Journal of Dentistry 153 (2025) 105539

Contents lists available at ScienceDirect

Journal of Dentistry
journal homepage: www.elsevier.com/locate/jdent

Review article

Factors associated with the success of restorative treatment for root caries:
A systematic review with meta-analysis
Yulin Wen , Xiaowei Zhao , Samantha Kar Yan Li , Edward Chin Man Lo , Chloe Meng Jiang *
Faculty of Dentistry, The University of Hong Kong, Hong Kong, China

A R T I C L E I N F O A B S T R A C T

Keywords: Objectives: The aims of this systematic review were to estimate the success rates of root caries restorations, and to
Root caries identify possible factors associated with the success of root caries restorations.
Restorative treatment Data and Sources: Literature search was conducted in three databases, PubMed, MEDLINE and Web of Science to
Systematic review
identify clinical studies reporting on the success of restorative treatment for root caries. Factors that may in­
Success rate
fluence clinical outcomes of the restorative treatment were summarized and analyzed.
Factor
Older adults Study selection: A total of 4116 records were identified, and 9 articles were eligible to be included. A total of 415
participants with 1174 root caries restorations were analyzed. The short-term (6-month) estimated success rates
(ESRs) (95 % CI) of root restorations with per protocol (PP) and intention-to-treat (ITT) approach were 95.1 %
(92.9 % to 97.3 %) and 90.5 % (86.8 % to 94.1 %), respectively. The long-term (24-month) ESRs were 82.2 %
(73.9 % to 90.6 %) and 73.8 % (64.0 % to 83.5 %) by using the PP and ITT approaches, respectively. Resin
composite restorations had a higher ESR compared to glass ionomer cement restorations at 12-month follow-up
(96.3 % vs. 86.9 %, p = 0.037), but no significant differences were found at 6-month and 24-month follow-ups (p
> 0.05). Besides filling material, no other factors with statistically significant association with treatment success
could be identified, including study setting, material mixing method, moisture control method, use of liner or
not, and restorative technique.
Conclusion: The overall success rate of root caries restorations is good in the short-term (6-month) but decreases
over time, with an approximately 20 % failure rate after two years. Complete/partial loss of restoration and gross
marginal defect were the main reported reasons accounted for failure. No studied factor has clear evidence to
support its association with long-term (24 months or longer) success of root caries restoration.
Clinical Significance: This review provides up-to-date evidence on the overall success rates of root caries resto­
rations. No compelling evidence shows any studied factor is related to the long-term success of root caries
restoration, whilst there is a tendency to observe higher success rates of resin composite compared to glass
ionomer cement restorations.

1. Introduction physical and mental capacity, as well as growing risk of diseases. Oral
health, as an integral part of general health, is an essential component of
Population ageing is a global phenomenon and the increasing num­ overall health and well-being [2]. Under such context, it is worth
ber of older adults bring challenges to the healthcare system, such as investigating how to provide appropriate dental care for older adults to
amplified healthcare needs and shortage of workforce. As reported by promote their oral health.
the World Health Organization (WHO), by year 2030, one in every six Root caries is any carious lesion on dental root surfaces either cavi­
people worldwide will be aged over 60 years, and the number of older tated or non-cavitated, which usually involves both cementum and
people (aged 60 or more) will increase from 1 billion in 2020 to 1.4 dentine while excluding the adjacent enamel [3]. Similar to coronal
billion in 2030. By year 2050, the proportion of the world’s older pop­ caries, root caries is a biofilm-mediated multifactorial dynamic disease
ulation over 60 years will nearly double from 12 % in 2015 to 22 % in that results in a net loss of dental hard tissues [4]. Root caries can be
2050 [1]. With biological changes, ageing leads to gradual decrease in associated with pain, hypersensitivity, compromised oral function, tooth

* Corresponding author at: 3F, The Prince Philip Dental Hospital, 34 Hospital Road, Sai Ying Pun, Hong Kong, China.
E-mail address: [email protected] (C.M. Jiang).

https://doi.org/10.1016/j.jdent.2024.105539
Received 25 October 2024; Received in revised form 13 December 2024; Accepted 19 December 2024
Available online 20 December 2024
0300-5712/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Fig. 1. The flow diagram of the review.

Table 1
Main characteristics and profiles of the included studies.
Author [Ref] Year Location Study design Follow-up Outcome measure Sample size (female/ Age range No. of
period male) (mean) restorations

Hu [25] 2005 mainland NRSI 24 months ART criteria 15 (7/8) 37~76 (63) 146
China
Lo [23] 2006 Hong Kong RCT, parallel 12 months Modified USPHS+ ART criteria 103 (72/31) N.R. (78.6) 162
Gil-Montoya 2014 Spain RCT, parallel 24 months ART criteria 28 (14/14) >65 (81.5) 81
[27]
Hayes [24] 2016 Ireland RCT, three- 12 months Modified USPHS 85 (31/54) ≥65 (71.9) 305
arms
Cruz Gonzalez 2016 Colombia RCT, parallel 6 months Modified criteria based on 64 (32/32) 60~101 (74.9) 174
[21] former studies
Koc Vural [28] 2020 Turkey RCT, split- 60 months Modified Havemann 39 (25/14) 18~67 (39.6) 100
mouth
Koc Vural [29] 2021 Turkey RCT, split- 36 months Modified USPHS 33 (23/10) 37~89 (52.6) 110
mouth
Mahmoud [22] 2022 Egypt RCT, parallel 12 months Modified USPHS 22 (N.R.) ≥50 (62.5) 44
Ratnayake [26] 2022 New Zealand RCT, split- 6 months Modified USPHS 26 (10/16) ≥18 (N.R.) 52
mouth

NRSI, Non-randomised study of intervention; RCT, randomized controlled trial; N.R., not reported.

