You are on page 1of 3

Multifactorial risk assessment for survival of

abutments of removable partial dentures based on


practice-based longitudinal study
Abstract
Objectives: Predicting the tooth survival is such a great challenge for evidence-based dentistry. To prevent
further tooth loss of partially edentulous patients, estimation of individualized risk and benefit for each residual
tooth is important to the clinical decision-making.
While there are several reports indicating a risk of losing the abutment teeth of RPDs, there are no existing
reports exploring the cause of abutment loss by multifactorial analysis. The aim of this practice-based
longitudinal study was to determine the prognostic factors affecting the survival period of RPD abutments using
a multifactorial risk assessment. Methods: One hundred and forty-seven patients had been previously provided
with a total of 236 new RPDs at the Osaka University Dental Hospital; the 856 abutments for these RPDs were
analyzed. Survival of abutment teeth was estimated using the KaplanMeier method. Multivariate analysis was
conducted by Coxs proportional hazard modelling. Results: The 5-year survival rates were 86.6% for direct
abutments and 93.1% for indirect abutments, compared with 95.8% survival in non-abutment teeth. The
multivariate analysis showed that abutment survival was significantly associated with crown-root ratio (hazard
ratio (HR): 3.13), root canal treatment (HR: 2.93), pocket depth (HR: 2.51), type of abutments (HR: 2.19) and
occlusal support (HR: 1.90). Conclusion: From this practice-based longitudinal study, we concluded that RPD
abutment teeth are more likely to be lost than other residual teeth. From the multifactorial risk factor
assessment, several prognostic factors, such as occlusal support, crown-root ratio, root canal treatment, and
pocket depth were suggested. Clinical significance: These results could be used to estimate the individualized
risk for the residual teeth, to predict the prognosis of RPD abutments and to facilitate an evidence-based clinical
decision making.

Introduction
When designing and providing dental prosthesis, it is very important to estimate the individualized risk
and benefit of the prosthesis for the residual teeth. In planning dental treatment for patients with tooth
loss, the potential impact of replacement prostheses on dental health must be considered carefully. The
concept of biological price is frequently described, and the replacement of missing teeth should be
balanced with the potential for a prosthesis to contribute to dental and periodontal diseases. This is
particularly relevant to the teeth used as abutments for fixed and removable partial dentures (RPDs).
Various reports have referred to the association between replacement prostheses, particularly RPDs, and
dental diseases. The abutment teeth of RPDs were reported to be at higher risk of periodontitis, dental
caries and root fracture than other teeth. Longitudinal studies have also shown that RPD abutments were
at the increased risk of loss. Tooth loss, especially in the case of abutment teeth, is intrinsically involved
in complex relationship with many factors in the long term. However, while this mandates a multifactorial
analysis using practice-based research to examine the significant risk factors determining tooth loss,
such analysis has not been reported yet for RPD abutment teeth. As long as RPDs remain a common
treatment option for partially edentulous patients, it is imperative to know the specific prognostic factors
dictating the survival of RPD abutments, and their relative contribution to the duration of tooth survival.
This knowledge facilitates the development of prosthodontic treatment strategy and the evidence-based
prediction of long-term prognosis for those abutment teeth. The absence of any such risk assessment
denies us a practical way for predicting the survival period of each abutment depending on their
individual characteristics. This longitudinal retrospective cohort study aimed to fill this void by examining
the survival of RPD abutments in longitudinal clinical cases, and exploring the prognostic factors dictating
survival and their relative contribution to tooth loss.

Methods
Study population
We targeted all patients provided with RPDs between January 2002 and December 2003 in Removable
Prosthodontics department of Osaka University Dental Hospital, Japan. The protocol of this study was
approved by the School of Dentistry Ethics Committee (No. H22-E2). Patients were included if they had
been provided with a clasp-retained, cobaltchromium designed and tooth supported RPD which is
covered by Japanese medical insurance and had used it for 2 years or more, and were excluded if their
dentures were immediate RPDs that required fixing, and dentures with complex designs such as
maxillofacial prostheses, attachment-retained or lingual-plate-connected dentures. In addition, we
excluded patients who had not received conservative periodontal intervention or maintenance at least
once a year during the observation period. RPDs were provided by prosthodontists certificated by Japan
Prosthodontic Society. Periodontal maintenance was performed by dentists in the preventive or
periodontal departments. Data were gathered from the dental records, and patients were examined by
the attending prosthodontists at the time of RPD provision. These data included general and oral status,
and RPDs design.

