Review Article
Diagnostic Accuracy of 5 Dental Pulp Tests:
A Systematic Review and Meta-analysis
Anshul Mainkar, DDS,* and Sahng G. Kim, DDS, MS†
Abstract
Introduction: The aim of this systematic review was to
investigate and compare the diagnostic accuracy
including sensitivity, specificity, adjusted accuracy,
A n accurate pulpal diag-
adjusted positive predictive value (PPV), and adjusted
negative predictive value (NPV) of cold pulp testing
(CPT), heat pulp testing (HPT), electric pulp testing
(EPT), laser Doppler flowmetry (LDF), and pulse oximetry
(PO). Methods: Three electronic databases were
searched from January 1964 to December 2016. True-
positive, false-positive, true-negative, and false-
negative values were extracted from data in each study.
Sensitivity, specificity, adjusted accuracy, adjusted PPV,
and adjusted NPV were calculated from those values, if
not presented. A random effects model was used to
calculate pooled estimates of sensitivity, specificity,
adjusted accuracy, adjusted PPV, and adjusted NPV.
Results: A total of 125 articles were identified, and
28 studies were included for the final review. The pooled
estimates of sensitivity for CPT, EPT, HPT, LDF, and PO
were 0.87, 0.72, 0.78, 0.98, and 0.97, respectively.
Those of specificity were 0.84, 0.93, 0.67, 0.95, and
0.95, respectively. Those of adjusted accuracy were
0.84, 0.82, 0.72, 0.97, and 0.97, respectively. For
adjusted PPV, they were 0.81, 0.89, 0.62, 0.94, and
0.94, respectively, and for adjusted NPV, they were
0.87, 0.80, 0.79, 1.00, and 0.99, respectively. Conclu-
sions: LDF and PO were the most accurate diagnostic
methods, and HPT was the least accurate diagnostic
method. EPT showed high accuracy when testing vital
teeth (specificity = 0.93) but low accuracy when assessing
nonvital teeth (sensitivity = 0.72). CPT had moderate ac-
curacy when evaluating vital (specificity = 0.84) and non-
vital (sensitivity = 0.87) teeth. (J Endod 2018;44:694–
702)
Key Words
Cold test, electric pulp test, heat test, laser Doppler
flowmetry, pulp test, pulse oximetry
From the *Division of Endodontics, University of Connecticut Health, Farmington, Connecticut; †Division of Endodontics, College of Dental Medicine, Columbia Uni-
versity, New York, New York.
Address requests for reprints to Dr Sahng G. Kim, Division of Endodontics, College of Dental Medicine, Columbia University, New York, NY 10032. E-mail address:
sgk2114@columbia.edu
0099-2399/$ - see front matter
Copyright ª 2018 American Association of Endodontists.
https://doi.org/10.1016/j.joen.2018.01.021
694 Mainkar and Kim
nosis, which is based
on the neurovascular
Significance
An understanding regarding the diagnostic accu-
racy of pulp testing methods helps clinicians to
changes of the pulp, is a
reach a correct diagnosis and choose the most
prerequisite for successful
effective treatment. This systematic review provides
endodontic treatment. The
quantitative analysis of the diagnostic accuracy of 5
most commonly used
pulp testing methods.
method for the diagnosis
of pulp conditions is pulp
sensibility testing such as cold pulp testing (CPT), heat pulp testing (HPT), and electric
pulp testing (EPT). Although pulp sensibility testing has been considered reliable by most
clinicians, it has inherent limitations such as its reliance on patients’ subjective responses
(1–3) and dentists’ interpretations (1, 3) and its inability to detect pulpal blood flow
(1–3). In order to overcome the shortcomings of pulp sensibility testing, laser Doppler
flowmetry (LDF) and pulse oximetry (PO) have emerged as true pulp vitality testing (3–
5). These pulp vitality testing methods can detect pulpal blood flow without relying on
the patients’ responses and are thought to provide more accurate pulp status (3–5).
However, they suffer from technical challenges such as patients’ head movement (6), non-
pulpal noise (7, 8), signal detection limits (9), and the need for custom-made probes (10).
Understanding the diagnostic accuracy of pulp testing methods would help clini-
cians to reach a correct diagnosis and choose the most effective treatment. Diagnostic
accuracy includes sensitivity, specificity, accuracy, positive predictive value (PPV), and
negative predictive value (NPV). Sensitivity and specificity describe the intrinsic ability of
a diagnostic test to correctly identify necrotic pulp and vital pulp, respectively. They are
independent of disease prevalence, which refers to the probability of disease in a spe-
cific population at a given time. On the other hand, accuracy, PPV, and NPV are
prevalence-dependent descriptors of a diagnostic test. Accuracy measures how accu-
rately a diagnostic test predicts and excludes pulp necrosis. PPV indicates the probabil-
ity of a tested tooth to truly have pulp necrosis when a test result is positive, whereas NPV
describes the probability of a tested tooth to truly have vital pulp when a test result is
negative. The predictive values, although they are not as often used to compare different
pulp testing methods as sensitivity and specificity, serve as a basis for assessing the diag-
nostic test results in the clinical setting.
For an accurate pulp diagnosis, clinicians should acquire a thorough knowledge
concerning the diagnostic values of each pulp test before its use. It is noteworthy that a
large variation exists regarding the diagnostic accuracy of different pulp testing methods
among individual studies. For example, Dastmalchi et al (10) showed that the sensitivity
of CPT was 53% and the specificity of HPT was 55%, whereas Villa-Chavez et al (11)
showed that the sensitivity of CPT was 88% and the specificity of HPT was 100%. Further-
more, there is a lack of consensus regarding the results of different pulp testing methods
JOE — Volume 44, Number 5, May 2018

