J. Mark Thomason Guido Heydecke Jocelyne S. Feine Janice S. Ellis. 168 c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Transcription

1 J. Mark Thomason Guido Heydecke Jocelyne S. Feine Janice S. Ellis How do patients perceive the benefit of reconstructive dentistry with regard to oral health-related quality of life and patient satisfaction? A systematic review Authors affiliations: J. Mark Thomason, Janice S. Ellis, School of Dental Sciences, Newcastle University, Newcastle upon Tyne, UK J. Mark Thomason, Guido Heydecke, Jocelyne S. Feine, Faculty of Dentistry, McGill University, Montréal, Canada Guido Heydecke, Department of Prosthodontics, School of Dentistry, University Hospital, Freiburg, Germany Correspondence to: Prof.J.MarkThomason School of Dental Sciences Newcastle University Newcastle upon Tyne NE2 4BW UK Tel.: þ Fax: þ j.m.thomason@ncl.ac.uk Key words: oral health-related quality of life, prosthodontics, quality of life, reconstruction, satisfaction, systematic review Abstract Background: Reconstructive dentistry encompasses an enormous range of treatment modalities from the restoration of single teeth to the reconstruction of the whole dentoalveolar complex in edentulous patients. Some treatment modalities have been assessed in terms of quality-of-life (QoL) outcomes and satisfaction Objectives: The aim of the present investigation was to search and review studies published between 1996 and 2006 in which the impact of the treatment was measured in terms of QoL outcome, ideally, oral health-related quality of life (OHRQoL). Patient satisfaction was also accepted as an outcome. Materials and methods: The primary search engine used was NICB PubMed based on MeSH headings. Hand searching of the cited references in the included papers identified a number of additional studies. The primary focus of the search was to link treatment to QoL outcomes. Results: The majority of included studies involved the treatment of edentulous patients, particularly the mandible. The preponderance of the studies comparing conventional dentures (CDs) and implant-supported overdentures (IODs) were randomized-controlled trials (N ¼ 18). There was compelling evidence that patients were more satisfied with IODs than CDs. There was strong evidence that OHRQoL can be significantly improved using IODs. Evidence suggesting that one retention system is superior to another needs further clarification. Although high satisfaction ratings have been reported for maxillary implant prostheses, the overall ratings given to the maxillary implant prostheses were not significantly greater than for CDs. There was only sparce information regarding QoL or satisfaction outcomes for the majority of other forms of reconstructive dentistry. Conclusion: Apart from the restoration of the edentulous mandible with IODs or CD, where there is an accumulating body of evidence on the effect of treatment choice, there are many procedures for which there are little or no such data at all. As yet, the entire range of reconstructive treatment has witnessed insufficient investigations relating treatment to its effect on QoL or satisfaction. This is an area that needs to be expanded as a way of quantifying the effect of treatment choices. To cite this article: Thomason JM, Heydecke G, Feine JS, Ellis JS. How do patients perceive the benefit of reconstructive dentistry with regard to oral health-related quality of life and patient satisfaction? A systematic review. Clin. Oral Impl. Res. 18 (Suppl. 3), 2007; doi: /j x The aim of the literature search was to identify and review studies published between 1996 and 2006 in which the impact of the treatment being considered was measured in terms of quality-of-life (QoL) outcome ideally with a clear and specific 168 c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

2 reference to oral health-related QoL (OHRQoL). The questions to be considered were as follows: How do patients perceive the benefit of reconstructive dentistry (OHRQoL)? Within this framework, three additional questions were considered Table 1. Levels of evidence for studies of therapeutic effectiveness I Evidence from at least one published systematic review of multiple well-designed randomized-controlled trials II Strong evidence from at least one published properly designed randomized-controlled trial of appropriate size and in an appropriate clinical setting III Evidence from published well-designed trials without randomization, single-group pre post, cohort, time series or matched case controlled studies IV Evidence from well-designed experimental studies from more than one centre or research group V Opinions of respected authorities based on clinical evidence, descriptive studies or reports of expert consensus committees In patients with severe attrition, particular interventions influence OHR- QoL? Does OHRQoL influence dietary selection and/or intake? Which clinical and surrogate outcomes demonstrate the best correlation to oral health quality of life? Patient satisfaction was also accepted as an appropriate outcome as this links directly with the primary research question how do patients perceive the benefit and would allow the consideration of a possible surrogate for QoL to be considered (sub question 3). Hypothesis The null hypothesis there is no difference in terms of OHRQoL or patient satisfaction between different forms of reconstructive dentistry in the edentulous, partially dentate or in the restoration of individual teeth. Methodology Search strategy The Dental literature was searched and was completed on 4 August The primary search engine used was NICB PubMed, based in the National Library of Medicine and the National Institute of Health, USA. The reference lists of the retrieved articles were themselves screened for further references. The overall Strategy for the search used is shown in Table 1. All searches were based on MeSH headings to facilitate the breadth of the search. The search was limited to the last 10 years and included the following search limitations: clinical trial; meta-analysis; randomized controlled trial (RCT); review; case reports; controlled clinical trial; multicentre study. Inclusion and exclusion criteria and levels of evidence The primary focus of the search was to