fracture and even tooth loss. However, root caries is distinct from cor­ enamel. Thus, root surfaces are more prone to demineralize in com­
onal caries due to the differences in dental hard tissues. The mineral in parison to tooth crown [5]. Besides, there are differences between root
dentine, which contain elevated carbonate content, is more soluble than and coronal caries in terms of microbiology, critical pH for

2
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Table 2
Main characteristics of the placed restorations in the included studies.
Study subgroup Material Mixing method Study setting Auxiliary moisture control* Lining Technique

Hu, 2005a GIC Encapsulated Clinic No No Conventional treatment

Hu, 2005b GIC Encapsulated Clinic No No ART
Lo, 2006a RMGIC N.R. Outreach Yes No Conventional treatment
Lo, 2006b GIC Hand mixing Outreach Yes No ART
Gil-Montoya, 2014 RMGIC N.R. Outreach No No ART
Hayes, 2016a GIC Encapsulate Clinic Yes No Conventional treatment
Hayes, 2016b RMGIC Encapsulated Clinic Yes No Conventional treatment
Cruz Gonzalez, 2016a RMGIC Encapsulated Outreach Yes Yes Conventional treatment
Cruz Gonzalez, 2016b RMGIC Encapsulated Outreach Yes Yes ART
Koc Vural, 2020a Resin Composite N/A Clinic Yes No Conventional treatment
Koc Vural, 2020b Resin Composite N/A Clinic Yes Yes Conventional treatment
Koc Vural, 2021a RMGIC Hand mixing Clinic Yes No Conventional treatment
Koc Vural, 2021b Resin Composite N/A Clinic Yes No Conventional treatment
Mahmoud, 2022 GIC Encapsulated Clinic Yes No Conventional treatment
Ratnayake, 2022 GIC Hand mixing Clinic N.R. No ART

N/A, not applicable; N.R., not reported.

GIC, Glass ionomer cement; RMGI, resin-modified glass ionomer; ART, atraumatic restorative treatment.
*Auxiliary moisture control - Yes: use of cotton roll plus other moisture control method; No: use of cotton roll only.
a, b, indicated each study group of the study.

demineralization, and histopathology of carious lesions [5]. Root caries based on available clinical evidence, and to identify associated factors.
is more prevalent in older adults who have gingival recession and root
exposure [6]. The reported prevalence of untreated root caries varied 2. Material and method
among populations in different places. It ranged from 8 % (Finland) to
74 % (Brazil) in community dwellers, and from 30 % (Hong Kong) to 96 The present review was reported following the Preferred Reporting
% (Vietnam) in institutionalized older adults [7]. With the anticipated Items for Systematic Reviews and Meta-Analysis (PRISMA) guideline
ageing of population, older adults are retaining more natural teeth, thus, [16]. The population, intervention, comparison and outcome (PICO)
treatment of root caries is expected to be a main challenge in dentistry. question was defined. Population (P) was adults with root caries lesions,
Various options are available to treat root caries lesions, using both intervention (I) was any restorative treatment for root caries lesion, and
non-restorative and restorative approaches, depending on individual’s outcome (O) was the success rate of root caries restoration at follow-up
conditions such as extent and cleanability of the lesion, and patient’s examinations. Since the included clinical studies had their own research
acceptance of different treatments. Non-invasive approaches, such as questions, there were no straightforward control groups among all the
high concentration (5000 parts per million) fluoride toothpaste or gel, included studies. Thus, comparing group (C) was not applicable in the
sodium fluoride varnish, chlorhexidine varnish, and silver diamine present review.
fluoride (SDF) have been shown to be effective in arresting root caries Literature search was conducted in three databases, PubMed, MED­
lesions [8,9]. Despite this, deeper or widely extended non-cleansable LINE and Web of Science by two independent researchers. The software
cavitated lesions may still need to be restored. However, there is no Covidence was used to manage the identified references. Clinical studies
consensus on how to restore root caries lesions in terms of filling ma­ published from January 2004 to April 2024 reporting on the success of
terial, restorative technique and case selection [10]. Glass ionomer restorative treatment of root caries were included in this review. Pub­
cement (GIC), resin-modified glass ionomer (RMGI), and resin com­ lications were excluded if (1) not being a clinical study, e.g., review, case
posite are commonly used filling materials for placing root caries res­ report and laboratory study; (2) not on relevant topic, e.g., coronal
torations. In vitro studies have shown that GIC and RMGIC have superior caries, non-carious cervical lesions, primary dentition, and non-
cariostatic effect and potential for secondary caries prevention, and restorative treatment; (3) main outcome data were missing, e.g., no
antimicrobial effects, while there is better tooth strengthening effect for data on the success of root caries restorations; (4) description of the
resin composite and adhesive systems [11]. Despite these laboratory lesions was unclear, e.g., “cervical lesions” or “non-occlusal caries”;
findings, there is insufficient clinical evidence to support any material to and/or (5) not written in English. The keywords used in the literature
be the most appropriate one for restoring root caries lesions [11,12], and search were [root caries] AND [restoration OR restorative OR filling OR
all materials have rather high failure rates [13]. In terms of operative treatment]. The search strategies are shown in Appendix 1.
technique, atraumatic restorative treatment (ART), which only involves The Newcastle-Ottawa Scale (NOS) was adopted to assess the risk of
the use of hand instruments and rarely needs local anaesthesia, can be a bias of the included studies [17]. It should be pointed out that, although
good treatment for root caries lesions. However, a recent review RCTs were included, the purpose of the present study was not to make
concluded that there was insufficient evidence to support any difference direct comparisons between the intervention and control groups.
regarding the restoration longevity between ART and conventional Instead, indirect comparison with subgroup analysis and
treatment [14]. meta-regression was employed to investigate the associations between
At present, no consensus has been achieved on how to restore root the studied clinical parameters (potential factors) and the treatment
caries lesions and what factors influence the success of root caries res­ outcome (success of root caries restorations) [18]. Each study group of
torations. It is speculated that several factors may affect the success rates an RCT was treated as one cohort group. The NOS, which was designed
of restorative treatment, such as lesion extension (single vs multiple for assessing the quality of cohort studies, focusing on three major do­
surfaces), location (buccal, lingual, or approximal surface), tooth mains, selection, comparability and outcomes, were selected as the
(anterior vs posterior, upper vs lower), moisture control method, and assessment tool of the included studies. The NOS uses a star system to
restorative technique [15]. Thus, the present review aimed to search for assess the quality in which each domain can earn stars based on specific
relatively recent evidence (reported within 20 years) to identify factors criteria met, and the star distribution between domains was assessed
that affect the success of root caries restorations. The objectives of the based on Agency for Healthcare Research and Quality (AHRQ).
review were to estimate the success rates of root caries restorations The main data extracted from the included studies were the number