(12, 13) and which pulp testing method provides the best diagnostic
accuracy (14). This may explain why clinicians often confront clinical
situations in which the result of a pulp test on a tooth contradicts those
of the other tests.
There have been a few reviews published on pulp testing methods
(15–17) with qualitative analysis. To date, no studies have provided a
comprehensive, quantitative analysis on which diagnostic reasoning can
be established. Therefore, the aim of this systematic review was to
investigate and compare the diagnostic accuracy including sensitivity,
specificity, accuracy, PPV, and NPV of CPT, HPT, EPT, LDF, and PO
through a meta-analysis.
Materials and Methods
Search Strategy
Three electronic databases (PubMed, Scopus, and Web of Science)
were searched from January 1964 to December 2016 using 26 key words.
The selected key words were ‘‘pulp test,’’ ‘‘pulp tester,’’ ‘‘pulp testing,’’
‘‘pulpal test,’’ ‘‘pulpal tester,’’ ‘‘pulpal testing,’’ ‘‘electric pulp test,’’ ‘‘electric
pulp tester,’’ ‘‘electric pulp testing,’’ ‘‘vitality test,’’ ‘‘vitality tester,’’ vitality
testing,’’ ‘‘thermal test,’’ ‘‘thermal tester,’’ ‘‘thermal testing,’’ ‘‘cold test,’’
‘‘cold tester,’’ ‘‘cold testing,’’ ‘‘heat test,’’ ‘‘heat tester,’’ ‘‘heat testing,’’ ‘‘pulse
oximeter,’’ ‘‘pulse oximetry,’’ ‘‘laser Doppler flowmetry,’’ ‘‘laser Doppler
flowmeter,’’ and ‘‘LDF.’’ The search results were narrowed down by sub-
ject. For PubMed, the journal category of ‘‘Dental journals’’ was selected.
For Web of Science, the research area of ‘‘Dentistry oral surgery medicine’’
was selected. For Scopus, the subject area of ‘‘Dentistry’’ was selected.
Eligibility Criteria
The inclusion criteria were as follows:
1. A clinical study
2. Sample size given
3. Pulp diagnosis of a study sample confirmed by histologic analysis,
direct clinical observation (access cavity), or evidence of root canal
filling (only to confirm nonvital teeth)
4. The study provides a numeric value of or enough data to calculate at
least one of the following for identifying the vitality of a tooth: sensi-
tivity, specificity, accuracy, PPV, or NPV
5. A prospective or retrospective cohort study or cross-sectional study
6. A human study
7. Tetrafluoroethane (EndoIce [Coltene Whaledent, Mahwah, NJ]), a
propane-butane mixture (EndoFrost [Roeko, Langenau, Germany] or
FriscoSpray [Ad-arztbedarf GmbH, Frechen, Germany]), carbon dioxide,
dichlorodifluoromethane, and ethyl chloride (for CPT); gutta-percha
and a rubber cup (for HPT); and EPT, LDF, and PO pulp testing methods
The exclusion criteria were as follows:
1. Not a clinical study
2. No sample size given or an unclear sample size
3. Pulp diagnosis confirmation criteria not met
4. Study does not provide a numeric value of at least one of the
following for identifying the vitality of a tooth: sensitivity, specificity,
accuracy, PPV, or NPV
5. A review article
6. An animal study
7. Naylor instrument (thermoelectric stimulator), heated ball
burnisher, and ice pulp testing methods
Study Selection
Studies were searched initially by title and abstract and then full
text for inclusion in the systematic review. Further studies were identi-
JOE — Volume 44, Number 5, May 2018
Review Article
fied by manually searching the online university library catalog using
select key words and assessing the references in the included studies.
The authors checked the eligibility of the studies and resolved any dis-
agreements by collective discussion.
Studies that confirmed the vitality of teeth by a lack of symptoms
and/or normal radiographic signs without further confirmation with
one of the methods in the inclusion criteria were excluded because
pulpal necrosis often occurs without signs of pain (18) or radiographic
changes (19).
Some studies included both ‘‘teeth that met the diagnosis confir-
mation criteria’’ and ‘‘those that did not.’’ In such studies, if the results
for the subset of teeth that met the diagnosis confirmation criteria were
available, that subset of teeth was included in the current review. If it was
not possible to separate study results between the 2 groups, the study
was excluded.
The Extraction and Synthesis of Data
Quantitative data were compiled from each study and converted
based on the same definitions of sensitivity, specificity, accuracy, PPV,
and NPV. In this study, a positive result of a pulp test is considered to
be the diseased state (nonvital), and a negative result is considered
to be the healthy state (vital). Therefore, sensitivity is the percentage
of nonvital teeth that are correctly identified as nonvital, specificity is
the percentage of vital teeth that are correctly identified as vital, accuracy
is the percentage of all teeth that are correctly identified, PPV is the per-
centage of teeth that give a positive result that are actually nonvital, and
NPV is the percentage of teeth that give a negative result that are actually
vital. Data from studies that used the reverse definition were converted
to match this definition.
Using quantitative data reported in each article, true-positive (TP),
false-positive (FP), true-negative (TN), and false-negative (FN) values
were extracted from each study using the following calculations (20):
TP ¼ Sensitivity  Number of nonvital teeth
FN ¼ Number of nonvital teeth  TP
TN ¼ Specificity  Number of vital teeth
FP ¼ Sample size  Number of nonvital teeth  TN
The calculated TP, FP, TN, and FN were rounded to the nearest
whole number because it is not possible to have a portion of a tooth.
These rounded values were considered raw values. The raw values
were then used to calculate accuracy, PPV, and NPV using formula
definitions (11, 20) if they were not reported in the study. For
example, if a study reported the sensitivity and specificity, the data
reported in the study were used to determine raw TP, FP, TN, and
FN values, as shown in the calculation provided earlier. The raw
values were then used to calculate accuracy, PPV, and NPV using
the formulas. The validity of the raw values was confirmed by
comparing the calculated values with any reported values in the
studies. Studies were excluded from the review if a discrepancy in
data was noted.
Accuracy, PPV, and NPV can only be directly compared between
studies if the disease prevalence of both samples is the same. To facil-
itate comparison, accuracy and predictive values were standardized to
the total disease prevalence (42.8%) of all the studies included in the
systematic review using the following formulas (Prev: indicates disease
prevalence):
Diagnostic Accuracy of 5 Dental Pulp Tests 695
Review Article
Sensitivity  Prev: þ Specificity  ð1  Prev:Þ
Adj: Accuracy ¼
Sensitivity  Prev: þ ð1  SpecificityÞ  ð1  Prev:Þ þ Specificity  ð1  Prev:Þ þ ð1  SensitivityÞ  Prev:
Sensitivity  Prev:
Adj: PPV ¼
Sensitivity  Prev: þ ð1  SpecificityÞ  ð1  Prev:Þ
Specificity  ð1  Prev:Þ
Adj: NPV ¼
Specificity  ð1  Prev:Þ þ ð1  SensitivityÞ  Prev:
The total disease prevalence was calculated by summing the number
of nonvital teeth and dividing by the total number of teeth in all the
included studies. The number of teeth from each study was only counted
once even if the study included multiple pulp tests.
Studies with a sample of 100% nonvital teeth cannot have a valid
PPV (because the FP will always be 0) or NPV (because the TN will
always be 0). Similarly, studies with a sample of 100% vital teeth
Figure 1. A flowchart of the study selection.
696 Mainkar and Kim
cannot have a valid PPV (because the TP will always be 0) or NPV
(because the FN will always be 0). Therefore, NPV and PPV were
not included for any studies with a sample of 100% vital or nonvital
teeth. Furthermore, studies with either a sample of 100% vital or non-
vital teeth do not permit the calculation of adjusted accuracy, adjusted
PPV, or adjusted NPV because of a 0 in the denominator for the
‘‘TP + FN’’ (in the denominator of sensitivity) or ‘‘TN + FP’’ (in the
denominator of specificity) components of the adjustment equations,
respectively. The pooled estimates of sensitivity, specificity, accuracy,
PPV, and NPV were generated by the meta-analysis from the previously
described 5 diagnostic values obtained or calculated from the
included studies using a random effects model to account for the var-
iations between the studies. I2 was used to test the heterogeneity of the
studies. Comprehensive Meta-Analysis software (Version 2; Biostat,
Englewood, NJ) was used for all analyses.
JOE — Volume 44, Number 5, May 2018

