Table 3.

Summarization of all the studies included in meta-analysis

References	Age of patients	Number of implants	Follow-up	Objective parameters	Type of study	Sample size	Intervention	Primary end point	Outcome
Narhi TO (2001) Finland	Not reported	2-6 (Branemark, IMZ, Dyna)	72 months cumulative	Clinical examination and interview	In vivo-retrospective comparative analysis	Bar group: 11 Ball group: 5	Bar and ball attachments	The cumulative survival rate for the implants after 72 months was 90%. Loss of bone support correlated with peri-implant probing depth (r=0.29; P<0.02). No differences in mean bone loss between the subjects with ball-retained or bar-retained overdentures were found. The presence of plaque or peri-implant bleeding was not associated with the type of attachment	More mucosal changes were seen with bar superstructures. However, an increased possibility of peri-implant bone loss should be considered when un-splinted implants are used
Mericske-Stern R (2002) Switzerland	40-89 years	4-6 screw-type ITI implants (Straumann, Waldenburg, Switzerland), well spaced, in the anterior part of the maxilla	1-9 years average=4.1 years	Peri-implant parameters: plaque index, bleeding index, probing attachment level, radiographic distance, crestal bone level on radiograph	In vivo prospective case-control	Baseline sample: 173 After loading: 173	Dolder bar, ball anchors or telescopic copings	The 5-year cumulative survival rate of all implants was 94.2%. The peri-implant parameters gave evidence of healthy soft tissues and good oral hygiene. The increase in probing depths and attachment loss were significant (P<0.05). The mean marginal crestal bone loss was about 0.7 mm and was statistically significant at mesial and distal sites (P<0.001).	In planned maxillary overdenture treatment, it is possible to achieve a satisfactory survival rate of the implants
Assad AS (2004) Egypt	48-60 years	2 (Dyna Dental Implant; Dyna Dental Engineering, Bergen op Zoom, The Netherlands) with 13 mm length and 3.6 mm diameter	6 months, 12 months, 18 months	Clinical evaluation Implant mobility Pocket depth (sensor probe on 4 surfaces) Gingival index by Loe and Silness Plaque index Radiographic evaluation Marginal bone height Bone density	In vivo comparative analysis	Magnetic group: 5 Bar group: 5	Magnetic (mucosa supported) and bar (mucosa-implant supported)	Plaque index was significantly high in magnetic group, After 18 months bar supported showed significant increase in gingival inflammation	Mucosa supported (magnetic) overdentures showed less bone resorption distal to implants as compared to bar type. Both the attachments lead to increased functional load and hence affect bone density and resorption
Macentee M (2005) Vancouver , Canada	Average: 61-63 years	2 Nobel Biocare, Canada Inc. Richmond Hill, Ontario, Canada	1 month, 1 year, 2 years, 3 years	Patient satisfaction using 100 mm VAS	Randomized control trial (In vivo)	Baseline group: 100 Attachment group: 100	Bar clip, ball/spring	Over 3 years of follow up, mean numbers of repairs per subject differed significantly between groups, 6.7 repairs in ball spring group, compared to 0.8 in bar clip group (P<0.001)	Subjects were satisfied with new dentures. However, ball-spring attachments required more repairs
Kleis WK (2009) Germany	46-95 years	2 Osseotite® TGStandard implants (BIOMET3iImplant Innovations, Palm Beach Gardens, FL, USA)	12 months	OHIP-G 49 survey for patient satisfaction	In vivo prospective case control	Self-aligning locator group: 25 Traditional ball group: 31	Self-aligning locator, traditional ball attachment	After 1-year of clinical service, locator had 34 prosthetic complications especially the need for change of the male parts or activation because of loss of retention and TG-O ring had 14 prosthetic complications, most of them the change of the O-rings. Dal-Ro abutment had 7 minor complications	Prosthodontic maintenance was restricted to loss of retention for all systems. Locator shows higher rate of maintenance
Cristache CM (2009) Bucharest, Romania	42-84 years	2 Screw Type Standard Straumann Implants (Institut Straumann Ag, Basel, Switzerland)	5 years	Gingival index, plaque index, calculus, bleeding index, probing pocket depth, implant stability, radiographic outcome, prosthetic maintenance, complications and patient satisfaction	Randomized control trial	Ball attachment group (B): 23 Magnetic group (M):23	Retentive anchors (ball attachments) and magnets	The mean bone loss in the 1st year for B group 0.51±0.20 and in M group was 0.51±0.19, and the annual bone loss from 1 to 5 years was 0.05 for both the attachments	With the low initial cost of treatment in B group require high maintenance cost as compared to the M group, initial high cost and low maintenance cost. Both the attachments function well after 5 years