3
 Y. Wen et al. Journal of Dentistry 153 (2025) 105539

of root caries restorations placed at baseline, and the number of total

restorations examined as well as successful restorations observed at each

(50.3, 73.6)
(58.5, 80.3)

(39.3, 61.9)

(73.3, 94.2)

(68.6, 91.4)
(73.3, 93.5)
(65.0, 88.2)
follow-up. The success rate of root restorations was calculated by

(64.0,
83.5)
dividing the number of successful root restorations by the total number
of root restorations. To assess possible impacts of missing data in the
​
​

​
​
​
​

​
​

​
​
included studies, both per protocol (PP) and intention to treat (ITT)

<0.001
83.6 %
analyses were employed in meta-analysis. The PP approach proposes to

73.8
62.5
70.3

50.6

86.0

82.0
85.5
78.2
ITT
24 months ESR (95 %CI) (%)

include only those participants who completed the intervention and

​
​

​
​
​
​

​
​
stayed in the assigned group without any major protocol deviation, and
(62.1, 85.3)
(67.0, 87.9)

(46.2, 70.2)

(82.1, 98.6)

(76.4, 96.4)
(81.1, 97.8)
(71.4, 92.9)
the ITT principle states that any participants are analyzed based on

(73.9,
protocol no matter they received the assigned intervention or dropped

90.6)
out [19]. In the present study, the PP approach assessed the success rate
​
​

​
​
​
​

​
​

​
​
of root caries restorations based on the participants who remained in the

<0.001
82.7 %
study and received follow-up examination, and those dropped out dur­

82.2
75.0
78.8

58.6

93.5

89.1
92.2
84.3
ing follow-up were excluded from analysis. While ITT approach esti­
PP

​
​

​
​
​
​

​
​
mated the success rate based on all participants regardless of dropping
(59.2, 80.0)

(73.3, 94.2)

(68.6, 91.4)
(75.5, 94.7)
(65.0, 88.2)

out, and it was assumed that all dropped out cases had restoration
failure. Besides, potential factors that may influence the success of root
(74.9,
87.2)
caries restorations, e.g. filling material, mixing method for GIC and
RMGIC (encapsulated or manually mixed), moisture control method, use
​
​
​
​

​
​
​
​

​
​

​
​ of cavity lining or not, study setting (clinic or field) and restorative
47.7 %
0.105
81.0
70.4

86.0

82.0
87.3
78.2

technique (ART or conventional treatment) were recorded for further

18 months ESR (95 %CI) (%)

ITT

analysis.
​
​
​
​

​
​
​
​

​
​

Data analysis was conducted by using the software Stata (StataCorp

(60.9, 81.7)

(82.1, 98.6)

(76.4, 96.4)
(83.8, 98.8)
(71.4, 92.9)

LLC, Version 16.1, Texas, USA). The overall estimated success rate (ESR)
(79.8,
94.2)

of root caries restorations at 6-month, 12-month, 18-month, and 24-

month follow-ups were pooled by meta-analysis using the metaprop
​
​
​
​

​
​
​
​

​
​

​
​

command to compute the study specific confidence intervals using exact

74.9 %
0.003

method and Wald confidence intervals for the pooled estimate [20].
87.2
72.2

93.5

89.1
94.1
84.3
PP

Heterogeneity was measured by I-squared statistic and also p-value of

​
​
​
​

​
​
​
​

​
​

homogeneity test (Cochran’s Q test) for the groups (or subgroups) with
95.1)
94.3)
78.7)
75.8)
80.0)

(69.9, 88.7)
(69.9, 88.7)

(78.2, 96.7)

(70.9, 92.8)
(84.9, 98.9)
(84.9, 98.9)

(65.1, 97.1)

more than two studies. Subgroup analysis and meta-regression were

(77.7,

conducted to investigate the potential factors in influencing the success

(79.3,
(78.2,
(58.0,
(53.8,
(59.2,

88.7)

rate of root caries restorations. The level of statistical significance was

​
​

​
​

set as 0.05.
<0.001
80.3 %
83.2
88.9
87.8
69.1
65.4
70.4

80.5
80.5

90.0

84.0
94.6
94.6

86.4
ITT

3. Result
12 months ESR (95 %CI) (%)
Estimated success rate (ESR) of root caries restorations at 6, 12, 18 and 24 months of follow-up.

