A Levels in the Gingival Crevicular Fluid of Teeth with External Apical Root Resorption Robert Babak Marzban

University of Connecticut OpenCommons@UConn

SoDM Masters Theses School of Dental Medicine

June 1999 TNF- a Levels in the Gingival Crevicular Fluid of Teeth with External Apical Root Resorption Robert Babak Marzban

Follow this and additional works at: https://opencommons.uconn.edu/sodm_masters

Recommended Citation Marzban, Robert Babak, "TNF- a Levels in the Gingival Crevicular Fluid of Teeth with External Apical Root Resorption" (1999). SoDM Masters Theses. 86. https://opencommons.uconn.edu/sodm_masters/86 TNF- c LEVELS IN THE GINGIVAL CREVICULAR FLUID.

OF TEETH WITH EXTERNAL APICAL ROOT RESORPTION

Robert Babak Marzban

B.S. George Mason University

D.D.S. Medical College of Virginia

A Thesis

Submitted in Partial Fulfillment of the

Requirements for the Degree of

Masters of Dental Science

at the

University of Connecticut

1999 APPROVAL PAGE

Masters of Dental Science

TNF-ot LEVELS IN THE GINGIVAL CREVICULAR FLUID OF TEETH WITH EXTERNAL APICAL ROOT RESORPTION

Presented by

Robert Babak Marzban, D.D.S.

Major Advisor

Associate Advisor Edward F. Rossomando

Associate Advisor Kamran E.

Associate Advisor Ravindra Nanda

University of Connecticut 1999

ii ACKNOWLEDGMENTS

My ultimate goal to become a practicing member of the health profession will finally come to fruition. It took only 24 short years of education. Despite previous degrees and graduations, the completion of this Master's Thesis is the true mark for the end of my formal education. The long list of people deserving thanks for aiding me throughout this process should begin with my family.

No words could describe the appreciation that my family deserves. My wife

Tourang, who remained by my side for the last eight years, despite turbulent and trying times. You provided me with insight, support, and love in those time when it was needed most. My father, who instilled the desire and the ethics to work, learn, and succeed by example. My mother taught me to have passion for life and family. And my sister who taught me how to be kind and have compassion (I still need a great deal of work in this area).

I would like to thank Dr. Norton, my major advisor, for providing me with direction throughout this project. Without you, Dr. Rossomando, and Dr. Safavi, I would still be working on my lit review. For giving me the opportunity to obtain the best post- graduate education possible and for getting my name in Medline, I would like to thank

Dr. Nanda. Not only did you accept me into your program, but also into your family and circle of friends. Fighting through the long line of past and future residents I would like to extend my sincere gratitude to Dr. Andrew Kuhlberg who is the true resident's advocate, who taught me how to stay on the straight and narrow, who helped keep my

iii sanity, and who shared his office with me despite my poor choice in music. Thanks also to Drs. Pae and Burleson who, along with Dr. Kuhlberg, kept statistical anarchy at a minimum.

Second to my great looks, my most cherished attribute remains to be my group of

Mends. Aside for being the sounding board for my complaints and brain storms, my co- residents who have grown to be excellent friends have encouraged me to work hard while playing hard. A special thanks goes to Dr. O'Hea, without whom my sample size could be enumerated on a single hand. I would also like to thank Dr. Wadhwa for providing me with access to his computer and for making sure that my nutritional requirements are not neglected. Finally, thanks to Drs. McKenna, Priebe, and Nadeau for their help in improving our sampling technique.

iv TABLE OF CONTENTS

ABSTRACT 1

INTRODUCTION 2

REVIEW OF LITERATURE 4 TERMINOLOGY 4 INCIDENCE 5 INDIVIDUAL SUSCEPTIBILITY 6 BIOLOGY OF ROOT RESORPTION IN RESPONSE TO TOOTH MOVEMENT 8 Hyalinization 8 Odontoclasts 9 Initiation ofRoot Resorption 10 Restoration ofresorbed root structure 12 Resistance to resorption 13 CLINICAL CORRELATIONS 15 Force magnitude 15 Treatment duration 17 Age 18 Systemicfactors and diet 19 Alveolar bone density 20 Root canal treatment, trauma, andpernicious habits 20 Vulnerability ofspecific teeth 21 Mechanics 22 Retention and post-retention 23 DENTAL HEALTH AFTER ROOT RESORPTION 24 CYTOKINES AND HARD-TISSUE METABOLISM 25 lnterleuln-1 26 Tumor Necrosis Factor 28 Involvement in dental pathology 29 TOOTH ERUPTION MODEL 31 CYTOKINE INVOLVEMENT IN ORTHODONTICS 33 GINGIVAL CREVICULAR FLUID 36 RATIONALE ]9 A NON-INVASIVE TECHNIQUE TO SAMPLE THE PDL IN HUMANS 39 PREVIOUS DATA 40 SPECIFIC AIMS 41 RESEARCH METHODS 43 STUDY POPULATION 43 DATA COLLECTION 43 CLINICAL DETERMINATION AND SCORING 45 SAMPLING FROM THE GINGIVAL SULCUS 47 QUANTIFICATION OF BEAD-BOIYND ANTIBODIES 48 DATA ANALYSIS 49 DIFFICULTIES SIGNIFICANCE 4 RESULTS 55 SAMPLE SELECTION 55 DESCRIPTION OF THE SAMPLE 56 DETERMINATION OF ROOT LOSS 58 TUMOR NECROSIS FACTOR-o: MEASUREMENTS 59 ZERO TNF- MEASUREMENTS 60 CORRELATION TO TNF-c 61 PATTERNS SPECIFIC TO PATIENTS 61 STEPWISE MULTIVARIABLE REGRESSION MODEL 62 DISCUSSION DESCRIPTION OF THE SAMPLE 65 DETERMINATION OF ROOT LOSS 66 TNF-c LEVELS IN THE GINGIVAL SULCUS 67 THE ASSOCIATION BETWEEN TNF-tz AND CLINICAL VARIABLES 69 The association between root loss and TNF-z values 70 Gingival conditions and TNF-cr levels 71 TNF-z in the gingival crevicularfluid 72 GENERAL DISCUSSION 74 FUTURE STUDIES 76 CONCLUSION 78 APPENDIX I: FIGURES APPENDIX II: TABLES ...... 98 APPENDIX III. CONSENT FORMS 161 INFORMED CONSENT FOR ADULTS 161 INFORMED CONSENT FOR DEPENDENTS 164 BIBLIOGRAPHY 167

vi LIST OF FIGURES

FIGURE 1. BONE METABOLISM. CONTRIBUTION OF CYTOKINES, HORMONES, AND

VARIOUS GROWTH FACTORS IN REGULATION AND STIMULATION OF BONE

METABOLISM 79

FIGURE 2. METHODS #1 AND #2 FOR DETERMINING ROOT LOSS 80 FIGURE 3. IMMUNOMAGNETIC BEADS, 5tM IN DIAMETER, ARE COATED WITH ANTI-TNFct ANTIBODIES AND USED TO CAPTURE TNF-c MOLECULES FROM

THE GINGIVAL SULCUS 81

FIGURE 4. ELISA IS USED TO QUANTIFY THE AMOUNT OF TNFo HARVESTED

FROM THE GINGIVAL SULCUS 82

FIGURE 5. DISTRIBUTION OF TEETH IN OUR SAMPLE 83

FIGURE 6. DISTRIBUTION OF NORMAL AND DEVIATING ROOT SHAPES 84

FIGURE 7. DISTRIBUTION OF THE PLAQUE INDEX 84

FIGURE 8. DISTRIBUTION OF THE GINGIVAL INDEX 85

FIGURE 9. DISTRIBUTION OF THE PERIODONTAL INDEX 85

FIGURE 10. TIME SINCE SUBJECTS LAST BRUSHED THEIR TEETH UNTIL THE TIME

OF SAMPLING (IN HOURS) 86

FIGURE 11. TYPES OF TOOTH MOVEMENT ATTEMPTED DURING THE PERIOD

PRIOR TO SAMPLING 87

FIGURE 12. FORCE APPLICATION DETERMINED AT THE VISIT PRIOR TO SAMPLING

(APPLIED) AND AT THE TIME OF SAMPLING (REMAINING) 88

FIGURE 13. TIME IN TREATMENT FROM INITIAL BANDING VISIT (MONTHS) 89

FIGURE 14. AMOUNTS OF ROOT LOSS IN OUR SAMPLE AS DETERMINED FROM

METHOD #1 (RL) 90

vii FIGURE 15. AMOUNTS OF ROOT LOSS AS DETERMINED FROM METHOD #3 (RRI)..91

FIGURE 16. TNF-c DISTRIBUTION IN OUR SAMPLE (PICOGRAMS) 92

FIGURE 17. TNF-et VALUES TABULATED AS AN INDEX 93

FIGURE 18. TNF-a VALUES TRANSFORMED USING THE NATURAL LOGARITHM 94

FIGURE 19. DISTRIBUTION: SAMPLES WITH ZERO TNF-a VALUES ARE OMITTED..95

FIGURE 20. DISTRIBUTION OF TABULATED DATA USING THE TNF INDEX AND

OMITTING SAMPLES WITH ZERO VALUES 96

FIGURE 21. DISTRIBUTION OF TRANSFORMED DATA USING NATURAL

LOGARITHM AND OMITTING SAMPLES WITH VALUES OF ZERO 97

viii LIST OF TABLES

TABLE 1. INCIDENCE OF ROOT RESORPTION IN NON-ORTHODONTICALLY

TREATED INDIVIDUALS 98

TABLE 2. INCIDENCE OF ROOT RESORPTION SUBSEQUENT TO ORTHODONTIC

TREATMENT 99

TABLE 3. REPORTED AMOUNTS OF ROOT RESORPTION SUBSEQUENT TO

ORTHODONTIC TREATMENT 100

TABLE 4. CYTOK1NES WITH EFFECTS ON BONE METABOLISM 101

TABLE 5. ACTIONS COMMON TO BOTH INTERLEUKIN -1 AND TUMOR NECROSIS

FACTOR 102

TABLE 6. MOLECULES FOUND IN GINGIVAL CREVICULAR FLUID 103

TABLE 7. COMPOSITION OF THE GINGIVAL SULCUS. 236 104

TABLE 8. CRITERIA FOR PATIENT SELECTION AND DISQUALIFICATION 105

TABLE 9. SAMPLE SUMMARY. THE ABOVE FORM WILL BE USED TO COLLECT

DATA AT EACH VISIT 106

TABLE 10. ROOT RESORPTION INDEX 107

TABLE 11. THE PERIODONTAL INDEX (PI).237 107

TABLE 12. GINGIVAL INDEX (GI).224 108

TABLE 13. PLAQUE INDEX.225 108

TABLE 14. DISTRIBUTION OF TEETH INCLUDED IN OUR SAMPLE. ALSO REFER TO

FIGURE 5 109

TABLE 15. DISTRIBUTION OF NORMAL AND DEVIATING ROOT SHAPES. ALSO

REFER TO FIGURE 6 110

ix TABLE 16. DISTRIBUTION OF THE PLAQUE INDEX. ALSO REFER TO FIGURE 7 110

TABLE 17. DISTRIBUTION OF THE GINGIVAL INDEX. ALSO SEE FIGURE 8 111

TABLE 18. DISTRIBUTION OF THE PERIODONTAL INDEX. ALSO SEE FIGURE 9 111

TABLE 19. TIME SINCE SUBJECTS LAST BRUSHED THEIR TEETH UNTIL THE TIME

OF SAMPLING (IN HOURS). ALSO REFER TO FIGURE 10 112

TABLE 20. TYPES OF TOOTH MOVEMENT ATTEMPTED DUR1NG THE PERIOD PRIOR

TO SAMPLING 112

TABLE 21. FORCE APPLICATION DETERMINED AT THE VISIT PRIOR TO SAMPLING

(APPLIED) AND AT THE TIME OF SAMPLING (REMAINING). ALSO REFER TO

FIGURE 11 113

TABLE 22. TIME IN TREATMENT FROM INITIAL BANDING VISIT (MONTHS) 113

TABLE 23. DESCRIPTIVE STATISTICS OF THE THREE METHODS USED TO ASSESS

ROOT RESORPTION 114

TABLE 24. DESCRIPTIVE STATISTICS OF THE DIFFERENCE BETWEEN REPEATED

MEASURES 115

TABLE 25. REPEATABILITY (ERROR) OF THE THREE METHODS FOR

DETERMINATION OF ROOT LOSS 116

TABLE 26. CORRELATION BETWEEN THE THREE METHODS USED TO ASSESS ROOT

LOSS 116

TABLE 27. DISTRIBUTION OF TUMOR NECROSIS FACTOR ALPHA (TNF-a) VALUES. 117

TABLE 28. DISTRIBUTION OF RAW TNF-c VALUES IN OUR SAMPLE (PICOGRAMS). 118

TABLE 29. TNF-ct VALUES TABULATED USING AN ORDINAL INDEX 0-5 119

TABLE 30. TNF-c VALUES TRANSFORMED USING THE NATURAL LOGARITHM 119

TABLE 31. T-TEST FOR TOOTH SELECTION 120 TABLE 32. T-TEST FOR ROOT SHAPE 121

TABLE 33. T-TEST FOR TIME SINCE BRUSHING 122

TABLE 34. T-TEST FOR THE TYPE OF TOOTH MOVEMENT 123

TABLE 35. T-TEST FOR THE TIME IN TREATMENT 124

TABLE 36. T-TEST FOR THE PLAQUE INDEX 125

TABLE 37. T-TEST FOR THE GINGIVAL INDEX 126

TABLE 38. T-TEST FOR THE PERIODONTAL INDEX 127

TABLE 39. T-TEST FOR ROOT LOSS (RL) 128

TABLE 40. T-TEST FOR CORRECTED ROOT LOSS (RL*) 129

TABLE 41. T-TEST FOR THE ROOT RESORPTION INDEX 130

TABLE 42. DISTRIBUTION WITH SAMPLES CONTAINING ZERO TNF VALUES

OMITTED 131

TABLE 43. DISTRIBUTION OF TNF VALUES WITH SAMPLES OF ZERO OMITTED... 131

TABLE 44. DISTRIBUTION OF TNF VALUES SORTED ACCORDING TO THE TNF

INDEX WITH ZERO VALUES OMITTED. FOR COMPARISON, REFER TO TABLE

29 132

TABLE 45. DISTRIBUTION OF TNF DATA WITH NO SAMPLES OF ZERO WHEN

TRANSFORMED USING NATURAL LOGARITHM. FOR COMPARISON, REFER TO

TABLE 30 132

TABLE 46. CORRELATION COEFFICIENTS (TABLES 46-48) 133

TABLE 47. CORRELATION COEFFICIENTS (TABLES 46-48) 134

TABLE 48. CORRELATION COEFFICIENTS (TABLES 46-48) 135

TABLE 49. CORRELATION COEFFICIENTS, OMITTING SAMPLES WITH TNF VALUES

OF ZERO (TABLES 49-51). FOR LEGEND SEE TABLE 51 136

xi TABLE 50. CORRELATION COEFFICIENTS, OMITTING SAMPLES WITH TNF VALUES

OF ZERO (TABLES 49-51) 137

TABLE 51. CORRELATION COEFFICIENTS, OMITTING SAMPLES WITH TNF VALUES

OF ZERO (TABLES 49-51) 138

TABLE 52. CORRELATION COEFFICIENTS BETWEEN ROOT LOSS AND RAW TNF

SCORES WEP CALCULATED FOR EACH PATIENT (N=19) 139

TABLE 53 DISTRIBUTION OF CLINICAL VARIABLES (TABLES 53-55) 140

TABLE 54. DISTRIBUTION OF CLINICAL VARIABLES (TABLES 53-55) 140

TABLE 55. DISTRIBUTION OF ROOT LOSS MEASURES AND TNF VALUES (TABLES

53-55) 141

TABLE 56. REGRESSION MODEL: Y= TNF, X= GI + MVMT + RRI 142

TABLE 57. REGRESSION MODEL: Y= TNF, X= GI + RRI 143

TABLE 58. REGRESSION MODEL" Y= TNF, X= GI 144

TABLE 59. REGRESSION MODEL" Y= TNF, X= RRI 1.45

TABLE 60. REGRESSION MODEL: Y= RANKED TNF, X= PLI + GI + RRI 146

TABLE 61. REGRESSION MODEL: Y= RANKED TNF, X= GI + RRI 147

TABLE 62. REGRESSION MODEL: Y= RANKED TNF, X GI 148

TABLE 63. REGRESSION MODEL: Y= RANKED TNF, X= RRI 149

TABLE 64. REGRESSION MODEL: Y= LN (TNF+ 1), X PLI + GI + RRI 150

TABLE 65. REGRESSION MODEL: Y= LN (TNF+ 1), X GI + RRI 151

TABLE 66. REGRESSION MODEL: Y= LN (TNF+ 1), X GI 152

TABLE 67. REGRESSION MODEL: Y= LN (TNF+ 1), X RRI 153

TABLE 68. REGRESSION MODEL: Y= TNF INDEX, X PLI + GI + RRI 154

TABLE 69. REGRESSION MODEL: Y=TNF-0'S, X=GI 155

xii TABLE 70. REGRESSION MODEL: Y=TNF-0'S, X=RRI 156

TABLE 71. REGRESSION MODEL: Y= RANKED TNF-0'S, X=GI + BRS + RRI 157

TABLE 72. REGRESSION MODEL: Y= RANKED TNF-0'S, X=GI 158

TABLE 73. REGRESSION MODEL: Y= RANKED TNF-0'S, X=BRS 159

TABLE 74. REGRESSION MODEL: Y= RANKED TNF-0'S, X=RRI 160

xiii ABSTRACT

Despite exhaustive research, the etiology and mechanism for the progression of extemal apical root resorption (EARR) subsequent to orthodontic treatment still remains evasive. The literature is replete with statistical correlates of clinical variables and root resorption while there is little understanding of the cell biology of the underlying mechanism, particularly in the human. This study evaluated the relationship between a cytokine, which is known to be a significant participant in hard tissue metabolism, and the incidence of root resorption in a population of patients who received comprehensive orthodontic treatment. The levels of Tumor Necrosis Factor- Alpha (TNF-a) in the gingival crevicular fluid of 85 teeth were measured and evaluated against the amount of

EARR and an array of clinical variables. The rationale for the use of an indirect technique in monitoring root resorption is discussed. With respect to gingival inflammation, the TNF-o data from this study reflected patterns seen in previous research involving other cytokines. Also noted was the possible contributory role of gingival inflammation as a confounding variable when evaluating the levels of cytokines which are also inflammatory mediators. Regression analyses did not show a consistent model where EARR and TNF-ot levels display a linear relationship. However, differences with regard to the severity of root resorption were found between those samples which had a measurable level of TNF-c and those where TNF-ct was absent (p<0.001). INTRODUCTION

Root resorption is a common sequela of orthodontic treatment. Although a small degree of resorption is almost universally seen, the longevity and the function of the affected teeth remains uncompromised. The benefits provided through orthodontic therapy typically outweigh the risk of slight root modification. Infrequently, the degree of root resorption exceeds what is expected from conventional orthodontic therapy and modification of treatment goals becomes a necessity. This progressive resorptive pattern is rare, and although there are some statistical correlations, it is largely considered unpredictable.

Over the last eighty years, a great deal of effort has been exerted in search of the etiology and predictability of root resorption through clinical variables. A review of literature uncovers a jumble of, sometimes conflicting, clinical information aimed at finding the one clinical procedure or combination of perameters responsible for eliciting the process. However, a reliable cause and effect relationship has not been found. Pure clinical studies have provided information regarding directions in which to further investigate.

Histologic studies are now becoming more commonplace in the root resorption literature indicating a more finite approach. With the advent of new technology in molecular biology and the existing database, it is possible to examine root resorption as it occurs and to describe the precipitative events. The biology behind bone metabolism, orthodontic tooth movement, and tooth eruption is better understood and may serve as a model from which to study root resorption. These processes require the interplay of various, but common, cells and their products. Admittedly, the list of biologic variables is even more vast than the clinical variables which may be examined in orthodontic root resorption. The purpose of this study is to examine if levels of a mediator which is important in all the above mentioned processes is predictive for the incidence and severity of root resorption. REVIEW OF LITERATURE

TERMINOLOGY

The investigation of the process of root resorption began with primary teeth. This research was extended to permanent teeth in the mid 1800's. The relationship between orthodontic treatment and root resorption of permanent teeth was first documented by

Ottolengui in 1914.1 Other studies soon followed in hopes of uncovering a specific etiology.2-7 An enumeration of the processes which cause root resorption includes: normal physiologic tooth movement (drift), pressure fi'om an adjacent impacted tooth, periapical or periodontal inflammation, tooth implantation or replantation, continuous occlusal trauma, tumors or cysts, metabolic or systemic disturbances, local functional or behavioral problems, untoward forces secondary to orthodontic treatment, and idiopathic. 8-15

It would seem that any force, whether natural, pathologic, or iatrogenic, produces activation of osteoclastic activity. Andreasen defines three types of external root resorption. Surface resorption is a self limiting process that occurs on small areas of the root surface. This is the resorption commonly seen after orthodontic treatment. It may be undetectable by x-ray and transient, when the damage is minimal. Surface resorption is followed with immediate repair by adjacent parts of the PDL where vital. Repair occurs by way of restoration of the root surface with a cementum like tissue. Inflammatory resorption indicates the progression of the initial resorption into dentinal tubules which lead to an infected necrotic area, either the pulp or a leukocyte zone. Replacement of resorbed tooth substance by bone is termed replacement resorption and results in sites of ankylosis. 16, 17 Tronstad further defines inflammatory resorption as transient or progresSive depending on the degree of the stimulus. Transient inflammatory resorption would be the response to minimal stimulation resulting in a situation similar to

Andreasen's surface resorption. Long-term stimulation from increased pressure, infection, sharp edges, or a disease state is described as progressive inflammatory resorption. 18

Surface or transient inflammatory resorption occurs with orthodontic tooth movement. However, both of these events are by definition transient, mostly undetectable by demal x-rays, and soon repaired. This cannot fully describe the majority of external root resorption which occurs with orthodontic tooth movement. Tronstad provides a more practical approach to describing root resorption in three categories.

Progressive inflammatory resorption is tissue-pressure related. Upon removal of the stimulus, the process is halted. This process may results in apical root shortening and may be severe in some cases. Cervical resorption occurs commonly, but is infrequent subsequent to orthodomic therapy. The process is thought to be a result of injury to the cervical attachment apparatus. Replacement resorption is described as stated above by

Andreasen. 16, 17

INCIDENCE

Approximately 90% of non-orthodontically treated permanent teeth present with small areas of external root resorption. 19 Harris and Robinson found that 7-10 percent of their non-treated sample exhibited obvious signs of apical root resorption, with severe occurrences in 1-2 percent.20 The incidence of root resorption, as reported in the literature, is highly variable for non-treated and orthodontic patients (Tables 1-3). This is partly due to the variation in the criteria for identifying root resorption between investigators. For example, histologic resorption can easily be found even in the smallest amount, in comparison to detection by dental x-rays. When detected, investigators differ in their criteria for severity of resorption to qualify for incidence. In addition, there differences in the site of examination (apical, cervical, etc.).

Root resorption incurred from orthodontic tooth movement occurs on all surfaces of the root. However, it is most frequently noticed at the apex due to detection by way of dental x-rays, hence the term extemal apical root resorption (EARR). Also, most resorption sites are subsequently restored, except when occurring at the apex of the root.

Mild generalized resorption throughout the dentition is seen in almost every orthodontically treated case ranging 0-3 mm. 19, 21-38 Most resorption sites are relatively small (1 mm wide and 0.1 mm deep) and are soon repaired. Resorption beyond

3mm occurs in approximately 5-10% of orthodontically treated patients.

INDIVIDUAL SUSCEPTIBILITY

Individual susceptibility continues to be the most commonly cited reason for the variability seen in degrees of root resorption after orthodontic treatment. There are differences in susceptibility in relation to macroscopic tooth shape and size, microscopic root shape, and differences in tissue response and activity. Even the process of root resorption seems to vary in the same individual from site to site and from time to time.

Some researchers attribute the variability in the extent of EARR to genetic predisposition. A heritable component for crestal bone loss has been found in twin studies. 39 A few studies have described genetic inheritance of the predisposition for root resorption, however a specific mode of inheritance was not found. 13, 31 A recent study found evidence for a "substantive genetic factor in susceptibility to EARR." Although, the authors admit that there is an obvious genetic relationship of malocclusions between the siblings studied, and therefore, a common regimen of treatment. The similarity in treatment between the siblings may also be a contributing factor to EARR in addition to an "innate genetic predisposition". The authors discussed two independent pathways effecting EARR. They speculated that the most important determinant comes from innate genetic susceptibility, possible involving genotypically determined biochemical events.

The type of malocclusion requiting a specific treatment regime is a lesser, but also important determinant.40

Conflicting conclusions have been drawn regarding gender and its correlation to root resorption. Many studies found a greater severity of resorption in females. 8, 13, 31,

41 Other studies cite an even ratio of incidence and severity between the two sexes and attribute the increased severity found in females to their earlier maturation. 8, 9, 30, 42 BIOLOGY OF ROOT RESORPTION IN RESPONSE TO TOOTH MOVEMENT

Hyalinization

Tooth movement requires bone remodeling. In orthodontic treatment we rely on appliances which deliver forces and create mechanical stresses which are transferred from the tooth to the PDL and in turn, to alveolar bone. This stress induces biochemical responses which evoke structural changes allowing for tooth movement by way of bone remodeling.

As a tooth is displaced in the socket, early movement occurs by PDL compression and tension. This elicits an acute inflammatory response characterized by periodontal vasodilatation and migration of leukocytes out to the periodontal ligament capillaries.43-

45 Cellular and vascular changes are followed by fibrillar alterations. Due to intolerable pressure, cessation of circulation occurs and degenerative changes of the periodontium leads to hyalinization.

With continued application of pressure, cells soon surpass the point of irreversible degeneration.46-48 The adjacent undamaged PDL in the border areas undergoes active hyperemia characterized by high oxygen pressure with rapid blood flow, minimal lymphatic flow, and minimal loss of oxygen, vitamins, and proteins. These conditions are well suited for hard tissue resorbing cells, in contrast to passive edematous hyperemia which occurs with high protein content in the tissue fluid inducing osteoblastic activity.49, 50

The primary hyalinized zones in humans seldom last longer than three weeks.

The continuity between fibrils and cementum is not disrupted with normal orthodontic treatment unless there is a localized focus of heavy pressure where hyalinization is prolonged.50 Pulpal blood flow disturbances and, very rarely, pulp necrosis may occur.

However, neither have been found to potentiate or elicit resorption of the root.29, 51

Odontoclasts

Odontoclasts are the cells responsible for the resorption of roots. They are similar to the osteoclast, large, pleomorphic (50-100 um), multinucleated cells, and most probably ofhematopoietic origin. 58, 59 The cytoplasm is strongly basophilic, granulated, and contains two to ten centrally located nuclei of various shapes and sizes.