link treatment to QoL outcomes with specific reference to OHRQoL. As OHRQoL is not listed as a search, MeSH heading and the wider QoL heading was used to avoid the possibility of excluding appropriate studies. In order to be as inclusive as possible the search also included satisfaction as an acceptable surrogate for QoL. This would give an opportunity to discuss the possibility of a satisfaction outcome acting as a surrogate for OHRQoL. While the primary focus for the search was on systematic reviews, Cochrane reviews and meta-analyses of RCTs that used either QoL or satisfaction as an outcome following prosthetic intervention (equating to level 1a for the agency for Health\Care Policy and Research; AHCPR 1992), the criteria were widened to include evidence levels I IV as agreed in the workshop protocol. These map to the current levels of evidence guidelines of the Agency for Health Care Policy and Research. Types of studies A systematic review of multiple well-designed RCTs, RCTs, well-designed trials without randomization, single-group pre post, cohort, time series or matched case controlled studies and evidence from welldesigned experimental studies from more than one centre or research group that considered the effect of rehabilitation or reconstructive dentistry in terms of its impact on OHRQoL or patient satisfaction were included. Participants. Adult patients who underwent some form of reconstructive dentistry for the rehabilitation of edentulousness, partial edentulous, tooth substance loss or discolouration and associated conditions. Interventions. To include Complete dentures Implant-supported complete dentures (removable and fixed) Removable and fixed partial dentures (FPDs) Veneers/crowns Outcomes QoL, OHRQoL Patient satisfaction (with a range of parameters). Sampling and data processing Initial screening based on the outline of the study described in the abstract was undertaken by a panel comprising three of the authors (J. M. T., G. H., J. S. E.). Each paper was assessed by two panel members. Where agreement regarding exclusion was not unanimous, the final decision was taken by the third panel member. Each of the remaining papers was then graded according to the criterion outlined in Table 1. Studies graded at levels I IV evidence were retained for the review. Hand searching of the cited references in the included papers identified a number of additional studies that were included in the review if they met the other criteria. Results The results of the primary search strategy for the period from 5 August 1996 to 4 August 2006 are illustrated in Table 2. Using this strategy, 420 papers were initially identified, of which 83 were review articles and as such contained no new data. c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard 169 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 /

3 Table 2. Synopsis of search strategy and outcome using NICB PubMed, based in the National Library of Medicine and the National Institute of Health, USA Dental prosthesis 5908 Patient satisfaction[mesh] 6439 Dental implants 1817 Personal satisfaction[mesh] 282 Prosthodontics 6289 AND Consumer satisfaction[mesh] 6857 Or 420 Quality of life[mesh] 14, Hits 337 without reviews Oral health[mesh] 323 Health status indicators[mesh] 24,330 Dental health surveys[mesh] 1181 Or 44,396 These papers were excluded. The initial screening based on the assessment of the abstracts, followed by the grading of the full papers based on the outline of the study (viz excluding papers that were not graded at levels I IV evidence) excluded a further 275 papers. Sixteen additional articles that appeared to be potentially relevant after combining the terms and initial abstract review were excluded after assessment of the complete paper because they contained no data on patient-based outcomes. These papers are listed in the excluded section of the reference list. Identification of cited references allowed the inclusion of a further nine papers. These 74 studies were filed electronically. Data included the year of publication, number of participants, type of prosthetic condition and restoration. The type (questionnaire, interview) and number of patient-based measurements [e.g., Oral Health Impact Profile (OHIP) or custommade single-item measures of chewing ability or general satisfaction] were recorded and are reported in annotated form in the tables below. The edentulous mandible The overwhelming majority of the studies meeting the inclusion criteria agreed for the review involved the assessment of treatment of edentulous patients. The majority of these involved treatment of the edentulous mandible. Most studies comparing conventional dentures (CDs) and implantsupported overdentures (IODs) were almost exclusively based on RCTs (N ¼ 18) (evidence level 2); these are annotated in Table 3. Only three papers were not based on RCTs; those comparing different implant systems are annotated in Table 4. Seven RCTs comparing rehabilitation of edentulous patients with either mandibular-supported overdentures opposing an upper complete denture (IODS) or conventional upper and lower dentures (CDs) were identified (Table 3). These studies were represented by a total of 18 papers (Bouma et al. 1997; Awad & Feine 1998; Kapur et al. 1998, 1999; Geertman et al. 1999; Meijer et al. 1999, 2003; Awad et al. 2000a, 2000b, 2003a, 2003b; Raghoebar et al. 2000, 2003; Roumanas et al. 2002; Heydecke et al. 2003a, 2005a; Thomason et al. 2003; Allen et al. 2006). Comparing implant overdentures, preprosthetic surgery and conventional complete dentures Bouma et al. reported an improvement in specific QoL measures including fewer psychological problems at 12 months post-treatment. Despite improvement in dental health-related QoL, no impact on general QoL was detected (Bouma et al. 1997). Significantly better satisfaction scores were reported for the two surgical groups at 1 year (Raghoebar et al. 2000). At 5 and 10 years, the authors reported results for satisfaction data but not QoL data. In particular, at 5 years, satisfaction with the mandibular denture was greater in the two surgical groups and there were significantly fewer complaints in the IOD group (Raghoebar et al. 2000). A similar finding was reported at 10 years (on