The 47 °C for One Minute Rule: Origin, Generalizability, and a Thermal-Dose Reformulation for Implant Osteotomy
- Paper type
- Critical Review (with a theoretical reformulation)
- Status
- Original Work · Working Paper
- Keywords:
- bone heat threshold · thermal dose · CEM43 · Arrhenius damage integral · implant osteotomy · drilling temperature · reporting standard
Contents
Abstract
Background. Implant surgery teaches that bone heated to about 47 °C for 1 minute is injured. The figure is usually traced to rabbit experiments in which a titanium chamber was heated at a controlled temperature for a fixed time. It is applied to human jaw osteotomies, where temperature is a transient shaped by drill, irrigation, and bone. Later animal studies place the injury threshold between 47 and 55 °C, and randomized trials of low-speed drilling without irrigation report no marginal bone penalty.
Objective/Hypothesis. We examine whether a single temperature-time point is an adequate injury criterion for osteotomy, and hypothesize that part of the disagreement among studies arises because different temperature-time histories are summarized by one peak value. A cumulative thermal dose (cumulative equivalent minutes at 43 °C, CEM43, and an Arrhenius damage integral) would allow the studies to be compared on one scale.
Proposed approach. A critical review traces the rule to its origin, appraises supporting and contradicting studies at abstract level, and sets out the dose mathematics with defined symbols. An illustrative computation, run for this paper and not empirical, converts several exposures to dose. A minimum reporting dataset for drilling temperature recordings is proposed, together with a re-expression protocol for published studies.
Expected contribution. An originality search found that CEM43 and Arrhenius models have already been combined for ultrasonic skull-bone grinding, so the contribution is narrowed to a historical and critical appraisal of the dental rule, a worked mapping of the rule onto dose, and a reporting standard. Tissue constants for human jaw bone are not established.
Conclusions. The 47 °C/1 min rule is best read as one constant-temperature exposure measured in rabbit tibia, not as a validated drilling limit. Dose expression is a plausible but untested way to reconcile the literature; its value over reporting peak temperature plus time above a cutoff must be shown empirically.
1. Introduction
Heat generated during site preparation is a standard concern in implant dentistry. Common teaching holds that bone temperatures above roughly 47 °C maintained for one minute produce thermal injury that can compromise osseointegration, and this single figure underlies the insistence on copious irrigation, intermittent drilling, sharp drills, and low drilling speeds. The rule is compact, memorable, and easy to audit with a thermocouple, which also makes it easy to apply beyond the conditions in which it was derived.
Three features of the rule deserve scrutiny. First, it describes an exposure, not a surgical event: the founding experiment heated a chamber at a set temperature for a set duration [1]. A drilling transient has no plateau. Its temperature rises over seconds, peaks, and decays, and the time spent near the peak is typically far shorter than one minute. Second, the threshold is a point on an unknown surface. Injury depends on both temperature and time, and a single point cannot say whether 52 °C for five seconds is more or less damaging than 47 °C for one minute. Third, the clinical literature that is supposed to validate or refute the rule, namely trials comparing irrigated and unirrigated drilling, rarely records temperature at the bone with the time resolution needed to place any exposure relative to the threshold.
Hyperthermia oncology faced the same problem in the 1980s. Sapareto and Dewey proposed converting any time-temperature history into an equivalent number of minutes at a reference temperature of 43 °C, an approach that does not assume that all tissues share the same heat sensitivity [2]. The resulting quantity, CEM43, is now a common exposure metric in hyperthermia and has been proposed as a basis for safety thresholds elsewhere, for example for radiofrequency exposure in magnetic resonance imaging [3]. A parallel tradition describes thermal injury as a first-order rate process with Arrhenius kinetics; the two formulations differ in range and generality, and methods exist to convert between them [4].
This manuscript is a critical review with a theoretical reformulation. It does not report experiments or patient data. Its aims are to (i) trace the dental 47 °C/1 min rule to the experiments that produced it and state what those experiments can and cannot support; (ii) set the animal, in vitro, and clinical evidence against it; (iii) formalize a thermal-dose expression of osteotomy heating and compute, under stated assumptions, what the rule’s own exposure implies for other temperature-time combinations; and (iv) propose how drilling temperature recordings should be reported so that dose can be computed and studies compared. Because the originality of the dose idea was a stated concern, the search methods and the prior work that narrowed the contribution are described explicitly.
2. Background and critical appraisal of the evidence
Search and appraisal approach
PubMed and Europe PMC were searched on 17 November 2026 using several vocabularies: thermal dose and CEM43, Arrhenius damage, thermal necrosis threshold, and drilling temperature, each combined with bone, osteotomy, drilling, grinding, and implant. ClinicalTrials.gov was searched for trials of drilling protocols. Known foundational and review papers were retrieved by title. Embase, Scopus, and Web of Science were not accessible. Full texts were not available; appraisal is at abstract level and every statement about design, sample size, and bias should be verified against full texts. Two foundational papers (a 1972 monograph on heat and bone tissue [5] and the 1983 vital-microscopic rabbit study [6]) have no PubMed abstract, so only their titles were seen and nothing is claimed about their contents.
Origin of the rule
The experiment most often cited for the threshold used a dividable titanium implant placed in the tibial metaphysis of rabbits, which allows a numerical estimate of ingrowing bone. Heating the test implants to 47 °C or 50 °C for 1 minute significantly reduced bone formation in the chamber, whereas 44 °C for 1 minute had no significant effect [1]. The same group had earlier published a vital-microscopic study of temperature threshold levels in rabbit bone [6], and an earlier monograph addressed thermal properties of bone and threshold levels for thermal injury [5]. The sample size of the 1984 study could not be verified from the abstract.
