During the 1990s, the use of systematic reviews, or SRs, became the preferred process for reviewing large volumes of clinical and scientific data in medicine and other health fields. Although this evidence-based approach is expensive and time-consuming, it is more objective than traditional literature reviews. It is scientifically sound and patient-focused, incorporates clinical experience, stresses good judgment and yields helpful information for use in clinical decision making.^2–5 Participants in the AAP Workshop on Contemporary Science in Clinical Periodontics used SRs to develop consensus statements that included implications for practice and research.

The AAP’s SR process began with development of a protocol that guided all phases of the project.⁶ Reviewers trained in SR methodology formulated clinically relevant, focused questions and developed formal search criteria for each question. To capture the entire body of evidence, the reviewers conducted systematic searches of online databases and conducted hand searches of journals. Reviewers also sought unpublished data by contacting authors, journal editors and industry experts. To ensure consistency, a single bio-statistician performed all statistical analyses.

The grade given to each consensus statement does not mean that a therapy should or should not be used in the care of a particular patient.

Fifteen SRs were conducted as a part of three general topics included in the conference: host modulation, anti-infective therapy and tissue engineering. This article outlines the methods used for conducting SRs, highlights findings from the SRs and discusses consensus statements developed at the conference that have implications for clinical practice and research.

	METHODS

TOP ABSTRACT METHODS CONFERENCE FINDINGS DISCUSSION CONCLUSIONS REFERENCES

 
For each SR, the reviewers determined the level of evidence according to a hierarchy that reflects the relative value of the information (Table 1).⁶ For example, a well-designed randomized clinical trial provides better evidence than does a case-control study.^7–9 In addition, the reviewers used predetermined criteria to assess each study for internal validity (such as rationale, structure, statistical approach) and external validity (such as relevance and applicability to typical practice) for inclusion. They then summarized the findings for each focused question in an SR.⁶

At the workshop, consensus groups evaluated the SRs, produced consensus statements and assigned each statement a grade (termed "strength") that reflected the underlying level of evidence (Table 2). The grade given to each consensus statement does not mean that a therapy should or should not be used in the care of a particular patient. Instead, it means that, given our present level of knowledge, there is evidence that a particular therapy with a strong recommendation may yield a predictable result and should be considered in the appropriate circumstances. Therapies that are graded "moderate," "limited" or even "insufficient" may be of value and may be considered, but further studies are needed to better confirm appropriate applications and predictability. Indeed, multi-center, large-scale, placebo-controlled studies do not exist for many of this conference’s topics, nor for much of the care that is provided by dentists in general. Therefore, the absence of data necessary to support a strong recommendation may indicate the need for additional research and is not an indication that a particular therapy is flawed.

	CONFERENCE FINDINGS

TOP ABSTRACT METHODS CONFERENCE FINDINGS DISCUSSION CONCLUSIONS REFERENCES

Atherosclerosis. Atherosclerosis is a major factor in the development of cardiovascular disease, or CVD, and infection and inflammation have been implicated in its pathogenesis.¹⁴ Risk factors for atherosclerosis-related disease (such as increased levels of C-reactive protein, fibrinogen and leukocytes) have been identified in patients with periodontitis, suggesting that localized infections resulting in inflammation and tissue devolution in the periodontium have a systemic effect associated with initiation and/or progression of atherosclerosis.^15–17

In the SR, four of five case-control studies found a positive association between indicators of poor oral health and CVD; 11 of 15 cross-sectional studies supported a modest association between periodontitis and CVD, after researchers controlled for CVD risk factors such as smoking. Four studies reported a positive association between periodontitis and stroke.¹⁸ Additional studies are needed to better define the strength of these associations.

Nosocomial pneumonia. Community-acquired pneumonia typically is caused by aspiration of resident oropharyngeal bacteria such as Streptococcus pneumoniae, Haemophilus influenzae and Mycoplasma pneumoniae. In contrast, nosocomial pneumonia, which occurs in institutional settings such as hospital intensive-care units and nursing homes, usually is caused by pathogens that colonize the oropharynx, such as Pseudomonas aeruginosa, Staphylococcus aureus and enteric gram-negative bacteria.¹⁹ Poor oral hygiene is common in institutional settings and may have a role in the initiation and/or progression of nosocomial pneumonia.