Their presence in bone is infrequent, and when found, they are almost always near mineralized matrix. Extensive golgi complexes reside near each nucleus with mitochondria, lysosomes and free ribosomes abundant in the cytoplasm. These motile cells also feature a ruffled border, similar to the osteoclast, constituting a complex folding of the membrane, which is separated by a narrow space from mineralized matrix the sealing zone. They are attracted to mineralized tissues like denuded root areas where vital cementoblasts have been lost. Osteoclast have the capability of demineralizing calcified tissue and degrading organic matrix. 18, 60 Odontoclasts, however, are unable to resorb unmineralized matrix.29, 51

The cytodifferentiation of the odontoclast has been studied by Sahara et al. They found that tartarate resistant acid phosphatase (TRAP) activity correlates with cytodifferentiation and function of the odontoclast. When following this process, it has been noted that TRAP positive mononuclear cells make contact with predentin surfaces 10 on primary teeth. These cells spread out, fuse with each other, and develop into the familiar multinucleated cells with specialized cellular components such as raffled borders and clear zones. Upon termination of resorption, the ruffled border is lost and the cells lose their attachment to the dentin surface. The authors concluded that odontoclasts differentiate from TRAP positive progenitor cells. After attachment to the resorption surface, these cells differentiate into mature odontoclasts which are able to resorb predentin as well as detain.

Initiation ofRoot Resorption

Orthodontic root resorption occurs in areas which are more sensitive to local changes and where physiologic bone resorption originates. Specifically, root resorption begins around hyalinized tissue, and is mediated by cells from the adjacent healthy periodontal ligament. 52-55 The population of cell (first three days) involved in this response are different than those involved in the later phase. Around the periphery of the hyalinized zone, fibroblast-like cells from the adjacent vital parts of the PDL begin removal of necrotic tissue. Multinucleated giant cells infiltrate and remove the main part of the necrotic tissue, including the outer necrotic layer of the root surface. 56, 57

During the earlier phase of resorption, thinning of the cementum occurs more rapidly directly beneath the hyalinized zone than further out to the periphery. Resorption and repair of the PDL and root surface in the periphery is presumably mediated by cells originating from the vital PDL. These cells have lower resorbing potential than those. originating from alveolar bone and marrow. Initiation of a resorptive attack on the root 11 surface experimentally by creating mechanical injury, has been shown to occur more frequently and aggressively if cells repopulating the area come from alveolar bone and marrow. Removal of alveolar bone in this area apparently necessitates repopulation by cells from the PDL and causes less resorption of the roots. 61-64

After removal of the alveolar bone wall and superficial cemental resorption in the peripheral areas, root resorption stops and repair begins even in the presence of continued orthodontic forces. By contrast, necrosis of the whole width of the PDL in the central parts of the hyalinized zone and bone resorption necessary for tooth movement requires presence of clast cells from the marrow. This theory of competitive healing and observation of different cell populations at various times and location indicates that the factor that promotes extensive clastic activity on the root surface may be the source of the cells that repopulate the damaged area. 57

Direct root resorption begins with a hyalinized zone of the periodontium around which areas of resorption soon appear. Indirect resorption begins from Howship lacunae, undermining its way to the surface of the root.24-26, 46-48, 65-67 Elimination of the hyalinized PDL tissue takes 20-25 days, but if still stimulated, the resorption of root structure may continue after this time. 50 At one week, the zone of hyalinization is at its maximum after a single force application. The extent of the of the zone corresponds to the size of resorbed root surface by three weeks. 57 It takes 10 to 35 days for resorption lacunae to appear on the roots, however, this process is not yet visible in radiographs.23,

28, 29, 50, 51 12

Restoration ofresorbed root structure

Upon removal of the applied forces, the resorptive process ceases. The repair process which ensues provides a cementum fill for the lacunae.25, 50 This repair process delays further resorption. Complete healing may take up to 6 weeks in absence of orthodontic forces.26 It has been found that by allowing a tooth to relapse, the resorption process can be halted. 50 By contrast, continued resorption occurs two to three weeks after application of active force when the tooth is retained in the new position. This may be due to passive stress in the PDL, which in the presence of remaining necrotic tissue, elicits continued resorption. The clinical implication is that reactivation of an appliance in the presence of necrotic tissue would likely cause continued resorption of the roots. 57

When the damage to root structure is small, cementoblasts from the vital portions of the periodontal ligament provide repairs. In response to excessive forces, the large amount of damage overwhelms the reparative ability of the cementum and PDL, forcing bone marrow cells and alveolar bone to provide healing. The healing process, when hosted by cell populations from marrow and alveolar bone leads to osteoclastic activation.41, 42

Progressive insults to the root will develop areas of thinner cementum and in instances where root resorption has already begun, the root surface may be devoid of cementum altogether. 68 Until the repair process replenishes the cementum coveting for the root surface, the process which has overwhelmed the repair mechanism will continue to do irreversible damage.24-26 This is exhibited in teeth with previous resorption. 13

Harris observed an increase of 15 % more root loss in teeth with previous signs of resorption prior to beginning orthodontic treatment.69

In examination of the reparative potential of resorbed roots, 28% of the teeth studied showed signs or repair by the first week. By 8 weeks, 75% had begun repair.

The cementum was almost exclusively of the cellular type. During the first 4 weeks of retention, repair with the resorption cavity walls only partially covered with cementum was the most frequently seen. Between 4 to 8 weeks, functional repair with coverage of the total surface of the resorption cavity walls dominated. Individual variation in healing potential was great. 70

Deep lacunae were covered by a layer of cementum which was not thick enough to restore the root to its original anatomy. This results in development of an irregular root shape. When the apical portion of the root was resorbed, the repair cementum did not restore the original length of the tooth. Instead, rough edges are remodeled and a layer of cementum was laid over the apex. The PDL space undergoes bone remodeling in the alveolus following the new root contour.25, 26, 49, 50, 52-57, 65, 71

Resistance to resorption

The process of resorption proceeds more readily through dentin and bone than cementum. The cementum lining the roots of teeth is somewhat resistant to remodeling and resorption. Matrix tissues such as precementum, predentin, and osteoid have been thought to provide resistance to resorption, however, recent data show that continuous pressure can cause resorption of these areas also. 10, 25, 26, 65, 72 The effectiveness of 14

Sharpey's fibers as a prevemative structure is debatable but corroborated by some. 60, 73

Root surfaces are covered with mounds which are formed by .continued mineralization of attached periodontal fibers after the production of cementoid has stopped. Acellular cementum has 100 percent Sharpey fiber content versus actively forming cellular cementum with 50 percent.

Most bisphosphonates inhibit bone resorption. Hydroxyethylidene-1, 1- bisphosphonate (HEBP) inhibits both bone formation and resorption. A single injection in HEBP into rats can inhibit formation of acellular cementum. In its stead, an atypical hyperplastic cementum is formed. In comparison to roots with acellular cementum, these roots resorbed readily in response to orthodontic forces with no differences in the numbers of multinucleated cells present. 74, 75

Rygh provides plausible explanations for the resistance of cementum to resorption. First, there is less vascularity in the cementum when compared to bone.

Bone is constantly remodeled and thus the cells responsible for resorption are always present in the PDL. On the cementum side of the PDL this is not true, as cementum is constantly deposited throughout life. Also, cemental tissue is more mature than bone tissue and thus less susceptible to chemical changes. Furthermore, unmineralized cementum or cementoid appears to provide resistance to resorption by odontoclasts. 50

During the removal of the hyalinized zone, the cementum needs modification to re-establish connection with the PDL. This requires removal of cementoid and the more mature collagen, allowing a small breakthrough in the compromised area. The degree of breakdown products from the hyalinized tissue which may have importance as a signal 15 for the development of resorbing cells is unknown. If rapid progression of root resorption is noted, a rest period with removal of orthodontic forces is recommended.26, 50, 76, 77

Studies by Andreasen show evidence that the cells in the innermost layer of the

PDL provide a protective mechanism against root resorption. The PDL consists of cementoblasts, fibroblasts, osteoblasts, endothelial, and perivascular cells that supply the various protective mechanisms and the potential for repair of the tissues adjacent to it.

The vitality of the PDL seems to provide protection for cementum. 16, 17 This protection is also extended to the epithelial rests of Mallassez which are also present in the PDL in close proximity to the cementum and may be responsible for substances which prevent fusion of cementum to bone. 16, 17, 78 The mechanism becomes apparent when examining autotransplantation studies. The greatest likelihood of success in an autotransplantion is obtained when the PDL remains undisturbed. Those teeth are least likely to undergo resorption if the PDL is left unharmed.79

CLINICAL CORRELATIONS

Force magnitude

In theory, concentration of stress at the apex of the tooth would be closely related to resorption of roots. This would explain why intrusive mechanics are commonly blamed for severe incisor resorption. 34, 66, 80 However, other movements have also been shown to cause resorption. 3, 10, 26, 35-38, 81, 82 Generation of high forces has been noted to closely correlate with the appearance of resorption lacunae.24-26, 28, 41,

42, 51, 65, 83, 84 This is part of the reasoning for the use of optimal forces in tooth 16 movement. Many early studies investigating the relationship of force to tooth movement advocated the use of light and continuous force systems. Their findings established that heavier forces did not provide any advantage in the rate of tooth movement and would instead lead to a larger amount ofhyalinization, undermining resorption, and possibly a greater degree of root resorption.22, 28, 29, 51, 85-87

In review of a an excellent series of studies, force was found to increase the rate of tooth movement but not the degree of resorption. 88 In comparison to a continuous 50 gram force, a 100 gram force applied by use of the same mechanism was not found to have a significant difference in regards to tooth movement or EARR. 89 At 200 grams, a significantly increased amount of tooth movement was seen. However, the degree of

EARR still was not significantly different than what was seen with the 50 gm force. 90 A comparison of continuous and interrupted continuous forces showed that continuous forces were more effective in producing tooth movement but again, there was no difference in the amount of root resorption. 91 The experimental period in these studies was limited to 7 weeks. Perhaps a longer examination time or a larger differential in the amount of applied forces may have provided more insight into the effects of force application with respect to root resorption.

When examining the experimental teeth at various time points (1 to 7 weeks), increasing amounts of resorption was noted as time increased. Early signs were visible by one week and increased to 20 times more by seven weeks. Resorption half way to the pulp was found by the third week. The authors noted a great amount of individual 17 variation in EARR irrespective of the amount of tooth movement that occurred with the continuous 50 grn force. 92

In studying palatal expansion where high force systems are generated Vardimon et al found that compression of the PDL is excessive, thereby inhibiting "physioimmune system" and potentiating EARR. They concluded that EARR has three determinants: impulse, critical barrier, and environmental density. Impulse was found to be the predominant factor with two components, force and time. During short time periods, force magnitude is largely responsible for EARR severity in a logarithmic function. With longer treatment, the time during which force is applied becomes the deciding factor. 83

Treatment duration

There seems to be some relationship between clinical treatment length and root resorption.4, 5, 11, 12, 25, 29, 33, 35, 41, 42, 51, 76, 80 As early as two weeks after force application, resorption lacunae develop in almost half of orthodontically treated teeth.28, 29, 51 Levander and Malmgren found that if radiographic resorption is not detected after 9 months of treatment, the amount resorbed by the end of treatment will most likely not by excessive. 76, 77, 93 Because orthodontic appliances may be present for long periods without applying pressure to specific teeth, many researchers are hesitant to claim a cause and effect relationship when considering overall treatment time. 10, 34,

94 18

Age

The progression of age causes many changes in periodontal tissues. Alveolar bone increases in density, the PDL space narrows, and both undergo a decrease of vascularity and aplasia. Bone remodeling and osteoblastic activity are decreased in adults and thus the initiation of tooth movement is delayed in comparison to the adolescent patient. This is attributable to the increased time to reach the mobilization phase in response to hyalinization during tooth movement.25, 26, 65

In a study of alveolar bone tumover in rats, older rats were found to have higher numbers of osteoclasts (5 per mm) in comparison to younger rats (2.5 per mm). The rate of skeletal tumover, gauged by levels of alkaline phosphatase and acid phosphatase, was higher in the younger rats. The authors concluded that in an older system there is a higher recruitment of bone cells to compensate for a slower skeletal turnover.95

The layer of non-calcified cementiod coveting the root surfaces of younger teeth was found to be thicker on less mature roots and thus providing a protective barrier against resorption. 22, 24, 50 These factors, in part, explain the findings of studies which show correlation between early treatment and smaller degrees and decreased prevalence of root resorption.4, 5, 11, 21-26, 35-38, 60, 65, 80, 96

Orthodontic movement of teeth with incompletely formed apices has been associated with decreased root length upon completion of tooth formation and dilaceration.4, 5, 27, 29, 51, 96 Other researchers dispute the resistance of less mature roots to resorption. 9, 80 Harris and Baker found that prior to commencing orthodontic therapy, adults had shorter roots than adolescents. This corroborated the findings of 19 earlier research which noted root loss progression with age under normal conditions. 31 In application of identical treatment techniques to samples of adolescents and adults, no significant difference was found in the amount of further root resorption. The association between age and root loss from tooth movement was superficial and could be attributed to other dental factors (loss of dentition or periodontal problems).97 In another study,

Harris stated that "loss of stability from adjacent teeth, use of fewer remaining teeth, and loss of anchorage in bone are significant predictors of external apical root resorption".20

Systemicfactors and diet

Immunologic factors, systemic disease, and diet have been related to the resorption of roots. The role of the immune system in root resorption has been investigated, however, conclusive evidence has not yet been found. 98 Early investigators believed systemic disease such as endocrine imbalances and metabolic diseases of bone to be the major etiologic factor associated with root resorption. 6, 99

More recent data indicates an association of abnormal hormone levels from endocrine diseases of the pituitary, thyroid, and parathyroid with root resorption but only as a secondary causative factor. 5, 25, 100 Even a secondary relationship between hormone level and root resorption is disqualified by some research. 99, 101 Deficiency of dietary calcium and vitamin D were originally thought to cause the resorption of roots. 102

Recently it has been concluded that nutrient imbalances may also be only a secondary influencing factor. 35, 36 In a controlled study, lactating rats with a calcium deficiency were noted to have an expected increase in parathyroid hormone secretion and thus a 20 decrease in bone density. Tooth movement in these rats was greatly facilitated and low levels of root resorption was noted. 103

Alveolar bone density

Movement through denser bone was found to correlate with higher amounts of root resorption. Lamellar bone or cortical bone is difficult to resorb, in addition increased amounts of marrow space in less dense bone allows easier recruitment of resorptive cells for bony remodeling.26, 33, 103-106 Kaley and Phillips found that the chance for root resorption is 20 times greater when the maxillary incisors are in close proximity to the palatal cortex. 107 Other researchers have found no relation between root resorption and proximity to the palatal cortex. 37, 38, 108 In a study of the effects of rapid palatal expansion on tooth roots, Barber and Sims noted that although severe resorption occurred over the entire buccal root surface, over 35% of the root surface, even heavier resorption occurred in the cervical third. The authors speculated that the close apposition of this part of the root against the buccal cortical plate could be to blame. 109

Root canal treatment, trauma, andpernicious habits

Although a few early studies indicate otherwise, 110 current belief indicates that teeth treated endodontically undergo decreased amounts of resorption, presumably due to the increased density of dentin.37, 38, 111 A common reaction to traumatic injury is external root resorption and recommended treatment of these teeth includes calcium hydroxide treatment followed by instrumentation and obturation at a later time. Internal 21 resorption is treated the same way, in hopes to increase dentin hardness and thus increase resistance to resorption.

Roots of teeth which experience resorption in response to trauma are at increased risk for further resorption when moved orthodontically. 31, 32, 106 However, no increased resorption is seen in orthodontically teeth which remain unaffected by trauma.

A waiting period of 4-5 months is recommendable for those teeth which have experienced trauma, during which no signs of inflammatory resorption become present.

112 Occlusal trauma from restorations, prosthetics, and even on natural teeth with interferences and prematurities produce jiggling forces which will elicit thinning of the cemental layer over roots with possible progression to root resorption. 113 Nail-biting, mouth breathing, tongue thrust, clenching have all been positively associated with increased root resorption. 33, 35, 69, 113, 114

Vulnerability ofspecific teeth

Although all teeth have been shown to display some degree of root resorption, some teeth experience greater amounts of shortening than others. Many studies relate this variability to difference in sensitivity.2, 3, 9, 10, 31, 80, 81, 115 The amount of root movement is thought to be a great contributing factor. 7, 10, 34, 82, 94, 116 In order of decreasing severity, the most frequently affected teeth are maxillary incisors, mandibular incisors, distal root of the mandibular first molars, and mandibular and maxillary second premolars.10, 12, 26, 30, 94, 116 Also, root shape has been determined to be an important factor in susceptibility. Roots which are dilacerated, blunt, or pipette-shaped, 22 undergo significantly more resorption than their normally shaped counterparts. 13, 34-38,

76, 80, 93 Teeth with naturally short roots those which are not resorbed were previously thought to be more apt to experience resorption than longer teeth. In an evaluation of force systems required for tooth movement, it becomes clear that longer teeth need greater force values to be moved and the actual displacement of the root apex is larger during tipping or torquing movemems. Newer studies corroborate this logic. 37,

38, 42, 76, 93

Mechanics

A great deal of controversy exists regarding the method and mechanics of tooth movement in regards to root resorption. Studies involving correlations between root resorption to orthodontic mechanics are phenomenological in nature. They are statistical and do not describe the cause. The only consistent finding from this vast literature is that applications of high moments delivered over long periods of time seem to increase the chance of clinical root resorption. As stated before, maxillary incisors are the most commonly affected tooth with regard to root resorption. This may be due to an inherent sensitivity of these teeth to resorption as a.result of more extensive treatment, i.e. intrusion, extent of movement, and movements which concentrate forces at the apex.21-

24 But, even the correlation between the extent of root movement and intrusion to root resorption is debated by some. 10, 34 Gottlieb did not observe root shortening or apical blunting during incisor intrusion up to 4 mm (mean of 2.36 mm) using 15 to 20 grams of force per incisor. 117 Using 25 grams of force per incisor, Dermaut and DeMunck 23 obtained 3.6 mm of intrusion. They observed 2.5 mm of root resorption but found no correlation between the amount of intrusion or duration of applied imrusive force with increased amounts of root resorption. 34 In addition, jiggling forces created from repeated trauma by functioning on inclined planes or by using intermaxillary elastics and removable appliances, have been shown to contribute to root resorption. 5, 7, 11, 35, 36,

41 Early research indicated that jiggling forces caused by removable appliances may be harmful to the roots. 7 However, by use of fixed appliances, roots cannot function individually which is also hypothesized to be harmful.2, 35, 36

Comparisons of Begg versus edgewise techniques have been conflicting as to which is less likely to illicit root resorption since both techniques have been reported to be culpable.41, 104-106, 112 Although more tipping movements are expected by use of

Begg mechanics, an accurate comparison between these two techniques is impossible since throughout treatment so many different variables become apparent no matter which technique is employed. Orthodontic treatment with or without extractions has not been found to differ with regard to the potential for root resorption. 80, 82

Retention andpost-retention

If severe root resorption is seen during treatment, either modification or termination of treatment should be considered. Upon termination of active treatment, cessation of the resorption process should occur. 10, 18, 19, 24, 82, 105, 118-120 At this point repair begins, which in itself may cause a minor reduction in root length as the rough edges at the root apex are remodeled.26, 111, 121 This process may take 24 approximately 6 weeks after active forces are discontinued 26, however, instances where active resorption continues for a prolonged time have been reported. 122 Traumatic occlusion would be highly suspect. 11 Retainers with active forces should also be examined when continued resorption is seen. Movement as a result of orthodontic relapse has been de-valued as a contributor to resorption26, however, with no other contributing factors, even light muscle forces should draw some attention. 105

As mentioned above, if radiographic resorption is not detected by 9 months of treatment, the amount or EARR at the end of treatment will most likely not be excessive. 76 However, if rapid progression to EARR is noted in mid-treatment, a pause with removal or orthodontic forces is recommended. 26, 76, 93, 111 To demonstrate this, a group of investigators compared incisors where EARR was noted after 6 months of treatment. The control group continued the original treatment plan whereas in an other group treatment was paused for 2-3 months. The experimental group had significantly less resorption at the end of treatment than those in which the original treatment plan was continued without a pause. 77

DENTAL HEALTH AFTER ROOT RESORPTION

It is well accepted that some root resorption occurs in every orthodontically treated tooth. Most resorption sites are repaired to restore the original anatomy of the root. Although defects are filled in by cementum, the initial reattachment of Sharpey's fibers is not as extensive as original which may indicate a less effective attachment apparatus. 109 In the cases with apical root shortening, root loss is rarely greater than 2 25 mm. Loss of 2 mm of apical root length would cause loss of total attachment area of the root by 5-10 percent. 10 Zachrisson concluded that 2 mm of root shortening is not detrimental to the longevity and of the tooth. 123 More severe root shortening, 4 to 6 mm, is equivalent to 20 to 35% decrease of attachment. 10 Marginal attachment loss, specifically, alveolar bone height is considered more significant to support and function that loss of root length. 10, 29, 32, 94, 115, 116, 123-125

CYTOKINES AND HARD-TISSUE METABOLISM

The term cytokine can include interleukins, tumor necrosis factors, interferons, colony stimulating factors, growth factors, neuropeptides, monokines, and lymphokines.

Cytokines are a group of generally glycosylated peptide regulatory factors, produced transiently by most nucleated cells. Even at low molecular weights (<80kDa), they are very potent, acting in the 10-13 to 10-10 M range. They interact in a complex cell-tissue communication network modulating growth and differentiation. 126 They have pleiotropic overlapping biological activities, which are mainly autocrine or paracrine. 127

High affinity receptors, constitutively or inducible, are expressed on target cell surfaces usually in low numbers and may be cytokine-specific or shared by distinct cytokines.

The chemical affinity of receptors for a wide range of cytokines has been measured to be generally around 10-50 pM. 128 This corroborates studies that measure cytokine concentrations required for 50% of maximal growth rate for cell cultures which are also in the 10-50 pM range. 129, 130 26

The interaction between cytokines and cells is dependent on a number of factors.

This mode of communication is governed by cell secretion rates, transport processes, and ultimately by genetic and biochemical processes. Secretion rates may be also be modulated by many different means. The maximum secretion rate is an important factor and is estimated by evaluation of molecules with specific half-lives and functions to perform. In an analysis of mRNA and RNA polymerase, the maximum secretion rate of a cell was estimated to be 2300-8000 molecules per second. Based on the above variables, the effective intercellular communication distance is mathematically estimated to be 250 um or approximately 50 cell radii. This extrapolates to a 10-30 minute time constant for intercellular communication by cytokines. 131

Originally a single molecule was thought to mediate bone resorption. Osteoclast activating factor (OAF) was named as the factor released by cultured peripheral blood mononuclear cells which had been activated with dental plaque antigens and stimulated bone resorption in organ cultures. 132 Later it was found that this effect was mediated by a number of molecules from various cell types. More recently, the number of cytokines which are involved in bone metabolism has grown extensively. Table 4 and figure 1 illustrate some of the factors which play significant roles in bone metabolism.

lnterleukin-1

Interleukin -1 (IL-1) was the first immune cytokine positively identified as having an action in the metabolism of bone. The name interleukin was conferred due to 27 regulatory functions on the immune system. Since then, many other functions have been attributed to this family of proteins.

Two related proteins have been identified, IL-1 ot and IL-1 , with almost identical actions. The two proteins come from distinct genes on chromosome 2. Both ha,e 3 l kDa precursors and share 26% amino acid homology. The mature IL-1 ot molecule contains

159 amino acids and weighs 17.5kDa with pI 5. IL-1 weighs in at 17.3 kDa, with 153 amino acids and pI 7. In most cell types IL-11 is synthesized and secreted in a greater amount over IL-lc. 133 IL-I is more active in bone metabolism. It may stimulate bone resorption while inhibiting bone formation. 134 Although the amino acid homology between the two forms is limited, both interact with the same plasma membrane receptors. The affinity of the receptor for IL-1 does not correlate with the magnitude of the biological response and the post-receptor events are unclear. Cyclic AMP (cAMP) accumulation, protein phosphorylation, and other events have been measured following stimulation of cells. 135 A third member of the IL-1 family has been discovered, and although containing 26% homology with the other two members, it performs antagonistic functions, thus named IL-1 receptor antagonist protein (IRAP or IL-lra). 136

As noted earlier, IL-1 has pronounced effects on the immune system. Its effects on T and B cells mainly include induction of growth and differentiation other factors such as IL-2, IL-4, IL-6 and interferon gamma. Interleukins may activate natural killer (NK) cells when working in concert with other factors. Chemotactic effects have also been found in these family of proteins. Apart from immune functions, IL-I has been found to have major roles in the synthesis of acute phase proteins in the liver, in the coagulation 28 process in the endothelium, induction of fever at the level of the hypothalamus by prostaglandin E-2 (PGE2) production, involvement in hemopoiesis, and various metabolic effects. Powerful metabolic effects on cartilage, bone, and fibrous connective tissue are well documented. 137, 138

IL-1 induced stimulation of bone resorption is seen as an increase of multinucleated cells in marrow spaces. This resorption can be inhibited by calcitonin, but is unaffected by indomethacin, suggesting that the resorption is osteoclast mediated but that prostaglandin production is not required. 139 Other studies have found various results regarding indomethacin inhibition in bone resorption. The role of prostaglandins in bone resorption is well accepted, however, a prostaglandin-independent mechanism of

IL-1 induced resorption is also possible. In fact IL-1 is the most potent resorbing agent of bone, active at concentrations of 3 X 10. 139

Tumor Necrosis Factor

Tumor necrosis factor (TNF) also exists in an o and [ forms. Both proteins are coded for on chromosome 6 and share a 30% amino acid homology. TNFo is cleaved to form a 17 kDa mature protein which, in the active form, exists as a trimer. 140, 141 TNF has been found to interact with two receptors. Again, the post-receptor activities are unclear.142 A major source of TNF is believed to be the monocyte and macrophage, although it may be produced by a variety of cells. TNF production may be stimulated by lipopolysaccharides, interferons, IL-1, and many other cytokines. 29

The pleiotropic actions of TNF are similar and sometimes overlapping with IL-1

(table 5). Both cytokines are produced by the same cell types and have synergistic effects. TNF is involved in most of the pathways listed above for IL-1. In addition, much attention has been attributed to TNF because of its cytotoxic effects on certain tumor lines. 139-141, 143

TNF stimulates bone resorption in a way similar to IL' 1. At very low concentrations IL-113 and TNFc are implicated in bone remodeling through specific receptors on bone cell populations.45, 138, 144 In vitro studies show increased numbers of osteoclasts in marrow cultures in response to TNF. 144 Both IL-1 and TNF have been shown to induce osteoclastic activity in vivo. 145 In addition to osteoclast recruitment,

IL-1 and TNF have been found to inhibit osteoblast maturation, collagen synthesis and osteocalcin production.