following the per-protocol analysis), although significance was lost if using the intention-to-treat principle (Raghoebar et al. 2003). It should be noted that after year 1, subjects outside the IOD group had the opportunity to have IODs. In a series of papers reporting an RCT comparing IODs and CDs based on 102 subjects with well-controlled diabetes, Kapur et al. (998) reported no significant differences between groups at 6 months post-treatment (N ¼ 89) but the withingroup changes were significantly greater in the IOD group (P ¼ 0.028). A similar finding was reported at 1 year but not in a subgroup at 2 years (N ¼ 46). Interestingly, at 6 months a sub-group (N ¼ 68) reported a decline in perceived chewing ease and eating frequency, which was more common and greater in the CD than in the IOD group. Both types of study dentures resulted in declines in perceived taste and texture acceptability of almost all test foods but failed to affect food choices (Roumanas et al. 2002). By contrast, chewing experience (as well as masticatory performance) was substantially better in the implantretained overdenture group compared with a complete denture group (Geertman et al. 1999). Both patient satisfaction and QoL data are reported in a series of studies based on a population of middle-aged subjects (N ¼ 102, age years) (Awad & Feine 1998; Awad et al. 2000a, 2000b, 2003a; Thomason et al. 2003; Heydecke et al. 2005a). At 2 months post-treatment, the implant group reported significantly lower post-treatment OHIP scores (P ¼ ), indicating better QoL (Awad et al. 2000a). The IOD group reported less post-treatment looseness for all parameters (eating, speaking, kissing and yawning) and less unease during kissing and during sexual activity than CD subjects but the relationship between sexual activity items and OHIP was weak (Heydecke et al. 2005a). Eighty-nine percent of the variance in patient satisfaction can be explained by gender, patient s rating of comfort, stability, aesthetics and the ability to chew as reported in 1998 by the Montreal Group (Awad & Feine 1998). The difference between groups at 6 months was greater than at 2 months, and at 6 months the differences for ability to speak were also significant between the groups (Thomason et al. 2003). In an attempt to investigate the effect of treatment preference on outcome, the level of satisfaction with the original dentures and level of education were significant predictors of preference. Patients were more likely to express a preference for implant treatment if their ratings of current denture satisfaction were low (Awad et al. 2000b). Greater satisfaction with IODs compared with CDs was reported after 1, 5 and 10 years. Subjects in the CD group were offered the opportunity to receive implants but despite this, the mean satisfaction score of the CD group (including patients who later received implants) was still lower than that of the IOD (Meijer et al. 1999, 2003). Improvements in both 170 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 / c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

4 Table 3. Studies comparing conventional dentures (CDs) and implant-supported overdentures (IODs) Study Summary Primary outcome measure Generalizability A Pre-prosthetic surgery (PPS), IODs and CDs Bouma et al. (1997) Other reports on the above population Raghoebar et al. (2000) Randomized-controlled trial (RCT), N ¼ 90 (IODs ¼ 40, PPS ¼ 30, CDs ¼ 30) Comparison of implant-retained mandibular overdentures IODs and two conventional treatments (PPS and CDs) RCT, N ¼ 90 (IODs ¼ 40, PPS ¼ 30, CDs ¼ 30) (Cawood classes IV and V, mean mandibular height 20.7 mm) Validated self-administered questionnaire Functional ability (chewing, speaking, etc.), patient satisfaction, and quality of life (QoL) (psychosocial functioning) Functional ability (chewing, speaking, etc.), patient satisfaction and QoL (psychosocial functioning). 1 and 5 years Raghoebar et al. (2003) RCT, N ¼ 90 (IODs ¼ 40, PPS ¼ 30, CDs ¼ 30) Functional ability (chewing, speaking, etc.), patient satisfaction and QoL (psychosocial B Diabetic patients Kapur et al. (1998) Other reports on the above population Kapur et al. (1999) RCT diabetic patients, N ¼ 102. Comparison of CD and IOD. Randomized to receive a new maxillary denture and CD or IOD. Treatment was completed for 89 RCT diabetic patients, N ¼ 89 completed treatment 68 provided longitudinal data Comparison of the benefits perceived by patients with a new maxillary denture and a mandibular CD or an IOD Roumanas et al. (2002) RCT diabetic patients, N ¼ 68 provided longitudinal data. Assessment of the impact of IODs and CDs on foods choice of diabetic patients C Middle aged Awad et al. (2000a) Other reports on above population Awad & Feine (1998) RCT, N ¼ 102, years. Maxillary CD and mandibular overdentures retained by two implants and a bar attachment with CDs Data to 2-month post-treatment RCT, N ¼ 102, years. Relationship between patient s ratings of general satisfaction and perception of different aspects of their mandibular prosthesis investigated 12 months after treatment, the average scores for almost all specific QoL measures had improved significantly in all three groups Patients experienced fewer restrictions in their social activities and had fewer psychological problems because of their full dentures. While no impact was detected on general QoL, all three dental treatments had a positive effect on dental health-related QoL Significantly better scores were seen in the two surgical groups (IOD, PPS) than in the CD groups. At 5 years complaints of the lower denture were significantly better in the IOD group. No significant differences were observed between the PPS and CD groups. Both in the short and long term, denture satisfaction appears to be most favourable in the IOD group when compared with the PPS and CD groups. IODs are a satisfactory treatment modality for edentulous patients even when not severely resorbed At 10-year evaluation, the intention-to-treat analysis revealed no significant differences between the three groups but the per- functioning). 