Cune M (2009) Netherland	33-56 years	2 (F Rialit 2, Friadent)	10 years	VAS for patient satisfaction, marginal probing depths, bleeding index and radiographic marginal bone loss were assessed	Crossover clinical trial	Ball group: 7 Bar group: 7	Ball socket attachment: Frialit-2 Friadent Bar- clip, round bar in conjunction with metal omega shaped IMZ clip, Friadent	The differences in the mean scale and VAS scores did not reach a statistically significant level neither at the initial or after 10 years	The probing depth for ball attachments were quite shallower than that of the bar group
Alsabeeha N (2011) New Zealand	Not reported	1	1 year	Mechanical testing: Hardness and elastic modulus measured using nanoindentation technique Clinical performance: Prosthodontic maintenance events Scanning electron microscope	In vivo prospective case control	Ball attachment group: 24 Locator group: 12	Large ball attachment, Standard 2.25 ball attachment, locator attachment	The titanium nitride-coated particles, regardless of system, appeared unchanged and did not require any maintenance. Extensive wear was evident in the uncoated titanium alloy particles and Dalla Bona-type gold alloy matrices, resulting in high maintenance (15 activations). Minimal wear was observed in the plastic matrices with minimal maintenance (2 replacements). The locator nylon matrices showed extensive deformation and deterioration with a substantial need for maintenance (16 replacements).	Large ball attachment systems of titanium nitride-coated particles and plastic matrices reflect favorable wear behavior and clinical performance. These attachments are recommended for patients receiving mandibular single-implant overdentures
								The performance of the particles was related to hardness, while that of the matrices was related to the creep response
Malmstrom H (2015) Rochester, New York	18 years or older	2 Zimmer Tapered Screw-Vent implants (Zimmer Dental Inc, Carlsbad, Calif; 3.75 mm in diameter, 11 mm or 13 mm in length) were placed at intraforaminal area at either the canine or lateral incisor location	2 weeks, 3 months, 1 year, 2 years	To determine the implant success: 1. Gingival index by Loe and Silness, 2. Plaque index, 3. Pocket depths at 6 sites, 4. Mobility measured manually, 5. Bone loss on radiographs Overdenture prosthesis Evaluation: Stability, retention and esthetics on a scale of 0 to 10 Patient satisfaction: VAS scale of 100 mm	In vivo prospective case-control	Baseline: 45 Locator group: 45	Locator attachment (median force pink locator male)	Though there was 11% incidence of replacement of locator male in 2 years, the subjects’ overall satisfaction with fit and ability to chew hard foods with their mandibular overdentures improved significantly (P<0.05) following the denture attachment to the dental implants	97.7% success rate
Geckili O (2015) Istanbul, Turkey	55-72 years	2 (Astra Tech)	6 months, 1 year, 2 years	Measurement of momentary retention forces using dynamic testing machine Patient satisfaction and quality of life 100 mm VAS questionnaire and OHIP-14.	In vivo prospective case-control	Momentary retentive force group: 55 Patient satisfaction group: 55	Ball or locator attachments	No statistically significant association between momentary retention forces and VAS scores (P>0.05). But, higher retention forces presented significantly better quality of life, OHIP score 14 (P<0.05)	Higher retention surely provides better quality of life, it doesn’t affect patient satisfaction
Kappel S (2015) Heidelberg, Germany	Mean age: 73.4 years	2 (BEGO Semados®, BEGO Implant GmbH and Co. KG, Bremen, Germany).	3, 6, 12, 24 months	Implant survival was defined on the basis of clinical stability, plaque index gingival index, radiographic marginal bone loss analysis, prosthetic complication and patient satisfaction using VAS	Randomized clinical study	Bar group: 23 Locator group: 23	Dolder bar and locator	Marginal bone loss after 2 years for locator was 6 mm at both mesial as well as distal surface and 10 mm for bar attachment	Survival was 89.1% and 93.5% for the bar and locator groups, respectively. Survival of the original dentures was 93.5% and 95.7% for the bar and locator groups, respectively