Three properties of this evidence bear on generalization. The exposure was applied to an implant at a defined temperature for a defined minute, so the experiment tests a constant-temperature step, not a transient. The tested levels were 44, 47, and 50 °C, so the threshold lies between 44 and 47 °C at the earliest; the “47 °C” figure is the lowest tested temperature at which an effect was seen, not an interpolated critical temperature. And the outcome was bone formation in a chamber in a rabbit long bone, which is a regenerative readout in a species and site that differ from human cortical jaw bone and from the bone-implant interface after drilling. None of this makes the finding wrong. It means the finding supports the statement that sustained heating at 47 to 50 °C impairs bone regeneration in this model, which is narrower than the statement that any drilling episode reaching 47 °C is harmful to human jaw bone.
Orthopaedic reviews restate the threshold as established. One review of cortical bone drilling states that temperatures above 47 °C cause thermal osteonecrosis that contributes to screw loosening and implant failure [7]. Within implant dentistry, a systematic review of heat generation and drill wear during site preparation synthesized the factors that modify heat [8] (a corrigendum was later issued [9]), and reviews of orthopaedic drilling catalogue the parameters involved [10].
Later animal and in vitro tests of the threshold
A systematic review of preclinical in vivo studies of thermal bone necrosis, undertaken for thermal explantation of implants, included 25 animal studies in 455 animals and found threshold values ranging from 47 to 55 °C for 1 minute; at 50 °C for 1 minute, bone-implant contact decreased significantly and infrabony pockets increased, and cancellous bone was more damaged than cortical bone [11]. The authors concluded that no clear threshold was available. This review addressed a heating-to-remove-implants question and not drilling safety, but it shows that the figure moves with species, site, bone type, and endpoint.
Individual studies illustrate the spread. In sheep iliac crest, a study of 15 implant sites in low-density bone found no significant difference from control at 50 °C and a lower bone-implant contact at 60 °C, without failures [12]. In a study of 48 rats, bone-implant contact fell significantly only at 50 °C [13]. These are small animal experiments with short follow-up and different heating methods, and their main risk of bias is that the heating protocol (implant-borne heating at fixed temperature) again differs from drilling.
In vitro work complicates the picture further. Bone cells exposed to heat-shock for durations occurring in surgical cutting showed an immediate necrotic response at 45 and 47 °C, most severe at 60 °C, and a longer-term apoptotic response that was stronger in osteocyte-like than in osteoblast-like cells; yet mild heat-shock of up to 47 °C enhanced differentiation and mineralization in mesenchymal stem cells [14]. The study is in vitro and uses cell lines and murine progenitors, so its relevance to a living cortical wall is uncertain, but it shows that exposures near the rule’s temperature can be injurious to some cells and stimulating to others, depending on duration and cell type.
Clinical evidence: drilling without irrigation
A systematic review of low-speed drilling without irrigation versus conventional drilling included 13 studies of mixed design (in vitro, animal, and clinical) and reported that temperatures stayed below 47 °C in both arms [15]. This is informative but also a reminder that such trials may not test the rule at all: if neither arm approaches the threshold, an absence of a difference is expected. A randomized trial of 25 patients with 30 implants compared irrigated high-speed drilling with unirrigated low-speed drilling and found mean 12-month bone loss of 0.83 and 0.62 mm, respectively, without a significant difference [16]. A later randomized single-blinded parallel trial in 66 patients with single posterior edentulous sites compared high-speed (800 rpm) drilling with irrigation against low-speed drilling without irrigation, with marginal bone level measured on periapical radiographs over 12 months, and reported no difference in bone loss [17]; the same research group has also compared patient-centered outcomes between the two protocols [18]. These trials are small, short, use a radiographic marginal-bone endpoint that is insensitive to subclinical necrosis, and come from a connected group of investigators; independent replication is limited, although registry records indicate other groups are now testing biological or low-speed protocols (for example NCT07937384, completed, and NCT06992557, not yet recruiting; results not located).
Drilling experiments also show that temperature and duration vary jointly with technique. In vitro, a larger drilling force lowered the maximum cortical temperature and shortened the time spent above 50 °C [19], and operator technique and bur design altered temperature during preparation for self-tapping screws [20]. Systematic reviews continue to address specific modifiers, including surgical guides [21], piezosurgery [22], and drill material [23]. The point for the present argument is that technique changes the whole temperature-time curve, and a peak summary discards most of the information that distinguishes one technique from another.
What the literature does not contain
No study located applied a time-integrated injury metric to dental implant osteotomy, and none reported human in vivo dose-response relationships between recorded bone temperature and histological or radiographic outcome. Temperature recording itself is an active methodological area, with thermocouple, infrared thermography, and fiber Bragg grating approaches in current use [24, 25, 26], and bone-tumor hyperthermia has produced its own review of temperature monitoring [27]. In the thermal-damage literature generally, a survey of damage thresholds in normal tissues reported that a large share of papers had to be discarded because the time-temperature history at the site of damage assessment was not recorded, and recommended that future studies record it [28]. Osteotomy research appears to share this weakness.
| Study | Design and size | Finding relevant to the rule | Main bias or limitation | Level |
|---|---|---|---|---|
| Eriksson and Albrektsson 1984 [1] | Rabbit tibial bone chamber; n not verified | 47 or 50 °C for 1 min reduced bone formation; 44 °C did not | Constant-temperature implant heating, not drilling; species and site | B |
| Kniha et al. 2020 [11] | Systematic review, 25 animal studies, 455 animals | Threshold 47 to 55 °C for 1 min; cancellous more affected | Heterogeneous heating methods; abstract-level appraisal | B |
| Trisi et al. 2015 [12] | Sheep iliac crest, 15 sites | 50 °C not different from control; 60 °C lowered BIC | Small n, low-density bone | C |
| Kniha et al. 2022 [13] | Rat tibia, 48 animals | BIC fell significantly only at 50 °C | Implant-borne heating; species | C |
| Dolan et al. 2012 [14] | In vitro cell lines and progenitors | Necrosis at 45 to 47 °C; mild heat promoted MSC mineralization | Cells, not tissue | C |
| Pellicer-Chover et al. 2017 [16] | RCT, 25 patients, 30 implants | Unirrigated low-speed drilling not different in 12-month bone loss | Small n; temperature not stated in abstract | C |
| Bernabeu-Mira et al. 2025 [17] | RCT, 66 patients, 12 months | No difference in marginal bone loss | Radiographic endpoint; related investigators | C |
| Babbar et al. 2021 [29] | Finite-element and thermography study, ultrasonic skull grinding | CEM43 and Arrhenius thermal dose model for bone grinding | Not dental; validation of constants not seen in abstract | C |
Note. Levels are operational categories used in this manuscript, not GRADE: A strong, B moderate, C preliminary, D mechanistic or theoretical, E speculative. BIC, bone-implant contact; MSC, mesenchymal stem cell; RCT, randomized controlled trial.