Appraisal of Study Quality
The qualities of all included studies were assessed by using the
modified Quality Assessment of Diagnostic Accuracy Studies tool
(21). The studies were rated high, moderate, or low according to study
design, random or consecutive sample selection, blinding of the pulp
tester to the results of other tests, blinding of the pulp tester from signs
and symptoms, test description details, sample size, and diagnostic ac-
curacy presented as sensitivity and specificity.
Results
Electronic and manual searching generated 42,562 articles, of
which 125 articles were included for full-text review for relevance after
title and abstract screening. A total of 30 studies met the inclusion
criteria and were selected for data extraction. Two studies were
excluded because of the discrepancies between the calculated diag-
nostic accuracy values from raw TP, TN, FP, and FN values and the
reported values in the studies (13, 22). For the final review and
meta-analysis, 28 studies were included (Fig. 1). Generally, high hetero-
geneity (I2 > 50) was identified among all studies except those for LDF.
A random effects model was used to account for the variations among
studies.
The pooled diagnostic accuracy values of CPT, EPT, HPT, LDF,
and PO were obtained from the raw TP, TN, FP, and FN values for
each study (Tables 1 and 2). A summary of the pooled diagnostic ac-
curacy values is presented in Table 3. For the pooled sensitivity, LDF
and PO showed the highest values (0.975 for LDF and 0.973 for PO)
without a statistically significant difference (P > .05) followed by CPT
(0.867), HPT (0.778), and EPT (0.720). These 3 pulp sensibility tests
were statistically significantly different from LDF in pooled sensitivity
(P < .05), but CPT and HPT were not statistically different from PO
(P > .05). Among the pulp sensibility tests, CPT showed statistically
higher sensitivity than EPT and HPT (P < .05). For the pooled spec-
ificity, LDF and PO also showed the highest values (0.950 for LDF and
0.954 for PO, P > .05), which were statistically different from CPT
(0.843) and HPT (0.665). EPT showed no statistical difference
from LDF and PO in pooled specificity.
The values of accuracy, PPV, and NPV cannot be used for compar-
ison because the disease prevalence varies among the individual studies
and affects the values of accuracy, PPV, and NPV. Therefore, adjusted
accuracy, adjusted PPV, and adjusted NPV were calculated with the total
disease prevalence (42.8%) of all included studies in order to compare
the diagnostic accuracy of the 5 pulp testing methods. LDF and PO
showed significantly higher pooled estimates of adjusted accuracy
and adjusted NPV than all pulp sensibility tests (P < .05). No statistically
significant differences were found among the pulp sensibility tests in the
pooled values of adjusted accuracy or adjusted NPV. For the adjusted
PPV, the pooled value of EPT was identified to be not statistically
different from those of LDF, PO, and CPT (P > .05) and higher than
that of HPT (P < .05). LDF was not statistically different from PO in
adjusted accuracy, adjusted PPV, and adjusted NPV (P > .05).
All of the studies were rated low in the quality of the study except 2
studies, of which 1 study (3) was determined to be of moderate quality
and the other (11) to be of high quality based on the modified Quality
Assessment of Diagnostic Accuracy Studies criteria (Fig. 2). For most
studies, the low-quality categorization was because of no or unclear
blinding of the pulp tester to the other pulp tests and from signs and
symptoms. Only 2 studies incorporated blinding into the protocol (3,
11). Only 2 studies met the random or consecutive sample selection
(11, 39). Of these 2 studies, 1 study (39) was determined to be of
low quality because of no blinding to the other test results. The other
(11) met all criteria and was rated high. The study of moderate quality
JOE — Volume 44, Number 5, May 2018
Review Article
(3) met all criteria except the random or consecutive sample selection.
The pulp diagnosis was confirmed clinically (using endodontic access)
in 17 studies (3, 4, 10–12, 24, 27, 28, 31, 32, 34–38, 40, 41),
histologically in 7 studies (23, 24, 27, 31, 34, 40, 41), and by the
presence of root filling in 8 studies (3, 5, 24, 25, 29, 34, 41, 42).
Four studies used 2 confirmation methods (endodontic access and
histology or endodontic access and the evidence of root filling) (23,
24, 34, 41) (Tables 1 and 2).
Discussion
This systematic review investigated and compared the sensitivity,
specificity, adjusted accuracy, adjusted PPV, and adjusted NPV of 5
dental pulp tests including CPT, EPT, HPT, LDF, and PO. There have
been many quantitative systematic analyses for the diagnostic accuracy
of a multitude of diagnostic tests used in the medical field using the same
method as in this investigation (43, 44). Two systematic reviews have
been published about the diagnostic accuracy of pulp testing
methods (14, 15), but no study has quantitatively determined the
diagnostic accuracy of the pulp testing methods. The pooled values
presented in the current review may provide clinicians with a more
accurate understanding of the sensitivity, specificity, adjusted
accuracy, adjusted PPV, and adjusted NPV of each pulp testing method.
Different definitions of ‘‘positive’’ and ‘‘negative’’ have been used in
studies. Some studies refer to ‘‘positive’’ as the disease state (necrosis)
and ‘‘negative’’ as the healthy state (vital) (3, 5, 11, 24, 27, 31, 32, 35,
38, 40), which agrees with the definition in our study. However, others
define ‘‘positive’’ as the healthy state and ‘‘negative’’ as the diseased state
(4, 10, 12, 23, 25, 28, 29, 30, 33, 34, 36, 37, 39, 41, 42, 45). For this
reason, TP, TN, FP, and FN values were converted based on our
definition for the quantitative analysis. For 2 histologic studies that
did not specify them (26, 46), the descriptors referring to the
presence of any necrotic tissue in the pulp are considered ‘‘positive’’
and the other descriptors as ‘‘negative.’’ Therefore, clinicians should
identify what value (positive or negative) is used to describe disease
when they review diagnostic accuracy from scientific articles in order
to avoid misinterpretation of sensitivity or specificity, PPV, and NPV.
Sensitivity and specificity are most commonly used to compare
different diagnostic methods because they represent the inherent testing
ability (15–17). In our systematic review, 27 studies provided sensitivity
and specificity values, but only 13 studies provided predictive values.
However, in the clinical settings, predictive values can be more useful
to clinicians because they represent the probability of a test to
provide the correct diagnosis (11). PPV is the probability that teeth
with a positive test result (no response to a test) are truly nonvital,
and NPV is the probability that teeth with a negative test result (response
to a test) are truly vital. From a clinician’s perspective, if the test shows
high predictive values, PPV provides a basis for performing the end-
odontic treatment relying on the high probability of pulp necrosis in
the positively tested tooth, whereas NPV provides a rationale for avoiding
the unnecessary treatment by ruling out pulp necrosis in the negatively
tested tooth.
In this systematic review, adjusted PPV and adjusted NPV as well as
adjusted accuracy were investigated because, unlike sensitivity and
specificity, their values are affected by disease prevalence (11, 20).
For example, PPV increases if the disease prevalence of a population
increases, and NPV increases if the disease prevalence of a
population decreases. The shift in accuracy depends on if the pulp
test is more accurate with vital or nonvital teeth. In this study, the
predictive values were based on a disease prevalence of 42.8%,
which was calculated from all of the included studies and was also
within the range of previous studies (39%–45%) (11, 35). The
Diagnostic Accuracy of 5 Dental Pulp Tests 697
 698