1, 5 and 10 years protocol analysis showed that the IOD group was the most satisfied Questionnaires given at baseline and 2 and 6 No significant difference between the distribution of responses to months. Use of denture for eating, chewing the three questions in the two groups at baseline or 6 months. comfort, overall satisfaction However, the change score (6-month baseline) showed a significantly higher percentage improvement in the IOD group compared with CDs Two questionnaires with categorical Mean scores and percentage distributions of longitudinal data for responses were used; questionnaire I showed perceptual improvements with both types of One to 13 questions to ascertain a patient s study dentures. Improvements were higher in the IOD than in the CD absolute assessments of original dentures at group but there was no significant difference entry and study dentures at 6- and 24- Significant difference was found in the change in percentage months distributions for perceptual chewing ability in favour of the IOD Two to 11 questions that assessed the group (although lost in a sub-sample at 24 months) relative change perceived by patients with With the comparative questionnaire, a higher percentage of patients study dentures in the IOD group than in the CD group perceived improvements with study dentures from their original dentures significant for chewing ability and less difficulty in chewing hard foods Perceptions of taste and texture No significant differences were found for frequency in either group, acceptability, and chewing ease on a fourpoint although there was a tendency for perceived deterioration in nominal scale and eating frequency on function. Both types of study dentures resulted in declines in a five-point scale for 13 specific measures perceived taste and texture acceptability of almost all test foods. Declines in perceived chewing ease and eating frequency were more common and greater in the CD than in the IOD group. Both types of study dentures failed to affect food choices Oral Health Impact Profile (OHIP) Implant treatment associated with lower post-treatment OHIP scores indicating better QoL General satisfaction visual analogue scale (VAS), (also comfort, stability, ability to chew, speech aesthetics, ability to clean) Multiple regressing modelling revealed that gender, patient s rating of comfort, stability, aesthetics, and ability to chew were the most important factors relating to denture satisfaction explaining 89% of the variance in satisfaction those who considered ability to chew the most important rated their satisfaction the highest c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard 171 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 /

5 Table 3. Continued Study Summary Primary outcome measure Generalizability Awad et al. (2003a) RCT, N ¼ 102, years. Data at 2 months Thomason et al. (2003) RCT, N ¼ 102, years. Data at 6 months Awad et al. (2000b) RCT, N ¼ 136, years. Investigation of the differences between subjects who have treatment preferences and those who do not Heydecke et al. (2005a) RCT, N ¼ 102, years. Data at 2 months D Geertman et al. (1999) RCT, N ¼ 95 three groups transmandibular bar (Bosker), two implants linked with a bar CD E Meijer et al. (1999) RCT, N ¼ 121, IODs (n ¼ 61) CDs (n ¼ 60). Comparison of patient satisfaction after 10 years Other reports on the above population Meijer et al. (2003) F Seniors Awad et al. (2003b) Other reports on the above population Heydecke et al. (2003a) RCT, N ¼ 121, IODs (n ¼ 61) CDs (n ¼ 60). Comparison of patient satisfaction after 10 years. One year after placement of the denture, unsatisfied patients of the CD group were given the opportunity to receive implants RCT, N ¼ 60, years. Conventional maxillary denture opposing mandibular overdenture (two implants ball attachment) or conventional mandibular denture. Data to 2 month post treatment RCT, N ¼ 60, years. Baseline, 2- and 6-month data General satisfaction VAS, (also comfort, stability, ability to chew, speech aesthetics, ability to clean) of original dentures and new General satisfaction was significantly higher in the overdenture group than for CD group. The implant group reported significantly higher ratings for comfort, stability, and ease of chewing prostheses General satisfaction VAS, (also comfort, Between-group differences greater at 6 months than at 2 months stability, ability to chew, speech aesthetics, Significant difference for ability to speak at 6 months ability to clean) Preferences may influence the outcome of Level of satisfaction with the original dentures and level of education treatment, especially in trials when patients were significant predictors of preference. Compared with subjects cannot be blinded to the type of treatment who rated their satisfaction with their current condition as low, received and the outcome is based on medium rating had an odds ratios (ORs) 0.31 [95% confidence patients evaluations of therapy interval (CI): ] for a preference for implant treatment, and Before receiving treatment, subjects were 0.11 (95% CI: ) for those who rated in the high range required to complete a satisfaction VAS Subjects with high levels of education were significantly less likely to questionnaire regarding their satisfaction have a preference for either conventional or implant treatments for their current prostheses and to indicate (OR ¼ 0.18, 95% CI: and OR ¼ 0.2, 95% CI: , the treatment they would prefer respectively Social Impact Questionnaire, OHIP Significant improvement I implant group for looseness when eating, Questionnaire relating to denture complaints, chewing ability and denture satisfaction Validated five-factor questionnaire 12 questions on mandibular denture seven questions on maxillary denture. Other questions chewing ability tough and hard food, satisfaction score speaking, kissing and yawning. The IOD group reported less posttreatment looseness for all parameters (Po0.0001). IOD subjects less uneasy kissing and during sexual activity than CD subjects. Relationship between sexual activity items and OHIP weak Masticatory performance, as well as chewing experience, was substantially better for the implant-retained overdentures compared with the complete denture group. No significant differences emerged between the TMI and the IMZ groups. No relationship between chewing efficiency and chewing experience 14 of the CD group (23%) chose an IOD after year 