In order to identify potentially relevant human randomized controlled clinical trials published in English, a comprehensive systematic search was performed across several electronic databases, namely PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, Scopus, and Embase. The search included publications up to December 2024 and utilized a mix of Medical Subject Headings (MeSH) and free-text keywords. The search strategy, initially designed for PubMed and adjusted for use in the other databases, combined terms as follows: (Periodontal Diseases OR periodontal disease OR periodontitis) AND (tetracycline OR doxycycline OR minocycline) AND (local OR topical) AND (randomized controlled trials OR clinical trials) (Supplementary Appendix 1). In addition, the reference lists of selected systematic reviews and individual studies were manually examined to locate any potentially relevant articles not captured by the electronic database. This systematic review was conducted in accordance with the PRISMA 2020 guidelines for systematic reviews and meta-analyses (Supplementary Appendix 2), and its protocol was registered in the PROSPERO database (Registration No: CRD42021247300) (Figure 1).

Although the redefinition of periodontal diagnoses (e.g., Stage/Grade) was established by the 2017 World Workshop on the Classification of Periodontal and Peri-Implant Diseases (43), a plethora of published literature from 1990 to 2024 labeled cases as “chronic periodontitis” and described interventions as “SRP”. Since, SRP remains the gold standard non-surgical therapy and therefore, in order to ensure comprehensiveness and consistency with the available RCT literature, we retained the terms “SRP” and “chronic periodontitis” in our search strategy and data inclusion criteria while acknowledging the updated classification framework.

The inclusion criteria were defined according to the PICO framework:

Population (P): Human participants aged 18 years or older diagnosed with chronic or adult periodontitis.

Intervention (I): SRP (whether full-mouth, localized, or performed in single or repeated appointments) combined with the administration of a locally delivered tetracycline-based antibiotic (e.g., tetracycline, doxycycline, or minocycline) at manufacturer-recommended concentrations or dosages.

Comparison (C): SRP alone or combined with a vehicle/placebo.

Outcomes (O): Primary outcomes included changes in PPD and CAL. Secondary outcomes included BOP, PI, and GI, which provide additional information on inflammation and oral hygiene status.

Any adverse effects reported by patients or examiners were also collected when available. Only human randomized controlled clinical trials (RCTs) with either parallel or split-mouth designs were included. A minimum follow-up period of six months was required for inclusion. Studies were excluded if they involved systemic antimicrobials as an intervention, evaluated local antimicrobials as monotherapy, employed non–sustained-release delivery vehicles, or were conducted exclusively in diabetic populations.

A systematic search of electronic databases was conducted to identify relevant studies. Two independent reviewers (NSS, SLJ) evaluated study eligibility by first screening titles and abstracts, followed by a detailed review of the full texts of potentially eligible studies. Studies lacking adequate data for meta-analysis were included in the systematic review but excluded from the meta-analysis. Data extraction was carried out using Microsoft Excel (Microsoft, Redmond, WA, USA), collecting the following details: author information (name and publication year), participant demographics, study design and characteristics, methods for clinical outcome assessment, intervention specifics (type and dosage of tetracycline used), changes in clinical indices, and any documented adverse effects.

Quality assessment was conducted independently by the same two reviewers. The risk of bias within individual randomized controlled trials was evaluated using the Cochrane Collaboration Risk of Bias Tool, which is widely regarded as the standard instrument for assessing methodological quality in intervention studies (44). This tool was chosen because it provides a structured framework to assess bias across key domains: selection bias (random sequence generation and allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessors), attrition bias (handling of incomplete outcome data), and reporting bias (selective outcome reporting). Each domain was rated as having a high (H), low (L), or unclear risk (UR) of bias (45).

