Lessons Learned from the Cracked Tooth Registry – a Three-year Clinical Study in the Nation’s Network

Abstract

Background:

This article reviews the most salient lessons learned from a large, multi-site, 3-year observational study of posterior teeth with cracks conducted in the National Dental Practice-Based Research Network.

Studies Reviewed:

Eight articles published over a 6-year period (2017–2022) describing clinical characteristics of posterior teeth with cracks and their treatment and outcomes are reviewed and discussed to answer three common questions faced by clinicians: Which cracked teeth will get worse? When should practitioners intervene? What is the best treatment?

Results:

Though cracks in teeth are prevalent, few will fracture (3%) or show crack progression in three years (12%). Characteristics that guide the clinician to treatment include active caries, biting pain, and to a lesser degree, having a crack detectable with an explorer, connecting with a restoration, or that blocks trans-illuminated light; the main treatment chosen is a complete crown. Of those treated (36%), few (14%) will need retreatment, but will still survive, despite having an internal crack as well.

Implications:

While cracked teeth often pose a dilemma to clinicians, they are generally very good at deciding which teeth to treat, and when, and which to monitor.

Introduction

The disposition of teeth with cracks is a common and almost everyday dilemma for dental practitioners. This is especially true for teeth for which symptoms are not currently being experienced, but there are significant concerns over the future structural stability of the tooth. While cracked teeth have been a topic of laboratory and clinical studies for decades, many questions regarding their treatment remain unanswered.

Investigators in the National Dental Practice-based Research Network, an organization of dental patient care practices participating in clinical research studies and funded by the National Institute of Dental and Craniofacial Research of the National Institutes of Health, conducted a large-scale, three-year, prospective cohort study to address some of these questions. The original goal was to enroll 3,000 patients ages 19–85 from 150–300 Network practices, each of which had at least one posterior tooth with a crack, either symptomatic or not, and assess characteristics at the patient-, tooth- and crack-level at baseline and at annual follow-ups. The study, which is the largest prospective study conducted to follow outcomes of teeth with cracks, was conducted from April 2014 through December 2018. It is note worthy that 91% of all enrolled patients attended at least one recall: 96% of these attended year 1, 86% attended year 2, 80% attended year 3, and 74% (1912 patients) attended all three annual recall appointments.

The main objective of the CTR (Cracked Tooth Registry) was to identify patient-, tooth- and crack-level characteristics that were associated with the initial symptom status of the tooth and to follow these patients for 3 years to determine changes, such as increase or decrease in symptoms, crack progression, restorative treatment, need for endodontic therapy, fracture, extraction, and the characteristics that were associated with these changes. In particular, the study attempted to answer three critical questions: Which cracked teeth will get worse? When should a practitioner intervene? What is the best treatment for a cracked tooth?

The CTR resulted in eight peer-reviewed publications.^1–8 A review of these individual publications highlighting their most clinically relevant and informative findings, as well as an overall summary, is presented below.

Methods

The full details of the study design are provided in the individual eight publications. In brief, 209 independent practitioners in six geographic regions distributed throughout the U.S. enrolled a convenience sample of 2,858 patients from April 2014 through April 2015. Enrolled patients had at least one vital posterior tooth with a visible external crack, and the tooth may or may not have been symptomatic (pain to cold, pressure, spontaneous or a combination). For the purpose of this study, a crack was defined as “an obvious break of the external contiguous structure of the tooth, discernible by visual examination, but involves no loss of tooth structure (e.g. lost cusp).” The patients were recalled annually for three years, and at each visit, as at the baseline visit, characteristics were collected at the patient-, tooth- and crack-level to monitor progress and predictability of outcomes for the tooth. Clinicians were asked to treat patients based on their own experience and comfort, without any guidance about timing or type of any intervention. Patients were either identified as whether to monitor or treat, and if the latter the type of treatment. This assessment was repeated at each of the 3 yearly recall visits. For teeth that were treated, when possible, an assessment was made of the presence of internal cracks, i.e., within the dentin.

Results

Table 1 contains a summary of important characteristics at the three levels recorded at baseline. Overall, 45% of the teeth with cracks experienced one or more symptoms. The study was designed to oversample teeth that showed pain symptoms, so this is not a true prevalence level. It was interesting that the presence of the crack was observable on a radiograph for only 2% of the teeth.

Table 1.