1. Patients in the IOD group were significantly more satisfied than patients in the CD group, although the score at 5 years was lower. The mean satisfaction score of the CD group (including patients who later received implants) was still lower than that of the IOD group As above In the CD group, 24 patients (40%) chose an IOD between 1 and 10 years. Patients in the IOD group were significantly more satisfied than patients in the CD group after 1 year, after 5 years and after 10 years. The mean satisfaction score of the CD group (including patients who later received implants) was still lower than that of the General satisfaction VAS, (also comfort, stability, ability to chew, speech aesthetics, ability to clean) OHIP, OHIP-EDENT IOD group despite opportunity for re-treatment with IODs General satisfaction significantly better with IODs at 2 months also higher ratings for comfort, stability and ability to chew OHIP-EDENT IOD subjects had significantly fewer oral healthrelated QoL (OHRQoL) problems OHIP-20, SF-36 IOD subjects had significantly fewer OHRQoL problems measured by the OHIP-20. No significant differences using SF Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 / c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

6 Both groups reported improvement in OHRQoL and denture satisfaction. There were no significant post-treatment differences between the groups, but a treatment effect may be masked by application of intention-to-treat analysis. The OHIP change scores were significantly greater for patients receiving implants than for those who refused them OHIP, and a validated Denture Satisfaction Scale RCT, N ¼ 118. IODs (n ¼ 62) or CDs (n ¼ 56). Comparison of implant-retained mandibular overdentures (IODs) and conventional complete dentures (CDs). Our a priori hypothesis was that implantretained mandibular overdentures would be significantly better than conventional complete dentures Q Edentulous population not requesting implants Allen et al. (2006) Non-RCTs Fartash et al. (1996) Case series, N ¼ 86 VAS questionnaires at baseline and 6 months Patient satisfaction with this form of oral rehabilitation was high in all but two patients who experienced discomfort Positive within-group changes for self-confidence, stability, chewing ability, speech and aesthetics (P-values not given) Watson et al. (1997) Case series, N ¼ 127 based at five centres Anamnestic dysfunction index score 97% considered overdentures better than average sustained over 5 years. Improved stability and retention and improved chewing ability. Increased frequency of reported improved speech over the 5 years. Patients who withdrew rated overdentures less favourably (Helkimo) Effectiveness by questioning patients based on frequency of removal Treatment outcome VAS OHRQoL improved in 96% of the subjects and the OHIP-G49 medians reached the level of OHRQoL in the general population 1 month after treatment. However, the OHIP-G49 medians were below the population norms 6 12 months after treatment. The problem rate for those with removable/complete dentures was 1.9 times the problem rate in fixed prosthodontic patients OHIP-G (German version) at baseline and 1 and 6 12 months after treatment. The sum of OHIP-G item responses (OHIP-G49, range 0 196) characterized OHRQoL impairment in OHIP-G49 medians were compared with the OHRQoL level in a general population sample (n ¼ 2026). A multivariable binomial regression analysis, controlling for the effects of baseline OHRQoL and follow-up wave, was used to compare the level of impaired OHRQoL in different prosthodontic treatment groups John et al. (2004) N ¼ 107. Convenience sample of 42 patients treated with fixed prosthodontics, 31 patients treated with removable dentures and 34 patients treated with complete dentures OHRQoL and satisfaction have been reported in both a CD and an IOD group [RCT, N ¼ 118. IODs (n ¼ 62) or CDs (n ¼ 56)]. While there were no significant post-treatment differences between the groups, the authors reported that the treatment effect may have been masked by application of intention-to-treat analysis. The OHIP change scores were significantly greater for patients receiving implants than for those who refused them, but analysis per-protocol was underpowered to show a difference in the subgroups (Allen et al. 2006). The Montreal Group also reported on the differences between CDs and IODs using ball attachments in a senior population N ¼ 60, years. (Awad et al. 2003b; Heydecke et al. 2003a). Satisfaction was again assessed using visual analogue scale (VAS). At 2 months, general satisfaction was significantly better with IODs as were ratings of comfort, stability and ability to chew. This was paralleled by the observation from the OHIP-EDENT that IOD subjects had significantly fewer OHR- QoL problems (Awad et al. 2003a). Similar QoL results were reported with OHIP at 6 months post-treatment but differences were not apparent with the SF36. The general findings of improved satisfaction with IODs compared with CDs are supported by two case series (Fartash et al. 1996; Watson et al. 1997) and by one convenience sample study for OHRQoL (John et al. 2004). Other implant trials comparing attachments, etc. Additional information on implant (only)- based treatments on QoL and patient satisfaction in edentulous patients can be drawn from a series of studies at different centres comparing variables in implant type, number and attachment methods. The first of these (Naert et al. 1997, 1999, 2004) comprised an RCT comparing three attachment types and reported satisfaction over 10 years. There were initial improvements in satisfaction reported from the old prosthesis for each of the three treatment options (bar, ball or magnet group two implants) at 1 year (Naert et al. 1997). At 5 years, there was no significant difference for satisfaction between different attachment groups, although the magnet group expressed a c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard 173 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 /