3. Hypothesis and theoretical framework
Formal statement
Hypothesis H1. Let T(t) be the temperature history at a defined bone location during an osteotomy. The probability and extent of thermal injury at that location is a monotone function of a time-temperature integral D[T(t)], not of the peak temperature alone. Consequently, published studies that appear to disagree about the “threshold” because they report different peak temperatures, durations, and sites will show smaller between-study disagreement when their exposures are re-expressed as D with explicit uncertainty.
Null hypothesis H0. Peak temperature, or peak temperature together with time above a fixed cutoff, predicts injury as well as D, so the integral adds no information.
The hypothesis is about exposure expression. It does not state that a specific numerical dose is harmful in human jaw bone; that requires data that do not yet exist.
Assumptions
- Injury at a location is determined by the local temperature history, and the history can be measured or modeled at that location.
- Over the temperature range of interest, a single dominant rate process governs injury. Pearce noted that the equivalent-dose formulation describes a single process over a limited temperature range and is not suited to multiple higher-temperature events, whereas Arrhenius formulations can describe many processes [4]. This assumption is the most fragile one and is examined in section 7.
- Repair, adaptation, and heat-shock responses during and after exposure are not modeled. Mild heat has been reported to stimulate mesenchymal progenitors in vitro [14], so a monotone dose-injury relationship may not hold at low dose.
- Constants in the dose formulations are borrowed from hyperthermia convention or calibrated to the rule itself. No tissue-specific constants for human jaw bone are established, and none of the values below should be read as such.
Two dose formulations
Cumulative equivalent minutes at 43 °C. With T(t) in degrees Celsius and time t in minutes,
CEM43(τ)=∫_0^τ R^( (43−T(t))) dt ≈ ∑_i^ R^( (43−T_i)) Δt_i
where T_i is the mean temperature in sampling interval i of duration Δt_i (minutes), and R is a dimensionless constant. Conventional hyperthermia practice takes R = 0.5 for temperatures at or above 43 °C, meaning that each degree of increase halves the time needed for an equivalent effect, and R = 0.25 below 43 °C. These conventional values are stated here as assumptions and were not taken from the abstracts reviewed. Sapareto and Dewey introduced the equivalent-minute concept to express any thermal exposure as an exposure time at a reference temperature of 43 °C, and explicitly noted that it does not require tissues to have the same heat sensitivity [2]. In practice CEM43 is used with tissue-specific thresholds; for example, a workshop on radiofrequency exposure in magnetic resonance proposed limits expressed in CEM43 for different thermoregulatory conditions [3].
Arrhenius damage integral. With absolute temperature T_K(t) in kelvin and time in seconds,
Ω(τ)=∫_0^τ A exp(−E_a/(R_g T_K(t)))dt, damage fraction=1−e^(−Ω)
where A is the frequency factor (s^-1), E_a is the activation energy (J mol^-1), and R_g = 8.314 J mol^-1 K^-1 is the gas constant. The damage fraction expression assumes a single first-order irreversible process. Pearce compared the two formulations and reviewed methods to convert between them, noting that Arrhenius formulations are quantitative and can be compared with histologic markers of injury [4].
Link between the two. Over a narrow range around a temperature T_0 (kelvin), the rate ratio per kelvin is exp(E_a/(R_gT_0^2)), which equals 1/R. Hence E_a≈ln(1/R) R_g T_0^2. For R = 0.5 at 47 °C (320.15 K) this gives E_a of about 591 kJ mol^-1 (computed).
Calibration to the rule. To ask what the rule itself implies, the Arrhenius frequency factor was set so that Ω = 1 for 47 °C held for 60 s (ln A = 217.8, A in s^-1). This is a calibration to the rule and not a measurement. It makes the 47 °C/1 min exposure equal to 16 CEM43 under R = 0.5 and equal to Ω = 1 under the calibrated Arrhenius model, so that all other numbers below are statements about the rule’s implications, not about bone.
Illustrative computation under stated assumptions — not empirical data
The script (dose_calc.py, saved with this manuscript) computes CEM43 with R = 0.5 at or above 43 °C and 0.25 below, and Ω with the calibrated constants above. Table 2 gives constant-temperature exposures; the output is reproduced exactly as computed.
| Temperature (°C) | Duration (s) | CEM43 (min) | Arrhenius Ω (calibrated to 47 °C, 60 s = 1) |
|---|---|---|---|
| 41 | 1800 | 1.88 | 0.433 |
| 43 | 600 | 10 | 0.603 |
| 44 | 60 | 2 | 0.123 |
| 47 | 60 | 16 | 1 |
| 47 | 1 | 0.267 | 0.0167 |
| 50 | 1 | 2.13 | 0.131 |
| 50 | 5 | 10.7 | 0.654 |
| 50 | 60 | 128 | 7.85 |
| 55 | 1 | 68.3 | 3.73 |
| 55 | 10 | 683 | 37.3 |
| 60 | 1 | 2180 | 96.1 |
Note. The Arrhenius calibration forces the rule to equal Ω = 1 and is therefore not independent evidence. The rabbit result placed 44 °C for 1 min (2 CEM43) in the no-effect range and 47 and 50 °C for 1 min (16 and 128 CEM43) in the effect range [1]. The 43 °C, 600 s row gives Ω = 0.603, i.e., a long mild exposure approaches the rule’s dose; whether bone responds this way is unknown.