Meta-analyses were performed by pooling the mean differences in treatment effects from baseline to follow-up. Data on primary outcomes from the selected studies were combined and evaluated using weighted mean differences (WMDs) accompanied by 95% confidence intervals (CIs). For both test and control groups, mean changes in clinical indices—such as probing pocket depth (PPD) and clinical attachment level (CAL)—were obtained from the studies. When these differences were not directly reported, they were derived from baseline and end-point values following methods outlined in prior literature (34, 35, 46). The difference (Δ) was computed using the equation:ΔPPD=PPDBL−PPDEndwhere PPDBL is the mean PPD at baseline, and PPDEnd is the mean PPD at the end of follow-up. This calculation was extended to other indices, including ΔCAL, ΔBOP, ΔPI, and ΔGI. Variances for these differences were estimated using,VarΔ=VarBL+VarEnd−2×r×σBL×σEndwhere r (assumed 0.5) represents the correlation between baseline and follow-up measures, and σBL and σEnd are the standard deviations at baseline and end of follow-up, respectively (34, 35, 46). Heterogeneity among studies was quantified with the Cochrane Q test and I² statistic (25% = low, 50% = moderate, 75% = high), and effect estimates were combined under both fixed-effect/common-effect (Mantel–Haenszel–Peto) and random-effects (DerSimonian–Laird) models, reporting the latter as significant heterogeneity was present. Subsequently, subgroup analysis and meta-regression were performed. All computations were performed in R (version 4.4.2; R Core Team, 2024), with statistical significance defined at P ≤ 0.05.

For assessment of publication bias across trials, funnel plot asymmetry was examined and formally tested using Egger's regression method (47) when at least ten studies were available. If bias was indicated, the trim-and-fill approach (48) was applied to estimate and impute potentially missing studies and derive an adjusted pooled effect size. Robustness of the findings was further evaluated through sensitivity analyses—by leave-one-study-out approach, in which each study was omitted in turn to assess its individual contribution to the overall effect.

The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) study flow, which delineates the process of article inclusion, is shown in Figure 1. Publications released until 2024 that were in the English language underwent a screening process. Records were screened, and following an examination of the titles and abstracts of the identified articles, 52 articles were selected; of these, 43 articles were incorporated into the systematic review, while the remaining 9 were excluded after a thorough analysis of the full texts (Supplementary Appendix 3). For the meta-analysis, 33 articles met the eligibility criteria. Although 52 articles were reviewed for data extraction, they correspond to 43 independent studies due to the fact that clinical data and clinical data coupled with microbiological data were presented in two distinct publications: Cortelli et al. (49, 50); Drisko et al.; Michalowicz et al. (26, 51); Eickholz et al.; Ratka-Krüger et al. (21, 52); Goodson et al.; Socransky et al. (53, 54); Newman et al.; Wilson et al. (28, 55); Wong et al.; Wong et al. (56, 57). Thus, the publications that addressed the clinical data were utilized (21, 26, 28, 49, 53, 56). The investigations conducted by Tomasi et al. (58); Tomasi and Wennstrom (59) explored the same data from varying perspectives, with the former providing a comprehensive discussion and the latter examining the data in relation to different degrees of furcation involvement. Moreover, another study derived its data from three distinct papers: Colombo et al.; Gonçalves et al.; Rodrigues et al. (60–62), two of which primarily concentrated on microbiological data, while Gonçalves et al. (61) addressed clinical data. Nevertheless, the research by Rodrigues et al. (62) engaged in a subgroup analysis based on the smoking status of the cohort.

When stratified by smoking status, in the medium term, both smokers and non-smokers showed comparable gains when tetracycline acted as an adjunct to SRP, indicating that smoking does not blunt the initial adjunctive effect. However, over the long term, only non-smokers maintained significant PPD reduction, whereas smokers lost this additional benefit. Attachment level gains tended to persist in both groups in the short term. The long-term improvement of CAL in non-smokers was of higher magnitude than in smokers, but with no statistical significance, likely reflecting fewer studies or greater variability in that subgroup.

A critical finding of our research is the observation that smokers demonstrate less favorable responses to periodontal therapy compared to non-smokers. Taken together, these findings suggest that while smokers can achieve similar short-term improvements with adjunctive tetracycline, smoking may undermine the durability of pocket depth reduction. The mechanisms underlying this compromised healing response in smokers are multifactorial, involving impaired vascular function, altered immune responses, and potentially increased prevalence of resistant periodontopathogens (63).