7 Table 4. Studies comparing implant-supported overdenture (IOD) attachment/support systems Study Summary Primary outcome measure Generalizability A Naert et al. (1997) Randomized-controlled trial (RCT), N ¼ 36 (12 to each of bar, ball or magnet group two implants) Other reports on the above population Naert et al. (1999) RCT, N ¼ 36 (12 to each of bar, ball or magnet group two implants) Naert et al. (2004) RCT, N ¼ 36 (12 to each of bar, ball or magnet group two implants) B Stellingsma et al. (2003) RCT, N ¼ 60 (transmandibular, augmentation and four implants or four short implants) Comparison of satisfaction and psychosocial aspects in the three groups Other reports on the above population Stellingsma et al. (2005) RCT, N ¼ 60 (transmandibular, augmentation and four implants or four short implants) Comparison of masticatory function in the three groups C Wismeijer et al. (1997) RCT, N ¼ 110 (IOD-2 implants and ball attach, two implants and bar, four implants and bar) Other reports on the above population Timmerman et al. (2004) RCT, N ¼ 103 (IOD-2 implants and ball attach, two implants and bar, four implants and bar) D Walton et al. (2002) RCT, N ¼ 64 (IOD with or without cast frame and by bar or ball) Other reports on the above population MacEntee et al. (2005) RCT, N ¼ 68 at 3 years (IOD with or without a cast frame and by a bar or ball) E Visser et al. (2005) RCT, N ¼ 60 (mandibular height mm supported by two or four implants) F Karabuda et al. (2002) RCT, N ¼ 36 (randomly allocated to the Dolder bar group or the ball attachment overdenture group) Q Tang et al. (1997) RCT X-over, N ¼ 16, long-bar overdenture (LBOD) supported by four implants and two implant hybrid denture R Cune et al. (2005) Case series led up to RCT X-over, N ¼ 18. Comparison of mandibular overdentures using magnet, bar-clip and ball-socket attachments; and (2) assess the relation between maximum bite force and patient satisfaction Questionnaire at 12 (and presumably 36 months) Questionnaire at 12, and 60 months (ninepoint scale, dichotomous scale and open text) Patient satisfaction similar in the three groups although the lowest score was for magnet stability. Significant improvement from old prosthesis but no further changes up to 3 years Patient satisfaction was rated similar for all groups, although the magnet group showed lower retention forces. All patients would repeat the same treatment even though the majority of the magnet group would prefer a more retentive solution Visual analogue scale (VAS) at 10 years Prosthesis stability and chewing comfort for the overdenture were rated significantly lower for the magnet group compared with the ball and bar groups. Prosthesis stability of the maxillary denture was rated significantly lower in the bar group compared with ball and magnet groups Denture satisfaction (11-point scale), psychosocial aspects (Groningen activity restriction scale dentistry and scale for denture patients; Bouma et al. 1997) explored with a battery of questionnaires Masticatory function was assessed before and after treatment using a questionnaire, a masticatory performance test and a structured interview Patient opinion and social functioning baseline and 16 months Patient opinion and social functioning baseline and 19 months and 8.3 years Satisfaction assessed on VAS at baseline, 1 month and 1 year Subjects received two implants and were assigned randomly to one of four groups. Satisfaction assessed on VAS at baseline, 1 month and 1 and 2 years and 3 years (N ¼ 68) Patient satisfaction 54 items on six scales patients clinically assessed at 1, 2, 3, 4 and 5 years with their overdentures Patient satisfaction with the treatment was recorded using a questionnaire (presumable before and some time after treatment) Satisfaction VAS, and categorical scales for physical and psychosocial function and general health Questionnaire on denture complaints at baseline, 3 months of the new denture without attachments, after 3 months of function with each of the attachments (within-subject comparison). Patients were also asked to express their overall appreciation on a VAS Significant improvement in satisfaction within each of the three groups no differences among groups (the augmentation procedure was the least favoured of surgical procedures The patient-based masticatory function improved significantly following treatment no significant differences between groups All patients generally satisfied no difference between groups suggesting therefore the use of the simplest form of treatment All treatments significantly increased scores for satisfaction and functioning improvement is stable for years. Patients with two ball attachment less satisfied with retention at 8 years than those with bar Within-group improvement significant from baseline. Authors advocate bar-clip design as no difference was detected in satisfaction but more repairs were required for ball attachments. Within-group improvement significant from baseline. No significant between-group differences. The mean numbers of adjustments per subject and clinical times did not differ significantly between the two groups. Over 3 years, significantly more repairs occurred in the ball attachment group (Po0.001) No differences in satisfaction were observed between the groups No significant difference between the two attachment types on patient satisfaction Most factors except cleaning rated better with LBOD all subjects chose LBO. All scores for two IOD within 10 mm of LBOD Mandibular implant-supported overdenture treatment reduced various denture complaints. Patients strongly preferred bar-clip (10/ 18 subjects) and ball-socket attachments (7/18 subjects) over magnet attachments (1/18 subjects). Maximum bite force was not correlated to scale or VAS score 174 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 / c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