Review Article
TABLE 1. Diagnostic Accuracy of Pulp Sensibility Tests in the Included Studies
Adjusted Adjusted Adjusted Confirmation
Study for CPT Testing tool N TP FP FN TN Sensitivity Specificity accuracy PPV NPV method Quality
Mainkar and Kim

Bruno et al, 2009 (23) Propane-butane mixture 20 17 0 3 0 0.850 — X X X Histologic, Low

(EndoFrost) clinical
Chen and Abbott, 2011 (24)* Propane-butane mixture 184 13 15 2 154 0.920 0.910 0.914 0.884 0.938 Clinical, RCF Low
(EndoFrost)
Chen and Abbott, 2011 (24)* CO2 184 14 5 1 164 0.950 0.970 0.961 0.960 0.963 Clinical, RCF Low
Cooley and Robison, 1980 (25) Dichlorotetrafluoroethane 2 2 0 0 0 1.000 — X X X RCF Low
Dastmalchi et al, 2012 (10) Propane-butane mixture 24 6 7 3 8 0.660 0.530 0.586 0.512 0.676 Clinical Low
(FriscoSpray)
Dummer et al, 1980 (26) Ethyl chloride 75 16 14 16 29 0.500 0.670 0.597 0.531 0.642 Histologic Low
Evans et al, 1999 (27) Ethyl chloride 53 49 0 4 0 0.920 — X X X Clinical Low
Farid et al, 2015 (28) Tetrafluoroethane 75 22 8 3 42 0.880 0.840 0.857 0.805 0.903 Clinical Low
Fuss et al, 1986 (29) CO2 10 10 0 0 0 1.000 — X X X RCF Low
Fuss et al, 1986 (29) Dichlorodifluoromethane 10 10 0 0 0 1.000 — X X X RCF Low
Fuss et al, 1986 (29) Ethyl chloride 10 10 0 0 0 1.000 — X X X RCF Low
Garfunkel et al, 1973 (30) Ethyl chloride 109 28 21 19 41 0.595 0.660 0.632 0.567 0.685 Histologic Low
Gopikrishna et al, 2007 (3) Tetrafluoroethane (EndoIce) 80 34 3 8 35 0.810 0.920 0.873 0.883 0.866 Clinical Moderate
Hori et al, 2011 (31) EndoFrost 56 11 12 4 29 0.733 0.707 0.718 0.652 0.780 Clinical Low
Kamburoglu and Paksoy, 2005 (32) Propane-butane mixture 93 40 1 3 49 0.930 0.980 0.959 0.972 0.949 Clinical Low
Moody et al, 1989 (33) Ethyl chloride 50 9 7 3 31 0.750 0.815 0.787 0.752 0.813 Histologic Low
Peters et al, 1994 (34) CO2 95 89 0 6 0 0.937 — X X X Clinical, RCF Low
Petersson et al, 1999 (35) Ethyl chloride 59 24 3 5 27 0.828 0.900 0.869 0.861 0.875 Clinical Low
Saeed et al, 2011 (36) Propane-butane mixture (EndoIce F) 74 71 0 3 0 0.960 — X X X Clinical Low
Saeed et al, 2011 (36) Tetraflouroethane (Green 74 68 0 6 0 0.920 — X X X Clinical Low
EndoIce [Coltene Whaledent,
Cuyahoga Falls, OH])
Villa-Chavez et al, 2013 (11) TFE 110 44 1 6 60 0.880 1.000 0.949 1.000 0.918 Clinical High
Weisleder et al, 2009 (12) CO2 150 57 21 7 65 0.890 0.760 0.816 0.735 0.902 Clinical Low
Weisleder et al, 2009 (12) Tetrafluoroethane (EndoIce) 150 59 21 5 65 0.920 0.760 0.828 0.742 0.927 Clinical Low
Pooled diagnostic accuracy values 0.867 0.843 0.840 0.807 0.871
95% CI 0.810–0.909 0.773–0.895 0.769–0.893 0.722–0.871 0.808–0.915
Heterogeneity (I2) 86.826 87.265 88.346 89.927 86.975
Study for EPT
Bruno et al, 2009 (23) Electric pulp tester 20 10 0 10 0 0.500 — X X X Histologic, Low
clinical
Chen and Abbott, 2011 (24)* Electric pulp tester 184 11 3 3 166 0.790 0.980 0.899 0.967 0.862 Clinical, RCF Low
Cooley and Robison, 1980 (25) Electric pulp tester 2 2 0 0 0 1.000 — X X X RCF Low
Cooley and Lubow, 1984 (42) Electric pulp tester 30 28 0 2 0 0.930 — X X X RCF Low
Dastmalchi et al, 2012 (10) Electric pulp tester 24 2 6 7 9 0.220 0.600 0.437 0.292 0.507 Clinical Low
JOE — Volume 44, Number 5, May 2018