Despite the generally compromised outcomes in smokers, our findings suggest that certain tetracycline formulations may partially counteract smoking's negative impact on periodontal healing (103). Locally applied controlled-release tetracycline fiber has demonstrated particular promise in this regard, improving clinical outcomes in smokers beyond what would be expected with mechanical debridement alone (104). Similarly, smokers treated with minocycline microspheres in addition to SRP exhibited pocket depth reductions of 1.19 mm at 9 months, compared to only 0.90 mm with SRP alone (105).

However, it must be acknowledged that the evidence for adjunctive antibiotic therapy specifically in smokers remains incomplete. A previous systematic review concluded that evidence for additional benefits of adjunctive antibiotics in smokers with chronic periodontitis is insufficient and somewhat inconclusive (106). Chambrone et al. pooled seven RCTs in heavy smokers, reporting an additional 0.81 mm PPD reduction and 0.91 mm CAL gain at sites with baseline PPD ≥5 mm, whereas systemic antimicrobials showed no significant benefit in this population (37). A smoking-specific review of local doxycycline reported a 1.10 mm CAL gain at six months from two trials (38). These findings highlight that smokers, who typically respond poorly to conventional therapy, can still achieve substantial periodontal improvements with local antimicrobials, often outperforming systemic regimens. However, the need for further well-designed studies specifically targeting this patient population should be highlighted.

The meta-analysis provides evidence that tetracycline, doxycycline, and minocycline, when used as local adjuncts in the treatment of chronic periodontitis, significantly enhance key clinical outcomes, notably PPD and CAL, though efficacy varies by drug and formulation. These findings align with the established antimicrobial and host-modulatory properties of tetracycline-class antibiotics, which are known to combat periodontal pathogens and reduce inflammation. Tetracycline fibers provided robust and sustained benefits in PPD and CAL. Doxycycline products (Atridox, Ligosan) showed notable medium-term improvements but were less consistent over the long term. Minocycline's broader efficacy may be attributed to its higher lipid solubility and enhanced tissue penetration, potentially allowing greater interaction with periodontal tissues. Additionally, minocycline's ability to inhibit matrix metalloproteinases, enzymes involved in tissue destruction in periodontitis, may contribute to its superior performance across multiple clinical parameters. Moreover, the absence of data on outcomes such as BOP for doxycycline and tetracycline formulations, and PI and GI for tetracycline fibers, underscores the need for further research to fully evaluate these adjuncts' comprehensive impact.

However, the presence of moderate to high heterogeneity in some analyses (e.g., I² = 85.13% for minocycline in medium-term CAL, I² = 73.39% for minocycline in medium-term PPD) indicates variability in study designs, patient populations, or treatment protocols, which necessitates cautious interpretation. For instance, differences in drug delivery methods, dosages, or patient characteristics could influence outcomes. The lower heterogeneity in some doxycycline analyses (e.g., I² = 0.00% for long-term CAL) suggests more consistent results, possibly due to standardized formulations like Atridox or Ligosan. The number of studies and participants also varied significantly, with minocycline supported by a larger evidence base (e.g., 27 studies for medium-term PPD with 1,386 controls and 1,426 test participants) compared to tetracycline (8 studies for medium-term PPD) and doxycycline (7 studies for medium-term PPD). This disparity may contribute to the robustness of minocycline's findings but also highlights the need for more extensive research on tetracycline and doxycycline, particularly for outcomes like BOP, PI, and GI, where data were limited or absent.

A recent randomized trial directly comparing tetracycline collagen fibers to 2% minocycline HCl gel showed both to produce equivalent PPD reductions and CAL gains by 3 months (107). This finding is consistent with previous analyses suggesting that the key factor is the presence of local antibiotics rather than the specific formulation. In practice, tetracycline fibers (e.g., Actisite), resorbable collagen fibers, 2% tetracycline ointment, 2% minocycline microspheres, and 10% doxycycline gel all achieve similarly modest adjunctive benefits. No consistent head-to-head study has demonstrated a clinically significant difference among these agents. Thus, while pharmacokinetics (release profile, substantivity) differ by product, our data and the literature suggest that they deliver comparable adjunctive efficacy (39). Formulation choice may therefore be guided by handling characteristics or patient preference rather than expected therapeutic advantage.