8 wish for more retention (Naert et al. 1999). At 10 years, satisfaction (general) did not differ significantly between the groups; however, stability and chewing comfort for the overdenture were rated significantly lower for the magnet group compared with the ball and bar groups. Prosthesis stability of the maxillary denture (CD) was rated significantly lower when opposed by a barsupported prosthesis compared with ball and magnet groups (Naert et al. 2004). Stellingsma et al. reported that patients were more satisfied with the new prosthesis (RCT N ¼ 60 transmandibular, augmentation and four implants or four short implants) than with the old prosthesis. They also reported no difference among the three treatment groups for satisfaction (Stellingsma et al. 2003) or with chewing ability (Stellingsma et al. 2005). Similar results have been presented in other studies [RCT N ¼ 110 (IOD two implants and ball attach, two implants and bar, four implants and bar)] (Wismeijer et al. 1997; Timmerman et al. 2004), (RCT N ¼ 68) (IOD with or without a cast frame and by bar or ball) (Walton et al. 2002; MacEntee et al. 2005), (RCT N ¼ 60 mandibular height mm supported by two or four implants) (Visser et al. 2005), (RCT N ¼ 36 Dolder bar or ball attachment group) (Karabuda et al. 2002). These observations raise the question as to the design of the studies in terms of whether they were actually powered to test differences in satisfaction between the different treatment options or whether in fact there was no difference. In these latter studies, seeking to compare different forms of a similar treatment (e.g., different implant attachments), this information was not readily available within the manuscripts. While it is possible to say, therefore, that no difference was found between groups it is not possible to say that there is no difference between the respective treatments being considered. Although not specifically stated in the paper, the cross-over study examining long-bar overdenture (LBOD) and hybrid implant-supported mandibular overdentures undertaken by the Montreal group (Tang et al. 1997) used the same design as a previous study (de Grandmont et al. 1994) powered (40.99) to see a VAS difference of 10 mm or more. The authors demonstrated that factors except cleaning were rated better with the LBOD and that all subjects chose the LBOD. Interestingly, the scores for two IOD were all within 10 mm of the LBOD (and a significant difference was still detected). In a further study, 18 patients wore a new CD before converting it to one of: magnet, bar-clip or ball-socket attachment overdenture as a random cross-over study. In the case series part of the study (CD to IOD) the mandibular IOD treatment reduced various denture complaints. At the end of the study, patients strongly preferred bar-clip (10/18 subjects) and ball-socket attachments (7/18 subjects) over magnet attachments (1/18 subjects). Maximum bite force was not correlated to scale or VAS score (Cune et al. 2005). In a study attempting to examine the relationship between professional clinical evaluation of the oral condition and their prostheses and patient s perceived satisfaction, Heydecke et al. (2003b) reported on a patient sample of 60 from an RCT comparing mandibular implant overdentures and CDs. Patients rated satisfaction before and after treatment on VAS and the treating prosthodontist rated the dentures for the same categories. None of the clinical variables were significantly correlated with patient satisfaction before or after treatment. In addition, the prosthodontists scores were not significantly correlated with patient scores for any question. The authors concluded that the clinicians assessments of the quality of denture-supporting tissues are poor predictors of patient satisfaction with mandibular implant or conventional prostheses. Summary of implant prostheses in the edentulous mandible The majority of the studies in Table 2 were RCTs with an evidence level II. There is compelling evidence that patients were more satisfied with IODs than CDs. Improvements in satisfaction compared with the old prosthesis are seen for both treatment alternatives (within-group changes), but the improvements are significantly greater for the IODs. A similar picture is seen for OHRQoL. Most studies report significant withingroup improvements for both IODs and CD with significantly greater improvements for the IOD. In some studies, for instance those from the Montreal group, separate values for within-group changes and between-group differences in satisfaction with a variety of parameters are reported including: aesthetics, chewing ability (Awad et al. 2003a); stability, comfort (Naert et al. 1999; Awad et al. 2003b); speech (Thomason et al. 2003); food choice (Roumanas et al. 2002); and impact on social and sexual activities (Heydecke et al. 2005a). The trend within the studies has been for simpler rather than more involved treatment. Patients who have previously experienced removable prostheses may choose a removable prosthesis rather than a fixed prosthesis; those who rated stability and chewing ability as being more important chose fixed prostheses, while those who rated ability to clean and aesthetics as being more important selected removable overdentures (de Grandmont et al. 1994). The evidence that OHRQoL measured using the OHIP can be significantly improved by mandibular prostheses stabilized by two implants in edentulous patients compared with a CD is strong (Awad et al. 2003b). These findings are supported by a range of other reports (Bouma et al. 1997; Awad & Feine 1998; Awad et al. 2000a, 2000b; Allen et al. 2001a, 2001b, 2006; Sloan et al. 2001; Allen & Locker 2002; Heydecke et al. 2003a; John et al. 2004; Strassburger et al. 2004). The reduction in the negative impacts on OHRQoL was seen in all sub-scales of the OHIP at a significant level (Awad et al. 2000a). The evidence suggesting that one retention system is significantly better than another needs further clarification. It is not clear whether all the studies outlined were powered to show these differences. For the magnet systems tested to date, there is reasonable evidence to suggest that they result in less patient satisfaction than bar or ball retention systems. LBODs (4 þ implants) were preferred to short-bar OD by patients but the difference in satisfaction, although statistically significant may not be clinically significant (Tang et al. 1997). c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard 175 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 /