Dummer et al, 1980 (26) Electric pulp tester 75 4 0 28 43 0.125 1.000 0.625 1.000 0.604 Histologic Low
Evans et al, 1999 (27) Electric pulp tester 53 46 0 7 0 0.870 — X X X Clinical Low
Farid et al, 2015 (28) Electric pulp tester 75 22 9 3 41 0.880 0.820 0.846 0.785 0.901 Clinical Low
Fuss et al, 1986 (29) Electric pulp tester 10 10 0 0 0 1.000 — X X X RCF Low
Gopikrishna et al, 2007 (3) Electric pulp tester 80 30 3 12 35 0.710 0.920 0.830 0.869 0.809 Clinical Moderate
Hori et al, 2011 (31) Electric pulp tester 55 12 3 3 37 0.800 0.925 0.871 0.889 0.861 Clinical Low
Ingolfsson et al, 1994 (4) Electric pulp tester 11 7 0 4 0 0.636 — X X X Clinical Low
Johnson et al, 1970 (39) Electric pulp tester 361 20 4 15 322 0.571 0.988 0.810 0.972 0.755 Histologic Low
Kamburoglu and Paksoy, 2005 (32) Electric pulp tester 93 36 2 7 48 0.837 0.960 0.907 0.940 0.887 Clinical Low
Karayilmaz and Kirzioglu, 2010 (5) Electric pulp tester 59 54 0 5 0 0.915 — X X X RCF Low
JOE — Volume 44, Number 5, May 2018

Kontakiotis et al, 2015 (40) Electric pulp tester 11 11 0 0 0 # # X X X Clinical Low

Moody et al, 1989 (33) Electric pulp tester 50 10 11 2 27 0.830 0.710 0.761 0.682 0.848 Histologic low
Peters et al, 1994 (34)† Electric pulp tester 95 53 0 42 0 0.558 — X X X Clinical, RCF Low
Peters et al, 1994 (34)† Electric pulp tester 95 53 0 42 0 0.558 — X X X Clinical, RCF Low
Petersson et al, 1999 (35) Electric pulp tester 59 21 3 8 27 0.720 0.900 0.823 0.843 0.811 Clinical Low
Reynolds et al, 1966 (45) Electric pulp tester 56 2 0 1 53 0.670 1.000 0.859 1.000 0.802 Histologic Low
Saeed et al, 2011 (36) Electric pulp tester 74 72 0 2 0 0.970 — X X X Clinical Low
Seltzer et al, 1963 (46) Electric pulp tester 129 15 7 13 94 0.536 0.931 0.762 0.853 0.728 Histologic Low
Shahi et al, 2015 (41) Electric pulp tester 70 22 6 7 35 0.759 0.854 0.813 0.795 0.825 Clinical, RCF Low
Villa-Chavez et al, 2013 (11) Electric pulp tester 110 38 0 12 60 0.760 1.000 0.897 1.000 0.848 Clinical High
Weisleder et al, 2009 (12) Electric pulp tester 150 48 7 16 79 0.750 0.920 0.847 0.875 0.831 Clinical Low
Pooled diagnostic accuracy values 0.720 0.928 0.817 0.888 0.804
95% CI 0.647–0.783 0.877–0.959 0.770–0.856 0.818–0.934 0.759–0.842
Heterogeneity (I2) 88.971 84.836 76.355 88.237 73.814
Study for HPT
Bruno et al, 2009 (23) Heated gutta-percha 20 18 0 2 0 0.900 — X X X Histologic, Low
clinical
Dastmalchi et al, 2012 (10) Rubber cup 24 5 6 4 9 0.550 0.600 0.579 0.507 0.640 Clinical Low
Dummer et al, 1980 (26) Heated gutta-percha 75 24 28 8 15 0.750 0.350 0.521 0.463 0.652 Histologic Low
Garfunkel et al, 1973 (30) Heated gutta-percha 109 26 17 21 45 0.550 0.720 0.647 0.595 0.681 Histologic Low
Hori et al, 2011 (31) Heated gutta-percha 55 13 10 2 30 0.867 0.750 0.800 0.722 0.883 Clinical Low
Petersson et al, 1999 (35) Heated gutta-percha 59 25 13 4 17 0.860 0.570 0.694 0.600 0.845 Clinical Low
Villa-Chavez et al, 2013 (11) Heated gutta-percha 110 43 0 7 60 0.860 1.000 0.940 1.000 0.905 Clinical High
Pooled diagnostic accuracy values 0.778 0.665 0.723 0.619 0.785
95% CI 0.647–0.869 0.485–0.807 0.578–0.833 0.490–0.733 0.670–0.868
Heterogeneity (I2) 85.755 88.040 86.973 77.317 82.268
—, either 100% vital or nonvital teeth (predictive values are not valid); CI, confidence interval; FN, false negative; FP, false positive; NPV, negative predictive value; PPV, positive predictive value; RCF, root canal filling; TN, true negative; TP, true positive; X, adjusted values cannot be
calculated because of a 0 in the denominator of formula.
*Additional data obtained through personal communication (Dr. Paul V. Abbott, September 2015, written communication); pulp vitality was confirmed clinically (endodontic access) or histologically or by the presence of root canal filling.
†
Peters et al, 1994 used two different electric pulp testers.
Diagnostic Accuracy of 5 Dental Pulp Tests