Baseline data – characteristics of cracked teeth on three levels for 2,858 patients

Patient level: Mean age 54 (range 19–85) 63% female 85% non-Hispanic white 77% had some dental insurance 86% some college education 62% reported gum chewing 66% reported clenching/grinding or pressing 81% reported some stress 39% reported they were stressed weekly

Tooth level: 82% molars 59% mandibular 92% of teeth had a restoration 22% had root exposure 24% had at least one wear facet through enamel 11% had caries 9% had NCCL (non-carious cervical lesion) 35% had 3 or more cracks; very few had only 1

Crack level: 81% were stained 73% connected with restoration 69% detectable by explorer 65% blocked transilluminated light 94% were vertical 44% occlusal, 51% lingual 36% involved two or more surfaces

The eight articles addressed specific topics, namely characteristics associated with symptoms, types of symptoms, how and when dentists decide to treat a cracked tooth, treatment recommendations, presence of internal cracks within the tooth, rate of change in symptoms and crack characteristics, longevity of pain related to cracks, characteristics predicting crack progression and tooth fracture, and outcomes of cracked teeth over three years. The following paragraphs address the findings described in each of the eight articles.

When should I expect to see symptoms associated with a cracked tooth?

While it is very common for dentists to see teeth with cracks in their patient’s mouths, it is far less common for there to be associated pain. The objective of this study was to determine what patient-, tooth- and crack-level characteristics were associated with the presence of any symptoms, either pain to cold, pain on biting, spontaneous pain or some combination, as assessed by the practitioner.¹ To assess the independent associations of the different characteristics with symptoms, a stepwise, multivariable, generalized estimating equation (GEE) regression model was used. After adjusting for demographics, it was determined that the patient having clenching/grinding habits, the tooth being a molar, having caries, a wear facet through enamel, and blocking transilluminated light or involving the distal surface were all associated with the tooth showing symptoms, while the crack being stained was less likely to show symptoms. Associations showed that the patients could be divided into low and high risk groups, with those in the low risk having the characteristics of being men, white, 65 years or older, and do not clench or grind teeth, and the high risk group being women, Hispanic or nonwhite, younger than 65 years, and clench or grind teeth. The symptom probability increased when the crack was on a molar, blocked transilluminated light, and was on the distal surface, to a maximum probability of 77% of patients in the high risk group, and 44% in the low risk group (Figure 1). However, in either risk category, a patient with the three characteristics noted had a 20% greater chance of their cracked tooth being symptomatic.

Figure 1 –

Symptom probability resulting from sequentially adding tooth-level and crack-level characteristics for both low-risk demographic groups (men, white, 65 years or older, do not clench jaws or grind teeth) and high-risk demographic groups (women, Hispanic or nonwhite, younger than 65 years, clench jaws or grind teeth).

Types of symptoms (pain) present?

The CTR oversampled for teeth that showed symptoms in an attempt to better understand the relationship between type of pain and specific characteristics. The objective of this study was to determine which characteristics at the patient-, tooth- and crack-level correlate with the specific types of symptoms exhibited for the cracked teeth, including spontaneous pain, pain on biting, and pain from a cold stimulus.² The study showed that 37% of all of the cracked teeth in the study experienced pain to cold, 16% to biting, and 13% spontaneous (Figure 2). Of the 1,297 who experienced any pain, 65% had only one type of pain, 24% had two types ad 11% had all three types. For patients who experienced only one type, 78% were to cold, 14% were to biting, and 8% were spontaneous. Perhaps the most surprising finding from the study was that rather than pain on biting, which is typically considered one of the most important considerations in diagnosing a cracked tooth, pain to cold was the most common symptom experienced.

Figure 2 –

Distribution of percent of cracked teeth having at least one type of pain, comparing those with cold pain alone, biting pain alone, spontaneous pain alone, and each combination of pain, including all three. Percentages add up to 100 for each pain category.

Many characteristics were independently associated with the different types of pain. The highest odds ratios were obtained when the crack was detectable on a radiograph for pain on biting or spontaneous, or when all three types of pain were present. Molar cracks were associated with most combinations of pain. Patients who clenched or ground their teeth had higher odds of experiencing pain, mainly to biting. Cracks on the distal surface were correlated with pain on biting, spontaneous pain, and to biting and spontaneous, as well as when all three symptoms were present. If a crack blocked transilluminated light, the tooth was more likely to be associated with pain on biting, biting and spontaneous, and all three symptoms. Presence of an NCCL (non-carious cervical lesion), exposed roots and stained cracks were inversely associated with symptoms. No crack characteristic was associated with pain to cold alone, though cold was the most frequent type of pain.

How quickly do symptoms/cracks change?