Cepa S (2016) Riegel, Germany	Mean age: 65.2 years	2 (Ankylos, Dentsply Implants, Mannheim, Germany	1, 2, 3 years	Bleeding index, modified plaque index, gingival index, probing depth, radiographic evaluation of marginal bone loss	Randomized clinical trial	Ball group: 12 conus Group: 13	Ball attachments (Ankylos ball heads, Dentsply implants, Mannheim, Germany) or prefabricated conus attachments (Ankylos SynCone, Dentsply Implants, Mannheim, Germany) were used for prosthesis retention`	Compared to the conus group (0.97, 0.66), the mean Modified Plaque Index of the ball attachments (0.39±0.39) was significantly lower (P=0.024) at baseline	Patient satisfaction was 64% for the ball- and 100% for the conus-retained IODs. For the conical attachment, the recommendation of the manufacturer is to use four interforaminal implants to support a removable prostheses
Albuquerque RFD (2018) São Paulo, Brazil	Mean age: 73.2	2 regular tissue-level 4.8 mm diameter Straumann implants in the canine regions of the mandible, conventionally loaded	1 week, and 3, 6, and 12 months	Attachment retention as minimum tensile load, 100 mm VAS questionnaire for patients response	Randomized crossover clinical trial	Ball group (RA): 12 Locator group (LA): 12	Ball attachments (retentive anchor, ref. 048.439, Straumann, Basel, Switzerland) with gold matrices. cylindrical attachment system (LA - Locator, Zest Anchors Inc., Escondido, CA)	Mean retention of worn attachments and patient satisfaction with denture retention assessed in the preintervention phase were 3.2 N (SD 4.9) and 23.5 mm (IQR 6.5-65.5), respectively. Overall mean retention along the study was higher for RA than LA (difference of 5.0 N, 95% CI: 2.5-7.6; P=0.0005), declining significantly overtime (P<0.0001), more steeply for the cylindrical attachment	Ratings of retention decreased significantly overtime for both systems and earlier for LA than RA. Satisfaction was lower when retention was either too low or too high, although there was large variation in this association. At the end of the study, all participants chose to remain with the attachment system that they had received last
Boven GC (2019) Netherland	Mean age: 60.1-63.8 years	4 dental implants 3.5 mm in diameter (Nobel Active Narrow Platform, Nobel Biocare AB, Gothenburg, Sweden) in maxilla	1 year	Radiographic bone level, implant survival, overdenture survival, peri-implant hygiene, soft tissue conditions, and patient satisfaction	Randomized clinical trial	Bar group: 25 Locator group: 25	Bar (Nobel Biocare AB), locator attachment	Marginal bone loss was estimated 0.58±0.71 for locators and 0.31±0.47 for bar group over 1 year	Implant survival was 96.7% in locator group and 97.9% in bar group
Park JH (2019) Seoul, Korea	Mean age 61.6±10.6	4 roughened surface dental implants (NRline®, Dentium) 3.1-5.0 mm in diameter and 7-13 mm in length were placed in each anterior maxilla	1 year	PI, BOP and GI and were assessed at definitive denture delivery (T0) and at 3 months (T3) and 12 months (T12) after definitive prosthesis delivery	Randomized control trial	Bar group: 20 Ball group: 20	a bar group (Hader bar, Preci-Horix, Alphadent NV) and a ball group (Mini-ball attachment, Dentium	The mean marginal bone loss after 1 year of loading was 0.34±0.88 mm, and there were no significant differences between the two groups. Plaque index, gingival index and bleeding on probing were significantly higher in the bar group (P<0.001), and the implant success rate of the bar group was significantly lower than that of the ball group (P=0.028)	The bar group was more vulnerable than the ball group with respect to maintaining peri-implant tissue health
Taha NEKS (2019) Goiania, Brazil	Not reported	A cylindrical Morse taper implant (TitamaxCM Cortical, NeodentImplants), 3.75 mm diameter, was inserted in the mandible midline, 1-2 mm infrabony	1 week, 3 months	Patient satisfaction with the mandibular denture according to six items: comfort, speech, chewing ability, retention, stability and overall satisfaction. Each item was measured as scored by the patient in a VAS ranging from 0 to 100	Randomized cross over clinical trial	Ball group: 9 Equator group: 9	The ball-type attachment (Bola Mini CM, Neodent Implants) The “stud” type Equator attachment (Equator CM, Neodent Implants)	No significant differences were observed when comparing the ball and Equator at the initial (P=0.330) and final (P=0.08) periods	The use of a single midline implant to retain a mandibular overdenture significantly improves patient satisfaction irrespective of the attachment used

VAS: Visual analogue scale, PI: Plaque index, BOP: Bleeding on probing, GI: Gingival index, SD: Standard deviation, CI: Confidence interval, OHIP: Oral Health Impact Profile, IMZ: Intramobile Cylinder Implant System, RA: Retentive Anchors for ball attachment group, LA: Locator Attachment group