Involvement in dental pathology

Both IL-1 and TNF are implicated in the destruction of supporting structures in periodontal disease. 146 This is in part due to the ability of these molecules to directly induce secretion of collagenase by fibroblasts, and their facilitation of bone and cartilage resorption. 144 Increased levels of both molecules are present in the periodontal tissues and/or gingival crevicular fluid of sites exhibiting periodontal disease. 147-152

Following treatment of diseased sites, levels of these, and other cytokines levels decrease. 153 30

Infection of the dental pulp may result in the formation of a periapical lesion and bone resorption. The impetus for this process is thought to be the presence of immune complexes, lipopolysaccharides, or other bacterial components any of which may trigger the production of inflammatory.cytokines. 154, 155 Macrophages, fibroblasts, neutrophils, and osteoclasts in the area of the infected pulp are stimulated to synthesize and secrete cytokines, specifically, IL-1 a and TNF tx. 156

Generally, TNF is also a potent inducer of bone resorption, with the ct form being more potent than the 13 form. Although, neither are as potent as IL-1. An increase in IL-1 may be a later and more significant potentiator of bone resorption. However, an increase in TNF may be incipient and portend an IL-1 increase. Furthermore, IL-1 and TNF stimulate endogenous PG production (and a slew of other cytokines) which would abrogate the metabolic process. 157

IL-1 and TNF are synergistic, both stimulating PGE-2 production and bone resorption (table 5). Many important synergistic effects are shared by these two molecules. This includes regulation of phospholipid metabolism, induction of acute phase proteins, prostacyclin synthesis, promotion of growth in fibroblasts, expression of human major histocompatibility complex antigens, induction of fever, and synthesis of hematopoietic growth factors, activation of inflammatory leukocytes, modification of vascular permeability, and induction of bone resorption. 143, 158-160

Previous studies using in vitro models illustrate the interactive effects of cytokine combinations on PGE production by fibroblasts. Using murine calvarial organ cultures, a ten-fold increase in stimulation of PGE-2 production is seen when stimulated by IL-1 and 31

TNF in concert. This is in comparison to stimulation by 1 uL ofIL-1 or 50 ng/mL of

TNF which produced 5-7 mg/mL of PGE-2 when used individually. 161 Unfortunately, quantitative evaluation of PGE production in response to cytokines in the in vivo model remains a difficult task, but these data would help describe feedback mechanisms in other cell types. 162

In summation, IL-1 and TNF are one of the earliest cytokines to be associated with bone metabolism. The vast amount of research indicates that these molecules induce bone resorption by affecting proliferation and migration of osteoclasts, by inducing the release of other factors to aid in the process, and by influencing the differentiation of osteoblasts. 138 The previously mentioned effects in addition to the synergistic effects that these molecules have with other proteins/cytokines (PG's) which also may induce bone resorption make them invaluable markers for bone metabolism.

TOOTH ERUPTION MODEL

In an examination of the cell biology of root resorption, the resorption of deciduous roots during tooth eruption may serve as a model. Marks and Cahill noted that the demal follicle is necessary for the eruption process. 163 The follicle itself, consists of a connective tissue sac which envelops the tooth before its eruption. If the follicle is removed, eruption does not occur. Conversely, if the contents of the follicle are substituted with inert material, leaving the follicle itself, eruption still occurs. 164 During the process of eruption, monocytes and macrophages influx into the follicle and their numbers generally peak at the time of active eruption. Osteoclast numbers also increase, 32 collecting around the coronal area of eruption crypt. As the monocyte population decreases so does the osteoclast numbers. 165 Therefore, to allow tooth eruption through bone, monocyte and macrophage influx is necessary for osteoclast formation. 166

Cells participating in tooth eruption and root resorption need to have a method of communicating with each other. They do this by the creation of biochemical molecules, chemical signals that target specific cell surface receptors. In their investigation of the molecular signaling pathway for tooth eruption, Wise and Lin found four possible candidates.166 The first and most likely trigger for tooth eruption is colony stimulating factor 1 (CSF-1), also called macrophage colony stimulating factor. Kodama found that

CSF-1 played an essential role in osteoclast differentiation. 167, 168 In a study using osteopetrotic rats, where a reduction of the number of osteoclasts prevented tooth eruption, injection of CSF-1 triggered an osteoclast increase resulting in tooth eruption.167, 169 Normal rats injected with CSF-1 experienced accelerated tooth eruption and increase of osteoclast and monocyte numbers. 170 The CSF-1 gene is expressed in the rat dental follicle and once translated, has an autocrine effect on the

CSF- 1 gene. 171

The presence of interleukin-1 alpha (IL-1 ct) has been noted in the stellate reticulum. Because of its ability to enhance CSF-1 gene transcription, it is hypothesized that a diffusion ofIL-1 ct into the dental follicle may be the trigger for CSF-1 transcription. 166 Epidermal growth factor (EGF) is suspected to be the regulator of IL-1 ct expression. Following EGF injections in rats, levels oflL-let in the stellate reticulum increase. 172 Furthermore, these injections also enhance the expression of IL-la 33 receptors in the dental follicle. Although EGF does not stimulate CSF-1 synthesis directly, its effect on IL-1 a transcription and IL-1 ct receptors provides an indirect mechanism for tooth eruption. Transforming growth factor beta-1 (TGF-1) is also considered as an initiator of tooth eruption. This is due to its ability to enhance IL-1 c translation and its role in chemoattraction of monocytes. It has also been immunolocalized in the stellate reticulum. 173 Its transcription has found to be enhanced incubation with EGF. 174 The signal for tooth eruption could begin with TNFc, EGF, or

TGF-131, any or all of which could enhance IL-1 c synthesis. It should be noted that

TNFet is a proximal molecule in the cytokine cascade leading the expression of the interleukins. TNFct is a product of macrophages which, as was mentioned, are an essential part of the normal dental follicular eruption and root resorption of deciduous teeth. Therefore, osteoblastic activation could occur from the effects of CSF-1, IL-1 , or both. 166, 172

CYTOKINE INVOLVEMENT IN ORTHODONTICS

The importance of Prostaglandin E (PGE) in bone and cementum remodeling has been described by Raisz and others. 175, 176 Prostaglandins, in general, are potent modulators of physiological and pathological reactions. In animal studies, local administration of PGE increases the rate of tooth movement. 177, 178 Conversely, administration of prostaglandin inhibitors may slow tooth movement presumably by inhibiting cyclooxygenase activity and thus PG synthesis. 179, 180 In a study of rats injected with PGE2, tooth movement was significantly enhanced, however, findings 34 regarding its role in root resorption were inconclusive. 181 A variety of cells have been observed to increase synthesis and secretion of PGE in response to mechanical or chemical stimuli. 134, 182-184 Under these conditions gingival fibroblasts have been observed to synthesize PGE and cAMP. Increased PGE production is also seen when IL-

1 [ is administered to stressed PDL fibroblasts. 134 The link between cytokines mediating bone remodeling during tooth movement has been frequently demonstrated.

Orthodontic appliances apply forces to teeth and in turn, to the periodontal ligament. These mechanical events illicit a biochemical process which allows bone remodeling for tooth movement. This process may be regulated by acid phosphatase, hormones 185, neurotransmitters 44, and cytokines (IL-1 c and [, IL-2, tumor necrosis factor c, and interferon ,). 43-45, 134, 162, 183-191 In examination of human gingival crevicular fluid, increased levels ofIL-l, IL-6, TNF-ct, and EGF were seen 24 hours after force application. These cytokine levels plateaued suggesting down regulation after initial stress. 191 Grieve also found an increase in cytokine levels in the gingival crevicular fluid of orthodontically moved teeth. At one and twenty four hours after initial activation, IL-1 levels rose to 8.9 pg and 19.2 pg respectively. The IL-1 levels at the control teeth measured 2.0 and 2.9 pg. PGE rose to 108.9 and 97.9 pg at 24 and 48 hours compared to control levels of 61.8 and 70.8 pg.192 The lagging peak of PGE compared to IL-1 indicates a stimulatory effect by IL-1 which has been also noted in other studies. 162, 183, 184, 186, 187 The drop in cytokine levels with time can be attributed to lack continuous force systems used in these studies. 35

It is speculated that these cytokines originated from cells in the PDL in their efforts as mediators of early tooth movement. In order to trace the origins of cytokines involved in tooth movement, distalized molars from cats were immunohistochemically examined at different time intervals. Levels of calcitonin gene-related peptide, IL-1 et,

IL-113, and TNF-x increased significantly in the alveolar bone marrow cavity cells near the sites of orthodontically stressed PDL, particularly where compressed. After one week, the channels leading from the PDL into neighboring marrow cavities were populated by cells with high levels of these cytokines. After four weeks, cessation of bone remodeling was seen, and the marrow cells resumed baseline levels of cytokine formation. This would indicate that marrow cells respond to the orthodontic forces expressed onto the PDL by way of cytokine production which enables participation in bone remodeling and thus tooth movement.45 In summation, early sources for these cytokines may be fibroblasts, macrophages, osteoclasts, osteoblasts, cementoblasts, and cementoclasts. All of these cells stain for cytokines at 12 to 24 hours, however osteoclasts and adjacent mononucleated cells are the only active cells 14 days later in the process.45 Again, the lack of continuous forces may explain the drop in the number of mediators.

In evaluation of stimulated blood monocytes taken from individuals with severe

EARR and control subjects, no difference was found in the levels ofIL-1 [ and TNFc. A wide range of stimulated and unstimulated IL-113 and TNFct production was seen. It was concluded that no intrinsic difference exists in peripheral monocyte production of these molecules between patients whom exhibit severe EARR and those who do not. 193 36

In another study, root resorption was inhibited in rats by topical application of a bisphosphonate. 194, 195 The use ofbisphosphonate will suppress fibroblastic activity and cause a decrease in periodontal ligament width. 196 Finally, a third study established a relationship between adhesion molecules and aggressive osteoclastic activity on root surfaces in a feline model. 197

The cytokine levels may be the same for bone resorption or root resorption in the

PDL and ultimately in the crevicular fluid. There appears to be no specific marker for resorbing cementum as opposed to bone. It is likely that the array of markers would change or present a different profile for normal tooth movement as compared with concomitant extensive root resorption. This would be seen in the PDL and could be seen in a crevicular fluid assay.

GINGIVAL CREVICULAR FLUID

A thin sulcular epithelium allows passage of fluid from gingival connective tissue into the gingival crevice. This fluid aids in cleansing material from the sulcus, contains plasma proteins which promote epithelial attachment to the tooth, and contains antibodies for antimicrobial protection. Gingival crevicular fluid contains various elements, mostly derived from surrounding periodontal tissues, which may not always correlate to levels found in blood serum (table 7).

In 1958, Brill and Krasse demonstrated that chemicals injected into the bloodstream can be collected from the gingival sulcus in three minutes.201 They concluded that there is a passage of material from the blood vessels into the connective 37 tissue and through the epithelium into the sulcus. They noted that in health, gingival crevicular fluid is scarce in volume and presents as a transudate thereby being difficult to collect.202 In the Brill study fluorescein was used, however, a number of molecules have been shown to be able penetrate the sulcus. Those substances include large or charged molecules such as albumin203 endotoxin204 and histamine205 Passage of these substances indicates permeability to substances of molecular weight up to one million.206

Later researchers reclassified gingival crevicular fluid (GCF) as an inflammatory exudate. They claimed that increased amounts of GCF from the Brill studies 203,207 could have been elicited by an increase in capillary permeability due to damage from intrasulcular insertion of paper strips for the collection of GCF.208, 209 Different methods have been explored in quantifying the amount of GCF and its coments at various normal to pathological states. Paper strip assays of slightly inflamed gingiva measured

0.1 mg of GCF in 3 minutes.210 Isotope dilution method used in the interproximal molar areas with a mean gingival index less than 1 measured 0.43 to 1.56 ml.211 As stated before, excess amounts of GCF is present as an exudate. Presence in a sulcus which appears normal could be explained by the fact that even a clinically healthy sulcus will exhibit inflammation when examined microscopically. GCF measurements in individuals with periodontitis measured at 0.52 ml at diseased sites and 0.28 ml at non- diseased sites in the same individual. These measurements were taken with the paper strip method over 30 seconds. 153 Using the same method over 20 seconds, GCF measurements were made on healthy, gingivitis, and periodontitis patients. The volumes 38 were 0.255, 0.540, and 0.370 ml respectively.212 Increased fluid production could also be due to mastication of coarse foods, toothbrushing, gingival massage21 O, ovulation213

contraceptives 214, and smoking215. RATIONALE

A NON-INVASIVE TECHNIQUE TO SAMPLE THE PDL IN HUMANS

The orthodontic literature is replete with clinical studies dealing with statistical correlates of root resorption and a host of independent variables. There is little understanding of the cell biology of root resorption, particularly in the human. Without extraction and histologic evaluation, the only method of diagnosis for root resorption is the radiograph.

Unfortunately, significant modification of root shape is required to be detectable by x-ray. 198, 199 Early detection would require periodic radiographic screening with periapical radiographs of the incisors taken at least every year. 12, 26, 29, 35, 94, 119 A full mouth periapical series is required to ascertain the pattern and severity of resorption.

Resorption of the facial or lingual surfaces may still escape radiographic detection.

Therefore, development of a non-invasive method for evaluation should be very useful.

This would provide a more detailed accounting of the balance between bone and root resorption, and because of its benign nature, would allow for early detection of more pernicious resorption. In addition, it would allow for more frequent monitoring without untoward side effects.

Because the site of the root resorption is at the apex or along the surface of the dental roots, monitoring for changes in biochemicals would require either tissue samples or development of an indirect technique. The co-investigators have developed an indirect

39 40 technique. The details of our previous findings with this technique have been described.200 The site for obtaining the samples is the gingival sulcus.

As the sulcus is relatively remote from sites where resorption is active, the question becomes how to assay cytokines and growth factors formed at these sites.

Absorption assays may be performed on filter paper strips 190-192, or by the more efficient immunomagnetic technique.200 This project will allow us to directly study cell biological changes in an average human population undergoing orthodontic treatment.

We have selected a molecular signal that is known to play a significant role in bone metabolism (table 6).

PREVIOUS DATA

Assuming that the clast cells are involved in resorption are in sufficient number to initiate the process, production of TNF, other cytokines and phosphatase should be an early event in this resorptive process.216-220 Therefore, for example, one should observe an increase in TNF during active root resorption beyond that seen in a control

Previous studies using IL-1 as a representative marker molecule have shown IL-1 is present in the normal healthy sulcus. Upon application of an orthodontic force, the level of IL- 1 and other molecules increased significantly. 183, 184, 186, 192 TNF is present in the normal healthy sulcus in a range of 17.1-482.8 pg with a mean of value of

159.4 SD+/- 25.4.216, 217 Orthodontic tooth movement has been shown to increase these levels up to two fold.200 41

Another study employed the immunobead technique to follow TNF levels in deciduous teeth on a group of patients in a private dental practice. Seventeen deciduous tooth sites were sampled, thirteen with normal exfoliation (root resorption) and four believed to be ankylosed by clinical criteria. The TNF values found in the patients with normal exfoliation ranged from 29-149pg/site with a mean of 66pg/site. The positive values in the teeth considered ankylosed ranged from 12-82pg/site with a mean concentration of 38pg.

A larger study is now being conducted using the same criteria. After comparison of the data collected so far, a difference is already obvious. Statistical analysis has not yet been completed, but the difference between experimental and control samples show an encouraging trend that differences can be determined using this immunobead technique.

SPECIFIC AIMS

Present models suggest that the root resorption process should be preceded or accompanied by an increase in the level of biochemical modulators. Based on these models, we hypothesize that if the resorptive process is increased, there should be a measurablle change in the level of these biochemical modulators. The following are our specific research aims for this project:

1) To measure TNFc levels present in gingival crevicular fluid using the immunobead

technique. 42

2) To validate the sampling technique using clinical situations where a difference in TNF

levels are known to exist.

3) To compare crevicular fluid levels of TNFa in various clinical settings among

orthodontic patients.

4) To compare differences in TNFc levels in different stages of orthodontic treatment.

5) To determine if the TNFa levels of in the crevicular fluid correlate with the degree of

root resorption.

6) To determine if this process can be useful in diagnosis of root resorption. RESEARCH METHODS

STUDY POPULATION

Subjects will be asked to participate from the University of Connecticut School of

Dental Medicine. Race, ethnicity, or gender are not exclusionary criteria. We will attempt to include a minimum of 20 human subjects in the permanent or late mixed dentition.

Table 8 comains the full listing for exclusion and inclusion criteria. Exclusion criteria will preclude patients with a history of chronic inflammation, chronically taking anti-inflammatory medication (i.e. aspirin), smokers, and patients with an elevated caries index because subclinical periodontitis may be more likely and mask the specificity of our results. 190, 191,200

Subjects will be identified as having undergone some degree of root resorption during orthodontic treatment. The extent of resorption will be evaluated by use of a periapical radiograph and scaled as described below (clinical determination of root resorption). The involved teeth, their adjacent, and contralateral counterparts will all be evaluated.

DATA COLLECTION

A host of variables will be examined for each patient at each visit. In addition to nominal and continuous data, some variables will be tabulated using ordinal indices. The

43 44 criteria used for categorization is described below. Information will be collected at each visit as shown in table 9.

Although not all collected data will be relevant to the currem investigation, some of the variables have been found to statistically correlate with root resorption and will be recorded for furore examination. Demographic data will include patient name and date of collection, patient number, age 4, 5, 11, 21-26, 35, 36, 60, 65, 80, 96, 97, sex 8, 13, 31,

41, contributing medical history 5, 6, 25, 35, 36, 96, 102, 103, familial history of EARR

39, and time since last tooth brushing. The latter will be recorded as recent tooth brushing may alter the composition of gingival crevicular fluid.

Information about each sampled tooth will also be collected. Description of root morphology 13, 34, 37, 38, 76, 80, level of EARR 76, history of trauma 35, 36, 112, previous root canal treatment 37, 38, 110, 111, presence of pathology, restorations, periodontal index221, and sulcus depth.

Information about orthodontic treatment will also be collected. This will include the type of movement 3, 10, 26, 34-38, 66, 80-82, force applied 24-26, 28, 29, 41, 42, 65,

83, force remaining at the time of assay, time since force application 83, whether the force was continuous or interrupted 22, 28, 29, 85-87, occlusal trauma 113, use of intermaxillary elastics 5, 7, 11, 35, 36, 41, use of rapid maxillary expander and time since

83, presence of an openbite, digital or other pernicious habits 35, 42, 70, 84, 113, 114,

222, 223, treatment time so far 4, 5, 11, 12, 25, 29, 33, 35, 41, 42, 51, 76, 80, 116, and 45 expected treatment time remaining. Also, the measured TNF level for each tooth will be recorded following the appropriate laboratory assays.

CLINICAL DETERMINATION AND SCORING

Radiographic assessment of root morphology will be made from periapical radiographs. Quantification of root loss will be performed by use of three methods.

Method #1 involves the direct measurement, to the nearest tenth of a millimeter, of entire length of the tooth as viewed from pre and post treatment periapical x-rays. The difference in tooth length from pre to post treatment is defined as root loss (RL). In method #2 we will attempt to correct for magnification error between x-rays by multiplying the ratio of crown heights against the post treatment tooth length. Corrected root loss (RL*) will be calculated as follows (figure 2): Method #2: RL* T T (C/Cz) where T is the measurement of the pre-treatment tooth length, T2 is the post treatment tooth length, and C/Cz is a ratio of pre-treatment to post treatment crown lengths

(correction factor). A third method will also be used for the assessment of root resorption. 35 Method #3 employs a subjective measurement, the root resorption index

(RRI) based on an ordinal scale of 0-4. 76, 107, 112, 116 A score of 0 indicates the lack of any visible root modification, whereas scores of 1-4 indicate increasing amounts of root resorption as determined from periapical x-rays (table 10). Although not as precise, this scale allows for a simple classification system of severity. The subjective assessment 46 will attempt to categorize the degree of root modification for each tooth in an ordinal scale. The first seventy measurements and all root resorption index determinations will be repeated by the same investigator at a different time. The new values will be compared against the original recordings to determine measurement error.

The Periodontal Index (PI) estimates the severity, pocket formation, and masticatory function of each tooth. 221 The scale has a true biologic gradient but lacks objective specificity about the degree of inflammation. The PI is the most common classification system used with epidemiological studies and is currently used in the

National Health Survey. X-rays or probes are non used so the actual periodontal problem may be underestimated (table 11).

The Gingival index was developed solely for the purpose of assessing the severity of gingivitis and its location in four areas: the distofacial papilla, facial margin, mesiofacial papilla, and entire lingual gingival margin (table 12). A periodontal probe is used to assess bleeding potential. It is a subjective assessment of inflammation, and requires manipulation to assess bleeding.224

The Plaque Index is a subjective assessment of plaque accumulation on the dentition. Although in its inception, it is intended for use on a per patient basis. A specific tooth in each quadrant of the mouth is to be evaluated by running probe across tooth surface. A total score is then assigned for the patient in order to evaluate overall oral hygiene.225, 226 In this study, a plaque score will be recorded for each tooth (table

13). 47

SAMPLING FROM THE GINGIVAL SULCUS

Paramagnetic beads (M-450 coated with sheep anti-mouse antibodies) will be obtained from Dynal, Inc., Lake Success, NY (figure 3). The beads are manufactured so as to be spherical shape and relatively uniform in size, each being 5 um in diameter.

Anti-TNFot antibodies will be attached to the beads and to be used for capture. The procedure for attachment of antibodies to the beads has been described.218,227

The immunomagnetic sampling procedure has been previously described.200 The sulcus of the tooth to be tested will be gently irrigated with water at 370 C for thirty seconds and air dried with a dental syringe. To determine the levels of the markers, a stock solution will be prepared containing lx106 of beads coated with antibodies for the markers. The immunobeads are introduced into the sulcus as a slurry prepared with a known number of beads. Approximately 2 mi of the slurry is deposited into the sulcus by use of a pipette with a plastic tip with care so as not to damage or irritate the gingiva.

The flat tip of the pipette should be placed just at the free gingival margin and the slurry dispensed slowly. The slurry has a brown color allowing the dispensing to be easily monitored by the appearance of the brown color along the neck of the tooth. This procedure will sample the entire gingival sulcus since the beads quickly distribute themselves in all four areas of the crevice (mesial, distal, buccal and lingual).

The slurry will remain in the sulcus for 1 minute, the mean time it takes for the gingival sulcus to fill with crevicular fluid.200 A special instrumem has been devised for removal of the beads from the sulcus. This harvester has a sterilizable spatula tip and a shank which houses four disk shaped permanent magnets. The magnets generate a field 48 of 185-188 gauss at the tip. The immunobeads will be recovered with a gentle sweeping motion of the tip at the gingival margin.228

QUANTIFICATION OF BEAD-BOUND ANTIBODIES

After harvesting, the beads appear as a brown mass on the tip. The tip is inserted into a 1.5 ml Eppendorf microcentrifuge tube containing PBS/BSA buffer and gently vortexed to suspend the beads. The beads are easily separated from the solution by placing a magnet against the side of the tube.200, 216

The immunobeads are cleaned of detritus by repeated resuspension in fresh solution and separation using a permanent magnet. At this point, a purified concentration of beads is accumulated with their attached monoclonal antibodies bound to their respective molecules.

Recovery may be reduced by loss of beads in the sulcus, during the purification process, and in the laboratory phase (assay). More than 99% of the beads are routinely recovered from the sulcus.216 Remaining beads will be either dislodged and lost during post harvesting irrigation or washed out GCF. A few remaining beads may be phagocytosed by macrophages (Kupfer cells).

Recovered beads are counted using a haemocytometer and analyzed for antigen by an ELISA (figure 4). The general procedure for the ELISA assay uses the reagents provided in commercially available kits (Predicta TNF alpha kit, Genzyme Corporation, 49

Cambridge). As part of our validation process, we have compared the results obtained with TNF in a plate ELISA to those of the bead ELISA. The results of the comparison of the two methods show close similarity both in quantity and quality.

The amount of the antigen will be determined by calibration of the recorded absorbance to a standard curve made for each assay. The lower limit sensitivity is determined by_adding two standard deviations of all zero standards to the mean absorbance of all zero standards. This value is used to separate positive measurements from zero in each assay. If the sample absorbance is greater than this value, the measurement is considered positive. If the measurement is less than or equal to this value, it is considered to be zero. The level of antigen is expressed in pg antigen/sample site.

DATA ANALYSIS

As this method captures the antigen directly from the sulcus without the collection of sulcular fluid, it is not necessary to measure fluid volume, and the antigen values will be expressed directly as an amount (pg) per site. Also, the total number of beads recovered from each sulcus will be determined (see above). Thus, the amount of antigen will be expressed as pg antigen/number of beads recovered/site. These data will be analyzed for differences in the cytokine profile of the crevicular fluid in our sample of patients. We will then compare levels of antigen in various clinical conditions to determine if correlations exist.

All measured variables will be first treated with descriptive statistics. As described above, raw root resorption amounts and the root resorption index will each be 50 measured twice at on separate occasions. The difference in measurements between the two time intervals will describe measurement error from the periapical films. The three methods will be compared against each other for measurement error and reproducibility

(reliability).

The various clinical (independent) variables will be similarly treated with descriptive statistics. TNF data will also be analyzed descriptively. The data will be evaluated to check for the possibility of pooling samples in order to differentiate between subsets of the sample population. Differences between groups will be evaluated by

ANOVA and t-tests. Separate means will also be constructed for data collected for contralateral and opposing teeth. We will test the equalities of the means and, if not equal, evaluate their differences. This will be performed by a one-way analysis of variance (ANOVA) to test the equality of this mean data across the three study groups.

Should the ANOVA reveal statistically significant results, we will use multiple comparison procedures to isolate specific pairwise differences. Prior to implementing the

ANOVA, we will perform data transformations if the extent of variation across groups indicate this to be appropriate. The clinical variables will be separately evaluated for correlation to TNF values. Stepwise linear regression models will be used to evaluate the contribution of each variable to TNF found in the sulcus.

As TNF values are expected to be highly variable, grouping of data for t-tests may not be sufficient to analyze this data. Instead, each subject will serve as both the experimemal and control. The differences in root resorption index and amount will be evaluated against TNF values in each patient. The remaining clinical variables will be 51 treated the same way as necessary. All statistical analyses will be with the advice and assistance of the University of Connecticut Health Center Biostatistical Facility.

We have investigated the power of the ANOVA model to be used in our study.

By way of example, existing normative pilot data for TNF (159 pg/106 beads +/-25) indicate that it will be possible to detect small to moderate effects of sizes with high power. For example, specifying an alpha level of 0.05 for the ANOVA, we should detect a mean difference of 20% with 52 subjects per group; a mean difference of 30% with 21 subjects per group; a mean difference of 40 % with 14 subjects per group; and a mean difference of 50% with 8 subjects per group. The percent mean difference is based on comparing the largest and smallest of the group means with the power at least 0.70 for all tests. These values depend upon the reliability of the pilot data. DIFFICULTIES

The potential problems of recruiting patients for clinical research also apply to our project. Locating patients who meet our criteria with the appropriate radiographs is difficult. Compounding the problem of sustaining an adequate sample size is the requirement of locating enough patients with the more severe resorption pattem. The literature indicates that mild resorption should be common in orthodontic patients.

However, those patients with a more severe resorption pattern are rare. Their presence in our sample will help to stratify our results and demOnstrate possible differences.

The basic science phase of this project presents with its own difficulties.

Originally, our intent was to examine a panel of cytokines from the gingival crevicular fluid. IL-1, PGE-2, and acid phosphatase levels were to be measured along with TNF, as they are also important mediators in bone metabolism, tooth movement, and tooth eruption. Each mediator has a significant and, to some degree, unique role.