In the SR, some studies reported an increased risk of nosocomial pneumonia in dentate patients and in those whose dental plaque was colonized by respiratory pathogens.¹⁹ One study reported significant associations between pneumonia and the number of carious teeth, the frequency of tooth brushing and being dependent on others for oral care.²⁰ In addition, meta-analysis of oral hygiene intervention trials provided moderate evidence that intervention significantly reduces the odds of pneumonia in institutionalized patients. As with CVD, additional studies are needed.

Adverse pregnancy outcomes. Preterm delivery and low birth weight are significant causes of infant morbidity and mortality. Bacterial vaginosis is a known risk factor for these and other adverse outcomes of pregnancy.²¹ Infections in other areas of the body have been hypothesized to have a role in neonatal morbidity and mortality.

In the SR, moderate evidence suggested an association between periodontitis and adverse pregnancy outcomes, but causality was unclear.²² Owing to study heterogeneity, meta-analysis was not possible; however, preliminary evidence suggested that periodontal intervention might reduce adverse pregnancy outcomes.²² Discrepancies among studies may be related to different study populations and to other risk factors. For example, African-American women have both a higher incidence of adverse pregnancy outcomes and a higher prevalence of periodontitis, but confounding factors such as socioeconomic status must be more fully considered before accurate conclusions can be drawn.

Evidence suggests that periodontitis can be a systemic exposure that may contribute to the development or progression of other diseases and conditions.

Summary. Evidence suggests that periodontitis can be a systemic exposure that may contribute to the development or progression of other diseases and conditions. The consensus groups suggested that patients and health care providers be informed that periodontal intervention may prevent the onset or progression of atherosclerosis-induced diseases, such as myocardial infarction and ischemic stroke, and may help to prevent adverse pregnancy outcomes. Additional intervention studies are needed to help clarify these associations and could lead to formulation of more detailed recommendations for clinical practice.

Anti-infective therapy. Because bacterial infection is the primary cause of periodontal diseases, reducing the amount of subgingival plaque is an essential component of therapy. Nonsurgical, mechanical therapies, including scaling and root planing, or SRP, usually are considered the first steps in anti-infective management of periodontal diseases. SRP usually is effective, but bacteria quickly recolonize subgingival environments and re-form biofilms. A number of mechanical treatments have been used in therapy, and various anti-infective pharmacological agents have been used adjunctively to better manage the subgingival flora.

Local anti-infective therapy: pharmacologic agents. Locally delivered anti-infective agents generally are administered in doses much smaller than those used in systemic delivery. These agents include antibacterial gels, microspheres and irrigants.

In the SR, moderate evidence suggested that adjunctive therapy with minocycline gel or minocycline microspheres enhances probing depth reduction, and that adjunctive therapy with chlorhexidine chips or doxycycline gel enhances gain in clinical attachment level, or CAL.²³ Strong evidence suggested that the adjunctive use of these locally delivered intracrevicular agents does not result in significant adverse events. Strong evidence also suggested that, compared with SRP alone, therapist-delivered chlorhexidine irrigation during SRP does not improve probing depth, enhance CAL gain or reduce bleeding on probing.

Further research is required to determine effectiveness and duration of action of local anti-infective therapy in all forms of periodontitis.

Local anti-infective therapy: mechanical and manual approaches. SRP is a common professional approach to the initial treatment of the periodontal diseases. The relative efficacy of various methods—including manual versus machine-driven approaches with or without adjunctive agents—has not been determined.

The reviewers performed an SR to identify the most effective therapies on the basis of both clinical and patient-centered outcomes. The SR focused on studies that compared the efficacies of manual and mechanical approaches with and without irrigation. Nine randomized clinical trials were included in the review, and there was moderate evidence that manual methods and mechanically driven instrumentation were equally effective. Furthermore, there was moderate evidence from a limited number of studies that subgingival irrigation provides no benefit beyond that achieved with mechanical therapy alone.

Systemic agents. A number of systemic antibiotics have been used adjunctively in the treatment of periodontitis. In the SR, strong evidence indicated that the use of systemic antibiotics as an adjunct to SRP results in CAL improvement lasting at least six months regardless of baseline probing depth, and that CAL improvement is greater in patients with aggressive periodontitis than in those with chronic periodontitis.²⁴ Moderate evidence suggested that tetracycline and metronidazole provide statistically significant CAL improvement, but there was insufficient evidence to determine the optimal dosage and duration of this or other systemic antimicrobial regimens.

