The present review evaluated the current scientific evidence to determine whether, in diabetic patients with periodontitis, NSPT with or without antimicrobial adjuvant therapy, was effective in significantly reducing HbA1c over at least six consecutive months. Results from the ten selected studies demonstrated an overall modest, though non-statistically significant effect of NSPT on glycaemic control, marking a -0.11% (95% CI -0.31 to 0.09) reduction in HbA1c compared to that of the non-treatment group, at 6 months. Our findings are consistent with those of previous meta-analysis (20, 46, 47). For instance, Simpson et al. (2015) in their meta-analysis based on five studies (464 participants who received NSPT and 362 controls) also reported a low and non-significant reduction in HbA1c of -0.02% (95% CI -0.2to 0.16; P = 0.84)(20). Wang et al.(2014) in their meta-analysis including four studies, found a MD in HbA1c of 0.30% (95% CI: -0.69 to 0.09; P = 0.13) between subgingival debridement (n = 413) and non-treatment (n = 341) at 6 months (47). Furthermore, Corbella et al. (2013) discovered a -0.31% (95% CI: -0.74 to 0.11; P = 0.15) MD reduction in three studies with 147 and 88 participants in the intervention (NSPT) and control groups, respectively (46).The modest and not significant HbA1c reduction at 6 months post-treatment could indicate a transient nature of the NSPT effect. Interestingly, D'Aiuto et al. (2018) by implementing intensive periodontal therapy (including periodontal surgery) demonstrated a long-lasting reduction of HbA1c over 12 months (22). One possible explanation for the absence of significant NSPT effectiveness is that other parameters which may influence patients' long-term glycaemic control, such as the control of hypertension and dyslipidemia, weight loss, sedentary lifestyle, and dietary measures, were not sufficiently considered in the clinical trials included in the present review (48).
To the best of our knowledge, with 1128 participants, our review had the largest number of participants with a comparison between NSPT and no treatment or non-active treatment in terms of 6 month HbA1c. Moreover, the HbA1c assessment at 9 months is particular to this review. Simpson et al.(2015) compared different therapeutic modalities with NSPT in the control group while our review included essentially no treatment in the control group, oral hygiene instructions alone, and/or supragingival scaling (20). Of the clinical trials collected, the Engebretson et al.(2013) trial with 514 participants accounted for 30% of the estimated final effect size. Remarks on this study have been shared in previous reviews following comments by Borgnakke et al. in 2014(20, 47). According to Borgnakke, in the study by Engebretson et al. (2013), glycaemic control was near optimal for 60% of the included patients, which may have played a role in the final assessment of the HbA1c level after intervention(49). The data on periodontal parameters, while highlighting a persistence of a certain level of inflammation (BOP at baseline and at 6 months from 60–41% in the intervention group), do not necessarily reflect a failure in the intervention’s application (20).
The strengths of this review include the absence of language restrictions in the selection of studies, the robust statistical analyses, and the low heterogeneity in the assessment of the primary endpoint, as well as the risk of bias assessment with the new Cochrane tool. Accordingly, a strong point of this review was that 40% of studies had an overall low risk of bias [Chen et al.(2012); Engebretson, et al.(2013); Koromantzos et al. (2011); Zhang et al. (2013)], 20% had some concerns about the risk of bias [Botero et al.(2013); Kaur et al.(2015)] and 40% had a high risk of bias [Katagiri et al. (2009); Mizuno et al. (2017); Stewart et al. (2001); Li et al. (2011)]. The high risk of bias mainly emerged from a deficiency in the randomization process, where some shortcomings in evidence of concealment or randomness in the randomization sequence appeared. In contrast to the previous assessment tool, the lack of blinding of reviewers here did not interfere with the assessment of the primary endpoint and therefore was not in itself a source of bias.
Another strong point of this review is the low level of heterogeneity for the primary endpoint (I2 = 25%, p = 0.22). This could be related to the overall characteristics of participants in the included studies. Participants were T2DM patients (except for the study of Botero et al.(2013)in which patients with T1DM were included ), with a diagnosis of at least 2 months old and HbA1c levels ranging between 7 to 8% (except for two studies, Stewart et al. (2001) and Kaur et al. (2015) who had a baseline HbA1c > 8%). The glycaemic equilibrium was acceptable for almost all included studies, due to the chronic comorbidities present at the inclusion of their participants (microvascular or macrovascular complications of diabetes, reduced life expectancy) (35, 40, 41, 47). Only the study by Mizuno et al. (2017) had an imbalance in the sex ratio in favor of men (42).Conversely the periodontitis definition was inconsistent across studies and residual levels of inflammation remained after intervention (BOP > 20%) (18, 41).
In the present study, NSPT improved the periodontal status of diabetic patients for over 6 months, in line with results from previous studies (50). However, the use of Armitage classification in all studies did not allow the assessment of the decrease in severity of the periodontal damage in terms of a decrease in the number of pocket depths > 6 mm or the gain of attachment level at the most affected site (24). As a result, the gains reported clinically are difficult to assess, as they are less than a millimeter in length (decrease in PD of -0.38 mm, gain in CAL of 0.46 mm).Furthermore, while Chen et al. (2012) reported an improvement in CRP and TNF- levels following NSPT, there was insufficient data to draw any conclusions (37, 51).
However, the study results have some limitations that should be noted. Although the number of participants included is large compared with previous reviews, only one included study has been conducted since the Cochrane meta-analysis in 2015 (20, 42). The duration of follow-up was limited to 6 months for eight studies, and two had at least 9 months of follow-up but with a high or some concerns level of bias. Moreover, sample sizes of some studies were too small to identify a mean difference in HbA1c of the order of 0.2% between the intervention and control groups, and the reporting of some outcomes was partial or difficult to use, requiring mathematical calculations [Katagiri et al. (2009); Mizuno et al. (2017);Engebreston et al. (2013)]. Assessment of the effect of NSPT on T1DM was not possible because only one study reported T1DM among the included diabetic patients (38). Furthermore, 7 of the 10 studies had a baseline HbA1c of less than 8%. The inclusion of patients with significant imbalance would allow for a better assessment of the effect of NSPT, given that NSPT does not have a significant effect on improving HbA1c in patients with good glycaemic control (52). Conversely, considering scaling as a non-active treatment here and its analysis after association with the control group of the 3-arm studies [Li et al. (2011); Botero et al. (2013);Koromantzos et al. (2011)] raises some concern, as we know that scaling has an impact on periodontal parameters (PI, BOP and periodontal pockets with depths of less than or equal to 4 mm) (25, 53).
The findings of the present meta-analysis of the literature contribute to increasing the level of evidence on the NSPT effect on glycaemic control in diabetic patients and highlight the need for further studies. Firstly, clinical trials with longer monitoring periods (> 6 months) are necessary to assess the effect of NSPT on glycaemic control. In this review, two studies reported a follow-up of over 6 months, and important elements of diabetic follow-up such as the occurrence of complications, systemic inflammation, and oxidative stress were only minimally addressed in the studies. The study NSPT associated with adjunctive therapies in clinical trials with large sample sizes in diabetic patients with a baseline HbA1c level of more than 8%will be able to clarify their effect on glycaemic control. Interventions encompassing quality of life and hygienic-dietary measures would be of interest to provide better understanding and monitoring for diabetic patients considering the complex mechanisms of this disease. The literature remains poorly supplied in the study of periodontal therapeutics and their effects on other forms of diabetes, including T1DM.