Three observations follow. First, the rabbit experiment itself spans a factor of 8 in dose between its two effect-producing exposures (16 and 128 CEM43) and a factor of 8 between the no-effect and the lower effect exposure (2 and 16 CEM43). The data therefore bracket a dose threshold between about 2 and 16 CEM43 in that model, with the threshold “at 47 °C” being a consequence of the temperatures chosen by the experimenters. Second, dose rises steeply with temperature: each degree doubles it under R = 0.5, so a one-degree sensor error produces a twofold dose error, and a five-degree error a thirty-two-fold error. Third, the time to reach the rule’s dose shortens rapidly with temperature. Under R = 0.5 the time needed to accumulate 16 CEM43 is 240 s at 45 °C, 60 s at 47 °C, 7.5 s at 50 °C, 1.875 s at 52 °C, 0.234 s at 55 °C, and 0.007 s at 60 °C (computed). The values above about 50 °C extrapolate the equivalent-minute relation beyond the range for which it is normally used, and the Arrhenius form is preferable there, but neither has been validated for bone.
Synthetic drilling transients. To show how peak temperature and dose can diverge, five hypothetical profiles were constructed with a 37 °C baseline, a linear rise to a peak over a rise time, and an exponential decay with time constant τ. These profiles were not derived from heat-transfer physics or from any recording and are chosen only to show the arithmetic. Results were:
- A (peak 50 °C, rise 3 s, τ = 10 s): 3.3 s at or above 47 °C; CEM43 3.4 min; Ω = 0.215.
- B (peak 50 °C, rise 3 s, τ = 30 s): 8.6 s at or above 47 °C; CEM43 8.79 min; Ω = 0.557.
- C (peak 55 °C, rise 2 s, τ = 5 s): 3.8 s at or above 47 °C; CEM43 41 min; Ω = 2.34.
- D (peak 60 °C, rise 1 s, τ = 3 s): 3.1 s at or above 47 °C; CEM43 578 min; Ω = 27.4.
- E (peak 45 °C, rise 10 s, τ = 60 s): never at or above 47 °C; CEM43 0.851 min; Ω = 0.0699.
Profiles A and B share the same peak but differ by a factor of about 2.6 in dose; both are below the rule’s dose of 16 CEM43, even though both exceed 47 °C briefly and would be flagged by a peak-based criterion. Profiles C and D exceed the rule’s dose within seconds. Profile E never reaches 47 °C but accumulates a measurable dose from a long, mild elevation. Under the rule’s own dose, therefore, a peak criterion misclassifies in both directions: it flags exposures that are small in dose and is silent about long moderate exposures. A criterion of “peak above 47 °C” and a criterion of “dose above 16 CEM43” disagree on A and B. Whether either matches biology is the empirical question.
A report that “the maximum temperature was 49 °C” is compatible with a range of doses spanning more than an order of magnitude, depending on duration. Without duration, dose cannot be bounded tightly, and the sign of the disagreement between a study and the rule cannot be determined.
Mechanism and where the formulation may fail
Thermal injury to bone involves several processes with different kinetics: protein denaturation in osteocytes and osteoblasts, apoptosis activated over hours to days, and vascular injury that limits subsequent repair. In vitro, osteocyte-like cells show a stronger longer-term apoptotic response than osteoblast-like cells after short heat exposures, and heat exposures of the same type can promote differentiation in progenitors [14]. A single Ω or CEM43 collapses these into one exposure variable. It is therefore best seen as a predictor of the extent of immediate necrosis, not of the net healing outcome. Cancellous bone showed more damage than cortical bone in the animal review [11], implying that constants, or the relation between exposure and damage, depend on tissue microstructure and perfusion.
A further difficulty is spatial: dose varies with distance from the drill wall, while a thermocouple records one point. A finite-element study of bone drilling that estimates a thermally affected zone as a function of drill diameter, speed, feed force, and predrilling illustrates a route to estimating volumes [30], although the abstract reviewed does not state which injury metric it applied.
Numbered, falsifiable predictions
- P1. In a set of drilling or heating studies in which T(t) at bone has been recorded or can be reconstructed, between-study heterogeneity in injury outcome (histological necrosis, bone-implant contact) will be lower when exposure is expressed as Ω or CEM43 than when expressed as peak temperature. This is refuted if peak temperature explains the outcomes as well as or better than dose.
- P2. In controlled constant-temperature experiments of the 1984 type, exposures matched in dose but differing in temperature and duration (for example 47 °C for 60 s and 50 °C for about 7.5 s under R = 0.5) will produce similar injury. This is refuted if the short, hotter exposure is consistently less (or more) injurious than predicted, which would show that the time-temperature exchange rate is wrong for bone.
- P3. Recorded drilling transients in irrigated and unirrigated low-speed protocols will, in most sites, fall below the dose equivalent to 47 °C for 1 minute; the exceptional sites with higher dose will show more injury. This is refuted if sites with doses above the rule’s dose show no more injury than low-dose sites.
- P4. Sensor lag and sensor position will bias recorded dose downward more than recorded peak temperature, because dose is a convex, exponential function of temperature. This is refuted if dose and peak are affected proportionally.
- P5. Cortical and cancellous bone will require different effective constants (R or Ea) to fit histological injury. This is refuted if one set of constants fits both.
- P6. At low dose, outcomes such as stability gain will not be worse than at near-zero dose. This is the weakest prediction, based on in vitro reports of heat-induced osteogenic responses [14], and it is refuted by a monotone worsening of outcomes with dose.
4. Proposed methodology
The methodology has three components that build on one another: a review and re-expression protocol applied to existing literature, a minimum reporting dataset for future recordings, and a validation experiment that would give dose a biological meaning. Only the first component can be carried out without new data collection. No component has been performed.
Review and re-expression protocol
Eligibility. In vitro, animal, and clinical studies of drilling or controlled heating of bone, in which temperature was measured and a biological outcome (histology, bone-implant contact, implant stability, marginal bone change, failure) was reported. Studies reporting temperature only, without outcome, would be retained in a separate stratum for the reporting audit.