Variables

Variables set as patient-related factors were: gender (male/ female), age (<65 or _65 years), lifestylerelated disease (with: having a medical history of at least one of hypertension, diabetes mellitus or
dyslipidemia/without: having no medical history of these diseases) and occlusal support (A +
B1/B2/B3/B4/C, based on the Eichner classification). Tooth-related factors were: jaw (upper/lower), type of
tooth (incisor/canine/premolar/molar), existing root canal treatment (with/without), pocket depth (PD: _3
mm/4 mm/ 5 mm/_6 mm), crown-root ratio (<1.0/1.01.5/_1.5) and type of abutment (direct: abutment in
contact with direct retainer/indirect: abutment in contact with indirect retainer). Root canal treatment
and crown-root ratio were determined from the radiographs taken at the time of prosthetic diagnosis.

Statistical analysis
KaplanMeier survival analysis was performed to show the survival curve of direct and indirect
abutments, as well as the other residual teeth. The survival distribution was then compared with a logrank test. P-values less than 0.05 were considered to be statistically significant. The Bonferroni correction
methods for counteracting the problem of multiple comparisons were used. Coxs proportional hazard
analysis was used to test bivariate and multivariate associations between each variable and the
abutment survival time. For the multivariate model, variables for which the bi variable p-value less than
0.25 were considered as prognostic variables by the stepwise backward selection (adoption criterion: p <
0.05, exception criterion: p < 0.10). Cases where data for the prognostic variables were missing were
deleted. We defined the entry-point as the date of provision of RPDs and the end-point as either the date
of the last visit to the hospital, which was treated as a censoring, or the date of abutment tooth loss
(defined as extraction of the tooth or changes to metal or resin coping of the over denture). Data were
analyzed using PASW Statistics 18 software (formerly SPSS; IBM Company, Tokyo, Japan).

Results
Demographics
One hundred and forty-seven patients satisfied the inclusion criteria and had been provided with a total
of 236 RPDs. The total numbers of RPD abutments were 856 and the study sample contained a further
1114 residual (non-abutment) teeth (Table 1).

Clinical outcomes
During the observation period, 13.7% of the abutments were lost (in contrast to 4.4% of non-abutment
teeth), including 17.9% of direct and 8.5% of indirect abutments.

Survival curves
The KaplanMeier survival curve is shown in Fig. 1. The 5-year survival rate was 95.8% for non-abutment
teeth, 93.1% for indirect abutments and 86.6% for direct abutments. In the log rank test, significant
differences between these three groups were observed by multiple comparison tests.

Bivariate analysis
Bivariate analysis by using Coxs proportional hazard modelling indicated that significant variables were
occlusal support, root canal treatment, pocket depth, crown-root ratio and type of abutment (Table 2). In
the case of occlusal support, because the statistical difference between A-B2 and B3-C was apparent, we
divided subjects into two groups and reanalyzed. For the same reason, in the section of pocket depth and
crown-root ratio, subjects were also divided into two groups, respectively. Results for these analyses are
also shown in Table 2.

Multivariate analysis
Variables selection for the multivariate model was performed by the backward selection technique. The
final Coxs proportional hazard model indicated that crown-root ratio (hazard ratio (HR): 3.13), root canal
treatment (HR: 2.93), pocket depth (HR: 2.51), type of abutments (HR: 2.19) and occlusal support (HR:
1.90), and were significant prognostic factors in the abutment survival period (Table 3).

Discussion
This longitudinal prospective cohort study indicates the expected survival trends of RPD abutments,
identifies several significant prognostic factors related to their survival and generates numeric hazard
ratio (HR) values to quantitatively estimate the extent to which these factors influence their survival.
These novel findings can help us to predict the survival prospects for abutment teeth at the time of
diagnosis based on their individual characteristics. The result of the KaplanMeier analysis showed that
the survival rate of abutment teeth was significantly lower than that of non-abutment teeth. A recent
clinical study, following 100 patients after periodontal therapy over 10 years, showed that 18% of RPD
abutment teeth were lost, compared with only 6% of non-abutment teeth. Other previous research has
suggested that being an abutment of RPDs was a significant risk factor for tooth loss. This is likely due to
the continuous and repetitive mechanical stress with which these teeth are loaded, the attendant higher
risk of damage to the periodontal tissue. It was also reported that the presence of RPD retainers can
contribute to deterioration in dental hygiene around abutment teeth. However, survival curves of RPD
abutments based on large number of longitudinal clinical cases have not previously been reported. Much
of the published clinical research evaluated only the frequency with which abutment teeth are lost, but