Prior systematic reviews have also compared various tetracycline formulations, but among other antimicrobial agents (e.g., chlorohexidine and metronidazole). Herrera et al. provided detailed effects on PPD reduction: Actisite fibers (0.729 mm; n = 7), Atridox doxycycline (0.800 mm; n = 2), and Arestin minocycline microspheres (0.279 mm; n = 6) (35). While a comparatively higher number of RCTs were analyzed for Atridox (n = 4) and Arestin (n = 17) in our study, we identified only five eligible Actisite trials. Methodological differences in study selection criteria likely explain the discrepancy (e.g., inclusion of systemic antibiotics or duplicate publications). Nadig and Shah included one doxycycline, six minocycline, and three tetracycline-fiber trials but did not report agent-specific WMDs, limiting head-to-head comparison (36). Matesanz-Pérez et al. noted that tetracycline fibers (0.850 mm; n = 1) produced greater pocket reduction than sustained-release doxycycline (0.562 mm; n = 1) or minocycline (0.500 mm; n = 1) in medium-term follow-up, though only one study has been analyzed (34). Our review, which incorporated more trials for each agent and extended follow-ups, yielded a PPD reduction of 0.740 mm in 6–9 months (n = 8) and 0.427 mm in 12 + months (n = 3) for tetracycline fibers. Our doxycycline PPD outcomes (6–9 months; n = 7, WMD = 0.454 mm, 12 + months; n = 7, WMD = 0.143 mm) were similar to the 0.562 mm (n = 1) and 0.100 mm (n = 1) reported by Matesanz-Pérez. These comparisons underscore that all tetracycline formulations improve clinical outcomes, but fiber- or gel-based vehicles may confer marginally greater pocket closures when examined over extended periods.

Nonetheless, clinicians should consider delivery system characteristics, as these may influence long-term outcomes, though the overall adjunctive effect remains modest. Our results highlight that the choice of local antibiotic should consider the specific delivery system's longevity and release profile. Clinicians might expect somewhat greater long-term pocket closure with a prolonged-release minocycline ointment or doxycycline systems, while acknowledging that any adjunctive drug effect is modest.

Systematic reviews in this field vary widely in scope, affecting the robustness and comparability of their findings. Our meta-analysis included 42 randomized controlled trials (RCTs), emphasizing longer follow-ups (6–9 and ≥12 months) and a broad selection of tetracycline fibers, minocycline, and doxycycline products. By comparison, Nadig and Shah focused solely on tetracycline-based local delivery, but only considered 10 RCTs, in contrast to 52 studies analyzed in ours, which may have led to overestimated effect sizes due to the limited evidence base (36). Herrera et al. reviewed 50 investigations reported in 59 papers, encompassing various local antimicrobials (35), while Matesanz-Pérez et al. analyzed 52 studies across 56 papers with a similar focus (34). Although these larger reviews offer comprehensive overviews, they also introduce greater heterogeneity in study designs, follow-up durations, and formulations.

Hanes and Purvis (108); Bonito et al. (33) reported mean pocket reductions on the order of 0.3–0.6 mm with local antibiotics. In contrast, Nadig and Shah's meta-analysis of 10 trials (588 sites) found ∼1.2 mm PPD reduction and ∼1.0 mm CAL gain with adjunctive tetracycline vs. control (36), likely driven by their use of standardized mean differences and inclusion of unpublished data. Our WMD for PPD reduction (0.516 mm) and CAL gain (0.336 mm) in 6–9 months and 0.371 mm and 0.310 mm in 12 + months respectively, more closely align with broader reviews. For instance, Herrera et al. reported a WMD of 0.365 mm for PPD reduction at 6–9 months and 0.190 mm at 12–60 months, with CAL gains of 0.263 mm and 0.090 mm, respectively (35). Matesanz-Pérez et al. found a WMD of 0.407 mm for PPD reduction and 0.310 mm for CAL gain (34), closely matching Herrera et al. for shorter-term outcomes. These variations in WMD likely reflect differences in follow-up duration, statistical methods, and specific antimicrobial agents evaluated.