​
​

​
95.1)
94.3)
97.4)
94.0)
81.7)

(73.4, 91.3)
(74.6, 92.2)

(85.5, 99.5)

(76.9, 96.5)
(89.9, 99.9)
(89.9, 99.9)

(65.1, 97.1)

Fig. 1 shows the literature search and screening process. A total of

4116 records were identified from the literature search. After removal of
(85.5,
(79.3,
(78.2,
(82.4,
(75.0,
(60.9,

93.7)

duplicates (n = 1237), title and abstract of the remaining 2879 records

were screened for eligibility. A total of 2851 records were excluded due
​
​

​
​

to not being a clinical study (n = 1250) and not reporting on restorative

<0.001
77.1 %
89.6
88.9
87.8
92.1
86.4
72.2

83.8
84.9

95.7

89.4
98.1
98.1

86.4

treatment of root caries (n = 1601). The remaining 28 potentially

PP

eligible articles were retrieved for full-text reading, and 19 articles were
​
​

excluded for various reasons. Finally, nine articles were included in the
(86.8, 94.1)
99.9)
99.2)
85.8)
83.6)
99.2)

(80.6, 95.4)
(79.0, 94.5)
(67.8, 85.0)
(59.5, 81.2)

(86.3, 99.5)
(90.3, 99.9)

(65.1, 97.1)
(69.7, 91.8)

present review, reporting on eight RCTs and one non-randomized study

p value indicates the heterogeneity across the included studies.
100.0)

100.0)

of intervention (NRSI). The majority of the included studies (n = 7) were

(92.5,
(88.6,
(67.0,
(63.2,
(89.6,

(92.9,

(93.5,

assessed as good quality, while two were evaluated as fair (Appendix 2).
​
​

Table 1 summarizes the main characteristics of the included studies.

<0.001
88.0 %

A total of 1174 root restorations in 415 adults were included in the

100.0

100.0

90.5
98.6
96.0
77.4
74.1
96.3

89.6
88.3
77.2
71.2

96.0
98.2

86.4
82.7
ITT

present review. The included studies were conducted in different loca­

6 months ESR (95 %CI) (%)

tions, including Colombia [21], Egypt [22], Hong Kong [23], Ireland
(92.9, 97.3)

[24], mainland China [25], New Zealand [26], Spain [27] and Turkey
99.9)
99.2)
99.1)
96.5)
99.7)

(80.6, 95.4)
(80.3, 95.3)
(85.1, 97.3)
(69.5, 89.9)

(86.3, 99.5)
(90.3, 99.9)

(65.1, 97.1)
(73.3, 94.2)

[28,29]. The follow-up periods varied from 6 months to 60 months. The

100.0)

100.0)
(92.5,
(88.6,
(87.6,
(80.7,
(91.3,

(92.9,

(93.5,

United States Public Health Service (USPHS) criteria was the most
commonly adopted criteria for assessing root restorations, and ART
​
​

criteria was also adopted by some studies. Several potential associated

<0.001
68.7 %
100.0

100.0

factors were reported by the included studies, including filling material

95.1
98.6
96.0
95.6
90.6
97.5

89.6
89.5
92.9
81.3

96.0
98.2

86.4
86.0
PP

(GIC, RMGIC, CR), mixing method of GIC/RMGIC (encapsulated vs hand

mixing), moisture control method (auxiliary vs cotton roll only), study
Koc Vural, 2020b

Koc Vural, 2021b

Koc Vural, 2020a

Koc Vural, 2021a

Ratnayake, 2022
Mahmoud, 2022

setting (clinic vs field), use of liner or not, and restorative technique

Hayes, 2016b
Hayes, 2016a
Gil-Montoya,

(ART vs conventional treatment) (Table 2). In conventional treatment,

Cruz, 2016b
Cruz, 2016a
Hu, 2004b
Hu, 2004a

Lo, 2006b
Lo, 2006a

the treatment procedure used powered rotary equipment and was

p value*
I-square
Overall
2013
Table 3

Study

different to that of ART which used hand instruments only.

Meta-analysis showed that the pooled ESRs of root caries restoration

4
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Fig. 2. Forest plot of ESR of root caries restorations with different filling materials at each follow-up period.

with the PP and ITT approaches were similar, with only a slightly higher reported reasons for restoration failure. Although individual studies
with the PP approach. The pooled ESRs (95 % CI) using the PP approach used different criteria to evaluate the root caries restorations at follow-
were 95.1 % (92.9 % to 97.3 %) and 82.2 % (73.9 % to 90.6 %) at 6- up examinations, restoration loss accounted for >80 % of all failures
month and 24-month follow-ups, respectively (Table 3). When using while marginal defects accounted for over 60 % of all failures. In addi­
the ITT approach, the ESRs (95 % CI) were 90.5 % (86.8 % to 94.1 %) tion, presence of secondary caries, severe marginal discoloration,
and 73.8 % (64.0 % to 83.5 %) at 6-month and 24-month follow-ups, replacement by crown restoration and tooth loss were also reported
respectively. The findings of meta-regression showed that resin com­ reasons accounted for restoration failure (Table 5).
posite restorations had a significantly higher pooled ESR compared to
GIC restorations at 12-month follow-up (96.3 % vs 86.9 %, p = 0.037), 4. Discussion
but no significant differences were found at 6-month and 24-month
follow-ups (p > 0.05) (Fig. 2). Besides filling material, no other factors In the present study, the overall pooled ESR of root caries restora­
with statistically significant association with treatment success could be tions was over 90 % at 6-month follow-up, but it decreased over time.
identified, including mixing method of GIC/RMGIC (encapsulated vs After two years, approximately 20 % of the restorations failed, which
hand mixing), moisture control method (auxiliary vs cotton roll only), may be considered unsatisfactory. These findings are consistent with
study setting (clinic vs field), liner or not, and restorative technique those reported in earlier studies [11,30]. Due to the paucity of available
(ART vs conventional treatment) (p > 0.05). Since the factors adhesive RCTs, we cannot conduct meta-analysis on direct pair-wise comparison
systems (self-etch vs total etch) and beveling of cavity margin (yes vs no) for each study factor. To investigate if the studied clinical characteristics
were only reported in two studies [27,28], subgroup analysis and (potential factors) were associated with the treatment outcome (success
meta-regression were not performed on these two factors. Table 4 shows of root caries restorations), we employed meta-regression analysis, as an
the ESRs of root caries restorations at 12-month follow-up according to extension to the traditional meta-analysis, in the current systematic re­
the analyzed factors. view. Similar to simple regressions, meta-regression aims to predict the
Besides, dislodgement of restoration (complete or partial loss of outcome variable (success rate of root caries restoration) by the values of
restoration) and gross marginal defect (over 0.5 mm) were the main explanatory variables (potential factors), so as to explore the