The few studies that investigated adverse events reported that gastrointestinal symptoms occurred, but were infrequent. Superinfection—that is, overgrowth of opportunistic pathogens such as fungi—is another potential adverse effect. In addition, the development of antibiotic resistance is a major concern whenever systemic antibiotics are used.²⁵ The emergence of bacterial strains that are resistant to antibiotics not only is potentially hazardous for the patient undergoing treatment, but also can have serious untoward implications for the global community.²⁵

A clinician’s decision to recommend local or systemic antibiotics is one of clinical judgment.

Microbial identification. The benefits that accompany the adjunctive use of systemic antibiotics in periodontal therapy, along with evidence that implicates specific bacteria in the etiology of periodontitis, have spurred scientific investigations to evaluate roles for microbial analyses in treatment decision making. Of the 83 articles reviewed for this SR, the reviewers included 24, most of which were case reports. There was insufficient evidence to support microbial identification in treating chronic periodontitis. There was limited evidence to support the use of microbial identification in the management of aggressive periodontitis and in nonresponsive cases of chronic periodontitis. The absence of high-level evidence supporting the use of microbiologic testing in periodontal treatment planning is not necessarily an indication that these procedures are of no value. Rather, it suggests that better investigations are needed to evaluate the clinical outcomes of therapeutic regimens based on microbial identification.

Summary. The evidence shows that manual and mechanically driven approaches are equally effective in improving clinical outcomes. Subgingival irrigation with a variety of agents as an adjunct to SRP was no more effective than SRP alone. Evidence suggests that certain locally delivered anti-infective agents may provide an adjunctive benefit, and that adjunctive use of certain systemic antibiotics results in CAL improvement, especially in patients with aggressive periodontitis. These adjunctive effects are small, as compared with the magnitude of change resulting from SRP. There is insufficient evidence to support specific recommendations regarding choice or dose of antibiotics, duration of their use or their use as monotherapies. Furthermore, because the use of systemic antibiotics can lead to superinfection and development of resistant bacterial strains, this form of anti-infective therapy merits caution. Evidence suggests that microbial identification may be of limited value in the treatment of aggressive periodontitis and nonresponsive chronic periodontitis. A clinician’s decision to recommend local or systemic antibiotics is one of clinical judgment that is likely to incorporate the phase of treatment, the patient’s preferences or both.

Host-modulating agents. Periodontal tissue destruction is mediated by the host response to infection. Three approaches to host modulations have been used: antiproteinases, which inhibit matrix metalloproteinases; anti-inflammatory drugs, which inhibit the production of cytokines and prostaglandins; and bone-sparing agents, which inhibit the activation of osteoclasts.

The antiproteinase used in treating periodontal disease is doxycycline, a tetracycline that inhibits matrix metalloproteinases. Subantimicrobial doses of doxycycline, or SDD, are used as an adjunct to SRP at a dosage of 20 milligrams twice daily. In this SR, meta-analysis of studies of nine to 12 months’ duration yielded strong evidence supporting the use of SDD in combination with root planing.²⁶ Compared with SRP alone, the combination of SRP and SDD provided statistically significant improvements in both mean CAL and probing depth. Throughout the existing studies, adverse events and side effects were not significantly different from those seen with conventional treatment alone. Research has not demonstrated the development of resistant bacterial strains. Studies of longer duration are needed.

Moderate evidence indicated that the use of systemic nonsteroidal anti-inflammatory drugs, or NSAIDs, can slow bone loss, but long-term use carries the risk of significant adverse effects such as gastric ulcers and bleeding. There was limited evidence to support the use of topical NSAIDs, and insufficient evidence for the use of bone-sparing agents such as the bisphosphonates.

Summary. Based on nine- to 12-month clinical trials, there is strong evidence that SDD provides small but statistically significant improvements in CAL and probing depth without significant side effects. The clinician’s decision to use adjunctive SDD therapy remains a matter of clinical judgment that may include consideration of the phase of treatment as well as the patient’s status and preferences.

Tissue engineering: natural teeth. Regeneration of tooth-supporting structures is a major goal of therapy in patients with periodontal osseous defects. Amelogeninlike factors; root surface modification; bone replacement grafts, or BRGs; and guided tissue regeneration, or GTR, are among the techniques used in periodontal reconstructive surgery. Recession of the gingival tissues has been addressed with various grafting techniques and guided tissue regeneration.

Amelogeninlike factors. Amelogenins are proteins involved in the formation of tooth enamel and tooth attachment structures. Enamel matrix derivative, or EMD, an amelogeninlike factor, has been used to promote periodontal regeneration. EMD is applied topically to exposed root surfaces during surgical treatment of periodontal osseous defects.

