Search and study selection

A total of 1043 studies were identified initially, only 8 cross-sectional study met the criteria and were included in our study [13,14,15,16,17,18,19,20] , no RCT, cohort study or case-control study was included (Fig. 1).

Fig.1Study quality assessment

Two tools were used to assess the quality of all studies: the AHRQ checklist for cross-sectional study and a quality tool for TDM studies. According to the AHRQ checklist, 7 studies were determined to be of moderate quality, while 1 study was deemed to be of high quality. According to the quality tool for TDM studies, 3 studies scored 8.35, 2 studies scored 7.35, and 3 studies scored 6.85 (Table 1).

Table 1 Characteristics of the included studiesStudy characteristics

8 Studies comprising 1151 children with epilepsy were included in this study. Studies were published from 2001 to 2021, two of them were retrospective study and 6 were retrospective study. Children in 7 studies were from China and 1 study were from Japan. 2 studies further grouped children into younger children (aged 2 to < 12 years) and older children (aged 12 to 18 years). Blood samples in all studies were collected after reaching the maintenance dose of LTG, 7 studies determined trough concentration, while 1 study determined peak concentration. Children in 3 studies received LTG monotherapy, 2 studies received LTG concomitant with VPA, and 3 studies received LTG concomitant with other ASMs. Details are shown in Table1.

Estimation methods of reference range

2 studies only estimated the lower limit of reference range, and 6 studies estimated both upper limit and lower limit of reference range (Table 2).

Table 2 Calculating methods of reference rangeLTG monotherapy

3 studies estimated reference range for LTG monotherapy. 2 of these studies estimated both the upper and lower limit concentrations, while1 study only estimated lower limit concentration.

For lower limit, 3 studies used the same method. Wang [16] and Wang et al. [20] enrolled 110 and 339 epileptic children aged 2–18 years respectively, and both divided children into two groups: (1) younger children (aged 2 to < 12 years) and (2) older children (aged 12 to 18 years). Zhao et al. [17] enrolled 200 epileptic children aged 0–18 years. In these 3 studies, children who experienced a reduction in seizure frequency (Seizure Reduction Rate, RR) of ≥ 50% compared to their baseline were classified as responders, while those who did not were classified as non-responders. The researchers established the lower limit concentration by performing a receiver operating characteristic curve (ROC), which based on the clinical response and LTG concentration.

For upper limit, both Wang [16] and Wang et al. [20] developed concentration-effect curves based on RR and concentration of LTG to calculate a threshold concentration beyond which therapeutic effect no longer increases significantly. This threshold concentration was confirmed as the upper limit. Zhao  [17] did not estimate upper limit.

LTG concomitant with VPA

2 studies estimated reference range for LTG concomitant with VPA [15, 19]. 1 of these studies estimated both the upper and lower limit concentrations [19], while 1 study only estimated lower limit concentration [15].

For lower limit, He et al. [15]enrolled 72 epileptic children treated with LTG concomitant with VPA. Children who experienced a RR of ≥ 50% compared to their baseline were classified as responders, while those who did not were classified as non-responders. In this study, children were divided into two groups based on a preset LTG level (≥ 2 μg/L or < 2 μg/L) and their clinical response was compared. The results showed that the non-responders in the < 2 μg/L group was obviously higher compared to the ≥ 2 μg/L group (χ2 = 5.3731, P = 0.02), and all patients in the ≥ 2 μg/L group were responders. Therefore, this study recommended the preset LTG level as a reference range for LTG in combination with VPA. Iwasaki et al. [19] enrolled 32 epileptic children aged 2.9–14.1 years that treated with LTG concomitant with VPA. Children who experienced a RR of 50% compared with baseline were defined as effective cases, and the others were defined as ineffective cases. This study discovered a significant difference in LTG blood concentrations between effective and ineffective cases. The concentration range in effective cases did not overlap with that of ineffective cases. The study confirmed the lower limit concentration as the average minus the standard deviation of concentrations in effective cases.

For upper limit, Iwasaki et al. [18] confirmed it as the average plus the standard deviation of concentrations in effective cases.

LTG concomitant with other ASMs

Bao et al. [13] included 44 epileptic children aged 3–14 years that treated with LTG concomitant with other ASMs. In this study, children who experienced a RR of ≥ 50% compared to their baseline were defined as effective cases, those who did not were classified as non-responders. The reference range was estimated by calculating the 95% confidence interval of concentration in all effective cases.

Li and Huang [14] included 111 epileptic children aged 3–18 years that treated with LTG concomitant with other ASMs. This study estimated the 95% confidence interval of concentration in all cases, and this 95% confidence interval was considered as the reference range.

Zhao et al. [18] included 266 epileptic children aged 0–18 years that treated with LTG concomitant with other ASMs. This study estimated the percentile interval (P2.5 to P97.5) of concentration in all cases, and this interval was considered as the reference range.

Data type

6 studies estimated the reference range using efficacy data [13, 15,16,17, 19, 20], while 2 studies estimated the upper limit and lower limit of reference range only using concentration data [14, 18]. No studies used safety data.

LTG monotherapy

For LTG monotherapy, 3 studies estimated the upper and lower limit of reference range based on both concentration data and efficacy data. 2 of these studies [16, 20]assessed efficacy of LTG based on the average monthly number of epileptic seizures over 3 months, while time frame of another study was unclear [17] (Table 3).

Table 3 Methods of efficacy evaluation for studies that used efficacy dataLTG concomitant with VPA

For LTG concomitant with VPA, 2 studies estimated the upper and lower limit of reference range based on both concentration data and efficacy data. One of these studies assessed efficacy of LTG based on the average monthly number of epileptic seizures over 28 days [19], however, time frame of another study was unclear [15] (Table 3).

LTG concomitant with other ASMs

For LTG concomitant with other ASMs, 2 studies estimated the upper and lower limit of reference range only using concentration data. 1 study used both concentration data and efficacy data, while time frame that assessed efficacy of LTG was unclear [13] (Table 3).

Results of reference rangeLTG monotherapy

Reference range recommended for LTG monotherapy in different studies were basically consistent. For younger children (aged 2 to < 12 years), Wang [16] recommended 3.99–8.97 mg/L and Wang ML recommended 3.29–9.08 mg/L. For older children (aged 12 to 18 years), Wang HX recommended 2.67–8.56 mg/L and Wang et al. [20] recommended 2.06–8.43 mg/L, which were lower than the recommendation for younger children. Zhao et al. [17] recommended ≥ 2.64 mg/L for children (Table 4).

Table 4 Results of reference rangeLTG concomitant with VPA

Reference range recommended for LTG concomitant with VPA varied across different studies. He et al. [15] recommended ≥ 2 mg/L, Iwasaki et al. [19] recommended 8–11.5 mg/L (Table 4).

LTG concomitant with other ASMs

Reference range recommended for LTG concomitant with other ASMs varied across different studies, they were 1–9.9 mg/L [19], 2.20–16.24 mg/L [18], and 3.09–5.90 mg/L [13] respectively (Table 4).