Eight randomized clinical trials were included after application of eligibility criteria. From the 103 studies acquired, 43 were excluded because they were older than 10 years. Summary of studies and methodology were included in (Table 2, Fig. 1).

Four trials [2, 18, 19, 25] (50%) are RCTs and four (50%) of them are ex-vivo clinical trials [21,22,23] where volunteers utilized intraoral appliances to simulate clinical conditions. Randomized clinical trials included a total of 202 patients. [2, 18, 19, 25] Two of them treated 1074 teeth which is a huge sample size [2, 18]. The ex-vivo studies involved 62 volunteers in 4 ex-vivo studies [21,22,23].

Six out of Eight (75%) studies utilized CO2 laser in remineralization or prevention of early enamel caries [2, 18,19,20, 22, 23]. One of the remaining two studies used Erbium doped Yttrium Aluminum Garnet (Er:YAG) laser. The other compared three different laser systems; Er:YAG, neodymium-doped yttrium aluminium garnet laser (Nd:YAG) and CO2 laser [21, 23]. All the included studies used non-contact mode except Correa-Afonso et al. [23] utilized Nd:YAG laser in contact mode.

Four studies of eight (50%) used output power of less than 1 Watt following the Low Level Laser Therapy (LLLT) protocol [2, 18, 19, 23]. Other studies [20, 23] used output power ranging from 1–5 W, whereas one study used medium power of 2 W [22]. Regarding duration of laser application, majority of the included studies used short application time from 9–30 s (Table 2). On contrary, Rechmann et al. [25] used 1.5–2 min application time. The longest application time was reported in Raghis et al. [19] study in which 10 min application was utilized.

Five studies (62.5%) tested individual laser irradiation [18, 19, 21, 23, 25]. There are some studies followed combination treatment of laser application and fluoride containing remineralizing agent. Other studies investigated if there were synergetic effect of laser application on the remineralization potential of fluoride agents.

Four of RCTs (40%) revealed a significant positive effect of using laser irradiation to prevent caries occurrence or progression in a different clinical situation such as newly erupted first permanent molars, fissures, around restorations and orthodontic brackets. The clinical follow-up periods ranged from 3 to 18 months [2, 18, 19, 25].

Two of ex-vivo studies [21,22,23] utilized CO2 laser for prevention of enamel demineralization reported a synergetic effect between CO2 laser irradiation and the use of fluoride product. Takate et al. [20] found significant higher inhibition of mineral loss in enamel slabs when treated individually or in a combination of low power CO2 laser and 1.23% acidulated phosphate fluoride (APF) solution. The application of 1.23% APF solution after low power CO2 laser treatment showed maximum inhibition of mineral loss. Besides, compared to the control, the CO2 laser irradiation with a specific set of laser parameters (0.3 J/cm2 /5 μs/226 Hz) either alone or in combination with a fluoride gel (AmF/NaF) significantly decreased enamel mineral loss (Table 2) [20].

Gabriel et al. [22] cross sectional microhardness test reported that individual CO2 laser application reduced enamel demineralization, and no additional benefits to the combined CO2 laser and fluoride treatment. (Table 2).

Regarding the ex-vivo studies which utilized Er:YAG and Nd:YAG laser, Colucci et al.[21] reported that Er:YAG laser might control the progression of carious lesions around restorations margins. However, Er:YAG laser irradiation was not effective in preventing caries formation adjacent to restorations. Afonso et al. [23] revealed that Er:YAG laser irradiation did not increase enamel resistance to demineralization in pits and fissures. Nevertheless, this study found that Nd:YAG and CO2 laser were effective in increasing enamel acid resistance (Table 2) [23].

Three clinical studies were included in the first meta-analysis. Besides, three in situ/ex-vivo studies were included in the second meta-analysis. The outcome of meta-analysis on the effect of LLLT with CO2 laser on incidence of WSLs was presented in Fig. 2. The repeated study names in the figure exhibited different cohort studies and different follow up periods within the same study. The overall standardized mean difference was 0.21 [ 95% confidence interval (CI): 0.15–0.30, p < 0.00001]. This indicates that the incidence of new WSLs in patients who received low power CO2 laser treatment was highly significantly lower than placebo groups. The heterogeneity was considerable (I2 = 71%) (Fig. 3).

Forest plot of meta-analysis on incidence of enamel carious lesions in patients received Low Level Laser Therapy (0.5–2 W) utilizing CO2 laser compared to placebo. * Repeated study names are different subgroups (incidence of new lesions in successive follow periods within the study; 3 months to 18 months)

The results of meta-analysis on the effect of CO2 laser irradiation on enamel microhardness were illustrated in Fig. 4. The laser power ranged from 0.4–5 W. The overall standardized mean difference was 49.55 [ 95% confidence interval (CI): 37.74, 61.37, p < 0.00001]. This indicates that microhardness of enamel receiving CO2 laser irradiation is highly significantly lower than control untreated enamel. The heterogeneity was substantial (I2 = 48%).

Forest plot of meta-analysis comparing microhardness of enamel received CO2 laser treatment compared to control untreated enamel

RevMan 5.4 windows version (RevMan 5.4, The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) was used to obtain a risk of bias summary and graph (Figs. 5a and 2b). The assessment included the following domains: selection bias (randomization, allocation concealment, unit of randomization issues), performance bias (blinding of participants, operators, examiners), detection bias (blinding of outcome assessment), attrition bias (loss to follow-up and missing values or participants), reporting bias (unclear withdrawals or absence of non-significant reported outcomes) and other bias which included the authorship of the sponsor in data reporting or in outcome data management and analysis. Bias was assessed as a high, low, or unclear judgment.

a Risk of bias graph. b Risk of bias summary