Incorporating the other cytokines into the existing study would provide useful information and increase the significance of this study. However, time requirements and funding for a separate laboratory phase each mediator prohibited the greater venture.

TNF was chosen as the study molecule for reasons previously described.

Previous research involving TNF values from gingival crevicular fluid have shown variability which could not be explained by clinical conditions. The same difficulty exists for incidence of root resorption. Investigators attribute this variability to individual variation within their respective samples. By comparing differences with each

52 53 patient separately, we hope to eliminate the problem of variation between individuals.

However, intra-subject variation still exists as a problem.

The method used to capture and quantify the TNF molecules, the immunobead technique, presents with its own difficulties. Loss of a significant portion of the beads may result in extreme TNF values. Unless the recovered beads are counted in every sample, extreme outliers may have to be discarded. TNF measurements of zero or those above 1500pg should be judiciously interpreted if at all.

Finally, the presence of TNF in the sulcus as a result of an event ongoing at the apex of the root seems unlikely. If the TNF is synthesized and secreted for a specific purpose (root resorption, bone remodeling, etc.), then it should be used for that purpose rather than being dumped out into the sulcus. Cells with receptors for this molecule are far more abundant elsewhere. Also, if the presence of TNF in the sulcus is intentional, its removal by our sampling technique should cause a deficiency somewhere. SIGNIFICANCE

Clinically, the use of bead-bound antibodies would be an accurate test to

determine levels of any number of molecules in the gingival sulcus and an index of what may be occurring metabolically in the periodontal membrane. Also, it could be a

sensitive and definitive method for the diagnosis of root resorption and its possible severity. Our results will be evaluated to determine if the technique allows for the discrimination of root resorption in various phases of orthodontic treatment. Finally, and most important, if specific molecules can be identified that participate in or precipitate root resorption, one can envision the use of antagonistic drugs as a pharmacologic prevention of root resorption in the future.

Three outcomes are possible: 1) an elevation of TNF levels is observed in relation to x-ray detectable root resorption, 2) there is no change in cytokine levels as detected by our methods, 3) there is a spectrum of increases and decreases in relation to the time of onset of the untoward root resorption.

54 RESULTS,

SAMPLE SELECTION

The original sample consisted of fifty nine subjects selected on the basis of the inclusion criteria as previously described. Forty eight of these subjects were also involved in another project to evaluate the progression of root resorption. For the purposes of the other project, these subjects had periapical radiographs taken with a standardized measurement jig on a three month basis. The other eleven patients were selected on the basis of the inclusion criteria and mid-treatment periapical radiographs indicating extemal apical root resorption on at least one tooth. All patients were involved in comprehensive orthodontic treatment in our clinic, had appropriate periapical radiographs at the start of treatment, and presented for treatment in the permanent or late mixed dentition.

Of the sample of fifty nine subjects, nineteen patients were included in the final patient sample pool. The remaining were excluded either on the basis of their refusal to participate in our study (twenty eight), termination of orthodontic treatment (four), or lack of evidence of radiographic root resorption on any teeth (seven). Our final sample consisted of ten males and nine females. The mean age of the sample was 14.2 years with a standard deviation of 4.9 years. The youngest patient was 11 years old, the eldest was 26, and the median age was 12.

A total of 110 samples were taken from the nineteen subjects, averaging 5.8 teeth per patient. Since only effected teeth- those with radiographic evidence of root

55 56 resorption- and comparable control teeth were sampled, the number of samples per patient varied. At a minimum, four samples (4 teeth) were evaluated from each patient.

The maximum was 10 samples taken from the same patient. All sampled teeth were maxillary teeth, and most (78%) were incisors. Table 14 and figure 5 display the distribution of the teeth included in our sample.

DESCRIPTION OF THE SAMPLE

A multitude of independent variables were recorded for each tooth as described previously in table 9. The emire sample had no familial history of root resorption and no significant medical history. None of the patients were treated with rapid maxillary expansion, reported any pernicious habits, or displayed an openbite malocclusion. Only one tooth in our sample (110 teeth) had a history of previous trauma (2.5 years prior to treatment). Three teeth had been previously restored (minimum of one year prior to sampling) with moderate to conservative restorations. None displayed any signs of periapical pathology or had been treated with root canal therapy.

From pre-treatment periapical x-rays, most teeth (80%) displayed an unremarkable root anatomy whereas 20% demonstrated abnormal root morphology (table

15 and figure 6). The most frequent type of deviation in root form was blunting (15.3%).

Pipette shaped roots and dilacerations were rare in this sample at 2.4% and 1.2% respectively.

Various indices were used to assess gingival conditions. Three teeth (two percent) had sulcus depths beyond three millimeters. In all three instances, probing depths measured at 4mm and were in sites of gingival inflammation (pseudopocketing). 57

No teeth in our sample showed loss of attachment or bony support. Tables 16-18 and figures 7-9. describe the plaque index, gingival index, and periodontal index for the sample. Only one tooth displayed severe plaque accumulation while none displayed severe gingival or periodontal index scores. Most teeth had mild to moderate plaque scores (86%), gingival index scores (84%), and periodontal index scores (82%).

The time between when patients reported brushing their teeth and when the tooth was sampled is displayed in table 19 and figure 10. All patients had brushed their teeth between two to eleven hours prior to sampling.

A subjective assessment of the type of orthodontic movement and the mechanics involved directly prior to sampling was perfolxned to evaluate for possible correlation with dependent variables. Table 20 and figure 11 depict that 81% of the samples were involved in some form of orthodontic tooth movement directly prior to sampling, the most common type being tipping (53%). Twenty six percent of the samples were translated directly prior to being sampled while only 2% were rotated.

Force application to the sampled teeth was recorded both at the appointment prior to sampling and also at the time of sampling, representing the force remaining from the activated appliance (table 21 and figure 12). Nineteen percent of the sample was not orthodontically moved and thus had no force application. Moderate forces of 100-250 grams were applied to 59% of the sample and light forces (<100grns) were used for the remaining 22%. Heavy forces (>300grns) were not used throughout the sample. At the time of sampling, only light forces (74%) or no force at all (26%) still remained.

The time in treatment in our patient sample ranged from 0 to 36 months (table 22, figure 13). The majority of the sample (85%) had been in treatment for a year or less. 58

The remaining 15% of the sample were into the third year of treatment. Forty percent of the sample was in their 12th month of treatment.

DETERMINATION OF ROOT LOSS

Three methods were employed in measuring root loss (table 23, figures 14 and

15). Method # 1 involved direct measurement of the entire tooth length from pre and post treatment periapical x-rays. The difference in length from pre to post treatment was defined as root loss (RL). In our sample, the mean root loss as determined from method

# 1 was 0.69 mm, with a standard deviation of 0.63 mm (standard error 0.0684), and a range from -1.7 to 2.8 mm. In method #2, the length of the crown, which should remain relatively constant throughout treatment, was used to correct for magnification error. The corrected root loss (RL*) was calculated to be 0.69 +/-0.63 mm (SE 0.0687). The range was -1.69 to 2.86 mm. The mean of the root resorption index (RR1) from method #3 was

0.86 with a standard deviation of 0.80, standard error of 0.0872, and range from 0 to 3.

Direct measurements of the crown and/or the entire tooth for the first 70 samples, as in methods #1 and #2, were repeated at a separate interval by the same investigator to assess the reliability (reproducibility). Subjective assessment of root loss in method #3

(root resorption index), was performed for the entire sample at two separate intervals.

Reliability between the two intervals, where the same measuremem was made, was assessed by measuring the correlation between the two sets of measurements. Table

24 displays the descriptive statistics and correlation coefficients for the difference in repeated measures. The correlation between all direct repeated measures was greater than

0.98. Repeated RRI measures showed a correlation of 0.89. Descriptive statistics and 59 correlation coefficients were also calculated for the difference in repeated RL and RL* measures (table 25). RL and RL* both display a slightly higher error than RRI (0.81 and

0.82). Correlation coefficients between the three methods were calculated. The correlation between the corrected (RL*) and the uncorrected (RL) root loss was 0.99.

The correlation coefficients between the root resorption index and method 1 and method

2 were both 0.72.

TUMOR NECROSIS FACTOR-a MEASUREMENTS

Of the 110. samples taken, only 85 .were used for data analysis regarding TNF-c levels from the gingival crevice. Data from the other twenty five samples were discarded because of technical problems encountered in the laboratory phase during quantification of the captured TNF-ct. The first column of numbers in table 27 displays descriptive statistics for TNF-ct levels. The mean for the sample of 85 was 232.9 pg +/- 422.4 (SE

45.81). Zero TNF-ct values were found for 47% of the sample. The remaining 53% were distributed in a highly skewed manner ranging up to 2944 pg (table 28 and figure 16).

Multiple methods were employed to evaluate the TNF data. TNF values were transformed to ordinal data using an index ranging from 0 to 5. The TNF index reduced the skewness of the data (table29 and figure 17) having a mean of 1.27 +/- 1.57 (SE

0.1702). Transformation of TNF data, using its natural logarithm, created difficulties in that the samples with values of zero could not be included (zero has no natural log). To overcome this problem, all TNF values were increased by one (picogram). This allowed 60 calculation of the natural logarithm for all samples (table 30, figure 18) further reducing the skewness.

ZERO TNF-ct MEASUREMENTS

As mentioned above, zero TNF-a values were found in 40 of the 85 samples

(47%). In order to evaluate for discriminatory variables in this subset of the samples, student's t-tests were carried out for a number of variables (Tables 31-41). With respect to the Gingival Index, the Periodontal Index, and all three methods of assessing root loss, samples with positive TNF-a levels showed a significant difference from those with none at all (p<0.01). Due to the differences found between the subsets in our sample, some statistics were calculated separately for the subset with positive TNF values.

In calculating the descriptive statistics displayed in table 42, the samples with

TNF values of zero were omitted. This decreased the sample size to 43, increased the mean to 437.5 pg, with an increase in standard deviation and error to 508.5 and 77.55 respectively. The same procedure was carried out for TNF index values producing a mean of 2.35 +/- 1.39 (SE 0.2128). With the samples containing zero TNF values removed, the natural logarithm could be calculated directly. The frequency tables and histograms for all three methods are illustrated as tables 43-45 and figures 19-21. 61

CORRELATION TO TNF-tx

Correlation coefficients for TNF data and clinical variables are displayed in tables

46-48. As the distribution of TNF values was not normal (Poisson distribution),

Spearman's ranked correlation coefficients were also calculated. Correlation coefficients for RL, RL*, and RRI to raw TNF-c levels were 0.22, 0.21, 0.37 respectively. Moderate to mild correlations were also found for the Gingival Index (0.49), the Periodomal Index

(0.42), and the type of tooth movement (0.22). When using Spearman's ranked correlation, the coefficients for the GI, RL, and RL* increased to 0.67, 0.35, and 0.43.

The PI and RRI coefficients dropped to 0.42 and 0.36, whereas the type of tooth movement showed no correlation to ranked TNF data.

Similar calculations were made, omitting the samples containing TNF values of zero (tables 49-51). More variables displayed moderate levels of correlation with raw and ranked TNF data. The time since the patient last brushed their teeth, force application, the type of tooth movement, and the time in treatment had coefficients of

0.29 to 0.49). RL and RL* showed almost no correlation with TNF (raw or ranked data).

The correlation of RRI to TNF dropped to 0.32 and 0.29.

PATTERNS SPECIFIC TO PATIENTS

The variation in samples, specifically with regards to TNF values, created difficulties in pooling data and finding baseline measurements to use for comparison. As described in the methods section, we attempted to analyze data collected for each patient individually. A more complete regression model could not be constructed due to the lack 62 of data points for each patient. Also, some variables were constant across samples from the same patient (time in treatment, brushing, etc.). An analysis of the association of

TNF to root loss was performed for each patient by measuring the correlation between root loss (RL) and raw TNF values (table 52). Of the 19 patients in our sample, 5 showed a moderate to high correlation between TNF and RL. The remaining 14 patients had mild, none, or a negative correlation between the variables.

STEPWISE MULTIVARIABLE REGRESSION MODEL

As it was the purpose of this project to evaluate the effect of root resorption, and other clinical variables, on the levels of TNF-ot in the gingival sulcus, it was deemed that a regression model would be more appropriate for data analysis. By using a stepwise multivariable regression model, the variation in the TNF-et data (the dependent variable) attributable to the independent variables could be calculated. Thirteen clinical variables and all three methods of assessing root loss were used as the independent variables (tables 53-55):

Variables were dropped in a stepwise backwards elimination manner if the p- value of their coefficient of determination exceeded 0.05. The GI, RRI, and type of movement remained to explain for 37% of the variation in raw TNF-c values (table 56).

The coefficient of determination for GI and RRI alone was 31% (table 57). The Gingival

Index and the Root Resorption Index explained for 29% and 14% of the variability in

TNF levels respectively (table 58 and 59). Collinearity of the variables was examined 63 from correlations already performed (tables 46 48). No redundancy could be attributed to the R values from the three variables (GI, RRI, and movement).

As mentioned before, TNF-a values were not normally distributed. Traditionally, data transformation for regression of non-normal data involves using the natural logarithm of either both variables, or only the independent variable(s). For the purposes of our data, we required transformation of only the dependent variable. Again, several methods were used to evaluate the legitimacy of our regression model.

The first method involved the use of ranked TNF values, followed by stepwise regression as before. The Gingival index and the RRI still retained a significant p-value.

However, the type of tooth movement was not found to retain a significant p-value, whereas the Plaque Index did (table 60). Combined, the three variables explained 56% of the variation in ranked TNF-c scores. The Plaque Index was dropped, lowering the Rz value to 0.52 (table 61). Individually, GI and RRI had Rz values of 0.31 and 0.20 (tables 62 and 63).

The second method to normalize the data involved using the natural logarithm

(In)of the TNF-ot values in the regression model. Table 64 displays that, PLI, GI, and

RRI were left with p-values less than 0.05 once again, and explained 54% of the of the variation in lnTNF values. GI and RRI retained a Rz value of 0.51 together, and 0.42

(GI) or 0.19 (RRI) individually (tables 65-67).

The TNF index, described above, was used in our third method. Table 68 and further analysis showed that although the R value for PLI, GI, and RRI is slightly lower 64 at 0.49, use of the TNF index showed essentially the same pattem as using lnTNF or ranked TNF values.

When performing multivariable stepwise regression on the data without the samples with zero TNF- values, no combination of variables could be found which would retain a significant coefficient of determination. GI and RRI could explain for

36% and 10% of the TNF variation when examined individually (tables 69 and 70).

When the positive TNF data are ranked, GI, RRI, and the time since brushing retained significant R2 values of 0.58 (table 71). Individually, they could explain for 43%, 8%, and 16% of the variation in ranked TNF values (tables 72-74). Using the lnTNF and TNF index data, a similar pattern was found where GI maintains a moderate association with

TNF and the association between RRI and TNF wavers between weak to almost none. DISCUSSION

DESCRIPTION OF THE SAMPLE

The lack of severe apical resorption in our sample precludes conclusions regarding the association of increased TNF-ct levels in the face of fulminating root resorption. According to the literature, this type of resorption (beyond 3mm) may be expected in approximately 10% of orthodontic patients. Furthermore, selection of particular teeth, maxillary incisors and/or teeth with deviating root forms, may increase the likelihood of locating the teeth which experience a more aggressive resorption pattern. Our sample consisted mainly of maxillary incisors (80%) and included a fair number of teeth with deviated root forms (20%) Aside from the selection of patients with other possible contributory factors such as familial history of root resorption, significant medical history, pernicious habits, and recent trauma, not a great deal more may be done in order to produce a sample that should to include more patients with severe root resorption. Furthermore, the possible contributory factors listed above, and those mentioned in the introduction, do not guarantee location of those severely effected teeth.

Most patients in our sample were adolescents. Some studies indicate that younger patients are less prone to root resorption as their roots have only recently completed formation or are still in the process.4, 11, 22-24, 35-38, 60, 65, 96 This explains why a few samples showed an increase in root length throughout treatment, and possibly why the amounts of root resorption throughout our sample were small. It must be mentioned

65 66 that there is no conclusive evidence regarding an association between age and severity of root resorption in the literature. 97

The simplest and most direct means to increase the number of teeth with severe resorption in our sample is to select a larger sample. In our study, the original 59 patients selected would have yielded in excess of 250 samples (4-5 samples per patient). If the

5% incidence seen in other studies held true for ours, we would still only find 13 teeth with severe resorption. Enforcement of the exclusion criteria, termination of treatment, and disinterest for participation in the study by the patient (usually the parent) reduced our patient pool to 19. Furthermore, of the 110 samples collected, 25 had to be discarded.

Some were due to the loss of too many beads (and captured TNF-ct) during the purification or the collection phase. Others were simply discarded due to errors in technique during the assay. Laboratory difficulties reduced our sample size to 85 teeth.

Again, applying the 5% incidence rate for severe root resorption, we would expect approximately 4 teeth with the more severe resorption pattern. We found two teeth with resorption near or just beyond 2.5mm.

DETERMINATION OF ROOT LOSS

Root loss was calculated by three means The first method was a direct comparison of pre-treatment to post-treatment root length. In the second method, we attempted to correct for magnification error by using the length of the crown (which should be more stable than the root) to establish a ratio to aid in calculation of the relative tooth length in either film. These two methods produced very low measurement error and 67 greater than 80% correlation between repeated measures of the first 70 samples. The measurement error in this study resembles similar studies which also employ periapical films. The two methods also correlate highly with each other, producing essentially the same error and the same data. In light of this, the first method is preferred as it is quicker and simpler. Method #3 incorporated a subjective assessment of root loss using a calibrated ordinal index. Repeated measures of the Root Resorption Index actually produced a higher correlation (0.89) between repeated measures than the direct root loss methods #1 and #2. This method provided a simple but effective way to assess root resorption which may be invaluable in epidemiological studies.

TNF-c LEVELS IN THE GINGIVAL SULCUS

The samples which were used to collect data provided a high range of TNF-c levels with almost 50% registering at zero. For the purpose of data analysis, the highly skewed TNF values were transformed using multiple techniques. An index was first used to compress the range of TNF values. The natural logarithm provided a more normal distribution across our sample. The transformations were similar, both reducing the skewness and compressing the data by reducing the range. TNF values were also ranked, which is the preferred method for evaluating correlations in a sample which is not normally distributed. Samples with zero values were removed in order to view the positive data separately and the previous data transformations were performed to repeat similar results. Removal of the zero values normalized the data somewhat, mostly in the case of the TNF natural logarithms. 68

The high number of samples with zero TNF values creates some concem regarding the contribution of these samples to our data analysis. A reading of zero TNF-c in a sample from the gingival crevicular fluid may occur by various means.

1) Sampling error and lab error accounted for the loss of twenty-five samples. It is

possible that samples which registered at zero were also affected by lab error.

Our sampling technique may have failed to register the presence of TNF-c in the

sulcus, reading as a false negative. This may reflect errors in the collection

technique from the sulcus, a loss of the bead-bound antibodies during purification

or processing, or an error during the assay.

2) There may have been no TNF-c in the sulcus at the time of sampling. This can

occur if the TNF was washed away by tooth brushing or rinsing (still a false

positive). A student's t-test failed to find a significant difference between the zero

TNF samples and those which registered positive for TNF with respect to when

the patient last brushed their teeth (table 33).

3) Finally, we may have had readings of zero simply because there was no reason for

TNF to be present in those samples (an accurate reading). A statistical difference

(p < 0.01) was found between the zero and positive TNF samples with respect to

GI, PeI, RL, RL*, and RRI (tables 37-41). These variables, alone or in

combination, may contribute to the presence, or lack of TNF-c in the gingival

sulcus. Previous research involving TNF-ot harvested from gingival crevicular

fluid describes similar results where zero or very low levels of the marker were

found. 153, 212, 229 In light of these findings, clinical perameters such as 69

gingival inflammation and root resorption provide the most plausible explanation

for the finding of zero TNF levels. All five of these variables correlate with one

another to a mild degree requiting further analysis to pinpoint the most significant

contributor to the dichotomy in our data (tables 46-48).

THE ASSOCIATION BETWEEN TNF-c AND CLINICAL VARIABLES

Although correlation coefficients were calculated for all recorded variables, in order to investigate the relationship between them and TNF-et (the dependent variable) levels a regression model was deemed more appropriate. Several models were constructed in which a backwards stepwise analysis was performed. Four methods were used in evaluating TNF data: raw TNF-ot levels, ranked TNF values, transformed values using natural logarithms, and the TNF Index.

In all models, the Gingival Index maintained a level of significance with a coefficient of determination of 0.29-0.45 The Root Resorption Index contributed to 14-

20% of the variation in TNF levels when considering the entire sample. When omitting the samples with values of zero TNF, the RRI failed to be consistent, or provided a low coefficient of determination (R < 0.10) in the various methods of interpreting and transforming the TNF data. A few other variables became significant in the various trials, but failed to prove as consistent as GI and RRI. 7O

The association between root loss and TNF-t values

The direct measures of root loss, RE and RL*, failed to show a significant coefficient of determination in relation to TNF data. The correlation between either direct measure and the RRI was moderate to high and significant (0.72, p<0.001). All three methods were designed to measure root loss. If they correlate with one another and are essentially different rulers to measure the same variable, they should have the same relationship with TNF- levels. Due to the fact that R values for RRI with respect to

TNF data were low and the correlation with other measures of root loss (which failed to show a statistical relationship with TNF) were high, the association between RRI and

TNF values is questionable.

According to the previously mentioned t-tests (tables 39-41), all three methods of determining root loss showed statistically significant difference when disceming between no TNF and some TNF in the sulcus. It would be difficult to explain the significance in all three methods and the significant R2 value for the regression model of RRI and TNF as a result of type I errors. When considered together, the data precludes a conclusion of a linear dependent relationship between the degree of root loss observed in this study and the TNF-c values observed. However, the coincidence of data between zero and positive samples in relation to root loss indicates some association.

Scientifically, the relationship between root loss and TNF- in the gingival crevice is not fully explained in the past literature. Root resorption occurs in the periodontal ligament, which is sealed off from the gingival crevice by junctional epithelium. Although this attachmem is not impervious to molecules like TNF-c, the 71

most likely source for its presence in an otherwise healthy sulcus remains to be the

gingiva (and its blood supply).

Gingival conditions and TNF-tz levels

Several studies have found that molecular signals harvested from the crevicular fluid reflect the progression and severity of periodontal conditions (table 6). The high correlation between TNF-tx levels and the Gingival and Periodontal index in our study validates our sampling technique. The correlation was higher for the gingival index probably due to the greater range with which it presented throughout our sample. This range helped to stratify the data in respect to the gingival index and in comparison to the periodontal index (or other variables).

Regression models indicated a moderately strong dependent relationship between

TNF-ct levels in the crevicular fluid and the Gingival Index. Depending on the how the

TNF data was used, the Gingival Index alone could be used to explain 30-42% of the variation in TNF values.

If TNF-ct levels in the gingival crevicular fluid were totally dependent on gingival conditions only, then, most of the data points relating the two variables should lie on the regression line. In that case, we should be able to predict gingival inflammation, or have a quantitative measure of its severity by measuring TNF-a levels.

Perhaps, the Gingival Index is not totally accurate in describing gingival inflammation.

This may easily be true, as it is a simple index based on four categories. Because TNF-ct may be measured on a continuous axis, as it was in this study, the linear dependent 72 relationship between the two may be "fogged" and display a lower coefficient of determination. The only other explanation would be that other variables may contribute to TNF-t levels in the gingival sulcus and only part (maybe most) of the variability is due to the variation in gingival conditions.

TNF-a in the gingival crevicularfluid

Periodontal literature confirms the relationship between periodontitis and crevicular molecular signals. 153, 230 The actions of molecules such as IL-1 and TNF are well recognized in bone metabolism, specifically, in the destructive process involving the breakdown of alveolar bone in periodontitis. Similar to root structure, alveolar bone is relatively remote from the sulcus. Albeit, radiographic root resorption occurs at the apex, and alveolar bone loss occurs at the crest of the bone, very near to the sulcus. However, the process of transfusion from the PDL to the sulcus can and does occur in periodontitis, and may occur (over a slightly longer distance) during root resorption. Furthermore, root resorption occurring at the apex is probably also progressing on the axial walls of the root at all levels, and cytokines mediating this process may be produced at these higher levels thus closing the gap to the gingival crevice. Finally, the markers which appear in the sulcus need not originate or even be intended for the PDL. Stimulus to increase production of these factors is a local phenomena, however, it may reach slightly further than intended. Nearby cells in the gingiva may be similarly affected and begin production of mediators for metabolic functions in the periodontium related to the 73 processes occurring at the level of the bone. In other words, the presence of TNF-c in the gingival sulcus, in relation to root resorption, is far from remote.

The more interesting question involves the rationale for the presence of TNF-o in the gingival sulcus. When stimulated, cells synthesize and secrete molecules such as

TNF-tx for purposes of communication and metabolic function. These molecules have an intended purpose and target. They should be taken up by target cells or broken down when not utilized. The rationale for the presence of"excess" cytokines in the gingival sulcus may provide answers regarding metabolic processes and pathologies in the region in and by itself.

1) Human physiology does allow room for errors and excess. These molecules

are secreted for the purposes of metabolism and communication. A cell

population in a person who is overly susceptible (genetically) to specific

stimuli may act overzealously to minor insults. A decreased threshold to

trigger the production of cytokines or an overly active production mechanism

for these molecules may be responsible for excessive amounts of these

molecules. When secreted in excess and when not fully utilized or

metabolized, they may "wash out" into the gingival sulcus. These excess

molecules will then have no actual metabolic purpose and serve only as

markers for metabolic functions if so investigated. This mechanism dictates

that the excess secretion of these molecules reflects a hyperactive production

mechanism where, in our model, excessive resorption of the root will occur, 74

and enough molecules are secreted so that the excess will also be reflected in

the wash out in the sulcus.

2) The other explanation involves a mechanism where the normal uptake or

feedback loop is disabled. If the intended targets are not able to respond to or

bind these molecules they will .accumulate in excess and wash out into the

sulcus. The overall effect.in thispathway will be the reverse of the first. If

the target cells do not bind these molecules, they may remain inactive. In our

.model, increased levels of wash out will reflect a decreased amount o.f reso.rptive activity.

As we were not able to find a dependable linear model to predict root resorption from crevicular TNF-x levels, it is difficult to know which pathway is responsible for the presence of cytokines in the gingival sulcus. Better understanding of the interaction of these molecules with their targets and the subsequent actions of the targets with regards to root resorption is also required to validate either model.

GENERAL DISCUSSION

It is well accepted that with orthodontic treatment some degree of root resorption occurs. As most sites of root resorption are repaired, orthodontic treatment provides benefits which far outweigh the risks. In fact, minor loss of root structure, even when left un-repaired, is not a significant dental health problem in orthodontic patients. Rarely, this resorption is severe and may result in the loss of periodontal support which may compromise the longevity of the tooth. The mechanism or sequence of events which 75 cause a more pernicious resorption pattern is yet unknown. No specific, or even set of clinical criteria have been identified as contributory factors to this more severe occurrence of root resorption.

Previous investigators have identified the cell population and mediators involved in tooth movement and root resorption. Basic steps involved in the biology of tooth movemem have been studied. For example, the levels of specific mediators and the number of clastic cells increases as a response to the application of mechanical stresses to the PDL. Also, similar processes and increases in mediators is seen where a pathologic state exists. Given these circumstances, one would expect levels of mediators to increase where root resorption is progressing unchecked by a proper repair response.