To address comparative efficacy between systemic and local antimicrobials, Teughels et al. reported a WMD of 0.485 mm for PPD reduction at ≥6 months with systemic antimicrobials (109), roughly similar to the 0.365–0.407 mm reductions documented for local antimicrobials by Herrera et al. and Matesanz-Pérez et al. (34, 35). However, systemic regimens carry higher risks of adverse events and antibiotic resistance, whereas local delivery targets specific periodontal sites and avoids systemic side effects, reducing the emergence of resistant flora. Though Herrera et al. (13) found that adjunctive systemic antibiotics (e.g., amoxicillin/metronidazole) improve PPD and CAL vs. SRP alone, estimates vary by drug and patient population. Our effect sizes for local tetracycline (∼0.5 mm PPD reduction and ∼0.3 mm CAL gain) are comparable to systemic regimens but without systemic side effects. Prior meta-analyses consistently show that tetracycline-class local adjuncts provide moderate yet significant PPD and CAL improvements, and our long-term results fall squarely within those ranges. The greater magnitude of benefit seen with local delivery in smokers further supports favoring local over systemic antimicrobials in this population.

The significant improvements in PPD, CAL, BOP, PI, and GI suggest that local tetracycline-class antimicrobials are valuable adjuncts to SRP, particularly for patients with persistent or recurrent periodontal pockets. The similar medium-term benefits in smokers and non-smokers indicate that these therapies can be broadly applied, enhancing outcomes across diverse patient populations. However, the potential for reduced long-term PPD benefits in smokers highlights the need for tailored treatment plans. Clinicians should consider more frequent monitoring and maintenance visits for smokers, alongside smoking cessation programs to optimize long-term outcomes.

The choice of antimicrobial product may also influence results. Tetracycline fibers (Actisite) and minocycline ointment demonstrated robust and sustained effects, suggesting they may be preferred for patients requiring long-term management. Doxycycline products, while effective in the medium term, showed less consistent long-term PPD benefits, which clinicians should weigh when selecting treatments. The minimal adverse effects reported enhance the appeal of local antimicrobials, as they offer a safe adjunctive option compared to systemic therapies. However, practical considerations, such as the time required for application and patient adherence to post-treatment instructions (e.g., avoiding certain foods), should be addressed to ensure efficacy.

For smoking patients specifically, locally delivered tetracyclines may partially overcome the diminished response to conventional therapy typically observed in this population. While not completely negating smoking's detrimental effects, these agents can help narrow the therapeutic gap between smokers and non-smokers.

Local antimicrobial delivery systems can reduce the need for surgical interventions while maintaining clinical outcomes and potentially offering cost savings and shorter treatment times compared with conventional mechanical therapy (110). Heasman et al. highlighted that systemic antimicrobials can also be cost-effective (111); however, these analyses often overlook the potential impact of increased bacterial resistance on patient management.

While locally delivered tetracycline-class antimicrobials are effective, their use must be considered in the context of both clinical and economic factors. Their benefits may be most pronounced in high-risk groups, such as smokers, patients with recurrent disease, or those with furcation lesions, where modest clinical gains could prevent more extensive and costly interventions. However, locally delivered agents generally involve higher costs relative to their short-term clinical gains, particularly when applied as sustained-release formulations. Evaluating their efficacy is further complicated by the need to consider product costs—including the agent, number of teeth treated, number of applications, wastage, clinician time, and patient visits—against the expected benefit. Cost analyses indicate that treating a single tooth at a dedicated visit represents the most expensive scenario for all local delivery systems, whereas treating multiple teeth offers substantial savings (112). Existing studies are also limited by short follow-up periods and rarely incorporate broader patient-centered outcomes, such as tooth retention, function, aesthetics, or quality of life. These limitations highlight the need for future research that integrates clinical outcomes with comprehensive economic assessments, including evaluations of effectiveness in furcation areas, to support evidence-based decision-making in periodontal therapy.