In the SR, meta-analysis of Level I and Level II data provided strong evidence supporting the adjunctive use of EMD.²⁷ Compared with open-flap débridement, or OFD, alone, the combination of OFD and EMD provided highly consistent and statistically significant CAL gain and probing depth reduction in patients with periodontal osseous defects. One study analyzed serum antibody levels and found no elicitation of antibody response.²⁸ EMD appears to be safe for single and multiple administrations and does not impede periodontal wound healing.²⁹

Root surface modification. The chemical modification of the root surface is one of the oldest and most frequently attempted regenerative techniques. Many materials have been used, including citric acid, tetracycline and ethylenediaminetetraacetic acid.

The SR, based on 20 of 21 controlled observational studies, showed that moderate evidence indicates that the use of chemical root modifiers during periodontal surgery does not provide adjunctive benefits in CAL gain or probing depth reduction.³⁰

Bone replacement grafts. BRGs are used widely for correction of periodontal osseous defects. Various graft materials have been used, including allogenic bone, autogenous bone, calcium phosphate ceramic, bioactive glass, and others such as coralline calcium carbonate.

In the SR, the reviewers evaluated studies comparing BRGs with OFD and found strong evidence that use of BRGs in the treatment of intra-bony defects increases bone level and CAL and reduces crestal bone loss and probing depth.³¹ In particular, the use of demineralized freeze-dried bone allograft, or DFDBA, as a periodontal regenerative material is supported strongly by both clinical studies and studies that assessed histological outcomes.

Guided tissue regeneration. GTR is an important technique in periodontal reconstructive surgery. GTR delays apical migration of the gingival epithelium and allows tissue from the periodontal ligament and surrounding bone to repopulate the space adjacent to the denuded root surface.

In the SR, strong evidence supported the following findings.³² Compared with OFD, use of barrier membranes improves CAL and probing depth in both intrabony and furcation defects. In intrabony defects,

– use of a BRG in addition to a barrier does not improve clinical outcomes;

– morbidity associated with use of bioabsorbable membranes is no greater than that associated with OFD;

– use of a barrier may be associated with increased recession of the gingival margin;

– gingival recession is similar with bioresorbable and nonresorbable barriers;

– intensive postoperative care protocols result in shallower probing depths.

In furcation defects,

– use of a BRG in addition to a physical barrier improves clinical outcomes;

– use of coronally advanced surgical flaps to cover the barrier improves clinical outcomes.

Treatment of gingival recession. The SR assessing the surgical treatment of gingival recession showed strong evidence—from 31 randomized clinical trials and two independent systematic reviews—that several surgical procedures are successful in covering exposed root surfaces. There is strong evidence that coronally positioned flaps with autogenous connective-tissue grafts result in greater root coverage and increased keratinized tissue compared with GTR procedures.

Summary. There is strong evidence supporting the use of EMD in regenerative surgical treatment of periodontal osseous defects. BRGs improve clinical outcomes of surgery for intrabony defects, and there is strong evidence supporting that GTR enhances clinical outcomes of periodontal surgery for both intrabony and furcation defects. There is strong evidence supporting the use of EMD, BRG and GTR rather than OFD procedures to increase bone level, reduce crestal bone loss, increase CAL and reduce probing depth in the treatment of intrabony defects. There is moderate evidence that chemical root modification provides no clinically significant benefit to periodontal regeneration. The clinical application of these regenerative procedures depends on the clinical judgment of the therapist and the specific needs of the individual patient. There is strong evidence that gingival recession can be treated with connective-tissue grafts and coronally positioned flaps, resulting in decreased root sensitivity and improved esthetics.

Tissue engineering: implants. Dental implant studies traditionally have focused on implants placed in native bone. Clinicians have used a number of methods to reconstruct the alveolar ridge in patients with trauma-induced defects, extensive bone resorption or other anatomic limitations. Inadequate alveolar bone height is a common problem in the placement of implants in the posterior maxilla, and most often it is addressed by grafting the floor of the maxillary sinus.

Two SRs were conducted. In one SR, moderate evidence suggested that implants placed at sites of socket bone augmentation are successful under functional loads, and that success rates for implants placed at sites of horizontal bone augmentation are similar to those for implants placed in native bone.³³ In the second SR, strong evidence indicated that the use of the lateral window technique for sinus bone augmentation provides sufficient bone for implant placement, with implant survival rates comparable to those seen in native bone.³⁴ Strong evidence also indicated that membrane coverage of the lateral window enhances implant survival.