9 The edentulous maxilla There were no data available on the effect on OHRQoL of rehabilitation with either fixed or removable prostheses in the maxilla (Table 6). Improvement in the OHIP-49 score (15 points) has been reported with the use of conventional removable dentures (Allen & Locker 2002). Five papers report the effect on patient satisfaction of maxillary rehabilitation with either fixed or removable implant-supported prostheses (Watson et al. 1997; Naert et al. 1998; de Albuquerque et al. 2000; Heydecke et al. 2003b, 2003c). A high satisfaction rating was reported for four splinted implants supporting a removable prosthesis in a series of case reports (Naert et al. 1998). Using a crossover trial design, the Montreal group compared patient satisfaction with maxillary long-bar implant overdentures with and without palatal coverage opposed by a fixed mandibular implant-supported prosthesis (ISPs) (de Albuquerque et al. 2000). Although the ISPs was rated higher than the CD for chewing ability (carrot and sausage), and for chewing function, the overall ratings given to the maxillary implant prostheses were not significantly greater than for new conventional maxillary prostheses. With only 13 subjects, the study was no longer powered for VAS differences of 10 mm at ao0.05 (14 required). The authors reflected that with the very small size of the difference between groups, a sample of 100 subjects would have been required and, as such differences of this magnitude were not clinically significant. The authors suggested that removable maxillary implant prostheses should not be considered as a general treatment of choice in patients with good bony support for maxillary conventional prostheses (de Albuquerque et al. 2000). The decision to use implants in the maxilla for subjects satisfied with their maxillary CD would need to be made on other grounds, such as the preservation of bone. Set against this finding, the same team demonstrated that LBODs were rated significantly higher than fixed prostheses in the maxilla not only for general satisfaction but also for speaking (and ease of cleaning) (Heydecke et al. 2003c) (Tables 6 and 7). Nine of the 13 subjects remaining in this crossover study chose to retain the removable prosthesis. These patient-based findings for speech were fully congruent with those for third-party assessment of speech qualities with the two prostheses (Heydecke et al. 2004) (Table 7). In a nonrandomized parallel-arm study (N ¼ 20), 10 patients were treated with a fixed prosthesis and 10 with removable IODs. In this case, although there were no significant differences between groups for the major satisfaction outcomes, the within-group changes for aesthetics, taste and speech were greater in the IOD group (Zitzmann & Marinello 2000) (Table 6). Additional papers that consider the effect of implant treatment in the edentulous maxilla are also outlined in Table 7. Summary The studies in table had an evidence level of II III. High satisfaction ratings have been reported for maxillary implant prostheses. Nevertheless, the overall ratings given to the maxillary implant prostheses were not significantly greater than for new conventional maxillary prostheses. There is little to suggest that maxillary implant-supported prostheses should be chosen as a general treatment in patients with good maxillary support over CDs. RCTs have shown preference for removable prostheses (over fixed) but little difference between those with and without palatal extension. Other reports The sensitivity to change in OHIP with implant and CD rehabilitation has been addressed by Allen et al. (2001a) (Table 5 and 6). The authors investigated three groups of patients (implant group IG, n ¼ 26, CD in those seeking implants CDG1, n ¼ 22, and CD in those seeking CD CDG2, n ¼ 35) with the objective of assessing the sensitivity to change of the OHIP. Following treatment, a significant improvement in satisfaction with oral prostheses and OHRQoL was reported by IG and CDG2 subjects, although the level of improvement was more moderate for CDG1 subjects. The OHIP change scores were correlated with denture satisfaction change scores. The authors concluded that the sensitivity to change of the OHIP was good and was not improved by using statement weights (Allen et al. 2001a). Ring et al. (2005) reported on a study investigating how individual s responses may change with time; a phenomenon referred to as a response shift (Table 8). An individual may change his or her internal standards, or conceptualization of the target construct as a result of external factors such as a treatment or a change in health status. The authors pre-supposed that this may have important implications for assessing the effects of treatments as a change in QoL that may include a response shift, a treatment effect or a complex combination of both. An Individualized QoL (IQoL) measure (SEIQoL), together with a then-test, was used to determine whether response shift would influence the measurement of treatment efficacy in edentulous patients. IQoL was measured using the SEIQoL-DW in 117 patients at baseline (T1) and 3 months (T2) after receiving high-quality CDs. At the end of the study, the unadjusted SEIQoL index scores revealed no significant impact of treatment at 3 months. However, the then-test at 3 months revealed that patients retrospectively rated their baseline IQoL as significantly lower (Po0.001) than they had rated it at the time (then-test baseline: 69.2). Comparison of the 3-month scores with this re-adjusted baseline indicated a significant treatment effect (then-test baseline: 69.2; 3 months: 73.2, P ¼ 0.016). Eighty-one percent of patients nominated at least one different IQoL domain at 3 months. The data indicated a degree of reconceptualization and reprioritization. The authors suggested that assessment of the impact of treatments using patient-generated reports must take account of the adaptive nature of patients. Food choice One paper designed to test a primary hypothesis relating improved satisfaction with improved food selection was identified (Allen & McMillan 2002) (Table 8). Using the patient group described above, before treatment, all subjects were asked whether they ate a variety of hard and soft foods, to indicate the degree of difficulty they experienced when chewing these foods, and to rate their satisfaction with various aspects of their maxillary and mandibular complete dentures. Subjects who 176 Clin. Oral Impl. Res. 18 (Suppl. 3), 2007 / c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard