most dangerous risks so thatthey could be avoided in the future.

1 – METHODS AND MATERIALS

Between December 2005 and May 2009,41 patients consulted us at Garanciere forprospective orthodontic treatment thatwould require use of mini-screws.

We included in the study only the29 patients who had, in fact, benefited from

placement of mini-screws and follow-uporthodontic care at the Faculty, 22 womenand 7 men from 12 to 61 years of age, withan average age of 40. It is important to notethat 13 of them received preliminary period-ontal treatment (fig. 1).

Article available at http://www.jdao-journal.org or http://dx.doi.org/10.1051/odfen/2010407

http://www.jdao-journal.org

http://dx.doi.org/10.1051/odfen/2010407

In total, we placed 95 AbsoAnchor@

(Dentos@) self-drilling, cylindro-conical,titanium alloy mini-screws were avail-able in a variety of:– lengths, ranging from 6 to 12 mm.

We used the 10 mm mini-screwsmost frequently, in 70% of thecases;

– diameters, ranging from 1.2 mm to1.6 mm. We used the 1.3 mm sizefor 67% of our sample;

– models. We chose the SH, smallhead mini-screw more often thanthe CH, circle head.Intrusion was the type of tooth

movement that our orthodontistsmost frequently decided needed theassistance of mini-screw anchorage.But they decided on the type, number,and placement of the screws in

collaboration with the oral surgeon.Primarily they more often elected touse mini-screws in the maxilla thanthe mandible, buccally than palatally,and in the posterior rather than theanterior. More precisely, mini-screwswere placed between the secondpremolar and the first premolar 40%of the time. In 93% of casesthe attached gingiva was respected.(table 1).

Orthodontic treatment was neverstarted before necessary periodontaltreatment had been completed.

Fourteen different operators, all ofthem experienced in implantology,placed the mini-screws. The mini-screw placement protocol began witha chlorhexidrine mouth rinse, injectionof local anesthetic, and preliminary

CE

CS

MO

RU

BR

IC

Figures 1 a to eInitial intraoral photos of M.T. taken after periodontal prophylaxes. An over-retained upper right temporary canine,distal to the upper right permanent canine had a 3 mm maxillary midline deviation to the left.

ARNAUD COSTI

402 Costi A. Successes and failures with orthodontic mini-screws: a retrospective clinical study of 95 mini-screw implantations

drilling of the site with a 0.9 mm bur.In conformity with the manufacturer’srecommendations, the operators in-serted the mini-screws with the aid ofa manual screw driver furnished in thekit (fig. 2 a to c).

After inserting the mini-screws wetook periapical radiographs to be surewe had not breached the integrity ofany adjacent roots.

The patients were assigned to14 CECSMO students. Sixty-onemini-screws were inserted and loadedimmediately and 41% of them wereleft in place for more than 6 months.

Operators usually, in 74% of thecases, used a direct connection ofscrew to appliance rather than anindirect one (fig. 3 a and b).

CE

CS

MO

RU

BR

IC

NUMBER OF SCREWS

INDICATION

intrusion

uprighting

antero-posterior traction

50 (53%)

8 (8%)

37 (39%)

ARCHmandibular

maxillary

19 (20%)

76 (80%)

SURFACEpalatal

buccal

35 (37%)

60 (63%)

SECTORposterior

anterior

89 (94%)

6 (6%)

TYPE OF GINGIVAattached

free

89 (94%)

7 (7%)

Table IIndication and site of mini-screw insertion.

Figures 2 a to cAfter having removed the upper right temporary cuspid, the student orthodontist had two mini-screws placed forpatient M.T., both CH 1.3 x 10 mm between the upper right first and second bicuspids, one buccal and the otherpalatal. After preparing the sites with a .9 mm bur (a) Dr. Rosec, the oral surgeon, inserted each mini-screw with theaid of a screwdriver until all the threads were buried.

SUCCESSES AND FAILURES WITH ORTHODONTIC MINI-SCREWS

J Dentofacial Anom Orthod 2010;13:401-412 403

Each patient had a four page infor-mation booklet containing on the:

– First page a short medical historyof the patient reporting age, sex,previous medical and dental his-tories, state of oral hygiene,facial type, local constraints, andany painful sites;

– Second page an account of thesurgical situation: type of pre-surgical preparation; informationon the implant site, its locale, itsosseous density, and the qualityand quantity of surrounding softtissue; the surgical protocol; andany noted defects;

– Third page an assessment by theorthodontist covering indicationsfor treatment, the bio-mechanicalappliance and force applications

being considered, and data ontreatment timing, anticipatedduration, possible problems andcomplications;

– Fourth page devoted to the mini-screws stipulating their shapeand dimensions.

All this data was keyed into a digitaltable with the Microsoft Office Excel@

2007 program with which we couldcalculate the overall failure rate. Thenwe made a univariate analysis ofvariance of the potential causes offailure in this study. This was donewith the aid of the R DevelopmentCore Team@ 2009 program composedby Doctor Charpentier, a dentist andstatistician at the Pitie-SalpetriereHospital in Paris.

2 – RESULTS

2 – 1 – Calculating the failure rate

Of the 95 mini-screws that we putin place, 23 were lost, which consti-tutes a failure rate of 24.2%. Thecauses of failure were:

– 3 fractures;– 7 loss of stability, 5 immediately

and 2 later;– 13 losses of stability associated

with inflammation of surrounding

CE

CS

MO

RU

BR

IC

Figures 3 a and bThe orthodontist delivered force to patient M.T.’s appliance with elastic chains, changed every two weeks, to retractthe upper right cuspid into the space left by the over-retained temporary canine. Re-opening of space for the upperright first molar was accomplished at the same time. (a) Next the patient was asked to supply the force necessary foren masse retraction of the upper incisors (b) by placing elastics between the between the buccal mini-screw anda hook soldered on the arch wire just distal to the upper right lateral incisor.

ARNAUD COSTI


soft tissues, 3 immediately and10 later.

It is interesting to note that:– 8 screws were lost prematurely,

that is before any force wasapplied to them,

– 15 screws began to loosen laterwhen they were subjected toorthodontic traction, 3 were con-nected indirectly, 12 were con-nected directly (fig. 4).

This failure rate of 24.2% is about1.5 times higher than the 16.4% figurethat Schatzle9 reported in his meta-analysis and is situated beyond theupper limit of the 95% confidencelevel (95% Cl: 13, 4–20.1%).

2 – 2 – Univariate analysisof variance

An examination of the data sug-gests that the factor ‘‘patient’’ isimportant in view of the unequal

distribution of screws between pa-tients.

A comparison between the func-tions of distribution of the number ofscrews with the number of failuresprovides another illustration of thisinequality: the four patients with themost failures accounted for 52%although they had received only 22%of the screws.

And the small size of the sampleand the poor distribution of some ofthe factors being studied, notably thefactors clustered around ‘‘patient’’,make this intervening variable one ofconfusion.

The comparisons that we are aboutto make will accordingly be presentedin a mixed model format, the onlyaleatory factor being an intercept on‘‘patient.’’

In decreasing order of significance,where p < .01, we have retained thefollowing variables: number of screws

CE

CS

MO

RU

BR

IC

1. Immediate 2. Later 1. Immediate 2. Later

FRACTURES LOSS OF STABILITY LOSS OF STABILITYAND INFLAMMATION

Nature of failures

Nu

mb

er

of

fail

ure

s

PREMATURE – without connectionLATER – indirect connection

LATER – direct connection

12

10

8

6

4

2

0

1

2

8

12

3

3 3

Figure 4Distribution of failures in relation to their nature and the

type of connection.



in the system considered, angle ofscrewing in relation to the long axis ofthe tooth to which it was connected inan application of force, duration of itsbeing employed, and type of connec-tion (tab. 2).

It flows from this analysis that therisk of failure would increase when:• The number of screws in the system

diminishes: the risk is maximum forthe first screws not put to work; thenthe risk diminishes for the systemsof 1, 2, 3, 4, and, finally, 5 screwswhere the failure rate is nil;

• The screw is inserted at 900 ormore with respect to long axis ofthe tooth rather than 40-450;

• The duration of the orthodonticemployment is short: failures pri-marily occurred primarily before

force was applied to the screw orduring the first three months of itsemployment;

• The type of connection applied littleforce to the screw: failures oc-curred most often before any ortho-dontic force was applied, and morefor indirect connections than fordirect.Of the 13 mini-screws that were not

put to work, 8 had already failed andcouldn’t be used for that reason. Thiscomes to a failure rate of 8/13 or 62%.This high but logical proportion offailures occurred with a few para-meters that we deemed interesting:the number of screws in an activeforce system was zero, the duration oftheir employment was zero, and theywere never connected.

CE

CS

MO

RU

BR

IC

VARIABLE LEVEL SUCCESS FAILURESHARE OF FAILURE

PER LEVEL

Number of screws

in the force system

0

1

2

3

4

5

5

8

14

13

31

5

8

2

4

2

4

0

62%

18%

20%

11%

10%

0%

Angle of screwing

with respect to the

long axis of tooth

40–45�90�

58

14

10

3

15%

48%

Duration of

employment

(month)

0

< 3

3 to 6

> 6

5

18

14

35

8

8

3

4

62%

31%

18%

10%

Connection0

indirect

direct

5

9

58

8

3

12

62%

25%

17%

Table 2Success and failures of the variable.

ARNAUD COSTI


An analysis of these parameterswould make more sense if we limitedit to the sub-group of screws effec-tively employed. It would, accordingly,

be interesting to complete our calcula-tions by a statistical analysis of thissub-group.

3 – DISCUSSION

3 – 1 – Critical analysisof the results

With the univariate analysis wewere able to establish a relationshipof the fairly high failure rate of 24.2%and certain variables.

First, it seems that the more screwsthere are in a given orthodontic set-upthe fewer failures there will be. Itwould also appear that when theintensity of the force is spread over alarge number of screws it imparts aweaker destabilizing action on each ofthem. Chen2 had already formulatedthis hypothesis in one of his articles.And quite recently Kim4 confirmed thisidea by showing in his multivariateclinical study that splinting screwstogether was the only factor capableof expressing a positive influence onthe success rate of screws placed inthe palate.

The second parameter that comesinto focus from our study is the angleof the screwing with respect to thelong axis of the tooth. An angulation of40-50� would be preferable, for exam-ple, than one equal to or greater than90�, a better primary stability support-ing these results.

In effect, when the gingiva is notattached at a high point, an insertion ofa screw at a 90� angle in the keratinizedgingival suggests that it is close to theinter-dental papilla where the lesseramount of bone is of poorer quality than

it is a deeper level. And if, to compen-sate for this, the screw is aimed at thebase of the buccal sulcus it risks beingdestabilized by the muscular activity ofthe cheeks and lips.

Pickard8 also asserts that screwsangulated at 45� to the direction of theorthodontic traction rather than 90� orin ‘‘tent stake’’ fashion have greaterstability. He explains this phenomen-on by its more equal sharing ofconstraints with the bone surroundingthe implanted screw. Wilmes12 re-ports that animal studies have shownthat an insertion angle of 60–70� isideal but that this angulation has to bereduced when the screw is placednear roots of teeth. It is true that if theangulation is too acute, the operatormight slip in inserting the screw, asPark7 has observed. In addition, theuppermost thread or threads may notbe set in cortical bone.

We also found that our mini-screwfailures occurred essentially in the firstfew months after their insertion, 8 rightafter surgery, 8 in the first 3 months oftheir being put to use, which repre-sents 2/3 of the failures (16 of 23).From this it seems clear that operatorsshould delay treatment until after a‘‘natural selection’’ of screws hastaken place during the first fewmonths after they have been placed.

Paradoxically, the mini-screws thatwe connected directly to appliancesfailed slightly less frequently. 17.1%

CE

CS

MO

RU

BR

IC



of the time, than mini-screws that weconnected indirectly to appliances,whose failure rate was 25%. This isthe reverse of most reports in theliterature1, where indirect connectionshave been associated with slightamounts of anchorage loss. It ispossible that operators have con-nected slightly mobile mini-screwsindirectly as a precautionary measurethat failed to accomplish its purpose insome cases.

Summing up, we found that theunivariate analysis merely indicatedpaths worthy of future investigation.In order to prove the accuracy of ourobservations and the hypotheses theysuggested it will be necessary to setup as many prospective studies asthere are variables to test and to dothis with a scrupulously identical pro-tocol for all the different groups.Moreover, we have to take certainparticularities specific to this studyinto consideration.

3 – 2 – Particularities of this study

As with any retrospective study, wehave to ask ourselves what level ofproof our results have achieved.

The first possible bias to consider isthe recruitment and follow-up of pa-tients, most of whom had consultedthe Periodontology Service seekingcorrection of migrations of teeth thatresulted from loss of other dental unitsas well as generalized weakening ofperiodontal support. This sample ofpatients is special because they wereat greater than average of sufferinginflammation or infection around theimplant sites. So the results weobtained cannot necessarily be extra-polated to a larger population.

Furthermore, there were almost asmany practitioners, 14 orthodontistsand 14 oral surgeons, as there werepatients, which it makes it question-able that the protocol can be repli-cated or is, in fact, reliable; especiallysince none of the professional partici-pants in the study had really had anyprevious experience in the field. And,as we know, there is for every surgicalprocedure, as is true for any endeavor,a learning curve of apprenticeship4.

In a prospective study carried out in2004 at the University of Aarhus inDenmark Melsen6 obtained a failurerate of 10.5%. Of the 19 failuresreported, 16 occurred in the firstweeks of the project, all of which thechief investigator attributed to theinexperience of the dentists whoplaced the mini-screws and especiallyto the difficulty they encountered inkeeping the axis of the screwing-inconstant. This study showed thatwhen the screwdriver is held incor-rectly, like a pen, the failure rate isgreater than when it is grasped withthe whole hand, the palm exertingpressure.

In our study we used only screws ofthe AbsoAnchor@ Dentos@ brand,which many authors, especially thosefrom Asia, have employed with suc-cess. But in our study we used onlyself-setting screws with a diameter of1.6 mm or less, the majority of whichwere 1.3 mm in diameter. Our havingutilized a pre-insertion drilling or aprepared hole in more than half ofthe cases certainly avoided the occur-rence of additional fractures. But, onthe other hand, it may have contrib-uted to a loss of primary stability insome cases where the screw’s dia-meter was 1.6 mm or less, particularly

CE

CS

MO

RU

BR

IC

ARNAUD COSTI


in the maxilla. Moreover, Kim3 con-siders that pre-drilling is not onlyunhelpful, but perhaps harmful forprimary stability when the diameterof the screw if 1.6 mm or less, thepreparatory bur used always beingsmaller sized than the mini-screwschosen for their sites1.

3 – 3 – Some thoughtson the notion of failure

Some authors11 believe that anyscrew that becomes mobile must beconsidered a failure. While it is truethat the short term viability of anymobile screw is not promising, if it hasfulfilled its role there is no reason thatit should be recorded in the categoryof failures.

The determining factors are themoment when it began to showmobility and the extent to which it ismobile. We must bear in mind thatimplanted orthodontic mini-screwshave a limited raison d’etre and periodof utility: to facilitate the movement ofa tooth or groups of teeth by servingas temporary anchorage reinforce-ment and the time required to accom-plish that goal (fig 5 a to e).

In other words, the success oforthodontic movement does not ne-cessarily depend on the success ofthe mini-screw: in our study the failureof one mini-screw was associatedwith satisfactory displacement ofteeth and the failures of 5 mini-screwswere associated with partial, butacceptable orthodontic movement.

CE

CS

MO

RU

BR

IC

Figures 5 a to eIntraoral photos taken at the end of orthodontic treatment by Dr. Rosec, for patient M.T. The maxillary center line hasbeen corrected and spaces have been opened in the upper right and lower left first molar regions in which implantscan be placed.



This means either that even thoughthese screws were lost prematurely,they played their roles satisfactorily orthat recourse to mini-screws was notindispensable in these cases (table 3).

On the other hand, of course, thesuccess of a mini-screw does notnecessarily mean that the orthodontictreatment has succeeded: 8 screwsdeemed to be successful were asso-ciated with failure of the orthodontictherapy and 30 with improvementsthat were only partial. A mini-screwremains nothing more than an aid totherapy like any other aid and it is theorthodontist’s responsibility to antici-pate and overcome biomechanicaldifficulties.

In addition, Liou5 and later Wang10

have demonstrated that, unlike teeth,mini-screws subjected to orthodonticforce can move through bone without

becoming mobile. Such displace-ments, of the head as well as the tipof the screw, can extend to 1.5 mm.To explain this phenomenon theseauthors have pointed out that thefibro-conjunctive nature of the bone-implant interface allows a certain free-dom of movement in the direction ofthe source of the force application.This means then that mini-screwanchorage is not absolute as wasformerly thought.

Finally, the concept of failure isrelative: the criteria that authors useto define it differ considerably. For thenotion of failure to have clinical sig-nificance success rates should besystematically related to the lengthof utilization: so that we would speakof a survival rate of 6 months, 12months, and, eventually 18 months.

4 – CONCLUSION

Mini-screws offer orthodontists anappreciable source of temporary sup-plementary anchorage. However, theirfailure rate is not negligible.

The retrospective clinical study thatwe carried out over the last threeyears at the Faculty of Dental Surgery

of Paris 7, compiled a failure rate of24.2%, which is higher that the 16.4%figure reported in the literature.

The majority of our failures wereassociated with a loss of primarystability, a third of them occurringbefore the application of orthodontic

CE

CS

MO

RU

BR

IC

VARIABLE LEVEL SUCCESS FAILURE

TOOTH

MOVEMENT

insufficient

partially satisfactory

satisfactory

8

30

34

17

5

1

Table 3Success of tooth movement in relation to success of the mini-screws.

ARNAUD COSTI


force and a second third duringthe first three months of mechano-therapy.

The univariate analysis we con-ducted stressed the importance ofcertain parameters, notably that themore widely force application is dis-tributed to a range of mini-screws thelower the failure rate will be. More-over, mini-screws inserted perpendi-cular to bone or in ‘‘tent stake’’

fashion have higher than average fail-ure rates.

Certainly the inexperience of theoperators in our study accounts for ahigh proportion of the failure ratebecause as in any implant techniquethere is a learning curve.

Consequently, offering studentspractical training in the placement ofmini-screws will contribute greatly toan increase in future success rates.

REFERENCES

1. Chen Y, Kyung HM, Zhao WT, Yu WJ. Critical factors for the success of orthodonticminiimplants: a systematic review. Am J Orthod Dentofacial Orthop 2009;135:284–291.

2. Chen YJ, Chang HH, Huang HC, Lai EHH, Yao CCJ. A retrospective analysis of thefailure rate of three different orthodontic skeletal anchorage systems. Clin Oral ImplRes 2007;18:768–775.

3. Kim JW, Ahn SJ, Chang YI. Histomorphometric and mechanical analyses of thedrillfree screw as orthodontic anchorage. Am J Orthod Dentofacial Orthop2005;128:190–194.

4. Kim YO, Yang SM, Kim S, Lee JY, Kim KE, Gianelly AA, Kyung SH. Midpalatalminiscrews for orthodontic anchorage: factors affecting clinical success. Am J OrthodDentofac Orthop 2010;137: 66–72.

5. Liou EJW, Pai BCJ, Lin JCY. Do miniscrews remain stationary under orthodonticforces? Am J Orthod Dentofacial Orthop 2004;126:42–47.

6. Melsen B, Laursen MG. Factors in the decision to use skeletal anchorage. In: Nanda R.Temporary Anchorage Devices in orthodontics. StLouis, Missouri : Mosby Elsevier,2009, 73–89.

7. Park YC, Lee KJ. Biomechanical principles in miniscrewdriven orthodontic. In: NandaR. Temporary Anchorage Devices in orthodontics. StLouis, Missouri : Mosby Elsevier,2009, 93–144.

8. Pickard MB, Dechow P, Rossouw PE, Bushan PH. Effects of miniscrew orientation onimplant stability and resistance to failure. Am J Orthod Dentofac Orthop 2010;137:91–99.

9. Schatzle M, Mannchen R, Zwahlen M, Lang NP. Survival and failure rates oforthodontic temporary anchorage devices: a systematic review. Clin Oral Implants Res2009;20(12):1351–1359.

10. Wang YC, Liou EJW. Comparison of the loading behavior of selfdrilling and predrilledminiscrews throughout orthodontic loading. Am J Orthod Dentofacial Orthop 2008;133:38–43.

11. Wiechmann D, Meyer U, Buchter A. Success rate of miniand microimplants used fororthodontic anchorage: a prospective clinical study. Clin Oral Impl Res 2007;18:263–267.

12. Wilmes B, Drescher D. Impact of insertion depth and predrilling diameter on primarystability of orthodontic miniimplants. Angle 2009;79:609–14.

CE

CS

MO

RU

BR

IC

We offer our warm thanks to Dr. CHARPENTIER for his statistical analysis, to Dr. CAMY for producingthe questionnaire, and to Dr. ROSEC for participating in treatment of the clinical case.



References exactes du logiciel d’analyse statistique

R Development Core Team� (2009). R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-90005107-0,URL http://www.R-project.org.

Douglas Bates and Martin Maechler (2009). lme4: Linear mixedeffects models using S4classes. R package version 0.999375-31.

http://CRAN.Rproject.org/package=lme4.

CE

CS

MO

RU

BR

IC

ARNAUD COSTI


successes and failures with orthodontic mini-screws: a ... · with orthodontic mini-screws: ......

Documents