The loss of root structure occurs generally and to a mild degree during tooth movement. When root loss occurs excessively or at an aggressive rate, the balance between resorption and repair is impaired. Macroscopically, this is seen as the loss of significant amounts of root structure. Closer examination by histologic methods or biochemical sampling shows increased resorptive activity by clastic cells which are being mediated by increased levels of cytokines.

Our study incorporated a sample where resorption of root structure was clinically insignificant and unchallenging to the longevity of the teeth examined. The amount of root resorption was limited to mild occurrences, limited to less than 2.5 mm. TNF-a production reflected a balance between resorption and repair (deposition) or what is typical for the gingival status. The variation in TNF-ot production in the gingival sulcus 76 failed to distinguish differences in the severity of root resorption. This indicates that minor resorption of root structure are not recognized as a pathologic state.

FUTURE STUDIES

Ideally conducted, this study should incorporate a larger sample. This would increase the chance to incorporate a larger number of teeth with a more severe root resorption. Abutment teeth for rapid palatal expanders can be used, as they have been shown to experience significant amounts of resorption due to the high forces generated by the expander. Teeth planned for extraction could be used as samples. This would allow histologic, as well as biochemical examination.

The relationship between gingival conditions and the production of crevicular cytokines is well studied in the literature and demonstrated in the present study: All attempts should be made to control for these variables. For example, a strict protocol of oral hygiene should be implemented in the pool of patients. This can include administration of antibacterial agents (Triclosan, Chlorhexidine, etc.) to control or reduce gingival inflammation. Elimination of the contribution or affect of these significant variables allows a much clearer look at the affects of root resorption.

Animal models have been used in the past to evaluate root resorption. Most histologic and immunohistologic studies involve the use of an animal model.

Unfortunately, it is just as difficult to incorporate large numbers of teeth with pernicious root resorption in animal models. Closer examination of the involvement and actions of cytokines in root resorption should be pursued in these animal models. Also, the cell 77 population involved in the process has been studied, but a quantitative assessment of the cells still needs to be performed. CONCLUSION

1) The results of this study indicate that assessment of root loss may be performed

reliably and efficiently by use of an ordinal index.

2) A great amount of individual variation exists in the amounts or presence of Tumor

Necrosis Factor-x in the gingival sulcus.

3) The sampling technique for measuring Tumor Necrosis Factor-a in the gingival

sulcus is valid and derived TNF values are consistent with studies of gingival

inflammation.

4) The lack of gingival inflammation or the lack of root resorption may result in the

absence of Tumor Necrosis Factor-c in the gingival sulcus.

5) The results of this study suggest that within the realm of clinically insignificant root

resorption, the levels of Tumor Necrosis Factor -or in the gingival sulcus cannot be

used to approximate the amount of root resorption occurring.

78 Figure 1 Bone: :metasm. C+ont:: but:ion of c):o.kes, ho:ones, d vous groh c..o.rs in regtion ,d :udon of ne .taiism, Ten om, !on SH, Role ofc:}okMes= N ccM disorders ofne :[Rto i992 #349]

79 8O

F:ige 2..= Methods #I. d #2 :ib:r de,:eig root: loss. 8i

.m, bead

coated, sus!nded: i:n buffer soiution

Figure 3.. i:nNmno:nmgnetic beads. 5 g,m. in diete.r, e coated, wit,h ti.'l'NFa anti.dies and used to capture T'-c,., moleeui:es om the gingivaI suicus.. 82

F:igure, 4, lz,. ,ISA is 'ed,-t:o qut:i" the aot ofTNF este,d :om, the g,mg,, ai 83

Frequency Distribution of Sampled Teeth Cumulative %

18 100.00%

16 90.00%

80.00% 14 70.00% 12 60.00% 10 50.00%

40.00%

30.00%

20.00%

10.00%

.00% 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Figure 5. Distribution of teeth in our sample. 84

Distribution of Root Forms

Norm al B lu nted D ilacerated Pipette

Figure 6. Distribution of normal and deviating root shapes. N=85

Frequency Plaque Index Cumulative %

4O 100.00%

35 90.00% 80.00% 30 70.00% 60.00% 50.00%

15 40.00% 30.00% 10 20.00% 10.00% .00% 0 2 3

Figure 7. Distribution of the Plaque Index. 85

Gingival Index Frequency Cumulative %

40 100.00%

90.00% 35 80.00% 30 70.00%

25 60.00%

20 50.00%

40.00% 15 30.00% 10 20.00%

10.00%

.00% 0 1 2 3

Figure 8. Distribution of the Gingival Index.

Frequency Periodontal Index Cumulative % 70 100.00%

90.00% 60 80.00%

50 70.00%

60.00%

50.00%

40.00%

20 30.00% 20.00% 10 10.00%

.oo% 0 1 2 3 More

Figure 9. Distribution of the Periodontal Index. 86

Frequency Time Since Last Tooth Brushing Cumulative %

25 100.00%

90.00%

20 80.00%

70.00%

15 60.00%

50.00%

10 40.00%

30.00%

20.00%

10.00%

.00% 1 2 3 4 5 6 7 8 9 10 11 More

Figure 10. Time since subjects last brushed their teeth until the time of sampling (in hours). 87

Types of Tooth Movement

No Movement Translation Tipping Rotation

Figure 11. Types of tooth movement attempted during the period prior to sampling. 88

Force Application Ii Applied Remaining

None Light Moderate Heavy Force

Figure 12. Force application determined at the visit prior to sampling (applied) and at the time of sampling (remaining). 89

Frequency Time in Treatment --B- Cumulative %

35 100.00% 90.00% 30 80.00% 25 70.00% 60.00% 50.00% 40.00%

10 30.00% 20.00% 10.00% .00% 0 4 8 12 16 20 24 28 32 More Months

Figure 13. Time in treatment from initial banding visit (months). 90

Amounts of Root Loss

Figure 14. Amounts of root loss in our sample as determined from method # 1 (RL). 91

Amounts of Root Loss

0 1 2 3 More RRI

Figure 15. Amounts of root loss as determined from method #3 (RRI). 92

Frequency TNF Values Cumulative % 40 100.00%

90.00% 35 80.00% 30 70.00% 25 60.00%

20 50.00%

40.00% 15 30.00% 10 20.00%

10.00%

.00% 0 300 600 900 1200 1500 More

Figure 16. TNF-ct distribution in our sample (picograms). 93

TNF Index Frequency -e- Cumulative %

40 100.00%

90.00% 35

80.00% 30 70.00%

25 60.00%

20 50.00%

40.00% 15

30.00% 10 20.00%

10.00%

.00% 0 1 2 3 4 5

Figure 17. TNF-c values tabulated as an index. 94

Frequency Ln TNF+I Cumulative %

40 100.00%

90.00% 35 80.00% 30 70.00%

25 60.00%

20 50.00%

40.00% 15

30.00% 10 20.00%

10.00%

.00% 0.00 0.89 1.78 2.66 3.55 4.44 5.33 6.21 7.10 More

Figure 18. TNF-a values transformed using the natural logarithm. 95

TNF Distribution with No Zero Values Frequency ---e- Cumulative %

20 100.00%

18 90.00%

16 80.00%

14 70.00%

60.00%

50.00%

40.00%

30.00%

20.00%

2 10.00%

.00% 300 600 900 1200 1500 More

Figure 19. Distribution: samples with zero TNF-c values are omitted. 96

TNF Index with No Zero values Frequency

18 100.00%

16 90.00%

80.00% 14

70.00% 12

60.00% 10 50.00%

40.00%

30.00%

20.00%

10.00%

.00% 2 3 4 5

Figure 20. Distribution of tabulated data using the TNF index and omitting samples with zero values. 97

[m Frequency Ln TNF (no zero values) .._.. Cumulative %

14 100.00%

90.00% 12 80.00%

10 70.00%

60.00%

50.00%

40.00%

30.00%

20.00%

10.00%

0 .00% 2.64 3.53 4.42 5.31 6.20 7.10 More

Figure 21. Distribution of transformed data using natural logarithm and omitting samples with values of zero. APPENDIX II: TABLES

Reference Sample Size Patients with Teeth with Method of Resorption Resorption Evaluation (percent) (percent) Ketcham ',293 2012 0.5 PA Becks '396 72 32 PA Rudolph '364 4560 12 PA Henry and 90.5 HIST Weinman ,5119 Massler and 708 100 86.4 PA Malone '549 Massler and 301 100 86.4 PA Perreault '54 31 Stenvik. and 35 0 HIST Mjor '7.0 29 Plets et al. '87 50 27.5 27.5 PA 32 Goldson and 42 4 PA Henrikson '75 33 Harry and Sims 10 0 0 PA/HIST '82 28 Dermaut and 15 0 0 PA Munck '86 34

Table 1. Incidence of root resorption in non-orthodontically treated individuals.

98 99

Reference Sampl Treatment Patients w/ Teeth with Method of e Size Resorption Resorption Evaluation (percent) (percent) Ketcham '29 3 500 Fixed 19 PA Becks '39 6 72 Unknown 73 PA Rudolph '36 4 439 LL 57 PA Hemley '41 81 195 Fixed 21 3.5 PA Massler and Malone 81 Unknown 100 93.3 PA '54 9 Phillips 55 10 69 Edgewise 31 incisors PA 83 Phillips '55 10 62 Edgewise 92 92 CEPH Deshields '69 12 52 Edgewise 99 82 CEPH Stenvik and Mjor '70 35 Intrusion 93 HIST 29 Sjolien and Zachrisson 59 Edgewise PA '73 115 Plets et al. '87 32 45 Fixed 46 46 CEPH Goldson and 42 Begg 100 77 PA Henrikson 75 33 Hollender et al. '80 94 12 Edgewise 50 PA Ronnerman and 23 Activator/ 39 PA Larsson 81 120 Edgewise Harry and Sims '82 28 10 Intrusion 100 100 PA/HIST Kennedy et al. '83 30 32 Edgewise 26.5 PA Kennedy et al. '83 30 32 Serial Exo 20.5 PA and fixed Kennedy et al. '83 30 32 Serial Exo 6 PA Copeland and Green 45 Fixed CEPH '86 121 Dermaut and Munck 20 Intrusion 86 PA '86 34 Sharpe et al. '87 116 18 Edgewise 89 20.1 PA Sharpe et al. '87 116 18 Edgewise 83 13.3 PA Levander and 153 Fixed 56 PA Malmgren '88 76 Linge and Linge '91 485 Mixed 16.5 PA 36

Table 2. Incidence of root resorption subsequent to orthodontic treatment. 100

Reference Sample size Treatment Mean Method of type resorption analysis Phillips '55 ] 0 62 Edgewise 1.4 CEPH Sjolien and 59 Edgewise 0.5-1.8 PA Zachrisson '73 I15 Plets et al. '87 45 Fixed 1.78 CEPH 32 Goldson and 42 Begg 6 % with PA Henrikson '75 >2mm 33 Hollender et al. 12 Edgewise >2mm PA '80 94 Ronnerman and 23 EW and 1-3 PA Larsson '81 activator 120 Linge and 719 Mixed 0.7 PA Linge '83 35 Copeland and 45 Fixed 2.93 CEPH Green '86 121 Dermaut and 20 Intrusion 2.5 PA Munck '8634 Goldin '89 42 17 Edgewise 1.35 CEPH McFadden et 38 Intrusion max 1.84mand PA/CEPH al. '89 80 0.61 Linge and 485 Mixed 2.5+ mm PA Linge '9136 Mirabella and 343 Fixed 1.47 PA Artun '95 37

Table 3. Reported amounts of root resorption subsequent to orthodontic treatment. 101

Cytokine Bone Bone Implicated in Potential formation resorption Sources Interleukin- 1 + +++ Myeloma Leukocytes Squamous cancer Osteoblasts Periodontal disease Tumors Rheumato d arthritis Osteoporosis Interleukin-6. Myeloma As above Interleukin-9 :F+ ++ Leukocytes Osteoblasts + +++ Cancer, autoimmune disease Leukocytes Periodontal disease Osteoblasts Rheumatoid arthritis, osteoarthritis Osteoporosis, multiple sclerosis TNF + +++ Myeloma Leukocytes Tumors EGF no effect +++ Tumors? TGF c +++ Squamous cancer Tumors TGF ++ ++ Osteoporosis Osteoblasts. Cancer ? Bone Matrix Leukocytes PDGF ++ ++ Rheumatoid arthritis Platelets Osteoblasts CSF ? + Osteopetrosis Osteoblasts Stromal cell Leukocytes IFN 7 Rheumatoid arthritis Lymphocytes IGF-1 +++ no effect Osteoporosis Osteoblasts Bone matrix IGF'2 +++ no effect Osteoporosis Osteoblasts Bone matrix FGF +++ no effect ? Bone matrix Osteoblasts

Table 4. Cytokines with effects on bone metabolism. Taken from Ralston SH. Role of cytokines in clinical disorders of bone metabolism.231 102

General Bone metabolism

Regulation of phospholipid metabolism Prostaglandin synthesis (synergistic) Induction of acute phase proteins Induction of other cytokines Prostacyclin synthesis Increase osteoclast recruitment Expression of human major Increase osteoclast proliferation histocompatibility complex antigens Increase osteoclast activity Fever induction Inhibit osteoblast maturation Role in rheumatoid arthritis, Decrease collagen synthesis osteoarthitis, endotoxin shock, Increase osteocalcin production tuberculosis, multiple sclerosis, Promotion of fibroblast growth glomerulonephritis Induction of collagenase secretion Activation of inflammatory leukocytes Increased in periodontal disease process Modification of vascular permeability

Table 5. Actions common to both Interleukin-1 and Tumor Necrosis Factor. 103

GCF Normal value Gingivitis Periodontitis Orthodontic Molecular Tooth Signal Movement Acid 3X increase 232 Phosphatase

233 Alkaline 20 mU/m1233 60 mU/ml 233 90 mU/ml Phosphatase

234 152 234 TM IL 1-13 68 ng/ml 131 ng/ml 404 ng/mL 0.88 pg/ug 153 TM 153 192 36 ng/ml 194 ng/ml 81 ng/mL 19.2 pg/site 1.7 pg/site 31 pg/site 1 O0 pg/site3s TM 149 0.38 pg/ug 38 pg/mL 192 2.9 pg/site 235 10 pg/site 191 TM IL 6 0.020 pg/ug 0.063 pg/ug

149 TM 0.41 pg/ugTM 3.20 pg/mL 1.13 pg/ug PGE 2 20ng/ml 40 ng/ml 120 ng/ml 1.5X 233 192 5 .ng/m1233 6 ng/m1233 10 ng/ml increase 9 62 pg/site192 109 pg/site

191 EGF 0.18 pg/ugTM 0.82 pg/ug

Osteoclacin 3 5 pg/s 90 pg/s 115 ng/m1233 55 ng/m1233 110 ng/m1233 total amt 10X

Table 6. Molecules found in gingival crevicular fluid. 104

Cellular components Bacteria, desquamated Mostly from adjacent epithelial cells, periodontal tissues lymphocytes, monocytes, macrophages, polymorphonuclear neutrophils

Carbohydrates Glucose hexosamine, 3-4 times blood levels hexuronic acid

Electrolytes Potassium, sodium, Levels correlate with calcium inflammation

Proteins Immunoglobulins (IgA, Lower than blood protein IgG, IgM), Complement levels components (C3 and No significant C4), Plasma Proteins correlation with (albumin, fibrinogen, severity of gingivitis, etc.) pocket depth, or bone loss. Enzymes Acid phosphatase, beta glucuronidase, lysozyme, cathepsin D, proteases, alkaline phosphatase, lactic dehydrogenase Bacterial and metabolic Endotoxins, lactic acid, products urea, hydroxyproline, hydrogen sulfide

Table 7. Composition of the gingival sulcus. 236. 105

Inclusion criteria Exclusion criteria Comprehensive fixed orthodontic Chronic inflammation of the gingiva reatment High caries index Permanent or late mixed dentition Chronic use of anti-inflammatory Pre-treatment periapical radiographs of medication involved teeth Smokers Mid-treatment periapical radiograph Deep periodontal pockets" 5+mm exhibiting modified root anatomy Refusal to participate in the study

Table 8. Criteria for patient selection and disqualification. 106

Patient: Date: Number: Age: Sex: M F Familial history: Y N Medical history: Y N RME Y N Y: Timesince Habit: Y N Y: Tongue Nailbiter Tongue Openbite" Y N Y: Skeletal Dental Both Tooth # 7 8 9 10 Root loss Direct measurement

R*:mm Corrected EARR: 0-4 01234 01234 01234 01234 Irregular, blunt, 1/5-1/3, 1/3+ Root NA NA NA NA Norm vs. blunt, narrow, morphology pipette, dilacerated Trauma NY NY NY NY Yes: time since Pathology NY NY NY NY yes: type RCT NY NY NY NY yes: time since Restoration NY NY NY NY yes." type and time since Plaque 0123 0123 0123 0123 Index Gingival Ind 0123 0123 0123 0123 Periodontal Scale 0-8, need x-ray Index Sulcus N P None: 1-3 mm, Depth Pocket: In mm Brushing Last time in hours Movement ITRTr ITRTr ITRTr ITRTr Intrude, tip, rotate, translate Force LMH LMH LMH LMH Light 100gm, Moderate applied 100-250, Heavy 250+ Force ( at L M H LMH L M H L M H Force remaining at the assay) time ofassay Time since Last visit in weeks Cont / Int CI CI CI CI Type offorce applied Occlusal N Y N Y N Y N Y trauma Tx time In months Time to go In months TNF

Table 9. Sample summary. The above form will be used to collect data at each visit. 107

Score Changes noted in root anatomy No discernible erosion or blunting and complete apexification Irregular root, contour, notably para-apically Root blunting with minor resorption Severe root resorption, loss of one fifth to one third of the original length Extreme resorption exceeding one third of the original length

Table 10. Root resorption index. Adapted from Levander and Malmgren:76

PI score Clinical condition Negative: no overt inflammation Mild gingivitis" Overt inflammation in the free gingiva Gingivitis: Inflammation completely circumscribing the tooth with no apparent break in the attachment or loss of function Radiograph: early notch-like resorption of the alveolar crest Pocket formation: Epithelial attachment has been broken. Tooth is firm and has not drifted. Horizontal bone loss up to half of the root. Advanced destruction beyond one half of the root with loss of masticatory function. Mobility and drift.

Table 11. The Periodontal Index (PI).237 108

GI score Clinical conditions Normal gingiva Mild inflammation, slight change in color, slight edema, no bleeding on palpation Moderate inflammation, redness, edema, and glazing, bleeding on palpation Severe inflammation, marked redness and edema, ulcerations; tendency to spontaneous bleeding

Table 12. Gingival Index (GI).224

Plaque Index Clinical condition No plaque in gingival area A film of plaque adhering to the free gingival margin and adjacent area of the tooth Moderate accumulation of soft deposits within the gingival pocket and on the gingival margin and or adjacent tooth surface, which can be seen by naked eye Abundance of soft matter within the gingival pocket and or on the gingival margin and adjacent tooth surface.

Table 13. Plaque index.225 109

Tooth # Frequency Percentage Cumulative % 1 1.2 1.18% 1 1.2 2.35% 2 2.4 4.71% 0 0.0 4.71% 6 7.1 11.76% 15 17.6 29.41% 17 20.0 49.41% 17 20.0 69.41% 10 16 I8.8 88.24% 11 6 7.1 95.29% 12 0 0.0 95.29% 13 2 2.4 97.65% 14 2 2.4 100.00% 15 0 0.0 100.00%

Table 14. Distribution of teeth included in our sample. Also refer to figure 5. N=85 110

Root Shape Count Percentage Normal 68 80.0 Blunted 13 15.3 Dilacerated 1 1.2 Pipette 2 2.4

Table 15. Distribution of normal and deviating root shapes. Also refer to figure 6. N=85

Plaque Index Frequency Cumulative % Percentage 11 12.94% 12.9 36 55.29% 42.4 37 98.82% 43.5 1 100.00% 1.2

Table 16. Distribution of the plaque index. Also refer to figure 7. 111

Gingival Index Frequency Cumulative % Percentage 12 14.12% 14.1 38 58.82% 44.7 33 97.65% 38.8 2 100.00%

Table 17. Distribution of the Gingival Index. Also see figure 8.

Periodontal Index Frequency Cumulative % Percentage 0 15 17.65% 17.6 63 91.76% 74.1 7 100.00% 0 100.00% More 0 100.00%

Table 18. Distribution of the Periodontal Index. Also see figure 9. 112

Hours Frequency Cumulative % Percentage 1 0 .00% 2 10 11.76% 11.8 3 4 16.47% 4.7 4 4 21.18% 4.7 5 11 34.12% 12.9 6 25 63.53% 29.4 7 6 70.59% 7.1 8 7 78.82% 8.2 9 0 78.82% 0.0 10 8 88.24% 9.4 11 10 100.00% 11.8 More 0 100.00% 0

Table 19. Time since subjects last brushed their teeth until the time of sampling (in hours). Also refer to figure 10.

Type of Tooth Movement Frequency Percentage No Movement 16 18.82 Translation 22 25.88 Tipping 45 52.94 Rotation 2 2.35

Table 20. Types of tooth movement attempted during the period prior to sampling. Also refer to figure 11. 113

Force Applied Rmaining Frequency Percentage Frequency Percentage None 16 18.8 22 25.9 Light 19 22.4 63 74.1 Moderate 50 58.8 0 0.0 Heavy 0 0.0 0 0.0

Table 21. Force application determined at the visit prior to sampling (applied) and at the time of sampling (remaining). Also refer to figure 11.

Months Frequency Cumulative % Percentage 0 .OO% 0.0 14 16.47% 16.5 24 44.71% 28.2 12 34 84.71% 40.0 16 0 84.71% 0.0 2O 0 84.71% 0.0 24 0 84.71% 0.0 28 9. 95.29% 10.6 32 0 95.29% 0.0 More 4 100.00% 4.7

Table 22. Time in treatment from initial banding visit (months). Also refer to figure 13. 114

Distribution RL RL* Mean 0.6871 0.6894 0.8588 Standard Error 0.0684 0.0687 0.0872 Median 0.6 0.6 1 Mode 0.2 0 0 Standard Deviation 0.6307 0.6338 0.8041 Sample Variance 0.3978 0.4017 0.6465 Kurtosis 2.5060 2.5548 -0.9318 Skewness O. 1074 O. 1528 0.4043 Range 4.5 4.5497 3 Minimum -1.7 -1.6888 0 Maximum 2.8 2.8609 3 Sum 58.4 58.6011 73 Count 85 85 85 Confidence Level(95.000%) 0.1360 0.1367 0.1734

Table 23. Descriptive statistics of the three methods used to assess root resorption. Root loss (RL) in method # 1 is the difference in the length of the entire tooth as measured from pre to post treatment periapical radiographs. Corrected root loss (RL*) in method #2 incorporates a ratio using crown lengths to remove magnification error. The Root Resorption Index (RRI) from method #3 is a subjective measurement of root loss based on an ihdex ranging from 0 to 4. Zero score indicate no root loss, whereas a score of four indicates loss of root structure beyond 1/3 of the original length. 115

Difference between Pre-tx Post-tx Pre-tx Tooth Post-tx Tooth repeated measures Crown length Crown length Length Length (Error) Mean -0.1000 -0.0314 -0.1843 -0.1729 Standard Error 0.0468 0.0467 0.0691 0.0584 Median -0.2 -0.1 -0.2 -0.2 Mode -0.5 -0.1 -0.4 -0.3 Standard Deviation 0.3919 0.3907 0.5780 0.4884 Sample Variance 0.1536 0.1526 0.3341 0.2385 Kurtosis -1.1532 5.6252 -0.3900 -0.5615 Skewness 0.2929 1.9189 -0.5049 -0.1033 Range 1.5 2.2 2.4 2 Minimum -0.8 -0.7 -1.7 -1.2 Maximum 0.7 1.5 0.7 0.8 Sum -7 -2.2 -12.9 -12.1 Count 70 70 70 70 Confidence Level 0.0918 0.0915 0.1354 0.1144 (95.000%) orrelation 0.9840 0.9886' 0.9955 0.995 coefficient

Table 24. Descriptive statistics of the difference between repeated measures. The first 70 direct measurements of the crown length and the entire tooth were repeated to assess reproducibility. Reliability between the two intervals, where the same measurement was made, was assessed by calculating the correlation between the two sets of measurements. 116

Difference between RRI (entire repeated measures RL* RRI sample) Mean 0.0186 0.0177 0.0286 0.0636 Standard Error 0.0530 0.0530 0.0454 0.0414 Median 0.1 0.0979 0 0 Mode 0.1 0.0992 0 0 Standard Deviation 0.4437 0.4435 0.3 796 0.4343 Sample Variance 0.1969 0.1967 0.1441 0.1886 Kurtosis 1.1429 1.0891 4.3419 2.2920 Skewness 0.4005 0.3875 0.3126 0.3478 Range 2.4 2.3956 2 2 Minimum -0.9 -0.9073 1 1 Maximum 1.5 1.4884 1 1 Sum 1.3 1.2378 2 7 Count 70 70 70 110 Confidence 0.1039 0.1039 0.0889 0.0812 Level(95.000O%) correlation 0.8122 0.8158 0.8872 0.8931

Table 25. Repeatability (error) of the three methods for determination of root loss.

RL RL* RRI RL 1.0000 0.9996 0.7244 RL* 0.9996 1.0000 0.7246 RRI 0.7244 0.7246 1.0000

Table 26. Correlation between the three methods used to assess root loss. 117

TNF TNF Index In TNF+I Mean 232.86 1.27 2.95 Standard Error 45.8136 0.1702 0.3159 Median 29 1 3.4012 Mode 0 0 0 Standard Deviation 422.3807 1.5690 2.9121 Sample Variance 178405.5 2.4616 8.4805 Kurtosis 20.1074 -0.1829 1.8088 Skewness 3.7713 1.0337 0.0931 Range 2944 5 7.98786 Minimum 0 0 0 Maximum 2944 5 7.98786 Sum 19793 108 250.704 Count 85 85 85 Confidence Level(95.000%) 89.793 0.334 0.628

Table 27. Distribution of Tumor Necrosis Factor alpha (TNF-) values. 118

TNF- Frequency Percentage Cumulative % 0 40 47.1 47.06% 300 20 23.5 70.59% 600 16 18.8 89.41% 900 4 4.7 94.12% 1200 3 3.5 97.65% 1500 1 1.2 98.82% More 1 1.2 100.00%

Table 28. Distribution of raw TNF-ot values in our sample (picograms). Also refer to figure 16. 119

TNF TNF index Count Percentage Cumulative % (picograms) 0 40 47.1 47.06% 100-199 1 17 20.0 67.06% 200-399 2 9 10.6 77.65% 400-599 3 7 8.2 85.88% 600-999 4 8 9.4 95.29% 1000+ 5 4 4.7 100.00%

Table 29. TNF-tx values tabulated using an ordinal index 0-5. Also refer to figure 17.