One of the expectations when beginning the CTR study was that tooth pain would increase over time as crack progression occurred. This article quantified symptom and crack changes over the first year of observation for teeth that did not receive treatment (n = 1850), and correlated them with characteristics.³ It was surprising to find that only 6% of the patients experienced an increase in the number of external visible cracks in one year, typically one more crack. There was no association between an increase in number of cracks and any increase in pain. For 32% of the patients, a change in their tooth pain was observed, with decreases in pain being more than twice as common as increases (23% vs 10%; more than half being a decrease in cold pain), and decreases in pain being strongly associated with the baseline recommendation to monitor the tooth instead of treat. Overall, the results showed that untreated symptomatic teeth with cracks were relatively stable over one year, and a reduction in symptoms was more common than in increase (Figure 3).

Figure 3 –

Pain at baseline (Y0) and Y1 visit among 1850 patients who had no treatment (Tx) performed prior to Y1 visit.

Does a painful cracked tooth remain so?

The results of the three-year follow up to the previous article reinforced the one-year outcomes, showing untreated teeth to be fairly stable.⁴ This study also showed that treatment resolved pain, especially biting and spontaneous pain. For the 710 patients whose cracked tooth was treated at baseline or prior to year 1, the percent experiencing any pain at the year 1 recall decreased from 64% to 27%, biting pain decreased from 26% to 5%, spontaneous pain decreased from 25% to 2% and cold pain decreased from 49% to 23%. A similar trend at year 2 was observed for the 124 patients whose tooth was treated at year 1 or prior to year 2. It was also surprising that much of the pain in teeth not treated also decreased. Of the 586 teeth that were painful at enrollment but were not treated at any time during the study, the pain resolved in 58%. Virtually all biting (92%) and spontaneous (95%) pain resolved; 68% of cold pain resolved.

How do dentists recommend treating a tooth with a crack?

The CTR presented an opportunity to investigate the characteristics of posterior teeth with at least one crack that influence a practitioner’s decision about treatment, as well demographic, geographic and practice-type influences on recommendations.⁵ Of the 2,858 patients enrolled, 1,040 (36%) were recommended for treatment at baseline. Treatment recommendations were for restoration (98%), endodontics (3%), endodontics and restoration (2%), extraction (0.2%), and other noninvasive treatment, such as an occlusal device (1%). Odds for recommending restoration were highest for the presence of caries (OR=67.9), having biting pain (OR=7.3), radiographic evidence of the crack (OR=4.8), and presence of spontaneous pain (OR=2.9). Pain to cold, having dental insurance, the crack being detectable by explorer or blocking transilluminated light, or the crack connecting with a restoration were also associated with a recommendation for treatment, but at odds less than two-fold. The most common reason to recommend restoration was concern over compromised tooth structure (Figure 4). The most common restoration recommendation was a complete crown (61%), followed by intracoronal (35%) and partial crown (3%), and this was independent of region or type of practice.

Figure 4 –

Reasons for recommendation of restorative treatment of teeth with cracks by dental practitioners (N=1018).

How severe are treated cracks?

One of the dilemmas encountered by a practitioner for a tooth with a crack is estimating its severity or its penetration depth into the tooth. The CTR provided the possibility to assess cracks within dentin once the tooth was prepared for restoration, which most often was a complete crown. Data was collected for 435 teeth.⁶ Of these, 89% had at least one internal crack, and 46% had two or more cracks in dentin (Figure 5). Having pain on biting, an external visible crack connecting with a restoration, or the external crack extending onto the root were each associated with two-fold greater odds for having an internal crack. For 126 teeth for which there was only one internal crack and one external crack, the internal crack could be verified as an extension of the external crack for 100 of these, and the only external crack characteristic that was associated was the crack connecting with a restoration. This study suggested that an observation of an external crack that results in invasive treatment of the tooth will typically expose additional cracks within the dentin.

Figure 5 –

Distribution of 435 treated teeth according to number of internal cracks. Percentages do not total 100% because of rounding.

What predicts tooth fracture or crack progression?

One of the most significant concerns with teeth with cracks is their structural stability, as noted earlier, and trying to predict those that may show crack progression or eventual fracture. In this study, we reported that for 2601 patients that attended at least one recall, only 78, or 3%, subsequently developed a fracture (91% were partial fractures). Partial fracture was defined as the loss of a portion of tooth structure coronal to the periodontal attachment (e.g., loss of a cusp), and total tooth fracture included both the coronal and radicular tooth structure below the periodontal attachment (e.g., a fracture that renders the tooth non-restorable). In addition, only 12% showed evidence of crack progression (more cracks or more surfaces involved on untreated teeth) over 3 years.⁷ The majority of the changes occurred within the first two years (Figure 6). Maxillary teeth that had a wear facet through enamel, a crack detectable with an explorer, crack on the facial surface and in a horizontal direction were baseline characteristics associated with tooth fracture. Males and teeth with multiple cracks were associated with crack progression.

