We conducted and reported the results following the Cochrane Handbook and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [25]. We registered it at PROSPERO (CRD42018112231).
Focused question
Does non-surgical periodontal therapy influence systemic inflammation and metabolic markers in haemodialysis and/or peritoneal dialysis patients with chronic periodontitis?
Search strategy
A comprehensive electronic search of PubMed, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL), China National Knowledge Infrastructure (CNKI), and Chinese Medicine Premier's Wanfang database (WFPD) was searched from their inception to July 2019 for eligible articles. We used the following MeSH terms and free text keywords: population [“renal dialysis” OR “peritoneal dialysis” OR “haemodialysis” OR “kidney failure, chronic” OR “end-stage renal disease” OR “renal insufficiency”] AND intervention [“periodontal therapy” OR “periodontal/dental debridement” OR “non-surgical periodontal debridement” OR “dental prophylaxis/instrumentation” OR “root planing”]. We also reviewed the reference lists of identified studies and pertinent reviews for additional citations. Databases were searched to include papers published in English and Chinese. The detailed search process is illustrated in the supplementary file S1. We did not contact the original authors for further information.
Eligibility criteria
Two reviewers (HY and XXX) independently evaluated the eligible studies that met the following criteria: (1) study design: randomized controlled clinical trial (RCT); (2) the ESRD patients diagnosed with CP; further inclusion were: the patients undergoing PD and/or HD; without other sources of inflammation such as pulpal infections and active caries; no periodontal treatment in past six months; (3) intervention group of NSPT including professional OHI, full-mouth SRP, (including ultrasonic and/or hand supra/subgingival biofilm and calculus removal), SRP plus local or systemic antiseptic therapy without surgical flap procedures; (4) control group of age and gender-matched no periodontal treatment, delayed treatment or only including OHI; (5) and the follow-up times ≥ 4weeks.
The exclusion criteria were reviews, studies without a comparison group, case reports, and conference abstracts.
Screening and data extraction
Two reviewers (HY and XXX) independently performed the study screening process. First, we excluded duplicated publications and scrutinized titles and abstracts to select literature that may meet the inclusion criteria. Then, we read the full text carefully to include eligible studies further. We implemented data extraction from included studies, and then another reviewer (XZL) checked the results for accuracy. The following information was extracted: the name of first author, publication year, sample size of participants, study methods (such as study design, follow-up duration), intervention, conclusions, clinical and biochemical measures including serum inflammatory markers: hs-CRP, IL-6, TNF-a; metabolic markers including nutritional markers (such as the albumin, Alb) and lipid metabolic markers (including total cholesterol, TC; triglycerides, TG; high-density lipoprotein cholesterol, HDL-C; and low-density lipoprotein cholesterol, LDL-C). Any inconsistent results regarding the eligibility of studies and extraction of data occurred between the two reviewers was resolved by a conversation with a third reviewer (XZL).
Quality assessment
Two reviewers (HY and BH) independently performed the quality assessment of the selected studies via the Cochrane Handbook for Systematic Reviews of Interventions for assessing the risk of bias. The tool included six domains: random sequence generation, allocation concealment, blinding of participants, outcome assessors, incomplete outcome data, selective reporting, and other biases. All six included issues were evaluated as low risk, high risk, or unclear. Any divergence between the two investigators was resolved by discussion with a third reviewer (QL).
Data analysis
We (HY and QL) performed data analysis and pooled in statistical meta-analyses using Stata software (version 15.0, StataCorp, College Station, TX). In the meta-analysis, we calculated overall effect size (ES) estimates using the standardized mean difference (SMD) and the upper and lower limits of the 95 % confidence intervals (CI) to assess overall efficacy from all the eligible studies. The heterogeneity was assessed by Q statistic (P< 0.10 indicating significant heterogeneity), and I squared (I2) statistic. An I2 value of more than 50% represented high heterogeneity; thus, the random effect model would be adopted. I2 less than 50% representing low heterogeneity, fixed-effects models were used. Statistical significance was declared if the P-value was <0.05. Publication bias and sensitivity analyses would have been conducted if the included trials were at least ten according to Higgins and Green [26].