Review Article
699
 700

Review Article
Mainkar and Kim

TABLE 2. Diagnostic Accuracy of Pulp Vitality Tests in the Included Studies

Study for LDF Testing tool N TP FP FN TN Sensitivity Specificity Adjusted accuracy Adjusted PPV Adjusted NPV Confirmation method Quality
Chen and Abbott, 2011 Laser Doppler flowmetry 179 14 8 0 157 1.000 0.950 0.971 0.937 1.000 Clinical, RCF Low
(24)*
Evans et al, 1999 (27) Laser Doppler flowmetry 67 67 0 0 0 1.000 — X X X Clinical Low
Ingolfsson et al, 1994 (4) Laser Doppler flowmetry 11 10 0 1 0 0.909 — X X X Clinical Low
Karayilmaz and Kirzioglu, Laser Doppler flowmetry 59 59 0 0 0 1.000 — X X X RCF Low
2010 (5)
Olgart et al, 1988 (37) Laser Doppler flowmetry 33 32 0 1 0 0.970 — X X X Clinical Low
Roebuck et al, 2000 (38) Laser Doppler flowmetry 11 10 0 1 0 0.909 — X X X Clinical Low
Pooled diagnostic 0.975 0.950 0.971 0.937 0.997
accuracy values
95% CI 0.926–0.992 0.907–0.974 0.933–0.988 0.891–0.965 0.957–1.000
Heterogeneity (I2) 30.698 0.000 0.000 0.000 0.000
Study for PO
Dastmalchi et al, 2012 (10) Pulse oximetry 24 9 1 0 14 1.000 0.930 0.960 0.914 1.000 Clinical Low
Gopikrishna et al, 2007 (3) Pulse oximetry 80 42 2 0 36 1.000 0.950 0.971 0.937 1.000 RCF Moderate
Karayilmaz and Kirzioglu, Pulse oximetry 59 48 0 11 0 0.814 — X X X RCF Low
2010 (5)
Shahi et al, 2015 (41) Pulse oximetry 70 30 1 0 39 1.000 0.975 0.986 0.968 1.000 Clinical, RCF Low
Pooled diagnostic 0.973 0.954 0.974 0.943 0.990
accuracy values
95% CI 0.796–0.997 0.909–0.978 0.934–0.990 0.895–0.970 0.954–0.998
Heterogeneity (I2) 77.246 0.000 0.000 0.000 0.000
—, either 100% vital or nonvital teeth (predictive values are not valid); CI, confidence interval; FN, false negative; FP, false positive; NPV, negative predictive value; PPV, positive predictive value; RCF, root canal filling; TN, true negative; TP, true positive; X, adjusted values cannot be
calculated because of a 0 in the denominator of formula.
*Additional data obtained through personal communication (Dr. Paul V. Abbott, September 2015, written communication); pulp vitality was confirmed clinically (endodontic access) or histologically or by the presence of root canal filling.
JOE — Volume 44, Number 5, May 2018
 Review Article
TABLE 3. Summary of Pooled Diagnostic Accuracy Values for 5 Pulp Testing Methods
Sensitivity Specificity Adjusted accuracy Adjusted PPV Adjusted NPV
CPT 0.867 0.843 0.840 0.807 0.871
95% CI (0.810–0.909)a (0.773–0.895)ab (0.769–0.893)a (0.722–0.871)ab (0.808–0.915)a
EPT 0.720 0.928 0.817 0.888 0.804
95% CI (0.647–0.783)b (0.877–0.959)ac (0.770–0.856)a (0.818–0.934)ac (0.759–0.842)a
HPT 0.778 0.665 0.723 0.619 0.785
95% CI (0.647–0.869)ab (0.485–0.807)b (0.578–0.833)a (0.490–0.733)b (0.670–0.868)a
LDF 0.975 0.950 0.971 0.937 0.997
95% CI (0.926–0.992)c (0.907–0.974)c (0.933–0.988)b (0.891–0.965)c (0.957–1.000)b
PO 0.973 0.954 0.974 0.943 0.990
95% CI (0.796–0.997)ac (0.909–0.978)c (0.934–0.990)b (0.895–0.970)c (0.954–0.998)b
CI, confidence interval; CPT, cold pulp testing; EPT, electric pulp testing; HPT, heat pulp testing; LDF, laser Doppler flowmetry; NPV, negative predictive value; PO, pulse oximetry; PPV, positive predictive value.
Different letters within the columns indicate statistically significant differences among groups.