5
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Table 4 Table 5
Meta-regression and pooled estimated success rate (ESR) of root caries resto­ Summary of restoration failure types of the included studies.
rations at 12-month follow-up according to the analyzed factors. Author [Ref] Definition of restoration failure Failure type
Factor ESR (95 % CI) Coefficient (in p
Hu [25] • Present, but defect over 0.5 Restoration dislodgement
(%) %) value
mm in depth; repair needed. accounted for 82.8 % and
Material ​ ​ ​ ​ • Present, but at margin and/or marginal defects for 17.2 % of
- GIC 86.9 (83.1, # ​ wear of the surface of 1.0 mm or all failures. There were no
90.7) more in depth; repair needed. instances of unsatisfactory
- RMGIC 87.5 (77.3, 1.7 0.786 • Not present, restoration has restoration wear or recurrent
97.7) (almost) completely caries observed. Teeth with
- Resin composite 96.3 (93.3, 9.2 0.037 disappeared; treatment needed. three or more restored cervical
99.3) • Not present, because other surfaces accounted for 79.3 % of
Mixing method for GIC/ ​ ​ ​ ​ treatment has been performed all failures.
RMGIC for whatever reason.
- Hand mix 95.4 (92.4, # ​ • Tooth not presents for
98.5) whatever reason.
- Encapsulated 86.6 (82.8, − 7.8 0.174 Lo [23] Sound restorations and The main reasons for
90.3) restorations with marginal restoration failure were gross
Auxiliary moisture control* ​ ​ ​ ​ defect or wear <0.5 mm were marginal defect and loss of
- No 84.8 (80.2, # ​ classified as survived. retention.
89.4) Otherwise, restorations were
- Yes 90.9 (85.0, 9.2 0.191 classified as failure cases.
96.8) Gil-Montoya • Present, marginal defect over The main reason for restoration
Setting ​ ​ ​ ​ [27] 0.5 mm in depth (for whatever failure was marginal defect over
- Outreach 85.3 (80.5, # ​ reason), gross defect of >1.0 0.5 mm.
90.1) mm in depth, at any one site;
- Clinic 90.6 (84.6, 8.6 0.235 wear and tear over the majority
96.7) of the restoration and >0.5 mm
Technique ​ ​ ​ ​ in depth at the deepest point;
- Conventional treatment 91.1 (86.0, # ​ repair is needed.
96.3) • Not present, restoration has
- ART 82.6 (73.4, − 8.4 0.111 (almost) completely
91.8) disappeared; treatment is
needed.
P-value was derived from meta-regression. Hayes [24] Complete loss of the restoration, Completely or partly missing of
#, indicates reference group. fracture of the restoration the restoration was the main
GIC, glass ionomer cement; RMGIC, resin modified glass ionomer cement. exposing the base of the cavity, reason of failure the three study
*Auxiliary moisture control - Yes: use of cotton roll plus other moisture control the presence of recurrent caries, groups, i.e., Biodentine, Fuji IX
method; No: use of cotton roll only; ART, atraumatic restorative treatment. replacement of the restoration GP Extra and Fuji II LC groups.
with another restoration,
patient experiencing pain in the
heterogeneity among the results of included studies [18]. Findings of the tooth, or loss of the tooth.
meta-regression show a higher ESR of resin composite root caries Cruz • Not present. In the ART restorations group (n
restoration compared to GIC at 12-month follow-up, but not in favor of Gonzalez • Present, but marginal defect = 64), 19 % exhibited defects
[21] greater than 0.5 mm over 0.5 mm, 17 % presented
any filling materials at longer-term follow-up (>18 months). One
• Present, but presence of with secondary caries, and no
possible reason is the small number of studies included in the secondary caries cases of complete restoration
meta-analysis, and thus statistically significant differences cannot be loss were observed.
identified. It is recommended to implement high-quality studies with In the conventional restorations
group (n = 84), 3.5 % exhibited
long-term follow-up of at least 24 months, to observe the status of root
defects over 0.5 mm, 1 %
caries restorations. Despite this, in line with a recent review [30], it is presented with secondary
observed a tendency towards better performance of adhesively bonded caries, and 3.5 % of the
resin composite compared to GIC of root caries restorations. However, if restorations were completely
moisture control during treatment is compromised, especially for lost.
Koc Vural Debonding; extraction of the Replacement by crown
dependent older adults with limited cooperation, placement of GIC
[28] treated tooth; replaced by crown restoration was the most
restoration may be preferable compared to resin composite [31]. restoration common reason of failure,
Other than filling material, we could not identify any other factors followed by debonding and
associated with success of root caries restorations. Although encapsu­ tooth extraction.
Koc Vural • Retention: Explorer penetrates Loss of retention and severe
lated GIC/RMGICs exhibit better mechanical properties than their
[29] into crevice that is of a depth marginal discoloration
manually mixed versions [32], there is insufficient available clinical that expose dentine or base; accounted for the major failure
evidence to support the superiority of encapsulated GIC/RMGICs in Restoration fractured or flaked type.
restoring root caries lesions. In terms of the operative technique for off, or outstanding depressions
placing GIC root caries restorations, no significant difference was found unable to be repaired.
• Marginal adaptation: Strong
between ART and conventional treatment. This finding is consistent
negative step present in major
with a previous systematic review on a direct comparison between the parts of the margin, not
use of ART and conventional treatment for root caries lesions [14]. Thus, removable; Marginal base or
ART can be an option to restore root caries lesions where resources are dentine exposed.
limited, for example, in outreach dental services with no access to • Marginal discolouration:
Strong discolouration in major
electricity and water (dental handpiece and saliva ejector cannot be parts of the margin, not
used). removable. Discolouration at
Besides the above-mentioned factors studied in the present review, the margin penetrates towards
other factors which may influence the success of caries restorations [33, the pulpal end.
34], such as operator effect, periodontal and endodontic status of the (continued on next page)