Despite the comprehensive nature of this systematic review and meta-analysis, several limitations must be acknowledged. First, the included studies exhibited substantial heterogeneity, with I² values as high as 74.43% for PPD in long-term studies. This variability may be attributed to differences in study design (split-mouth vs. parallel), patient populations (initial vs. maintenance), assessment methods (full-mouth vs. partial-mouth), and antimicrobial formulations, which could affect the precision of effect estimates. Second, most studies were assessed as having a high risk of bias due to inadequate blinding, randomization, or reporting of methodological details. Only three studies were rated as having a moderate risk of bias (21, 76, 89), underscoring the need for higher-quality trials in this area.

Third, Egger's test indicated significant publication bias for PPD and CAL, with adjusted effect sizes slightly different than the unadjusted estimates. The exclusion of grey literature and restriction to English-language publications likely amplified publication bias by overrepresenting positive or significant findings and omitting null or negative results often found in unpublished or non-English sources. Consequently, our effect estimates may be inflated, and their generalizability to non-English–speaking populations limited.

Fourth, seventeen studies did not report the smoking status of their cohorts, and among those that did, definitions of smoking varied (e.g., current vs. former smokers, smoking intensity). This limited the number of studies available for subgroup analyses and introduced potential inconsistency in the results. Fifth, some subgroup analyses were based on a small number of studies (e.g., four studies for smokers' long-term PPD), potentially reducing statistical power to detect differences.

Moreover, our search strategy, relying only on keywords, “SRP” and “chronic periodontitis,” might have overlooked recent studies using the updated classification, possibly omitting valuable contemporary evidence. As clinical practice shifts toward the 2017 framework, previous terminology may require clinicians to reinterpret results within the current diagnostic context, possibly reducing their direct relevance to modern practice; conversely, because relatively few studies to date have employed the current terminology, the research community is only gradually adapting, underscoring the need for additional investigations using the updated categorization.

Finally, the focus on clinical parameters omitted patient-reported outcomes or quality of life measures, which are increasingly important for a holistic evaluation of treatment effectiveness. The cost-effectiveness of adjunctive local tetracycline therapy was also not evaluated in the included studies, and given that these agents can be expensive and yield only modest clinical gains, it is essential to determine whether the additional expense is justified in routine practice, especially in resource-limited settings. These limitations suggest that while the findings provide valuable insights, they should be interpreted with caution, and further research is needed to confirm and expand upon these results.

This study provides robust evidence that tetracycline, doxycycline, and minocycline, when used as local adjuncts in the treatment of chronic periodontitis, significantly enhance key clinical outcomes, notably PPD and CAL, with benefits observed in both smokers and non-smokers and with split-mouth studies showing the largest effect sizes. Treatment phase did not moderate these benefits meaningfully, underscoring the broad applicability of these agents in both initial and maintenance periodontal therapy. Minocycline ointment emerged as the most consistent and sustained adjunct across all evaluated clinical parameters. While medium-term benefits are comparable, long-term PPD reduction may be reduced in smokers, likely due to impaired healing. Clinicians should integrate these adjuncts into treatment protocols, especially for high-risk patients, and tailor strategies for smokers with enhanced monitoring and smoking cessation support.

Future research should prioritize high-quality randomized controlled trials with larger sample sizes, longer follow-up periods, and standardized smoking status definitions. Mechanistic studies investigating smoking's impact on treatment efficacy and comparative studies evaluating different local agents in smokers are needed. Cost-effectiveness analyses and exploration of combined therapies, such as with host modulation or laser therapy, could further guide clinical practice and improve outcomes for periodontitis patients.

Furthermore, while the current review provides comprehensive insights into the efficacy of tetracycline-class local antimicrobials in the treatment of chronic periodontitis, there remains a notable gap in the literature regarding their specific impact on furcation lesions. Furcation lesions, which occur in multi-rooted teeth and represent a significant challenge in periodontal therapy, have been less frequently studied in the context of adjunct local antimicrobial use. Given the unique anatomical and clinical challenges presented by furcation involvement, future research should prioritize conducting dedicated studies to evaluate the effectiveness of these adjuncts in managing Grade II/III furcation areas. Such investigations could provide valuable data on whether sustained-release tetracyclines can overcome the unique anatomical challenges of furcation areas and translate into improved long-term tooth prognosis in these high-risk sites.