disease prevalence in this review allows for valid comparison of
accuracy, PPV, and NPV among studies using different sample disease
prevalence.
CPT has moderate diagnostic accuracy when evaluating both vital
and nonvital teeth. CPT has also moderate predictive values, but lower
than those of LDF and PO. It has been used as a primary pulp sensibility
testing method among clinicians, perhaps because of its generally good
diagnostic accuracy values and its ease of application (14–16). CPT was
determined to be almost equivalent to EPT in all diagnostic accuracy
values except sensitivity. To identify nonvital teeth, clinicians should
use CPT rather than EPT because the sensitivity of CPT is significantly
higher than that of EPT.
It is of note that EPT shows the lowest sensitivity among all pulp
testing methods although other diagnostic values are generally high.
EPT can be considered more reliable when identifying vital teeth than
nonvital teeth. This low pooled sensitivity results from a high FN (a
response that incorrectly indicates that teeth are vital) in the majority
of the studies. The pooled sensitivity of EPT was determined to be
72% in this systematic review, indicating that EPT could correctly iden-
tify 72% of nonvital teeth as a positive test (TP), but 28% of nonvital teeth
are incorrectly identified as a negative test (FN). On the other hand, the
specificity of EPT is equivalent to those of LDF and PO (no statistical dif-
ferences, P > .05), suggesting the usefulness of this method for identi-
fying vital teeth. EPT has been commonly used as an alternative to CPT
because of the high specificity and predictive values (15, 17). However,
it is recommended that EPT should be used in conjunction with CPT
because of its low sensitivity.
The least accurate pulp testing method is HPT based on the finding
that it shows the lowest values in all diagnostic accuracy categories
except sensitivity. It is thought that heated gutta-percha or a rubber
cup used for HPT in the included studies may not reliably or correctly
identify the vitality or nonvitality of teeth. Furthermore, a wide variation
in diagnostic values was found among the studies for HPT. Therefore,
HPT should not be recommended as a primary pulp testing method.
It may be useful to reproduce a patient’s chief complaint such as
heat hypersensitivity or to provoke a painful response for the diagnosis
of vital but diseased pulp (17).
LDF and PO are determined to be the most accurate diagnostic
methods in this systematic review and should be used by clinicians if
possible because they show consistently high diagnostic accuracy values
(sensitivity, specificity, adjusted accuracy, adjusted PPV, and adjusted
NPV) from all of the included studies with little heterogeneity. These
pulp vitality testing methods have been considered more accurate
than other pulp sensibility tests such as CPT, HPT, and EPT because
they can directly assess the presence of blood flow. In the previous sys-
tematic reviews, LDF was the most accurate and reliable for identifying
the pulp status (16) although it has shortcomings such as technique
sensitivity and application limitation requiring a pulp chamber above
the gingival margin (15).
The statistics for diagnostic accuracy are designed for a binary
diagnostic test such as pulp tests, which can detect whether a tooth
has a disease (pulp necrosis) or not. In the clinical setting, pulp sensi-
bility tests are also often used to assess the reversibility of inflamed vital
teeth based on the patients’ responses. For example, if a tooth has the
lingering pain or intense pain in response to CPT and HPT, it may be
diagnosed with irreversible pulpitis. However, it should be noted that
the statistics for diagnostic accuracy such as sensitivity, specificity, ac-
curacy, PPV, and NPV do not relate to the accuracy of detecting the

Figure 2. Quality assessment of the included studies using the modified Quality Assessment of Diagnostic Accuracy Studies tool.