6
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Table 5 (continued ) comparison in generating odds ratio (commonly used measure for
Author [Ref] Definition of restoration failure Failure type dichotomous variables), we treated each study group of RCTs and NRSI
as a cohort group to generate pooled ESRs of the different types of res­
• Presence of secondary caries.
• Postoperative sensitivity: Pain
torations. This is the reason that we employed the Newcastle-Ottawa
and replacement is necessary. Scale (NOS) to inspect the risk of bias of the included studies instead
Mahmoud • Partial loss and total loss of the Restoration loss and presence of of using the RoB 2 tool which focuses on a specific result from RCTs [37].
[22] restorations. secondary caries were the main It should be pointed out that we excluded some studies in this review
• Restoration fracture with reasons for failure.
because of their unclear inclusion criteria for root caries lesions. The
exposure of dentine base.
• Secondary caries. term cervical caries lesion was found to be used to refer to root caries
• Pain or tooth fracture. lesions [15]. Although the outdated, G.V. Black classification system is
Ratnayake • Present, marginal defect over The main reason for failure was commonly used and class V lesions are sometimes used to indicate root
[26] 0.5 mm in depth (for whatever gross marginal defects. caries lesion. However, it is argued that class V refers to the anatomical
reason) at any one site; repair is Furthermore, no secondary
needed. caries lesions were evident in
site of the lesion at cervical third of the tooth, but not necessarily on root
• Present, gross defect of >1.0 the test restorations (GIC plus surface. In fact, many cervical lesions are largely located in the cervical
mm in depth; repair is needed. chlorhexidine) after 6 months. third of the tooth crown if there is no gingival recession and root
• Present, wear and tear over the exposure. Neither class V nor cervical lesions are necessarily caries le­
majority of the restoration and
sions originated from root surfaces. Thus, we excluded those studies
>0.5 mm in depth at the deepest
point; repair is needed. with unclear definition of caries lesions, such as “class V restoration” and
• Not present, restoration has “carious cervical lesion”, and only studies with clear statement on root
(almost) completely caries restorations were included in this review.
disappeared; treatment is There are limitations of the present review. First, we only conducted
needed.
indirect comparisons among included studies as performing direct
comparisons between the specific intervention and control groups were
tooth, extension of caries lesion and patient-related risk factors, cannot not possible due to paucity of eligible studies. Thus, we can only describe
be investigated in the present review due to lack of studies and evidence. associations between study factors and outcomes, and the strengthen of
In fact, restorative treatment for root caries can be challenging, and very the evidence is weak. Second clustering effect (multiple restorations per
often, subgingival margin of the restoration is inevitable. In terms of patient) in the included studies was not adjusted. The unit of analysis in
moisture and tissue control, rubber dam is not always possible when the included studies was restoration but not person. The restorations in
restoring subgingival root caries lesions. The use of retraction cord may the same patient tend to be more similar than restorations in different
be considered to facilitate a better visual and operational field. Besides, patients, and this similarity, represented by the intraclass correlation
as for subgingival restorations, perio-restorative interface with inte­ coefficient (ICC), reduces precision and widen confidence intervals,
gration of the concept of biologic width should be emphasized [35]. compared with non-clustered samples. This should be taken into account
Poor placement of restorative margin within the biologic width is in study design or adjusted using appropriate statistical methods [38].
harmful to periodontal tissues and may also trap dental plaque. Third, long-term observations are lacking in this review. Majority of the
Appropriate techniques, surgically removing bone or orthodontically included studies provided follow-up data till 12 months, but a very small
extruding tooth, should be meticulously considered to correct biologic number of studies reported long-term observations. It is recommended
width violation by the margin of restoration. However, these factors to perform long-term evaluation (at least 24 months) in studies of root
were not reported in any included studies, and no attempts can be made caries restorations because information on long-term success of resto­
to study the issue at the moment. Apart from that, Gil-Montoya et al. rations is of high clinical importance to both patients and clinicians.
[27] identified several risk factors associated with restoration failure,
including a lower number of tooth brushings per day, absence of pros­ 5. Conclusion
thesis, posterior location of the tooth, and higher baseline plaque index.
All of these variables are directly or indirectly related to plaque control. The overall success rate of root caries restorations is good in the
However, Cruz Gonzalez et al. [21] found that at 6-month follow-up short-term (6-month), but decreases over time, with an approximately
failure of restorations was neither associated with oral hygiene nor the 20 % failure rate after two years. Complete/partial loss of restoration
presence of an antagonist tooth to the treated one. Therefore, more and gross marginal defect were the main reported reasons accounted for
well-designed clinical studies are needed to investigate the influence failure. No studied factor has clear evidence to support its association
and association of possible factors at the level of practice, patient, tooth, with long-term (24 months or longer) success of root caries restoration.
and lesion on the success of root caries restorations.
In the present systematic review, we tried to include all available CRediT authorship contribution statement
clinical evidence and not just limited to those from RCTs. NRSI is defined
as any quantitative study estimating the effectiveness of an intervention Yulin Wen: Writing – original draft, Investigation, Formal analysis,
that does not use randomization to allocate units [36]. Traditionally, Data curation. Xiaowei Zhao: Writing – review & editing, Supervision,
systematic reviews focus on RCTs because RCT is more likely than NRSI Formal analysis, Data curation. Samantha Kar Yan Li: Writing – review
to provide unbiased evidence on the differential effect of interventions. & editing, Supervision, Software, Formal analysis, Conceptualization.
Although some biases apply to both RCT and NRSI, the key advantage of Edward Chin Man Lo: Writing – review & editing, Validation, Super­
a high-quality RCT is its ability to estimate the causal relationship be­ vision, Methodology, Conceptualization. Chloe Meng Jiang: Writing –
tween an experimental intervention (relative to a comparator) and review & editing, Writing – original draft, Validation, Supervision,
outcome. Generally, NRSI are included when the research question Methodology, Formal analysis, Data curation, Conceptualization.
cannot be answered by available RCTs, and there are justifications of
including NRSI in a review [36]. In the present study, we found that Declaration of competing interest
available RCTs addressed the research question (factors associated with
the success of root caries restorations) indirectly and incompletely, and, None.
in addition, there is a paucity of evidence from RCTs. We decided to
include NRSI but with cautions in synthesis of results. Instead of direct