Sources. PubMed, Embase, Scopus, Web of Science, and Europe PMC, with the vocabularies used in this manuscript and additional terms from orthopaedic, neurosurgical, and biomechanical engineering literature. The present manuscript could not access Embase, Scopus, or Web of Science, which is a limitation of the originality search as well as of the evidence appraisal.
Extraction. For each study arm: species and bone site; bone type (cortical or cancellous) and density; heating method (drilling, implant-borne heating, probe); sensor type, diameter, distance from the osteotomy wall, depth, sampling rate, and any lag characterization; irrigation, speed, feed force; temperature as peak, mean, and any time series; time above stated cutoffs; outcome, assessment time, and measure.
Dose reconstruction. Studies fall into three tiers. Tier 1 reports a digitizable time series, from which CEM43 and Ω can be integrated directly. Tier 2 reports peak temperature and duration above a cutoff; dose is bounded by assuming limiting shapes (a rectangular plateau gives an upper bound, a triangular pulse a lower bound for the same peak and duration). Tier 3 reports only peak temperature; no dose bound can be formed, and the study contributes to the reporting audit but not to the re-expression. Uncertainty in constants would be propagated by sampling R and Ea (or A and Ea) from stated ranges, with the 1984 rabbit exposures and other controlled constant-temperature animal experiments used as calibration anchors, and by reporting results across the entire sampled range rather than at a single set of constants.
Primary analysis. For tiers 1 and 2, a mixed-effects regression of the outcome on log dose, with study as a random effect, compared with a regression on peak temperature and on peak temperature plus time above 47 °C. Models would be compared by cross-validated prediction error, and by the proportion of between-study variance explained, which is the quantity named in P1. Species and bone type would be covariates. The prespecified analysis would not pool animal and human outcomes without a species term.
Bias control. Two reviewers would extract independently; outcome and temperature data would be extracted by different people when feasible so that dose reconstruction is blind to outcome. Selective reporting of temperature would be examined by comparing methods sections with reported results.
Proposed minimum reporting dataset
Dose can only be computed if the history is recorded. Table 3 lists items that would let a reader compute CEM43 and Ω, assess sensor bias, and compare studies. The principle is that raw or digitizable temperature-time data at bone should accompany summary values.
| Domain | Items to report |
|---|---|
| Temperature record | Full T(t) trace for each site (supplementary file or digitizable figure), sampling frequency (Hz), baseline temperature, start and end of recording relative to drill contact and withdrawal |
| Derived exposure | Peak temperature; time above 43, 47, and 50 °C (s); CEM43 and Ω with stated R, A, and Ea; dose computed from the raw series, not from the summary |
| Sensor | Type (thermocouple, thermography, fiber Bragg grating), diameter, response time or lag characterization, calibration method and tolerance, placement distance from wall and depth below the cortical surface, orientation relative to drill axis |
| Drilling | Drill diameter and design, wear state, speed (rpm), feed force or feed rate, pecking pattern, irrigation (type, temperature, flow, internal or external), pre-drilling sequence, guide use |
| Bone | Species, site, bone type, density (HU or equivalent), cortical thickness, in vivo or ex vivo, tissue temperature at start |
| Outcome link | Histology or imaging endpoint, time of assessment, region matched to sensor location, blinding |
| Uncertainty | Estimated temperature error at the sensor, and the dose range implied by that error under the stated constants |
These items overlap partly with the descriptors already tabulated in drilling reviews, such as speed, feed rate, cooling, drill geometry, wear, and measurement method [7], and with measurement approaches in current use [24, 25, 26]. What is added is the requirement to report the trace and derived dose, and to quantify sensor-related uncertainty.
Validation experiment (prospective, not performed)
Re-expression cannot show that dose is biologically meaningful, because existing studies were not designed to vary it independently of peak temperature. A designed experiment would do so. In an ex vivo or animal bone model, exposures would be applied by controlled heating (for example an implant-borne or probe heater) in a factorial arrangement of temperature and duration, with dose-matched pairs on the same iso-dose line (for example 47 °C for 60 s and 50 °C for 7.5 s under R = 0.5), a sham-heated control, and dose levels at, above, and below the rule’s dose. Heating would be verified by two independent sensors, one of which has characterized lag. Outcomes would be osteocyte lacunar emptiness and bone-implant contact at pre-specified times, scored blind. Drilled sites would form a second arm in which the real transients are recorded and the dose predicted from the controlled-heating calibration is compared with the observed injury. Cortical and cancellous bone would be analyzed as separate strata, and in vivo work would require ethics approval, which has not been sought or granted for this manuscript. Deliberately heating human bone to injurious levels is not proposed; observational recording during routine surgery, with outcome follow-up, is the feasible human design.
Sample-size reasoning
For the controlled-heating experiment, the comparison of two dose-matched exposures is an equivalence question and the comparison of dose levels is a difference question. As an illustration of the arithmetic for the difference question, assume the outcome is bone-implant contact expressed as a percentage, with a between-sample standard deviation σ of 10 percentage points (an assumption, not a literature value), a difference Δ of 10 percentage points to be detected, a two-sided α of 0.05, and 80% power. The normal-approximation sample size per group is
n=(2 σ^2 (z_(1−α/2)+z_(1−β))^2)/(Δ^2)
with z_(1−α/2) = 1.96 and z_(1−β) = 0.84, which gives about 16 per group (computed with python; a t-based correction raises it by one or two). For the reporting audit, the number of studies needed to estimate the proportion that report a time series with precision d = 0.10 at 95% confidence and an assumed proportion p = 0.5 is n=z^2p(1−p)/d^2, about 97 studies. These figures are placeholders for planning and depend on assumptions that pilot data would have to confirm. Equivalence testing of dose-matched exposures would need a prespecified margin that no existing data justify, and would require larger samples than the difference test for the same σ.