JOE — Volume 44, Number 5, May 2018 Diagnostic Accuracy of 5 Dental Pulp Tests 701
Review Article
degree or reversibility of pulp inflammation although tests may provide
clinically meaningful information to arrive at a correct diagnosis.
Within the limitations of this systematic review and meta-analysis,
the following conclusions can be drawn. The most accurate pulp testing
methods are LDF and PO, and the least accurate method is HPT. EPT has
low sensitivity, suggesting that it is less likely to correctly identify nonvital
teeth, but has high specificity, suggesting that it is more likely to
correctly identify vital teeth. Among pulp sensibility tests, CPT has gener-
ally high diagnostic accuracy and can be considered a primary pulp
testing method.
Acknowledgments
The authors thank Dr Paul V. Abbott for providing additional
information for this systematic review.
The authors deny any conflicts of interest related to this study.
References
1. Ehrmann EH. Pulp testers and pulp testing with particular reference to the use of dry
ice. Aust Dent J 1977;22:272–9.
2. Mumford JM. Pain perception threshold and adaptation of normal human teeth.
Arch Oral Biol 1963;8:493–501.
3. Gopikrishna V, Tinagupta K, Kandaswamy D. Evaluation of efficacy of a new custom-
made pulse oximeter dental probe in comparison with the electrical and thermal
tests for assessing pulp vitality. J Endod 2007;33:411–4.
4. Ingolfsson AR, Tronstrad L, Hersh EV, et al. Efficacy of laser Doppler flowmetry in
determining pulp vitality of human teeth. Endod Dent Traumatol 1994;10:83–7.
5. Karayilmaz H, Kirzioglu Z. Comparison of the reliability of laser Doppler flowmetry,
pulse oximetry and pulp tester in assessing the pulp vitality of human teeth. J Oral
Rehabil 2011;38:340–7.
6. Ramsay DS, Artun J, Martinen SS. Reliability of pulpal blood flow measurements us-
ing laser Doppler flowmetry. J Dent Res 1991;70:1427–30.
7. Schnettler JM, Wallace JA. Pulse oximeter as a diagnostic tool of pulp vitality.
J Endod 1991;17:488–90.
8. Mills RW. Pulse oximetry—a method of vitality testing for teeth? Br Dent J 1992;172:
334–5.
9. Soo-ampon S, Vongsavan N, Soo-ampon M, et al. The sources of laser Doppler
blood-flow signals recorded from human teeth. Arch Oral Biol 2003;48:353–60.
10. Dastmalchi N, Jafarzadeh H, Moradi S. Comparison of the efficacy of a custom-made
pulse oximeter probe with digital electric pulp tester, cold spray, and rubber cup for
assessing pulp vitality. J Endod 2012;38:1182–6.
11. Villa-Chavez CE, Patino-Marın N, Loyola-Rodrıguez JP, et al. Predictive values of ther-
mal and electrical dental pulp tests: a clinical study. J Endod 2013;39:965–9.
12. Weisleder R, Yamauchi S, Caplan DJ, et al. The validity of pulp testing: a clinical
study. J Am Dent Assoc 2009;140:1013–7.
13. Jespersen JJ, Hellstein J, Williamson A, et al. Evaluation of dental sensibility tests in a
clinical setting. J Endod 2014;40:351–4.
14. Levin LG, Law AS, Holland GR, et al. Identify and define all diagnostic terms for
pulpal health and disease states. J Endod 2009;35:1645–57.
15. Mejare IA, Axelsson S, Davidson T, et al. Diagnosis of the condition of the dental
pulp: a systematic review. Int Endod J 2012;45:597–613.
16. Alghaithy RA, Qualtrough AJ. Pulp sensibility and vitality tests for diagnosing pulpal
health in permanent teeth: a critical review. Int Endod J 2017;50:135–42.
17. Levin LG. Pulp and periradicular testing. J Endod 2013;39:S13–9.
18. Michaelson PL, Holland GR. Is pulpitis painful? Int Endod J 2002;35:829–32.
19. Petersson A, Axelsson S, Davidson T, et al. Radiological diagnosis of periapical bone
tissue lesions in endodontics: a systematic review. Int Endod J 2012;45:783–801.
20. Kim KW, Lee J, Choi SH, et al. Systematic review and meta-analysis of studies eval-
uating diagnostic test accuracy: a practical review for clinical researchers - part 1.
general guidance and tips. Korean J Radiol 2015;16:1175–87.
702 Mainkar and Kim
21. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the
quality assessment of studies of diagnostic accuracy included in systematic reviews.
BMC Med Res Methodol 2003;3:25.
22. Cerrud CC. Accuracy of Pulse Oximetry, Cold Test and Electric Pulp Test in
Determining Pulpal Status [thesis]. Houston: University of Texas Health Science
Center at Houston; 2013.
23. Bruno KF, Alencar AHG, Estrela C, et al. Microbiological and microscopic analysis of the
pulp of non-vital traumatized teeth with intact crowns. J Appl Oral Sci 2009;17:508–14.
24. Chen E, Abbott PV. Evaluation of accuracy, reliability, and repeatability of five dental
pulp tests. J Endod 2011;37:1619–23.
25. Cooley RL, Robison SF. Variables associated with electric pulp testing. Oral Surg
1980;50:66–73.
26. Dummer PM, Hicks R, Huws D. Clinical signs and symptoms in pulp disease. Int
Endod J 1980;13:27–35.
27. Evans D, Reid J, Strang R, Stirrups D. A comparison of laser Doppler flowmetry with
other methods of assessing the vitality of traumatised anterior teeth. Endod Dent
Traumatol 1999;15:284–90.
28. Farid H, Khan FR, Pasha L, Shinwari MS. Are pulp sensibility tests still sensible?
J Ayub Med Coll Abbottabad 2015;27:874–7.
29. Fuss Z, Trowbridge H, Bender IB, et al. Assessment of reliability of electrical and
thermal pulp testing agents. J Endod 1986;12:301–5.
30. Garfunkel A, Ulmansky M. Dental pulp pathosis: clinicopathologic correlations
based on 109 cases. Oral Surg 1973;35:110–7.
31. Hori A, Poureslami HR, Parirokh M, et al. The ability of pulp sensibility tests to eval-
uate the pulp status in primary teeth. Int J Paediatr Dent 2011;21:441–5.
32. Kamburoglu K, Paksoy CS. The usefulness of standard endodontic diagnostic tests in
establishing pulpal status. Pain Clinic 2005;17:157–65.
33. Moody AB, Browne RM, Robinson PP. A comparison of monopolar and bipolar
electrical stimuli and thermal stimuli in determining the vitality of human teeth.
Arch Oral Biol 1989;34:701–5.
34. Peters DD, Baumgartner JC, Lorton L. Adult pulpal diagnosis: I—evaluation of the
positive and negative responses to cold and electrical pulp tests. J Endod 1994;20:
506–11.
35. Petersson K, Soderstrom C, Kiani-Anaraki M, et al. Evaluation of the ability of ther-
mal and electrical test to register pulp vitality. Endod Dent Traumatol 1999;15:
127–31.
36. Saeed MH, Mazhari NA, Al-Rawi NH. The efficacy of thermal and electrical tests to
register pulp vitality. J Int Dent Med Res 2011;4:117–22.
37. Olgart L, Gazelius B, Lindh-Stromberg U. Laser Doppler flowmetry in assessing vi-
tality in luxated permanent teeth. Int Endod J 1988;21:300–6.
38. Roebuck EM, Evans DJ, Stirrups D, Strang J. The effect of wavelength, bandwidth,
and probe design and position on assessing the vitality of anterior teeth with laser
Doppler flowmetry. Int J Paediatr Dent 2000;10:213–20.
39. Johnson RH, Dachi SF, Haley JV. Pulpal hyperemia—a correlation of clinical and
histologic data from 706 teeth. J Am Dent Assoc 1970;81:108–17.
40. Kontakiotis EG, Filippatos CG, Stefopoulos S, et al. A prospective study of the inci-
dence of asymptomatic pulp necrosis following crown preparation. Int Endod J
2015;48:512–7.
41. Shahi P, Sood PB, Sharma A, et al. Comparative study of pulp vitality in primary and
young permanent molars in human children with pulse oximeter and electric pulp
tester. Int J Clin Pediatr Dent 2015;8:94–8.
42. Cooley RL, Lubow RM. Evaluation of a digital pulp tester. Oral Surg Oral Med Oral
Pathol 1984;58:437–42.
43. Dai H, Zhang H, He Y. Diagnostic accuracy of PLA2R autoantibodies and glomerular
staining for the differentiation of idiopathic and secondary membranous nephrop-
athy: an updated meta-analysis. Sci Rep 2015;5:8803.
44. Meng Q, Shi S, Liang C, et al. Diagnostic accuracy of a CA125-based biomarker panel
in patients with pancreatic cancer: a systematic review and meta-analysis. J Cancer
2017;8:3615–22.
45. Reynolds RL. The determination of pulp vitality by means of thermal and electrical
stimuli. Oral Surg 1966;22:231–40.
46. Seltzer S, Bender IB, Ziontz M. The dynamics of pulp inflammation: correlations be-
tween diagnostic data and actual histologic findings in the pulp. Oral Sur Oral Med
Oral Pathol 1963;16:969–77.
JOE — Volume 44, Number 5, May 2018