Figure 6 –

Timing of fracture development and crack progression as the number of occurrences in each category by year. Crack progression refers to increase in number of cracks or number of surfaces, and can be ascertained only on untreated teeth (and could happen more than once over the course of the study).

What happens to teeth with cracks in 3 years?

All teeth, either treated or monitored, were assessed for up to 3 years and final outcomes were reviewed for evidence of failure, namely extraction, fracture or recommendation for retreatment.⁸ Only 37 teeth were extracted during the 3 years of the study, making the overall survival rate 98%. Only 14% of the 478 teeth that were treated restoratively as recommended at enrollment required any retreatment. In the case of retreatment, either an intracoronal restoration was replaced with a crown (N=41), or a previously placed crown was repaired or replaced (N=26). None of the 9 treated with root canal therapy required any retreatment. About 80% of the teeth originally recommended to be monitored (65% of all enrolled teeth) remained as such throughout the study, while 14% that were initially recommended to be monitored were later recommended to be treated. Less than 7% of the teeth that were recommended to be monitored were later treated at an interim visit (without a recommendation). Overall, 92% of teeth initially recommended to be monitored were never treated.

Overall, 37% (n=966) of the enrolled teeth were treated during the 3 years of the study. Of these, 59% displayed symptoms, with 29% having multiple types of pain. The most common characteristics for treated teeth were a crack that stained (82%), was detectable with an explorer (73%), connected with a restoration (78%), or blocked transilluminated light (70%), and 42% had all four of these characteristics. More than half (54%) had a crack on the distal surface, and nearly one-quarter had caries, a wear facet through enamel, exposed root, or multiple restorations. Treatment was most often full coverage, and full coverage restoration using an indirect technique was more successful than intracoronal restorations placed directly. Nearly half of the treatments (45%) occurred at baseline, and 29% within the first year. Of the teeth treated with restoration in which the internal aspect of the tooth could be visualized, 90% had a crack in dentin.

Of the 63% (n=1651) of enrolled teeth that were not treated, 64% experienced no pain at baseline, and 50% had no pain throughout the study period. Of these untreated teeth, 12% had a recommendation for treatment at baseline, largely due to pain (60%). Of those with pain, 71% resolved and their recommendation changed to monitor.

Nearly all of the enrolled teeth had a restoration already at the time of enrollment, and nearly two-thirds had more than one visible crack.

Discussion

This comprehensive study presented new and confirmatory evidence about the characteristics of teeth with cracks. New findings include the cracked tooth being symptomatic was associated with the tooth having caries or a wear facet through enamel, and the crack blocking transilluminated light. Confirmatory findings were symptom association with the patient exhibiting clenching/grinding habits^9,10, the crack being on a distal surface^11,12, and the tooth being on a molar^13,14, likely due to heavy occlusal forces¹⁵.

Surprisingly, the most common type of pain for a cracked tooth was to cold, and not pain on biting as is most commonly cited.^16–19 It is possible that other factors, such as caries, dentin hypersensitivity and leaking or defective restorations could be contributing factors.^17,20 Of the 1641 cracked teeth that were never treated, only 7% had biting pain at baseline, and 95% became free of biting pain at one or more recalls, possibly suggesting that a conservative approach to treating teeth with cracks is warranted. Interestingly, cracks that stained, and teeth with NCCLs and exposed roots were less likely to be associated with pain. Stained cracks and NCCLs are suggestive of entities that have been present for a long time,²¹ and allowing reparative dentin formation.²²

Another unique aspect of this study was the ability to follow the temporal progression of pain in cracked teeth. Unexpectedly, the teeth were relatively stable over three years, with less than one-third showing a change in pain, and by 2:1 these changes were relief of symptoms, possibly due to increased pain tolerance. Not surprising was that 95% of patients who attended an unplanned visit had pain, with 93% reporting biting pain. However, only 6% showed an increase in the number of cracks.