Feasibility
5. Competing explanations and how the design discriminates among them
The disagreement among animal studies and between animal studies and clinical trials can be explained in several ways other than exposure expression. Each explanation makes a distinct prediction.
Species, site, and bone type. Thresholds could differ because rabbit tibial metaphysis, sheep iliac crest, rat tibia, and human mandible have different perfusion, mineralization, and remodeling rates. The animal review found cancellous bone more affected than cortical [11]. If this is the main cause, then dose-matched exposures would still give different injury across species and bone types (P5), and re-expression would not reduce between-study heterogeneity after adjustment for species. The design discriminates by requiring the species and bone-type covariate in the primary model and by running the controlled-heating experiment in more than one bone type.
Measurement error. Sensor position, thermal contact, and lag may account for much of the variation in reported peaks. Drilling temperature is a rapidly changing signal, and a slow sensor reports a smaller and later peak. If this dominates, recorded doses will be systematically underestimated (P4) and the apparent threshold will vary with the sensor used. The design addresses this by requiring lag characterization and by the two-sensor arrangement in the validation experiment. Re-expression of tier 1 studies would be stratified by sensor type.
Outcome insensitivity and repair. Radiographic marginal bone loss at 12 months may not register early, localized necrosis that is repaired by remodeling, and heat may stimulate progenitors [14]. If so, trials would find no difference even when doses differ, and the lack of effect in unirrigated-drilling trials [16, 17] would be uninformative about the threshold. The design discriminates by using histology and implant-contact outcomes in the animal experiment and by treating radiographic bone loss as a separate, low-sensitivity stratum in the review.
Threshold not reached in the trials. Drilling without irrigation at low speed may simply keep temperature below the dose of concern. The systematic review reporting temperatures below 47 °C in both arms [15] supports this reading. If it is true, the trials do not conflict with the rule; they are outside its range. The re-expression would test this by computing dose from trials that recorded temperature (tier 1 or 2) and showing whether both arms fall below the rule’s dose (P3). If instead unirrigated arms exceed the dose without worse outcomes, the rule’s transfer to human jaw bone is questioned.
The dose formalism itself is inadequate. Hyperthermia dose was derived from cell survival, and Pearce points out its limited range and single-process nature [4]. If bone injury at drilling temperatures involves processes with different activation energies, a single dose will fail to order exposures. P1 would then be refuted, and the appropriate conclusion would be that peak temperature plus duration at several cutoffs is the more defensible report. This is a real possibility, and the paper’s reporting standard is designed to remain useful in that case because it asks for the trace and the duration above several cutoffs, not only the dose.
6. Clinical and engineering implications
The statements below are conditional on the hypothesis being supported by the validation work and are not current recommendations.
If dose proves better than peak temperature at ordering injury, investigators and manufacturers of temperature-sensing drills, irrigation systems, and guided-surgery sleeves could compare designs by dose per osteotomy. Guided drilling is said to generate more heat than free-hand drilling [21]. Technique studies such as those on drilling force [19], operator technique and bur design [20], piezosurgery [22], and drill material [23] could be pooled by dose.
If long, mild elevations contribute meaningful dose, then protocols that avoid high peaks but allow prolonged warm periods (for example slow, continuous drilling in dense bone with inadequate cooling) would be flagged, which a peak criterion does not do.
If dose is not better than peak plus duration, then the practical output is simpler: clinicians and researchers should report time above specified cutoffs, and the 47 °C figure should be described as one point on a time-temperature curve. In either case, the teaching statement that drilling must remain below 47 °C could be restated with its origin and limits, without discarding the preference for irrigation, which has other rationales (debris removal, drill life) not examined here.
For engineering, the main implication concerns sensing. Because a one-degree sensor error doubles dose under R = 0.5, sensor characterization (lag, placement, calibration tolerance) matters more for dose than for peak-only reporting, and the choice among thermocouples, thermography, and fiber-optic sensors should rest on dynamic accuracy.
7. Limitations
Evidence base. The appraisal rests on abstracts. Study sizes, outcome definitions, risk of bias, and the actual temperature protocols were not seen for most papers, and the sample size of the pivotal 1984 rabbit study could not be verified. Two foundational papers [5, 6] were known by title only. The thermal-dose search covered PubMed and Europe PMC; Embase, Scopus, and Web of Science were not accessible, and no patent or standards search was performed because no device or test method is claimed. Other applications of thermal-dose models to bone cutting may exist in engineering literature that these databases index poorly.
Dose model validity. The conventional R values and the Arrhenius form come from hyperthermia and thermal-damage practice, not from bone. The Arrhenius calibration used here is circular by construction: it was set to make the rule equal one unit of damage. Therefore Table 2 and the synthetic transients express what the rule implies, not what bone does. The equivalent-minute relation is described as covering a single process over a limited temperature range [4], while drilling peaks reach temperatures above that range in some reports, and bone injury is likely to involve several processes [14].
Spatial and thermal-property issues. Bone conductivity, perfusion, and irrigation heat sink were not modeled, and a validated dose-based volume criterion for jaw bone is not available [30].
Outcome linkage. Even a perfect dose estimate would map onto histological necrosis, and the relationship of necrosis to clinical failure or marginal bone loss is a further step. Radiographic bone loss at 12 months in small trials has limited power to detect effects of modest necrosis [16, 17], and trials may not test the rule if temperature stays below it [15].
Heterogeneity and ecological inference. Re-expressing study-level summary data as dose creates an ecological analysis. Between-study differences in species, site, and method will be confounded with dose, and a regression across studies cannot establish causation. The proposed validation experiment is the safeguard, not the review.
8. Conclusions
The 47 °C for 1 minute rule has a traceable origin in rabbit experiments with controlled, constant heating at 44, 47, and 50 °C [1]. It is a legitimate observation about sustained heating in that model. It is not a validated limit for human jaw osteotomy, where later animal work has placed the threshold between 47 and 55 °C [11] and where trials of unirrigated low-speed drilling have not shown a marginal bone penalty [16, 17], possibly because the exposures never reach the dose in question [15].