7
Y. Wen et al. Journal of Dentistry 153 (2025) 105539

Funding [18] J.J. Deeks, J.P.T. Higgins, et al., A. DG, Chapter 10: analysing data and undertaking
meta-analyses, in: JPT Higgins, J Thomas, J Chandler, M Cumpston, T Li, MJ Page,
et al. (Eds.), A. DG, Chapter 10: analysing data and undertaking meta-analyses,
NA Cochrane Handbook for Systematic Reviews of Interventions version 6.5 (updated
August 2024) (2023).
Supplementary materials [19] N. Porta, C. Bonet, E. Cobo, Discordance between reported intention-to-treat and
per protocol analyses, J. Clin. Epidemiol. 60 (7) (2007) 663–669.
[20] V.N. Nyaga, M. Arbyn, M. Aerts, Metaprop: a Stata command to perform meta-
Supplementary material associated with this article can be found, in analysis of binomial data, Arch. Public Health 72 (1) (2014) 39.
the online version, at doi:10.1016/j.jdent.2024.105539. [21] A.C. 1uz Gonzalez, D.J. Marin Zuluaga, Clinical outcome of root caries restorations
using ART and rotary techniques in institutionalized elders, Braz. Oral Res. 30 (1)
(2016).
References [22] R. Mahmoud, A. Ibrahim, A. ElZoghby, O. Shaalan, Clinical evaluation of Carbomer
compared with high viscosity glass ionomer in restoration of root caries in geriatric
[1] World Health Organization, Ageing and health, 2022. https://www.who.int/ patients: a randomized controlled trial, J. Inter. Oral Health 14 (2) (2022).
news-room/fact-sheets/detail/ageing-and-health. [23] E.C.M. Lo, Y. Luo, H.P. Tan, J.E. Dyson, E.F. Corbet, ART and conventional root
[2] M. Glick, D.M. Williams, D.V. Kleinman, M. Vujicic, R.G. Watt, R.J. Weyant, A new restorations in elders after 12 months, J. Dent. Res. 85 (10) (2006) 929–932.
definition for oral health developed by the FDI World Dental Federation opens the [24] M. Hayes, C. da Mata, S. Tada, M. Cole, G. McKenna, F.M. Burke, P.F. Allen,
door to a universal definition of oral health, Int. Dent. J. 66 (6) (2016) 322–324. Evaluation of biodentine in the restoration of root caries: a randomized controlled
[3] P. Hellyer, Root caries, Br. Dent. J. 231 (1) (2021) 32, 32. trial, JDR Clin. Trans. Res. 1 (1) (2016) 51–58.
[4] N.B. Pitts, D.T. Zero, P.D. Marsh, K. Ekstrand, J.A. Weintraub, F. Ramos-Gomez, [25] J.Y. Hu, X.C. Chen, Y.Q. Li, R.J. Smales, K.H. Yip, Radiation-induced root surface
J. Tagami, S. Twetman, G. Tsakos, A. Ismail, Dental caries, Nat. Rev. Dis. Primers 3 caries restored with glass-ionomer cement placed in conventional and ART cavity
(2017) 17030. preparations: results at two years, Aust. Dent. J. 50 (3) (2005) 186–190.
[5] N. Damé-Teixeira, C.C.F. Parolo, M. Maltz, Specificities of caries on root surface, [26] J. Ratnayake, A. Veerasamy, H. Ahmed, D. Coburn, C. Loch, A.R. Gray, K.M. Lyons,
root caries: from prevalence to therapy, S.Karger AG, 2017. N.C.K. Heng, R.D. Cannon, M. Leung, P.A. Brunton, Clinical and microbiological
[6] J. Zhang, D. Sardana, M.C.M. Wong, K.C.M. Leung, E.C.M. Lo, Factors associated evaluation of a chlorhexidine-modified glass ionomer cement (GIC-CHX)
with dental root caries: a systematic review, JDR Clin. Trans. Res. 5 (1) (2019) restoration placed using the atraumatic restorative treatment (ART) technique,
13–29. Materials (Basel) 15 (14) (2022).
[7] A.K.Y. Chan, M. Tamrakar, C.M. Jiang, E.C.M. Lo, K.C.M. Leung, C.H. Chu, [27] J.A. Gil-Montoya, R. Mateos-Palacios, M. Bravo, M.A. Gonzalez-Moles, R. Pulgar,
A systematic review on caries status of older adults, Int. J. Environ. Res. Public Atraumatic restorative treatment and Carisolv use for root caries in the elderly: 2-
Health 18 (20) (2021). year follow-up randomized clinical trial, Clin. Oral Investig. 18 (4) (2014)
[8] R.L. Slayton, O. Urquhart, M.W.B. Araujo, M. Fontana, S. Guzmán-Armstrong, M. 1089–1095.
M. Nascimento, B.B. Nový, N. Tinanoff, R.J. Weyant, M.S. Wolff, D.A. Young, D. [28] U.Koc Vural, S. Gokalp, A. Kiremitci, Effect of cavity lining on the restoration of
T. Zero, M.P. Tampi, L. Pilcher, L. Banfield, A. Carrasco-Labra, Evidence-based root surface carious lesions: a split-mouth, 5-year randomized controlled clinical
clinical practice guideline on nonrestorative treatments for carious lesions: a report trial, Clin. Oral Investig. 24 (2) (2020) 979–989.
from the American Dental Association, J. Am. Dent. Assoc. 149 (10) (2018) [29] U.Koc Vural, L. Kerimova, A. Kiremitci, Clinical comparison of a micro-hybride
837–849, e19. resin-based composite and resin modified glass ionomer in the treatment of