Ln (TNF +1) Count Percentage Cumulative % 0.00 40 47.1 47.06% 0.89 0 0 47.06% 1.78 0 0 47.06% 2.66 1 1.2 48.24% 3.55 2 2.4 50.59% 4.44 3 3.5 54.12% 5.33 11 12.9 67.06% 6.21 13 15.3 82.35% 7.10 13 15.3 97.65% More 2 2.4 100.00%

Table 30. TNF-o values transformed using the natural logarithm. Also refer to figure 18. 120

Tooth selection" Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 8.53333 8.475 Variance 5.39091 3.33269 Observations 45 40 Hypothesized Mean Difference df 82 t Stat 0.12943 P(T<=t) one-tail 0.44867 t Critical one-tail 1.66365 P(T<=t) two-tail 0.89733 t Critical two-tail 1.98932

Table 31. T-test for tooth selection. Student's two sample t-test for statistical differences in the teeth selected for sampling between samples with zero TNF-a levels and samples positive for TNF- 121

Root Shape: Positive for TF Zero TNF Two sample t-test, unequal variances Mean 0.2 0.35 Variance 0.34545 0143846 Observations 45 40 Hypothesized Mean Difference df 79 t Star 1.0987 P(T<=t) one-tail 0.13761 t Critical one-tail 1.66437 P(T<---t) two-tail 0.27523 t Critical two-tail 1.99045

Table 32. T-test for root shape. Student's two sample t-test for statistical differences in root morphology between samples with zero TNF-ct levels and samples positive for TNF-a. 122

Time since brushing" Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 6.71111 5.975 Variance 6.07374 8.84551 Observations 45 40 Hypothesized Mean Difference df 76 t Star 1.23353 P(T<--t) one-tail 0.11059 t Critical one-tail 1.66515 P(T<=t) two-tail 0.22118 t Critical two-tail 1.99168

Table 33. T-test for time since brushing. Student's two sample t-test for statistical differences in time since the patient last brushed their teeth between samples with zero TNF-c levels and samples positive for TNF-z. 123

Type of tooth movement: Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 1.2 1.525 Variance 1.11818 1.12756 Observations 45 40 Hypothesized Mean Difference df 82 t Stat -1.4112 P(T<=t) one-tail 0.08098 t Critical one-tail 1.66365 P(T<=t) two-tail 0.16197 t Critical two-tail 1.98932

Table 34. T-test for the type of tooth movement. Student's two sample t-test for statistical differences in the type of tooth movement between samples with zero TNF-c levels and samples positive for TNF-a. 124

Time in treatment: Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 12.4889 9.825, Variance 95.6192 56.0455 Observations 45 40 Hypothesized Mean Difference df 81 t Stat 1.41865 P(T<=t) one-tail 0.07992 t Critical one-tail 1.66388 P(T<=t) two-tail 0.15984 t Critical two-tail 1.98969

Table 35. T-test for the time in treatment. Student's two sample t-test for statistical differences in the time in treatment between samples with zero TNF-c levels and samples positive for TNF-c. 125

Plaque Index: Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 1.4 1.25 Variance 0.42727 0.60256 Observations 45 40 Hypothesized Mean Difference df 77 t Stat 0.95716 P(T<=t) one-tail 0.17074 t Critical one-tail 1.66488 P(T<=t) two-tail 0.34148 t Critical two-tail 1.99126

Table 36. T-test for the Plaque Index. Student's two sample t-test for statistical differences in the Plaque Index between samples with zero TNF-c levels and samples positive for TNF-c. 126

Gingival Index: Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 1.64444 0.85 Variance 0.32525 0.3359 Observations 45 40 Hypothesized Mean Difference df 82 t Stat 6.3555 P(T<--t) one-tail 5.5E-09 t Critical one-tail 1.66365 P(T<---t) two-tail 1.1E-08 * t Critical two-tail 1.98932

Table 37. T-test for the Gingival Index. Student's two sample t-test for statistical differences in the Gingival Index between samples with zero TNF-c levels and samples positive for TNF-c. 127

Periodontal Index- Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 1.04444 0.75 Variance 0.27071 0.19231 Observations 45 40 Hypothesized Mean Difference df 83 t Stat 2.83023 P(T<=t) one-tail 0.00292 t Critical one-tail 1.66342 P(T<=t) two-tail 0.00583 * t Critical two-tail 1.98896

Table 38. T-test for the Periodontal Index. Student's two sample t-test for statistical differences in the Periodontal Index between samples with zero TNF-ct levels and samples positive for TNF-c. 128

Root Loss (method 1)" Positive for TNF Zero TNF Two sample t-test, unequal variances :Mean 0.9222 0.4225 Variance 0.3886 0.2828 Observations 45 40 Hypothesized Mean Difference df 83 t Stat 3.98751 P(T<=t) one-tail 0.00007 t Critical one-tail 1.66342 P(T<=t) two-tail 0.00014 * t Critical two-tail 1.98896

Table 39. T-test for Root Loss (RL). Student's two sample t-test for statistical differences in the amount of root loss, as determined from method #1, between samples with zero TNF-t levels and samples positive for TNF-ct. 129

Root Loss* (method #2): Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 0.9238 0.4257 Variance 0.3929 0.2873 Observations 45 40 Hypothesized Mean Difference df 83 t Stat 3.948093 P(T<---t) one-tail 0.000082 t Critical one-tail 1.663420 P(T<=t) two-tail 0.000164 * t Critical two-tail 1.988960

Table 40. T-test for Corrected Root Loss (RL*). Student's two sample t-test for statistical differences in the amount of root loss, as determined from method #2, between samples with zero TNF-a levels and samples positive for TNF-c. 130

Root Resorption Index: Positive for TNF Zero TNF Two sample t-test, unequal variances Mean 1.1778 0.5 Variance 0.6040 0.4615 Observations 45 40 Hypothesized Mean Difference df 83 t Stat 4.28994 P(T<=t) one-tail 0.00002 t Critical one-tail 1.66342 P(T<=t) two-tail 0.00005 * t Critical two-tail 1.98896

Table 41. T-test for the Root Resorption Index Student's two sample t-test for statistical differences in the root resorption index, as determined from method #3, between samples with zero TNF-c levels and samples positive for TNF-a. 131

Samples with zero TNF omitted TNF TNF index Ln TNF Mean 437.05 2.35 5.53 Standard Error 77.5512 0.2128 0.1748 Median 320 2 5.7683 Mode 320 1 5.7683 Standard Deviation 508.5371 1.3953 1.1461 Sample Variance 258610.0 1.9468 1.3136 Kurtosis 13.6915 1.0011 0.0976 Skewness 3.1788 0.6031 -0.4550 Range 2931 4 5.4226 Minimum 13 1 2.5649 Maximum 2944 5 7.9875 Sum 18793 101 237.9680 Count 43 43 43 Confidence Level(95.000%) 151.997 0.417 0.353

Table 42. Distribution with samples containing zero TNF values omitted.

TNF Frequency Percentage Cumulative % 300 20 46.5 46.51% 600 14 32.6 79.07% 900 4 9.3 88.37% 1200 3 7.0 95.35% 1500 1 2 3 97.67% More 1 2.3 100.00%

Table 43. Distribution of TNF values with samples of zero omitted. For comparison, refer to table 28. 132

TNF Index Frequency Percentage Cumulative 17 39.5 39.53% 9 20.9 60.47% 6 14.0 74.42% 7 16.3 90.70% 4 9.3 100.00%

Table 44. Distribution of TNF values sorted according to the TNF index with zero values omitted. For comparison, refer to table 29.

Ln"TF Frequency Percentage Cumulative % 2.64 1 2.3 2.33% 3.53 2 4.7 6.98% 4.42 3 7.0 13.95% 5.31 11 25.6 39.53% 6.20 11 25.6 65.12% 7.10 13 30.2 95.35% More 2 4.7 100.00%

Table 45. Distribution of TNF data with no samples of zero when transformed using natural logarithm. For comparison, refer to table 30. 133

Tooth PLI GI PeI Brs Rt 0.00 1.00 PLI -0.02 -0.07 1.00 GI 0.00 -0.12 0.51 1.00 PeI -0.02 0.01 0.55 0.65 1.00 Brs 0.06 -0.07 0.08 0.13 0.04 1.00 MVMT 0,00 -0.20 0.05 0.00 0.13 -0.18 Fa 0.10 -0.13 0:31 0.10 0.34 -0.17 FR 0,01 -0.05 0.31 0.15 0.37 -0.40 Time 0.01 -0.01 0.22 0.18 0.43 -0,40 C/I 0,07 -0.03 0.32 0.12 0.36 -0.38 TxT 0.01 -0.04 0.-116 0.14 -0.03 0.62 TxL -0.01 -0.07 -0.09 -0.22 -0.24 -0.39 RRI 0.01 0.05 -0.04 0.24 0.29 0.10 RL 0.06 -0.05 0.01 0.29 0.26 0.12 RE,* 0.07 -0.05 0.01 0.29 0.26 0.13 TNF 0.07 -0.10 0.14 0.49 0.46 -0.09 lnTNF+I 0.05 -0.13 0.13 0.65 0.38 0.06 TNF-Ind 0.11 -0.10 0.13 0.63 0.46 -0.05 TNF 0.08 -0.12 0,14 0.67 0.42 0.01 ranked

Table 46. Correlation coefficients (tables 46-48). For legend, see table 48. 134

MVMT Fa FR time C/I TxT TxL MVMT 1.00 Fa 0.69 1.00 FR 0.60 0.80 1.00 Time 0.61 0.70 0.96 1.00 C/I 0.55 0.77 0.97 0:93 1.00 TxT -0.12 -0,08 -0.35 -0,38 -0.34 1.00 TxL 0.08 0.08 0,15 0.02 0.16 -0.68 1.00 RRI -0.08 -0.04 0.03 0.10 0.02 0.20 -0.36 RL -0.08 0.03 0.12 0. ! 1 0.12 0.10 -0.08 RL* -0.08 0.03 0.12 0.11 0.12 0.10 -0.09 TNF 0.22 0.17 0.17 0.20 0.16 0.02 -0.12 InTNF+I -0.06 -0.03 0.00 0.03 -0.02 0.10 -0.15 TNF-Index 0.09 0,12 0.15 0.19 0.14 0.03 -0.16 TNF ranked 0.00 0,03 0.08 0.11 0.05 0.08 -0.17

Table 47. Correlation coefficients (tables 46-48). 135

RRI' RL RL* TNF InTNF+I TNF TNF Index ranked RRI 1.00 RL 0.72 1.00 RL* 0.72 1.00 1.00 TNF 0,37 0.22 0,21 1.00 InTNF+I 0.44 0.38 0.37 0.69 1.00 TNF'Index 0.39 0.310 0.29 0.84 0.89 1.00 TNFranked 0,36 0,35 0.43 0.75 0.98 0.96 1

Table 48. Correlation coefficients (tables 46-48). Legend: Tooth-which teeth were included in the sample, Root- root morphology, PLI- Plaque Index, GI- Gingival Index, PeI- Periodontal Index, Brs- hours since the patient last brushed their teeth, MVMT- type of tooth movement, Fa- force applied to the tooth during the visit prior to the sampling visit, FR- force remaining at the time of sampling, Time- weeks since the last visit (the number of weeks that this force was applied), C/I- continuous or interrupted force, TxT- number of months in treatment, TxL- approximated time in treatment left, RRI- Root Resorption Index, RL- root loss, RL*- corrected root loss, TNF- raw TNF levels (picograms/site), lnTNF+ 1- raw TNF values plus one picogram tranformed using natural logarithm function, TNF Ind- TNF values categorized using the previously mentioned index, TNF ranked- ranked TNF values using the entire array of TNF data. 136

Tth Rt PLI GI PeI Brs Tth 1.00 Rt -0.08 1.00 PLI -0.05 -0.02 1.00 GI 0.02 -0.13 0.58 1.00 PeI 0.00 0.04 0.49 0.60 1.00 Brs 0.12 -0.09 -0.18 -0.15 -0.24 1.00 Mvmt -0:03 -0.22 0.26 0.51 0.44 -0.22 Fa 0i06 -0.16 0.31 0.37 0.48 -0.35 FR 0,00 -0.11 0.46 0.46 0.54 -0.47 Time 0,04 -0.10 0.41 0.49 0.63 -0.47 C/I 0.10 -0.09 0.48 0,42 0.53 -0.42 Tx T 0.03 -0.11 -0.15 -0.05 -0.31 0,65 Tx.L -0,13 0.10 0.04 -0.12 -0.11 -0.47 RL 0.08 -0.18 0.01 0.03 -0.01 0.21 RL* 0.08 -0.18 0.00 0.02 -0.01 0.22 RRI 0.11 -0.18 -0.13 0.16 0.16 0.11 TNF 0:09 -0.06 0.14 0.60 0.49 -0.29 Tnf-I 0.18 0.00 0.09 0.58 0.49 -0.39 InTNF+I 0.17 -0.09 0.15 0.68 0.46 -0.38 InTNF 0.17 -0.09 0.15 0.68 0.46 -0.38 TNF rank 0,19 -0.05 0.13 0.65 0.48 -0.41

Table 49. Correlation coefficients, omitting samples with TNF values of Zero (tables 49-51). For legend see table 51. 137

Mvmt Fa Fn Time C/I Tx T Tx"L Fa 0.66 1.00 FR 0.65 0.85 1.00 Time 0.65 0.77 0.96 1.00 C/I 0,55 0.79 0.95 0.91 1.00 Tx T -0.11 -0.22 -0.33 -0.35 -0.32 1.00 Tx L 0.05 0,11 0.10 -0,01 0.10 -0.72 1.00 RL 0.12 0.15 0.20 0.17 0.23 0.11 0.03 RL* 0.10 0.15 0.20 0.17 0.23 0.12 0.02 RRI 0.22 0.18 0.21 0.27 0.22 0.18 -0.18 TNF 0.49 0,37 0.32 0.37 0.32 -0.07 -0.08 Tnf-I 0.47 0.46 0.42 :0.49 0.43 -0.12 -0.14 InTNF+I 0.44 0,37 0,34 0.41 0.32 -0.16 -0.07 InTNF 0.44 0,37 0.34 0:41 0.32 -0.16 -0,07 TNF rank. 0,46 0.45 0.42 0,49 0.41 -0.12 -0.14

Table 50. Correlation coefficients, omitting samples with TNF values of Zero (tables 49-51). 138

RL RL* RRI TNF Tnf-I InTNF+I InTNF TNF rank RL 1.00 RL* 1.00 1.00 RRI 0.74 0.74 1.00 TNF 0,03 0.02 0.32 1.00 Tnf-I 0.01 0,01 0.28 0,80 1.00 InTNF+I 0,01 0.00 0.28 0.78 0.88 1.00 InTNF 0,01 0,00 0,28: 0.78 0.88 1.00 1.00 TNF rank 0.02 0,01 0.29 0.77 0.95 0.97 0.97 1.00

Table 51. Correlation coefficients, omitting samples with TNF values of Zero (tables 49-51). Legend" Tooth-which teeth were included in the sample, Root- root morphology, PLI- Plaque Index, GI- Gingival Index, PeI- Periodontal Index, Brs- hours since the patient last brushed their teeth, MVMT- type of tooth movement, Fa- force applied to the tooth during the visit prior to the sampling visit, FR- force remaining at the time of sampling, Time- weeks since the last visit (the number of weeks that this force was applied), C/I- continuous or interrupted force, TxT- number of months in treatment, TxL- approximated time in treatment left, RRI- Root Resorption Index, RL- root loss, RL*- corrected root loss, TNF- raw TNF levels (picograms/site), lnTNF+I- raw TNF values plus one picogram tranformed using natural logarithm function, TNF Ind- TNF values categorized using the previously mentioned index, TNF ranked- ranked TNF values using the entire array of TNF data. 139

Patient Coefficient Patient Coefficient TS 0.9115 DSM 0.3004 GM 0.8974 MM 0.1237 RP 0.8432 JA 0.0099 LR 0.7358 YP 0.0000 DS 0.7177 JO 0.0000 SM 0.4855 BG -0.0411 BS 0.4460 LU -0.0793 DM 0.4155 KM -0.3042 SG 0.3678 KZ -0.3674 HJ -0.5521

Table 52. Correlation coefficients between Root Loss and raw TNF scores were calculated for each patient (n=19). 140

Tooth RT PLI GI PeI Brs Mean 8.51 0.27 1.33 1.27 0.91 6.36 Standard Error 0.227 0.068 0.077 0.076 0.055 0.296 Median 9.00 0.00 1.00 1.00 1.00 6.00 Mode 8.00 0.00 2.00 1.00 1.00 6.00 Standard Deviation 2.091 0.625 0.714 0.697 0.503 2.725 Sample Variance 4.372 0.390 0.509 0.485 0.253 7.425 Kurtosis 1.439 8.053 -0.642 -0.865 0.914 -0.705 Skewness -0.149 2.737 -0.379 -0.425 -0.185 0.209 Range 12 3 3 2 2 9 Minimum 2 0 0 0 0 2 Maximum 14 3 3 2 2 11 Sum 723 23 113 108 77 541 Count 85 85 85 85 85 85 Confidence 0.451 0.135 0.154 0.150 0.108 0.588 Level(95.0%)

Table 53. Distribution of clinical variables (tables 53-55).

MVMT Fa FR time C/I TxT TxL Mean 1.35 1.41 0.74 3.51 0.73 11.24 13.98 Standard Error 0.116 0.084 0.048 0.236 0.048 0.957 0.947 Median 1.00 2.00 1.00 4.00 1.00 9.00 19.00 Mode 1.00 2.00 1.00 4.00 1.00 12.00 19.00 Standard Deviation 1.066 0.776 0.441 2.175 0.447 8.826 8.733 Sample Variance 1.136 0.602 0.194 4.729 0.200 77.896 76.261 Kurtosis -0.964 -0.778 -0.761 -0.914 -0.917 1.655 -1.261 Skewness 0.576 -0.872 -1.121 -0.833 -1.051 1.570 -0.573 Range 3 2 1 6 1 34 25 Minimum 0 0 0 0 0 2 0 Maximum 3 2 1 6 1 36 25 Sum 115 120 63 298 62 955 1188 Count 85 85 85 85 85 85 85 Confidence 0.230 0.167 0.095 0.469 0.096 1.904 1.884 Level(95.0%)

Table 54. Distribution of clinical variables (tables 53-55). 141

RL RL* RRI TNF InTNF+I TNF-Index TNF ranked Mean 0.69 0.69 0.86 232.86 2.95 1.27 44.059 Standard 0.068 0.069 0.087 45.8 0.316 0.170 2.398 Error Median 0.60 0.60 1.00 29.00 3.40 1.00 43 Mode 1.10 0.00 0.00 0.00 0.00 0.00 23 Standard 0.631 0.634 0.804 422.4 2.912 1.569 22.109 Deviation Sample 0.398 0.402 0.646 178405 8.480 2.462 488.806 Variance Kurtosis 2.506 2.555 -0.932 20.11 -1.809 -0.183 -1.418 Skewness 0.107 0.153 0.404 3.77 0.093 1.034 0.390 Range 5 5 3 2944 8 5 62 Minimum -2 -2 0 0 0 0 23 Maximum 3 3 3 2944 8 5 85 Sum 58 59 73 19793 251 108 3745 Count 85 85 85 85 85 85 85 Confidence 0.136 0.137 0.173 91 1 0 4.700 Level(95.0%)

Table 55. Distribution of root loss measures and TNF values (tables 53-55). 142

Regression Statistics Multiple R 0.6083 R Square 0.3701 Adjusted R Square 0.3467 Standard Error 341.3873 Observations 85

ANOVA df SS MS F Significance F Regression 3 5545892.325 1848630.775 15.8619 3.34426E-08 Residual 81 9440167.981 116545.284 Total 84 14986060.31

Coefficients Standard t Stat P-value Lower 95% Upper Error 95% Intercept -355.442 95.030 -3.7403 0.00034 -544.522 -166.362 GI 258.657 55.066 4.6972 0.00001 149.093 368.221 MVMT 96.452 35.068 2.7504 0.00734 26.677 166.226 RR 150.392 47.870 3.1417 0.00235 55.146 245.638

Table 56. Regression model" Y= TNF, X= GI + MVMT + RRI Backwards stepwise multivariable regression with raw TNF-ct values as the dependent variable and the Gingival Index, type of tooth movement, and the Root Resorption Index as the independent variables. 143

Regression Statistics Multiple R 0.5579 R Square 0.3112 Adjusted R Square 0.2944 Standard Error 354.79 Observations 85

ANOVA df SS MS F Significance F Regression 2 4664255 2332128 18.52723 2.3E-07 Residual 82 10321805 125875.7 Total 84 14986060

Coefficients Standard t Stat P-value Lower 95% Upper 95% Error Intercept -219.11 84.26 -2.60 0.01104 -386.73 -51.48 GI 261.36 57.22 4.57 0.00002 147.53 375.18 RR 139.59 49.58 2.82 0.00610 40.96 238.23

Table 57. Regression model: Y= TNF, X GI + RRI Backwards stepwise multivariable regression with raw TNF-c values as the dependent variable and the Gingival Index and the Root Resorption Index as the independent variables. 144

Regres'io'n Statistics Multiple R 0.5429 R Square 0.2947 Adjusted R Square 0.2775 Standard Error 449.20 Observations 43

ANOVA df SS MS F Significance F Regression 1 3457458 3457458 17.13499 0.000169 Residual 41 8272885 2O1777.7 Total 42 11730344

Coefficient Standard t Stat P-value Lower Upper s Error 95% 95% Intercept -218.267 143.7973 -1.51788 0.136719 -508.672 72.13784 GI 409.1723 98.84719 4.139443 0.000169 209.5462 6O8.7984

Table 58. Regression model" Y= TNF, X= GI Backwards stepwise multivariable regression with raw TNF-c values as the dependent variable and the Gingival Index as the independent variable. 145

Regression Statistics Multiple R 0.3688 R Square 0.1360 Adjusted R Square 0.1256 Standard Error 394.97 Observations 85

ANOVA SS MS F Significance F Regression 1 2038027 2038027 13.06424 0.000515 Residual 83 12948034 156000.4 Total 84 14986060

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 66.4848 62.8815 1.0573 0.2934 -58.5840 191.5536 RR 193.7231 53.5969 3.6144 0.0005 87.1211 300.3252

Table 59. Regression model" Y= TNF, X RRI Backwards stepwise multivariable regression with raw TNF-(z values as the dependent variable and the Root Resorption Index as the independent variable. 146

Regression Statistics Multiple R 0.7472 R Square 0.5582 Adjusted R Square 0.5419 Standard Error 14.9643 Observations 85

ANOVA df SS MS F Significance F Regression 3 22921.45 7640.483 34.1201 2.33E-14 Residual 81 18138.26 223.9291 Total 84 41059.71

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 18.1524 4.0852 4.4435 0.0000 10.0242 26.2807 PLI -6.8085 2.7192 -2.5039 0.0143 -12.2189 -1.3981 GI 22.8988 2.8657 7.9908 0.0000 17 1970 28.6005 RR 6.8266 2.1335 3.1997 0.0020 2.5816 11.0716

Table 60. Regression model" Y= ranked TNF, X PLI + GI + RRI Backwards stepwise multivariable regression with ranked TNF-ot values as the dependent variable and the Plaque Index, Gingival Index, and Root Resorption Index as the independent variables. 147

Regression Statistics Multiple R 0.7239 R Square 0.,241 Adjusted R Square 0.5124 Standard Error 15.4376 Observations 85

ANOVA SS MS F Significance F Regression 21517.58 10758.79 45.14455 6.02E-14 Residual 82 19542.13 238.3186 Total 84 41059.71

Coefficient Standard t Stat P-value Lower 95% Upper 95% Error Intercept 13.1083 3.6663 3.5753 0.0006 5.8148 20.4018 GI 19.0298 2.4897 7.6434 0.0000 14.0770 23.9826 RR 7.8845 2.1574 3.6547 0.0005 3.5928 12.1763

Table 61. Regression model" Y= ranked TNF, X GI + RRI Backwards stepwise multivariable regression with ranked TNF-c values as the dependent variable and the Gingival Index and the Root Resorption Index as the independent variables. 148

Regression Statistics Multiple R 0.6682 R Square 0.4465 Adjusted R Square 0.4399 Standard Error 16.547 Observations 85

ANOVA df SS MS F Significance F Regression 1 18334.46 18334.46 66.96341 2.81E-12 Residual 83 22725.25 273.7982 Total 84 41059.71

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 17.1166 3.7498 4.5646 0.0000 9.6583 24.5748 GI 21.2046 2.5913 8.1831 0.0000 16.0507 26.3585

Table 62. Regression model: Y= ranked TNF, X GI Backwards stepwise multivariable regression with ranked TNF-c values as the dependent variable and the Gingival Index as the independent variable. 149

Regression'Statistics Multiple R 0.4511 R Square 0.2035 Adjusted R Square 0.1841 Standard Error 19.997 Observations 43

ANOVA SS MS F Significance F Regression 1 4189.509 4189.509 10.47657 0.002395 Residual 41 16395.62 399.8931 Total 42 20585.13

Coefficient Standard t Stat P-value Lower 95% Upper 95' s Error Intercept 35.6748 4.8561 7.3464 0.0000 25.8677 45.4819 RR 12.2321 3.7791 3.2368 0.0024 4.6000 19.8643

Table 63. Regression model: Y= ranked TNF, X RRI Backwards stepwise multivariable regression with ranked TNF-ot values as the dependent variable and the Root Resorption Index as the independent variable. 150

Regression Statistics Multiple R 0.7363 R Square 0.5421 Adjusted R Square 0.5251 Standard Error 2.0068 Observations 85

ANOVA SS MS F Significance F Regression 3 386.1534 128.7178 31.96169 9.86E-14 Residual 81 326.2O75 4.027253 Total 84 712.3609

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept -0.4185 0.5478 -0.7638 0.4472 -1.5085 0.6716 PLI -0.8709 0.3647 -2.3881 0.0193 -1.5964 -0.1453 GI 2.9122 0.3843 7.5779 0.0000 2.1476 3.6769 RR 0.9611 0.2861 3.3592 0:0012 0.3918 1.5304

Table 64. Regression model: Y= In (TNF+I), X= PLI + GI + RRI Backwards stepwise multivariable regression using transformed TNF-c values (natural logarithm) as the dependent variable and the Plaque Index, the Gingival Index, and the Root Resorption Index as the independent variables. 151

R'egression' Statistics Multiple R 0.7140 R Square 0.5098 Adjusted R Square O.4979 Standard Error 2.0635 Observations 85

ANOVA df SS MS F Significance F Regression 2 363.1854 181.5927 42.64503 2.01E-13 Residual 82 349.1755 4.258238 Total 84 712.3609

Coefficient Standard t Stat P-value Lower 95% Upper 95% Error Intercept -1.0636 0.4901 -2.1703 0.0329 -2.0386 -0.0887 GI 2.4174 0.3328 7.2637 0.0000 1.7553 3.0794 RR 1 .O964 0.2884 3.8020 0.0003 0.5228 1.6701

Table 65. Regression model: Y= In (TNF+I), X= GI + RRI Backwards stepwise multivariable regression using transformed TNF-c values (natural logarithm) as the dependent variable and the Gingival Index and the Root Resorption Index as the independent variables. 152

Regresio'n Statistics Multiple R 0.6507 R Square 0.4234 Adjusted R Square 0.4165 Standard Error 2.2245 Observations 85

ANOVA SS F Significance F Regression 1 301.6301 301.6301 60.95306 1.57E-11 Residual 83 410.7308 4.948564 Total 84 712.3609

Coefficient Standa'd t Siat' P-value Lower 95% Upper 95" s Error Intercept -0.5062 0.5041 -1.0042 0.3182 -1.5089 0.4964 GI 2.7198 0.3484 7.8072 0.0000 2.0269 3.4127

Table 66. Regression model" Y= In (TNF+I), X= GI Backwards stepwise multivariable regression using transformed TNF-a values (natural logarithm) as the dependent variable and the Gingival Index as the independent variable. 153

Regression' Statistics Multiple R O.44O96 R Square 0.19445 Adjusted R Square 0.18474 Standard Error 2.62941 Observations

ANOVA df 'SS MS F Significance F Regression 1 138.5161 138.5161 20.03475 2.4E-05 Residual 83 573.8448 6.913792 Total 84 712.3609

Coefficient Standa'rd t Stat P-value Lower 95% UPper 95% s Error Intercept 1.57785 0.41862 3.76919 0.00031 0.74524 2.41O47 RR 1.59708 0.35681 4.47602 0.00002 0.88740 2.30676

Table 67. Regression model" Y= In (TNF+I), X= RRI Backwards stepwise multivariable regression using transformed TNF-ot values (natural logarithm) as the dependent variable and the Root Resorption Index as the independent variable. 154

Regression Statistics Multiple R 0.6984 R Square 0.4877 Adjusted R Square 0.4687 Standard Error 1.1436 Observations 85

ANOVA df SS MS F Significance F Regression 3 100.8464 33.61548 25.70426 8.81E-12 Residual 81 105.93 1.307778 Total 84 206.7765

Coefficient Standard t Stat P-value Lower 95% Upper 95 s Error Intercept -0.4386 0.3122 -1.4048 0.1639 -1.0597 0.1826 PLI -0.4622 0.2078 -2.2244 0.0289 -0.8757 -0.0488 GI 1.5432 0.2190 7.0469 0.0000 1.1075 1.9790 RR 0.4225 0.1630 2.5913 0.0113 0.0981 0.7469

Table 68. Regression model: Y= TNF Index, X PLI + GI + RRI Backwards stepwise multivariable regression using the TNF Index as the dependent variable and the Plaque Index, the Gingival Index, and the Root Resorption Index as the independent variables. 155

Regression Statistics Multiple R 0.6025 R Square 0.3630 Adjusted R Square 0.3475 Standard Error 410.80 Observations 43

ANOVA SS MS F Significance F Regression 1 3942702 3942702 23.3636 1.91E-05 Residual 41 6918918 168754.1 Total 42 10861620

Coefficients Standard t Stat P-value Lower 9'5% Upper 95% Error Intercept -359.048 176.212 -2.038 0.0481 -714.916 -3.180 G! 475.445 98.363 4.834 0.0000 276.798 674.093

Table 69. Regression model: Y=TNF-0's, X=GI Backwards stepwise multivariable regression using TNF-c values, where samples with values of zero were omitted, as the dependent variable and the Gingival Index as the independent variable. 156

Regression Statistics Multiple R 0.3190 R Square t).1018 Adjusted R Square 0.0799 Standard Error 487.81 Observations 43

ANOVA df SS MS F Significance F Regression 1 1105391 1105391 4.645345 0.03706 Residual 41 9756229 237956.8 Total 42 1O861620

Coefficient Standard t Stat P-value Lower Upper s Error 95% 95% Intercept 186.77 137.91 1.354 0.1830 (91.74) 465.28 RR 215.24 99.86 2.155 0.0371 13.56 416.92

Table 70. Regression model" Y=TNF-0's, X=RRI Backwards stepwise multivariable regression using TNF-c values, where samples with values of zero were omitted, as the dependent vail'able and the Root Resorption Index as the independent variables. 157

Regression Statistics Multiple R 0.7599 R Square 0.5775 Adjusted R Square 0.5450 Standard Error 8.4488 Observations 43

ANOVA SS MS F Significance F Regression 3 3804.723 1268.241 17.76691 1.99E-07 Residual 39 2783.905 71.38218 Total 42 6588.628

Coefficient stan'dard t Stat P-value Lower 95% Upper 95% s Error Intercept 10.7394 5.5237 1.9442 0.0591 -0.4333 21.9121 GI 10.9822 2.0794 5.2814 0.0000 6.7762 15.1882 Brs -1.7637 0.5391 -3.2714 0.002:2 -2.8541 -0.6732 RR 3.8779 1.7686 2.1926 0'.0344 0.3005 7.4553

Table 71. Regression model" Y= Ranked TNF-0'S, X=GI + BRS + RR_I Backwards stepwise multivariable regression using ranked TNF-c values, where samples with values of zero were omitted, as the dependent variable and the Gingival Index, hours since the patient last brushed their teeth, and the Root Resorption Index as the independent variables. 158

Regression Statistics Multiple R 0.6540 R Square 0.4277 Adjusted R Square 0.4137 Standard Error 9.5900 Observations 43

ANOVA df SS MS F Significan"ce F Regression 1 2817.925 2817.925 30.64016 1.98E-06 Residual 41 3770.703 91.96836 Total 42 6588.628

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 0.6240 4.1137 0.1517 0.8802 -7.6837 8.9317 GI 12.7107 2.2963 5.5354 0.0000 8.0733 17.3481

Table 72. Regression model" Y= Ranked TNF-0's, X=GI Backwards stepwise multivariable regression using ranked TNF-c values, where samples with values of zero were omitted, as the dependent variable and the Gingival Index as the independent variable. 159

Regression Statistics Multiple IR O.4059 R Square 0.1647 Adjusted R Square 0.1444 Standard Error 11.5855 Observations 43

ANOVA df SS MS F Significance F Regression 1 1085.434 1085.434 8.086722 0.006924 Residual 41 5503.194 134.2242 Total 42 6588.628

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 35.6087 5.1320 6.9386 0.0000 25.2445 45.9730 Brs -2.0601 0.7244 -2.8437 0.0069 -3.5231 -0.5971

Table 73. Regression model: Y= Ranked TNF-0's, X=BRS Backwards stepwise multivariable regression using ranked TNF-a values, where samples with values of zero were omitted, as the dependent variable and the hours since the patient last brushed their teeth as the independent variable. 160

Regression Statistics Multiple R 0.2866 R Square 0.0822 Adjusted R Square 0.0598 Standard Error 12.145 Observations 43

ANOVA df SS MS F Significance F Regression 1 541.3384 541.3384 3.670219 0.062384 Residual 41 6047.289 147.4949 Total 42 6588.628

Coefficient Standard t Stat P-value Lower 95% Upper 95% s Error Intercept 16.3684 3.4334 4.7674 0.0000 9.4346 23.3023 RR 4.7632 2.4863 1.9158 0.0624 -0.2580 9.7843

Table 74. Regression model" Y= Ranked TNF-0's, X=RRI Backwards stepwise multivariable regression using ranked TNF-(x values, where samples with values of zero were omitted, as the dependent variable and the Root Resorption Index as the independent variable. APPENDIX III. CONSENT FORMS

INFORMED CONSENT FOR ADULTS

Title: Levels of cytokines present in the gingival sulcus of orthodontically treated teeth

Investigators" EF Rossomondo DDS, PhD, Louis Norton DMD, Mike McKay DDS, and

Robert Marzban DDS

I understand that I have been asked to participate in a research study conducted by Dr.

Rossomando and colleagues of the University of Connecticut Health Center. The purpose of this study is to develop a device for the diagnosis of root resorption. Although

I will receive no direct benefit from this research, this work may lead to the development of better diagnostic tools for orthodontics.

My participation in this research will require the collection of fluid from between my gums and teeth. The fluid will be collected from around the gums by putting a small amount of very small protein-coated particles in the space between the gums and teeth.

The particles have an iron center so that they can be taken out with a small magnet. I understand that these procedures will take about five minutes and I might be asked to allow the researcher to take repeated monthly samples at my convenience. At that time the doctor will examine me for any signs of trouble from the first procedure. If the doctor finds any problem, the second procedure will not be done. These procedures will not require any anesthesia and have no known risks. In some cases, only about one-half of the beads are recovered. The other half do not remain and are washed out of the mouth.

These studies have been in progress for more than five years and no harmful effects have

161 162 been seen. Nonetheless, unexpected trouble from side effects, namely an irritaion of the gums, is a possibility. I understand that I will receive a complete examination of the mouth for signs of oral disease and if any problems are found I will be informed and directed to appropriate individuals who advise me on treatment.

Any reports or presentations about this research will not idemify me by name. Best efforts will be made to maintain confidentiality, although the Federal Food and Drug

Administration has the fight to examine research records involved in drug or diagnostic device development and testing. I understand that this study has not been approved by the Federal Food and Drug Administration.

I understand that I may withdraw from this study at any time without prejudice to care at the University of Connecticut Health Center.

The University of Connecticut Health Center (UCHC) does not provide insurance coverage to compensate me if I am injured during this research. However, I may still be eligible for compensation. If I am injured, I can file a claim against the State of

Connecticut seeking compensation. For a description of this process, or available compensation options, I may contact Rose Marie Howes, Executive Secretary,

Institutional Review Board (RIB) at the UCHC, at 679- 142.

The UCHC does not offer free care. However, treatment for a research related injury may obtained at UCHC for the usual fee.

If I have any questions about this study, of my fights as a research subject, I may call

Dr. Edward Rossomando (679-2622) at the University of Connecticut Health Center. 163

I, the undersigned, have understood the above explanation, give consent to my voluntary participation in this research project, and have received a copy of this consent

Signature of the investigator Date

Signature of the subject Date Witness Date 164

INFORMED CONSENT FOR DEPENDENTS

Title" Levels of cytokines present in the gingival sulcus of orthodontically treated teeth

Investigators: EF Rossomando DDS PhD, DDS, Louis Norton DMD, Mike McKay DDS,

and RobertMarzban DDS

I understand that I have been asked to allow my child to participate in a research study conducted by Dr. Rossomando and his colleagues of potential problems with orthodontic treatment. Although my child will receive no direct benefit from this research, this work may lead to the development of better diagnostic tools for orthodontics.

Participation in this research will require the collection of fluid from between gums and teeth. The fluid will be collected from around the gums by putting a small amount of very small protein-coated particles in the space between the gums and teeth. The particles have an iron center so that they can be taken out with a small magnet. I understand that these procedures will take about 5 minutes. These procedures will not require any anesthesia and have no known risks since this technique has been used in over 200 patients. In some cases only about one half of these beads are recovered. The other half do not remain and are washed out. These studies have been in progress for more than four years and we assume no harmful effects have ever been seen.

Any reports or presentations about this research will not identify my dependent by name. Best efforts will be made to maintain confidentiality, although the Federal Food and Drug Administration has the fight to examine research records involved in drug or diagnostic device development and testing. I understand that this study has not been approved by the Federal Food and Drug Administration. 165

I understand that my child may withdraw from this study at any time without prejudice to care at the University of Connecticut Health Center.

The University of Connecticut Health Center (UCHC) does not provide insurance coverage for compensation in the highly unlikely event of injury during this research.

However, my child may still be eligible for compensation. If my child is injured, I can file a claim against the State of Connecticut seeking compensation. For a description of this process, or available compensation options, I may contact Rose Marie Howes,

Executive Secretary, Institutional Review Board (IRB) at the UCHC, at 679-2142.

The UCHC does not offer free care. However, treatment for a research related injury may obtained at UCHC for the usual fee.

If I need additional information, I may contact Rose Marie Howes, Executive

Secretary of the IRB at 679-2142. The Executive Secretary can review the matter with me, identify the resources that may be available, and provide me with further inforation as to how to proceed.

If I have any questions about this study, or my child's fights as a research subject, I may call Dr. Rossomando at 679-2664 at the University of Connecticut Health Center.

I, the undersigned, have understood the above expl[nation, give consent to my child's voluntary participation in this research project, and have received a copy of this consent form.

Investigator Date

Guardian Date Witness Date 166

If the subject is a minor:

My researcher, Dr. Marzban or Dr. Norton, and parents have talked to me about being a part of a study that will collect fluid from around.my teeth. They are interested to know how much of a particle (called TNF) is in the fluid around my teeth. How much TNF they find may be a reason why teeth shorten while being moved. I understand the reason for the study and why I am being asked to take part in it. I have been told that as a subject in the study I will have the following performed: cotton rolls will be placed between my tooth and cheek and also between my tooth and tongue; my tooth will be dried with air; little particles will be placed near my tooth and gum with a long thin straw; after about a minute or two, the particles will be collected with a magnet that is smaller than a pencil. I understand that these procedures will not hurt. I know that I can ask questions about this study at any time. I also know that I can decide not to be in this study, or after entering the study I can decide that I want to be taken out of it.

Whatever I decide to do, I know my researcher will not be angry with me and that my dentist will continue to treat me as his patient.

Subject Date Witness Date

Investigator Date BIBLIOGRAPHY

Ottolengui R. The physiological and pathological resorption of tooth roots. Items of Interest 1914; 36:332-62.

2. Ketcham AH. A preliminary report of an investigation of apical resorption of permanent teeth. Int. J. Orthod. 1927; 13:97-127.

Ketcham AH. A progress report of an investigation of apical root resorption of vital permanent teeth. Int. J. Orthod. 1929; 15:310-28.

Rudolph CE. A comparative study in root resorption in permanent teeth. J. Am. Dent. Assoc. 1936; 23:822-6.

So Rudolph CE. An evaluation of root resorption occurring during orthodontic treatment. J. Dent. Res. 1940; 19:367-71.

Becks H. Orthodontic prognosis: evaluation of routine dentomedical examination to determine "good and poor risks". Am J Orthod 1939; 25"610-24.

Stuteville O. Injuries to the teeth and supporting structures caused by various orthodontic appliances, and methods of preventing these injuries. J Am Dent Assoc 1937; 14: 1494-507.

8 Marshall J. A comparison of resorption of roots of deciduous teeth with the absorption of roots of the permanent teeth occurring as a result of infection. Int J Orthod 1929; 15:417.

Massler M, Malone AJ. Root resorption in human permanent teeth. Am J Orthod 1954; 40:619-33.

10. Phillips J. Apical root resorption under orthodontic .therapy. Angle Orthod 1955; 25:1-12.

11. Dougherty HL. The effect of mechanical forces upon the mandibular buccal segments during orthodontic treatment. Am J Orthod 1968; 54:83-103.

12. DeShields RW. A study of root resorption in treated Class II, Division I malocclusions. Angle Orthod 1969; 39:231-45.

13. Newman WG. Possible etiologic factors in external root resorption. Am.J Orthod 1975; 67:522-39.

167 168

14. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment: Part 2. Literature review [see comments]. Am J Orthod Dentofacial Orthop 1993; 103" 138- 46.

15. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment: Part 1. Literature review [see comments]. Am J Orthod Dentofacial Orthop 1993; 103"62-6.

16. Andreasen J. Review of root resorption systems and models. In" Davidovitch Z, ed. The biological mechanisms of tooth eruption and root resorption, 1988"9-22.

17. Andreason J. Summary of root resorption. In: Davidovitch Z, ed. The biological mechanisms of tooth eruption and root resorption, 1988"399-401.

18. Tronstad L. Root resorption--etiology, terminology and clinical manifestations. Endod Dent Traumatol 1988; 4:241-52.

19. Henry J, Weinman J. The pattem of resorption and repair of human cementum. J Am Dent Assoc 1951; 42"270-90.

20. Harris EF, Robinson QC, Woods MA. An analysis of causes of apical root resorption in patients not treated orthodontically. Quintessence Int 1993; 24:417-28.

21. Reitan K. Tissue behavior during orthodontic tooth movement. American Jouma! of Orthodontics 1960; 46"881-900.

22. Reitan K. Effects of force magnitude and direction of tooth movement on different alveolar bone types. Angle Orthodontist 1964; 34:244-55.

23. Reitan K. Clinical and histologic observations on tooth movement during and after orthodontic treatment. American Journal of Orthodontics 1967; 53"721-45.

24. Reitan K. Biomechanical principles and reactions. In: Graber T, ed. Current orthodontic concepts and techniques. Philadelphia: 'WB Saunders, 1969:56-159.

25. Reitan K. Initial tissue behavior during apical root resorption. Angle Orthodontist 1974; 44"68-82.

26. Reitan K. Biomechanical principles and reactions. In: Graber T, Swain B, eds. Orthodontics current principles and techniques. St. Louis" CV Mosby, 1985"101-92.

27. Sharer W, Hine M, Levi B. A textbook of oral pathology. Philadelphia: WB Saunders, 1983"328-32.

28. Harry MR, Sims MR. Root resorption in bicuspid immsion. A scanning electron microscope study. Angle Orthod 1982; 52"235-58. 169

29. Stenvik A, Mjor IA. Pulp and dentine reactions to experimental tooth intrusion. A histologic study of the initial changes. Am J Orthod 1970; 57"370-85.

30. Kennedy DB, Joondeph DR, Osterberg SK, Little RM. The effect of extraction and orthodontic treatment on dentoalveolar support. Am J Orthod 1983; 84:183-90.

31. Massler M, Perrealt JG. Root resorption in permanent teeth of young adults. J Dent Child 1954; 21"158-64.

32. Plets J, Isaacson J, Speidel T, Worms F. Maxillary central incisor toot length in orthodontically treated and untreated patients. Angle Orthodontist 1974; 44:43-7.

33. Goldson L, Henrikson CO. Root resorption during Begg treatment; a longitudinal roentgenologic study.-. J Orthod 1975; 68"55-66.

34. Dermaut LR, De Munck A. Apical root resorption of upper incisors caused by intrusive tooth movement: a radiographic study. Am.J Orthod Dentofacial Orthop 1986; 90:321-6.

35. Linge BO, Linge L. Apical root resorption in upper anterior teeth. Eur J Orthod 1983; 5"173-,83.

36. Linge L, Linge BO. Patient characteristics and treatment variables associated with apical root resorption during orthodontic treatment. Am J Orthod Dentofac Orthop 1991; 99:35-43.

37. Mirabella AD, Artun J. Risk factors for apical root resorption of maxillary anterior teeth in adult orthodontic patients. Am J Orthod Dentofacial Orthop 1995; 108:48-55.

38. Mirabella AD, Artun J. Prevalence and severity of apical root resorption of maxillary anterior teeth in adult orthodontic patients. Eur J Orthod 1995; 17"93-9.

39. Michalowicz BS, Aeppli DP, Kuba RK, et al. A twin study of genetic variation in proportional radiographic alveolar bone height. J Dent Res 1991; 70" 1431-5.

40. Harris EF, Kineret SE, Tolley EA. A heritable component for external apical root resorption in patients treated orthodontically. Am J Orthod Dentofacial Orthop 1997; 111"301-9.

41. Kinsella P. Some aspects of root resorption in orthodontics. NZ Orthod J 1971:21-5.

42. Goldin B. Labial root torque: effect on the maxilla and incisor root apex. Am J Orthod Dentofacial Orthop 1989; 95:208-19. 170

43. Davidovitch Z, Shanfeld JL, Montgomery PC, et al. Biochemical mediators of the effects of mechanical forces and electric currents on mineralized tissues. CalcifTissue Int 1984; 36:$86-97.

44. Davidovitch Z, Nicolay OF, Ngan PW, Shanfeld JL. Neurotransmitters, cytokines, and the control of alveolar bone remodeling in orthodontics. Dent Clin North Am 1988; 32:411-35.

45. Davidovitch Z, Okamoto Y, Gogen H, et al. Orthodontic forces stimulate alveolar bone marrow cells. In: Davidovitch Z, Norton L, eds. Biological mechanisms of tooth movement and craniofacial adaptation, 1996.

46. Kvam E. Scanning electron microscopy of human premolars following experimental tooth movement. Trans Eur Orthod Soc 1972" 1-11.

47. Kvam E. Scanning electron microscopy of tissue changes on the pressure surface of human premolars following tooth movement. Scand J Dent Res 1972; 80:357-68.

48. Kvam E. Scanning electron microscopy of organic structures on the root surface of human teeth. Stand J Dent Res 1972; 80"297-306.

49. Rygh P. Ultrastructural changes in pressure zones of human periodontium incident to orthodontic tooth movement. Acta Odont Scan 1973; 31"109-122.

50. Rygh P. Orthodontic root resorption studied by electron microscopy. Angle Orthod 1977; 47" 1-16.

51. Stenvik A, Mjor IA. The effect of experimental tooth intrusion on pulp and dentine. Oral Surg Oral Med Oral Pathol 1971; 32:639-48.

52. Brudvik P, Rygh P. Non-clast cells start orthodontic root resorption in the periphery ofhyalinized zones. Eur J Orthod 1993; 15"467-80.

53. Brudvik P, Rygh P. The initial phase of orthodontic root resorption incident to local compression of the periodontal ligament. Eur J Orthod 1993; 15"249-63.

54. Brudvik P, Rygh P. Multi-nucleated cells remove the main hyalinized tissue and start resorption of adjacent root surfaces. Eur J Orthod 1994; 16:265-73.

55. Brudvik P, Rygh P. Root resorption beneath the main hyalinized zone. Eur J Orthod 1994; 16:249-63.

56. Brudvik P, Rygh P. The repair of orthodontic root resorption: an ultrastructural study. Eur J Orthod 1995; 17" 189-98. 171

57. Brudvik P, Rygh P. Transition and determinants of orthodontic root resorption-repair sequence. Eur J Orthod 1995; 17:177-88.

58. Marks SC. The origin of osteoclasts" evidence, clinical implications, and investigative challenge of an extra-skeletal source. J Pathol 1983; 12:226-56.

59. Raisz LG. Mechanisms and regulation of bone resorption by osteoclastic cells. In: Coe FL, Favus MJ, eds. Disorders of bone and mineral metabolism, 1992:287-311.

60. Jones S, Boyde A. The resorption of dentine and cementum. In" Davidovitch Z, ed. Biological mechanisms of tooth eruption and root resorption, 1988:335-54.

61. Aukhil I, Greco G, Suggs C, Tomey D. Root resorption potentials of granulation tissue from bone and flap connective tissue. J Periodontal Res 1986; 21 "531-42.

62. Aukhil I, Simpson DM, Suggs C, Pettersson E. In vivo differentiation of progenitor cells of the periodontal ligament. An experimental study using physical barriers. J Clin Periodontol 1986; 13"862-8.

63. Knoxs B, Aukhil I. Ultrastructural study of experimental cementum regeneration in rats. oumal of Periodontal Research 1988; 23"60-67.

64. Line SE, Poison AM, Zander HA. Relationship between periodontal injury., selective cell repopulation and ankylosis. J Periodont 1974; 45:725-730.

65. Reitan K. Mechanism of apical root resorption. Trans Eur Orthod Soc 1972"363-79.

66. Gaudet EL. Tissue changes in the monkey following root torque with the Begg technique. Am J Orthod 1970; 58:164-78.

67. Williams S. A histomorphometric study of orthodontically induced root resorption. Eur J Orthod 1984; 6:35-47.

68. Kronfeld R. The biology of cementum. J Am Dent Assoc 1938; 25:1451-61.

69. Harris EF, Butler ML. Pattems of incisor root resorption before and after, orthodontic correction in cases with anterior open bites. Am J Orthod Dentofacial Orthop 1992; 101:112-9.

70. Owman-Moll P, Kurol J, Lundgren D. Repair of orthodontically induced root resorption in adolescents. Angle Orthod 1995; 65:403-8.

71. Brudvik P, Rygh P. Root resorption after local injection of prostaglandin E2 during experimental tooth movemem. Eur J Orthod 1991; 13:255-63. 172

72. Wesselnick PR, Beersten W. Initiating factors in dental root resorption. In: Z D, ed. Biological mechanisms of tooth eruption and root resorption, 1988"329-34.

73. Jones SJ. The root surface: an illustrated review of some scanning electron microscope studies. Scanning Microsc 1987; 1:2003-18.

74. Alatli I, Hellsing E, Hammarstrom L. Orthodontically induced root resorption in rat molars after 1- hydroxyethylidene-l,l-bisphosphonate injection. Acta Odontol Scand 1996; 54:102-8.

75. Alatli I, Hammarstrom L. Root surface defects in rat molar induced by 1- hydroxyethylidene-l,1- bisphosphonate. Acta Odontol Scand 1996; 54:59-65.

76. Levander E, Malmgren O. Evaluation of the risk of root resorption during orthodontic treatment" a study of upper incisors. Eur J Orthod 1988; 10:30-8.

77. Levander E, Malmgren O, Eliasson S. Evaluation of root resorption in relation to two orthodontic treatment regimes. A clinical experimental study. Eur J Orthod 1994; 16:223-8.

78. Loe H, Waerhaug J. Experimental replantation of teeth in dogs and monkeys. Arch Oral Biol 1961; 3" 176-84.

79. Andreasen JO, Paulsen HU, Yu Z, Schwartz O. A long-term study of 370 autotransplanted premolars. Part III. Periodontal healing subsequent to transplantation. Eur J Orthod 1990; 12:25-37.

80. McFadden WM, Engstrom C, Engstrom H, Anholm JM. A study of the relationship between incisor intrusion and root shortening. Am J Orthod Dentofacial Orthop 1989; 96:390-6.

81. Hemley S. The incidence of root resorption of vital.permanent teeth. J Dent Res 1941; 20:133-41.

82. VonderAhe G. Postretention status of maxillary incisors with root-end resorption. Angle Orthod 1973; 43:247-55.

83. Vardimon AD, Graber TM, Voss LR, Lenke J. Determinants controlling iatrogenic external root resorption and repair during and after palatal expansion. Angle Orthod 1991; 61" 113-22; discussion 123-4.

84. Vardimon AD, Graber TM, Pitaru S. Repair process of external root resorption subsequent to palatal expansion treatment. Am J Orthod Dentofacial Orthop 1993; 103:120-30. 173

85. Storey E, Smith R. Force in orthodontics and its relation to tooth movement. Austral J Dent 1952; 56:11-18.

86. Burstone CJ, Groves MH. Threshold and optimum force values for maxillary anterior tooth movement. J Dent Res 1961; 39"695.

87. Hixon EH, Aason TO, Arango J, et al. On force and tooth movement. Am J Orthod 1970; 57:476-489.

88. Owman-Moll P. Orthodontic tooth movement and .root resorption with special reference to force magnitude and duration. A clinical and histological investigation in adolescents. Swed Dent J Suppl 1995; 105" 1-45.

89. Owman-Moll P, Kurol J, Lundgren D. Effects of a doubled orthodontic force magnitude on tooth movement and root resorptions. An inter-individual study in adolescents. Eur J Orthod 1996; 18:141-50.

90. Owman-Moll P, Kurol J, Lundgren D. The effects of a four-fold increased orthodontic force magnitude on tooth movement and root resorptions. An intra- individual study in adolescents. Eur J Orthod 1996; 18"287-94.

91. Owman-Moll P, Kurol J, Lundgren D. Continuous versus interrupted continuous orthodontic force related to early tooth movement and root resorption. Angle Orthod 1995; 65:395-401.

92. Kurol J, Owman-Moll P, Lundgren D. Time-related root resorption after application of a controlled continuous orthodontic force. Am J Orthod Dentofacial Orthop 1996; 110:303-10.