Expressing exposure as a thermal dose is a plausible way to make these observations comparable, and the computations here show that peak temperature and dose can diverge by an order of magnitude or more for transients with similar peaks. The idea is not original in bone cutting: a CEM43 and Arrhenius model has already been developed for ultrasonic skull grinding [29]. The remaining contribution of this manuscript is the appraisal of the dental rule, the explicit mapping of the rule onto dose, and a reporting standard that would let the question be tested. Whether dose improves on peak temperature and time above cutoff is an empirical question, and the manuscript specifies the results that would refute it.
Cite this paper
Jain S, Gupta A. The 47 °C for One Minute Rule: Origin, Generalizability, and a Thermal-Dose Reformulation for Implant Osteotomy. Working paper. Center for Implant Dentistry; 2026. Available from: https://bayareaimplantdentistry.com/clinician-research-and-publications/bone-heat-threshold-thermal-dose
References
- Eriksson RA, Albrektsson T. The effect of heat on bone regeneration: an experimental study in the rabbit using the bone growth chamber. J Oral Maxillofac Surg. 1984;42(11):705-11. doi:10.1016/0278-2391(84)90417-8 (opens in a new tab). PMID: 6593442 (opens in a new tab).
- Sapareto SA, Dewey WC. Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys. 1984;10(6):787-800. doi:10.1016/0360-3016(84)90379-1 (opens in a new tab). PMID: 6547421 (opens in a new tab).
- van Rhoon GC, Samaras T, Yarmolenko PS, Dewhirst MW, Neufeld E, Kuster N. CEM43°C thermal dose thresholds: a potential guide for magnetic resonance radiofrequency exposure levels?. Eur Radiol. 2013;23(8):2215-27. doi:10.1007/s00330-013-2825-y (opens in a new tab). PMID: 23553588 (opens in a new tab).
- Pearce JA. Models for thermal damage in tissues: processes and applications. Crit Rev Biomed Eng. 2010;38(1):1-20. doi:10.1615/critrevbiomedeng.v38.i1.20 (opens in a new tab). PMID: 21175400 (opens in a new tab).
- Lundskog J. Heat and bone tissue. An experimental investigation of the thermal properties of bone and threshold levels for thermal injury. Scand J Plast Reconstr Surg. 1972;9:1-80. PMID: 4661159 (opens in a new tab).
- Eriksson AR, Albrektsson T. Temperature threshold levels for heat-induced bone tissue injury: a vital-microscopic study in the rabbit. J Prosthet Dent. 1983;50(1):101-7. doi:10.1016/0022-3913(83)90174-9 (opens in a new tab). PMID: 6576145 (opens in a new tab).
- Augustin G, Zigman T, Davila S, Udilljak T, Staroveski T, Brezak D, et al. Cortical bone drilling and thermal osteonecrosis. Clin Biomech (Bristol). 2012;27(4):313-25. doi:10.1016/j.clinbiomech.2011.10.010 (opens in a new tab). PMID: 22071428 (opens in a new tab).
- Möhlhenrich SC, Modabber A, Steiner T, Mitchell DA, Hölzle F. Heat generation and drill wear during dental implant site preparation: systematic review. Br J Oral Maxillofac Surg. 2015;53(8):679-89. doi:10.1016/j.bjoms.2015.05.004 (opens in a new tab). PMID: 26051868 (opens in a new tab).
- Möhlhenrich SC, Modabber A, Steiner T, Mitchell DA, Hölzle F. Corrigendum to “Heat generation and drill wear during dental implant site preparation: systematic review” [Br. J. Oral Maxillofac. Surg. 53(8) (2015) 679–689]. Br J Oral Maxillofac Surg. 2016;54(1):117. doi:10.1016/j.bjoms.2015.11.005 (opens in a new tab). PMID: 27110618 (opens in a new tab).
- Pandey RK, Panda SS. Drilling of bone: A comprehensive review. J Clin Orthop Trauma. 2013;4(1):15-30. doi:10.1016/j.jcot.2013.01.002 (opens in a new tab). PMID: 26403771 (opens in a new tab).
- Kniha K, Heussen N, Weber E, Möhlhenrich SC, Hölzle F, Modabber A. Temperature Threshold Values of Bone Necrosis for Thermo-Explantation of Dental Implants-A Systematic Review on Preclinical In Vivo Research. Materials (Basel). 2020;13(16). doi:10.3390/ma13163461 (opens in a new tab). PMID: 32781597 (opens in a new tab).
- Trisi P, Berardini M, Falco A, Vulpiani MP. Effect of temperature on the dental implant osseointegration development in low-density bone: an in vivo histological evaluation. Implant Dent. 2015;24(1):96-100. doi:10.1097/ID.0000000000000204 (opens in a new tab). PMID: 25621555 (opens in a new tab).
- Kniha K, Hermanns-Sachweh B, Al-Sibai F, Kneer R, Möhlhenrich SC, Heitzer M, et al. Effect of thermal osteonecrosis around implants in the rat tibia: numerical and histomorphometric results in context of implant removal. Sci Rep. 2022;12(1):22227. doi:10.1038/s41598-022-25581-9 (opens in a new tab). PMID: 36564495 (opens in a new tab).
- Dolan EB, Haugh MG, Tallon D, Casey C, McNamara LM. Heat-shock-induced cellular responses to temperature elevations occurring during orthopaedic cutting. J R Soc Interface. 2012;9(77):3503-13. doi:10.1098/rsif.2012.0520 (opens in a new tab). PMID: 22915633 (opens in a new tab).
- Bernabeu-Mira JC, Soto-Peñaloza D, Peñarrocha-Diago M, Camacho-Alonso F, Rivas-Ballester R, Peñarrocha-Oltra D. Low-speed drilling without irrigation versus conventional drilling for dental implant osteotomy preparation: a systematic review. Clin Oral Investig. 2021;25(7):4251-4267. doi:10.1007/s00784-021-03939-z (opens in a new tab). PMID: 33893557 (opens in a new tab).