Summary. Implant success at sites of bone augmentation is similar to that of implants placed in native bone, particularly in maxillary sinus augmentations, socket bone augmentations and horizontal bone augmentations. To date, there is only limited evidence regarding vertical bone augmentation of implant sites. Studies are needed to evaluate the role of molecular, cellular and genetic tissue-engineering techniques to enhance the quality and quantity of bone in augmentation areas. Studies are needed to evaluate the roles of molecular, cellular and genetic tissue-engineering techniques in enhancing bone quality and quantity.

	DISCUSSION

TOP ABSTRACT METHODS CONFERENCE FINDINGS DISCUSSION CONCLUSIONS REFERENCES

 
Clinical utility of SRs. The clinical applicability of the results of these SRs should be based on an individualized assessment of a patient’s needs and values, along with the strength of the evidence considered in a context that includes cost-benefit comparison between treatment alternatives. Such analyses include a comparison of the costs (physiological, psychological, monetary and so forth) associated with various treatment alternatives superimposed on the likelihood of treatment success, treatment failure or something in between. For example, in treating patients with periodontitis, most practitioners would agree that a treatment plan that aims to achieve anatomical stability by fully eliminating periodontal inflammation might require more than SRP. Indeed, a surgical procedure to fully débride infected sites may be needed. Some clinical circumstances may suggest treatment plans that include even more extensive surgical, regenerative and chemotherapeutic regimens in an effort to optimize oral, and perhaps even systemic, well-being. The appropriateness of any treatment (or nontreatment) direction as well as any subsequent care is unique to each patient’s values and (sometimes changing) needs, in a context of appropriate and ethical cost-benefit comparisons.

Limitations of SRs. As useful as SRs can be, they have a number of limitations.

– SRs address specific clinical questions. Applying the results of an SR to what appear to be similar clinical questions may lead to inappropriate use of an SR’s conclusions. Therefore, such extrapolations merit caution.

– The qualification criteria for inclusion in an SR may result in only a few studies’ being evaluated. This limitation does not affect statistical analyses or SR evaluation of data quality or the strength grades of a resulting consensus statement, but it can make the validity of a single study’s a priori assumptions, methods and, ultimately, accuracy more critical to the veracity of the SR’s conclusions.

– SRs assess the quality of studies’ evidence on a scale based on the methods used to obtain study data (for instance, randomized, controlled clinical trials are considered more credible than case reports).⁶ The strength of the consensus recommendations is based on the quality of the evidence from the SRs. A consensus statement with a "strong" recommendation can be misconstrued to reflect a consensus opinion endorsing the widespread clinical utility of the intervention for patients who may benefit in some way from treatment. This is because, in the case of the periodontal diseases, the magnitude and longevity of the measured effects remain unknown. This is especially important in the management of chronic maladies such as periodontal diseases, an area of study in which most randomized, blinded, multicenter clinical trials usually are less than a year in duration. On the other hand, a new but largely uninvestigated intervention for which low-level data (such as from uncontrolled experiments that result in "limited" or "insufficient" consensus recommendations) suggest a major (or inexpensive) therapeutic effect may be discounted inappropriately.

Published SRs can go out of date quickly, especially in areas in which ample research is under way. A possible example related to one of the SRs reported in this article is the mid-2003 postconference publication of a periodontal intervention study in pregnant women.³⁵ This pilot study indicated that in women with periodontitis, low-risk periodontal therapy may decrease the incidence of what often are high-morbidity premature deliveries. Fortunately, straightforward methods are available to update existing SRs.

	CONCLUSIONS

TOP ABSTRACT METHODS CONFERENCE FINDINGS DISCUSSION CONCLUSIONS REFERENCES

 
SRs are an important component of evidence-based dentistry, an approach that integrates the systematic evaluation of clinically relevant scientific evidence with the oral and medical condition of the individual patient and the expertise of the individual clinician. The decision to use specific therapies depends on clinical judgment, the phase of treatment and the likelihood of success in light of the patient’s needs, values and preferences. SRs capture and describe existing evidence, are effective tools for the development of consensus reports and clinical protocols, and provide direction for future studies of disease etiologies and associations, of therapeutic techniques and technologies, and of the impact of periodontal diseases and their treatment on individuals and populations.