Factors most often leading to the decision to treat a cracked tooth were caries, pain and radiographic evidence of the crack, consistent with previous studies.^23,24 In addition, blocking transilluminated light led to treatment, consistent with the expectation that the crack may be extensive and into dentin.^9,25 Similar to a previous practice-based study,²⁶ recommended treatment was typically restorative, more than 60% of the time with a crown, and due to concern over the quality of the remaining tooth structure,^{23, 27} and to seal the tooth.^28–30

For the first time, internal cracks in teeth with external cracks were quantitated through visual evaluation after tooth preparation, elsewhere described as deconstruction.³¹ From over 400 teeth with at least one visible external crack, 9 of 10 also had at least one internal crack (~1/2 had more than one). The presence of an internal crack was associated with pain on biting,^17,32 and patients with biting pain had >2.5 fold odds of having an internal crack than those without biting pain. Internal cracks were more likely when the external crack connected with a restoration and the crack extended onto the root surface.

While the number of tooth fractures was small (78/2601; 3%), fracture was associated with a wear facet through enamel, consistent with similar reports of a relation between excursive interferences and cracks.²¹ Fracture was more common in maxillary molars, which is similar to some studies,^18,33 while others have shown mandibular teeth to be more likely to have cracks.^{12, 34–35} A tooth with a crack on the facial surface, detectable by explorer, and horizontal in direction, were more likely to fracture. Of 1889 patients with untreated cracked teeth after year one, 12.3% showed evidence of crack progression (either increased number or surfaces). Surprisingly, the factors associated with fracture were not associated with crack progression.

This large study showed a very high 3-year survival rate of 98% for cracked teeth, and only 14% of teeth treated with restorations needed retreatment, and these were mainly intracoronal restorations. The main reasons teeth were re-treated were related to caries, broken restorations, or a compromised tooth. The success rate for teeth treated with full or partial coverage, and especially full coverage, was greater than for those treated with intracoronal restorations, consistent with other studies.^36–38 None of the 18 teeth in the study treated with root canal therapy failed, consistent with excellent results from other studies showing 88–93% success rates of root canal treated cracked teeth at 1 year^39–40 and 84% at 5 years.⁴⁰

This series of publications based on the three-year Cracked Tooth Registry conducted in The Network provided valuable information for a large number of patients about the treatment and outcomes of teeth with cracks for patients visiting dental practices throughout the United States. Practices and patients were highly motivated to participate. High rates of patient recall were obtained, leading to more consistent date collection over time. The study demonstrated that critical information to the profession can be generated under “real-life”, day-to-day setting that is practice-based clinical research. The study does have limitations, especially in the fact that being a practice-based study where this research was being provided within the normal workflow of the practice, the amount of data that could be reasonably collected was limited. For example, we did not collect data on several potentially important variables, such as periodontal probing, occlusion as anterior guidance or group function, of size of restoration in the teeth when one was present. Therefore, while specific guidelines for the treatment of cracked teeth are not already available for the dental practitioner and could not be fully elucidated from this work, this study has verified that dental practitioners proved to be reasonably adept at deciding when and how to treat teeth with cracks, in order to keep these teeth functioning and the patient largely pain-free.

Conclusions

This practice-based research network, three-year study of nearly 3000 posterior teeth with at least one visible crack provides excellent guidance about the clinical care of cracked teeth, at least for the relatively short term. The study was designed to address three specific questions, and now provides some important answers.

1. Which untreated cracked teeth will get worse? The study showed that few teeth will fracture in 3 years (3%), but those that do will most likely be a maxillary tooth, have a wear facet in enamel, a horizontal crack, have the crack on the facial surface, and the crack will be detectable with explorer. Only 12% of teeth will show crack progression, most likely male patients and those with cracks involving multiple surfaces. Further, of teeth presenting with symptoms that do not receive treatment, the majority will have their symptoms subside and remain so for three years. Treatment resolves virtually all biting and spontaneous pain, and most cold pain, and these teeth will typically remain free of pain.

2. When should we intervene? The study showed that the clinicians most often intervened when the tooth had active caries, the patient was experiencing pain on biting, and the tooth had multiple crack characteristics of staining, being detectable by explorer, connecting with a restoration, and blocking transilluminated light. Of treated teeth (where treatment was mainly crowns), only 14% will need to be retreated, but will survive. Overall, 98% of cracked teeth will remain in the mouth for at least 3 years. Of those teeth that are recommended to be monitored, 92% will remain so for at least 3 years.

3. What is the best treatment? The study showed that the most successful treatment of a tooth with at least one visible crack will be a full crown. This recommendation is further supported by the fact that of the treated teeth where it was possible to examine the internal aspects, 90% of these teeth also had at least one internal crack, and most had two or more, and thus were more severely compromised structurally than could be determined by standard means of clinical assessment.