this type of censored data do not provide a good prognostic indication of tooth survival. The Kaplan
Meier method and Coxs proportional hazard regression analysis used in this study are representative
ways of performing survival analysis using the censoring. In the multivariate analysis of our data by using
Coxs proportional hazard regression model, we collected the objective information about potential
factors (both patient related and tooth-related), which can be evaluated easily and correctly by any
dentist. This multivariate regression analysis indicated several independently significant prognostic
factors. Occlusal support area was one of the significant prognostic factors. In a 28-year follow-up survey,
it was indicated that the crown-root ratio and the tooth survival time. PD is another reliable objective
index for evaluating periodontal condition. Matuliene et al. reported that, from multivariate analysis of
the association between PD and tooth loss, PD of 5 mm and over represented a significant risk factor,
compared with PD of 3 mm or less. The corresponding odds ratios in that study for PD = 4, 5, 6 and 7mm
and more were 1.6 (p = 0.034), 3.0 (p < 0.0001), 2.7 (p = 0.005) and 9.9 (p < 0.0001), respectively.
Consistent with this previous work, we showed that abutments with 5 mm PD and more were statistically
at higher risk of teeth loss. Our bivariate analysis showed that the HR of 4 mm PD compared with 3 mm
PD was 1.00 (p = 0.814), suggesting no difference in the risk of tooth loss.
The existence of previous root canal treatment also independently affected the survival time of abutment
teeth. So far it has been reported that the 4-year survival rate of 759 teeth following primary root canal
treatment were 95%. Conversely, in case of the 410 abutment teeth of RPD after root canal treatment,
the 5-year survival rate was only 51%. This difference suggests that a retainer for RPD might deteriorate
survival of endodontically treated teeth, rather than existence of root canal treatment itself. The
mechanical stress from the RPD must increase the risk of that tooth fracturing. There was a significant
difference of the survival rates between direct and indirect abutment teeth, indicating that the type of
abutments was also a prognostic factor. No existing reports have compared direct and indirect RPD
abutment teeth. RPD abutments continuously loaded the mechanical and bacterial stress from RPDs and
have a higher risk to damage the periodontal tissue than non-abutment teeth. Especially, direct
abutment teeth experience continuous and repetitive mechanical loading (including the rotational and
settling movements of the RPD) much more directly than indirect abutment teeth. This study also
calculated the HR of each factor. For instance, the HR of the type of abutment showed that indirect
abutments were likely to survive 2.19 times longer than direct abutments after controlling for other
factors. Importantly, by using Coxs hazard regression model, we can also calculate the individual
survival probability of the abutment teeth at any given point in time under each specific condition, giving
us previously unprecedented prognostic power. As the patients in this study were limited to those
attending a university hospital, and might therefore be a selective sample, it is possible that other
prognostic factors could arise in other trials, or that the quantitative differences determined here might
change. However, we included all the patients satisfying appropriate selection criteria over a 2-year
period: this type of continuous sampling minimizes any selection bias. All RPDs were provided by a
limited number of operators with advanced training in prosthodontics, after preparation of the mouth for
RPD treatment by suitably qualified dental professionals. The patients visited the hospital regularly for
periodontal maintenance, including scaling, root planning and tooth polishing, throughout the
observation period. Therefore, we are confident that the significant prognostic factors found in this
research are an accurate reflection of those affecting the survival of RPD abutments in general
population.

Conclusion
We conclude that RPD abutment teeth are more likely to be lost than other residual teeth. Occlusal
support, crown-root ratio, root canal treatment, pocket depth and type of abutment are related to the
survival time of RPD abutments. These results will help us to estimate the individualized risk and benefit
of the prosthodontic treatment for the residual teeth, to evaluate the prognosis of RPD abutments and
also to develop evidence-based dentistry in practice.

You might also like