93. Levander E, Malmgren O, Stenback K. Apical root resorption during orthodontic treatment of patients with multiple aplasia: a study of maxillary incisors. Eur J Orthod 1998; 20:427-34.

94. Hollender L, Ronnerman A, Thilander B. Root resorption, marginal bone support and clinical crown length in orthodontically treated patients. Eur J Orthod 1980; 2" 197- 205.

95. King GJ, Latta L, Rutenberg J, Ossi A, Keeling SD. Alveolar bone turnover and tooth movement in male rats after removal of orthodontic appliances. Am J Orthod Dentofacial Orthop 1997; 111:266-75.

96. Rosenberg HN. An evaluation of the incidence and amount of apical root resorption and dilaceration occurring in orthodontically treated teeth, having incompletely formed roots at the beginning of Begg treatment. Am J Orthod 1972; 61"524-5. 174

97. Harris EF, Baker WC. Loss of root length and crestal bone height before and during treatment in adolescent and adult orthodontic patients. Am J Orthod Dentofacial Orthop 1990; 98:463-9.

98. King G, Courts F. Changes in the titer of tooth root antibodies accompanying traumatic root resorption. In: Davisovitch Z, ed. Biological mechanisms of tooth eruption and root resorption, 1988:365-70.

99. Rosenberg EH, Guralnick WC. Hyperparathyroidism. A review of 220 cases, with special emphasis on findings in the jaws. Oral Surg 1962; 15:84-94.

100. Midgett RJ, Shaye R, Fruge JF. The effect of altered bone metabolism on orthodontic tooth movement. Am J Orthod 1981; 80:256-62.

101. Silverman S, Gordan G, Grant T, Steinbach H, Eisenberg E, Manson R. The dental structures in primary hyperparathyroidism. Studies in forty-two consecutive dentulous patients. Oral Surg 1962; i 5:426-36.

102. Becks H. Root resorption and their relationship to pathologic bone formation. Int J Orthod 1936; 22:445-82.

103. Goldie RS, King GJ. Root resorption and tooth movement in orthodontically treated, calcium deficient, and lactating rats. Am J Orthod 1984; 85:424-30.

104. Hall A. Upper incisor root resorption during Stage II of the Begg technique. Br J Orthod 1978; 5:47-50.

105. Ten Hoeve A, Mulie RM. The effect of antero-postero incisor repositioning on the palatal cortex as studied with laminagraphy. J Clin Orthod 1976; 10"804-22.

106. Remmelnick HJ. The effect of anteroposterior incisor repositioning on the root and cortical plate: a follow-up study. J Clin Orthod .1984; 18"42-9.

107. Kaley J, Phillips C. Factors related to root resorption in edgewise practice. Angle Orthod 1991; 61" 125-32.

108. Wainwright WM. Faciolingual tooth movement" its influence on the root and cortical plate. Am J Orthod 1973; 64:278-302.

109. Barber AF, Sims MR. Rapid maxillary expansion and external root resorption in man: a scanning electron microscope study. Am J Orthod 1981; 79"630-52.

110. Wickwire NA, McNeill MH, Norton L, Duell R. The effects of tooth movement upon endodontically treated teeth. Angle Orthod 1974; 44"235-42. 175

111. Remington DN, Joondeph DR, Artun J, Riedel RA, Chapko MK. Long-term evaluation of root resorption occurring during orthodontic treatment. Am J Orthod Dentofacial Orthop 1989; 96:43-6.

112. Malmgren O, Goldson L, Hill C, Orwin A, Petrini L, Lundberg M. Root resorption after orthodontic treatment of traumatized teeth. Am J Orthod 1982; 82:487-91.

113. Ramfjord SP, Ash MM.. Occlusion. Philadelphia: WB Saunders Company, 1986.

114. Odenrick L, Karlander EL, Pierce A, Kretschmar U. Surface resorption following two forms of rapid maxillary expansion. Eur J Orthod 1991; 13:264-70.

115. Sjolien T, Zachrisson BU. Periodontal bone support and tooth length in orthodontically treated and untreated persons. Am J Orthod 1973; 64:28-37. i i 6. Sharpe W, Reed B, Subteiny jD, Polson A. Orthodontic relapse, apical root resorption, and crestal alveolar bone levels. Am J Orthod Dentofacial Orthop 1987; 91:252-8.

117. Gottlieb BS. The effects of an intrusive base arch on tooth position" A radiographic study: University of Connecticut, 1979.

118. Tronstad L. Root resorption- a multidisciplinary problem in dentistry. In: Davidovitch Z, ed. The biological mechanisms of tooth eruption and root resorption, 1988:293-302.

119. Jacobson O. Clinical significance of root resorption. Am J Orthod 1952; 38"687- 96.

120. Ronnerman A, Larsson E. Overjet, overbite, intercanine distance and root resorption in orthodontically treated patients. A ten year follow-up study. Swed Dent J 1981; 5:21-7.

121. Copeland S, Green LJ. Root resorption in maxillary central incisors following active orthodontic treatment. Am J Orthod 1986; 89"51-5.

122. Gholston L, Mattison G. An endodontic-orthodontic technique for esthetic stabilization of externally resorbed teeth. Am J Orthod 1983; 83:435-40.

123. Zachrisson BU, Alnaes L. Periodontic condition in orthodontic treated and untreated individuals: II Alveolar bone loss: radiographic findings. Angle Orthod 1974; 44:48-55. 176

124. Baxter DH. The effect of orthodontic treatment on alveolar bone adjacent to cemento-enamel junction. Angle Orthod 1967; 37"35-47.

125. KalkwarfKL, Krejci RF, Pao YC. Effect of apical root resorption on periodontal support. J Prosthet Dent 1986; 56:317-9.

126. Gowen M. Bone cell biology and the regulation of bone tumover. Cytokines and bone metabolism: CRC Press, 1992" 1-70.

127. Balkwill FR, Burke F. The cytokine network [see comments]. Immunol Today 1989; 10:299-304.

128. Aggarwall BB, Gutterman JU. Human cytokines: Handbook for basic and clinical research. Boston: Blackwell Scientific, 1992.

129. Knauer DJ, Wiley HS, Cunningham DD. Relationship between epidermal growth factor receptor occupancy and mitogenic response. Quantitative analysis using a steady state model system. J Biol Chem 1984; 259"5623-31.

130. Gronthos S, Simmons PJ. The growth factor requirements of STRO-l-positive human bone marrow stromal precursors under serum-deprived conditions in vitro. Blood 1995; 85:929-40.

131. Francis K, Palsson BO. Effective intercellular communication distances are determined by the relative time constants for cyto/chemokine secretion and diffusion. Proc Natl Acad Sci U S A 1997; 94:12258-62.

132. Horton JE, Raisz LG, Simmons HA, Oppenheim JJ, Mergenhagen SE. Bone resorbing activity in supematant fluid from cultured human peripheral blood leukocytes. Science 1972; 177:793-5.

133. Clark BD, Collins KL, Gandy MS, Webb AC, Auron PE. Genomic sequence for human prointerleukin 1 beta: possible evolution from a reverse transcribed prointerleukin 1 alpha gene [published erratum appears in Nucleic Acids Res 1987 Jan 26;15(2)-868]. Nucleic Acids Res 1986; 14:7897-914.

134. Ngan PW, Crock B, Varghese J, Lanese R, Shanfeld J, Davidovitch Z. Immunohistochemical assessment of the effect of chemical and mechanical stimuli on cAMP and prostaglandin E levels in human gingival fibroblasts in vitro. Arch Oral Biol 1988; 33:163-74. 135. Dower SK, Kronheim SR, March CJ, et al. Detection and characterization of high affinity plasma membrane receptors for human interleukin 1. J Exp Med 1985; 162:501-15. 177

136. Hannum CH, Wilcox CJ, Arend WP, et al. Interleukin-1 receptor antagonist activity of a human interleukin-1 inhibitor. Nature 1990; 343"336-40.

137. Gowen M, Mundy GR. Actions of recombinant interleukin 1, interleukin 2, and interferon- gamma on bone resorption in vitro. J Immunol 1986; 136"2478-82.

138. Gowen M, Wood DD, Ihrie EJ, McGuire MK, Russell RG. An interleukin 1 like factor stimulates bone resorption in vitro. Nature 1983; 306"378-80.

139. Gowen M. Interleukin 1 and tumor necrosis factor. Cytokines and bone metabolism, 1992"71-107.

140. Beutler B, Cerami A. The biology of cachectin/TNF--a primary mediator of the host response. Annu Rev Immunol 1989; 7:625-55.

141. Vilcek J, Lee TH. Tumor necrosis factor. New insights into the molecular mechanisms of its multiple actions, j Bioi Chem 1991; 266"7313-6.

142. Schall TJ, Lewis M, Koller KJ, et al. Molecular cloning and expression of a receptor for human tumor necrosis factor. Cell 1990; 61 "361-70.

143. Le J, Vilcek J. Tumor necrosis factor and interleukin 1" cytokines with multiple overlapping biological activities. Lab Invest 1987; 56"234.

144. Bertolini DR, Nedwin GE, Bringman TS, Smith DD, Mundy GR. Stimulation of bone resorption and inhibition of bone formation in vitro by human tumour necrosis factors. Nature 1986; 319:516-8.

145. Konig A, Muhlbauer RC, Fleisch H. Tumor necrosis factor alpha and interleukin- 1 stimulate bone resorption in vivo as measured by urinary [3H]tetracycline excretion from prelabeled mice. J Bone Miner Res 1988; 3"621-7.

146. Meikle MC, Heath JK, Reynolds JJ. Advances in understanding cell interactions in tissue resorption. Relevance to the pathogenesis of periodontal diseases and a new hypothesis. J Oral Pathol 1986; 15:239-50.

147. Masada MP, Persson R, Kenney JS, Lee SW, Page RC, Allison AC. Measurement of interleukin-1 alpha and -1 beta in gingival crevicular fluid: implications for the pathogenesis of periodontal disease. J Periodontal Res 1990; 25" 156-63.

148. Preiss DS, Meyle J. Interleukin-1 beta concentration of gingival crevicular fluid. J Periodontol 1994; 65"423-8. 178

149. Yavuzyilmaz E, Yamalik N, Bulut S, Ozen S, Ersoy F, Saatci U. The gingival crevicular fluid interleukin-1 beta and tumour necrosis factor-alpha levels in patients with rapidly progressive periodontitis. Aust Dent J 1995; 40:46-9.

150. Ebersole JL, Singer RE, Steffensen B, Filloon T, Kornman KS. Inflammatory mediators and immunoglobulins in GCF from healthy, gingivitis and periodontitis sites. J Periodontal Res 1993; 28:543-6.

151. Yakovlev E, Kalichman I, Pisanti S, Shoshan S, Barak V. Levels of cytokines and collagen type I and type III as a function of age in human gingivitis. J Periodontol 1996; 67:788-93.

152. Heasman PA, Collins JG, Offenbacher S. Changes in crevicular fluid levels of interleukin-1 beta, leukotriene B4, prostaglandin E2, thromboxane B2 and tumour necrosis factor alpha in experimental gingivitis in humans. J Periodontal Res 1993; 28:241-7.

153. Tsai CC, Ho YP, Chen CC. Levels of interleukin-1 beta and interleukin-8 in gingival crevicular fluids in adult periodontitis. J Periodontol 1995; 66"852-9.

154. Schonfeld SE, Greening AB, Glick DH, Frank AL, Simon JH, Herles SM. Endotoxic activity in periapical lesions. Oral Surg Oral Med Oral Pathol 1982; 53"82- 7.

155. Torabinejad M, Kiger RD. Experimentally induced alterations in periapical tissues of the cat. J Dent Res 1980; 59:87-96.

156. Stashenko P, Wang CY, Tani-Ishii N, Yu SM. Pathogenesis of induced rat periapical lesions. Oral Surg Oral Med Oral Pathol 1994; 78:494-502.

157. Beresford JN, Gallagher JA, Gowen M, et al. The effects of monocyte- conditioned medium and interleukin 1 on the synth.esis of collagenous and non- collagenous proteins by mouse bone and human bone cells in vitro. Biochim Biophys Acta 1984; 801:58-65.

158. Van Dyke TE, Lester MA, Shapira L. The role of the host response in periodontal disease progression: implications for future treatment strategies. J Periodontol 1993; 64:792-806.

159. Kjeldsen M, Holmstrup P, Bendtzen K. Marginal periodontitis and cytokines" a review of the literature. J Periodontol 1993; 64:1013-22.

160. Tatakis D. Interleukin-1 and bone metabolism: a review. Journal of Periodontology 1993; 64"416-431. 179

161. Ralston SH, Grabowski PS. Mechanisms of cytokine induced bone resorption: role of nitric oxide, cyclic guanosine monophosphate, and prostaglandins. Bone 1996; 19:29-33.

162. Saito S, Ngan P, Saito M, Lanese R, Shanfeld J, Davidovitch Z. Interactive effects between cytokines on PGE production by human periodontal ligament fibroblasts in vitro. J Dent Res 1990; 69:1456-62.

163. Cahill DR, Marks SC, Jr. Tooth eruption: evidence for the central role of the dental follicle. J Oral Pathol 1980; 9"189-200.

164. Marks SC, Jr,, Cahill DR. Experimental study in the dog of the non-active role of the tooth in the eruptive process. Arch Oral Biol 1984; 29:311-22.

165. Marks SC, Jr., Cahill DR, Wise GE. The cytology of the dental follicle and adjacent alveolar bone during tooth eruption in the dog. Am J Anat 1983; 168:277-89.

166. Wise GE, Lin F. The molecular biology of initiation of tooth eruption. J Dent Res 1995; 74:303-6.

167. Kodama H, Yamasaki A, Nose M, et al. Congenital osteoclast deficiency in osteopetrotic (op/op) mice is cured by injections of macrophage colony-stimulating factor. J Exp Med 1991; 173:269-72.

168. Kodama H, Nose M, Niida S, Yamasaki A. Essential role of macrophage colony- stimulating factor in the osteoclast differentiation supported by stromal cells. J Exp Med 1991; 173:1291-4.

169. Iizuka T, Cielinski M, Aukerman SL, Marks SC, Jr. The effects of colony- stimulating factor-1 on tooth eruption in the toothless (osteopetrotic) rat in relation to the critical periods for bone resorption during tooth eruption. Arch Oral Biol 1992; 37:629-36.

170. Cielinski MJ, Jolie M, Wise GE, Ando DG, Marks SC. Colony-stimulating factor -1 is a potent stimulator of tooth eruption in the rat. In: Davidovitch Z, ed. Biological mechanisms of tooth eruption, resorption and replacement by implants, 1994:429- 436.

171. Wise GE, Lin F. Regulation and localization of colony-stimulating factor-1 mRNA in cultured rat dental follicle cells. Arch Oral Biol 1994; 39:621-7.

172. Wise GE, Lin F, Marks SC, Gorski JP. The molecular basis of tooth eruption. In: Davidovitch Z, ed. Biological mechanisms of tooth eruption, resorption and replacement by implants, 1994:383-390. 180

173. Wise GE, Fan W. Immunolocalization of transforming growth factor beta in rat molars. J Oral Pathol Med 1991; 20:74-80.

174. Lin F, Wise GE. Effect of epidermal growth factor on expression of transforming growth factor-beta 1 mRNA in stellate reticulum cells of rat mandibular molars. Dev Dyn 1993; 198:22-7.

175. Raisz LG, Martin TJ. Prostaglandins in bone and mineral metabolism. In" Pecks WA, ed. Bone and mineral research. Vol. 2, 1984:286-310.

176. Klein DC, Raisz LG. Prostaglandins: stimulation of bone resorption in tissue culture. Endocrinology 1970; 86"1436-40.

177. Yamasaki K, Shibata Y, Fukuhara T. The effect of prostaglandins on experimemal tooth movement in monkeys (Macaca fuscata). J Dent Res 1982; 61"1444-6.

! 78. Yamasaki K, Shibata Y, Imai S, Tani Y, Shibasaki Y, Fukuhara T. Clinical application of prostaglandin E1 (PGE 1) upon orthodontic tooth movement. Am J Orthod 1984; 85:508-18.

179. Sandy JR, Harris M. Prostaglandins and tooth movement. Eur J Orthod 1984; 6:175-82.

180. Mohammed AH, Tatakis DN, Dziak R. Leukotrienes in orthodontic tooth movement. Am J Orthod Dentofacial Orthop 1989; 95:231-7.

181. Sinha PK, Nanda RS. The relationship between orthodontic tooth movement and root resorption in PGE2 and non-PGE2 treated sides in a rat model. In: Davidovitch Z, Norton LA, eds. Biological mechanisms of tooth movement and craniofacial adaptation, 1996.

182. Yeh CK, Rodan GA. Tensile forces enhance pro.staglandin E synthesis in osteoblastic cells grown on collagen ribbons. CalcifTissue Int 1984; 36:$67-71.

183. Saito M, Saito S, Ngan PW, Shanfeld J, Davidovitch Z. Interleukin 1 beta and prostaglandin E are involved in the response of periodontal cells to mechanical stress in vivo and in vitro. Am J Orthod Dentofacial Orthop 1991; 99"226-40.

184. Saito S, Ngan P, Rosol T, et al. Involvement of PGE synthesis in the effect of intermittent pressure and interleukin-1 beta on bone resorption. J Dent Res 1991; 70:27-33. 185. Takahashi N, Kukita T, MacDonald BR, et al. Osteoclast-like cells form in long- term human bone marrow but not in peripheral blood cultures. J Clin Invest 1989; 83:543-50. 181

186. Ngan P, Saito S, Saito M, Lanese R, Shanfeld J, Davidovitch Z. The interactive effects of mechanical stress and interleukin-1 beta on prostaglandin E and cyclic AMP production in human periodontal ligament fibroblasts in vitro" comparison with cloned osteoblastic cells of mouse (MC3T3-E 1). Arch Oral Bio1.1990; 35:717-25.

187. Saito S, Rosol TJ, Saito M, Ngan PW, Shanfeld J, Davidovitch Z. Bone-resorbing activity and prostaglandin E produced by human periodontal ligament cells in vitro. J Bone Miner Res 1990; 5:1013-8.

188. Saito S, Saito M, Ngan P, Lanese R, Shanfeld J, Davidovitch Z. Effects of parathyroid hormone and cytokines on prostaglandin E synthesis and bone resorption by human periodontal ligament fibroblasts. Arch Oral Biol 1990; 35:845-55.

189. Saito S, Ngan P, Saito M, et al. Effects of cytokines on prostaglandin E and cAMP levels in human periodontal ligament fibroblasts in vitro. Arch Oral Biol 1990; 35"387-95.

190. Uematsu S, Mogi M, Deguchi T. Increase of transforming growth factor-beta 1 in gingival crevicular fluid during human orthodontic tooth movement. Arch Oral Biol 1996; 41"1091-5.

191. Uematsu S, Mogi M, Deguchi T. Interleukin (IL)-1 beta, IL-6, tumor necrosis factor-alpha, epidermal growth factor, and beta 2-microglobulin levels are elevated in gingival crevicular fluid during human orthodontic tooth movement. J Dent Res 1996; 75 "562-7.

192. Grieve WG, 3rd, Johnson GK, Moore RN, Reinhardt RA, DuBois LM. Prostaglandin E (PGE) and interleukin-1 beta (IL-1 beta) levels in gingival crevicular fluid during human orthodontic tooth movement. Am J Orthod Dentofacial Orthop 1994; 105:369-74.

193. Rossi M, Whitcomb S, Lindemann R. Interleukin-1 beta and tumor necrosis factor-alpha production by human monocytes cultured with L-thyroxine and thyrocalcitonin: relation to severe root shortening. Am J Orthod Dentofacial Orthop 1996; 110"399-404.

194. Igarashi K, Mitani H, Adachi H, Shinoda H. Anchorage and retentive effects of a bisphosphonate (AHBuBP) on tooth movements in rats. Am J Orthod Dentofacial Orthop 1994; 106:279-89. 195. Igarashi K, Adachi H, Mitani H, Shinoda H. Inhibitory effect of the topical administration of a bisphosphonate (risedronate) on root resorption incident to orthodontic tooth movement in rats. J Dent Res 1996; 75"1644-9. 182

196. Wesselink PR, Beertsen W. Ankylosis of the mouse molar after systemic administration of 1- hydroxyethylidene-l,l-bisphosphonate (HEBP). J Clin Periodontol 1994; 21:465-71.

197. Shigeyama Y, Grove TK, Strayhom C, Somerman MJ. Expression of adhesion molecules during tooth resorption in feline teeth: a model system for aggressive osteoclastic activity. J Dent Res 1996; 75:1650-7.

198. Chapnick L. External root resorption: an.experimental radiographic evaluation. Oral Surg Oral Med Oral Pathol 1989; 67:578-82.

199. Andreasen FM, Sewerin I, Mandel U, Andreasen JO. Radiographic assessment of simulated root resorption cavities. Endod Dent Traumatol 1987; 3"21-7.

200. Lowney JJ, Norton LA, Sharer DM, Rossomando EF. Orthodontic forces increase tumor necrosis factor alpha in the human gingival sulcus. Am J Orthod Dentofacial Orthop 1995; 108:519-24.

201. Brill N, Krasse B. The passage of tissue fluid into the clinically healthy gingival pocket. Acta Odontol Stand 1958; 16:223.

202. Brill N, Bronnestam R. Immunoelectrophoretic study of tissue fluid from gingival pockets. Acta Odontolo Stand 1960; 18:95.

203. Tolo K. Transport across stratified nonkeratinized epithelium. J Periodontal Res 1971; 6:23 7-42.

204. Schwartz J, Stinson FL, Parker RB. The passage of tritiated bacterial endotoxin across intact gingival crevicular epithelium. J Periodontol 1972; 43:270-6.

205. Egelberg J. Cellular elements in gingival pocket fluid. Acta Odontol Scand 1963; 21:283.

206. Selvig K. Structure and metabolism of the normal periodontium, Position paper. International Conference on Research in the Biology of Periodontal Disease, 1977.

207. Tolo KJ. A study of permeability of gingival pocket epithelium to albumin in guinea pigs and Norwegian pigs. Arch Oral Biol 1971; 16:881-8.

208. Loe H, Holm-Pederson P. Absence and presence of fluid from normal and inflamed gingiva. Periodontics 1965; 3:171.

209. Weinstein E, Mandel I, Salkind A. Studies of gingival fluid. Periodontics 1967; 5:161. 183

210. Cimasoni G. Crevicular fluid updated. In: Myers A, ed. Monographs in Oral Science. Vol. 12. New York: Karger, 1983.

211. Challacombe SJ. Passage of serum aimmunoglobulin into the oral cavity. In: Lohner T, Cimasoni G, eds. Borderland between caries and periodontal disease. Vol. 2, 1980.

212. Hirose K, Isogai E, Miura H, Ueda I. Levels ofPorphyromonas gingivalis Fimbriae and inflammatory cytokines in gingival crevicular fluid from adult human subjects. Microbiol Immunol 1997; 41:21-6.

213. Lindhe J, Attstrom R, Bjom AL. Influence of sex hormones on gingival exudation in gingivitis-free female dogs. J Periodontal Res 1968; 3"273-8.

214. Lindhe J, Bjom AL. Influence of hormonal contraceptives on the gingiva of women. J Periodontal Res 1967; 2:1-6.

215. McLaughlin WS, Lovat FM, Macgregor ID, Kelly PJ. The immediate effects of smoking on gingival fluid flow. J Clin Periodontol 1993; 20:448-51.

216. Rossomando EF, White L. A novel method for the detection of TNF-alpha in gingival crevicular fluid. J Periodontol 1993; 64"445-9.

217. Bien M. The dynamics of periodontal fluid composition. Arch Oral Biol 1983; 3 04:123-42.

218. Rossomando EF, White LB, Hadjimichael J. Immunomagnetic separation of tumor necrosis factor alpha. II. In situ procedure for the human gingival space. J Chromatogr 1992; 583:19-26.

219. McCulloch CA, Nemeth E, Lowenberg B, Melcher AH. Paravascular cells in endosteal spaces of alveolar bone contribute to perigdontal ligament cell populations. Anat Rec 1987; 219:233-42.

220. McCulloch CA, Heersche JN. Lifetime of the osteoblast in mouse periodontium. Anat Rec 1988; 222:128-35.

221. Russell AL. A system of classification and scoring for prevalence surveys of periodontal disease. Journal of Dental Research 1956; 35"350.

222. Odenrick L, Brattstrom V. The effect of nailbiting on root resorption during orthodontic treatment. Eur J Orthod 1983; 5"185-8.

223. Odenrick L, Brattstrom V. Nailbiting: frequency and association with root resorption during orthodontic treatment. Br J Orthod 1985; 12"78-81. 184

224. Loe H, Silness J. Periodontal disease in pregnancy. Acta Odontol Scand 1963; 21:533.

225. Silness J, Loe H. Periodontal disease in pregnancy. 2. Acta Odontol Scand 1964; 22:121.

226. Silness J, Loe H. Periodontal disease in pregnancy. 3. Response to local treatment. Acta Odontol Scand 1966; 24:747-59.

227. Rossomando EF, Kennedy JE, Hadjimichael J. Tumour necrosis factor alpha in gingival crevicular fluid as a possible indicator of periodontal disease in humans. Arch Oral Biol 1990; 35:431-4.

228. Rossomando EF, White LB, Hadjimichael J, Schaffer D. Immunomagnetic separation of tumor necrosis factor alpha. I. Batch separation from the human temporomandibular fluid. J Chromatogr 1992; 583"1 I-18.

229. Tsai CC, Hong YC, Chen CC, Wu YM. Measurement of prostaglandin E2 and leukotriene B4 in the gingival crevicular fluid. J Dent 1998; 26"97-103.

230. Mathur A, Michalowicz B, Castillo M, Aeppli D. Interleukin-1 alpha, interleukin- 8 and interferon-alpha levels in gingival crevicular fluid. J Periodontal Res !996; 31:489-95.

231. Ralston SH. Role of cytokines in clinical disorders of bone metabolism. Cytokines and bone metabolism: CRC Press, 1992"361-381.

232. Insoft M, King GJ, Keeling SD. The measurement of acid and alkaline phosphatase in gingival crevicular fluid during orthodontic tooth movement. Am J Orthod Dentofacial Orthop 1996; 109:287-96.

233. Nakashima K, Roehrich N, Cimasoni G. Osteocalcin, prostaglandin E2 and alkaline phosphatase in gingival crevicular fluid" their relations to periodontal status. J Clin Periodontol 1994; 21:327-33.

234. Kao RT, Curtis DA, Richards DW, Preble J. Increased imerleukin-1 beta in the crevicular fluid of diseased implants. Int J Oral Maxillofac Implants 1995; 10:696- 701. 235. Tzannetou S, Efstratiadis S, Nicolay O, Grbic J, Lamster I. Interleukin-lbeta and beta-glucuronidase in gingival crevicular fluid from molars during rapid palatal expansion [In Process Citation]. Am J Orthod Dentofacial Orthop 1999; 115:686-96.

236. Carranza F. Defense mechanisms of the gingiva. Glickman's Periodontal Periodontology, 1990:96-98. 185

237. Spolsky V. The epidemiology of gingival and periodontal disease. In: FA C, ed. Glickman's Clinical Periodontology: WB Saunders, 1990"302-329.