[9] A.K.Y. Chan, M. Tamrakar, C.M. Jiang, Y.C. Tsang, K.C.M. Leung, C.H. Chu, cervical caries lesions: 36-month, split-mouth, randomized clinical trial,
Clinical evidence for professionally applied fluoride therapy to prevent and arrest Odontology 109 (2) (2021) 376–384.
dental caries in older adults: a systematic review, J. Dent. 125 (2022) 104273. [30] J. Schmidt, S. Proesl, E. Schulz-Kornas, R. Haak, H. Meyer-Lueckel, G. Campus,
[10] M. Hayes, P. Brady, F.M. Burke, P.F. Allen, Failure rates of class V restorations in M. Esteves-Oliveira, Systematic review and network meta-analysis of restorative
the management of root caries in adults – a systematic review, Gerodontology 33 therapy and adhesive strategies in root caries lesions, J. Dent. 142 (2024) 104776.
(3) (2016) 299–307. [31] S. Paris, A. Banerjee, P. Bottenberg, L. Breschi, G. Campus, S. Doméjean,
[11] W. Tonprasong, M. Inokoshi, M. Shimizubata, M. Yamamoto, K. Hatano, K. Ekstrand, R.A. Giacaman, R. Haak, M. Hannig, R. Hickel, H. Juric, A. Lussi,
S. Minakuchi, Impact of direct restorative dental materials on surface root caries V. Machiulskiene, D. Manton, A. Jablonski-Momeni, R. Santamaria,
treatment. Evidence based and current materials development: a systematic F. Schwendicke, C.H. Splieth, H. Tassery, A. Zandona, D. Zero, S. Zimmer,
review, Jpn. Dent. Sci. Rev. 58 (2022) 13–30. N. Opdam, How to intervene in the caries process in older adults: a joint ORCA and
[12] M. Hayes, P. Brady, F.M. Burke, P.F. Allen, Failure rates of class V restorations in EFCD expert delphi consensus statement, Caries Res. 54 (5–6) (2020) 1–7.
the management of root caries in adults - a systematic review, Gerodontology 33 [32] L. Al-Taee, S. Deb, A. Banerjee, An in vitro assessment of the physical properties of
(3) (2016) 299–307. manually- mixed and encapsulated glass-ionomer cements, BDJ Open 6 (1) (2020)
[13] H. Meyer-Lueckel, V. Machiulskiene, Rodrigo A. Giacaman, How to intervene in 12.
the root caries process? Systematic review and meta-analyses, Caries Res 53 (6) [33] M. Laske, N.J.M. Opdam, E.M. Bronkhorst, J.C.C. Braspenning, M. Huysmans, Risk
(2019) 599–608. factors for dental restoration survival: a practice-based study, J. Dent. Res. 98 (4)
[14] G. Göstemeyer, C. da Mata, G. McKenna, F. Schwendicke, Atraumatic vs (2019) 414–422.
conventional restorative treatment for root caries lesions in older patients: meta- [34] M. Jiang, Y. Fan, K.Y. Li, E.C.M. Lo, C.H. Chu, M.C.M. Wong, Factors affecting
and trial sequential analysis, Gerodontology 36 (3) (2019) 285–293. success rate of atraumatic restorative treatment (ART) restorations in children: a
[15] R.J. Wierichs, E.J. Kramer, H. Meyer-Lueckel, Risk factors for failure of class V systematic review and meta-analysis, J. Dent. 104 (2021) 103526.
restorations of carious cervical lesions in general dental practices, J. Dent. 77 [35] B. Nugala, B.S. Kumar, S. Sahitya, P.M. Krishna, Biologic width and its importance
(2018) 87–92. in periodontal and restorative dentistry, J Conserv Dent 15 (1) (2012) 12–17.
[16] M.J. Page, J.E. McKenzie, P.M. Bossuyt, I. Boutron, T.C. Hoffmann, C.D. Mulrow, [36] B.C. Reeves, J.J. Deeks, J.P.T. Higgins, B. Shea, P. Tugwell, G.A. Wells, G. on behalf
L. Shamseer, J.M. Tetzlaff, E.A. Akl, S.E. Brennan, R. Chou, J. Glanville, J. of the cochrane non-randomized studies of interventions methods, including non-
M. Grimshaw, A. Hróbjartsson, M.M. Lalu, T. Li, E.W. Loder, E. Mayo-Wilson, randomized studies on intervention effects. Cochrane Handbook for Systematic
S. McDonald, L.A. McGuinness, L.A. Stewart, J. Thomas, A.C. Tricco, V.A. Welch, Reviews of Interventions, 2019, pp. 595–620.
P. Whiting, D. Moher, The PRISMA 2020 statement: an updated guideline for [37] Higgins J.P.T., Savović J., Page M.J., Elbers R.G., Sterne J.A.C., Chapter 8:
reporting systematic reviews, BMJ 372 (2021) n71. assessing risk of bias in a randomized trial, in: Higgins JPT, Thomas J, Chandler J,
[17] G.A. Wells, B. Shea, D. O’Connell, J. Peterson, V. Welch, M. Losos, P. Tugwell, The Cumpston M, Li T, Page MJ et al. (Eds.), Cochrane Handbook for Systematic
Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies Reviews of Interventions version 6.5. Cochrane, 2024. 2019.
in Meta-Analyses, 2014. https://www.ohri.ca/programs/clinical_epidemiology/o [38] M.S. Litaker, V.V. Gordan, D.B. Rindal, J.L. Fellows, G.H. Gilbert, Cluster effects in
xford.asp. a national dental PBRN restorative study, J.Dent. Res. 92 (9) (2013) 782–787.