- Pellicer-Chover H, Peñarrocha-Oltra D, Aloy-Prosper A, Sanchis-Gonzalez JC, Peñarrocha-Diago MA, Peñarrocha-Diago M. Comparison of peri-implant bone loss between conventional drilling with irrigation versus low-speed drilling without irrigation. Med Oral Patol Oral Cir Bucal. 2017;22(6):e730-e736. doi:10.4317/medoral.21694 (opens in a new tab). PMID: 29053645 (opens in a new tab).
- Bernabeu-Mira JC, Peñarrocha-Diago M, Canullo L, Camacho-Alonso F, Rojo-Sanchis J, Peñarrocha-Oltra D. Peri-Implant Marginal Bone Changes Using High-Speed Drilling With Irrigation Versus Low-Speed Drilling Without Irrigation: A Randomized Clinical Trial With 12 Months of Follow-Up. Clin Oral Implants Res. 2025;36(6):698-709. doi:10.1111/clr.14416 (opens in a new tab). PMID: 39950686 (opens in a new tab).
- Bernabeu-Mira JC, Peñarrocha-Diago M, Peñarrocha-Diago M, Romero-Gavilán F, Camacho-Alonso F, Peñarrocha-Oltra D. Comparison of patient-centered outcomes measures between low-speed drilling without irrigation and high-speed drilling with irrigation: A randomized clinical trial. Clin Oral Implants Res. 2024;35(1):21-30. doi:10.1111/clr.14189 (opens in a new tab). PMID: 37800659 (opens in a new tab).
- Bachus KN, Rondina MT, Hutchinson DT. The effects of drilling force on cortical temperatures and their duration: an in vitro study. Med Eng Phys. 2000;22(10):685-91. doi:10.1016/s1350-4533(01)00016-9 (opens in a new tab). PMID: 11334754 (opens in a new tab).
- Kerawala CJ, Martin IC, Allan W, Williams ED. The effects of operator technique and bur design on temperature during osseous preparation for osteosynthesis self-tapping screws. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 1999;88(2):145-50. doi:10.1016/s1079-2104(99)70108-3 (opens in a new tab). PMID: 10468456 (opens in a new tab).
- Saxena V, Dhawan P, Rani S. Effect of Guided Implant Drilling on Bone Temperature Changes During Implant Osteotomy: A Comprehensive Systematic Review. Cureus. 2024;16(9):e70216. doi:10.7759/cureus.70216 (opens in a new tab). PMID: 39463655 (opens in a new tab).
- Jain N, Dhawan P, Rani S, Kalra J. Heat Generation and Pain Assessment in Piezosurgery Versus Conventional Drilling for Implant Placement: A Systematic Review. Cureus. 2024;16(10):e71396. doi:10.7759/cureus.71396 (opens in a new tab). PMID: 39539850 (opens in a new tab).
- Augusto Alves Bento V, Marcela de Luna Gomes J, Davi Del Rei Daltro Rosa C, Pedro Justino de Oliveira Limirio J, Micheline Dos Santos D, Coelho Goiato M, et al. Steel drills versus zirconia drills on heat generation at the surgical site of dental implants: A systematic review and meta-analysis. Saudi Dent J. 2024;36(1):1-10. doi:10.1016/j.sdentj.2023.09.001 (opens in a new tab). PMID: 38375375 (opens in a new tab).
- Barriga-Yauri G, Alvitez-Temoche D, Mauricio F, Espinoza-Carhuancho F, Calderon I, Medina J, et al. Influence of Infrared Thermography Predictors in Dental Implant Osteotomies: An Ex Vivo Study. Int Dent J. 2025;75(3):1683-1690. doi:10.1016/j.identj.2025.03.009 (opens in a new tab). PMID: 40163954 (opens in a new tab).
- Hu S, Wu M, Zhang Z, Guan Y, Ma S, He S, et al. Application of fiber bragg grating sensing to thermal variation during dental implant site preparation: an in vitro study. BMC Oral Health. 2026;26(1). doi:10.1186/s12903-026-08704-w (opens in a new tab). PMID: 42204520 (opens in a new tab).
- Balkanlioğlu AC, Acikan İ. Effect of Osteotomy Depth on Intraosseous Temperature During Implant Drilling: A Standardized CAD/CAM-Guided Thermocouple Study. Int J Oral Maxillofac Implants. 2026;0(0):1-22. doi:10.11607/jomi.11735 (opens in a new tab). PMID: 42090571 (opens in a new tab).
- De Tommasi F, Massaroni C, Grasso RF, Carassiti M, Schena E. Temperature Monitoring in Hyperthermia Treatments of Bone Tumors: State-of-the-Art and Future Challenges. Sensors (Basel). 2021;21(16). doi:10.3390/s21165470 (opens in a new tab). PMID: 34450911 (opens in a new tab).
- Yarmolenko PS, Moon EJ, Landon C, Manzoor A, Hochman DW, Viglianti BL, et al. Thresholds for thermal damage to normal tissues: an update. Int J Hyperthermia. 2011;27(4):320-43. doi:10.3109/02656736.2010.534527 (opens in a new tab). PMID: 21591897 (opens in a new tab).
- Babbar A, Jain V, Gupta D, Agrawal D. Finite element simulation and integration of CEM43 °C and Arrhenius Models for ultrasonic-assisted skull bone grinding: A thermal dose model. Med Eng Phys. 2021;90:9-22. doi:10.1016/j.medengphy.2021.01.008 (opens in a new tab). PMID: 33781484 (opens in a new tab).
- Chen YC, Tsai YJ, Hsiao HY, Chiu YW, Hong YY, Tu YK, et al. Assessment of Thermal Osteonecrosis during Bone Drilling Using a Three-Dimensional Finite Element Model. Bioengineering (Basel). 2024;11(6). doi:10.3390/bioengineering11060592 (opens in a new tab). PMID: 38927828 (opens in a new tab).

